Watson John 2006 Mastering The Chess Openings Vol 1 EGambit OCR

Mastering the
Chess Openings
Volume 1
John Watson
First published in the UK by Gambit Publications Ltd 2006
Copyright © John Watson 2006
The right of John Watson to be identified as the author of this work has been asserted in accordance
with the Copyright, Designs and Patents Act 1988.
All rights reserved. This book is sold subject to the condition that it shall not, by way of trade or
otherwise, be lent, re-sold, hired out or otherwise circulated in any form of binding or cover other
than that in which it is published and without a similar condition including this condition being im
posed on the subsequent purchaser.
ISBN-13: 978-1-904600-60-2
ISBN-10: 1-904600-60-3
DISTRIBUTION:
Worldwide (except USA): Central Books Ltd, 99 Wallis Rd, LondonE9 5LN, England.
Tel +44 (0)20 8986 4854 Fax +44 (0)20 8533 5821.E-mail:[email protected]
Gambit Publications Ltd, 99 Wallis Rd, LondonE9 5LN,England.

E-mail: [email protected]
Website (regularly updated): www.gambitbooks.com
Edited by Graham Burgess

Typeset by Petra Nunn
Cover image by Wolff Morrow
Printed in Great Britain by The Cromwell Press, Trowbridge, Wilts.
10 9 8 7 6 5 4 3 2 1
Gambit Publications Ltd

Managing Director: Murray Chandler GM
Chess Director: Dr John Nunn GM
Editorial Director: Graham Burgess FM
German Editor: Petra Nunn WFM
Webmaster: Dr Helen Milligan WFM
Contents
Symbols 6
Dedication 6
Acknowledgements 6
Bibliography 7
Introduction 9
1 The Nature of Chess Openings: Fundamentals 11
2 Opening Ideas and Positional Features 23
3 The Significance of Structure 36
Section 1: Open Games
4 Introduction to 1 e4 and the Open Games 87
5 Giuoco Piano 91
Blockading the Pawn 94
Chasing the Pieces 97
A Technical Approach: 5 d3 98
6 Two Knights Defence 101

The Calm 4 d3 102
White targets f7: 4 lt:Jg5 104
Interposition with the Pawn 105
Interposition with the Bishop 109
Central Play: 4 d4 110
7 Philidor Defence 114

Surrender of the Centre 115
Restraint 116
Counterattack 117
The e5 Strongpoint 118
8 Ruy Lopez 122

Chigorin Defence 126
Modem Chigorin 130
Keres Defence 132
Breyer Defence 133
Zaitsev Variation 137
4 MASTERING THE CHESS OPENINGS
M011er Defence 140

Open Variation 144
Exchange Variation 151
Marshall Attack 158
9 King's Gambit 166

King's Gambit Declined 167
P iece-Play 167
Central Expansion 169
King's Gambit Accepted 169
Section 2: Semi-Open Games
10 Introduction to the Semi-Open Games 173
11 Sicilian Defence 175

Introduction to Systems with 2 . . . d6 178
Dragon Variation 1 81
Classical Dragon 182
Yugoslav Attack 1 85
Najdorf Variation 194
The ilg5 Attack 196
Poisoned Pawn Variation 199
Najdorf Sozin Attack 202
Classical 6 ile2 System 208
English Attack 212
Introduction to Systems with 2 . . . e6 216
Sicilian Four Knights 219
Paulsen System 221
Taimanov Variation 227
Sozin Attack (and the Classical Sicilian) 236
Sozin with Kingside Castling 239
Velimirovic Attack 241
Accelerated Fianchetto 244
Alapin Variation 248
12 Caro-Kann Defence 251

Caro-Kann Two Knights 251
Exchange Variation 254
Panov Attack 256
Advance Variation 261
13 French Defence 270

Tarrasch Variation 271
Recapture with the Pawn 273
Recapture with the Queen 278
Tarrasch with 3 tiJf6
. . . 283
Classical Variation 294
CONTENTS 5
Winawer Variation 298

Positional Lines 299
French Poisoned Pawn 305
14 Pirc Defence 313

Austrian Attack 315
The i.c4 Variation 327
Classical Variation 32 8
Index of Players 332

Index of Openings 335
Symbols
+ check
++ double check
# checkmate
!! brilliant move
good move
!? interesting move
?! dubious move
? bad move
?? blunder
Ch championship
Cht team championship
Wch world championship
Wcht world team championship
Ech European championship
Echt European team championship
Ct Candidates event
IZ interzonal
z zonal
ECC European Clubs Cup
OL olympiad
jr junior event
tt team event
1 -0 the game ends in a win for White
lfi_lh the game ends in a draw
0- 1 the game ends in a win for Black
(n) nth match game
(D) see next diagram
Dedication
To Maura, the Light of My Life
Acknowledgements
Special thanks to Graham Burgess for his patience and help throughout this project.
Bibliography
Periodical Publications
ChessBase Magazine (up to 1 1 2)
New in Chess Magazine; New in Chess
Informator (up to 95); Sahovski Informator
Websites
ChessPublishing; Kosten, A; www.chesspublishing.com
The Week in Chess; Crowther, M.; www.chesscenter.com/twic/twic.html
Jeremysilman.com; Silman, J. ; www.jeremysilman.com
ChessCafe.com; Russell, H.; www.chesscafe.com
CDs and DVDs

Bangiev, A; Philidor Defence; ChessBase 2002
Henrichs, T.; Queen's Gambit Orthodox Defence- Exchange Variation D31/D35-D36;
ChessBase 2004
Opening Encyclopaedia 2005; ChessBase 2005
Petronijevic, Z.; Caro-Kann Panov Attack BJ3-Bl4; ChessBase 2004
Ripperger, R.; Giuoco Piano C50-C54; ChessBase 2004
Ripperger, R.; Isolated Queen Pawn; ChessBase 2003
Books and Articles

Aagaard, J.; Easy Guide to the Panov-Botvinnik Attack; Gambit/Cadogan 1 998
Alburt, L., Dzindzichashvili, R. & Perelshteyn, E.; Chess Openings for Black, Explained;
CIRC 2005
Baburin, A; Winning Pawn Structures; Batsford 1 998
Beliavsky, A & Mikhalchishin, A; The Two Knights Defence; Batsford 2000
Bosch, J.; Secrets of Opening Surprises 1 -3; New in Chess 2003-5
Bucker, S . ; 'Only a Storm in a Teacup?' (Marshall Attack); ChessCafe 2006
Burgess, G. ; The Taimanov Sicilian; Gambit 2000
Burgess, G.; Nunn, J. ; & Emms, J. ; The Mammoth Book of the World's Greatest Chess Games;
Robinson 1 998
Chemin, A & Alburt, L.; Pirc Alert!; CIRC 2001
Collins, S . ; Understanding the Chess Openings; Gambit 2005
Davies, N.; Play I e4 e5!; Everyman 2005
Dearing, E.; Play the Sicilian Dragon; Gambit 2004
de Firmian, N.; Chess Openings the Easy Way; Batsford 2003
Donaldson, J. & Silman, J.; Accelerated Dragons; Cadogan 1 998
Emms, J.; Play the Open Games as Black; Gambit 2000
Emms, J.; Starting Out: The Sicilian; Everyman 2002
Euwe, M. & Kramer, H.; The Middlegame, Books I and II; David McKay 1 964
Fine, R.; The Ideas Behind the Chess Openings [3rd Edition]; McKay 1 989
Flear, G.; Open Ruy Lopez; Everyman 2000
Flear, G. ; The Ruy Lopez Main Line; Everyman 2004
Gallagher, J.; Starting Out: The Pirc/Modern; Everyman 2003
Golubev, M.; The Sicilian Sozin; Gambit 2001

Hansen, Ca.; Improve Your Positional Chess; Gambit 2005
Hillarp Persson, T.; Tiger's Modern; Quality Chess 2005
Jacobs, B . ; Mastering the Opening; Everyman 2001
Jacobs, B . ; Starting Out: The French; Everyman 2003
Johansson, T.; The Fascinating King's Gambit; Trafford 2004
Kallai, G.; Basic Chess Openings [l e4]; Cadogan 1 997
Kindermann, S. & Dirr, E.; Franzosisch Winawer - Band I: 7 'WIg4 0-0; Chessgate 2005
Kindermann, S . ; The Spanish Exchange Variation -A Fischer Favourite; Olms 2005
Kinsman, A. ; Spanish Exchange; Batsford 1 998
Krnic, Z. (ed.); ECO B - 4th Edition; Sahovski Informator 2002
Marovic, D . ; Dynamic Pawn Play in Chess; Gambit 2001
Matanovic, A. (ed.); ECO C - 4th Edition; Sahovski Informator 2004
Nielsen, P.H. & Hansen, Ca.; Sicilian Accelerated Dragon; Batsford 1 998
Nunn, J. ; Burgess, G. ; Emms, J. & Gallagher, J. ; Nunn's Chess Openings; Gambit/Everyman 1 999
Nunn, J.; Grandmaster Chess Move by Move; Gambit 2005
Pachman, L.; Modern Chess Strategy; Dover 1 9 7 1 (based upon 1 963 English Edition)
Panczyk, K. & Ilczuk, J.; Ruy Lopez Exchange; Everyman 2005
Pedersen, S . ; The French: Tarrasch Variation; Gambit 2005
Pinski, J. ; The Two Knights Defence; Everyman 2003
Psakhis, L.; French Defence: Steinitz, Classical and Other Systems; Batsford 2004
Raetsky, A.; Meeting I e4; Everyman 2002
Sammalvuo, T.; The English Attack; Gambit 2004
Silman, J. ; Amateur's Mind; Siles 1 999
Silman, J.; Reassess Your Chess Work Book; Siles 2001
Soltis, A.; Pawn Structure Chess; McKay 1 995
Taimanov, M.; Sicilian: Paulsen; Batsford I 984
Uhlmann, W.; Winning With the French; Batsford 1 995
Watson, J.; Play the French [3rd Edition]; Everyman 2003
Watson, J.; Chess Strategy in Action; Gambit 2003
Watson, J. & Schiller, E.; Survive and Beat Annoying Chess Openings - The Open Games;
Cardoza 2003
Zeller, F.; Sizilianisch im Geiste des !gels; Kania 2000
Introduction
The initial moves of a chess game hold a particular fascination for those who play the game. This is
reflected in the fact that chess-players at all levels devote the greatest part of their study to what are
called 'openings' . Put simply, openings are sequences of early moves; we'll discuss exactly what
qualifies as an opening as we go along. Players normally study the openings that may potentially
appear in their own games. After all, nobody wants to incur a disadvantage before the game warms
up, and every chess-player would like to gain an advantage over his opponent right out of the
blocks.
Thus we find in the chess literature vast numbers of books about particular openings and open
ing systems. More has been written about the initial phase of the game than about any other chess
topic, whether the middlegame, endgame, history, strategy, attack or defence. There are also
encyclopaedias, magazines, CDs, DVDs, videos, and websites devoted solely to opening moves.
We refer to such material in general as 'opening theory' or simply 'theory ' . Within most fundamen
tal openings there are seemingly countless subsystems (called 'variations') and still further divi
sions of material into ' subvariations' . It is not uncommon to see large books devoted exclusively to
variations or even subvariations. Fortunately, openings are usually named, so we can communicate
about them without explicitly having to restate, for example, the first nine moves played by both
sides.
Among these myriad books and products, very few are devoted to explaining the ideas, strate
gies, and interconnections of chess openings taken as a whole. That is, individual theoretical books
concentrate upon a single opening's moves and variations, and most discuss why some of those
moves are good or bad. A fair number of these books will also examine basic strategies underlying
the opening in question, which is important and beneficial. But few give a feeling for the common
threads that underlie opening play or the reasons why opening strategies can differ so radically. In
the book before you (and Volume 2 of this project), I seek to provide a durable standpoint from
which to view the opening phase of the game. Then, regardless of the uncertainties of theory, you
should be able to find your way through many of the problems posed by unfamiliar moves.
As I began work on this book it became obvious that even in two large volumes it wouldn't be
possible to cover every opening, nor even the most significant variations of every opening, and still
achieve the insights that I hoped to convey. On the other hand, I have sought here to provide a start
ing-point for players of all strengths to be able to understand these openings. Regardless of what
anyone says, that simply can't be done without particulars, i.e. investigation of moves, alternatives,
and annotated examples. What's more, those particulars must be comprehensible within some
framework of general chess knowledge. In the end, I decided to begin the book with three chapters
covering fundamental ideas of opening play. The first chapter presents elementary concepts shared
by all openings. In the next two chapters, I incorporate motifs and structures that will inform your
study as you proceed to specifics.
The greater part of the book is devoted to a selection of individual openings (king's pawn open
ings in the case of this book; Volume 2 will focus on queen' s pawn openings). These openings are
examined from the ground up, which is to say that each chapter begins with an explanation of the
very basics of strategy. I shall often show what happens when you play alternatives that are inferior
to the generally approved moves. As the chapter progresses, established variations are explored,
sometimes in considerable detail, in order to establish the ideas and themes that characterize each
opening and to investigate the extent to which they resemble other opening complexes. At the begin
ning of each section I've paid special attention to move-order issues. Students are often perplexed
by move-orders, which frequently determine whether they get the opening position that they' re
aiming for.
Choosing which systems and variations to investigate proved an extremely difficult task. I de
cided to concentrate upon the most 'important' openings, that is, the ones which are and have been
the centre of theory and practice for decades. Obvious examples are the Ruy Lopez, Sicilian De
fence and Queen's Gambit. Within those and other major opening systems, I have selected a lim
ited number of variations that are, I believe, enlightening in strategic terms. I have also examined
some less prominent openings which not only have uniquely interesting properties but also lend
themselves to comparisons with more popular systems. You may find that structures and ideas
from superficially contrasting openings overlap more than you think. Finally, I explore how these
openings and their variations fit into the general contours of a chess game. It is important to under
stand that the games and analysis do not always represent current theory; they are intended to illus
trate underlying properties of the opening.
What are the rewards for studying openings and understanding the ideas associated with them?
Well, it's always nice to gain an early advantage over your opponent, as I mentioned above. But
such study has more valuable and far-reaching effects: it benefits your general chess knowledge in
a way that reading abstract books on strategy can't. The more thorough your investigation into
openings, the better your understanding of the play that occurs after the opening. To begin with,
many characteristics of openings, including typical strategies and tactics, endure throughout the
middlegame, so your deeper understanding of them will translate to your overall success. In addi
tion, the typical pawn-structures established by an opening will persist as we enter into simplified
positions and even endgames.
This book assumes a basic level of playing competence. Nevertheless, those who know the rules,
have played a bit, and are willing to put some effort into their chess study will do well. You need not
have advanced much beyond the initial playing stage to understand the basic ideas presented here.
All of Chapter 1 , most of Chapter 2, and the introductions to the chapters on individual openings
are designed to help in that regard. I have also woven fundamental ideas into the analysis of specific
openings, attempting to begin my presentation at a lower level and then proceed to the more ad
vanced concepts needed for substantial improvement.
After years of exploring the initial phase of the game, I have come to an important and, I think,
encouraging conclusion: every well-established opening is playable. That is not to say that all
openings lead to full equality, nor that all speculative gambits will lend themselves to acceptable
outcomes. But with sufficient study and understanding, any opening system that masters play, even
on a periodic basis, will serve you well enough to get you to the rniddlegame in decent shape. Un
der those circumstances, the result of the game will not be decided by your choice of the first 5 - 1 0
moves, whether against a club opponent o r i n top competition. Players o n all levels have an under
standable tendency to follow the latest fashions, and that can lead to the notion that openings not
currently being played are substandard. It's much more likely that those openings are simply out of
favour or running into difficulties against some esoteric move within a complicated variation.
There are many variations and even whole opening systems that have been declared inferior but
were then taken up again by the world's best players. When in doubt, look up the number of grand
masters who play one 'bad' opening or another. This will encourage you to approach your explora
tions with an open mind.
I hope that this book will reward your careful study and give you a new perspective on openings
and on the game of chess itself.
1 The Nature of Chess Openings:
Fundamentals
The first moves of a chess game can be played moves with names that are in general usage,
in random fashion, or they can be organized so we avoid dealing with such near-irrational se
as to form a coherent strategy. Chess is above quences as 1 a4 e5 2 f3, which do not fall within
all a game of logic and planning, so the player the category of openings as I have defined
who coordinates his moves towards an end will them. There are very few meaningful openings
almost always defeat an opponent whose moves that are unnamed, but I shall touch upon them if
have no purpose or are inconsistent. This book the occasion arises.
concerns itself with initial moves that make Most of this book is divided into major open
sense together and attempts to explain the rea ings which can be identified within four moves
soning underlying those moves. or fewer; for example, the Ruy Lopez ( 1 e4 e5 2
The first order of business will be to clarify tllf3 tllc6 3 i.b5), or the Griinfeld Defence ( 1
the scope of our investigation and to orientate d 4 tllf6 2 c 4 g 6 3 tllc 3 d5), each o f which then
ourselves in the world of openings. Then we subdivides into 'variations' . Named variations
shall look at some rudimentary ideas underpin of openings can be of almost any length; for ex-
ning successful opening play. ample, the Closed Variation of the Sicilian De
fence has just two moves: 1 e4 c5 2 tllc 3; and
What is an Opening? the Exchange Variation of the Ruy Lopez (also
known as the 'Spanish Game' ) consists of the
Generally speaking, an opening is defined by the four moves 1 e4 e5 2 tllf3 tllc 6 3 i.b5 a6 4
introductory moves of a chess game. An open i.xc6. Lasker's Variation of the Queen's Gam
ing begins on move one. The obvious question bit is distinguished by the seven moves 1 d4 d5
that suggests itself is surprisingly difficult to 2 c4 e6 3 tllc 3 tllf6 4 i.g5 i.e7 5 e3 0-0 6 tllf3
answer: how do we decide on what move an h6 7 i.h4 tlle4 (D).
opening ends and the middlegame begins ?
There i s n o general agreement among players
or authors about this; in many cases it turns out
to be a subjective judgement informed by play w
ing experience. In this book I shall define open
ings (and their variations) as sequences of moves
that are specifically named, with the name in
common chess usage and sometimes referring
to a complex of related positions. The advan
tage of using this convention is that we can
know precisely at which move an opening or
variation ends. For instance, the 'English Open
ing' is defined by a single white move: l c4.
The 'Sicilian Defence' consists of l e4 c5. And
the variation called the 'Najdorf Variation of
the Sicilian Defence' is delimited by the moves But some variations stem from other varia
1 e4 c5 2 tllf3 d6 3 d4 cxd4 4 tllxd4 tllf6 5 tllc 3 tions, which can stem from still others, and so
a6. By defining the word 'opening' to designate forth. For example, the Chinese Variation of the
Dragon Sicilian evolves from this move-order: more moves that are of equal worth, so we can
1 e4 c5 (this is the 'Sicilian Defence' ) 2 l2Jf3 d6 not decide in practice or even with hindsight
3 d4 cxd4 4 l2Jxd4 lLif6 5 l2Jc3 g6 (the moves whether one move actually achieves more than
thus far are known as the 'Dragon Variation' ) 6 another. Even if a hypothetical supercomputer
i..e 3 i..g7 7 f3 l2Jc6 8 �d2 (these first eight could solve the position, the end result of either
moves define the 'Yugoslav Attack ' ) 8 . . . 0-0 9 move would usually be the same - for example,
i..c4 (some authors refer to this as the ' i..c4 Yu a draw. Thus a player may have a wide choice
goslav Attack') 9 ... i..d7 1 0 0-0-0 and now with that is more a matter of taste and playing style
10 . . .J:;tbS (D), we have arrived at the 'Chinese than of objective quality. We can contrast this
Variation of the Dragon Sicilian' . situation with another part of the game - the
endgame. In most endgames, particularly those
with just a few pieces on the board, we can es
tablish precisely what the ultimate effect of a
particular move would be. Consequently, very
few moves will be made simply because they
suit someone' s style of play.
You should also note that players can usually
make one or two inaccuracies in the opening
and still not be punished with a lost position.
By contrast, a single mistake in a king and
pawn endgame, for example, may be fatal, and
punishment can come quickly for even a small
endgame inaccuracy. Thus, many reasonable
looking decisions in the endgame are unam
biguously right or wrong and can be demon
If some of this is confusing, you shouldn't strated to be so. In the opening, however, a
worry: it will become clear as we work our way player has more leeway, which means that he
through the book. is able to approach positions more creatively,
In this general scheme the word 'theory' is without needing to calculate variations out to a
used to indicate specific moves that have been win or loss. This in tum allows players of any
previously played or analysed, and are known strength to come up with worthwhile new
by a significant portion of the chess commu opening moves. Openings are also more for
nity, usually via publications or databases. In giving with respect to static features of play:
most but not all cases we can think of theory as the earlier in the game that you take on a bad
representing the end of the opening phase of bishop or pawn weakness, for example, the
the game but not the opening itself. Theory can more likely it is that you can solve the associ
therefore extend far into the game because peo ated problems. Furthermore, there are many
ple all over the world repeatedly play the same opening positions that are chaotic and defy
opening and consistently add to what is known useful generalization.
about it. Theoretical discussions sometimes deal It should not be surprising that the middle
with the 20th move of a variation or even fur game shares features with both the opening and
ther into the game, but most opening theory endgame. Middlegame play tends to include
typically ends on a move in the teens, and the more immediately critical decisions than open
theory of a lesser-known variation may end af ing play and middlegame mistakes are fre
ter only six or seven moves. quently life-threatening. An inaccurate attack
The opening has certain characteristics that or defence can lead to instant defeat and posi
distinguish it from the other parts of the game, tional problems tend to be harder to resolve. On
especially from the endgame. In the opening a the other hand, most middlegame moves will
large majority of the pieces and pawns are still not radically alter the strategic character of the
on the board. In this situation, it is quite possi position. Even allowing for the heightened
ble that in each position there are two, three or possibility of irreparable error, the majority of
THE NATURE OF CHESS OPENINGS: FUNDAMENTALS 13
middlegame positions are still flexible enough with your plans, not to mention your next move.
to support more than one functional move and, I have placed this section about the centre be
sometimes, more than one strategy. fore the one on development of the pieces be
Setting these details aside, what is extremely cause it provides a foundation for everything
important and should be a part of your chess that follows in this book. As you read the chap
thinking is this: most features of a game, out ters on specific openings you will run into more
side of material loss or catastrophic setback, commentary about the centre than about any
can be changed or will evolve of their own ac other subject, so it's important to familiarize
cord as the game goes from opening to end yourself with the related concepts.
game. Mastering the opening is to some extent The four squares in the middle of the board
recognition of this fact and adaptation to it. (e4, d4, e5 and d5) are traditionally called the
'centre'. The value of the centre can be seen by
imagining a piece on a central square on an
Elementary Properties of Openings
otherwise empty board. Queens, bishops and
We now look at just a few fundamental fea knights all control more points from the centre
tures of opening play. These are presented on a than if they were placed on a non-central square.
very basic level to provide some tools and vo When we then include the bordering squares
cabulary with which you can advance to the (e3, d3, c4, c5, d6, e6, f4, f5), we sometimes use
next chapters and at least partially understand the phrase 'extended centre'.
specific opening discussions. The experienced
player may want to skip this material alto
gether.
The terms and ideas presented here are used
throughout the book. For this first chapter, the
assumption is that you know the rules of the
game, can follow chess notation, and know ba
sic chess terms such as 'file', 'diagonal', 'pin',
and so forth. You should also understand the
relative value of the pieces and how much
'material' both sides have in terms of relative
strength (counting points is the best way to
start). Finally, you should have played enough
to be comfortable with a discussion of chess
formations. A vast array of ideas and advice for
the inexperienced is given by books, electronic Notice that if pawns occupy the bordering
material, and web sites; what I'm presenting in squares they can contribute to control of the cen
stead is an extremely abbreviated version of in tre proper (the middle four squares). Normally
troductory material. Some of what you'll be when I speak of the 'centre', I'll be referring to
reading involves definitions of terms, which the four inside squares, but you may also want
will probably bore you but are necessary if to think about the border squares when I speak
you're going to understand the fun parts later. about 'central control'.
There's more jargon that you'll get used to as
The Centre you see specific examples. One player's pawns
on central squares are said to be 'his centre' .
Every opening has unique characteristics in For instance, we might say that White's centre
cluding pawn-structure, typical tactics, and di in the top diagram overleaf consists of the white
verse methods of attack and defence. But all pawns occupying d4 and e5.
openings have one consideration in common In the position in the lower diagram, Black
when it comes to organizing one's pieces: cen may be said to have a broad centre (or 'central
tral configuration and control. The centre is a front'), describing his pawns on c5, d5, e5 and
primary concern in deciding how to proceed f5.
available in front of the 2nd rank and both

knights can go to three such squares.
Here are some examples of the ideal centre
versus some not-so-ideal centres. Suppose a
game begins with these moves:
1 d4 a5
This is a common beginner's move, hoping
to bring a rook out via a6.
2 e4
Now if 2 . . .J:;ta6, White will simply capture
the rook.
2 h5 3 tllc3 tlla6 4 tllf3 g6 5 .i.c4 J:;tb7 6
•••
.ilf4 tllh6 (D)
To confuse things a bit more, the phrase

'centre pawns' also denotes pawns that occupy
any square along the central corridors from e2
to e7 or d2 to d7. Don't worry: none of this need
be memorized. It may just help a little as you go
along.
On occasion I shall use the phrase 'ideal cen
tre', which refers to having pawns on e4 and d4
when you are playing with the white pieces, or
pawns on e5 and d5 when you play Black. We
call that the ideal centre because of all the pos
sible first two moves, the advances d4 and e4 as
White (or . . . d5 and . . . e5 as Black) give your
pieces the most freedom to move about, and
therefore to have the greatest influence on the We can see how Black has neglected the cen
game. The player with the ideal centre can also tre. In fact, none of his pieces control d4, e4, d5
more easily add to his control of the four central or e5. White has by far the better position.
squares. For instance, he might place his pieces Even if Black plays more reasonably and de
as in the following diagram. velops his pieces in the centre, he can get in
White controls the central square e5 three trouble for lack of central control. A simple ex
times (with two pieces and a pawn), d5 three ample, again using the ideal centre for White:
times, d4 twice and e4 once. His pieces are de 1 e4 e6 2 d4 d6 3 tllf3 tllf6 4 tllc3 tllc6 5
veloped and active. Both bishops have six moves .ilc4 .ile7 6 .ilf4 .ild7 (D)
runs 4 .i.gS [indirectly controlling e4 and d5 by

pinning the defender of those squares] 4 .i.e7 •..
[indirectly controlling d5 and e4 by unpinning

the defender] 5 e3 [d4] 5 0-0 [a useful move,
•••
but doesn' t control a central square] 6 l2Jf3 [d4

and e5] 6 l2Jbd7 [e5] 7 l:tcl c6 [d5] 8 .i.d3 (D)
•••
[e4].
At least Black has established some influence

over the centre, but neither of his pawns has
reached the fourth rank; his pieces are cramped
behind their own lines. Compare Black's bish
ops, which have only two retrograde moves
available to them, with White's bishops, which
can reach 8 squares apiece. And while Black's
knights are actively placed, White's superior
centre can chase them away by means of the d) The Nimzo-Indian Defence: 1 d4 [e5]
pawn advances d5 or e5. White has a distinct 1 l2Jf6 [e4 and d5] 2 c4 [d5] 2 e6 [d5] 3 lZ'lc3
••• •••
advantage. What went wrong for Black? He [e4 and d5] 3 .i.b4 [indirectly controlling e4
•••
needed to challenge White's centre with his and d5 via a pin on the c3-knight]. One typical
own pawns, bringing one of them to d5 or e5 to line proceeds 4 e3 [d4] 4 cS [d4] 5 i.d3 [e4]
••.
break up White's ideal centre and establish ter S ... l2Jc6 [d4 and e5] 6 lZ'lf3 [d4 and e5] 6 dS •••
ritory of his own. [e4] 7 0-0 [unpinning the c3-knight, which re
Let's take a look at a variety of common gains its influence on e4 and d5] 7 0-0 8 a3
•••
openings with respect to central control. You .i.xc3 [eliminating the knight's control of e4
will see the universal emphasis on controlling and d5] 9 bxc3 [d4] 9 dxc4 10 i.xc4 [d5]
•.•
central points. For each move of a pawn or 10 �c7 [e5] 11 i.d3 [e4] 11 eS [d4] 12 �c2
••• .••
piece I have indicated the corresponding cen [e4, indirectly] 12 l:teS (D) [e5].
•••
tral squares that it controls (or helps to control)

in brackets:
a) In what is called the Italian Game, note
that every move for both sides controls at least
one main central square: 1 e4 [controlling d5]
1 eS [d4] 2 lZ'lf3 [d4 and e5] 2 lZ'lc6 [d4 and
.•. ..•
e5] 3 i.c4 [d5]. Black typically responds with

the 'Giuoco Piano' , 3 i.cS [d4], or 3 lZ'lf6 [e4
••• .••
and d5], the Two Knights.

b) The Ruy Lopez (or 'Spanish' ) goes 1 e4
[d5] 1 eS [d4] 2 lZ'lf3 [d4 and e5] 2 l2Jc6 [d4
••• •.•
and e5], and now 3 .i.bS attacks a piece that

controls d4 and e5, thus indirectly reducing
Black' s influence over them.
c) The Queen' s Gambit Declined: 1 d4 [e5]
1...dS [e4] 2 c4 [d5] 2 e6 [d5] 3 lZ'lc3 [e4 and
•.. In some openings, one or both sides fian
d5] 3 l2Jf6 [e4 and d5]. A traditional line now
••• chetto their bishops ('fianchetto' means to bring
" ... WllllJI
Ulitiltllll &�'NG�
• •TIP•rll 'n/E CHESS OPENINGS
a bishop to g2, b2, g7 or b7); this move is also In order to develop efficiently, it's often de
for the sake of central control; for instance: sirable to move each piece only once or twice
e) The English Opening: 1 c4 [d5] 1. e5 .• until they're all in useful positions. Also, be
[d4] 2 l2Jc3 [e4 and d5] 2...l2Jc6 [d4 and e5] 3 g3 careful about bringing the queen out early in
g6 4 i..g2 [e4 and d5] 4....ilg7 [d4 and e5 ] 5 d3 the game, because she is sometimes subject to
[e4] 5...d6 [e5] 6 lLif3 [d4 and e5] 6...f5 [e4] 7 attack and will have to retreat. The difference
0-0 l2Jf6 [e4 and d5]. between the queen and other pieces in this re
gard is that the queen can't be exchanged for
To be fair, half of the initial moves of a most other pieces (the exception being for an
knight and all those of the middle four pawns other queen) without losing a lot of material, so
control some central square, so one might think in many situations she has to run away from the
that central control practically takes care of it threat of capture and waste time.
self. But the openings above show that the mas Here's a short game that combines the con
ters who developed them intended to occupy cepts of centre and development:
and control central squares in a continuous and
harmonious way. To a strong player, a particu Estrin - Libov
lar central structure calls out to the pieces and Moscow 1 944
indicates where they should go. Then the pawns
and pieces control the key squares while they 1 e4 e5 2 lLif3 l2Jc6 3 .ilc4 i.c5
are safely defended and work together. This co So far every move has contributed to both
ordination of pieces leads to the next subject. development and central control.
4 c3
Development Now White tries to occupy the centre with
pawns. If he succeeds, that will determine the
Another critical but simpler opening idea is best available squares for his other pieces.
called 'development' . This refers to moving 4...l2Jf6 5 d4 exd4 6 cxd4 (D)
pieces (not including pawns) off their initial
squares and putting them 'in play ' . Just count
ing the number of pieces that you have moved
is the simplest measure of development. Of
course it's essential to consider the 'quality' of
development, that is, how well the pieces are
placed. There are some principles of good de
velopment, which are unfortunately limited by
the context of each position, first and foremost
by the pawn-structure. Nevertheless, as you first
get used to playing chess you will do well most
of the time to:
a) get as many pieces developed (off their
initial squares) as possible, preferably early in
the game;
b) bring those pieces to active squares where White has achieved the ideal centre, but Black
they have good scope (without subjecting them is slightly ahead in development, in the sim
to attack, of course); and plest sense of the number of pieces that are out
c) coordinate your developed pieces with in play.
the centre, working with pawns to control as 6...i.b6?
many central squares as you can. This retreating move allows White' s centre
Usually you can't achieve everything that to advance. Black needs to gain time to get his
you want to, but by keeping these principles in king castled into safety. The way to do that is
mind you will have a better chance of gaining 6 . . . .ilb4+!.
the advantage. 7 d5 ti'Je7?
Another backward move that allows White to

win more time. 7 . . .tb.a5 attacks White's bishop
on c4, but after White retreats the bishop by 8
i.d3, Black has to be careful because White is
about to play the move b4, winning the trapped
knight.
8 e5 tb.e4 9 d6! (D)
White's king i s sheltered and relatively safe.

The squares f3, g3 and h3 are all protected
from a piece intrusion, nor can the king be di
rectly attacked along diagonals by a bishop or
queen, or along files by a rook or queen. If
Black does manage to capture one of White's
pawns, that reduces the king's safety, but at
least the other two pawns are still around for
The centre is the key to most openings, and the king to hide behind. Black's uncastled
White's just keeps moving forward. king, however, is subject to checks, perhaps by
9 tb.xf2
•.• a knight on d6 or c7, from a bishop on b5 or c6,
Black indulges his greed by both taking a a rook on e l or a8, or a queen from several di
pawn and setting up a double attack on White's rections.
queen and rook. Nevertheless, pawn-cover for the king may
10 �h3 tb.xhl 1 1 i.xf7+ <>t>rs 12 i.gs! be more than overshadowed by the aggressive
The bishop pins the knight which is already placement of the opponent's pieces towards the
under attack. kingside (or queenside, if that is where castling
12 . cxd6 13 exd6 1-0
. . is contemplated).
Black resigns because he will lose his queen A position from a famous game illustrates
after dxe7+ or i.xe7+. Ouch! the point:
The moral of the story is that Black ne
glected to challenge White's centre and then
had to move his knights too many times in the
opening.
King Safety
One of the most important guidelines in chess

is to protect your king from harm. This elemen
tary consideration is sometimes forgotten. It
can strongly affect the proper conduct of the
opening stage of the game.
The most common method of enhancing a
king's security is castling, but it should be done
with eyes wide open. The goal is usually to pro
vide pawn-cover for the king, as in this skeletal Em. Lasker - Bauer
view: Amsterdam 1 889
A lot of White's pieces are pointed towards If only the queens have been exchanged, or if
the king so even its well-positioned pawn de only the queen and one or two pairs of other
fenders can't save it: pieces have been exchanged, then the king can
14 l2Jh5! lZJxh5 15 i..:xh7+!! @xh7 16 �xh5+ still be hounded before a true ending arrives.
@gs 17 .ilxg7! rJitxg7 (D) This is the sort of decision that comes with expe
rience.
Space and Its Properties
The amount of territory that is under one's con

trol, generally referred to as space, is a concept
that is deceptively hard to understand. The first
point to be made is that having space is an ad
vantage more often than not. It gives you more
room to organize your forces and with luck it
will frustrate your opponent, who will have dif
ficulty getting his forces out. When you control
more territory you can often move your pieces
from one theatre of action to another more
quickly than your opponent can, and thus attack
See how White has ripped away Black's pro- on that front before he can defend. Some great
tective pawn-cover? players have spent their careers playing open
18 °ifg4+ rJith7 ings that emphasized the control of space over
l 8 ...rJitf6 1 9 �g5# is already checkmate! any other factor, even the assumption of weak
19 .l::tf3 nesses in their position or difficulties with their
The last type of piece joins the attack. White's development.
idea is 20 J:;th3+ .ilh4 21 J:;txh4#. Notice how the In many situations I shall simply assume
rook can only have an effect on Black's king without explanation that the side with space has
when the pawns in front of the king are gone. an advantage, although in other cases space
19...e5 20 l:Ih3+ �h6 21 .l::txh6+ @xh6 22 may be a problem that needs to be overcome!
'ifd7 For example, the possessor of more territory
This fork finishes off the combination by has more of the board to defend. That may seem
winning a piece. White is well ahead in mate trivial, but some positions are well-known for
rial now and even comes back to complete the the property that the player with less space ties
attack on Black's king: down the one with more space by constantly
22 .ilf6 23 �xb7 rJitg7 24 J:;tfl J:;tab8 25
••. threatening to change the pawn-structure in his
'ifd7 i:;tfd8 26 �g4+ rJitrs 27 fxe5 .ilg7 28 e6 own favour if his opponent tries to do anything.
.l::!. b7 29 °ifg6 f6 30 J:;txf6+ .ilxf6 31 �xf6+ rJit eS Several variations of the Sicilian Defence, the
32 �h8+ @e7 33 �g7+ rJitxe6 34 �xb7 most popular opening in chess, include lines in
and Lasker won shortly thereafter. which something of that nature occurs.
Since space is usually defined as a portion
Strong players have no fear of leaving their of the board that is delineated by pawns, one
kings in the centre if that is the safest place on question that needs to be answered is whether
the board, or if by doing so the king contributes those pawns are true boundaries or simply a
to the defence of weak or potentially weak temporary construct that can be neutralized.
squares. Sometimes an opening is even based For example, pawns can be overextended in
upon the useful position of the king. Also, when the opening such that the squares behind the
an opening becomes rapidly simplified, the king pawns are compromised and pawn advances
may remain in the centre to assist with the end do not correspond to control of space. Con
game. Centralized kings will generally be strong sider this position from the King's Indian Four
pieces in an endgame, but here one must beware. Pawns Attack:
the centre. You can see how positively that af

fects his development and activity.
Regardless of the mediocre quality of play in
this example, the lesson remains: if you seize a
large hunk of the centre in the first moves of the
game, make sure that you can defend the pawns
that control that territory. The concept of space
advantage is only significant when the pawns
and pieces begin to assume more settled posi
tions.
By contrast, look at this example from one of

the main lines of the same King's Indian De
fence:
White's pawn penetration into Black's posi
tion defines White's territory and he has an in
disputable advantage in space. He also occupies
more of the centre. But occupation and control
are two different things and the possibility of
undermining the advanced pawns can make
them unstable. For instance, the play from the
diagram might continue 7 ... tb.d7 8 h4?! (White
stakes out even more territory and tries to attack
the king by playing h5; however, he is making
too many pawn moves when he should be de
fending the space that he has grabbed in the
centre) 8... c5! (this is referred to as 'undermin
ing White's centre'; regardless of what White
does, his pawns will be cleared away) 9 exd6
l:te8 IO dxe7 l:txe7+ 11 i.e2 cxd4 12 tb.xd4 It's already fair to say that Black has staked
lb.b6 13 tb.f3 i.f5 14 �xd8+ l:txd8 (D). out territory and has space on the kingside
whereas White has space on the queenside.
Surely enough, a few moves later we might see
something like:
At this point Black controls every central

square and threatens ...tb.b4, while at the same
time White's bishops are running into their
own pawns. It turns out that Black has a win There's no question of who has secured terri
ning position because, ironically, he controls tory on which side of the board.
In numerous openings we'll be talking about the right colour can exchange a knight from
who has a space advantage, and what it means afar. Thus the two bishops can do more than
in terms of the assessment of the position. simply control squares.
There follow some elementary properties of
Piece Characteristics the pieces, and advice regarding their use in the
opening. Most readers will find them almost
Some fairly elementary terminology disguises self-evident, but this chapter is primarily de
much more complex issues that will come up signed to help the inexperienced player become
in the next few chapters. But it's worth dis comfortable with ideas that we'll be referring to
cussing a few representative terms with respect later.
to pieces.
First of all, we have a couple of terms to de 1. Bishops like open diagonals and should
scribe knights and bishops. They are called usually be developed accordingly. You may
'minor pieces', in contrast to the rooks and also use your bishop to pin an enemy piece, or
queen, which are called 'major pieces'. I shall to unpin your own. Exchanging your bishop for
regularly refer to the advantage of the 'bishop a knight is reasonable, but do so only to gain
pair' or 'two bishops' in this book. This reflects some advantage (or if forced to), otherwise you
the fact that in every stage of the game, includ will be surrendering the advantage of the
ing the opening, having two bishops on the bishop-pair for no return.
board versus two knights or a bishop and a Although there are many exceptions to this
knight more often than not constitutes a mean in various openings, try not to let your bishops
ingful advantage. That emphatically qualified become trapped behind their own pawns with
statement reflects the fact that, in a consider out good reason. Having said that, limiting a
able minority of cases, the player who pos bishop's activity may be necessary to ensure
sesses two knights or a knight and bishop will that your knights, rooks and remaining bishop
have the advantage over, or at least stand equally secure good positions.
with, his opponent who possesses two bishops.
Nevertheless, those instances are in the minor 2. Knights also need as much freedom of
ity, and when the bishop-pair is a recognizable movement as possible, but only to the extent
advantage I shall often point that out. Likewise, that they don't unduly interfere with the ac
if the bishops are hemmed in and/or the knights tivity of other pieces. For that reason, you may
are in excellent positions, that will frequently see knights developed on the second rank or on
be mentioned. Much of the time, however, I the side of the board with their first move, in
hope that the reader will come to notice all stead of to one of the 'ideal' squares f3, c3, f6 or
these imbalances on his or her own. c6. Knights are particularly fond of outposts,
So why are the two bishops so good in tan which arise in many openings. An outpost is a
dem? First and foremost, because they cover place in the opponent's pawn-structure where
squares of both colours. The bishop is a power your piece cannot be attacked by a pawn. To
ful, long-range piece that in a sense 'should' be have significance, an outpost should be on at
better than the knight because it can attack from least your 4th rank, and preferably on the 5th or
afar; but unlike a knight, a bishop can only 6th rank. From an outpost on a central file, a
travel on one colour. With two bishops that dis knight can exert considerable influence on sev
advantage is partially corrected. But another eral squares in the enemy position while main
considerable advantage is that the possessor of taining defensive coverage. Here's an example
the bishops can exchange one or even both of of an outpost that's occupied by a knight (see
the knights under favourable circumstances, i.e. upper diagram on following page):
dictate when and where he can exchange other The defining feature of the outpost is that the
pieces to advantage. It is difficult for the short knight can't be captured by a pawn. Its influ
hopping knight to track down and exchange a ence would be further strengthened by a rook or
bishop that is performing magnificently (or ful queen on the d-file, or by another knight on c3
filling some essential function), but a bishop of or e3.
This is a real-world example: Black has an

outpost on d5. Some students will say that
White's knight has an outpost on e5, but notice
that if Black's knight moves from f6, he could
then attack the knight with a pawn by ...f6. In
stead, the e5-square is sometimes called a
'support-point' because it is supported by his
d4-pawn and unlikely to be driven away by a
pawn in the near future. Knights are sometimes
just as happy to reside on a support-point as
they are to occupy an outpost.
The outpost and support-point are examples
of structural configurations, a subject that we
expand upon in Chapter 2 and still more Chap
ter 3.
3. Rooks like to have open files, preferably

ones that extend vertically as far as possible
into the enemy camp. Early pawn exchanges
will sometimes let rooks breathe and have im
mediate effect upon the game. In the opening,
assuming that you castle in one direction or an
other, your rooks may well end up on half-open
files (ones blocked by your own pawns). If they
can be centralized so as either to defend your e
or d-pawns or to assist in their advance, that's
also not a bad role. Doubling rooks (placing
one behind the other on an open file) used to
In the lower diagram Black's knight is on an be uncommon in the opening stage, but since
outpost that is unsupported by his own pawns openings extend further and further into what
yet not subject to attack by the opponent's was previously called the middlegame, you'll
pawns. Notice that Black could also occupy definitely run across that situation. Likewise
this outpost with a dark-squared bishop, rook, with the placement a rook on the seventh rank,
or queen. Support for the eS-knight could come which isn't generally possible until after the
from rooks on the open file, another knight, a opening, but does occur, usually to assist in an
bishop on f6 or d6, and various placements of attack. Rook-lifts to the third rank, on the other
Black's queen. hand, happen relatively frequently; often they
will move horizontally to help with an attack on
the opponent's king. Another common rook-lift
in the opening is to the second rank, because a
rook which moves horizontally along the second
rank can defend extremely sensitive squares
such as the ones immediately in front of the
king. This 'second-rank defence' is essential
against some attacks, and such rooks may also
be able to swing to the e- and d-files to support
the centre.
4. Apart from wide-open games in which

the centre pawns are blown off the board
early on, the queen tends to stay at home or
to lurk behind her pawns and pieces in the early because active pieces tend to force slower ones
stages of the opening. Increasingly, advanced onto the defensive, resulting in the creation of
players are bringing the queen out early but in a weaknesses in the enemy camp. The balance
judicious manner to control more of the board - that generally exists between attack and de
a practice that you'll see in this book. There's fence in chess will break down if one player is
nothing wrong with exchanging queens in the working with direct threats and gaining more
opening, but there's also no reason to go out of control of the board. Gathering momentum like
your way to do so, as so many young players this is called 'having the initiative'. As long as
do. the aggressor is able to force his opponent to
keep reacting to threats, he will maintain his
Activity and Initiative initiative. Sometimes the initiative peters out,
especially if handled poorly; it can even change
I shall refer time and again to a player's active hands. In this book, you will run across an as
pieces and to activity in general. This is a con sessment of mine that reads simply 'Black has
cept that may encompass a coordination of the initiative'. While it is ambiguous how much
forces, but to a first approximation simply ex advantage that confers on Black, the initiative
presses the mobility and reach of one's pieces. constitutes an advantage in and of itself.
Active pieces control more squares. Such pieces This chapter has covered terminology and
aren't necessarily involved in a direct attack but general ideas that I hope will serve you well. Re
can serve to harass opposing forces, support a member that most of what is discussed in these
pawn advance, and generally accrue more terri first three chapters will be applied and rein
tory. You will see that in opening play the active forced in the investigation of specific openings
player tends to get the better game, in part that occupies the larger portion of this book.
2 Opening Ideas and Positional
Features
In this chapter we'll begin by considering some search for an advantage? Books from the first
general and even philosophical issues about half of the 20th century particularly stressed the
opening play. We'll then tum to special topics need for equalizing before all else. They often
involving different types of centres and proper implied that the advanced, mature player would
ties of pieces and pawns. Much of the chapter focus on neutralizing White's first-move ad
will be devoted to pawns and weaknesses, open vantage, whereas the impatient youngster who
ing the investigation of 'positional' chess and tried to bowl over his opponent would be pun
setting the stage for its more detailed discus ished by a seasoned master. This attitude may
sion in Chapter 3. have slowly evolved out of experiences with
the openings that were played in the middle of
Black's Goals in the Opening the 19th century, openings which gradually lost
favour after players became more 'scientific'.
Chess books have traditionally said that Black's Most games of that day began with I e4 e5, and
goal in the opening is to obtain equality. A pop the apparent failure of ambitious counterattacks
ular variant of this is that Black must first se by Black reinforced the philosophy of 'equal
cure equality and only later search for chances ity first'. For example, interest dropped in the
to gain the advantage. There are certainly open more exotic King's Gambit lines such as that of
ings in which that is likely to be the case, but in the famous Andersson-Kieseritzky 'Immortal
many openings Black also has the choice to Game' : 1 e4 e5 2 f4 exf4 3 .tc4 �4+ 4 @fl
play aggressively and endeavour to steal the ad b5?! (D).
vantage from White right away. In cases where
he falls short of that goal, energetic opening
play by Black may still lead to a position so
complex and unclear that to speak of equality is
meaningless. Sometimes we say 'dynamically
balanced' instead of 'equal' to express the view
that either player is as likely as the other to
emerge from complications with an advantage.
This style of opening play has become preva
lent in modem chess, with World Champions
Fischer and Kasparov as its most visible practi
tioners.
Both approaches to playing Black are valid,
and the distinction between them contributes
to the diversity of styles amongst contempo
rary players. Of course, we should remember Also pushed to the periphery were 1 e4 e5 2
that White has always had a better percentage .tc4 .tc5 3 b4 .txb4 4 f4 and 1 e4 e5 2 f4 exf4
score than Black. But is that due to Black's ac 3 tt:\f3 g5 4 .tc4 g4 5 tt:lc3 (maybe not so horrid
ceptance of a small disadvantage in the course but abandoned nevertheless).
of playing directly for equality, or does it re Similarly, the adventurous Evans Gambit
sult from Black becoming overextended in his stayed around for a while, but after 1 e4 e5 2
lt:lf3 lllc6 3 .i.c4 .i.c5 4 b4 .i.xb4 5 c3 .i.a5 6

0-0 the likes of 6 ...'i!Vf6 ! ? were largely replaced
by safer defences such as Lasker' s 6 ...d6 7 d4
.i.b6. In addition, provocative openings such as
the Philidor Countergambit ( 1 e4 e5 2 lllf3 d6 3
d4 f5), and the Schliemann Defence to the Ruy
Lopez (1 e4 e5 2 lt:lf3 lll c6 3 .i.b5 f5) were held
to be dubious or were at any rate supplanted by
more careful strategies. Lastly, responses to 1
e4 which favoured confrontation over equality
also failed to gain a foothold until their playa
bility was established. Most masters didn't take
seriously such moves as 1 ...lt:lf6, l ...d6 and
1 ...g6, nor was 1...d5 approved of by the lead
ing masters. In fact, the latter has only been latter half of the 1 9th century and into the 20th,
convincingly revived in the last ten years. the Sicilian Defence grew to be played in a re
As an alternative to l ...e5, the solid Caro spectable 5% of such encounters. At first it was
Kann ( l...c6) gained popularity after 1 900, pri White who failed to play aggressively in the
marily as an equalizing weapon. In the same centre, typically choosing the Closed Sicilian
'equality-first' vein, French Defence players (2 lll c3) or 2 f4. As players then turned to the
employed the unambitious move ...dxe4 (e.g., 1 Open Sicilian with 2 lllf3 and 3 d4, Black
e4 e6 2 d4 d5 3 lll c3 dxe4), and the French De tended to play active, developing moves, until
fence generally lacked the dynamic character the Scheveningen Variation with its backward
that it later acquired. (To this day, in fact, ...dxe4 central structure (...e6 and ... d6) was brought to
systems are chosen by leading grandmasters, general attention in the 1920s by prominent
often as a way to simplify the play and equal players such as Euwe. Soon, various new inter
ize). When players did essay upon 1 d4 instead pretations of the Open Sicilian became estab
of 1 e4, l ...d5 was the overwhelming response lished as main lines. But the extent to which
by Black, with the various 'Indian' defences Black could disrespect the basics of develop
(beginning with I...lt:lf6) held in low esteem. ment and space in favour of other factors be
Looking back, we can see that the legitimate came apparent only much later. During the
desire to establish a pawn presence in the centre 1 940s and 1 950s new interpretations of the Si
greatly influenced the choice of and attitudes cilian Defence ushered in a modern age of dy
towards opening play. The Sicilian Defence ( 1 namism; players and theoreticians developed
e4 c5) neglects to move a centre pawn (see the the fundamental structures and piece-play that
next paragraph), whereas defences to 1 e4 such are used today by nearly every major player.
as 1 ...lllf6, l ...d6 and 1 ...g6 all concede the Dynamic variations of traditional openings also
ideal or at least favourable centre to White. So gained popularity; e.g., the Winawer Variation
do several of today's dynamic and/or unbalanc in the French Defence and the Marshall Attack
ing replies to 1 d4. For instance, the King's In in the Ruy Lopez. The Alekhine Defence and
dian Defence allows White to occupy the centre Pirc Defence had accumulated masses of the
directly in the main lines after 1 d4 lt:lf6 2 c4 g6 ory and stalwart grandmaster adherents by the
3 lll c3 .i.g7 4 e4 d6 (D). time Fischer used both openings in his 1 972
Compare the related discussion at the begin World Championship match versus Spassky;
ning of Chapter 3. today these openings are played less than oth
The Sicilian Defence ( 1 e4 c5), which ac ers in high-level chess but certainly retain their
counts for nearly 20% ( !) of all top-level grand legitimacy.
master games played today, was at first a more After that lengthy digression, no one will be
ambiguous case, with a curious evolution. Al surprised to find that either of Black's ap
though one sees only a handful of modern-style proaches to the opening is valid, that is, he can
treatments in top-level games throughout the play for equality or aspire to achieve a dynamic
OPENING IDEAS AND POSITIONAL FEA TURES 25
imbalance. Some players just starting out, how into complex areas that are not essential for the
ever, may not have heard about the latter op inexperienced player to master.
tion.
1. The 'vanishing centre'. As the name im
Wh ite's Goals in the Opening plies, all or most of the centre pawns are ex
changed or captured. They leave a gap in the
White has choices similar to Black's, assuming middle of the board through which pieces can
that he has the same opportunities. White can move in a more-or-less unobstructed fashion.
work patiently to hold on to his inherent ad The vanishing centre tends to favour the side
vantage, usually by suppressing his opponent's with the better development, and tactics can
counterplay and 'accumulating small advan easily dominate the play; for example, in the
tages'. Or White can seek dynamic situations in Danish Gambit with 1 e4 e5 2 d4 exd4 3 c3
which he tries to take the initiative and keep dxc3 4 .i.c4 cxb2 5 .i.xb2 d5 6 .i.xd5 'Llf6! 7
Black on his heels. Finally, White can plunge .i.xf7+ ! @xf7 8 �xd8 .i.b4+ 9 �d2 .i.xd2+ 10
into two-sided slugfests and hope to express his 'Llxd2, and the game settles down. But if devel
theoretical advantage or superior skills in that opment is about equal and the game hasn't been
environment. Once again all of these methods reduced to disorderly skirmishing, then vulner
are admissible. But for White there is a differ able points and pawn weaknesses can be mag
ent twist. Curiously, it is sometimes easier for nified because they are so accessible.
Black to launch an effective attack and to define
the quality of early play than it is for White to 1 e4 e5 2 d4 exd4 3 c3 d5 4 exd5 �xd5 5
do the same. Black has the advantage of know CLJf3!?
ing his opponent's moves ahead of time. If he 5 cxd4 is the main move.
chooses to play a solid game it may be impossi 5 .tg4 6 .i.e2 (D)
.•.
ble for White to attack aggressively. Of course

the reverse is also true: White can play 1 d4, 2
'Llf3 and 3 .i.f4 against most openings, or, for
example, 1 'Llf3, 2 g3, 3 .i.g2, 4 d3 and 5 0-0
against practically anything. But most players
aren't interested in giving away the advantage
of the first move with such conservative moves
and so will choose to play more ambitiously.
Paradoxically, this can let Black set the pace in
certain openings.
Central Types
Several very important central formations will

be explored in detail in the next chapter. Among
them will be centres characterized by: 6 ... d3! 7 .txd3 .i.xf3 8 gxf3
a) isolated pawns; White's doubled pawns are a serious disad
b) majorities and minorities; vantage. His bishop-pair on an open board offers
c) restrained central pawns; and some degree of compensation, but probably not
d) pawn-chains. enough, since it's easy for Black to develop his
Most other types of centre that have practical pieces.
significance will be represented somewhere in
the main body of the book. It's useful to look at Here's an illustration from a d-pawn open
some of those central formations to get a feel ing:
for how they can be analysed and assessed. Be 1 d4 'Llf6 2 c4 g6 3 'Llc3 .i.g7 4 e4 d6 5 'Llf3
aware that the material in this chapter will be 0-0 6 .i.e2 e5 7 dxe5 dxe5 8 �xd8 .l:Ixd8 9
gin at an elementary level but quickly move 'Llxe5 'Llxe4 10 'Llxe4 .i.xe5 (D)
also has the bishop-pair. White has a mobile

majority on the kingside, which can theoreti
cally be used to create a passed pawn. But that's
far down the road and in the middlegame, espe
cially with the vanished centre, one would ex
pect that Black's two bishops would be more
effective than the bishop and knight. The prob
lem is that White controls more space and Black
has no centre pawn with which to drive White's
pieces away. The position is about equal.
The vanished centre shows up in old gambit

lines which were popular 1 00 or more years
ago. Some of these lines have never been per
This is a well-known version of the Exchange manently stowed away. A case in point:
Variation of the King's Indian Defence where 1 e4 e5 2 l2Jf3 l2Jc6 3 j,_c4 lDf6 4 d4 exd4 5
a series of early exchanges has decimated the 0-0 l2Jxe4 6 l:tel d5 7 j,_xd5 'iVxd5 8 lbc3 'iVh5
centre. According to theory, Black stands 9 l2Jxe4 j,_e6 10 j,_g5 h6 1 1 j,_f6! 'iVg6 12 l2Jh4
slightly better. White has weaknesses on d3 °iVh7 13 c3 dxc3 14 bxc3 (D)
and d4 that can be occupied by Black's minor
pieces, whereas White can't find good squares
other than f6 to exploit in Black's position. If
White waits around, Black will occupy the d4-
square by ...lbc6-d4, so White should move
quickly and play 1 1 j,_g5 j,_xb2 12 j,_xd8! j,_xal
13 j,_xc7 lbc6 14 0-0 j,_g7, when Black has
only a small positional advantage.
1 e4 e5 2 ltJf3 ltJc6 3 j,_b5 a6 4 j,_xc6 dxc6 5

0-0 j,_d6 6 d4 exd4 7 'iVxd4 f6 8 l:tel l2Je7 9 e5
fxe5 10 lbxe5 0-0 (D)
This is all theory, that is, published knowl

edge. The centre has been cleared out and there's
no way to make a simple assessment. Only a lot
of brainpower, computer analysis and corre
spondence chess can solve this sort of thing; in
fact, only those things got chess researchers
this far ! Which brings me to another point: my
aim in this book is to have you understand strat
egy, including typical methods for both sides to
handle attacking positions. It's often possible to
indicate recurrent themes and some connec
tions among them. However, I shall rarely ana
This position and ones like it have occurred lyse chaotic positions like this one featuring
regularly in the Ruy Lopez Exchange Varia moment-to-moment variations in tactical events.
tion. The centre pawns have been swept away The correct moves are so unpredictable that
but static factors are still controlling the play. they really can't be 'explained' except on a
Black has the weaker pawn-structure but he case-by-case basis. You may be able to find out
OPENING IDEAS AND POSITIONAL FEATURES 27
more about them in books that make specific pawn on d6 or e6, respectively. Generally,
detailed investigations; better yet, you can try White has somewhat the better game by virtue
to work them out for yourself ! of his greater control of space, but Black has a
compact structure and an open file aiming at
2 . We have already seen and discussed cases White's 4th-rank pawn, so the advantage can
of the 'ideal centre' (also known as the 'classi range from tiny to moderately significant.
cal centre'), in which one side has pawns on e4 Here's an illustration taken from the 'classi
and d4 (or e5 and d5). Normally the ideal centre cal' Philidor's Defence:
constitutes an advantage, but that's only true if 1 e4 e5 2 tt:\f3 d6 3 d4 exd4 4 tt:lxd4 tt:lf6 5
it has some positive effect on the position; for tt:lc3 .te7 6 .tc4 0-0 7 0-0 l:Ie8 8 .tf4 .trs 9 f3
example, tying down the opponent's pieces, ad (D)
vancing with tempo, creating a passed pawn,
and/or serving as the pivot point from which
pieces can launch an attack. Otherwise the op
ponent might be able to attack the centre pawns
from afar with little risk. In a typical situation
Black restrains White's ideal centre but can't
break it down. This imbalance arises in certain
variations of the Queen's Gambit Accepted,
Slav, Grlinfeld, and this main line of the Semi
Tarrasch:
1 d4 d5 2 c4 e6 3 tt:lc3 tt:\f6 4 tt:\f3 c5 5 cxd5
tt:\xd5 6 e4 tt:lxc3 7 bxc3 cxd4 8 cxd4 .tb4+ 9
i.d2 .txd2+ 10 �xd2 0-0 1 1 .tc4 tt:lc6 12 0-0
b6 13 .l:!.adl .tb7 (D)
White has a pleasant advantage because he

controls more space and has freer development.
A surrender of the centre occurs in the old

main line of the Caro-Kann Defence:
1 e4 c6 2 d4 d5 3 tt:lc3 dxe4 4 tt:lxe4 .trs 5
tt:lg3 .tg6 6 tt:\f3 tt:ld7 7 h4 h6 8 h5 .th7 9 .td3
.txd3 10 �xd3 �c7 1 1 .td2 e6 (D)
White has won some famous battles from

this position, but the moves ...e6 and ... .tb7 in
conjunction with ... tt:la5 can serve to restrain
White's ideal centre, while ...�d6 covers key
squares, so the position is only a little bit better
for White.
3. The formation arising from what is called

the 'surrender of the centre' appears in many
different openings. It involves a single white Black has less space but a safe position and
central pawn on e4 or d4 facing a lone black no weaknesses. He can also try to break down
the d4-based centre with the move ... c5. White A similar illustration of the advance g4 as a
has the easier game, but against accurate play disincentive to ...d5 is 7 ...ll'ic6 8 �d2 .i.e7 9 g4
he will retain little if any advantage. This type d5?! 10 g5 tt:lxd4 11 �xd4 tt:lh5 12 f4!. White
of 'restraint centre' will be discussed at some is threatening 13 .i.e2, and 12...dxe4 (12 ...h6 13
length in Chapter 3. exd5 hxg5 14 fxg5 .i.xg5 15 0-0-0) 13 �xd8+
�xd8 14 .te2 g6 15 0-0-0+ leaves Black strug
gling.
Flan k versus Centre
8 g4 dS!? (D)
It's always hard to assess whether a flank pawn
advance in the opening is strong or weak. It's
often said that a centre has to be safe in order to
justify a pawn advance. That is true in many sit
uations; e.g.:
1 e4 cS 2 ll'if3 d6 3 d4 cxd4 4 ll'ixd4 ll'if6 S
ll'ic3 g6 6 .i.e2 .i.g7 7 .i.e3 0-0 8 0-0 tt:lc6 9
tt:lb3 aS 10 a4 .i.e6 11 g4 (D)
This is too early an advance. The centre
should be secured by 11 f4 with the idea 11...d5?
12 f5.
9 gS tt:\fd7 10 exdS .i.xgS 1 1 .i.xgS �xgS

At this point the position doesn't look that
bad for Black, but a simple move illustrates
how effective it was to drive Black's knight
away from f6.
12 �d2! �xd2+ 13 @xd2 tt:lb6 14 .l:!.el
White stands comfortably better. Black will
have serious weaknesses after 14 ... 0-0 15 dxe6
fxe6 16 .i.d3. In these examples, the flank de
fends the centre.
1 1 dS!
..• As long as you're aware that each situation
A central counterattack creates a threat on e4 has to be assessed on its own merits, you should
and makes the g-pawn look foolish out there always consider responding to a flank attack
doing nothing. with a central counterattack, and vice-versa.
12 exdS tt:lxdS 13 tt:lxdS .i.xdS 14 c4 .i.e6 But neither response should be made into a
lS tt:lcS �c8 16 ll'ixe6 �xe6 rule.
Black stands better because White's central
position is weak and the g4-pawn renders his Weaknesses
kingside difficult to defend.
The word 'weakness' refers to problems with
But the reverse is also true: flank pawn moves pawns and pawn-structures. Some terms relat
will frequently drive a piece away from a square ing to pawns still need to be defined, which
on which that piece controls the centre and/or we'll do presently. First, however, I want to
threatens to support a central advance. Another make a broader comment. Pawn weaknesses
line of the Sicilian Defence is a case in point: are to be avoided at any stage of the game if you
1 e4 cS 2 tt:\f3 d6 3 d4 cxd4 4 tt:lxd4 tt:\f6 S get nothing in return for them, and understand
tt:lc3 a6 6 .i.e3 e6 7 f3 .i.e7 ing pawn-structures (a subject much wider than
pawn weaknesses) is more important than any discussion and many examples of isolated
other factor in understanding chess. But that in pawns. In some very typical situations, their
sight should not be confused with a general advantages are famously in balance with their
phobia towards weaknesses. Generally they are disadvantages, which is why so many players
not as important in the opening as they are later rush to take them on and others to play against
in the game. Tarrasch' s dictum 'Before the them.
endgame the gods have placed the middlegame'
is part of the explanation, yet it is not the whole
story here. As the middlegame progresses and
considerable simplification has occurred (or is
imminent), a player must be particularly con
cerned with current weaknesses, and eventu
ally with what an endgame might bring if that
pawn-structure persists. Sometimes this calls for
radical action. But in the opening stage (partic
ularly within the first 10 moves or so) structural
weaknesses are generally more of an immedi
ate defensive problem than one which must be
attended to for the sake of the ending. They can
be incorporated into an overall approach to a
position that works extremely well; e.g., a terri
bly weak pawn may temporarily provide pro White has three isolated pawns, on a4, c3
tection from the opponent ' s play and allow you and e3. Black has one isolated pawn on b7. The
to gain the advantage. That holds true because pawns on a4, c3 and b7 are on open files and
of the ever-changing nature of most openings thus relatively more exposed than the pawn on
and middlegames. Especially players who are e3, which is masked by an opposing pawn on
beginning to gain experience with chess should the same file.
not overestimate the drawbacks of weaknesses
such as doubled, isolated or backward pawns 2. A backward pawn is one that has at least
and thus ignore good opportunities for attack or one pawn of its own colour on an adjacent file,
other positive activity. I find that students gen but that neighbouring pawn is situated one rank
erally err on the side of caution in this respect, or further ahead of its compatriot.
when they could aggressively pursue the initia
tive. So yes, try to avoid unnecessary weak
nesses and take advantage of those in your
opponent' s position, but don' t make decisions
that are too focused on just this one aspect of
the game. Y our pawn-structure may be telling
you other important things about how to handle
the position as a whole.
There follow some definitions and short ex

planations of pawn types and properties. In
Chapter 3 we investigate and evaluate these in
much greater detail.
1. An isolated pawn is one that has no

pawns of its own colour (i.e. friendly pawns) on In this well-known position from the Sicilian
any adjacent file. In practice, we are especially Defence, Black's d6-pawn is a backward pawn.
concerned with such a pawn when it's on an Often the square in front of the backward pawn
open file. In Chapter 3 you will find a lengthy serves as an outpost for the opponent, as it does
here (see Chapter 1 for a description of the out

post). We care most about backward pawns on
open file, as is Black's on d6. Backward pawns
are usually weak, but not always so.
3. A doubled pawn is one that resides on

the same file as another of your pawns. As
usual, doubled pawns on an open file are
weaker than those that are masked by enemy
pawns. Doubled pawns can be weak or strong,
but most of the time isolated doubled pawns
on an open file are a serious disadvantage,
both because they are hard to defend and be
cause there is a wonderful outpost in front of
the pawns, just asking for an opposing piece to Assessing Black's bishops is the easiest task.
occupy it. Here is a well-known situation in Black's pawns on e6 and d5 are on light squares,
which a knight is stationed in front of doubled so his bishop on d7 is 'bad' and the one on e7 is
f-pawns. 'good'. It happens that all of Black's other
pawns are on light squares as well, but except
for the c- and f-pawns, which are of limited im
portance, they aren't factors in the way we as
sess whether a bishop is good or bad.
White's light-squared bishop may look use
less because it is blocked by pawns on f3, h3
and g3, while even those pawns on c2, c3 and
a4 might provide obstacles. But it is a 'good'
bishop because White's centre pawns are on
dark squares. By contrast, the a5-bishop has
two nice open diagonals and can even reach the
wonderful outpost on e5. Nevertheless, it is a
'bad' bishop because it is on the same colour as
the central pawns. The point is that a 'good'
bishop can be a poor or even dysfunctional
piece whereas a 'bad' bishop may be the best
4. Pawns that block the path of pieces are piece on the board. However, those situations
always a problem, and the most famous of such are exceptional. In a considerable majority of
problems involve 'good' and 'bad' bishops. I'll cases a 'good' bishop really is the one that
be using those terms throughout the analysis serves you the best (and that you don't want to
section, so I should attempt a definition. A exchange!), while a 'bad' bishop tends to be ob
'bad' bishop is one whose central pawns are on structed and passive. This generalization goes
the same-coloured squares as the bishop; con back to the extraordinary importance of the
versely a 'good' bishop lives on the squares that centre.
are of the colour opposite to its central pawns. Bad bishops can serve as decent defenders
Notice the emphasis on central pawns. By far but they can be particularly unhelpful when op
the most important pawns in determining the posed by a good knight (see following dia
'goodness' or 'badness' of a bishop are the d gram):
and e- pawns. Adjacent c- and f-pawns can be 13 dS?!
...
factored in if they seem relevant to the bishop's In a fairly conventional Sicilian position,
overall mobility, but these pawns must be given Black plays the standard ... d5 break, thinking to
considerably less weight. Let's look at this situ free his pieces. But he may not have considered
ation in the abstract: the full consequences of a general liquidation.
formation of the d-pawn, knight and bishop is

one that frequently arises. Now watch how
White exploits the dark squares, his advantage
in space, and superior mobility.
21 as bxaS 22 �xaS .l:!.tb8 23 b3 �b6 24
�d2 as 2S .l:!.al Ji.d7 26 .l:!.fel g6?!
White gains a crucial kingside square after
26...a4 27 bxa4 .l:!.xa4 28 .l:!.xa4 ii.xa4 29 tt::l f5!.
A better try is 26...�c5.
27 �f4 .l:!.e8 28 h4! �b4 29 .l:!.xe8+ .l:!.xe8 30
'it>h2 a4 31 bxa4 ii.xa4 32 c3 �c4 33 hS! ii.d7
34 h6 .trs 3s �d6 1-0
There's nothing to be done about �f6; e.g.,
35...�xc3 36 �f6 �c7+ 37 'it>hl @f8 38 tt::l c6,
V. Gurevich - Zakharov etc.
Azav 1995
Fianchetto Themes and Prophylaxis

14 exdS tt::lbxdS lS tt::lxdS tt::lxdS 16 ii.xdS!
exdS Bad bishops can serve some productive roles
This is a type of end-position that can result that are not always obvious. The word 'prophy
from a number of other openings, such as a laxis' in chess has to do with the prevention of
French Defence with 3 tt::l d2 c5 or a number of an opponent's plans and desired-for continua
Queen's Gambits in which Black plays ...c5. tions, the latter including freeing moves and
The simplification that has occurred favours moves that serve a productive purpose, whether
White, who now succeeds in getting rid of defensive or aggressive in nature. Although the
Black's good bishop. concept of prophylaxis can also embrace a wider
17 Ji.f4! Ji.d6 set of meanings, those are the relevant ones for
After 17 ...�xf4 18 �xe7, Black lacks a really most discussions about openings.
good square for his c8-bishop so he has a tough Fianchettoed bishops, for instance, can be
time getting his rooks out. In the meantime, af bad and still serve prophylactic purposes. By
ter .l:!.fel , all of White's pieces would be ac way of illustration, one might wonder why
tively placed. Black spends two moves to fianchetto his
18 ii.xd6 �xd6 19 �d2 ii.g4 20 f3 ii.e6 (D) bishop in the King's Indian Defence and then
plays ...e5 to block it off! And why does Black
in that defence often go to lengths to avoid ex
changing that bishop? Shouldn't it be consid
ered the epitome of a poorly-placed bishop?
To the contrary, King's Indian fans tend to
think of that piece as their most precious pos
session. Let's see a simplified example:
1 d4 tt::lf6 2 c4 g6 3 tt::lc3 Ji.g7 4 e4 d6 S tt::lf3
0-0 6 ii.e2 es 7 0-0 tt::la6 8 dS tt:Jcs 9 �c2 as 10
tt::lel tt::lfd7 (D)
Black is planning ...f5. If one's analysis were
based solely upon attacking Black's centre, one
might play the weak move 11 f4?, leading to
1 l ...exf4 12 ii.xf4 tt::le5, but then the g7-bishop
is not only a powerful piece but supports the
White has achieved the desired 'good knight outpost on e5 in front of White's backward
vs bad bishop' position, which enables him to pawn. So one can say that the g7-bishop 'pre
control play on both sides of the board. This vents' 11 f4 (and the idea of f4 generally). Or, in
point is the Nimzo-Indian Defence, in which

the first three moves all control light squares (1
d4 tbf6 2 c4 e6 3 tbc3 i..b4), and several main
lines continue with ...b6, ...i..b7 and ...tbe4
(strictly speaking this last move aims at dark
squares, although it 'plays on' light squares and
prepares another light-square move, ...f5). In
doubled-pawn variations such as 4 a3 i..xc3+ 5
bxc3, we might see Black play ... b6, ...i..a 6,
...tbc6-a5 and ...d5, which is truly playing on a
colour complex. The following game combines
complementary themes of backward pawns,
outposts, and playing on a colour complex.
a sequence such as 11 tbd3 f5 12 tbxc5 tbxc5, Taimanov - Karpov

White shouldn't play 13 exf5? !. He might do Moscow teams 1 973
this for the sake of avoiding Black's dangerous
attack that follows from 13 f3 f4. But 13 exf5? ! 1 d4 tbf6 2 c4 e6 3 tbc3 i.. b4 4 e3 c5
i.. xf5 14 'iVd l runs up against 14...e4!, when Karpov departs for a move from the light
the g7-bishop has gone from a passive onlooker square strategy but he will soon return to it.
to a major force. In this and similar positions, 5 i.. d3 0-0 6 tbf3 d5 7 0-0 dxc4 8 i..xc4
the dark-squared bishop serves as a prophylac cxd4 9 exd4
tic measure versus White's exf5, which might Now White has an isolated pawn on d4.
otherwise hamper Black's plans. I should add 9 b6 10 'iVe2 i..b7 1 1 .l:i.dl tbbd7 12 i..d2
.•.
that in some cases where Black replies to exf5 .l:i.cS 13 i.. a6?! (D)
with ... gxf5, that will also allow him to play a
favourable ...e4 and free his bad bishop. What's
the lesson? That a bad bishop can discourage
moves that would otherwise hurt his cause.
It doesn't take a fianchettoed bishop to fill
that role, of course. In the Closed Ruy Lopez
when White constructs a pawn-structure with
e4 and d5 and places his bishop on c2, one
might say that White's bad bishop on c2 has a
natural prophylactic effect against the move
...f5, because then exf5 brings the bishop into a
kingside attack. If Black has a pawn on c7 (with
the same piece placement), then the move ...c6
can be answered by dxc6 and i..b 3, taking over
the open a2-g8 diagonal. For these ideas see,
for example, the Breyer Defence or Zaitsev This is the key move to the early part of the
Variation in the Ruy Lopez (Chapter 8). opening. A colour complex takes on stronger
meaning when a bishop residing on the colour
Colour Complexes opposite that of the centre pawns (i.e., a good
bishop) is exchanged. Thus White risks losing
In a great number of openings, one player or control of the light squares.
both will concentrate his forces either largely or 13 ... i..xa6 14 'iVxa6 i..xc3 15 bxc3 (D)
exclusively on squares of one colour or the Now White has assumed a backward pawn
other. This is particularly logical in Black's on an open file (often the only way a backward
case because he doesn't have time to keep up pawn is defined), and Black has an outpost on
with White on both colour squares. One case in c4, in front of that pawn. Instead, 15 i..xc3? !
would put a very bad bishop on c3 whose po

tential to be freed by the move d5 is almost
non-existent, especially after Black places a
knight on the truly powerful outpost on d5.
The second key move. Karpov sacrifices a

pawn just to occupy the outpost and maintain a
blockade! Ripperger offers the insightful line
17 ...�b7 (protecting a7) 18 c4 l:i.fc8 19 .if4
.l:!.c6 20 h3 a6 21 �b3 b5 22 c5 llld5, when
The c3-pawn can be either weak or strong, Black has restrained White's d-pawn but at the
the latter depending upon two possibilities: cost of a protected passed pawn on c5. After 23
a) the c3-pawn does such a good job of sup .id6, the position looks about equal.
porting d4 that it allows White the time to orga 18 �xa7 �c6 19 �a3
nize a kingside or central attack; Black was threatening to trap the queen by
b) the pawn can advance to c4. ....l:!.a8.
Taimanov wants to pursue the latter idea, 19 ... .l:!.cS 20 h3 h6 21 .l:!.bl l:!.a4 22 �b3 llld5
counting upon the superiority of his bishop (D)
over Black's d7-knight (which incidentally Light-square domination! This is a particu
doesn't have many prospects right now be larly good illustration of favourable play on a
cause it is restricted by White's d4-pawn). A colour complex.
favourable change of structure might come
about, for instance, if White can play c4 fol
lowed by .ll b4. The problem is that Black
strikes first.
15....l:!.c7
Black protects the a-pawn and would like to
play ...�c8 followed by ...llld5. He has already
set his eyes on light-square weaknesses on c4,
a4, and possibly that on d5.
16 l:!.acl
White aims to make the move c4. Trans
forming a backward pawn into a hanging pawn
is more often than not a good idea. If there's no
real possibility of dynamic play, however, it's
usually easier to defend a pawn on the third
rank than on the fourth. The opening stage is over and Karpov has
16 ...�cS 17 �a4 (D) more than enough positional compensation for
17 �xc8? .l:!.fxc8 fixes the pawn permanently a pawn. The rest of the game is very accurately
until it can be won, which won't take long to played until the last moves before the time
happen. control and demonstrates the strength of the
17 .l:!.c4!
.•. blockade and associated outpost:
23 .l:i.dcl .l:i.c4 24 .l:i.b2 f6 25 .l:i.el @f7 26 'iid l Nevertheless, when I speak of a position in
lbf8 27 .l:i.b3 lbg6 28 'ii bl .l:i.a8 29 .l:i.e4 .l:i.ca4 30 which 'Black dominates the dark squares',
.l:i.b2 lbf8 31 'iid3 .l:i.c4 32 .l:i.el .l:i.a3 33 'iibl there's usually a persistence of structural weak
lbg6 34 .l:i.cl lbxc3 35 'iid3 lbe2+ 36 'iixe2 ness that I'm referring to. For instance:
.l:i.xcl+ 37 ..llx cl 'iixcl+ 38 @h2 (D)
Domination of a colour complex doesn't nec

38....l:i.xf3!? essarily mean a winning position but it probably
Certainly an intimidating move when there's constitutes a serious advantage, often compen
not much time left. Objectively 38 ...lbf4! would sation or more for the exchange, which was sac
have left Black with a large positional advan rificed in this example by capturing a knight on
tage. f3. White still has his dark-squared bishop, but
39 gxf3 lbh4 0-1 he has lost the dark squares anyway.
White should play on (perhaps he Jost on
time?) with 40 d5 !, although Black still has the Internal Weaknesses
advantage after, for example, 40...'iif4+ 4 1
@h l exd5 4 2 'iie3 'iif5. An important situation arises when one or
This game is typical in that the structure re both sides have 'internal weaknesses'. This
sulting from the opening is indicative of whether means that they have unoccupied squares on
players will be concentrating upon a certain their third or fourth ranks that cannot be de
colour throughout the game. fended by other pawns. Often these weak
nesses are somewhat masked by a pawn-front,
There are quite a few other openings with a but they can also be exposed when a pawn
lasting orientation towards playing on one col front disappears or breaks down. Generally,
our. Consider the main lines of the Dragon I'll refer to internal weaknesses in the centre
Variation of the Sicilian: Black's central pawns of the board, i.e. White's squares e4, d4, e3
are situated to control dark squares, and his and d3, or Black 's on e5, d5, e6 and d6. Weak
most active pieces control dark squares: the all nesses on the flank squares are normally of
important g7-bishop, the c6-knight, his queen less note, but those created by a pawn advance
on a5 or c7 (more often than not), and even the in front of one's king are a huge exception; for
c8-rook has its greedy eye on c3. Black 's f6- instance, an attack by f4-f5, g4-g5 and h4 can
knight has a tendency to go to d7 and augment create critical weaknesses on f4, g4, h4, f3, g3
control of the dark squares e5 and c5. Only the and h3. Players tend to be very careful about
queen's bishop doesn't participate, but it has in exposing their kings in such a fashion. In my
herent difficulties in that respect. White nor experience, less advanced players fail to rec
mally castles queenside, when Black's most ognize this type of weakness, especially if the
devastating attacks seem to land on the squares squares in question are not immediately at
c3 and b2. tacked or occupied.
A typical example of a complex of internal

weaknesses arises with an advanced centre. In
the last chapter we saw a King's Indian Four
Pawns Attack in which the front of the centre
collapsed and the internal weaknesses were ex
posed. It's worth taking the time to look back at
that example, especially the final diagram. The
weaknesses remain regardless of whether the
pawns that mask them disappear.
The following game is a classic between two
of the greatest players of all time:
Karpov - Kasparov
Moscow Wch ( 16) 1 985
Black is even ready to reinforce d3, which
1 e4 c5 2 tt::lf3 e6 3 d4 cxd4 4 tt::lxd4 tt::lc6 5 tt::lb5 hardly needs it. In nearly every critical variation
d6 6 c4 tt::lf6 7 tt::llc3 a6 8 tt::la3 d5!? analysed later it proved to be the difference.
A shocking gambit prepared by Kasparov The d4-square, which is also weak, isn't occu
for this match. pied by a piece until much later, but White's
9 cxd5 exd5 10 exd5 tt::lb4 1 1 .lle2 .llc 5?! 12 loss of control over it allowed Black to proceed
o-o? o-o 13 .tr3 .tr5 (D; without impediment.
24 .tg2 �f6 25 a3 a5 26 axb4 axb4 27 �a2
.\lg6 28 d6
The forward guard has to be sacrificed. White
is hopelessly tied up, the more so after Black's
next move.
28 ... g4 29 �d2 @g7 30 f3 �xd6 31 fxg4
�d4+ 32 @hl tt::lf6 33 .l:!.f4 tt::le4 34 �xd3 (D)
What does Black have for his pawn? Greater

activity, to be sure, and White's a3-knight is a
very poor piece, but most of all White has seri
ous internal weaknesses in his own camp, d4
and d3. They are both on the closed d-file yet
still of major importance.
14 .\lg5 .l:!.e8 15 �d2 b5 16 .l:!.adl tt::ld3 (D)
There it is. The d3-square has no protection White finally captures the knight that has
and the knight will radiate influence from its been on his own third rank for 1 8 moves ! But at
position almost until the end of the game. this point the damage has been done and it's
17 tt::labl h6 18 .llh4 b4 19 tt::la4 .id6 20 way too late to save the game.
.tg3 .l:!.c8 21 b3 g5! 34 tt::lf2+ 35 .l:!.xf2 .ixd3 36 .l:!.fd2 �e3 37
..•
More space. .l:!.xd3 .l:!.cl 38 tt::lb2 �f2 39 tt::ld2 .l:!.xdl+ 40

22 .txd6 �xd6 23 g3 tt::ld7 tt::lx dl .l:!.el+ 0-1
3 The Significance of Structure
A Sim ple Question:

Pawns or Pieces?
An inexperienced player, having struggled with

a number of opening sequences, might legiti
mately ask: "Is it more important at the begin
ning of a game to establish my position with
numerous pawn moves, or should I be develop
ing my pieces as quickly as possible?" This
question is not so easily answered, perhaps not
even by those more familiar with the game.
In chess history, new openings that don't
stake a claim to the centre have been regarded
with suspicion, and one of the first reactions is
to refute such openings with the construction of
a large centre, soon to be followed by its ad The related Modern Defence, 1 e4 g6 2 d4
vance. Thus the Alekhine Defence was chal .i.g7, was similarly met by 3 lbc3 d6 4 f4 or 3
lenged by 1 e4 lbf6 2 e5 lbd5 3 c4 lbb6 4 d4 d6 c3 d6 4 f4. Even in a uniquely positional open
5 f4, and the King's Indian Defence by 1 d4 ing such as the Benko Gambit, 1 d4 lbf6 2 c4 c5
lLif6 2 c4 g6 3 lbc3 �g7 4 e4 d6 5 f4 (D), each 3 d5 b5 4 cxb5 a6 you'll find quite a few early
called 'the Four Pawns Attack' in their particu games with bxa6 followed by f4 and e4 with the
lar opening. idea of e5. Likewise, when the English Defence
began to gain notice, attention was focused on
broad-pawn-front variations such as 1 c4 b6 2
d4 e6 3 e4 and 1 c4 b6 2 d4 �b7 3 lbc3 e6 4 e4.
B Recently the opening 1 d4 lbf6 2 c4 lbc6 (the
'Knights' Tango') has become respectable, but
it first had to be shown that the uninhibited ad
vance 3 d5 lbe5 4 e4 e6 5 f4 was not a threat to
the entire system. Returning to more conven
tional openings, it's easy to forget how often
early games with the Nimzo-Indian featured 1
d4 lbf6 2 c4 e6 3 lbc3 .i.b4 4 a3 �xc3+ 5 bxc3
followed by a set-up with e4 (e.g., 5...0-0 6 f3
c5 7 e4 with �d3, lbe2 and f4 to follow, estab
lishing a broad central front). Most of the varia
tions listed above are not bad, and some remain
The Modern Benoni faced the pawn on effective weapons to this day, but none are refu
slaught 1 d4 lLif6 2 c4 c5 3 d5 e6 4 lbc3 exd5 5 tations of the openings concerned.
cxd5 d6 6 e4 g6 7 f4 �g7 8 e5. In the early days After these impetuous attempts, attention
of the Pirc Defence, theory and practice con usually turned to a less ostentatious centre and
centrated primarily upon the Austrian Attack, quicker development. In the examples above,
i.e. 1 e4 d6 2 d4 lbf6 3 lbc3 g6 4 f4 (D), often we might find White playing, respectively, 1 e4
with an early e5. lLif6 2 e5 lbd5 3 d4 d6 4 lbf3 (versus the
THE SIGNIFICANCE OF STRUCTURE 37
Alekhine Defence) or 1 d4 lLif6 2 c4 g6 3 l2Jc3 But the distinction between a philosophy of

(or 3 lLif3 .i.g7 4 g3) 3... .i.g7 4 e4 d6 5 lLif3 (D) 'pawns-before-pieces' and one assigning equal
(versus the King's Indian Defence). priority to both has become increasingly more
subtle and context-dependent as time has gone
by. I already mentioned in Chapter 2 that when
the Open variations of the Sicilian were estab
lishing themselves in the first part of the 20th
century, there was a tendency on Black's part to
get his pieces out reasonably quickly. For in
stance, you would see 1 e4 c5 2 lLif3 lLic6 3 d4
cxd4 4 lLixd4 lLif6 5 l2Jc3 e6 and ... .i.b4 and/or
the freeing move ... d5 with rapid development.
Systems such as the Dragon Variation became
relatively popular; for example, 1 e4 c5 2 lLif3
d6 3 d4 cxd4 4 l2Jxd4 l2Jf6 5 l2Jc3 g6 6 .lle2 .i.g7
7 0-0 0-0 8 .i.e3 lLic6 (D).
There are also 1 d4 lLif6 2 c4 c5 3 d5 e6 4

l2Jc3 exd5 5 cxd5 d6 6 e4 g6 7 lLif3 (versus the
Benoni) and 1 e4 d6 2 d4 l2Jf6 3 lLic3 g6 4 lLif3
(D) (versus the Pirc Defence).
In that case four of Black's pieces are devel

oped within the first eight moves. We then often
see Black make several more piece moves before
touching another pawn (e.g., ... l2Jxd4, ....\le6,
.. .l:Ic8); this policy is clearly indicated by the ini
tial pawn-structure. In contemporary play, how
Today we see the more modest 1 d4 l2Jf6 2 c4 ever, we regularly see variations of the Sicilian
c5 3 d5 b5 4 cxb5 (4 lLif3) 4... a6 5 bxa6 and Defence in which the establishment of pawn
l2Jf3 followed by l2Jc3 and g3 (versus the Benko structure swamps rapid development, not least
Gambit); 1 c4 b6 2 d4 e6 3 l2Jc3 .i.b7 4 a3 or 4 of which is the most popular Sicilian system of
lLif3 (versus the English Defence); 1 d4 l2Jf6 2 them all, the Najdorf Variation: 1 e4 c5 2 lLif3 d6
c4 l2Jc6 3 lLif3 e6 4 l2Jc3 (versus the Knights' 3 d4 cxd4 4 lLixd4 lLif6 5 l2Jc3 a6 to be followed
Tango); and 1 d4 lLif6 2 c4 e6 3 l2Jc3 .i.b4 4 by more pawn moves such as ... e6 and ... b5.
'iVc2 (versus the Nimzo-Indian Defence). Even in the list of 'balanced' variations that I
Of course these are just a few examples, and gave two paragraphs back, things will shift dra
many other main-line pawn-structures support matically in one direction or another while still
fast piece development. In these variations the in the opening stage. In the King's Indian exam
pieces and pawns seem to be in mutual support ple, everything follows the harmonious model in
and one might easily conclude that this is the the main line 1 d4 l2Jf6 2 c4 g6 3 l2Jc3 .\lg7 4 e4
ideal situation. d6 5 lLif3 0-0 6 .lle2 e5 7 0-0 lLic6 8 d5 lLie7 (D).
should be reframed: which takes precedence in

any given position, pawn-structure or piece
w development? How should we organize our
thoughts so as to optimize our understanding?
In the examples of openings above, and in the
vast majority of opening variations in this book,
the pawn-structure is in fact the determinant of
appropriate piece placement and not the other
way around. The structure sets the overall pa
rameters of development, such that there may
be many ways to bring the pieces out but their
effectiveness (or lack thereof) depends upon
pawn configuration. That relationship is true
whether or not you throw all of your pawns for
All well and fine, with a nice balance be ward to begin the game, or only a couple of
tween piece moves and pawn advances. But in them; thus it lends itself to a more useful view of
this position Black will customarily embark opening play. Relevant questions now emerge:
upon a massive pawn advance that, in its deter is my centre breaking down before I can com
mined neglect of piece development, would put plete the development with which it was sup
a Four Pawns attacker to shame. As you may posed to assist? Am I creating weaknesses and
know, Black plays ...lbd7 first, and then that ad targets of attack for my opponent? Is there any
vance typically consists of ... f5-f4, ... h5 and way, given the pawn-structure before me (or
... g5-g4 and is frequently accompanied by un the one I am about to construct) that I can ar
developing moves such as ... �f8 and ...lbg8. In range all my pieces on useful squares where
a large number of lines, Black's a8-rook and c8- they don't interfere with each other? Given that
bishop will remain in their places until moves my pieces won't be able to reach their desirable
20 to 25 or even longer. So the initial moves of squares in time, can I change the structure so as
an opening are not always indicative of its bal to make their deployment timely and useful? In
ance between pawn moves and development. other words, the pawns usually determine the
Naturally there are times in which early piece harmony or lack thereof in your potential piece
development and related events will dictate configurations.
what structure becomes appropriate, but not of In addition, there is the crucial relationship
ten. between pawn-structure, which we tend to think
Furthermore, pawn-structures have primacy of in static terms, and dynamics. In a sense every
in terms of the weaknesses they create, which attack depends upon the structure the attacker
determine both where the opponent can attack inherits, but that is not a very useful disclosure.
and what squares he can usefully occupy. For What counts is whether we can associate iden
crucial periods of time, pawns block the devel tifiable dynamic elements with known struc
opment of pieces, or open lines for them. tures. The result may be compared with happily
Whether freeing moves are even available to recognizing an old friend (resulting in a combi
activate passive pieces is largely dependent nation or tactic that one can easily assess), or
upon pawn-structure. Pieces have only second running into vaguely familiar but enigmatic
ary roles in these areas of consuming interest companions (when combinative success may
for the player. depend upon intuition). Ultimately, of course,
So the obvious answer to the query in the the most brilliant and original attacks (and de
first paragraph, namely, that 'you should both fensive miracles) have their own capricious
advance pawns and develop at the same time, in character that can't be anticipated from previ
a mutually supportive manner' , is simplistic ous knowledge. In fact, the most awe-inspiring
and wanting in content. What's more impor combinations are precisely those that 'shouldn't'
tant, such a statement doesn't serve as helpful work within a particular structural context, and
advice for most players. I think that the question 'shouldn't' work given the pieces and pawns
available for action. Nevertheless, the majority I solated Pawns

of attacks will be informed by describable cate
gories of positions. We saw some broad characterizations of posi
Thus the precedence of pawn-structure, and tional features in the last chapter. Now I want to
the motivation for this chapter. It is generally look at the structural elements across the board
agreed now that pattern recognition and the that bear upon the opening stage. We'll begin
ability to process patterns in context is the with the fairly straightforward case of the iso
foremost determinant in chess strength (put lated pawn, also called the 'isolani', which we
ting aside competitive factors). The number of defined in the Chapter 2. Textbooks almost al
patterns one can recognize and associate with ways concentrate upon the isolated d-pawn,
other structures correlates to how well one un also called the 'isolated queen's pawn' (abbre
derstands and plays the game. Grandmasters viated as 'IQP'). Most authors do so to the ex
store and process many more pawn-structures clusion of isolated pawns anywhere else on the
with accompanying piece placements than the board, writing chapters and even whole books
average player does, if only because of their on this specific case. Granted, it's very impor
repeated exposure to them in preparation and tant to give the IQP its due because it can arise
over the board. With study alone it's possible from so many openings, and so early in the
for one to master a great number of standard game. Why is that? To generate an IQP in the
opening positions in the same way, and to un opening, it's generally necessary to have the
derstand their interaction with the subsequent moves d4 and ...d5 appear early on, and it's ex
play. Appreciation of why a strategy works in tremely likely that one or both of the moves c4
one position but not in a similar position is an or ...c5 were also played in the first stage of the
indispensable part of chess mastery. Further opening. To show this, let's take a list of several
more, if you recognize ideas and manoeuvres openings that lead to the same, well-known
from other openings that apply to the one that type of isolated queen's pawn position, and
you are playing, it will help you to focus on the sometimes to the very same position:
issues and inspire you to make better deci Queen's Gambit Accepted: 1 d4 d5 2 c4
sions. dxc4 3 CLJf3 CLJf6 4 e3 e6 5 .ll xc4 c5 6 0-0 cxd4 7
How might we improve our knowledge of exd4 .lle7 8 CLJc3 0-0 9 .l:!.e l .
pawn-structures? Obviously it's not possible Nimzo-Indian: 1 d 4 CLJf6 2 c4 e6 3 CLJc3 .llb4
to list them all and memorize their unique fea 4 e3 c5 5 CLJf3 0-0 6 .id3 cxd4 7 exd4 d5 8 0-0
tures. But there are formations and related is dxc4 9 .ll xc4 CLJc6 10 .ll g5 .lle7 11 .l:!.el .
sues that repeat themselves from opening to Alapin Sicilian: 1 e4 c5 2 c3 d5 3 exd5 'iVxd5
opening, very often constituting the basis for 4 d4 CLJf6 5 CLJf3 e6 6 .id3 CLJc6 7 0-0 cxd4 8
the fundamental strategy of each. In this chap cxd4 .lle7 9 CLJc3 °iVd8 10 .l:!.el 0-0.
ter I'll examine some pawn-structures and the Caro-Kann: 1 e4 c6 2 d4 d5 3 exd5 cxd5 4 c4
issues associated with them, choosing selected CLJf6 5 CLJc3 e6 6 CLJf3 .lle7 (6 ....ll b4 7 cxd5
areas most likely to impact one's understand CLJxd5 8 .id2 .lle7 9 .\ld3 CLJc6 10 0-0 0-0 1 1
ing of the game, or at least to grasp the com .l:!.el CLJf6 12 .ll g5 would be a typical transposi
mon elements of the opening. These are not tion; White can also play 8 'iVc2 followed by 9
strange or irregular formations; one idea is to .id3) 7 cxd5 CLJxd5 8 .\ld3 (or 8 .llc4 0-0 9 0-0
show how one might use the same approach to CLJc6 10 .l:!.el CLJf6) 8 ...0-0 9 0-0.
study other, more complex, structures. Hope Semi-Tarrasch: 1 d4 d5 2 c4 e6 3 CLJc3 CLJf6 4
fully their usefulness will extend to players of CLJf3 .lle7 5 cxd5 CLJxd5 6 e3 0-0 7 .\ld3 c5 8 0-0
a wide range of skills. This is not a middle cxd4 9 exd4 CLJc6 10 .l:!.el .
game book, however, and my main goal has This is the basic picture (see diagram over
been to make the discussion in the forthcom leaf) :
ing openings section more readily comprehen The most significant difference among these
sible. When presenting individual variations openings is the position of White's light-squared
and games, I'll often assume your familiarity bishop (it's on c4 or d3). Sometimes the queen
with this chapter. is already placed upon c2 or e2, and the king's
Advantages of the isolated d-pawn:
1. White will be able to develop more easily

and aggressively, having more space and open
lines for his bishops.
2. The IQP creates a support-point for a
knight (or other piece) on e5 .
3. The threat of the d-pawn's advance ties
Black's pieces to the defence of d5.
4. Black, with less space, will have difficulty
developing actively without making some con
cession such as creating a weakness or ceding
the bishop-pair.
5. White has good kingside attacking chances
rook is usually but not always moved to e l . The based upon the support-point on e5, the e-file,
basic position and its variants have been played and his bishops aimed in that direction.
thousands of times and investigated in depth. In
fact, more words have been written about the In terms of strategy, White will have several
IQP than about any other specific positional ways of proceeding. He will usually complete
feature in chess. Neither side can be said to his development by putting his queen on e2 or
stand inherently better, which is why both sides d3 (less frequently c2 or b3) and queen's rook
are willing to enter into these positions. With on d l . Then one of the first goals is to provoke a
out going into detail, here are the basic struc weakness on the kingside. To do this, he can
tural properties and strategies that should be play lbe5 and swing a rook to the kingside via
stressed. For ease of discussion, let's assume e3 . Or he can line up his bishop and queen to
that White is the possessor of the IQP before we create a threat on h7. Black will generally de
attend to specific examples. fend by keeping a knight on f6 and playing . . . g6
if necessary. With that set-up White can attack
Disadvantages of the isolated d-pawn: the dark squares by i..h6, work to soften up the
kingside by h4-h5 and/or play for d5, often by
1. The IQP is a relatively easier target than bringing his bishop back to the a2-g8 diagonal.
most pawns because it can only be protected The safe advance of the d-pawn to d5 beto
by pieces, several of which may be required kens success in most cases because it opens
for the task (as opposed to needing only a sin lines or broadens potential uses for almost all of
gle pawn) . Also, the d-pawn is almost always his pieces (notably, the rooks on d l and e l ,
on an open file potentially facing Black's rooks bishop on a2 and knight on c3) and breaks
and/or queen. down the defender at e6; it also liquidates the
2. Defence of the isolated d-pawn can tie isolani itself. After d5 White usually has the far
down White's pieces which might be used more superior pieces, and he often has tactical re
effectively elsewhere. sources that win material. The d5 break is prob
3. Black gains an influential outpost in front ably the most frequently successful plan. There
of the isolani, which means that it is very diffi are also set-ups with the moves l:.el -e3-g3; or,
cult to drive his pieces off that spot. more commonly, lbe5, ..llc4 and J:.el , intending
4. The IQP tends to be a more serious weak tactics such as lbxf7, particularly if Black's
ness in simplified positions, the more so in an rook is on e8 . These ideas and others only work
endgame. Notice that the mutual possession of because White's superiority in space permits
the open c-file increases the chances of simpli him to transfer his pieces rapidly, make threats,
fication. Nevertheless, Black must be skilful to and take Black out of his game plan. The more
make the right kind of simplification that doesn't pieces with which to attack, the better.
come with other disadvantages. Often a new Black's strategy is not excessively compli
equilibrium will result from exchanges. cated, although implementing it may be. His
first goal is to maintain the blockade on d5, usu Spassky Avtonomov

-
ally with a knight. Simply leaving a knight there Leningrad 1949

is often not enough, however, because White
may be able to capture the piece at a point 1 d4 d5 2 c4 dxc4 3 tt::lf3 tt::lf6 4 e3 c5 5 .ltxc4 e6
where . . . exd5 is forced, eliminating the threat 6 0-0 a6 7 'l'ie2 b5 8 .ltb3 tt::lc6 9 tt::lc3 cxd4? 10
to d4 and sometimes transforming the pawn �dl .ltb7 1 1 exd4 tt::lb4
structure in White' s favour. Thus, whether oc The d5-square is protected by four pieces
cupied or not, d5 itself needs to be reinforced. and a pawn.
Often Black's knights will go to f6 and d5 (via 12 d5! (D)
...tt::lb4-d5) or to d5 and e7. His c8-bishop will Anyway ! Can this be sound?
be developed to b7, either by . . . b6 or by ... a6
and . . . b5. A rook on d8 can also act to support a
piece on d5 or restrain White' s pawn advance
to d5. One of Black's goals is simplification:
the more pieces that are exchanged the less
likely it is that White can break through. More
over, the closer that Black can get to an endgame
the better his prospects usually are. Exchanging
White' s minor pieces is a high priority, because
they can have considerable range from squares
around the isolated pawn. Knights in particu
lar are dangerous when posted on e4, e5 and
c5; and even seemingly ' defensive' knights on
c3 and f3 can quickly come into action. Ex
changing White' s light-squared bishop is a real
coup for Black; whether on c4, d3, a2 or c2, it is 12 tt::lbxdS
••.
the piece most likely to be involved in a direct You can confirm that after 1 2 ...tt::lfxd5 1 3 a3 !
attack. By contrast, a rook on d l defending the and 1 2 ... .ltxd5 1 3 .ltg5 ! .lte7 14 .ltxf6 gxf6 1 5
isolani is much less likely to do any damage. a 3 White will win material.
For all that, simplification can be double 13 .ltgS! .lte7 14 .ltxf6 gxf6 15 tt::lxd5 .ltxd5
edged because sometimes it clarifies White's 16 .ltxd5 exdS 17 tt::ld4! Wf8 18 tt::irs (D)
attacking themes, especially if he has support
points along open files in conjunction with pawn
advances. A wonderful illustration of this is
seen in Chapter 5 on the Giuoco Piano (in the B
main line with 1 0...tt::lc e7).
All that is rather abstract, so here are some
examples of strategy by both sides. There are
literally thousands of isolated-pawn positions
in games between masters, many of which can
be found in books on the opening or middle
game. As indicated, these positions will be taken
from openings in which an IQP situation is nor
mally created (for instance, in the same open
ings listed above). What you will eventually
find is that isolated pawns are formed in a wide
range of positions, many of them appearing af Many a pawn or exchange has been sacri
ter the opening stage because of an exchange ficed to bring a knight to f5 . Here it's worth
on d4 or d5. more than a rook.
Here is a brief lesson about the main danger 18 ... hS 19 �xdS 'l'Wxd5 20 'l'ixe7+ Wg8 21
posed by the d-pawn: its advance. 'l'ixf6 1-0
This next game is not as easy to understand, There's the thematic break.
but expresses the same theme. 19 tlla5
.•.
We're still in the opening ! 19 . . . exd5 20

Yusupov - Lobron .l:.xe7 ! is a tactical device to remember, while
Nussloch 1996 19 . . . tllxd5 20 i.xd5 ! ? i.xg5 2 1 tllxg5 exd5 22
.l:.xe8+ .l:.xe8 23 tllxd5 �e5 24 �f3 f5 25 �b3
1 d4 tllf6 2 c4 e6 3 tllc3 i.b4 4 e3 0-0 5 i.d3 d5 is another typical idea. Now we see a not-so
6 tllf3 c5 7 0-0 cxd4 8 exd4 dxc4 9 i.xc4 b6 1 0 typical one:
.l:.el i.b7 1 1 i.d3 tllc6 12 a3 i.e7 (D) 20 dxe6! tllxb3
Capturing the queen by 20... .l:.xd3 loses to 2 1
exf7+ 'it>g7 (2 1 . . .'it>h8 2 2 .l:.xd3 tllxb3 2 3 .l:.xe7 !
.l:.xe7 24 i.xf6# is pretty) 22 fxe8� �xe8 23
.l:.xd3 tllxb3 24 .l:.de3 ! and White wins.
21 exf7+ @xf7 22 �c4+ @g7 23 tlle 5! tllg8
24 .l:.xd8 �xd8 25 �f7+ @h8 26 �xb3 �d4
27 .l:.e3! .:rs 28 i.xe7 1-0
28 ... tllxe7 29 ctJf7+ 'it>g7 30 .l:.xe7 follows.
The next example is a model treatment from

Black's viewpoint:
Korchnoi - Karpov
Merana Wch (9) 1 981
All the moves thus far are customary ones. 1 c4 e6 2 tllc3 d5 3 d4 i.e7 4 tllf3 tllf6 5 i.g5
13 i.c2 .l:.e8 14 �d3 g6! h6 6 i.h4 0-0 7 .l:.cl dxc4 8 e3 c5 9 i.xc4 cxd4
An instructive combination goes 14 ....l:.c8? 15 10 exd4
d5 ! exd5 16 i.g5 (threatening i.xf6) 16 ... tlle4 The isolated queen's pawn arises.
( l 6 . . . g6? 1 7 .l:.xe7 ! �xe7 1 8 tllxd5) 1 7 tllxe4 10...tllc6 11 0-0 ctJh5! (D)
dxe4 1 8 �xe4 g6 19 �h4 �c7 20 i.b3 h5 2 1
�e4 (threatening �xg6+) 2 1 . . .'it>g7 2 2 i.xf7 !
@xf7 23 i.h6! �d6 24 �c4+ @f6 25 .l:.ad l
ctJd4 26 �xd4+ �xd4 27 .l:.xd4 .l:.c5 28 h4 ! 1 -0
Petrosian-Balashov, USSR 1 974.
15 h4 �d6 16 i.g5 .l:.ad8 17 .l:.adl �b8
Unmasking the rook against White's d-pawn.
18 i.b3 a6? 19 d5! (D)
Black's goal is simplification, to draw the

sting out of White's attacking chances. The
knight went to h5 so that the bishop couldn't es
cape capture by going to g3. It also looks at f4.
12 i.xe7 tllxe7
The knight covers the key square d5 .
13 i.b3
1 3 �el would be the usual idea: get all the the d-pawn and c-pawn is that White has more
pieces out. On the other hand, with a pair of space; not so obvious is that a third-rank pawn
pieces off and more to come, the customary d5 is easier to defend than a fourth-rank pawn ! In
advance will only lead to liquidation, and prob this instance the pawn on c6 prevents White' s
ably not one that White would be happy with; isolated pawn from advancing while maintain
for instance, 1 3 d5? ! exd5 14 lllxd5 lllxd5 1 5 ing an outpost on d5 and the options of ... �d6,
J..xd5 ( 1 5 �xd5 �xd5 16 J..xd5 lllf4 1 7 J..c4 ... �d8 and ... lllf5 . Black also has a useful b-file
J..e6 ! runs into the same kind of problems as 1 5 that is typical of this structure. A lot comes
J..xd5) 1 5 . . .lllf4 1 6 J..c4 �xdl 1 7 .Ufxdl J..g4 down to activity here; for instance, will a white
and Black already stands slightly better. This is rook on the outpost c5, with the possible help of
based more on the specifics of this position than a bishop on a4, make up for Black' s pressure on
a statement about the move d5, however. The the d-pawn? Probably not, but those are the
h5-knight happens to serve a powerful function kinds of competing factors that arise. More on
due to the possibility of ...lllf4. Usually a move the isolated c-pawn will follow in the examples
like 1 3 d5 would lead to equality, which is still a below.
success for Black in opening play. Incidentally, after 19 �xc6, 19 ...lllxc6 20 d5
13...lllf6 exd5 21 J..xd5 is at best equal for Black, be
Again protecting the crucial d5-square. cause bishop versus knight with pawns on both
14 llleS sides of the board is usually difficult for the side
White does the right thing by occupying the with the knight.
support-point. 19 �d6 20 g3 �d8 21 �dl l:tb6!
..•
14 J..d 7!
•.• The opening is over and Black has restrained
The normal continuation 14 ... b6 followed by the pawn, while White has no outposts or at
. . . J..b7 would only be tempting sacrificial ideas tack. Thus Black has the advantage. From this
on f7, as described above. point on Karpov plays one of the best technical
lS �e2 �c8 16 llle 4!? games in world championship history.
More simplification. But ... J..c6 was coming 22 �el �d7 23 �cd3 l:td6 24 �e4 �c6 2S
anyway. �f4 llldS 26 �d2 �b6 27 .ltxdS l:txdS 28 �b3
16 lllxe4 17 �xe4 .ltc6 18 lllxc6 �xc6 19
•.• �c6 29 �c3 �d7 30 f4 b6 31 �b4 bS 32 a4
l:i.c3 bxa4 33 �a3 as 34 �xa4 �bS 3S �d2 eS 36
Take a look at 19 �xc6 bxc6! (D). fxeS l:txeS 37 �al �e8 38 dxeS �xd2 39 l:i.xaS
�c6 40 �a8+ Wh7 41 �bl+ g6 42 �fl �cs+
43 'it>hl �dS+ 0-1
In the next game, two younger superstars

present a different approach to the same type of
position:
Kramnik - Anand
Dortmund 2001
1 d4 dS 2 c4 dxc4 3 lllf3 e6 4 e3 lllf6 S .ltxc4 cS

6 0-0 a6 7 J..b3 cxd4 8 exd4 lllc6 9 lllc3 .lte7
10 .ltgS 0-0 (D)
Pretty much the same position that we're
used to.
This is our first example of what is a recur 1 1 �d2!?
ring type of position in the openings world. This is a somewhat different way of deploy
Black takes on an isolated c-pawn at the same ing White's forces. Kramnik has �f4-h4 in
time as White has an isolated d-pawn. In the mind.
general case, the obvious difference between 1 1 ...lllaS 12 J..c2 bS 13 �f4 �a7
An incredible defence! It can't quite save

Black, but everything else loses; for example,
17 . . . g6 1 8 "!J.fel or 1 7 . . . h6 18 i.xh6 gxh6 19
�xh6, with tllg5 and "!J.d3 to follow next.
Black plans . . ."!J.c7 or if possible . . ."!J.d7, to

stop d5 .
14 "!J.adl i.b7
Since 14 . . ."!J.d7 allows llle5, Anand wants to
play . . . i.xf3 and then . . ."!J.d7 with at least equal
ity. 18 "!J.fel
15 d5! (D) 1 8 ctJd4 ! ? would be a typical tactical idea,
Again, this sacrifice is intended to cut off looking hungrily at the f5-square.
Black's pieces and free White's own. 18 tllc6 19 g4!? �d6! 20 gxh5 �b4! 21
•••
h6!
Black has miraculously averted mate, but
now a queenless middlegame ensues in which
White's attack persists for another 10 moves.
Notice the knight getting access to the key f5-
square; as Kasparov has shown, this tends to
win almost by itself!
21. �xh4 22 tllxh4 tlle4 23 hxg7 "!J.c8 24
••
i.xe7 tllxe7 25 i.xe4 dxe4 26 "!J.xe4 'it>xg7 27

"!J.d6! JJ.c5 28 "!J.g4+ @h7 29 ctJf3! tllg6 30
tllg5+ 'it>g7 31 tllxf7 "!J.xf7 32 "!J.dxg6+ '.t>h7 33
"!J.6g5 "!J.xg5 34 l:rxg5 l:rc7 35 a3 b4 36 axb4
l:tcl + 37 'it>g2 "!J.bl 38 l:ra5 "!J.xb2 39 "!J.a4! 1-0
Lautier - Karpov
15 i.xd5!
.•. Monte Carlo (rapid) 1995
From here on Anand defends in heroic fash
ion. After 1 5 . . . exd5 1 6 �h4 g6 1 7 "!J.fel White 1 d4 tllf6 2 c4 e6 3 tllc3 i.b4 4 �c2 0-0 5 a3
threatens a killing "!J.xe7, and he wins after i.xc3+ 6 �xc3 b6 7 i.g5 i.b7 8 f3 d5 9 e3
1 5 . . .tllxd5 16 lllxd5 i.xd5 1 7 l:rxd5 ! due to ctJbd7 10 cxd5 exd5 1 1 i.d3 l:re8 12 llle2 c5 13
1 7 . . . exd5 1 8 i.xh7+ '.t>xh7 19 �h4+ 'it>g8 20 0-0 �e7 14 tllg3 l:tac8 15 i.f5 cxd4 16 �xd4
i.xe7 �xe7 2 1 tll g5 . A pretty combination, There's the isolani; Black really doesn't seem
perhaps the one that Anand missed when he al ready for it.
lowed White to play 1 5 d5 . 16 "!J.c4 17 �d2 tllc5 18 "!J.adl h6 19 i.xf6
•..
16 tllxd5 exd5! �xf6 20 i.bl

Again, not 16 ... tllx d5? 17 "!J.xd5 ! exd5 1 8 Threatening i.a2. White has the better bishop
i.xh7+, etc. and is restraining the IQP.
17 �h4 h5! ! (D) 20 tlle6 21 i.a2 JJ.c5
.•.
Lateral defence of the isolani is best if you But we also have instances of IQPs on
can maintain the rook's position. That often ap Black's side of the board that look somewhat
plies to the endgame as well. different:
22 llle2 i.a6! 23 l:i.fel i.xe2 24 l:txe2 �d8 French Defence: 1 e4 e6 2 d4 d5 3 llld 2 c5 4
2S �d3 g6 26 �ed2 (D) exd5 exd5 5 lllgf3 (or 5 .ltb5+ .ltd7 6 i.xd7+
lllxd7 7 lllgf3 lllgf6 8 0-0 .lte7 9 dxc5 lllx c5)
5 ... lllc6 6 .ltb5 .ltd6 7 dxc5 .ltxc5 8 0-0 llle7 9
lllb3 .ltd6 (D).
Hasn't Black merely simplified into a rotten

position?
26...d4!
His d-pawn is weak so Karpov finds a clever
way to liquidate it. Tarrasch Queen's Gambit: 1 d4 d5 2 c4 e6 3
27 .ltxe6 �xe6 28 exd4 l:tcdS lllc 3 c5 4 cxd5 exd5 5 lllf3 lllc6 6 g3 ( 6 e3 lllf6
Black is a full pawn down but now it's White 7 i.e2 cxd4 8 lllxd4 would be analogous to our
with the IQP, and he can't break down the examples from the white side) 6 ... lllf6 7 .ltg2
blockade ! .lte7 8 0-0 0-0 9 .ltg5 cxd4 10 lllxd4 (D).
29 �e4 �f6 30 Wf2 lt>g7 31 �d3 as 32 a4
bS 33 b3? bxa4 34 bxa4 �c6
Hitting c2 and a4. Suddenly White's got
some problems.
3S �a3? �d6!
Black is attacking both the important pawn
on h2 and the rook on a3 !
36 �e3 �xh2 37 f4 �h4+ 38 lt>gl �f6 39
�ed3 hS 40 �e3 h4 41 �e4 �8d6 42 �ld2
�fS 43 �f3 �e6 44 �d3 �xf4 4S dS �al+ 46
Wh2 �xf3 47 gxf3 �es+ 0-1
Remember that Black can also take on the

isolated queen's pawn. In fact, every d-pawn
opening above has some kind of reversed case,
but particularly the Semi-Tarrasch, which can Roughly the same ideas apply to handling
arise from a number of openings; e.g., 1 c4 lllf6 these openings: White should maintain close
2 lllc3 c5 3 lllf3 e6 4 e3 d5 5 cxd5 exd5 6 d4 control of d4 and seek carefelly-chosen ex-
lllc6 7 .lte2 i.e7 8 dxc5 .ltxc5 9 0-0 0-0, or 1 d4 changes. As mentioned above, he may be better
d5 2 lllf3 lllf6 3 c4 e6 4 lllc 3 c5 5 e3 lllc6 6 off exchanging minor pieces than rooks, be
cxd5 exd5 7 .ltb5 .ltd6 8 dxc5 .ltxc5 9 0-0 a6 1 0 cause rooks tend to be passive pieces as defend
.lte2 0-0 and s o forth. ers. The side with the isolani should follow the
reverse approach, exchanging rooks (if any Isolated c-Pawns

thing has to be exchanged) and keeping minor
pieces on the board. That's getting into the Isolated c-pawns are very common and we
realm of middlegame theory, however. At any shall see them frequently throughout this book.
rate, activity is at a premium: rooks on open They may arise a little later in the game than in
files, bishops attacking weak points, etc. And the standard isolated d-pawn openings, partly
of course if you can safely get . . . d4 in, your because they can easily stem from them. The
odds of a happy conclusion increase. Sicilian Defence offers some examples:
Sicilian Defence, Alapin Variation: I e4 c5 2
Isolated e-Pawns c3 ctJf6 3 e5 ctJd5 4 d4 cxd4 5 cxd4 d6 6 ctJf3
tllc6 7 i.c4 ctJb6 8 i.b5 dxe5 9 tllxe5 i.d7 1 0
The IQP isn't the only isolated pawn of interest i.xc6 i.xc6 1 I tllxc6 bxc6 (D).
in chess openings. First, we might ask why we
don't see more isolated e-pawns in the opening.
That's fairly easy: at some point an f-pawn
would have to advance and that's not part of
most openings, especially since there would
have to be another central capture at some point.
However, in the Sicilian Defence we do see a
situation that is rare in other openings, i.e. the
pawn-structure often leads to isolated e- and d
pawns on adjacent files. There are a great num
ber of lines like 1 e4 c5 2 tllf3 d6 3 d4 cxd4 4
ctJxd4 ctJf6 5 tllc 3 a6 6 i.e2 e5 7 ctJb3 i.e7 8
0-0 0-0 9 @h 1 ctJbd7 1 0 i.e3 "V/i/c7 1 1 f4 exf4 1 2
i.xf4 (D) involving the routine moves . . .e5, f4
and . . . exf4.
Sicilian Defence, Rossolimo Variation: 1 e4

c5 2 ctJf3 tllc6 3 i.b5 e6 4 c3 tllge7 5 d4 cxd4 6
B cxd4 d5 7 exd5 tllxd5 8 0-0 i.e7 9 tlle5 °VJifb6 1 0
i.xc6+ bxc6 (D) .
Sometimes Black has his pawn on e6 and

the advance f4-f5 can lead to the same struc
ture, that is, if White responds to . . .exf5 by
capturing with a piece (usually a knight, i.e.
tllx f5), or Black does the same after White's
fxe6 (for instance, by ... i.xe6). The character
istics of those positions are fairly consistent Here are some other examples:
and will be discussed in Chapter 1 I on the Si Queen's Gambit Declined: 1 d4 d5 2 c4 e6 3
cilian Defence. ctJc3 i.e7 4 ctJf3 ctJf6 5 i.g5 h6 6 i.h4 ctJbd7 7
e 3 0-0 8 �c l c 6 9 .ltd3 dxc4 1 0 .ltxc4 b 5 1 1 In the Sicilian Defence, the same capture
i.d3 a6 1 2 a4 bxa4 1 3 lllxa4. happens but White's b-pawn is in better shape,
Catalan: 1 d4 d5 2 c4 e6 3 lllf3 lllf6 4 g3 .lte7 at least theoretically, because it has the c-pawn
5 i.g2 0-0 6 0-0 dxc4 7 llle5 lllc6 8 .ltxc6 bxc6 in its vicinity. A different way for 'b-pawn ver
9 lllxc6 �e8 10 lllxe7+ �xe7 1 1 �a4 e5 1 2 sus a-pawn' to arise is in a position with a white
dxe5 �xe5 1 3 �xc4, a position that has been pawn on a3. Black plays . . . b5-b4, the b-pawn is
played repeatedly over decades. captured by the a-pawn, and a piece recaptures
Two Knights Defence: 1 e4 e5 2 lllf3 lllc6 3 on b4. Then Black's a-pawn is left i solated, and
.ltc4 lllf6 4 lllg5 d5 5 exd5 llla5 6 .ltb5+ c6 7 often White's b-pawn as well. This can occur in
dxc6 bxc6 8 .lte2. the Sicilian Defence, French Defence, King's
Semi-Slav: 1 d4 d5 2 c4 e6 3 lllc 3 c6 4 lllf3 Indian Defence, or other openings featuring a
lllf6 5 .ltg5 h6 6 .ltxf6 �xf6 7 e3 llld7 8 .ltd3 minority attack. Finally, it sometimes happens
dxc4 9 .ltxc4 g6 1 0 0-0 .ltg7 1 1 e4 e5 12 d5 that with Black's pawn on b5 and White's on
lllb6 1 3 .ltb3 .ltg4 14 h3 .ltxf3 1 5 �xf3 �xf3 a4, Black will play simply . . .bxa4, a common
1 6 gxf3 We7 1 7 dxc6 bxc6. idea in the Ruy Lopez and Sicilian Defence
Isolated c-pawns are often created in the (likewise with Black's pawn on b4 capturing
middlegame. For the most part we won' t see White's on a3).
that in this book, but the same concepts apply. Because of their distance from the centre of
action, isolated and even doubled a-pawns are
Isolated a-Pawns seldom worthwhile targets in the opening. Their
vulnerability shows itself more in the endgame.
Few isolated b-pawns arise in the opening, but Certain structures lend themselves to a-pawn
isolated a-pawns are quite common, because raids; e.g., ... �a5(+) and . . .�xa2 in the Ex
their creation requires only that a b-pawn cap change Gri.infeld Defence and certain Queen' s
tures towards the centre. One recurrent situa Gambit Exchange Variations; or, for instance,
tion arises in a number of openings when White when Black goes out of his way to capture
plays a4-a5 against Black's pawns on a6 and White's a4-pawn in the Winawer Variation.
b7. This is a 'one pawn holds two' situation in But usually isolated a-pawns situated on the
the sense that if Black plays . . . b5 (or sometimes first two ranks (such as a black pawn on a6 in
. . . b6), then White captures en passant and iso several openings) tend to be defensible until
lates Black' s a-pawn. the middlegame is in full swing. For example,
sometimes White captures a knight on a6 with
his light-squared bishop and the same issues
arise; for example, 1 d4 d5 2 c4 c6 3 lllf3 lllf6 4
lllc 3 dxc4 5 a4 llla6 6 e4 i.g4 7 .ltxc4 e6 8
.ltxa6 bxa6 9 �d3 i.xf3 1 0 gxf3 (now we have
two sets of doubled pawns; Black's are weaker,
of course, but he is compensated by the b-file
and a potentially safer kingside) 10 . . . a5 1 1 �c4
�c8 1 2 .l:.gl �b8 1 3 .l:.g5 �b4 1 4 �e2 .l:.xd4 1 5
�xa5 �c7 1 6 i.e3 �xa5 1 7 .ltxd4 with ap
proximate equality, Korchnoi-Conquest, Buda
pest 1 996.
The treatment of all these phenomena varies
so much from position to position that we'll
have to discuss them in context.
This type of situation occurs repeatedly in the
King's Indian Defence and Benoni, for example, Pawn -Chains
but watch for it in other openings. In many cases
White's c-pawn will be on c4 or off the board, so When authors give examples of pawn-chains
his b-pawn will be isolated or backward. they tend to be pawns adjacent to and facing
another pawn-chain, i.e. interlocking. The text in which the majority of its main variations
book example is the French Defence Advance have pawn-chains.
Variation, 1 e4 e6 2 d4 d5 3 e5 c5 4 ctJf3 tllc6 5 1 e4 e6 2 d4 d5 3 e5
c3. The line of pawns from b2 to e5 is called a The logic behind this move for White is that it
'chain' , and the directly interlocking pawns are claims space on the kingside and cramps the de
on e6 and d5, but of course Black' s pawn on f7 velopment of Black's pieces. After 3 e5, Black's
holds up the ones on e6 and d5 . Most books on king's knight cannot go to its 'best' square on
strategy discuss this French Advance Variation f6, and Black's queen's bishop, which was al
when they want an example of pawn-chains, ready blocked by its pawn on e6, is further in
and also the main lines of the King' s Indian De carcerated by the inability of the e6-pawn to
fence. Those are excellent starting-points. We advance. As mentioned, a variation well-suited
don' t always think in terms of pawn-chains for a discussion of chains continues:
even if they share classical properties, for in 3 c5 4 c3 tllc6 5 tllf3 (D)
...
stance, in the Slav with 1 d4 d5 2 c4 c6 3 ctJf3

ctJf6 4 e3 a6 5 c5, in which White' s pawn-chain
is lengthy indeed. But in fact . . . e5 is the natural
way to attack that chain, and of late we've even
seen the arduous b4, a4 and b5 by White to at
tack the base of Black' s pawn-chain at c6 (this
has occurred a bit more often in the line that
goes 4 tllc3 a6 5 c5 followed by i.f4, but that' s
another matter).
Furthermore, much of what relates to those
pawn-chains is relevant to a great number of
other 'pawn strings' that aren't fully or di
rectly opposed by other pawns. In accordance
with some other sources, I ' ll call these pawn
chains as well. For example, if you look at the
Modem Benoni ( 1 d4 ctJf6 2 c4 c5 3 d5 e6 4 The last two moves are natural in that 4 . . .tllc6
tllc 3 exd5 5 cxd5 d6 6 e4, especially with develops and exerts influence upon d4 and e5,
6 . . . g6 7 f3) you see short 'pawn duos' pointing whereas 5 tllf3 defends those points. Note first
in opposite directions. In several openings that if White had played 5 dxc5 he would have
only partially overlapping chains emerge but broken the chain, which would have weakened
have chain-like properties; for example, things the front of the pawn-structure at e5 . That pawn
like c3/d4/e5 versus f7/e6 and e4/d5 versus would then be subject to a greater threat of
c7/d6, and so forth. We see a truncated chain in capture, like an isolated pawn which can't be
some Ruy Lopez variations, when White plays defended with other pawns. It also could be
d5, thus forming an opposition of e4/d5 ver exchanged more easily due to insufficient re
sus c7/d6/e5). Furthermore, pawn-chains with sources for maintaining it. A direct attack could
doubled pawns at their base will emerge from come by the moves . . .'filc7 and . . . tllge7-g6. Or
exchanges. Almost all of these can be looked the offer to exchange could be pursued via the
at in the same terms as the traditional French pawn move ... f6.
and King ' s Indian chains ; for example, in This leads to the idea that if Black can break
methods of attacking and defending them. down the d4 point, sometimes called the 'base'
Study of their common and contrasting ele of the pawn-chain, he can cripple or destroy the
ments will help you to master this part of the pawn-structure itself. To what end? By getting
game. rid of the pawn on d4 and then winning or ex
changing the one on e5, a natural place would
Let's start with the traditional examples and appear on f6 for the knight currently doing
see what we can discover. We'll start out with nothing on g8, and the move . . . e5 would be
the French Defence, probably the only opening more feasible. With a little luck that advance
would lead to the liberation of the c8-bishop,

and in the meantime B lack would control the
action with his own 'ideal centre' of pawns on
e5 and d5 . This particular fantasy, for the mo
ment out of reach without White' s cooperation,
motivates Black's desire to break down the
chain at its base. As it turns out, locating the
base of a pawn-chain is more of a practical than
a theoretical determination; if Black played
. . . b5-b4, then White's pawn on c3 would be
called the base of the chain, and in the unlikely
event that Black played . . . a5-a4-a3, then b2
would be so designated. Essentially it comes
down to where one is most likely to succeed in
undermining the chain. those structures that can be attacked following
Returning to the French Defence and its 'ef this rule.
fective' base at d4, we can see why White is in For instance, few if any players refer to the
terested in maintaining his pawn there rather lines of pawns from f7 to d5 and f2 to d4 in the
than playing dxc5 or allowing it to be captured. Queen's Gambit Declined as pawn-chains, even
The two sides' conflicting goals might be played when White plays c5 (as Steinitz used to do
out by a variety of means. An example of the without provocation ! ) . For example, 1 d4 d5 2
further play is: c4 e6 3 lllc 3 .lte7 4 lllf3 lllf6 5 i.g5 h6 6 .lth4
...5 �b6 lllbd7 7 e3 0-0 8 �c l a6 9 c5 c6 (D).
Black attacks d4 again ; for the moment the
pawn is adequately protected.
6 .lte2
This develops pieces and prepares to castle.
Another theme can arise if White plays 6 i.d3
.ltd7 ? ! (6 . . .cxd4 is normal) 7 dxc5 .ltxc5 8 0-0,
when White gives up his supporting pawn but
in return gains the possibility of b4-b5, when he
can use the d4-square as an excellent support
point for his pieces.
...6 cxd4 7 cxd4 lllge7
Already White has to think about the health
of his base, the d4-pawn. If he plays the most
natural move on the board, 8 0-0?, that pawn is
unavoidably lost after 8 . . . lllf5.
Obviously White would not play 8 0-0? but Thinking in terms of pawn-chains isn' t our
would'instead protect the pawn by, say, 8 b3 habit in this case, because the traditional idea of
lllf5 9 .ltb2 (D). how to break up a chain, that is, at its 'base' ,
These moves are not necessarily the best, but doesn't apply. After 1 0 .ltd3, it's normally not
they illustrate the basic idea. I've avoided a dis on the cards for White to play b4-b5 (he's turn
cussion of move-order subtleties in order to get ing his eyes towards the king, a less trivial tar
the point across without unnecessary complica get). Black can attack in the centre by . . . e5
tions. (hardly with the idea of putting pressure on d4,
The concept of attacking the base, first system however) or attack the front of the chain by
atized by Nimzowitsch, rapidly spread through 1 0. . . b6 1 1 cxb6 c5 ! ? , a sound idea although
out the chess world and was treated as sort of a subject to tactical issues.
general principle of pawn-chains. It's interest What is the reality? Even in the French De
ing that what are labelled chains are precisely fence example above, the standard illustration
of attacking the base, Black will end up by at 1 e4 e6 2 d4 d5 3 ctJd2 tllf6 4 e5 tllfd7 5 i.d3
tacking the protected front of the chain. For ex c5 6 c3 tllc6 7 tlle2 cxd4 8 cxd4 f6
ample, after White successfully protects his The e5-pawn is attacked three times.
base by 8 b3 lllf5 9 i.b2, Black ' s next step is to 9 exf6 tllxf6 10 ctJf3 i.d6 1 1 0-0 (D)
attack the front of the pawn-chain by . . .f6. For
instance, one line goes 9 . . . i.b4+ 1 0 @fl i.e7
1 1 tllc3 0-0 12 g3 f6 (D) and White will soon
surrender the leading pawn by exf6.
In this instance Black made only a half

hearted attempt to attack the base of the pawn
chain and then successfully attacked the front
of it.
Or, in the same Advance Variation, the phe So perhaps the procedure should be to attack
nomenon is illustrated in the variation 3 e5 c5 4 the base and then the front? But then there' s the
c3 tllc6 5 tllf3 i.d7 6 i.e2 tllge7 7 0-0 tllg6 following unadulterated example of attacking
with the intention of . . .f6 next or soon thereaf only the front of the chain:
ter; for example, 8 tlla3 i.e7 9 lllc2 (White is 1 e4 e6 2 d4 d5 3 lll d2 tllc6 4 tllgf3 tllf6 5 e5
still concerned with protecting the base at d4) llld7 6 i.d3 f6 (D)
9 . . . 0-0 1 0 .l:.el cxd4 1 1 cxd4 f6 (D).
I should note that according to theory Black

Now the front of the pawn-chain disappears stands perfectly well in this position. Other
because of the three-way attack: 1 1 exf6 i.xf6. openings attack chains in this manner; for ex
This time Black ignored the base and came out ample, English Opening variations in which
fine. Black plays . . . e4 and White eliminates the front
Other French pawn-chain variations are pawn by f3 . There are also Ruy Lopez varia
clearer in that respect; e.g.: tions in which d5 is met by . . . c6 (e.g., the
Breyer Defence), and several King' s Indian

variations as well.
Clearly we need a broader way of looking at

this subject. Let's go to the King's Indian De
fence example that's always used in the books:
1 d4 tt:Jf6 2 c4 g6 3 lLlc3 J..g7 4 e4 d6 5 lLlf3
0-0 6 J..e2 e5 7 0-0 tt:Jc6 8 d5 tt:Je7 (D)
great attack as shown by one game that contin

ued as follows:
20 hxg3 fxg3 21 J..xg3 lLlh5 22 J..h2 J..e7 23
tt:Jbl J..d7 24 'iVel J..g5 25 lLld2 J..e3+
B)
9 ltJd2 (D)
A very different approach emerges from this
move in the same variation.
We'll quickly look at two very distinct ap
proaches to this position, but only in terms of
pawn-chains.
A: 9 lLlel
B: 9 lLld2
A)
9 lLlel tt:Jd7 10 J..e3 f5 1 1 f3 f4
Black ignores the first 'effective' base at e4,
the one that he attacked in the French Defence
situation. Indeed, 1 l .. .fxe4 1 2 fxe4 �xfl + 1 3
i.xfl lLlf6 1 4 J..f2 only helps White because
Black has no kingside targets to bite upon. By
playing . . .f4 instead, he extends the chain to f3
in preparation for the march of his g-pawn.
These are all normal moves, details of which In some games the traditional pawn race en-
will be given in the chapter on the King's Indian sues:
Defence in the next volume. 9...tt:Je8 10 b4 f5
12 J..f2 g5 Attack on the base.
This pawn is headed for the new base at f3. 11 c5
13 tt:Jd3 tt:Jf6 14 c5 lLlg6 15 �cl �f7 16 �c2 Likewise.
i.f8 17 cxd6 cxd6 18 °iVd2 g4 19 �fcl g3 (D) 11 tt:Jf6 12 f3 f4 13 tt:Jc4 g5 14 i.a3 ltJg6 15
.•.
So Black never did attack the base on e4 or b5 lLle8

on f3, neither of which was ever seriously White is threatening the base at d6 three
threatened. In fact, the pawn attack ran right times, so Black has to defend it.
by the chain with . . . g3 and puts no pressure 16 b6! (D)
whatsoever on it! But in spite of the g2-d5 A nice picture ! White transfers the base all
chain surviving in full health, Black has a the way down to Black's second rank, the
seen in the French Defence examples, there are

benefits to that approach.
9...a5 (D)
ultimate undermining theme. This pure form of

attacking the base of such a long chain almost
never occurs in any opening.
16 axb6 17 cxb6 cxb6 18 \\!Vb3 h5 19 .l:iabl
.•.
g4 (D) First Black defends against b4, in tum pre

venting the key move c5.
10 a3 ..lld 7 11 b3
1 1 .l:ibl would be answered by l 1 . . .a4 ! (two
pawns holding down one, a theme that pops up
periodically through this book) 1 2 b4 axb3 1 3
.l:ixb3 b6 and White will never get c5 in. After
11 b3, however, White is ready for .l:ibl , b4 and
c5.
l l ...c6 (D)
Now if only Black could play . . .h4-h3 , he

could duplicate White's achievement!
20 ttJxb6
Having destroyed the very back of the pawn
chain, White has a very good position, though
must be careful that the tactics don't get out of
control.
This example illustrates how important it is,
in a game with pawn-chains, to have at least one
file open for a rook to work with in a direct way A strike against the front of the pawn-chain.
next to the pawn-chain. Other pieces alone usu The first point is that the leader of the chain on
ally can't completely break down the oppo d5 will now be vulnerable if White plays c5.
nent's position. 12 .l:ibl \\!Vb8!? 13 b4 cxd5
Since White's pawn-chain is so impervious Sometimes Black skips this move and an
to assault in the foregoing variation, Black can swers b4 with . . . b5, a dynamic attack on the en
think about challenging the front of the pawn tire chain, which is at least interesting if not
chain, even when it's protected to the hilt. As entirely convincing.
14 cxd5 the crucial difference is that this is the side of

14 exd5 gives Black a type of kingside ma the board where Black' s king resides, so such
jority that we shall see more of as we proceed. a plan is unrealistic. A simple analysis (with a
. . .f5 will follow shortly. Suffice it to say that in little bit of calculation) also tells you an attack
general that situation is favourable to Black. on d4 won ' t get very far: not enough pieces
14 Jks 15 .ib2 axb4 16 axb4
•• and plenty of defenders. But if you're thinking
The pawn-chain has been neutralized, prov in terms of pawn-chain experience, you'll see
ing that Black needn't only play on the side of that Black should play to undermine White' s
the board where he has the undermining moves. pawn-structure b y 9 b5 ! followed b y . . . b4
•••
The same applies to White. Chess is not so and moves such as . . . .l:Ib8, . . . bxc3, . . . .ia6 and
one-dimensional that you aren't permitted to . . . lLlb6-a4 in some intelligent order. This can
think about more than one theme, at least not in be an effective idea as long as Black is alert to
the opening where we have so many pieces on the defence of his king.
the board.
With those ideas in mind, let's look at exam
What's the upshot of all this? Is the practical ples from the Caro-Kann Advance Variation.
player left without any guidance whatsoever?
Not at all, because the more positions you see Anand - Karpov
and play, the more tools that you acquire. As in Wijk aan Zee 2003
any other situation in chess, you have to make
an assessment of which positions call for which 1 e4 c6 2 d4 d5 3 e5 .if5 4 lLlc3 e6 5 g4 .ig6 6
treatment. For instance, notice that Black ad tt:Jge2 tt:Je7 7 f4 (D)
dressed the front pawn on the queenside and
never attended to the e4 base. How realistic is
that in general? Let's imagine a similar position
of a type that does arise in the French Defence: B
1 e4 e6 2 d4 d5 3 lLld2 tt:Jf6 4 e5 lLlfd7 5 f4 c5
6 c3 tt:Jc6 7 lLldf3 .ie7 8 g3 'iVa5 9 Wf2 (D)
7 c5!
•.•
As we saw above, once f4 is in, it's less likely

that 7 .. .f6 ? ! will do any good. White simply
shores up the centre by 8 .ie3, when 8 .. .fxe5 9
fxe5 gives White f4 for his knight. I should add
that in some lines in which White plays h4-h5
Here White has safeguarded his king (it can instead of f4, . . .f6 is the best defence.
even go to g2 if necessary) and his pieces are 8 lLlg3!? cxd4 9 lLlb5 lLlec6!
about to spring out to aggressive positions; A piece sacrifice to win the centre.
e.g., ..ild3, lLle2, with perhaps g4 and f5. How 10 f5 ..tc5 11 tt:Jd6+
likely is that plan to succeed? The structure i s Black's point is that after 1 1 fxg6 fxg6 he
analogous (d4/e5/f4 t o c4/d5/e4), s o Black's picks up a second pawn and threatens the total
procedure would have to do with . . . f6, perhaps decimation of White's centre by . . .lLlxe5. Then
preceded by . . . h5 , connected with . . . g5 . But 1 2 'i¥e2 0-0 prepares . . . tt:Jd7 winning the last
centre pawn, and then 1 3 g5 (to get ..llh 3 in)

runs into 1 3 . . . 'ifb6! 14 ..llh3 d3 ! 1 5 �xd3 ( 1 5
cxd3 \\!Vxb5) 1 5 . . . lllxe5 and everything falls
apart. Notice how this was a consequence of
. . . c5 and . . .cxd4, although by no means a neces
sary one, and in fact later games improved for
White before this point in the game.
11 ... ..llxd6 12 exd6 �xd6 13 .ig2
1 3 fxg6? ! fxg6 ! is strong (Black has the open
f-file, a big centre and three pawns for the
piece). In fact, ...fxg6 is usually the correct an
swer in the French and similar structures. Hav
ing said that, even 1 3 . . . hxg6 ! ? sets up the rogue
tactic 14 .ig2? .l:ixh2 ! .
13 f6
•.. outside the pawn-chain. Notice, however, that
Now Black threatens to escape with the Black lost a tempo by playing . . . c6-c5, and that
bishop. he' s made an extra move with his light-squared
14 fxg6 hxg6 15 0-0 bishop, which doesn't happen in the French.
White steers clear of . . . .l:ixh2 again. The point is that White is getting extra time to
15 ... llld7 16 .l:if2 0-0-0 17 c3 dxc3 18 bxc3 consolidate his space advantage and Black
ctJb6! (D) needs to break down the centre in some way be
fore he becomes permanently cramped. Thus:
7...�b6 8 \\!Va4!
This move would be worse than useless in
the Exchange French because Black would play
. . . ..td7.
8 ... c4!?
There springs up another pawn-chain ! This
takes all the pressure off White's base while
forming a new one. The plan is slow (and un
usual) but there are special considerations. First,
Black has to look at lines like 8 . . .cxd4 9 ctJxd4 !
intending ..lle 3 next, with lllxf5 another prom
ising idea; e.g., 9 . . . .ic5 1 0 lllxf5 exf5 1 1 b4
..lle7 1 2 ..lle 3 \\!Vd8 1 3 .l:idl . This is almost im
possible to prevent without real compromise;
Although Black has only two pawns for the for example, a pretty line runs 8 . . . lllh6 9 dxc5
piece, he more than makes up for it with the .ixc5 10 b4 .ie7 1 1 ..lle 3 �c7 1 2 b5 and here
mobile centre, c4 outpost and kingside attack. 1 2 . . .lllb8 1 3 b6+ �d7 1 4 �xa7 ! or 1 2 . . .�a5 1 3
White went on to win, but not because of the ..lld l ! �xa4 1 4 ..llxa4 llla5 1 5 b6+. There are
opening. Attacking the base was the correct de many other lines with tactical and positional
cision. problems. So Seirawan reasons that he'll keep
the position closed for a while, and by the time
Short - Seirawan White organizes g4 and f5 he'll be winning on
Tilburg 1990 the other side of the board.
9 ctJbd2?!
I e4 c6 2 d4 d5 3 e5 .if5 4 ..lle2 e6 5 lllf3 c5 6 What do we know about such positions? The
0-0 lllc6 7 c3 (D) base of the enemy pawn-chain is far, far away,
In this Caro-Kann Advance Variation we so it's not hard to see that the head must be at
have the equivalent of the French Advance tacked. Short knows this of course, but his tim
Variation but with Black's light-squared bishop ing is bad. 9 b3 ! is a good move, hitting the
vulnerable part of the chain, when Black would shorter chains, or ones with outposts? How to
cave in if he were to play 9 . . . cxb3 10 axb3 and assess them? The Benoni complex shows us a
activate all of White ' s pieces; e.g., 1 0 ... tt:Jge7 little variety. In the Czech Benoni it' s fairly
1 1 ..ta3 lLlg6 1 2 ..txf8 .l:txf8 1 3 .tbs a6 1 4 easy to see the nature of the pawn-chains:
tt:Jbd2 and Black is short of good moves. 1 d4 lLlf6 2 c4 c5 3 d5 d6 4 lLlc3 e5 5 e4 (D)
9 'iVa5 10 °iVdl h6! 11 �el b5
•••
Back to pawn-chain operations ! . . .b4 is next,

so White tries to do something about it.
12 b4?! (D)
Black would like to play for . . .f5 or ... b5,

White for f4 or b4. In practice, White's breaks
are more likely to succeed because of Black's
lack of space or good squares for his pieces;
12 tt:Jxb4! 13 cxb4 .llxb4
..• e.g., he lacks c5 for his knights, or anywhere ac
As in the last game, Black has two pawns for tive for his king's bishop (which is sometimes
the piece and the promise of much more after reduced to the exotic idea . . . h6 and . . . i.e7-g5).
. . . i.c3. Short finds an interesting reply in the In particular Black has trouble enforcing .. .f5 if
midst of these threats. White sets up a structure involving ..ild3, ltJf3
14 tt:Jfl !? tt:Je7 and h3.
Capturing the rook by 14 . . . i.xe l 15 lLlxe l Notice that the same pawn-structure in the
allows White to survive the pawn-rush. King's Indian Main Line is more bearable for
15 tt:Jg3 i.g6 16 �fl Black because with his bishop better-placed, he
Here instead of 1 6 ... i.c3?, as he played, Seir can get counterplay with . . .f5 before White
awan gives 1 6 . . . lLlc6 1 7 ii.e3 i.c3 1 8 �c l b4 1 9 squelches it.
lLlh4 i.h7 2 0 lLlh5 �g8 2 1 ..ilxh6 0-0-0 2 2 i.g5 A Benko Gambit pawn-chain analysis re
�d7 23 ltJf3 'i¥xa2 with an unstoppable mass of veals a little about the gambit's strengths. After
pawns. 1 d4 lLlf6 2 c4 c5 3 d5 b5 4 cxb5 a6 5 bxa6 fol
When faced with a long-term space prob lowed by Black's recapture of the pawn over the
lem, like the one that Short created for his op next moves, White is very seldom able to en
ponent, waiting around is the worst thing to force an attack at the effective base of Black's
do. Look at whether attacking the base or front pawn-chain at d6, and can only dream of
of the pawn-chain has any chance of succeed achieving a successful b4 (it does happen, but
ing, then whether the two in combination can only rarely) . Black on the other hand has al
be effective. If not, you must create your own ready eliminated the base of White' s pawn
counterchances by hook or by crook, which chain on c4, and the move . . . e6, cracking up the
structurally may amount to a radical advance front pawn at d5, characterizes most Benko
of your own. Gambit variations at one point or another.
Take the Alekhine Defence, which actually
Practically every opening system has its includes a lot of pawn-chains. Here's the Four
pawn-chain examples. What about some other Pawns Attack, producing a partial chain after 1
e4 CLJf6 2 e5 CLJd5 3 d4 d6 4 c4 CLJb6 5 f4 dxe5 6 23 CL:lg5 Wh8 24 \\!Vf7 ..lld 7 25 ..llxe6 l:.xd8
fxe5. Where to attack? Let' s see: 6 . . .CLJc6 7 26 \\!Vg8+! .l:ixg8 27 CLJf7# (1-0)
..lle 3 ..llf5 8 CL:lc3 e6 9 CLJf3 ..lle7 10 ..lle2 0-0 1 1 Hopefully this section will give you a feel for
0-0 f6 ! . In front, that' s the best plan ! These con what's happening when we encounter cases of
cepts cement themselves with study and experi pawn-chains in other openings throughout this
ence. Here' s a recent high-powered example: book.
Grishchuk - Ponomariov Dou bled Pawns and Related Pawn

Torshavn 2000
Captures
1 e4 CLJf6 2 e5 CLJd5 3 d4 d6 4 CLJf3 g6 5 ..llc4 Understanding of doubled pawns is essential to
CLJb6 6 ..llb3 ..llg7 7 a4 a5 8 CL:lg5 e6 9 f4 (D) playing openings and eventually mastering
This is a solid chain that must be taken care them. As above, I'll approach this subject with
of quickly, or Black must find counterplay else some standard examples and then try to intro
where, which is no easy task. duce some more complicated ideas for you to
chew over. Other structures will be discussed in
conjunction with individual openings.
Doubled pawns are a recurring motif in the
Nimzo-Indian Defence. After 1 d4 CL:lf6 2 c4 e6
3 CL:lc3 ..llb4, capturing the c3-knight produces
doubled pawns, whose structure is such that
the forward c-pawn is particularly vulnerable.
Without getting into the jargon, you can see that
a structure with pawns on c4, c3 and d3 is more
secure than one with pawns on c4, c3 and d4. In
the former case each pawn can be protected by
another, whereas in the latter the c4-pawn is un
supported. Here's a game with several thematic
ideas in a typical Nimzo-Indian:
9 dxe5 10 fxe5 c5
..• Geller - Smyslov
Base of the chain. USSR Ch (Moscow) 1949
1 1 c3 cxd4 12 0-0 0-0 13 cxd4 CL:lc6 14 CLJf3
f6 1 d4 CL:lf6 2 c4 e6 3 CL:lc3 ..llb4 4 a3 ..llxc3+ 5
Front of the chain. bxc3
15 CL:lc3! fxe5 White now has doubled pawns on c3 and c4.
It's a little late to tum around. The forward pawn is the target; note that if
16 ..llg 5! \\!Vd7 17 dxe5 CL:lxe5? White's d-pawn were on d3, his doubled pawns
The best chance is 17 . . . \\!Vxdl ! 1 8 .l:iaxdl would be protected.
CL:lxe5 19 CL:lxe5 l:.xfl + 20 Wxfl i.xe5 21 i.e3 5 CL:lc6 6 f3
••.
..llx c3 22 .l:td8+ Wf7 23 bxc3 and the bishops Having secured the advantage of the two
are worth more than a pawn, but Black can at bishops in compensation for his doubled pawns
least hope for survival. However, he should at c3 and c4, White wants to build a large centre
avoid 23 . . . CLJd7 24 Wel ! b6 25 l:.h8 Wg7? 26 and use his extra space to help in a kingside at
..llxe6 ! Wxh8 27 i.d4+. tack. The kingside is a particularly good target
18 CL:lxe5 .l:ixfl+ 19 \\!Vxfl \\!Vd4+ because Black's dark-squared bishop has been
Or 1 9 . . . ..llxe5 20 l:tdl ..lld4+ 2 1 Wh l . exchanged and can't guard vulnerable squares
20 Whl \\!Vxe5 21 ..lld8 \\!Vc5 22 CL:le4 \\!Vb4 around the king.
Black is also dead in the water following 6 b6 7 e4 ..lla6
.•.
22 . . .\\!Vf5 23 i.xb6 \\!Vxe4 24 l:.dl h5 25 .l:id8+ Black is taking aim at White' s weak c4-
Wh7 26 \\!Vf7. pawn.
8 ..llg5 2 1 lLlxc6? °iVd7.

And White begins to drift to the right. 21...h5! 22 'iVc2 lLle7 (D)
8 h6 9 ..th4 tt:Ja5
•••
There are more examples of this structure in

Volume 2.
10 'iVa4 'iVc8! 11 lLlh3 lLlh7?!
Better is 1 1 .. .°iVb7 ! 1 2 ..ild3 'iVc6 ! (D).
Summing up: every important light square is

covered by Black, whose knight will be in a
dominating position on c4, with another knight
coming to f5. To make things worse for White,
his dark-squared bishop is bad and his rooks are
By this means the c-pawn would have fallen, inactive. This is all the logical result of the open
although Black's advantage might not be enough ing, and of 1 5 . . . d5 in particular. After many ups
for a win after 1 3 'iVxc6 dxc6 1 4 e5 lLld7 1 5 and downs, the game was eventually drawn, but
Wf2 ii.xc4 1 6 .llc 2. Capturing the c4-pawn di Black has a winning position at this point.
rectly is one theme; what happens in the game
is related. Next, a classic game that illustrates typical
12 ..td3 0-0 13 e5 .l:Ie8 14 0-0 tt:Jf8!? 15 pros and cons of doubled pawns.
tt:Jf4!?
White should always maximize his kingside Portisch - Fischer
play in such positions and not worry much Sousse IZ 1967
about a pawn or two on the queenside. Thus 1 5
f4 ! d5 1 6 f5 was called for, attacking the pawn 1 lLlf3 tt:Jf6 2 g3 g6 3 c4 ..tg7 4 d4 0-0 5 .llg2 d6
chain. Notice that White's attack benefits greatly 6 tt:Jc3 tt:Jbd7 7 0-0 e5 8 e4 c6 9 h3 °iVb6 10 .l:tel
from the lack of Black's dark-squared bishop, Oddly enough there's an important main line
which was exchanged off on the fourth move. of this same variation that involves doubled
15 d5!
•.• pawns: 1 0 c5 ! ? dxc5 1 1 dxe5 lLle8 1 2 e6 ! fxe6
Black may not win the c-pawn but he wins 1 3 lLlg5 lLle5 ( 1 3 . . .lLlc7 ! ?) 14 f4 ltJf7 1 5 lLlxf7
the light squares. This is often the result of ..td4+ 1 6 Wh2 �xf7 (D).
fighting against doubled pawns: the squares This has arisen in several games. Black is left
they are on become more important than the with doubled c-pawns (resulting from a capture
pawns themselves. 1 5 . . . g5? isn't worth it after away from the centre), as discussed below, and
1 6 ltJh5 gxh4 1 7 tt:Jf6+. a masked isolated pawn to boot. Given his extra
16 cxd5 ..txd3 17 lLlxd3 exd5 18 f4 ltJg6! pawn and reasonable piece placement, how
Smyslov anticipates the idea of . . . lDe7-f5. ever, the position is about equal.
19 ..tg3 °iVf5! 10 ... �eS 11 d5 lLlc5 12 �bl a5
Now Black begins a series of moves de All conventional moves so far, except that
signed to conquer an entire colour-complex. Black' s . . . �e8 leaves him a tempo down com
This was discussed in Chapter 2. pared to some similar variations. Black nor
20 lLlb4 c6! 21 .ttael mally plays . . ...td7 and . . .cxd5 to cover a4. This
important central squares, both the black out

post on d4 and most importantly White ' s d5,
which is protected from intrusions. On the other
hand, Black' s dark-squared bishop has very lit
tle scope, so the advantage of two bishops is not
yet a factor, and he has no pawn-breaks other
than . . .f5, which White can keep under control.
16 CL:la4
The forward doubled pawn is usually the
more vulnerable one. Here White has no pros
pects of attacking it along an open file because
of his own c4-pawn, but he can focus pieces on
it in order to tie Black's pieces to its defe nce.
When a player's doubled pawn can be pro
variation is generally another good illustration tected by adjacent pawns then his pieces need
of how Black can play on the queenside in not be diverted to defend it. That's why iso
King's Indian Defence. lated doubled pawns are so much worse than
13 ..lle3 \\!Vc7!? 14 ..llxc5!? connected ones, assuming that other factors
The exchange o n c 5 to get doubled pawns aren' t at work.
can occur in many, many distinct positions of 16 ..trs 17 \\!Vb3 CL:lhs 18 \\!Ve3 \\!Va7
.•.
the King's Indian. White has to decide whether Black's pieces are passive and now White
to give up his best bishop in order to cripple could try to transfer his f3-knight to b3, but if
Black's pawn-structure. He usually declines the necessary Black can bring his knight to e6 or
bargain. Here, however, he's a little ahead in d7. What Portisch does instead is quite clever.
time and goes for it. 19 h4!
14 dxcS 15 dxc6 bxc6 (D)
.•. This has the obvious idea of ..llh 3, trying to
The first point is that 1 5 . . . \\!Vxc6 would give exchange his bad bishop for Black's good one
White a huge and favourable outpost on d5, one at c8. Bu.t White also sees that Black's best plan
that might be reinforced by CLJd2-fl -e3. is the manoeuvre . . .CL:lg7-e6-d4, which will leave
his kingside less defended against the moves h5
and hxg6.
19 CL:lg7 20 @h2 f6 21 ..llh3 ..llxh3 22 @xh3
.•.
CLJe6 23 hS (D)
What are the main characteristics of the po

sition? Black's doubled pawns are isolated, and
what's more he has an isolated pawn on a5.
We've already mentioned, however, that iso
lated a-pawns are usually not serious weak 23...gxhS!?
nesses until the endgame. What is typical about Black takes on yet another set of isolated
the doubled c-pawns is that they control very doubled pawns! And he gives up the valuable
square f5. But Fischer realizes that he will be l::tg6+ and l::txg5 with a third pawn and play
able to cause trouble down the g-file in con against Black' s exposed king. The game ends
junction with ... lLld4. Normally 23 ...lLld4 would with true equality.
solve all of B lack ' s difficulties but the opening 43 g4! 44 .l:txg7+ 'i/ixg7 45 'ilie8+ 'it>h7 46
.•.
of the h-file would cause a few problems after 'ilih5+ 'it>g8 1/z-1/z
24 hxg6 hxg6 25 l::th l .
24 l::th l .l:Iad8 25 'it>g2 'i/ig7 26 'it>fl The subject of doubled pawns is boundless
26 .l:Ixh5?? loses to 26 ... lLlf4+. but especially for the sake of opening investiga
26 'ilig4 27 l::th4 'ilig6 28 'ilie2 ..th6 29 b3
••• tion we can narrow our focus considerably and
l::td7 look at cases that significantly influence practi
29 . . ...llg5 30 lLlxg5 fxg5 3 1 l::t xh5 .l:If8 was cal play. Specifically, doubled c-pawns arise
also suggested, as in the game. Black is doing more often than any other type and they deter
fine in any case. mine the nature of the play in many of those
30 .l:Idl l::tx dl+ 31 'ilixdl l::td 8 32 'ilie2 ..llg5 games. For the sake of clarity I'll concentrate
33 lLlxg5 fxg5 34 .l:Ixh5 l::td 2! 35 'i/ig4! on them, with a brief look first at a particular
35 'ilixd2 'ilixh5 threatens . . . tt:Jd4. White's central situation.
knight has served a good function but now
looks out of play. Doubled Centre Pawns
35 h6 36 l::t h2 'it>g7 37 tt:Jc3 l::td3 38 lLldl!?
.••
White is ready t o take up an outpost by Doubled centre pawns arise much less often in
lLle3-f5. the opening than doubled c-pawns. They are
38...'ilif7 39 'it>g2 generally produced by exchanges of minor
But he never gets a chance. At this point 39 pieces on the third or fourth rank, and usually
tt:Je3?? loses to 39 . . . l::tx e3. don't allow of the choice of recaptures that we
39 'ilid7! 40 'i/if5 (D)
.•• saw above. Their effects on the position tend to
White decides to bail out. 40 lLle3 l::txe3 4 1 be ambiguous.
'iff5 ! (not 4 1 fxe3?? lLlf4+ 42 'it>f3 'ilidl +) i s an
other way to do so.
This is a position from Chapter 6 on the Two

Knights Defence; similar situations can arise
40...l::txdl 41 'ilixe5+ 'it>g8? from a number of 1 e4 e5 openings. White plays
Perhaps Fischer was trying to win, but this ..lle3 to challenge the enemy bishop on c5 (like
gives White a real attack. Black had a draw by wise with colours reversed, of course). Cap
4 1 . . .'it>g6 42 'ilif5+ 'it>g7 43 'ife5+ with perpet turing that bishop on e3 will help White to gain
ual check. central control (in particular of d4, which was a
42 l::txh6 ltJg7 43 l::tg6? potential support-point for Black's knight), and
An error in tum. Good winning chances were he will have the open f-file to work with. But
to be had by 43 'ilixc5, or by 43 'ilib8+ lLle8 44 the resulting centre e3/e4/d3 is generally not
mobile. What does that mean? After the ex ..lle6 ! ? 7 ..llxe6 fxe6 shows that it' s possible to
change on e3, White' s pawns are initially well adopt such pawns in that situation. The result
protected; it's usually difficult to get at the single ing position is one that current theory indi
weakness at e3. However, if White plays d4 cates is equal. See Chapter 7 on the Philidor
thereafter, the forward e-pawn will be unpro Defence, in particular the discussion of early
tected by another pawn and therefore vulnera move-orders.
ble, just as the c4-pawn was in the Nimzo-Indian It bears repeating that the exchange of queens
example above. And if the d-pawn advances fur by no means betokens entrance into an end
ther to d5, the e-pawn may not be able to move game, because there can be many active pieces
for the rest of the game. Both sides have to remaining on the board producing astonish
weigh whether one advantage or the other is ingly complex positions. The phrase 'queenless
more important. If Black isn't going to exchange middlegame' doesn't appear often enough in
on e3, one of his options is to leave the bishop chess discussion, written or otherwise. It de
where it is on c5. Normally the doubled pawns scribes an extremely large set of situations, often
that Black would get if White played ..llxc5 lasting for the bulk of the game. The conditions
wouldn't be harmful (see the discussion of c for a decisive result are still there, as shown by
pawns below); but that' s not always true. The literally thousands of games. But for our pur
same idea comes up in the Ruy Lopez after 1 e4 poses it's important to note that a lot of queen
e5 2 CLJf3 CLJc6 3 i.b5 a6 4 ..lla4 CLJf6 5 0-0 ..lle7 6 exchanges such as the one above result in
.l:te l b5 7 ..llb3 d6 8 c3 ..lle6, as well as in some queenless openings ! Although the boundaries
queen' s pawn variations; e.g., 1 d4 d5 2 CLJf3 e6 of the queenless opening, middlegame and end
3 ..llf4 CLJf6 4 e3 CLJbd7 5 i.e2 ..lld6 (D). ing are to some extent a matter of judgement,
variations in which the queens have been ex
changed within the first 10 moves are routinely
analysed by players and theoreticians for an
other 10 moves, and clearly belong to the terri
tory of the opening proper.
Finally, we run across 5th-rank doubled

pawns in just a few openings; for example, 1 c4
c5 2 CLJc3 CLJc6 3 g3 g6 4 ..llg2 i.g7 5 d3 CLJf6 6
e4 0-0 7 CL:lge2 d6 8 0-0 CLJd4 9 CLJxd4 cxd4 1 0
CLJe2 (D).
B
Again the choice arises of whether White
should:
a) exchange bishops on d6, allowing . . .cxd6
if Black wants to;
b) leave his bishop on f4, inviting ... i.xf4;
or
c) retreat to g3 .
In master play all three solutions are played.
This position is simplified but shows the basic
situation that arises in many variations.
Isolated doubled e- and d-pawns are rare
when the queens are off the board; neverthe Keene referred to the doubled pawn on d4 as
less, an opening line such as 1 e4 d6 2 d4 CLJf6 a 'dead point' , so called because it has little or
3 CLJc3 e5 4 dxe5 dxe5 5 \\!Vxd8+ �xd8 6 i.c4 no dynamic potential. It makes Black's central
play difficult because White will be ready to re leads to doubled pawns is when a knight on c3
spond to . . . e6 and . . .d5 by cxd5 and e5 (espe or c6 is captured by a knight or bishop. Then a
cially if the move f4 has been played), whereas basic decision often presents itself: whether
Black' s move ... e5 would restrict his own bishop one wants to recapture with a b-pawn ( 'strength
and isn't very helpful with respect to mobility. ening' the centre) or with the ct-pawn, opening
The opponent (in this case White) can play lines for development. There are plenty of situ
'around' the pawn by f4, intending f5 and g4, ations in which there is no choice; for instance,
and/or by b4. This is a theme worth remember 1 d4 lllf6 2 c4 e6 3 lbc3 .ib4 4 a3 .ixc3+ 5
ing as it arises fairly frequently in openings bxc3 or 1 e4 e6 2 d4 d5 3 lllc 3 .ib4 4 e5 c5 5 a3
such as the Closed Sicilian, King ' s Indian and .ixc3+ 6 bxc3, both cases where the prece
English Opening. It tends to occur in the move dence of structure is conspicuous. In neither
sequence above, with a knight on d4 (from case does either player have many pieces de
White' s point of view) being captured by a veloped, but by their pawn moves both sides
knight on e2 or f3. In many cases there would have established a structure that will serve and
be a bishop on e3 in the above case, say, by determine their development.
8 . . . llld7 9 .ie3 llld4 (D). In both of these cases the players were forced
to capture ' towards the centre ' , the advice
given to students everywhere. But it's more re
vealing to look first at recaptures requiring a
decision.
Ruy Lopez Exchange Variation
1 e4 e5 2 lllf3 lllc 6 3 �b5 a6 4 �xc6 (D).
In this instance White's knight can't capture

on d4 because of the fork, and most players will
avoid giving away their good bishop by .ixd4,
dead spot or not. So you will commonly see
players wait until a bishop comes to e3 before
occupying the outpost with their knight. It also
frequently happens that when a bishop arrives
at e3, White is just ready to play d4, so Black's
knight jump has a double purpose. Obviously a) One answer is 4 . . . bxc6, but this is rarely
all of this is true with colours reversed as well. chosen. This case has more to do with specifics
Naturally there are no absolutes and the than with general principles, but that in and of
dead-point structure isn't always bad, but one itself adds interest. The usual lines go:
should be careful that there are compensating al) 5 0-0 d6 6 d4 f6 7 lllc 3, when White
factors before adopting it. controls the centre and has a simple lead in de
velopment (three pieces to none). Black's f8-
Doubled c-Pawns bishop can't take part in the action, and his
pieces are cramped, not what you want when
Now let's move on to doubled c-pawns, which you have the bishop-pair.
are far more common than central ones. The a2) 5 lllc 3 d6 6 d4 exd4 (6 . . .f6 7 .ie3 threat
most frequent exchange in the opening that ens 8 dxe5 fxe5 9 lllxe5 ! , and otherwise 8 'i¥d2
and 0-0-0 will be pretty unpleasant for B lack) 7 preferred by top masters, when the play seems
�xd4 with a position much as in the Philidor to favour White slightly but Black has squares
Defence, but a tempo up for White due to 3 . . . a6. for his pieces and the extra centre pawn gives
It' s a bad sign if B lack has to surrender the cen him a certain leeway, so in the hands of a
tre in a centre-strengthening variation ! knowledgeable player 4 . . .bxc6 isn't bad. Nev
b) 4 . . . dxc6 is well-known and doesn' t re ertheless, we can't say that it' s fully satisfac
quire special analysis. What counts is that the tory.
recapture away from the centre affords wide a l 2) Black can equalize by capturing away
open play for the bishops. from the centre: 4 . . . dxc6 5 d3 �g7 and 6 0-0
b l ) The line 5 d4 exd4 6 �xd4 �xd4 7 lLif6 or 6 h3 e5 works out well him. White is not
lbxd4 �d7 and ... 0-0-0 illustrates Black's ideas. able to achieve an effective d4 or e5, so Black
He will gladly play with a pawn-structure such gets easy development for his pieces.
as . . .c5 and ... b6. a2) 4 0-0 �g7 5 .l:!.el e5 6 �xc6 and then:
b2) The generally-approved move 5 0-0 has a2 1 ) 6 . . . bxc6 ? ! 7 c3 lbe7 8 d4 cxd4 9 cxd4
other attributes, but again the fact that a varia exd4 10 lbxd4 0-0 1 1 lbc3 (D) is notoriously
tion such as 5 . . . �g4 6 h3 h5 7 d3 �f6 even ex better for White.
ists shows that Black has dynamic counter
play. In fact, White often plays c3 and d4 versus
the . . .c5/. . .e5 structure, allowing the doubled
pawns to be liquidated and therefore indicating
that they weren't the sole reason for playing 4
�xc6. A case in point: 8 lbbd2 lbe7 9 lbc4
�xf3 10 �xf3 �xf3 1 1 gxf3 lbg6 1 2 �e3 �d6
1 3 .l:!.fd l f6 14 'it>fl c5 1 5 c3 'it>f7 16 d4 ! cxd4 1 7
cxd4 .l:!.hd8 ? ! ( 1 7 . . .exd4) 1 8 .l:!.ac l �e7 1 9 d5
with an advantage for White, Glek-Winants,
2nd Bundesliga 1 997/8.
The Berlin Variation with 3 ... lbf6 4 0-0 lbxe4
5 d4 lbd6 6 �xc6 dxc6 7 dxe5 lLif5 8 �xd8+
'it>xd8 also shows that Black is willing to play
this pawn-structure. For more on this subject
see Chapter 8 on the Ruy Lopez. Even in the reversed position from the Eng
Overall, we can say that in this particular lish Opening, Black usually gets the better of
opening, Black's choice of developing and acti this position with one less move to use. The
vating his bishops by capturing away from the problem is that 1 l . . .d5? gives White too much
centre leads to better positions than if he de pressure after 1 2 exd5 cxd5 1 3 �g5 . But other
cides upon a more compact pawn-structure by wise Black's dark squares are weak and 1 l . . .d6
capturing towards the centre. presents a target down the d-file. Notice that
this Sicilian Rossolimo is similar to the Ex
Rossolimo Variation of the Sicilian Defence change Ruy Lopez that we just looked at, in that
both variations have lines in which the pawn
This positionally instructive opening is defined break d4 is paradoxically strong even though it
by 1 e4 c5 2 lLif3 lbc6 3 �b5, and has numer straightens out the opponent's pawns.
ous lines with �xc6. I'll pick a few. a22) By contrast, 6 ... dxc6 7 d3 �e7 has tra
a) 3 ... g6 and then: ditionally been considered equal with careful
a l ) 4 �xc6 and now: play. The d-file is handy for Black and White's
a l l ) 4 . . .bxc6 5 0-0 �g7 6 .l:!.el gives another move d4, a poor one, would only open up the
lead in development which particularly shows game for Black's bishops.
up after 6 . . .lbf6 7 e5 lbd5 8 c4 lbc7 9 d4 ! cxd4 b) 3 . . . e6 4 �xc6 and here:
10 �xd4 with space and the simple idea �h4 b l ) 4 ... bxc6 5 0-0 lbe7 6 :e l (these are
and �h6. Thus 6 . . . lbh6 with the idea .. .f6 is hardly forced moves, just illustrations of the
play) 6 ... lllg6 7 c3 iL.e7 8 d4 0-0 9 lllbd2 cxd4 follows. Having pawns on c2 and c3 is easy for
10 cxd4 f5 ! and B lack has freed all of his White to handle, just as . . . dxc6 was in the Ruy
pieces. Lopez. The difference is that in this Petroff line
b2) 4 . . . dxc6 is inferior because White will both sides have two bishops, so it' s unlikely
get a pawn to e5 that cramps Black's game; e.g., that White has anything special in the way of a
5 0-0 'Viiic7 6 e5 and moves such as b3, iL.b2, d3 permanent advantage. Nevertheless, taking with
and llld2-c4 can follow. If he had an extra cen the ct-pawn is more promising than 6 bxc3,
tre pawn (as he does after 4. . .bxc6), Black could which would leave White with a restricted cen
play . . .f6 and break up White's centre, but in tre in which one of his bishops wouldn't be able
this case exf6 would expose a weak pawn on e6. to assume an active role.
In making a decision how to recapture in the
Rossolimo Variation, a major consideration is Scotch Game
whether Black can achieve . . . e5 after taking
with the d-pawn. If so, White has no particular In the Scotch Game with 1 e4 e5 2 lllf3 lllc6 3
way to gain space, because now c3 followed by d4 exd4 4 lllxd4 lllf6 5 lllx c6, 5 . . . bxc6 is the
d4 merely opens the centre for Black's bishops. almost automatic recapture. This is still an un
But if Black captures with the b-pawn he has to resolved line, but Black's queenside structure
watch out that an early d4 doesn't leave him too doesn't hurt him in most lines. Two examples
far behind in development (he has no open d-file with this type of structure:
to challenge a white piece on d4). In particular,
the variations in which Black fianchettoes his Rublevsky - Bologan
bishop can put his development behind sched Dortmund 2004
ule.
1 e4 e5 2 lllf3 lllc6 3 d4 exd4 4 lllxd4 iL.c5 5
Petroff Defence lllxc6 'Viiif6 6 'Viiid2 bxc6 7 lllc3 llle7 8 llla4 iJ..b6
9 iJ.. d3 0-0 10 0-0 d6 1 1 'Viiie2 lllg6 12 lllxb6
1 e4 e5 2 lllf3 lllf6 3 lllxe5 d6 4 lllf3 lllxe4 5 axb6 (D)
lllc 3 (this move is fashionable at the moment,
but it is not the only example of doubled pawns
in the Petroff: 5 c4 ! ? is an interesting move
whose very purpose is 6 lllc 3 lllx c3 7 dxc3 ! ,
aiming for active development) 5 . . . lllxc3 6 dxc3
(D).
Thus White has given Black a compact struc

ture on the queenside but gained the bishop-pair
and a mobile kingside majority as well. But
Black has some advantages too. His bishop is
good and he has two useful files for his rooks.
Right off, . . . lllf4 is a positional threat.
Recapturing with the ct-pawn is the very point 13 f4 'Viiid4+!
of 5 lllc 3, to get White's pieces out quickly with 1 3 . . . l:!.eS is also reasonable.
additional pressure down the open d-file if 0-0-0 14 �hl f5!
Blockade. 1 1 . . . .l::tg 8 1 2 g3 .lii.h 3+, which is at least equal.

15 :dl ! �f6 16 exf5 �xf5 17 �xf5 �xf5 So Kasparov played 1 0 �d2 :gs ( 10 . . . f5 ! ?) 1 1
(D) llla4 �d6 1 2 0-0-0 �e6 1 3 lbc3 0-0-0 (it' s hard
for White to develop) 14 g 3 ! �g4 1 5 �e2 llle5
16 �f4 ( 1 6 f4 �xe2 1 7 lbxe2 lbg4 1 8 :ctn
�c5, targeting weaknesses, is equal) 1 6 . . . i..h3
( 16 . . . �xe2 1 7 lbxe2 �c5 with equality) 1 7
�h5 �c5 1 8 .l:!.xd8+ :xct8 1 9 lbdl .lii.g 2 2 0 :e l
and the game was drawn in short order.
5 lllxc6 bxc6 6 e5 lbe4!?
6 ... llld5 is the main continuation.
7 lbd2 lllc5 8 �e2 �e7 9 0-0 0-0 10 lllb3
lbxb3?! 11 axb3 (D)
18 �c4+
Otherwise, White's bad bishop and weak
squares down the e-file will give him a serious
disadvantage.
18...l:!.f7!?
Alternatively, 1 8...�f7 1 9 �xc6 :ae8 ! would
threaten . . J:te2, . . .lbxf4, etc. The opening has
been a success for Black.
19 �xc6 l:!.a5! 20 �e8+ :rs 21 �e2 lLixf4!?
22 �xf4 �xf4 23 l:!.fl :es? 112-112
After 24 �d3, 24 ... �e4 ! gets Black's rook to White has foreseen something analogous to
the 7th rank in an ending. the Rossolimo Variation above. Black can't
move his d-pawn without one problem or an
Morozevich - Bezgodov other. White's queenside complex actually pro
Russia Cup (Tomsk) 1998 tects him from intrusions on the b-file, and his
possession of the a-file is a bonus.
1 e4 e5 2 lLif3 lbc6 3 d4 exd4 4 lLixd4 lLif6 1 1...d5 12 exd6 �xd6?
Another set of choices confronts Black in the Notice that this is an example of the vanishing
main-line variation 4 ... �c5 5 lbxc6 �f6 ! 6 centre ! Since Black has no attack he has no real
�d2 dxc6 7 lbc3 lbe7 8 �f4. Without getting compensation for the weak c-pawns. 1 2 ... cxd6
too theoretical, it's relevant to observe that Black must be a little better. On the other hand, Black's
wants to take on another set of doubled pawns centre pawns would still be weak and White
after 8 . . .lbg6 ! ? 9 �xf6 gxf6, as in Kasparov could probe the kingside. There might follow
Topalov, Las Palmas 1 997. How to assess this 1 3 �d3 ! ? (or 1 3 �f3) 1 3 ... d5 14 :e l �d6 1 5
kind of thing? It takes some experience but also �h5 and now 1 5 . . .f5 ! ? 1 6 �g5 or 1 5 . . .g6 1 6
a little calculation. Black has a temporary lead �h6. Black's kingside is causing him serious
in development and ifhe could castle queenside problems. l:!.a4-h4 and �d2-c3 are productive
and/or exchange off his f-pawn by ... f5, he'd ideas.
leave White having to defend squares such as 13 :a4! (D)
c2 and f2. Thus slow moves from White are not 13 �f5 14 �d3
••.
dangerous. But 1 0 �d3 isn't much of a solution This time the theory that simplification helps
because it runs into 1 0. . .lbh4 1 1 @fl ( 1 1 0-0? White makes sense. That's one less piece for
l:tg8 12 g3 �h3 1 3 :ct l lLif3+ 1 4 '>ith l �xf2) Black to defend pawns with.
14 .ixd3 15 'ii'xd3 'ii'f6 16 g3 .:re8 17 @g2

.•. Black has battened down the hatches but
.ic5 18 l:lf4 'ii'e6 19 lldl .id6 White has many modes of attack on the weak
Versus 'ii'd 7. ened c-pawns in these sorts of positions, in
20 .i::r.c4 c5 cluding .ie3, .l::l ac l , llle4, 'i¥a5 and some cases
Now Black won't be able to defend the c even b3 and .ia3.
and a-pawns, especially the former. 13 llle4 .ih3 14 .l::ld l 0-0 15 llla5 lllb5
21 'ii'f3 .:ab8 22 .id2 f6 23 :al .i::r.bd8 24 The only defence for the c-pawn but the
.ie3 'i¥e5 25 .i::r.c3 :as 26 :as knight also heads for d4, a typical defence.
There goes the c-pawn. 16 i..e3 llld4 17 .ixd4 cxd4 18 lllxc6
26 �e4 27 .ixc5 'ii'xf3+ 28 @xf3 .ie5 29
••. As was the case with doubled pawns in the
.i::r.e3 .ixb2 30 .i::r.xa7 .i::r.xe3+ 31 @xe3 llxa7 32 Nimzo-lndian, it's very common to see the
.ixa7 �f7 33 @d3 l-O one in front be exchanged and the one behind
fall.
Finally, giving up a fianchettoed bishop on 18 ...�hS 19 l:!.acl .l::lfc8 20 'i¥a6 .if8 21 b4
g7 for a knight on c3 (or one on g2 for a knight .l:i.c7 22 b5 l:!.ac8 23 f3 h5 24 l:!.c2 'i¥d5 25 l:!.dcl
on c6) is a traditional technique that crops up i.. d7 26 llle7 .l::lxc2 27 l:!.xc2 .ixe7 28 l:!.xc8+
in many variations. The question is always .ixc8 29 'i¥xc8+ �h7 30 'i¥e8
whether the bishop-pair compensates for the White is winning a second pawn, after which
doubled c-pawns. By themselves the bishops the rest was easy for him.
usually aren't sufficient to offset the pawns,
but the capture has also seriously weakened Hamann - Geller
squares on the opponent's kingside. Getting a Copenhagen 1960
feel for this trade-off is more a matter of expe
rience, so here' s a small selection of a few very 1 d4 lllf6 2 c4 g6 3 lllc3 .ig7 4 e4 0-0 5 .ie2 d6
lightly annotated games. The first is a win by 6 lllf3 .ig4 7 0-0 lllfd7 8 .ie3 lllc6 9 d5 .ixf3
White in a variation that's arisen hundreds of 10 .ixf3
times: 1 0 gxf3 ! ? is definitely worth thinking about.
It keeps more queenside options open, and
Korchnoi - H. Bohm White's king is perfectly safe.
Wijk aan Zee 1 980 10 ... llla5 11 'i¥a4
White could also try keeping the position
1 c4 c5 2 lllc3 lllf6 3 g3 d5 4 cxd5 lllxd5 5 open for the bishops by 1 1 .ie2 .ixc3 1 2 bxc3
.ig2 lllc7 6 lllf3 lllc6 7 0-0 e5 8 d3 .ie7 9 e5 ! ? 1 3 f4 ( 1 3 dxe6 fxe6 14 f4).
llld2 .ie6 1 1....ixc3!? 12 bxc3 b6 13 i..e2 e5 14 g3
Later, 9....id7 became the main line, to avoid 1 4 dxe6 might be better. The rest of the
the doubled pawns: game gets one-sided as the knights dominate
10 .ixc6+ bxc6 1 1 'i¥a4 'i¥d7 12 lllc4 f6 (D) the bishops.
MASTERING THE CHESS OPENINGS
14 ltJcS 15 \i'c2 \i'd7 16 �h6 .l:!.fe8 17 a4

••• After this exchange sacrifice things are clear.
f6 18 'it>g2 :e7 19 h4 .l:!.ae8 20 �e3 lbab7 2 1 There is no dark-squared bishop remaining to
h5 g5 exchange one of Black's mighty knights, Black
Geller closes the kingside. It's hard to be has control of the c-file, and White's light
lieve that he can win on the queenside alone. squared bishop is awful.
22 g4 .l:!.c8 23 f3 lba5 24 .l:!.tbl lllcb7 25 .l::tb4 34 \i'd2 \i'c7 35 h6+ cJi;f7 36 .l:!.5b4 lbc5 37
c6 (D) .lii.b5 lbcb3 38 \i'h2 cJi;f8 39 \i'h5 \i'c8 40 l:!.fl
a6 41 �e2 \i'c5 42 .l:!.hl lllc l 43 .l:!.h2 \i'e3 44
�dl \i'f4+ 45 'it>f2 lLid3+ 46 cJi;g2 \i'cl 0-1
The coming ... lllf4 is about as strong a knight
move as you'll see.
Remember, though, that the fianchettoed

bishop is missing. There have been numerous
games where the opponent made that count.
Here's one example that almost explains itself:
Anikaev - A. Petrosian
Kiev 1973
1 c4 lbf6 2 lLic3 c5 3 g3 d5 4 cxd5 lbxd5 5 �g2

lbc7 6 d3 e5 7 \i'b3 lbc6 8 �xc6+ bxc6 9 lLif3
A key concept. In many Nimzo-Indians, this f6 10 \i'a4 .lii.d7 11 0-0 lbe6 12 lbe4 'li'b6 13
pawn is on c5 and in spite of Black's efforts lbfd2 'li'b5 14 \i'dl .lii.e7 15 lbc4 0-0 16 b3
there is no way to increase the pressure on the 'li'b8! 17 �a3? ! f5 18 lbc3 .l:!.f6 19 .l:!.cl .i:th6 20
doubled pawns. Black should always think about lba4 f4 21 e3?
keeping . . . c6 in reserve. It's surprisingly difficult for White to de
26 i.d2 \i'd8 27 \i'a2 .l:!.ec7 28 l:!.dl cJi;g7 29 fend. Perhaps 2 1 �b2 \i'f8 22 e4 ! ? is a good
'it>g3 cxd5 idea, to hit the weak e5-pawn and at the same
There it is, the attack on the back pawn that time prevent . . .\i'f5 .
we 've talked about. But how can Black break 21 \i'f8! (D)
.•.
down the defensive structure?

30 cxd5 .l:!.c5 31 .l:!.dbl .l:!.8c7
Black probably intends . . .\i'c8 before any
thing else, with the same ideas as in the game;
but as events have it he doesn't have to wait.
32 l:!.b5? l:!.xc3! 33 .lii.xc3 .l:!.xc3 (D)
22 f3
It's too late for 22 lbxe5 \i'f5.
22 fxg3 23 hxg3 \i'f5 24 .l:!.c2?! 'li'h3 25
•.•
l:!.g2 lbg5 26 \i'e2 e4! 27 dxe4 lbxe4 28 fxe4

.lii.g4 29 .l:!.h2 .lii.xe2 30 :xh3 .l:!.xh3 31 .l:!.f4
.ixc4 3 2 bxc4 l:lxg3+ 3 3 lit> f2 l:!.h3 3 4 .ixc5 This structure generally arises from two pawn
.ixc5 35 lllxc5 l:!.f8 36 e5 g5 37 l:!.xf8+ @xf8 38 exchanges on c5 and d5, but it can also come
llle6+ 'iit>e7 39 lllxg5 l:!.h2+ 40 @f3 l:.xa2 41 about when an isolated pawn is transformed by
lllxh7 a5 0-1 a piece exchange on c6.
Like 'isolated pawns' , the term 'hanging
It's worth adding that in the Accelerated pawns' is defined more broadly, but it doesn't
Fianchetto Sicilian the capture . . . .ixc3 often seem to extend beyond this single case when
comes without White having made a move like actually being discussed. That is understand
c4. The best example of this situation occurs af able, because so few analogous structures reg
ter 1 e4 c5 2 lllf3 lbc6 3 d4 cxd4 4 lllxd4 g6 5 ularly arise, at least in the opening. You could
lbc3 .ig7 6 lllb 3 .ixc3+ 7 bxc3, a line which call pawns on e4 and d4 'hanging' under cer
usually continues 7 ... lllf6 8 .id3, and now Black tain circumstances, but that's not convention
has the interesting choice between 8 . d6, in
. . ally done.
tending to blockade the c-pawns, and 8 . . . d5 9 Returning to the basic position, Black' s
exd5 'i¥xd5, when Black develops so quickly hanging pawns have advantages and disadvan
that White' s pieces tend to be tied down. The tages. Much as is the case with an isolated d
usual considerations with respect to Black's pawn, Black has the persistent possibility of
dark squares on the kingside apply in principle, breaking the position up by . . . d4, thus extend
but there is little chance that they will become a ing the range of his pieces, initiating favour
real factor. able tactics, and/or creating a powerful passed
If you want to continue investigating the is pawn. The hanging pawns also cover key cen
sue of captures away from and towards the tral squares and give Black's pieces somewhat
centre, there will probably be instances of both more manoeuvring room than White's. Finally,
in the openings that you play. The more that the e- and b-files can be used to create dynamic
you study these and get to experience them, chances.
the better a player you'll be in the widest From White's point of view there are many
sense. promising ways to attack this structure. Most of
them begin by restricting the advance of the d
Hanging Pawns pawn. White has a pawn, a knight (sometimes
two), and a rook or two on an open file to
The term 'hanging pawns' is habitually used to achieve this, with a bishop on b2 for good ef
refer to black pawns on c5 and d5 separated fect. Once the pawn is 'fixed ' , White can do
from Black's other pawns by at least a file on one of several things:
both sides. The hanging pawns are usually pit a) Attack it with his pieces; e.g., a bishop on
ted against a white pawn on e3 and open d- and g2, knight on c3 and/or f4, and rook(s) on an
c-files. Of course the same applies with colours open file. The queen and rooks are particularly
reversed. effective attackers of hanging pawns.
b) Advance a pawn to b4 or e4 to force a
desirable change in pawn-structure. If White' s
advancing pawn either captures Black's or vice
versa, an isolated pawn remains in Black's
camp. Or, if one of Black's pawns advances, it
creates a juicy outpost for White to the side of
it. For example: if White attacks with e4 and
Black responds with . . . d4, then the c4-square
is available for a piece.
c) Exchange pieces and simplify the posi
tion; as is the case with an isolated queen' s
pawn, this reduces the pawns' dynamic possi
bilities and makes them easier to put under
pressure.
In the following game White strives to fix

the hanging pawns and Black to use them dy
namically.
Seirawan - Short
Montpellier Ct 1985
1 d4 lLif6 2 c4 e6 3 lLif3 b6 4 lbc3 JLb7 5 JLg5

h6 6 JLh4 JLe7 7 �c2 c5 8 dxc5 bxc5 9 e3 0-0
10 iLe2 d6 (D)
This pawn-structure is fine, as has been dem
onstrated in many games. Black ultimately plays
. . . d5, which he could also do immediately; e.g.,
1 0 ...d5 1 1 cxd5 exd5 12 .l:!.d l lbbd7 followed by
...�b6 (or ...a6 first). A standard picture of the hanging pawn duo
c5/d5.
20 Ji.b5 .l:!.c7 21 �f4
Short may have been hoping for 2 1 lbxd5
.l:!.xd5 22 .l:!.xd5 lbxd5 23 l:!.xd5 lbd4 ! , when
Black is at least equal.
21...lba5 22 �a4 lbc6 23 Ji.e2 :cd7 24 �f4
a6!?
24 . . .lba5 25 lbe5 .l:!.d6 is probably OK as
well; and 24 . . . a5 would put the idea of b4 to rest
for a while.
25 iLfl �f8
Unfortunately, Black has no . . .d4 break and
there isn't much positive to do.
26 g3!? �e7 (D)
1 1 0-0 lbh5!?
Black gets rid of White's most dangerous
bishop, the one that could attack him from g3 or
capture on f6 at the right moment. This takes an
extra move (the knight will return to f6 while
Black gets . . .�e7 in) but he seems to have the
time to get away with it.
12 ii.xe7 �xe7 13 .i:tadl lbf6 14 .i:!.d2 lbc6
15 .i:!.fdl .i:tfd8
White has a little space and d-file pressure,
but the d6-pawn is typically safe and he has no
particular targets of attack.
16 h3 :d7 17 a3 .l:!.ad8
17 . . ..l:!.b8 is the other natural move, to take
advantage of the open file and potentially probe 27 JLg2
the holes left by a3 . White's reorganization is complete. The d5-
18 �a4 d5!? pawn holds firm, however.
A huge decision, changing the character of 27 �e6 28 'it>h2
•.•
the game, although not necessarily to Black's The danger lurking in the background is
detriment. Preventing b4 by 18 . . .a5 looks equal. shown by 28 �a4? ! 'it>h8 29 b4? d4! .
19 cxd5 exd5 (D) 28 lba5?
•••
Black, trying to win, disturbs the balance and b 3 ! ? .l::l ac8 1 9 bxc4 dxc4 2 0 l:!.c2 .ixc3 2 l l:!.xc3
permits simplification. After that, White forces tlld5 ! (it turns out that the pawn is not weak; at
serious positional concessions from Black's this point 22 .i::r.xc4? loses to 22 . . .lllc 3) 22 l:!.c2
position. c3 23 :de 1 .l::l c5 24 lllb 3 .i::r.c6 25 llld4 l:!.c7 ! 26
29 llle5 .l::ld6 30 'ii'a4! 'ii'xe5 31 'i¥xa5 :cs lllb5 l:!.c5 27 lllxc3?? lllx c3 28 .l::lx c3 l:!.xc3 29
32 llla4! .l::ldc6 33 .i::r.c2 Wi/e7 34 .i::r.d cl c4 (D) l:!.xc3 �2 ! 0- 1 .
35 l:!.dl
Once again all pieces are to be aimed at d5.
White still has to win the overprotected pawn
on that square or break through in some other
fashion, no easy task.
35 ...l:!.d8 36 .i::r.cd2
This attacks d5 ; in one more move, every
piece will be trained upon it.
36... .:cd6?!
36 . . . l:.cc8 keeps the possibility of lateral de-
fence by . . . l:!.c5 alive. The d-pawn is tough to
corral, but ultimately the threat of a break by e4
will overload Black; for example, 37 �gl (37
lllc 3 l:!.c5 ! ) 37 . . .llle4 3 8 l:!.d4 ! lllc5 39 tllxc5
.i::r.xc5 40 'i¥b4 a5 41 'i¥c3 f5 42 b 3 ! cxb3 43
An almost decisive concession. Sometimes 'i¥xb3 and Black is reduced to total passivity.
this advance is a reasonable trade-off because 37 lllc3 'i¥e6 38 l:!.d4 l:!.6d7 39 l:!.ld2 g6 40
White's vulnerable pawn on b2 is fixed. But llla4
here Black can't even begin to mount an attack Back to c 5 !
on that pawn, and his b7-bishop is too passive 40...'i¥e7 41 lllc5 l:!.c7 42 lllxb7
to make room for any dynamic compensation. A typical exchange of a horrible piece for a
Compare this position from O.Bemstein-Capa good one in order to eliminate the best de
blanca, Moscow 1 9 14: fender.
42 l:!.xb7 43 .ixd5 l:!.xd5 44 .i::r.xd5 lllxd5
.•.
45 'i¥xd5
and wins.
Here's the flip side:
Korchnoi - Karpov
Merano Wch ( 1) 1 981
1 c4 e6 2 lllc3 d5 3 d4 .ie7 4 lllf3 lllf6 5 .igS

h6 6 .ih4 0-0 7 e3 b6 8 1:!.cl .ib7 9 ii.e2 lllbd7
10 cxd5 exd5 11 0-0 c5 12 dxc5 bxc5 (D)
13 'i¥c2 l:!.c8
Obviously . . . d4 is on Karpov' s mind, in or
der to exploit White's queen's position.
In the Capablanca game Black has full equal 14 l:!.fdl 'i¥b6
ity because the b-file and b-pawn are just as This is a perfect spot for the queen. It supports
much a worry to White as the d-pawn and d-file . . .d4, will attack the b-pawn if . . .c4 is needed,
are to Black. The biggest difference is that and, not least, Black's rooks are connected.
Black has an active good bishop versus the very 15 'iYbl l:!.fd8 16 l:!.c2
bad one in Seirawan-Short. This well-known Korchnoi would like to double rooks on the
game (because of its cute finish) continued 1 8 d-file, as in Seirawan-Short.
25 ...dxe3 26 fxe3
White's pawn-structure is shattered, although
simplification would still leave him with some
chances. So Karpov takes aim immediately.
26 c4! 27 lbed4 'Yi'c7 28 lbh4
.••
Hoping for 28 . . . 'Yi'xg3?? 29 lLihfS+.

28 'Yi'eS 29 cJi;bl cJi;gS!
..•
There are always issues of accuracy. Karpov

avoids 29 . . . lLihS? 30 lLihf5+ gxf5 3 1 lbxf5+
with some play. Now Black wins with ease.
30 lbdf3 'Yi'xg3 31 .l:!.xd8+ .lii.xd8 32 °Yi'b4
.lii.e4 33 �xe4 lbxe4 34 .l:!.d4 lLif2+ 35 cJi;gl
lLid3 36 °Yi'b7 .l:!.b8 37 °Yi'd7 �c7 38 cJi;hl .l:!.xb2
39 .l:!.xd3 cxd3 40 °Yi'xd3 °Yi'd6! 41 'Yi'e4 °Yi'dl +
16 ... 'Yi'e6! 17 �g3 42 lbgl °Yi'd6 43 lbhf3 .l:!.bS 0-1
But now 17 .l:!.cd2? fails to 17 ...lbe4 ! l 8 lbxe4
dxe4 1 9 �xe7 exf3 20 �xd8 fxe2 2 1 .l:!.xd7 Majorities and M i norities
°Yi'g4 ! 22 .l:!. l d5 (the only move) 22 ... i.xdS 23
.l:!.xd5 °Yi'b4 and wins. This gives Black just The term 'pawn-majority' refers to one player
enough time to rid himself of White's bishop. having more pawns than his opponent in a par
17...llJhS 18 .l:!.cd2 lbxg3 19 hxg3 lLif6 20 ticular sector of the board, that sector being de
'Yi'c2 g6 21 'Yi'a4 fined by a number of adjacent files. Normally,
White's pieces begin to assume more active we only talk about a majority when the pawns
posts. in question are connected, i.e. there is no empty
21 ...a6 22 �d3 cJi;g7 23 �bl °Yi'b6! (D) file between them. Putting that into a real-world
context, here is a Griinfeld Defence in which
Black has a queenside majority (2 to 1, hence
forth '2: l ' ), White has a central majority (2: 1 ),
and the pawns are evenly divided on the king
side (3:3):
24 a3?
White is trying to avoid . . . °Yi'b4, but he under
estimates the strength of Black's next move:
24 d4!
•••
Everything depends upon whether Black can

get away with this advance. The other way to express this is that there
25 lbe2 are two connected sets of pawns ( ' pawn
A sad retreat. The idea of 24 . . . d4 is 25 exd4 islands ' ) , so that we have 2: 1 on the queenside
�c6 ! 26 'Yi'c2 (26 °Yi'c4 �xf3 27 gxf3 cxd4) and 5 :4 in the centre and kingside. I think that
26 . . . �xf3 ! 27 gxf3 cxd4 28 lba4 °Yi'b5 ! and the imparts less information, so I ' ll divide centre
knight falls. and flank pawns, with the exception that if
there is a single centre pawn on the board and

it is connected with other pawns on the flank, I
may group them together, an important case
being the 4 : 3 kingside set-up that we shall run
into in the course of discussing openings with
that pawn distribution.
One way of thinking about the Griinfeld De
fence main line above is that White's centre is
under pressure by direct threats and other in
conveniences from an enemy who has no tar
gets of attack in his own position. What' s more,
there aren't even prospective targets of attack in
the near future ! This sounds one-sided until
you take into account that White is protecting a
central majority, possibly the most valuable as the necessity of keeping White ' s pieces fairly
set in chess in the realm of pawns and struc rigidly poised to prevent B lack' s central expan
tures. How is that? First, two central pawns sion. Even though White' s centre pawn in this
control more central points than one, in itself example is more advanced than Black' s are, it
an advantage. Then, after a protracted struggle can still cover only one central square, namely,
to survive the constant threats to their lives d5 .
and/or their integrity, a central pawn-majority To illustrate this, we might ask why ... b5 is
can sweep across the board and scatter the op so effective in the Sicilian Defence (when Black
ponent's pieces, sometimes exacting material has pawns on e6 and d6). A large part of the rea
tribute along the way. Even more frequently a son is a well-timed . . .b4, of course, to drive
central majority can be transformed into a away the c3-knight (e.g., to e2) and then either
passed pawn that is difficult or impossible to put pressure on White's e-pawn or successfully
stop. That is precisely what happens when achieve a pawn-break in the centre.
things go wrong for Black in many variations of
the Griinfeld Defence. B arring such a trium
phant journey, a central majority has other ad
vantages. It can advance far enough to grant
abundant room for friendly pieces to roam, but
can also provide the maximum security to the
pieces behind it. There are even advantages to
having a central pawn-majority that resides on
the third rank. The most important situation in
which that occurs is in the Open Sicilian, in
which Black always has a central majority to
begin with, because White has played 3 d4
cxd4 4 lllxd4, as in this example (see following
diagram):
The pawns on d6 and e6 protect against
threatening incursions by putting all of White's But White often plays g4-g5 himself and
important 5th-rank squares under pawn super drives away the f6-knight (e.g., to d7) . Often
vision. They combine that with a threat to ad that has less effect as regards positional con
vance, when they would give Black's pieces siderations in the centre. What's the differ
freer play and begin to restrict White's. Such a ence? The central majority. Let's pretend that
majority can compensate for a space disadvan Black had only a pawn on d6 and White has
tage elsewhere, because the main value of a his usual central pawn on e4. Then driving
space advantage is the ability to shift forces away White's c3-knight might be of about the
about more easily, and that can be limited by same importance as White' s driving away the
f6-knight. Furthermore, the lack of an e-pawn of pawns (on d6 and e5), White ' s knights are
for Black would mean seriously weakened de denied e4 and d4, so that defending d5 is really
fence against White' s pieces occupying cen Black' s only practical concern, just as White
trally-oriented squares, specifically d5 and f5 . himself must watch out for . . . d5. Again, see
For example, if Black ' s e-pawn were missing, Chapter 1 1 on the Sicilian for various exam
then d5 would be an attractive outpost that ples.
would be further weakened if White could What are some other common central pawn
force Black' s knight off f6 by g4-g5 . In that majorities? White finds himself with this ma
kind of a position a knight on f5 is also notori jority in several variations of the Griinfeld De
ous for tearing Black ' s position to shreds. As fence such as the one mentioned above and in
it is, since Black ' s pawn is on e6, White ' s lim the important variation 1 d4 lZ'if6 2 c4 g6 3 lZ'ic3
ited central pawn presence in the Open Sicil d5 4 lZ'if3 j_g7 5 l'lb3 dxc4 6 l'lxc4. In the
ian also allows Black to use influential squares Queen's Gambit Exchange Variation White as
for his purposes, such as c5 and e5 for his sumes a 2: 1 majority on move four ( 1 d4 d5 2
knights. Then the knights will have fewer ob c4 e6 3 lLic3 lZ'if6 4 cxd5 exd5), and in the
stacles to reaching c4 or attacking e4. Queen's Gambit Accepted he gets it on move
Of course, in ' extra-positional' terms, White two ( 1 d4 d5 2 c4 dxc4). White also ends up
has the opportunity for violent attacks based with an extra centre pawn in many variations of
upon the pawn advances e5 and f5, and/or sac the English Opening in which Black plays . . .d5
rifices on f5, e6, d5 and b5 . With a single inac (an example would be 1 c4 c5 2 lZ'ic3 lZ'if6 3 lZ'if3
curacy by Black (or merely choosing the wrong d5 4 cxd5 lZ'ixd5). Finally, every Modem Benoni
variation), these attacks can be so powerful as to variation has Black accepting a 2: 1 deficit from
decimate the defence. Otherwise no one would the start ( 1 d4 lZ'if6 2 c4 c5 3 d5 e6 4 lZ'ic3 exd5 5
play White's side of an Open Sicilian. I simply cxd5).
want to demonstrate Black's underlying reason There aren't a great many 2:0 central majori
for accepting a cramped position. See Chapter ties in standard openings, although examples
1 1 on the Sicilian for other illustrations of how do exist. Take the Nimzo-Indian variation with
his central majority functions in diverse situa 1 d4 lZJf6 2 c4 e6 3 lZ'ic3 j.,b4 4 l'lc2 d5 5 cxd5
tions, such as the Paulsen and Dragon Varia l'ixd5 6 lZ'if3 l'if5 7 l'ixf5 exf5 (D) .
tions.
The next diagram shows another type of cen
tral majority in the Open Sicilian arising from 1
e4 c5 2 lZ'if3 d6 3 d4 cxd4 4 lLixd4 lZ'if6 5 lLic3
lZ'ic6 6 j_e2 e5 7 lZ'ib3 j_e6 (D).
Ironically, this position seems to be perfectly

playable for Black. White's difficulty is that
when he finally organizes f3 and e4, Black can
capture on e4 and will have an f-pawn in re
serve to restrain or even attack the centre. There
Although by comparison with the previous are a growing number of openings in which this
example, Black has a more vulnerable structure structure arises.
In conclusion, whether central majorities are This brings us to the minority attack, which
inherent to a specific opening or not, they are involves two pawns attacking three. It is fa
extremely important and tend to assert them mously effective in the Sicilian Defence, in
selves in the long run. Defenders must be sure volving . . . b5 and . . . b4, sometimes supported
to have a clear plan for neutralizing them, some by ... a5, driving away White's knight from c3
times by transforming the structure itself before and/or gaining open files. The exposure of
the pawn-majority can do any damage. White's queenside renders his majority irrele
The corresponding issue has to do with vant in most cases, at least in so far as creating
queenside majorities and minorities, since cen passed pawns is concerned.
tral majorities for one side almost always leave The most famous minority-attack structure
the other side with a queenside majority. Since is 2:3, 2: 1 and 3:3, sometimes called the Carls
most majorities can in principle be transformed bad pawn-structure.
into a passed pawn, it has been said that a
queenside majority is advantageous because the
resulting passed pawn will usually be an outside
passed pawn and thus of special value. That is, in
a king and pawn ending, one king will have to go
chasing after the queenside passed pawn in order
to stop it from promoting, while the other king
mops up on the enemy pawns on the kingside.
Unfortunately, several considerations interfere
with this optimistic scenario.
First, if both kings are centralized (as hap
pens in many endings) neither majority neces
sarily results in a passed pawn further 'outside'
than the other. Secondly, the hypothetical ad
vantage of the queenside majority is reversed if
the parties castle queenside. But since kingside Numerous books discuss the minority at
castling is the rule, a more compelling issue tack by b4-b5 in great detail because its appli
arises that especially impacts the opening (our cation is widespread, although not necessarily
area of concern, after all): the relation of ma in the pure form shown. The most important
jorities to king safety. Since there are more examples that directly conform to the model in
pieces on the board in the opening, the ad the diagram are in the Queen's Gambit Ex
vance of kingside pawns to create a passed change Variation and a few other variations of
pawn carries with it the risk of exposing one's the Queen's Gambit Declined. The Carlsbad
own king; obviously, doing the same with a pawn-structure also emerges in the Nimzo
queenside majority is safer. On the other hand, Indian Defence following 1 d4 llJf6 2 c4 e6 3
the results of a kingside advance may be to put lZ:Jc3 .tb4 4 Wic2 d5 5 cxd5 exd5 6 .tg5 h6 7
the opposing king in danger, whereas defence .txf6 Wixf6 8 a3 .txc3+ 9 Wixc3. Then Black
against a queenside majority doesn't require often feels compelled to play ... c6 in the face of
any compromise of the king's position ! c-file pressure, making White's minority attack
These many considerations suggest a sort of by b4-b5 all the more effective. Interestingly,
theoretical balance between the types of major the Caro-Kann has the same pawn distribution
ities, depending upon concrete features of the with colours reversed after 1 e4 c6 2 d4 d5 3
position. As a practical matter in the opening exd5 cxd5 4 .td3 lZ:Jc6 5 c3 llJf6; in fact, you
stage of the game, one shouldn't pay much at will find an example of a pure minority attack
tention to the matter of majorities and minori by Black in Chapter 12. The most thorough
ties, apart from their value in beginning to discussion of minority attacks in this set of
pursue a specific plan. The odds are that the books will naturally be linked to the Queen's
pawn-structure will be transformed prior to the Gambit Exchange Variation (covered in Vol
onset of the endgame. ume 2).
In addition to this there are related positions. threatening to advance to e5 , breaking up

For instance, a minority-attack situation comes Black' s pawn-structure and opening up the
up in the Griinfeld Defence after 1 d4 tllf6 2 c4 game in favour of White's more aggressively
g6 3 tllc 3 d5 4 tllf3 il.g7 5 .tg5 tlle4 6 cxd5 placed pieces. Naturally that's not the end of
tllxg5 7 tllxg5 e6 8 tllf3 exd5. Then White's the story. Unlike Black in the King's Indian,
strategy is based upon b4-b5, whether or not the Benoni player has the unrestrained bishop
Black gives him a target by playing ... c6. These on g7 and a clear shot at White's e-pawn along
positions share the same basic ideas but natu an open file. Moreover, he has a mobile queen
rally have their own subtleties. side majority that can cause considerable disar
The Modern Benoni provides a good exam ray in White's camp. But understanding the
ple of a central majority versus a queenside ma role of majorities and minorities explains a lot
jority: about these and other openings.
The Light-Sq uare Restra i nt

Structure
Because of their increasing popularity, we'll
take a look at structures with ... c6 and ...e6 ver
sus two white pawns, one on d4 and the other
on either c4 or e4. I'll call these 'restraint
structures' or a 'restraint centre ' , because their
function is to restrain the advance of White's
d-pawn. Four of many openings with versions
of this set-up are:
a) The Caro-Kann Defence: 1 e4 c6 2 d4 d5 .
Now several sequences produce the basic struc
ture; for instance, 3 tllc 3 dxe4 4 tllxe4 and now
White has the central majority, which sets either 4... .tf5 5 tllg 3 il.g6 (with . . . e6 to come)
the stage in and of itself. We know that central or 4... tlld7 5 tllg5 tllgf6 6 il.d3 e6 and similar
majorities are vitally important and generally lines. Another example is 1 e4 c6 2 tllf3 d5 3
underrated. Let's think about the King's Indian tllc 3 .tg4 4 h3 .txf3 5 l'ixf3 e6 6 d4 dxe4 7
Defence ( 'KID ') vis-a-vis the Benoni. They tllxe4 (D) . In these lines White retains his c
both take the same number of tempi to arrive at pawn but not his e-pawn.
their basic position, and in the Benoni Black's
bishop is on a powerful open diagonal whereas
in the King's Indian Black's bishop is blocked
by its own pawn. How can the King 's Indian as
an opening be considered the equal of or supe
rior to the Benoni? I think that the answer rests
mostly with the pawn-majority. In the King's
Indian Defence, Black and White go on pawn
chain assaults. White's attack consists of, for
example, c4, b4 and c5 with tlld2-c4 and cxd6.
What has White accomplished? He has spent
all those moves to create a weak pawn on d6,
but that pawn is only exposed to attack by
pieces, since Black's c-pawn has replaced his
d-pawn. However, in the Benoni Black's pawn
is already sitting alone on d6 without the expen b) The Scandinavian Defence: 1 e4 d5 2
diture of 6 or more moves by White to get it exd5 �xd5 3 tllc 3 �a5 4 d4 tllf6 5 tllf3 .tf5 6
there ! What's more, White's e-pawn is always .td2 c6 7 .tc4 e6. There are numerous variants
THE SIGNIFICANCE OF STR UCTURE 75
of this opening with the same structure, includ Generally, White's first goal is expansion in
ing lines with .....ig4, ... ..ixf3 and . . . e6. In the the centre, in the one case by c4 and d5, in the
Scandinavian, as in the Caro-Kann, White is other by e4 and d5 . These are difficult to achieve
left with a c-pawn but no e-pawn. given Black's pawn-structure, which is specifi
cally designed to prevent d5, and Black is ready
to play ... e5 or ... c5 at the first opportunity. But
White also has other resources, including using
the support-point at e5 (and sometimes at c5) to
make threats and favourably transform the cen
tral situation. Or he can expand on the wings.
In some of these variations, Black's light
squared bishop comes out in front of its pawns.
Then Black already has some freedom for his
pieces and can take more time to play for a
transformation of the pawn-structure. When the
bishop is stuck behind its pawns, as in the
Queen's Gambit or the Caro-Kann with 4. . . ltJd7,
Black needs to get ... e5 or ... c5 in as a freeing
move, preferably sooner rather than later, if he
is to equalize. The ... c5 move not only loosens
c) The Slav/Semi-Slav: in the traditional White's grip on the centre but if followed up
Slav lines we have 1 d4 d5 2 c4 c6 3 lLif3 lLif6 4 by ...cxd4 it claims the c5-square for Black's
lLic3 dxc4 5 a4 ..if5 6 e3 e6 7 ..ixc4, and in the pieces, often a knight. In that case we have
Semi-Slav Meran Variation 1 d4 d5 2 c4 c6 3 something similar to various French Defence
lLif3 lLif6 4 lLic3 e6 5 e3 ltJbd7 6 ..id3 dxc4 7 lines with 1 e4 e6 2 d4 d5 3 lLic3 (or 3 ltJd2
.llxc4. These are two of many examples of the dxe4) 3 . . .dxe4 (or 3 . . .lLif6 4 ..ig5 dxe4) 4 lLixe4
basic structure. By contrast with the first two lLid7 5 lLif3 ltJgf6 6 lLixf6+ lLixf6, where Black
openings, White remains with an e-pawn but no will generally play for . . .c5. If Black can play
c-pawn. . . . e5, he attacks the centre but also frees his
d) The Queen's Gambit Declined: in the light-squared bishop. It's better to show a few
Classical Capablanca and Lasker Variations, examples than to speak in generalities.
we have 1 d4 d5 2 c4 e6 3 lLic3 lLif6 4 ..ig5 ..ie7
5 e3 0-0 6 lLif3 ltJbd7 (or 6. . . h6 7 ..ih4 lLie4 8 Gulko - Lakdawala
.llxe7 'fllxe7 9 .l:!.c l lLixc3 10 .t:rxc3 c6 1 1 ..id3 USA Ch (San Diego) 2004
dxc4 12 ..ixc4 ltJd7) 7 .:te l c6 8 ..id3 dxc4 9
.llxc4. In this opening White again ends up 1 d4 d5 2 c4 c6 3 ltJf3 lLif6 4 lLic3 dxc4 5 a4
with an e-pawn but no c-pawn. ..if5 6 e3 e6 7 ..ixc4 ..ib4 (D)
Here's the . . . c6/. . .e6 structure. Thanks to

Black's control of the centre via ... .tb4 this
may be considered about equal.
8 0-0 0-0 9 l'le2 i.g4!?
9 ... .tg6 is a popular move, preventing e4 for
the moment. Then ...lZ'ibd7 and ... 'iVa5 , with the
idea . . .c5 or ...e5 might follow. But 9 ... .tg4 has
proven quite playable.
10 h3
White grabs the two bishops without delay.
This means that Black will have to do something
in the centre or simply stand worse. Knights are
often the equal of bishops in such positions; it
depends upon the timing.
10 .txf3 1 1 'iVxf3 lZ'ibd7 12 J:!.dl l:!.c8 (D)
..• This seems to win something but there are
tactical problems. Black could justify his two
knight strategy and ... l:!.c8 move by playing
1 5 ... lZ'ib6 ! . Then all of Black's pieces combine
with tempo and he can capture on d5 to better
effect. Still, never underestimate those bishops !
For instance: 16 i.a2 ( 1 6 .txb6 l'ixb6 1 7 l:!.abl
keeps more tension, but the opposite-coloured
bishops don't really help either side's attack
ing chances and therefore the position might
prove drawish in the end) l 6 . . . .txc3 17 bxc3
cxd5 1 8 exd5, and now 1 8 . . .J:!.fd8 isn't clear
because 1 9 c4 ! ? lZ'ixc4 20 .tg5 ! sacrifices a
pawn to maximize the bishops' power. The am
bitious 18 ... .l:i.xc3 19 .td2 l:!.xf3 20 .txa5 l:!.f4 ! ?
2 1 d6 i s also hard to assess. The bishops seem
Black plays a subtle move designed to an to balance out Black's extra pawn. These are
swer a potential d5 by White with ...cxd5. At raw chess fundamentals at work !
the same time ... l:!.c8 lends strength to the ad 16 bxc3 cxdS 17 .txdS (D)
vance . . . c5.
13 e4 eS
The standard idea: Black doesn't wait around
for the centre to become protected and stabi
lized; rather, he wants to break it up and estab
lish strong points for his pieces. The only other
positional solution in such lines is ... c5; one of
these pawn-breaks is just about obligatory.
14 i.e3 l'la5 15 dS! ? (D)
White could delay this thematic push, but
then he would have to deal with . . . exd4 and
. . .lLie5 .
Now we're in another typical and critical
struggle between two bishops with a passed
pawn versus immediate pressure by opportu
nistic knights. The issue is whether the bishops 17 lLicS
•••
can consolidate. The first point is that 17 ... l:!.xc3? 1 8 i.d2 !

15 i.xc3?!
••• .l:i.xf3 19 .txa5 traps Black's rook. On 17 ... lLixd5
18 �xd5 'fllc 7, White has 19 'fllf5 ! (or 19 a5) .tc4 .tc5 4 d3), when Black' s bishop is on c5
19 ... lLib6 20 l:!.c5 (20 .txb6 'fllxb6 21 'fllxe5) and White's bishop goes to e3. Then when
20 ... 'flld6 21 l:he5 g6 22 'flif4 with the idea Black plays . . . ..ixe3 and White recaptures by
22 ... lLixa4? 23 Wih.6 !. These lines show the bish- fxe3 we have the mirror image. I discuss this at
ops in their best light. some length in Chapter 5 .
18 c4 lLixd5 19 cxd5 Returning to our game, Black soon unneces
The opening is essentially over and White sarily straightens out White's pawns for him,
has won it because the restraint upon his centre and creates our restraint centre.
broke down. True, Black has the c-file and a 10 h6 1 1 .tf4 dxe4!? 12 dxe4 'flla5
.••
comfortable knight on c5 but as is so often the The ... c6/. . .e6 centre arises. Since he doesn't
case, the advantage of an ideal centre is trans face the bishop-pair, as he did in the above ex
formed into a powerful central passed pawn ample, Black has more time to organize ...c5 or
that wreaks havoc. . . .e5. Notice that White has no light-squared
19 lLixa4
••• bishop to enforce d5.
After a slow move the bishop and passed 13 'flle2 l:Ifd8 14 a3
pawn are too much; e.g., 1 9 ...l:Ifd8 20 'flig4 14 l:Ifd l �ac8 would be a typical restraint
'it>h8 21 J:Iac l b6 22 l:!.c4 and l:Idc l . position. Black can't undertake much but has
20 'fllf5! f6 dynamic counterplay if White tries to make
Or 20 ... l:Ife8 2 1 d6 l:Icd8 22 l:!.d5. progress. This resilience accounts for the re
21 d6 l:Icd8 22 'flle6+ l:If7?! newed interest in such structures. As this game
Losing, but after 22 ... @h8 23 d7 the pawn is shows, the drawback is that it's difficult, but not
strong, backed up by the advantage of bishop impossible, to get positive chances.
versus knight. l:Iac 1 -c8 is one problem. 14 ltJfS!? 15 h3 ltJg6 16 .th2
•.•
23 l:Idcl @rs 24 J:Ic8 b5 25 l:!.acl 1-0 Bishops in many openings are stuck on the
side of the board at g3 and h2. This one appar
Bogoljubow Kramer
- ently has good scope but it doesn't defend the
Travemiinde 1 951 d-pawn. Therefore 1 6 ..ie3 looks better, cen
tralizing and intending 16 ... ltJhS 1 7 lLie5 ! .
1 d4 lLif6 2 .tg5 d5 3 lLic3 c6 4 e3 .tf5 5 .td3 1 6...l:Id7 1 7 l:Ic2 ltJh7! (D)
.txd3 6 cxd3 e6 7 lLif3 .te7 8 0-0 0-0 9 .:tel
ltJbd7 10 e4 (D)
A great idea! Black wants to play ...lLig5 and

eliminate White's best piece on f3, the defender
Here's a case of the immobile centre that we of the d-pawn.
see in the mirror-image form of e3/e4/d4 versus 18 'flle3 .t:rad8 19 .tg3 ltJg5 20 lLid2
a black pawn on e5 in some I e4 e5 openings. White's protects his d-pawn indirectly and
For example, that situation typically arises in a plans a logical reorganization of the position.
Giuoco Piano with d3 ( 1 e4 e5 2 lLif3 lLic6 3 He almost achieves it.
20 'Yi'b6
•.• 16 h6 17 'Yi'b4 'Yi'b6 18 cS!? 'Yi'xb4 19 axb4
.•.
Not 20 . . . .l:i.xd4? 2 1 tllb 3. exd4 20 cxd4 .l:i.e4 21 i.f3!?

21 tlle2 cS! 21 .l:i.adl .l:lae8 22 Wfl is solid and equal.
Just in time, Black manages to get this move 21 .l:i.xd4 22 bS! tllxcS 23 bxc6 bxc6 24
•••
in with the help of tactics. Of course, the nor i.xc6 .l:lc8 2S .tbs i.d3!?
mally dangerous response d5 isn't remotely With White's two bishops gone, Black has
possible. won a pawn for very little, but White manages
22 tllc4 to scare up play.
White cedes a pawn but what else? 22 dxc5 26 .txd3 tllxd3 27 .l:i.edl .l:i.d7 28 f3 .l:lcd8 29
i.xc5 ! 23 'Yi'xc5 (23 .l:lxc5 .l:ld3 ! ) 23 . . .'Yi'xc5 24 Wfl tllb4 30 .l:i.xd7 .l:lxd7 31 .tf2 tllc6 32 .l:la6
l:!.xc5 .l:lxd2 threatens e2 and b2. .l:lc7 33 i.g3 .l:i.c8 34 .lif2 .l:lc7 3S i.g3 .l:i.c8 36
22 'Yi'c6 23 f3 cxd4
.•. .lif2 tllb 4!? 37 .l:lxa7 .l:i.cl+ 38 We2
Black's strategy has succeeded. . . . tllh7-g5 The game is about equal and was eventually
was quite a blow to White's position. drawn.
24 'Yi'd3 tllh7 2S f4 tllf6 26 tlld2 'Yi'xc2! 27
'Yi'xc2 d3 28 'Yi'c4 dxe2 29 .l:lel .l:lxd2 Djuric - Larsen
Black is winning. Copenhagen 1 979
G. Lee - Taulbut 1 e4 dS 2 exdS 'Yi'xdS 3 tllc3 'Vias 4 d4 tllf6 5

British Ch (Morecambe) 1 981 tllf3 i.fS 6 .tc4 tllbd7 7 'Yi'e2 e6 8 .td2 .tb4! 9
a3 0-0 10 0-0 i.xc3! 11 .txc3 'Yi'b6 (D)
1 e4 dS 2 exdS 'Yi'xdS 3 tllf3
Instead of the usual 3 tllc 3.
3 tllf6 4 d4 .trs s .te2 e6 6 0-0 c6 7 .tf4
•••
tllbd7 8 c4 'Vias 9 tllc3 (D)
This was a shocking idea at the time: giving

up the bishop-pair and accepting less space at
the same time ! But the . . . c6/ ...e6 structure is
very handy for such a position: knights are tem
In this situation, White has a c-pawn, not an porarily as good as bishops and ...c5 or ...e5 is
e-pawn as in the last two examples. Although not to be stopped forever.
both configurations arise regularly and have 12 .tb3 as 13 .ta4 c6 14 i.d2!
differences, Black's main strategy is still to get Rerouting from a passive square to a nice
. . . e5 or . . . c5 in, and White would like to play d5 . lengthy diagonal is logical.
9 .lib4 10 'Yi'b3 0-0 11 tlleS!? tllxeS 12
•.. 1 4 h6 lS i.e3 .l:i.fe8 16 c3
.•.
i.xeS tlld7 13 i.g3 eS! 16 c4 is the thematic move. Then Black might
He has to play this way to get counterplay. think about exchanging off his other bishop by
White will transform the centre in response. 16 ... .tg4 ! ? ( 1 6 . . . 'Yi'c7 would prepare ... e5 and
14 a3! .lixc3 lS bxc3 .l:i.fe8 16 .l:i.fel also makes sense) 1 7 h3 i.xf3 1 8 'Yi'xf3 'Yi'a6 ! ?
'Yi'xb7 is an option here and next move. 19 .l:i.ac 1 b 5 20 cxb5 cxb5 21 .tc2 tlld5 and this
IQP position is hard to assess, but I think that

Black can be satisfied.
16 .tg4
••.
Now ... ..ixf3 is threatened because if the

queen recaptures, ... �xb2 works. In what fol
lows Black makes the ... e5 break and exchanges
off a pair of bishops with full equality.
17 l:!.abl 'fllc7 18 h3 .th5 19 .tc2 e5! 20 g4
exd4 21 cxd4 .tg6 22 .txg6 fxg6 23 'flid3 ltJd5!
24 .txh6! gxh6 25 Wlxg6+ @rs 26 'flixh6+
@gs 21 Wlg6+ @rs 2s 'flih6+ @gs 11i-11i
Black seemed to stand perfectly well through
out.
There's a better-known version of this c4/d4 16 .tb3 'fllf6!?

structure: Considering what happens, there's really no
reason for this.
Matanovic - Petrosian 17 'flle2
Kiev (USSR-Yugoslavia) 1959 White could have admitted to his difficulties
and exchanged the knight on d5. But the posi
1 e4 c6 2 lLic3 d5 3 d4 dxe4 4 lLixe4 ltJd7 5 lLif3 tion seems so innocent.
ltJgf6 6 lLixf6+ lLixf6 7 .tc4 .tf5 8 'flle2 e6 9 17 .t:rd7 18 c3
•••
.llg5 .te7 10 0-0-0 .tg4! (D) The advance 1 8 c4 is way too committal and
weakens d4: 1 8 ... lLie7 1 9 .tc2 .t:rfd8 20 'flid3
ltJg6 and the pawn will fall.
18 b5!
.••
A simple idea designed to prevent c4, and

Black also has in mind a minority attack with
... b4.
19 g3 .t:rfd8 20 f4! ?
White stops ...e5, but that's not the only
pawn-break.
20 b4! 21 'fllf3 bxc3 22 bxc3 c5!
•.•
Once Black achieves this he already has the

advantage.
23 .t:re5
Black penetrates White's position after 23
c4?! ltJb4! 24 dxc5 Wif5+ 25 'it>al ltJd3. And 23
A familiar idea. A knight on d5 becomes as .txd5 .t:rxd5 threatens the d-pawn. White's king
strong as a bishop. is none too safe either.
11 h3 .txf3 12 'fllxf3 lLid5 13 .txe7 23 cxd4 24 .txd5 .t:rxd5 25 .t:rxd5 exd5! 26
•••
A very important point is that Black will get .t:rxd4 h6 (D)

a great attack if White tries to conserve his White has managed to exchange down into
bishpp-pair: 1 3 ..id2 b5 14 .tb3 a5 ! . an isolated queen's pawn position, but his king
1 3 'fllxe7 1 4 .t:rhel 0-0 1 5 @bl
•.• is too exposed.
15 .txd5 Wlg5+ 16 'it>bl cxd5 is equal. 27 g4
15 .t:radS (D)
••• 27 .t:rxd5 can be answered by 27 ... .l::tb 8+ 28
Here we have a d4- and c-pawn versus 'it>c2 Wlb6 ! and Black's attack will be too
. . . c6/. . . e6 again. Obviously White needs to play strong.
c4 if he's going to claim any advantage, but 27 �e7! 28 �f2 .l::tb8+ 29 'it>al �a3 30
.•.
Petrosian has a way of dealing with that. �c2 .l::te8 31 .t:rb4 d4! 32 .l::txd4 .l::te l+ 33 .l::td l
you might recognize that idea right away and

take action against it, as Iordachescu does.
w 10 a4!
The same technique as in the Meran.
10 a6
•••
Now ... c5 is prevented for a while.

1 1 .l:i.el .te7 12 lZ'ibd2
White seems to be planning a stock attack by
lZ'ie4 but Black's next move changes his mind.
12 'Yi'b6?! (D)
•.•
12 ... 0-0 13 lZ'ie4 'Yi'b6 is better since it takes

the bite out of c4 .
.l:i.xdl+ 34 'iVxdl 'Yi'xc3+ 35 @bl l'ixh3 36 a4

hS! 37 gxhS 'iVfS+ 38 @b2 'iVxf4 39 @b3 'iVfS
40 @c4 @h7 41 'Yi'd2 0-1
The opening of the following game com

bines this ...b5 idea with our earlier theme of
the fight between an isolated d-pawn and iso
lated c-pawn:
lordachescu - Wohl
Naujac sur Mer 2002
1 e4 lZ'if6 2 eS lZ'idS 3 d4 d6 4 lZ'if3 dxeS S lLixeS

c6 6 .tc4 lZ'id7 7 lZ'if3 e6 8 0-0 13 c4!
Again we've arrived at the . . .c6/... e6 re White takes on an isolated and fully block
straint structure, coming from a slightly unusual aded pawn on d4. But having seen this d4 ver
source. Now Black makes a very committal but sus c6 structure before (hopefully many times)
logical move: he assesses this as a favourable isolated queen's
8 ...bS!? 9 .td3 .i.b7 (D) pawn position. Black is well-developed, and if
he gets ...c5 in it will open up the b7-bishop and
activate his game. The issue then is whether
White can make use of any particular advan
tages that he has in advance of that freeing
move. The dark squares and aggressively-placed
pieces look good, so the first question is: where
is Black weak? The squares c5, e5 and d6 may
be vulnerable, and if you've foreseen the move
15 .tg5 before playing 1 3 c4, that should be
enough to convince you to go ahead.
13 bxc4 14 lZ'ixc4 'Yi'c7 15 .tgS! (D)
••.
15 cS?!
•.•
Black's position is still solid, so he shouldn't

allow the exchange of dark-squared bishops.
Other moves are 15 ...lZJ7f6 16 l:tc l 0-0 and
The bishop may not seem to be doing much 15 ... .tb4, just to get castled. The b4-square is a
here, but Black wants to play ... a6 and ... c5. If nice outpost for Black, who has a future ... .l:i.b8
you know the Meran Variation of the Semi-Slav in mind.
B w
16 l:Icl ! ? 0-0 17 il.xe7 0,xe7 18 0,ce5 l:Iad8 This creates a weakness on e6 but otherwise
19 b4 White can squeeze Black by expansion on ei
Admirably sticking to his purpose, even ther or both wings.
though the tactic 1 9 il.xh7+ �xh7 20 0,g5+ 19 exf6 il.xf6 20 �c4! �b6 21 b4 0,a6 22
'it>g8 2 1 0,xd7 ! does ultimately win after com 0,e4!
plications. White has the advantage. He can exploit the
19 0,xeS 20 0,xe5
.•. weakness on e6, or play for a well-timed 0,c5.
Black's c-pawn falls, and the opening is Spassky went on to win the game.
over. Previous knowledge of the properties of
... c6/... e6 restriction and the standard IQP posi Space and Structure
tion, as well as recognizing the similarity to the
Meran Variation, undoubtedly helped White to The relationship of space to structure is poten
find his way in this game. That is an illustration tially an immense subject, but I just want to
of what I call 'cross-pollination' , discussed be make a few comments about it. We know that
low. White is the one who will generally grab more
space in the opening (particularly in the major
Rather than trying to fight directly against openings discussed in this book). Several situ
the ... c6/...e6 complex, it's sometimes better to ations can arise for Black. In the Closed Sys
give up on d5 and transform the structure. In tem of the Ruy Lopez and several other double
this famous game White does so by using his e-pawn openings, Black's strongpoint on e5
support-points: (based upon the pawn-chain c7-d6-e5) and his
b5-pawn establish a sufficient command of ter
Spassky - Petrosian ritory that he doesn't usually feel the need to
Moscow Wch (13) 1966 acquire more. The Chigorin set-up with ...0,a5
and ... c5 is an exception, in that it is clearly
1 e4 c6 2 d4 d5 3 0,c3 dxe4 4 0,xe4 il.f5 5 0,g3 aimed at extending Black's territorial reach; but
il.g6 6 h4 h6 7 0,f3 0,d7 8 h5 .li.h7 9 il.d3 that this policy is not necessary is shown by the
.li.xd3 10 �xd3 �c7 11 il.d2 e6 12 �e2! 0,gf6 popular Breyer, Zaitsev, M0ller and Smyslov
13 0-0-0 0-0-0 (D) Variations (see Chapter 8 on the Ruy Lopez for
14 0,e5! 0,xe5 15 dxe5 0,d7 16 f4 examples). To some extent this is also true with
White stands well. He has more space and no the double d-pawn openings such as the Queen's
worries about the kind of central attacks that Gambit Declined and Slav. Nevertheless, in the
we've seen from Black. Of course, White still traditional Queen's Gambit variations Black
needs to break through Black's defences; he tends to play for ... e5 at some point, arguably
does so by creating another support-point on exchanging one type of territorial control (the
c5 . d5-pawn) for another that also activates his
16 il.e7 17 0,e4 0,c5 18 0,c3 f6! ?
•.. pieces. In the Dutch Variation of the Slav ( 1 d4
d5 2 c4 c6 3 lllf3 lllf6 4 lllc 3 dxc4 5 a4 iLf5 6 Variations, in which White already has control
e3), Black is generally in no hurry to play ... e5 of space with pawns on c4 and e4, Black finds it
or even ... c5, which also true of several of the a little more urgent to achieve ...b5 or ... d5, or at
other ... e6/... c6 restraint openings that we saw least threaten to do so.
above, especially since his queen's bishop is Almost every opening can be looked at_ in
outside his pawn-chain. this way, that is, how vital is it for the side with
By contrast, look at many of the other major less space (usually Black) to win space, and
d-pawn openings. In the King's Indian Defence how quickly? What about the need for White to
main lines (e.g., 1 d4 lllf6 2 c4 g6 3 lllc 3 ii.g7 4 take on more space quickly, or can he be pa
e4 d6 followed by ...0-0 and ... e5), once White tient? If you understand the urgency (or lack of
takes space in the centre, Black will seldom be it) in achieving these goals, you will have a
satisfied that the single central pawn on e5 fully much better feel for the logic and timing behind
represents his interests in that sector. Without the opening moves.
further pawn moves he will slowly be strangled
by White's central and queenside pawn ad Cross- Pol I ination
vances. Therefore you will almost always see a
rapid .. .f5, or in some cases an attempt to take Sometimes manoeuvres and positional ideas
over territory on the queenside by ... c6 or ...c5. will arise across openings that are not specifi
Likewise in the Modern Benoni (1 d4 lll f6 2 cally related, a phenomenon that I call 'cross
c4 c5 3 d5 e6 4 lllc 3 exd5 5 cxd5 d6 followed pollination' . We have seen repeated examples
by . . . g6, . . .ii.g7, . . . 0-0, etc.), Black can almost of structures that show up in various openings,
never be satisfied with the central control of and in a way everything that we've seen about
fered by his c5-pawn. In most variations he is structures to this point has involved cross
almost compelled to win more space by ... b5 or pollination, that is, every structure has been re
.. .f5 or get strangled by White's pieces and on lated to other structures. Here I'll briefly discuss
rushing pawns. In the Semi-Slav, a combina the process that may lead you to recognize such
tion of . . .dxc4 and ... b5, or ... dxc4, ...ii.d6 and similarities and therefore play an unfamiliar or
...e5 is customary before White extends his con only partly familiar variation with increased
trol over the central squares (note that Black's confidence. Grandmasters are very good at see
light-squared bishop is trapped behind his ing this type of relationship in subtle ways.
pawns). You'll gain a lot from the very process of using
What about the Sicilian Defence? In general, your study and experience from one position
if he has the ... e6/. . . d6 centre, Black is in a re and then applying it to another. All the more
markable lack of hurry to take on more space. reason to keep your opening knowledge broad
At most he will play ...b5, and if White stops and not overspecialized.
that by playing a4 it is hardly a matter of great As an example, you've probably wondered
concern. But look at White's various strategies whether to play with an isolated queen's pawn
against the Sicilian. It seems practically man in a given position. This requires judgements
datory to expand his reach over the board. Re based upon experience. We already know that
cently there are players who set up with f3, g4, the isolated pawn offers similar lessons across a
g5 and h4 (and even h5 and g6) against the ma wide range of openings. We even see standard
jority of Sicilian variations. Traditionally, f4 IQP positions that are essentially the same in the
has been a standard way of proceeding, with f5 Nimzo-Indian, Caro-Kann, Sicilian and Queen's
to follow or perhaps e5 (although the latter is Gambit. But you'll consistently be given the
sometimes more of a tactical device, because option of deciding whether a new IQP position
the pawn will seldom stay on e5 long enough to in a foreign position has more good features
be a true claimant of territory). These days than defects, and experience with other open
there are also more combinations of f4 and g4. ings will do more than an author's generalities
Barring those kingside moves, White will at can.
least play a4 to stake out some space on the A more interesting illustration of cross
queenside. In the Maroczy Bind and Hedgehog pollination relates to decisions about when to
bring your queen out, and whether you can do whether to play ... d5 and/or recapture with the
so productively at an early stage. If, as Black, queen on that square. From White's point of
you've captured some 'poisoned' pawns on b2 view we have such things as 1 c4 e5 2 g3 lllf6 3
or gambited them as White, you'll certainly get J.g2 c6 4 d4 cxd4 5 �xd4.
a better feel for when to take the risk in either Say that you're playing the French Defence
way. Here are a few examples that you might and start out 1 e4 e6 2 d4 d5 3 llld2 lllc 6, re
run into: cently a hot variation. Maybe you have some
1 e4 c5 2 fllf3 d6 3 d4 cxd4 4 lllxd4 lllf6 5 recent analysis on 4 lllgf3 and 4 J.b5 that you
lllc 3 a6 6 J.g5 e6 7 f4 �b6 8 �d2 �xb2 want to try out. When your opponent plays 4
1 d4 lll f6 2 lllf3 e6 3 J.g5 c5 4 e3 �b6 5 c3, you don't recognize the move, but search
lllbd2 �xb2 your pattern database and come up with 4 ... e5 !
1 e4 g6 2 d4 J.g7 3 lllc 3 d6 4 f4 c6 5 lllf3 5 exd5 �xd5. Pattern recognition could also
.\ig4 6 J.e3 �b6 7 �d2 �xb2 be involved if you play the Pirc Defence and
1 e4 e6 2 d4 d5 3 e5 c5 4 c3 fllc6 5 lllf3 �b6 are confronted with 1 e4 d6 2 d4 lllf6 3 J.d3 .
6 J.e2 cxd4 7 cxd4 lllh6 8 J.xh6 �xb2 Playing 3 ... e5 is fairly obvious, and then White
1 e4 c6 2 d4 d5 3 e5 J.f5 4 J.e3 �b6 5 llld2 plays 4 c3. What now? If you're attuned to the
�xb2 way that a pawn on c 3 prevents lllc 3, you
1 d4 lllf6 2 J.g5 c5 3 d5 �b6 4 fllc 3 �xb2 5 might see 4 ... d5 ! , with the idea 5 exd5 �xd5
.lid2 or 5 dxe5 lllxe4 (D). This looks fun and worth
1 d4 lllf6 2 J.g5 llle4 3 J.f4 c5 4 d5 �b6 5 a try.
llld2 �xb2 6 lllxe4 �4+ 7 �d2 �xe4 8 c3
Or, with colours reversed:
1 d4 d5 2 c4 J.f5 3 �b3 e5 4 �xb7
1 d4 d5 2 c4 dxc4 3 lllf3 lllf6 4 e3 J.g4 5
.lixc4 e6 6 �b3 J.xf3 7 gxf3 lllbd7 8 �xb7
1 d4 lllf6 2 c4 c5 3 d5 e6 4 fllc 3 exd5 5 cxd5
d6 6 e4 a6 7 a4 g6 8 lllf3 J.g4 9 �b3 J.xf3 1 0
�xb7 lllbd7 1 1 gxf3
They are of differing soundness and strength.
If you get a new position in which you are being
offered a b-pawn in the opening, you can make
a better decision by studying these.
Another question: when do you want to al
low your queen to come out with the move
. . . �xd5 or �xd4 within the first few moves of
the game? What about that rule that says the But then you notice 6 J.xe4 dxe4 7 �a4+
queen shouldn't come out too early? Maybe as followed by �xe4, shake your head, and play
a beginner you have seen or read about the Dan some other 4th move. This is where the stock
ish Gambit line 1 e4 e5 2 d4 exd4 3 c3 d5 4 exd5 of familiar positions comes in. Two weeks
�xd5. The c3-square is temporarily occupied later you happen to notice a grandmaster in
and thus there's time for Black to develop be this position as Black and after a short think he
fore his queen is attacked; for instance, 5 cxd4 plays 4 ... d5 anyway. There follows 5 dxe5
lllc6 6 fllf3 J.g4 7 J.e2 lllf6 8 lllc 3 J.b4 (this lllxe4 6 J.xe4 dxe4 7 �a4+ J.d7 ! 8 �xe4 J.c6
position also arises in the Goring Gambit) and (D) with plenty of compensation (two bishops,
Black has equality. Later you see similar ideas light squares, and direct attack on g2).
in the Sicilian Defence, where we have 1 e4 c5 Our grandmaster didn't give up on the line
2 c3 d5 3 exd5 �xd5 and c3 is occupied so that after he saw 7 �a4+; was this due to seeing fur
White can't place a knight there with tempo; of ther than the club player? Probably not, because
ten 4 d4 lllf6 5 lllf3 J.g4 will follow. Perhaps just about every grandmaster and international
the improving student will start to examine the master has seen this kind of sequence before.
c3-square as one strong criterion in deciding For example, there are a couple of classic games
think that it's fair to say that the reason that

Black can get away with this is that the knight
isn't all that well placed on c3, such that Black
can play moves like ...ii.f5, ...e6 and ... c6 at
some point, when White would prefer to have
his c-pawn free to advance and increase his
central control. Or the queen, when attacked,
may use the tempo 'lost' to make a second pro
ductive move. A good example comes up in the
line 1 c4 e5 2 g3 lllf6 3 lt:Jf3 e4 4 lt:Jd4 lllc6 5
lllc2 d5 6 cxd5 Vi'xd5 7 lllc 3 Vi'h5 ! intending
...ii.h3, when Black has an excellent game.
That leads to many other examples, such as
those in which a knight on c3 (or ... c6) is pinned,
with 1 c4 lllf6 2 lllc 3 e6 3 e4 d5 4 cxd5 exd5 5 so that a queen can come to d4 (or d5). A well
e5 llle4 6 lllxe4 dxe4 7 Vi'a4+ il.d7 8 Vi'xe4 known case is the Nimzo-Indian line with 1 d4
ii.c6. The broader your exposure to typical lllf6 2 c4 e6 3 lllc 3 ii.b4 4 l!Vc2 d5 5 cxd5
structures, the better you'll be able to handle Vi'xd5 ; and a related one is the Chigorin De
unfamiliar situations. See Chapter 14 on the fence with 1 d4 d5 2 c4 lllc6 3 cxd5 l!Vxd5 4 e3
Pirc Defence for more details about this varia e5 5 lllc 3 ii.b4. From the white side, we have a
tion; the next thing that happens is that it turns Sicilian Defence with 1 e4 c5 2 lllf3 d6 3 d4
into an Open Variation of the Ruy Lopez! cxd4 4 Vi'xd4 lllc6 5 il.b5, which we might
Along the same lines (bringing queens to d4 compare with a Philidor Defence 1 e4 e5 2 lt:Jf3
or d5), a tricky anti-Sicilian variation goes 1 e4 d6 3 d4 exd4 4 l!Vxd4 lllc6 5 ii.b5. In both cases
c5 2 lllf3 d6 3 c3 lllf6 4 ii.d3 ! ? lllc6, when the queen is allowed to stand her ground, but
some players may not want to face ... ii.g4, so often at the cost of the bishop-pair. Do you spot
they play 5 h3 . But with that pawn on c3, 5 ... d5 the main difference? In the Sicilian line, Black
should be considered, with the idea 6 e5 llld 7, keeps his central majority intact; in the Philidor
and nowWhite can play 7 ii.b5 Vi'b6 (a French Black surrenders the centre. After a while it be
Defence pattern), or he might enter into the se comes second nature to look for these situa
quence 7 e6 ! ? fxe6 8 lllg 5, a tactical ploy that tions, and advanced players do so.
arises in a good half-dozen other opening varia Cross-pollination between 1 d4 and 1 e4 is
tions. With experience in any of those, you may more common than you'd think. The chess
be helped by recognition of associated patterns player with some experience may have noticed
such as 8 ... lllf6 9 lllxh7 (9 ii.xh7 lllxh7 10 that the Benoni pawn-chain ... c5/... d6 versus
Vi'h5+ 'it>d7 1 1 lllxh7 'it>c7, etc.) 9 . . .l!Vd6 (or White's e4/d5 will often arise in the King's In
maybe 9 ... lllxh7 10 Vi'h5+ 'it>d7 1 1 il.xh7 b6, dian Defence, after, for instance, 1 d4 lllf6 2 c4
even if you haven't seen that one before). The g6 3 lllc 3 ii.g7 4 e4 d6 5 ii.e2 0-0 6 ii.g5 c5 7 d5
fact that you've seen and/or played other posi h6 8 ii.e3 e6 9 lllf3 exd5 10 cxd5 and in several
tions with the e6 move helps you to make more other major lines. But if you're playing the
accurate calculations and gives you confidence black side of a Ruy Lopez, you might consider
that the resulting positions should be fine for heading for this same structure by way of vari
Black. ous Closed lines. For example, in the Keres
Variation you may arrive at this main position:
There are plenty of other cases of an early 1 e4 e5 2 lllf3 lllc6 3 ii.b5 a6 4 ii.a4 lllf6 5 0-0
'it'xd4 (or ... l!Vxd5) in which the c3-square (or ii.e7 6 .l':tel b5 7 �b3 0-0 8 c3 d6 9 h3 lt:Ja5 10
c6-square for Black) isn't occupied. The sim ii.c2 c5 1 1 d4 llld7 ! ? 1 2 lllbd2 exd4 1 3 cxd4
plest of these is the Scandinavian Defence lllc6 14 d5 lllce5 (D).
(Centre Counter) 1 e4 d5 2 exd5 l!Vxd5, when 3 The Benoni structure has arisen and you al
lllc 3 forces the loss of a tempo with the queen, ready have the move ... b5 in. That's the key
still out early and subject to further attack. I move in nearly every Benoni and very often
flowering, is the phenomenon of semi-waiting

moves in the opening. That is, moves that serve
a definite purpose but only just so, and which
seem to need the opponent' s cooperation to
take on meaning. It is fascinating to see, how
ever, that these moves are a little more effective
than my description would imply, i.e. the oppo
nent hasn't really the luxury of doing nothing in
return without giving ground. A lot of these
ideas are unassuming; for example, develop
ment of pieces to the second rank that appear to
have five good answers and yet are hard to
meet. Or a sequence of moves that seems to
lose a tempo but puts the opponent' s pieces
White will prevent it. All Black needs to do somewhere they' d rather not be; for instance, in
next is 'fianchetto' his bishop by ... J.f6 and he d-pawn and c-pawn openings with an early ... e6
will obtain an excellent game. White doesn' t we see many new cases of ...i.b4+ followed by
want to allow this and plays 1 5 lllxe5, but as it ...J.e7, and ... J.a6 followed by ... J.b7. In the
turns out that frees Black' s game or at least Sicilian Defence and English Opening, Black
gives his pieces places to go. always seems to be playing ... J.c5 or ... i.b4
Cross-pollination will appear in contexts that followed by ... i.e7.
are not strictly structural, but relate to the scope I find the little rook's pawn moves to be par
of plausible structures. I think that a lot of this ticularly thought-provoking, and I suspect that
shows up in the opening preparation of players grandmasters are finding inspiration from such
and their borrowing of ideas from each other. moves' success in some openings to experiment
Grandmaster X will see a new move that Grand with them in others. These are not necessarily
master Y has played on the 1 8th move of a cer new moves but often obscure older ones which
tain variation of the Sicilian. Then he may apply later received general acceptance. For instance,
that move to his 14th move in a closely-related Kasparov's strengthening of Petrosian's little
variation. That is an interesting exchange of move 1 d4 lllf6 2 c4 e6 3 lllf3 b6 4 a3 ! ? led to an
ideas, of course, but it's more exciting to see explosion of games and investigations, and 4 a3
players latch on to the same moves or general has been going strong in the Queen's Indian
ideas across the range of openings. The number Defence for many years now. Variations such as
of older, well-known, positions in which White 1 c4 c5 2 lllf3 lllf6 3 d4 cxd4 4 lllxd4 lllc6 5
has recently found and played the move g4 can lllc 3 e6 6 a3 ! ? began to appear. Then some
not be coincidental. Whole articles have been years later players got serious about the mod
written about this move appearing in so many est-looking ... a6 within the first four moves in
new and interesting contexts. The list of open two variations of the Slav: 1 d4 d5 2 c4 c6 3
ings thus affected includes several variations of lllf3 lllf6 and here 4 e3 a6 or 4 lllc 3 a6. Not
the English Opening, the Semi-Slav, the Two only that, these two moves have now accumu
Knights Defence, the Bogo-Indian Defence, the lated analysis and playing experience that rival
Dutch Defence, the Caro-Kann Defence, and the main lines of some openings! In that case,
just about every variation of the Sicilian De Black wants to play ... b5 to gain space, or cap
fence! And I could make a similar although ture on c4 and then play ... b5 . He may also want
shorter list of openings in which Black has be to play his bishop out to g4 or f5 and not worry
gun to use the move ... g5. Obviously, once the about �b3, answering that move with .. J1a7 in
idea struck players' imaginations they began to some variations! Another example: the varia
look for it in every position. tions with 4 ... a6 in the Modem Defence are a
Something that has struck me about chess little insulting to the classical thinker, but re
from the last several decades, actually stretch freshing; e.g., I e4 g6 2 d4 J.g7 3 fllc 3 d6 and
ing back more than a century but only recently now 4 i.e3 a6, 4 f4 a6, 4 fllf3 a6 or 4 J.g5 a6,
and so forth. These all seem fully playable, in everywhere throughout them. In other words,
part because . . . c5 can follow and 'threaten' to experiments like these are successful only be
go into a favourable Sicilian Defence, an ex cause of the vast knowledge of traditional open
ample of cross-pollination. Recently players ings that lets players find old patterns in new
started looking at long-established openings contexts. The moral of the story is not to play the
and found a new idea or rediscovered it in older move a3 in every position (or any position! ), but
literature; for instance, 1 e4 e5 2 lbf3 lbc6 3 to realize that mastery of openings comes from a
lllc 3 lllf6 4 a3 ! ?. This is another waiting move broader set of structures and techniques that ap
that doesn't do much but achieves a little some pear across the board. While you study the tra
thing; for example, 4 ...ii.c5 5 lllxe5 ! , when the ditional openings, be sure to look at ideas from
resource ... ii.b4 isn't available after 5 ...lllxe5 6 every other source to reinforce what you're
d4. Or in the Pirc Defence, the remarkable 1 e4 learning.
d6 2 d4 lllf6 3 lllc3 g6 4 f4 ii.g7 5 a3 !?, prevent Furthermore, you can look at structural
ing the usual 5 ... c5 in view of 6 dxc5 l!Va5 7 b4, themes in the same way, comparing them from
and otherwise waiting for Black to make a com opening to opening. The more that you examine
mittal move, of which it turns out that many and compare outposts and support-points, for
have disadvantages. In the Sicilian Defence, 1 example, the more you will find yourself able
e4 c5 2 lllf3 lllc6 3 d4 cxd4 4 lllxd4 lllf6 5 lllc 3 to work with them. Ask simple questions when
e6 became a popular way to avoid major Sicil you play over games by grandmasters: when
ian theory; after a century of experience with are outposts on squares like e5, d5, e4 and d4
that line players noticed the possibility of 6 a3, similar, and how do they differ? Does the out
preventing 6 ... ii.b4 and again waiting to see post piece radiate influence and make counter
what Black is going to do. There' s a current in play fruitless? Can the outpost be maintained?
terest in 1 e4 c5 2 a3 (not to mention 2 llla3 ! ?), Can a piece on the outpost be exchanged off fa
and even a monograph devoted to it. Similar vourably in order to change the pawn-structure?
things have been going on with Black. In the Is there a situation in which the outpost can be
French Defence with 1 e4 e6 2 d4 d5 3 lllc 3 'played around' , leaving an impressive-looking
(and 3 llld2), grandmasters have been using but uninvolved piece occupying it? Similarly, is
3 ...h6 (the other rook' s pawn !), asking White to a piece on an outpost in front of doubled or
commit while preventing il.g5 and finding the backward pawns so powerful that it's worth a
move ... g5 useful in a remarkable number of rook, or will it just sit there and block one's own
positions. Likewise, Anand and many others play? Either result is possible.
have played 1 c4 e5 2 g3 lllf6 3 ii.g2 h6. Cross-pollination turns out to be an unlimited
It seems obvious that these sorts of ideas feed subject and contributes to the fact that we take so
off each other, with each new explorer inspired strong an interest in chess. There are examples
by the most recent discoveries. But if you look at throughout this book and in most sources of
the details of the newly discovered theory and chess information. Keep an eye out for them, es
practice of such lines, you will see that standard pecially as you study and play openings. You'll
structures from other chess openings appear find it a fun exercise, and helpful for your chess.
4 Introduction to 1 e4 and the Open
Games
Want to play a game of chess? I'll move first: These are not exactly compelling reasons for
1 e4 (D) 1 e4 to have ascended to the throne of the open
ings realm. Maybe we should think on an even
more fundamental level. What's the first goal of
opening play? To control the centre. And what's
the best way to do that? To set up an ideal cen
tre. There are only two moves involved in that
project: e4 and d4. To some extent, playing the
one creates the threat to play the other. Thus,
playing one of these two moves right away nar
rows Black's set of logical responses and in
some sense establishes a degree of control. At
that point there are various advantages to either
move, and indeed 1 d4 is White's second most
popular opening move by a landslide. The over
all preference for 1 e4 then comes down to
more subtle factors, and I may as well cite the
Advancing the e-pawn two squares is the old obvious fact that in the great majority of open
est and still the most popular way to begin the ings, 1 e4 prepares the way for kingside cas
game. Beginners who know little more than the tling more quickly than does 1 d4.
rules proudly play 1 e4 before they start losing Now things get a little more complicated.
their pieces. Chess in the movies is dominated Notice that the e4-pawn is undefended. Not
by e-pawn play. The majority of the world's top surprisingly, Black will often attack it and try to
ten players use 1 e4 more often than not. compel White to spend a move protecting his
What's so great about this move? On the pawn. This immediate vulnerability is not
most basic level, 1 e4 fights for control of the shared by other popular first moves by White
key central square d5, and it frees the fl -bishop such as 1 d4, 1 c4, or 1 lllf3. Hence Breyer's
to join the fray. Indeed, in the 1 e4 e5 openings proclamation that 'After 1 e4, White's game is
that dominated chess practice for so many years, in its last throes' ! That is melodramatic, of
we find the bishop being developed at an early course, but it does reflect the direction in which
stage. Surprisingly, however, that doesn't hold Black's defences will tend to go. He will gener
true for most of Black's other defences to 1 e4. ally create threats to White's e-pawn, usually
What other advantages stand forth? Well, mov by the move . . . lllf6 or by ... d5 . We find such an
ing the e-pawn also opens up the d l -h5 diago attack on White's e4-pawn in most of the major
nal for White's queen to come out on, although defences to 1 e4, usually within the first two or
she doesn't use that privilege much in the early three moves of the game. For example:
stages, so as not to become an object of attack. a) The Caro-Kann: 1 e4 c6 2 d4 d5;
White's queen does prevent or discourage cer b) The Alekhine: 1 e4 lllf6;
tain uncommon deployments of Black's pieces c) The Petroff: 1 e4 e5 2 fllf3 lllf6;
and pawns, such as rash advances involving d) The French: 1 e4 e6 2 d4 d5 (and 3 lllc 3
.. .f6 or .. .f5. lllf6 or 3 llld 2 lllj6, among other examples);
e) The Scandinavian: 1 e4 d5; of 1 e4 as an 'attacking' opening, and of 1 d4 as

f) The Pirc: 1 e4 d6 2 d4 t:fJ/6. a 'positional' opening doesn't really follow.
After 1 e4 e5 much the same holds; for ex The openings arising from 1 e4 e5 are called the
ample, 2 f4 exf4 3 t:fJf3 d5 (or 3 ... t:fJ/6). Or, after 'Open Games' because pieces tend to come out
1 e4 e5 2 t:fJf3 t:fJc6, we have 3 ii.c4 l:iJf6, 3 ii.b5 rapidly and at least part of the pawn-centre
t:fJ/6 and many other Ruy Lopez positions with tends to evaporate. In particular, the association
...t:fJf6 at a very early stage. of 1 e4 with aggressive play stems in large part
An exception to all this is the Sicilian De from the tradition of tactically-based annihila
fence: after 1 e4 c5, Black's move ...d5 is nor tions that spring from 1 e4 e5 . Openings deriv
mally inferior and in the main lines he doesn't ing from 1 e4 e5 (D) are also combative from at
usually get to play ...t:fJf6 until the 4th or 5th least one perspective, namely that even within
move (e.g., 2 t:fJf3 d6 3 d4 cxd4 4 t:fJxd4 t:fJf6), or the first few moves the players so often make
perhaps later (after 2 t:fJf3 e6 3 d4 cxd4 4 t:fJxd4 threats to pawns, pieces, or even the king.
t:fJc6 5 t:fJc3 a6 6 ii.e2 'W/c7 7 0-0 t:fJf6, for exam
ple). Nevertheless, attack on White's e4-pawn
is a consistent theme of Black's strategy in
these Sicilian variations; he very often follows
...t:fJf6 with moves such as ... ii.b7, ...t:fJbd7-c5
and the like. Finally, one of White's principal
alternatives to 1 e4 c5 2 t:fJf3 is 1 e4 c5 2 c3, to
which Black normally replies 2 ...d5 or 2 ... t:fJf6,
both attacking e4.
This observation may seem trivial, but in
how many queen's pawn openings (i.e., those
stemming from 1 d4) does Black attack the d4-
pawn at all? Certainly not early on in openings
like the following:
a) The Queen's Gambit Declined: 1 d4 d5 2
c4 e6 and, for example, 3 t:fJc3 t:fJf6 4 il.g5 ii.e7 We can see this reputation illustrated by nu
5 e3 0-0, etc.; merous standard variations. The move 2 t:fJf3
b) The Nimzo-Indian: 1 d4 t:fJf6 2 c4 e6 3 immediately threatens the e-pawn, and after
t:fJc3 ii.b4; 2 ... t:fJc6 3 ii.c4 t:fJf6, 4 t:fJg5 already attacks the
c) The major Indian defences that begin with king! Such things don't happen after 1 d4 d5.
I ...t:fJf6 and 2 ... e6 or 2 ... g6, with the exception It might be useful to look at some of the more
of the Benoni ( 1 d4 t:fJf6 2 c4 c5). This is not traditional openings after 1 e4 e5. We have 2 f4
universally the case, but for the most part it (the King's Gambit, an opening ubiquitous in
holds true. master chess of the 1 9th century). After 2 ...exf4
After 1 t:fJf3, of course, White's knight is ex White tries to attack the king down the f-file in
empt from direct attack by Black; in practical conjunction with ii.c4, t:fJf3 and 0-0. 2 d4 is an
terms, so is White's pawn after 1 c4. We there other way to attack the pawn, and after 2 ...exd4
fore have a fundamental difference between 1 3 c3, White is already trying to blast open the
e4 and other first moves. centre with tempo (after 3 ...dxc3, 4 ii.c4 ! ? cxb2
5 il.xb2 is the Danish Gambit, and 4 t:fJxc3 is
The Open Games the Goring Gambit). The Vienna Game with 2
t:fJc3 t:fJf6 3 ii.c4 t:fJxe4 4 l!Vh5 has all the ele
It is interesting that 1 e4 is commonly thought ments of attacking chess. Even the generally
to be an 'attacking' move. To some extent that calm Giuoco Piano main line, 2 t:fJf3 t:fJc6 3
derives from the very exposure of the e4-pawn i.c4 ii.c5 4 c3 (don't forget 4 b4, the Evans
to attack, which can lead to early confrontation Gambit) 4 ...t:fJf6 5 d4 exd4 6 cxd4 ii.b4+, can
and the kind of dynamism often associated with result in violent play after 7 t:fJc3 ! ? t:fJxe4 8 0-0
king's pawn openings. But the characterization ii.xc3 9 d5 and similar variations. There are
INTRODUCTION TO I e4 AND THE OPEN GAMES 89
plenty of other examples such as the wild Max disappearing centres, to be sure - the centre is
Lange with 2 lllf3 lllc6 3 i.c4 lllf6 4 d4 exd4 5 remarkably stable in most Sicilian variations
0-0 and after 5 ... i.c5 6 e5 (a typical advance; when you consider what's going on around it
see below) 6...d5 7 exf6 dxc4 8 l:Ie l + i.e6 9 but in the exuberant activity of the pieces.
lbg5 �d5 1 0 lllc 3 or 5 ... lllxe4 6 .l:tel d5 7 White' s energetic knights on c3 and d4 are of
i.xd5 �xd5 8 lllc 3 there are things being at ten complemented by bishops on g5, e3, d3
tacked all over the place! and/or c4; his queen goes to d2, e2 or f3 ; his
In all of these examples the centre opens up rooks to central files, and his pawns rush for
quickly with short-term tactical consequences. ward to attack from squares such as f4, f5, g4,
So isn't it clear that the Open Games are domi g5, h4, h5, etc.
nated by attacking chess? There's something
missing from this argument; you could make it
1 e4 versus 1 d4
in the year 1900 but not today. In contemporary
chess, most of the above variations are rarely So which is objectively better, 1 e4 or 1 d4? The
seen (although they are instructive and worth short answer is that it depends upon the prefer
experimenting with), partly because the quality ences of the individual player. To go any fur
of dynamism can easily peter out when accom ther, we should address the state of theory.
panied by too many exchanges. In fact, all of Many of us will remember that for some time 1
them put together aren't played nearly as often d4 was Garry Kasparov's main opening move,
as the Ruy Lopez ( 1 e4 e5 2 lllf3 lllc6 3 i.b5). played in order to generate attacks. Indeed, a
That is significant because in the most impor- significant portion of his most brilliant and ag
tant variations of the Ruy Lopez it frequently gressive games begin with 1 d4. Attackers like
occurs that not a single pawn is exchanged until Shirov also used d-pawn openings, as did a
well into the middlegame, nor do the pieces get younger and more aggressive Kramnik. Korch
near each other if they can help it. Looked at noi rarely deviates from his adherence to 1 d4/1
from that perspective, the king of e-pawn open c4 openings and of course many other top-level
ings doesn't act like an Open Game at all ! To be grandmasters use 1 d4 almost exclusively. Nev
sure, the variations described in this manner are ertheless, at this moment we see a distinct pref
'Closed' Ruy Lopez systems and do not encom erence for 1 e4 among most of the world's
pass the entire opening. Nevertheless, in most strongest grandmasters. Is that because 1 d4
games with the Ruy Lopez the dynamic action isn't an exciting move? Would you say that the
is delayed until after some serious manoeuv Exchange Variation of the Griinfeld, the Bot
ring has occurred, a type of play that becomes vinnik Variation of the Semi-Slav, the Exchange
increasingly fascinating as you become a better Variation of the Queen's Gambit Declined, the
player. A similar statement can be made about Taimanov Variation of the Benoni, and any
the Petroff Defence ( 1 e4 e5 2 lllf3 lllf6), the number of King's Indian lines, are not aggres
next most popular 1 e4 e5 opening at the inter sive attacking systems? In reality, what happens
national level. The Petroff shouldn't be de is that in different eras, individual defences
scribed as non-confrontational, but it tends to prove to be temporary barriers to the general
lead to fairly stable half-open structures in which use of 1 e4 or 1 d4 at the very highest levels. At
tactics play a lesser role. The Giuoco Piano ( 1 this moment in time I would say that the
e4 e5 2 lllf3 lllc6 3 i.c4 i.c5) and the Scotch Nimzo-Indian ( 1 d4 lllf6 2 c4 e6 3 lllc 3 i.b4) is
Game are examples of double e-pawn openings such a defence, with Black complementing its
that can produce either tactical or positional use with the Queen's Indian Defence or Queen's
struggles. I think that it's fair to characterize 1 Gambit Declined when confronted by 3 lllf3 .
e4 e5 as neither exceptionally dynamic nor se Recently, however, White has done reasonably
date. well against the Queen's Indian and it has tradi
It might be argued, in fact, that the Open Si tionally been possible to create chances against
cilian ( 1 e4 c5 2 lllf3 with 3 d4) has inherited the Queen's Gambit. Furthermore, White's
the mantle from double e-pawn openings in score against the Nimzo-Indian is somewhat
producing romantic attacking chess. Not with better than his score against other openings,
including those beginning with 1 e4. In the Ruy and 4 ii.c4. Just a bit further on in French De
Lopez, on the other hand, we currently see fence games, we have 1 e4 e6 2 d4 d5 3 lllc 3
White avoiding the Marshall Attack with, for ii.b4 4 e5, 1 e4 e6 2 d4 d5 3 lllc 3 lt:Jf6 4 ii.g5
example, an early h3 followed by moves such ii.e7 5 e5, and so forth. After 1 e4 e5, there are
as d3, a3 , lllc 3 and ii.a2 (see Chapter 8). Given moves such as 2 lllf3 lllc6 3 ii.c4 lllf6 4 d4
the unambitious appearance of this method of exd4 5 e5, or the more complex 2 lllf3 lllc6 3 d4
play (although it' s faring tolerably well so far), exd4 4 lllxd4 lZ:lf6 5 lllxc6 bxc6 6 e5, the latter
one wonders if the pendulum might swing back revived and brought into prominence by World
to 1 d4. Or perhaps players will amend their Champion Kasparov.
tastes some years hence, for unrelated reasons. Where do you find similar advances in the
That is part of the fun of following opening the practice of 1 d4 d5, or in any line beginning
ory. At any rate, the average player (and even with 1 d4? In a d-pawn opening, White seldom
'ordinary' master) need not worry about such plays d5 with a threat within the first six moves.
matters; either first move will produce games In fact, only in a few openings (such as 1 d4
with plenty of opportunities for victory. lllf6 2 c4 c5 3 d5) does the d-pawn even reach
Don't worry if the recital of names in the last the fifth rank, whether there is a threat or not.
few paragraphs befuddles you. My point is to It's true that d5 will fairly often occur in the
present 1 e4 from a broad perspective. It can be King's Indian Defence (e.g., 1 d4 lllf6 2 c4 g6 3
as much an option for positional players as for lllc 3 ii.g7 4 e4 d6 5 lllf3 0-0 6 ii.el e5 and now
attacking players. There are ways to fight for 7 d5 or 7 0-0 lllc6 8 d5); and similarly in a few
very small and lasting advantages against nearly lines of the Griinfeld. However, such d5 ad
every defence to 1 e4, and there are ways to try vances don't occur often after 1 d4 d5 and will
to decimate the opponent with slash-and-bum usually happen well past the first several moves
tactics. Most of the latter methods come up of the opening. In e-pawn openings, an analo
short of their goal against proper defence, or in gous situation would be the advance d5 in the
the face of counterattack by Black. Still, once Ruy Lopez, normally played after the 10th
the smoke has cleared, a bold attack may be just move.
as effective as any other approach at producing What does that mean? That by using 1 e4, at
a small but durable advantage. least in some openings, White has the option of
Rather than measuring degrees of aggres staking out a significant space advantage early
sion, a dispassionate investigation of e-pawn on. This is indeed an aggressive stance, but not
openings turns up a more interesting distinction one that involves open centres and multiple ex
between 1 e4 and 1 d4. This has to do with the changes - quite the contrary. And keep in mind
acquisition of space by the pawn advance e5, that when pawns are advanced they can become
which is prominent in the Semi-Open Garnes vulnerable; again we hark back to Breyer's 'last
(defences other than 1 .. .e5), but can also occur throes' . If you are an e-pawn player, you have
in double e-pawn openings. Consider that White to take that possibility into account when you
can play e5 on the third move of both the Caro advance your pawns. Failure to tie your oppo
Kann Defence ( 1 e4 c6 2 d4 d5 3 e5) and French nent down or make other difficulties for him
Defence ( l e4 e6 2 d4 d5 3 e5), and on the sec can sometimes leave you on the defensive. On
ond move of the Alekhine Defence ( 1 e4 lllf6 2 the other hand, an aggressive pawn presence in
e5). Against the Pirc Defence, e5 is a common the enemy camp can reward you with a winning
move in the variation 1 e4 d6 2 d4 lllf6 3 lllc 3 advantage. You will see examples of both of
g6 4 f4, and played in several lines with 4 lllf3 these results throughout the book.
5 G iuoco Piano
1 e4 e5 2 tbf3 tbc6 3 J.c4 (D) will sometimes transpose to 3 J.c4, for in
stance after 2 ... J.c5 3 tbf3 tbc6. The inde
pendent line 2 J.c4 tbf6 3 d3 c6 4 tbf3 d5 5
J.b3 J.d6 can lead to complex play, and of
course Black can play ... tbc6 on one of the
first few moves. One of the problems with 2
J.c4 is that Black has various ways to control
the direction of play. That interferes with
some players' desire to be in command as
White, particularly when facing a symmetri
cal variation such as l . ..e5 .
3...J.c5 (D)
The move 3 J.c4 has been used consistently

since the early days of chess in its modem form.
Although far behind the Ruy Lopez (3 J.b5) in
popularity, 3 J.c4 is White's second favourite
continuation. Placing the bishop on c4 agrees
with the principles of development and central
ization, and prepares to castle quickly. It is also
the move that most directly attacks Black's po
sition, in particular the sensitive f7-square. In
addition, White wants to control the central
d5-square and thus prevent Black's freeing move
. . . d5. In this respect 3 .ltc4 fulfils a positive po This development of Black's bishop is the
sitional role that, for instance, 3 J.e2 doesn't. oldest well-analysed response to 3 .ltc4. I'll use
As always, there are drawbacks, not obvious the generally accepted name 'Giuoco Piano' for
at first. Because the bishop on c4 makes no 3 . . .J.c5 ; it is also called 'the Italian Game' in
threat, Black himself is able to develop freely. recognition of the Italian players who published
That would also seem to be true of 3 .ltb5, analysis of the move in the late 1 6th and early
which also has no direct threat; but the latter 1 7th centuries.
move discourages a number of black set-ups With 3 ... J.c5, Black attends to White's move
that .ltc4 doesn't, by virtue of the potential d4, the idea of which is to form an ideal centre.
threat of .ltxc6 and tbxe5 . In the Giuoco Piano, Moreover, the move tbg5 is lurking in the back
moreover, we shall see that if Black does ground; since that would attack the f7-square
achieve the move ... d5, White may lose a tempo twice, Black wants to be ready to defend against
or suffer some positional disadvantage. These the threat by castling. The straightforward posi
considerations are rather abstract, and can only tion after 3 ...J.c5 contains a majority of the ba
be shown by example. sic classical ideas about development, centre
I should mention that the Bishop's Open and attack. That should motivate us to examine
ing, l e4 e5 2 J.c4, is a respectable choice that it in some depth.
4 c3 c) 7 .lid2 .ltxd2+ 8 �xd2 gives White supe

I shall concentrate upon this continuation as rior development in terms of quantity and qual
representing the purest intent of the opening: ity.
to establish an ideal centre and drive Black's We now return to 4 ... tllf6 (D):
pieces away with tempo. 4 c3 leads to play that
resembles other openings and is therefore of
general value. For organizational reasons, the
line 4 d3 tllf6 is discussed in Chapter 6 about
the Two Knights Defence. It will arise via the
move-order 1 e4 eS 2 tllf3 tllc6 3 .ltc4 tllf6 4 d3
.lies. The similar 4 c3 tll f6 S d3 is placed at the
end of this chapter.
4.. tllf6
.
With this move Black develops a piece and

counterattacks. Other moves allow White to ex
ecute his plan; for example, the line 4 ... d6? ! S
d4 exd4 6 cxd4 .ltb4+ (D) illustrates White's
central superiority.
5 d4
Certainly the most challenging continuation.
A less aggressive but also interesting alterna
tive is S d3. I'll discuss that more technical
move at the end of the chapter.
White has the instructive option of playing
S 0-0, when Black does best to capture by
S . . .tt:lxe4 and meet 6 d4 with 6 ... dS ! (file this
move away in your memory ! Black should al
most always play . . . dS when allowed to do so,
that is, if it's tactically sound) 7 dxcS dxc4 8
�xd8+ @xd8. From White's point of view,
this endgame is at best equal, and more likely
he will end up with a somewhat inferior posi
Then White has several good continuations: tion.
a) 7 @fl ! ? (this is the fancy way to get out 5...exd4 6 cxd4
of check; White threatens 8 dS, and when the The seemingly assertive 6 eS can again be
knight moves, 9 �a4+ picks up the bishop) answered by 6 ... dS ! (6 ...tt:le4 ? ! , with 7 �e2 dS
7 . . . .ltaS 8 dS tt:lce7 9 b4 ! (9 �a4+ c6 protects 8 exd6 0-0 in mind, is strongly answered by 7
the aS-bishop) 9 ... .ltb6 10 .ltb2 and the bishops .lidS) 7 .libs (7 exf6? dxc4 8 fxg7 :Ig8 leaves
are dominating the board. One can compare the all of Black's pieces active and ready to spring
Evans Gambit ( 1 e4 eS 2 tllf 3 tt:lc6 3 .Jlc4 .lies into action, whereas White is underdeveloped
4 b4 .ltxb4 S c3), in which something like this and losing badly in the centre; Black will cas
can arise but with Black having an extra pawn tle queenside in order to safeguard his king)
by way of compensation. 7 ...tt:le4 8 cxd4. Now Black can play either
b) Naturally 7 tt:lc3 , developing a piece, 8 . . . .ltb4+ or, more commonly, 8 ... .ltb6. In the
can't be bad: 7 . . . tllf6 8 dS .ltxc3+ (again, watch latter case play might go 9 0-0 0-0 1 0 .ltxc6?!
out for 8 ...tt:le7?? 9 �a4+, winning a piece; this (this slightly dubious capture is given in the
is a common trick in many openings, including books; the rationale is that Black was planning
those stemming from 1 d4) 9 bxc3. The result ... tt:le7) 1 0 ...bxc6 (D).
ing position favours White because of his dom At first it may look like the bishop is badly
inating centre. placed on b6 and Black suffers from weak
GIUOCO PIANO 93
pawns on the open c-file. However, he has the 7 ii.xd2+

...
bishop-pair and is ready to assault White's Recently the older 7 . . .lt'ixe4 8 i..xb4 lt'ixb4 9
centre by ...c5. There' s little White can do about i..xf7 + @xf7 10 'ilib3+ has again been tried for
that; for example, 1 1 b4 ! ? (to stop . . . c5; some Black, often leading to 10 ... d5 1 1 lt'ie5+, when
other moves are 1 1 'ilic2 i.. g4 ! and 1 1 h3 c5 ! 1 1 ...@e6 ! ? 1 2 'ilixb4 c5 results in complex play.
12 i..e 3 cxd4 1 3 lt'ixd4 'ilie8 ! , having in mind This line is unresolved; many players will not
. . . 'ilixe5 , or . . . c5 and . . . f6; lastly, 1 1 i..e 3 i.. a6 trust it because Black's king comes to the centre,
12 l:Ie l c5 1 3 dxc5 lt'ixc5 is good for Black) and others will embrace its adventurous char
1 l . ..a5 1 2 i..a3 axb4 1 3 i.. xb4 c5 14 dxc5 acter. In any event, it's refreshing that long
.ltxc5 15 i..xc5 lt'ixc5 1 6 'iiic2 lt'ie6 1 7 l::td 1 c5; discarded variations can spring to life again.
then Black has two passed pawns and a nicely 8 lt'ibxd2 d5 9 exd5 lt'ixd5 (D)
centralized position.
Notice the combination of 1 3 ... c5 and 1 7 ... c5.
This double-hammer with the c-pawns with the
intent to destroy White's centre is a common
theme. White should take that possibility into
account when playing .ltxc6. This type of posi
tion will frequently arise in other opening vari
ations.
6 .ltb4+ (D)
•••
It's worth a look to see how powerful the

possession of an unopposed ideal centre can be:
6 ... .ltb6? 7 d5 ! lt'ie7 8 e5 lt'ie4 9 0-0 0-0 10 'ilie2
lt'ic5 1 1 b4 lt'ia6 1 2 d6 cxd6 1 3 exd6 lt'ig6 14
.ltg5 'ilie8 1 5 .lte7 @h8 1 6 lt'ic3 lt'ixb4 1 7 l::ta el
lt'ic6 18 .ltxf8 'ilixe2 19 l::txe2 lt'ixf8 20 l::te8
@g8 21 lt'id5 g6 22 lt'ie7+ 1 -0 Euwe-Jutte, Am 10 'ilib3
sterdam 1 927. See also the sample game in White usually plays this immediately, in or
Chapter 1 . der to attack Black's blockading knight on d5
7 .ltd2 before it is fully secured and before Black's
The tactics that follow 7 lt'ic3 lt'ixe4 8 0-0 king reaches safety. There are two instructive
.ltxc3 9 d5 lead to some 20 moves of theory and alternatives, the second of which keeps the
are not dealt with here. Various books will sup game interesting for both players:
ply the details. 7 @fl , the 'Krakow Variation' , a) 10 0-0 is playable but allows Black more
should be met by the standard counterthrust options after 10 . . . 0-0, when 1 1 'ilib3 ? ! lt'ia5 !
7 . . . d5 ! 8 exd5 lt'ixd5 and it's not clear what the eliminates White's c4-bishop without compro
king is doing on f l . mise. A demonstration of how White can pit his
activity against Black's static advantages went

1 1 tt:le5 ! ? (objectively, the move 1 1 �c2 ! is
doubtless better; compare 10 �c2 in variation
'b') l 1 . . .tllxd4 ! ? ( l 1 . . .tllxe5 1 2 dxe5 .lte6 1 3
�3 .l:!.b8 is equal, but White's pawns are recon
nected in that case) 1 2 tt:lb3 ! tt:lxb3 1 3 .ltxd5 !
tt:lxa l ? ( 1 3 ...�f6 is best) 14 .ltxf7+ @h8 1 5
�h5 ! (D) with a terrific attack, Kluxen-Capa
blanca, Hamburg simul 1 9 1 1 .
Black' s blockade of d5 is a key element in his

attempt to keep the position under control. In
the position before us, White can't break the
blockade but can 'play around' that knight to
create threats. The exchange of dark-squared
bishops should favour Black, because simplifi
cation makes it harder for White to muster
forces for an attack. This raises the interesting
question of what degree of simplification tends
to negate the more active party' s compensation
The game continued 1 5 . . ..ltf5 ( 1 5 . . . h6 1 6 for his isolated pawn weakness. In this particular
.l:i.d l ) 1 6 �xf5 �f6 1 7 tt:lg6+ ! �xg6 ( 1 7 . . .hxg6 situation White still has significant resources, as
1 8 �h3+ is a trick worth knowing) 1 8 .ltxg6 we shall see. Although some further exchanges
.l:!.xf5 19 .ltxf5 g6 20 .lte4 1 -0, since the al - will seriously cut into his chances, others can in
knight will fall. If this hadn't been a simulta crease his pressure! It all depends upon piece
neous exhibition (as opposed to a serious tour play. For instance, isolated queen' s pawn posi
nament game), Kluxen's name would have gone tions from other openings like the Queen' s
down in history for beating the mighty Capa Gambit Accepted, Nirnzo-Indian, Caro-Kann,
blanca! As it stands, the game shows the appeal etc. , allow Black options of expanding on the
of the Open Games. queenside by ... a6 and ...b5 or fianchettoing with
b) The other alternative with a durable char . .. b6 and ... .ltb7. That sort of thing doesn't apply
acter is 1 0 �c2; for example, 10 ... tt:lce7 (if to our current variation, nor does Black appear
1 0... �e7+, then 1 1 @fl is good for White; like to have a way to disturb the equilibrium. If that's
wise 10 ... .lte6 1 1 0-0 0-0 1 2 .l:!.fel tt:ldb4 1 3 true, Black may have to leave his opponent
�b3 .ltxc4 1 4 tt:lxc4) 1 1 0-0 0-0 1 2 tt:le4 .ltg4 alone for a while, giving White crucial time to
1 3 .l:!.ac 1 .ltxf3 ! ? 14 tt:lg5 g6 15 tllx f3. So White try to improve his position. On the other hand,
has more than one way to maintain active play Black has no weaknesses to attack and will only
on the board. permit a weakness to be created if he can gain
Now let's return to 1 0 �b3 (D): something in return.
This venerable position is characterized by a After 10 �b3, Black has two basic plans: re
balance of classic positional factors: White's inforcing his blockade by 1 0... tt:lce7, or trying
greater activity and space, including pressure to force events by 1 0 ... tt:la5. We'll examine
down the c- and e-files, versus Black's firm both.
blockade of a potentially weak isolated pawn
on d4. In Chapter 3 , I discuss numerous other Blockading the Pawn
isolated queen's pawns in chess openings. How
does this IQP compare with those? As always, 10 tt:lce7 (D)
.•.
GIUOCO PIANO 95
The irritating effect of the pawn-push a5 shows

up in the beautiful game Rossolimo-Reissman,
Puerto Rico 1 967: 1 3 . . . b6 14 ctJe5 .ltb7 15 a5 !
l::tc 8?! ( 1 5 .. .f6 ! 16 ctJd3 @h8 is the consistent
strategy, guarding d5; Black should use his
own strengths) 16 ctJe4 iVc7 1 7 a6 .lta8 1 8 iVh3
ctJf4 19 iVg4 ctJed5 20 l::ta3 ! . Now Black's king
is under serious attack before he has the chance
to play ... c5 and free his a8-bishop. The game
continued 20 . . . ctJe6? (a poor move, but 20 . . . c5
2 1 g3 ctJg6 22 lLixg6 hxg6 23 iVh4 is also good
for White, with ideas including 24 ctJg5 ctJf6 25
l::tae3 !) 2 1 .ltxd5 cxd5 22 ctJf6+ @h8 23 iVg6 ! !
(D).
In this situation, the battle between piece
activity and positional factors revolves specif
ically around the isolated pawn and its block
ader. Some samples of the play follow.
O'Kelly - Euwe
Amsterdam 1950
1 1 0-0 c6 12 l::tfe l 0-0

Black has shored up d5 with no obvious dif
ficulties. However, White has energetic minor
pieces and can create significant problems.
First, he stakes out some territory.
13 a4
Gaining space is often the best policy when
there are no direct targets. White operates Naturally the brilliance of this move strikes
against ... b5, but also plans a5, serving the dou one first (reminiscent of the famous Levitsky
ble function of preparing an attack on b7 and Marshall ...iVg3 ! ! game) but an eye for detail
keeping a knight from b6. The other strategy is will also pick up Black's bishop stuck behind the
to emphasize piece-play, for instance by 1 3 d5-pawn. That was of course the blockading
lLie4. White can also develop immediately by square which was the pride and joy of Black's
1 3 l::tac 1 , as he has done in a few games; for exposition. The game continued 23 ...iVc2 (mate
ample, 13 . . . a5 ! ? (or 1 3 ...iVb6 14 iVa3 .lte6 1 5 on h7 is threatened, but accepting the queen
ctJe4, with nagging pressure involving moves sacrifice loses instantly: 23 .. .fxg6 24 lLixg6+
like ctJd6, lLic5 and ctJfg5, Rossolimo-O' Kelly, hxg6 25 l::th3#; 23 . . . hxg6 24 l::th3#; or 23 ... gxf6
Amsterdam 1 950; Black would most likely do 24 iVxf6+ lLig7 25 l::tg 3 l::t g8 26 ctJxf7+) 24
better to play 13 ...ctJb6 14 .ltd3 .ltf5 15 lLie4 l::th 3 ! 1 -0. Either the rook or queen threatens to
iVc7 with some kind of dynamic equality) 14 capture on h7 with checkmate on the next move,
lLie4 a4 15 iVa3 ctJf5, Renner-Gabriel, 2nd Bun but 24 ... iVxg6 25 ctJxg6+ fxg6 26 l::txh7# is
desliga 2000/1 ; at this point White had the op checkmate.
portunity for the transformation 16 .ltxd5 ! ? Instead of 1 3 . . . iVb6 or 13 . . .b6, a defence
iVxd5 1 7 l::tc5 iVd8 1 8 iVb4 when Black is tied which avoids weakening the queenside and
down; e.g., 1 8 ... l::te8 19 l::te5 ! . maintains the bishop on the h3-c8 diagonal is
13...iVb6! 1 3 . . . l::tb 8; for example, 14 a5 f6 ! ? (D). Any
Euwe's continuation is probably the most move of Black's f-pawn is double-edged: why
logical response. Simplification should help the would he allow a hole on e6? The answer is
defender and Black avoids weaknesses as well. that, by protecting e5, Black prepares ... .ltg4
without fear of tt:le5. He can also play ... @h8 tt:lg5 .ltf5) 1 6 . . . .l:!.ad8 1 7 tt:lc5 .ltc8 !. Here White
without worrying about a knight attack on f7. is running out of ideas whereas the d5-square is
Black thinks that he can afford the weakness the axis of the game. This line serves as a good
on e6 for the sake of quick development. An model for Black's play.
other, apparently safer, move is 14 ... h6, prepar
ing . . . .lte6 or ... .ltf5, but then 1 5 a6! b5 1 6
.ltxd5 tt:lxd5 1 7 tlle4 .l:!.b6 1 8 tllc5 tt:lc7 1 9 �c3 !
tt:lxa6 20 b4 establishes a huge clamp that is
worth more than a pawn.
14 ...\\Wxb3 15 tt:lxb3 .ltf5

Although developing the bishop is probably
satisfactory, it is not as clear as 1 5 ... !Id8, when
if White gets too ambitious we can see all of
Black's pieces coordinate to his benefit; e.g., 1 6
Garcia Femandez-Komeev, Madrid 2002 tt:lc5 ! ? :bs ! 1 7 tt:le5 (White can try 1 7 .:tac 1 b6
continued 15 tt:le4 ( 1 5 a6 b5 16 tt:le4 .ltg4 !) 18 tll d3 ! , but Black will equalize after 18 ... .ltd7
15 ...@h8 1 6 �a3 .ltg4 1 7 h3 .lth5 1 8 tt:lc5 :es 19 axb6 axb6; for example, 20 .llxd5 ! tt:lxd5 2 1
19 tt:le6 (a huge square but there are no targets !) tllfe5 tlle7 2 2 tt:lb4 f6 ! 2 3 tt:lxd7 �xd7 24 tt:lxc6
19 . . . �d7 20 tll c5 �c8 2 1 .ltxd5 tt:lxd5 22 tt:lxc6 25 �xc6 :xd4 26 f3 l:tb7 with an equal
.l:!.xe8+ �xe8 23 .l:!.el �d8. Black has achieved position) 1 7 ...'it>f8 ( 1 7 ... f6? 1 8 tt:led7 !) Rosso
the desired simplification and preserved the limo-Unzicker, Heidelberg 1949; now if White
precious blockade on d5. Now the re-entry of plays 1 8 !Ia3? ( 1 8 tt:lf3 is equal; that's the best
the knight by 24 tt:le6 is met by 24... \\Wg8 threat that White can do) 1 8 ... b6 1 9 axb6 axb6 20 tlle4,
ening ... l:te8, ... .ltf7, etc. It takes courage to Black repulses White by 20 .. .f6 2 1 tt:lf3 .lid7 !.
play this way, however. 16 tt:le5
14 a5!? (D) With the idea 17 a6.
Rossolimo presses on with a remarkable 16 tt:lb4!?
•.•
idea. It's amazing that White can permit Black 1 6 ... !Ife8 gives a more solid impression.
to exchange queens, which in theory should be Black is probably close to equality hereabouts
all that Black needs to consolidate his d5 out but it's hard to counter White's queenside pres
post and attack the d4-pawn. Here we have a sure. 16 ... a6? would create a strongpoint on c5
lesson about isolated queen's pawns: although which White could immediately occupy to good
it's not the rule, a great deal of simplification effect.
can be suffered by their owner if his pieces get 17 l:tacl
to favourable squares. Instead, 14 \\Wa3 is the After 17 tllxf7, Black's trick was l 7 ...tt:led5 !
matic, yet the black pawn-structure remains un with the threats of ... !Ixf7 and ...tllc 2.
challenged following 14 ... .lte6 ( 1 4 ... tllf5 has 17...tt:ledS 18 a6! (D)
been played but 1 5 .ltxd5 ! cxd5 16 tt:lb3 should White destroys the foundation of Black's
give White a small advantage due to his good light-square bulwark.
knights and Black's bad bishop) 15 a5? ! ( 1 5 18...bS 19 .ltxd5 cxd5 20 tt:lc6 tt:lxc6 21
tt:le4 i s double-edged) 1 5 ... �c7 1 6 tt:le4 ( 1 6 !Ixc6 !Ife8 112-112
GIUOCO PIANO 97
l::tc 5 and b4. Then 1 3 ... �5 ! is forced, but

White dominates after 14 �a3 ! (threatening
l::tc5 and Clie4) 14 ... b6 15 Clie4 Clib7 1 6 Clie5
.ltd7 1 7 Clic4 iVd5 1 8 0-0 0-0-0 1 9 �xa7 !
�xe4 20 d5 ! with utter destruction of Black' s
position to follow.
After the text-move, . . . Clib6 is threatened,
to rid White of his best attacking piece, the
c4-bishop. There follows an illustration of the
play.
Kupreichik - Aleksandrov
Bad Worishofen 2001
Pachman analysed this position and showed 12 0-0!? (D)

that White stands much better, confirming the Castling is the most interesting continuation.
general idea that White's knight is superior to Quite a few games have continued 1 2 �b3 ctJa5
Black's bishop: 22 l:Ixe8+ l:Ixe8 23 f3 (23 h3 is 1 3 �a4+ Clic6 14 �b3 Clia5 with a draw, an out
also good) 23 . . . l:Iel + (23 ...l:Ic8 24 l:Id6 i.e6 25 come that must have been satisfactory to both
Clic5) 24 @f2 l::tb l 25 Clic5 l:Ixb2+ 26 @g3 g5 players, probably even before the game started.
27 l::tc7 @g7 28 llxa7 lla2 29 l':.b7 i.c8 30 This indicates that in order for Black to try for a
l::txb5 i.xa6 3 1 Clixa6 l:Ixa6 32 l::txd5 with a win, he should play 1 2 ... Clice7 (see the previous
technically winning game. section on 10 ... Clice7). However, White doesn't
All this material is terribly instructive for the have to take this draw and can keep the play
developing player, and even masters might find alive by 1 2 0-0. He also had the earlier option of
the ideas intriguing. 10 0-0 0-0 1 1 �c2, mentioned above in the note
to White's tenth move.
Chasing the Pieces White's other options are 12 .ltb5, which is
playable but not so convincing after 12 ... .ltd7;
10 CliaS ll �a4+ (D)
..• and 12 Clie5 ? ! , which runs into surprising prob
lems after 1 2 ... 0-0! 1 3 Clixc6 �e8+ ! 14 @fl
ctJb6! 1 5 �b3 ( 1 5 �b5 bxc6 1 6 �c5 Clixc4)
15 ...�xc6, Karkocha-Swerin, corr. 1985. Black
B has a big advantage in view of White's terrible
king position.
1 1 Clic6
..•
This has long been thought to provide Black

with a drawing option, or at least an extremely
drawish one. Instead, 1 1 .. .c6? defends d5 but
fails for concrete reasons, because after 1 2
.ltxd5 ! �xd5 1 3 l::tc 1 White threatens both 12 0-0
...
The problem with 1 2 0-0 is supposed to be lack of weaknesses protect him from immediate
that 1 2 . . . tt:lb6 forks queen and bishop, but then attack.
there can follow 1 3 .l:i.fe l + .lte6 14 �a3 ! , after
which White prevents Black from castling and The older lines of the Giuoco Piano can still
14 . . . tll xc4 (what else?) 1 5 tt:lxc4 gives White challenge the chess understanding of both play
extremely well-placed pieces and Black still ers. No other opening serves better as a model
can't bring his king to safety. A simple plan is for classical double e-pawn chess. Those of lit
:Lad 1 followed by d5. If Black captures the d tle or moderate playing experience will find
pawn by 1 5 ... tt:lxd4, he is subject to a typical careful study and practice of this opening par
open-position attack after 16 .l:!.ad l tll xf3+ 1 7 ticularly valuable, and even experienced play
�xf3 . After l 7 ...�c8?! 1 8 �a3 White has pre ers could do worse than to investigate its unique
vented castling and threatens f4-f5 . Black properties.
should play the active 1 7 . . . �h4, though 1 8
�xb7 keeps an edge. A Technica l Approach : 5 d3
13 .ltxd5!?
Slightly passive. In the spirit of avoiding What if White doesn ' t want to engage in the
simplification, White should try 1 3 �c2! .lte6 kind of open struggle just described? Let's take
( 1 3 ...tllb6 14 .lid3 h6 1 5 �c5 ! ?) 14 .l:i.fel tlldb4 a look at what happens if he doesn't go in for
15 �c3, and the struggle between White's space the relatively forced moves that follow 5 d4 and
and Black's pressure on the IQP continues, a plays 5 d3 instead:
sample line being l 5 ... .ltxc4 16 tt:lxc4 tt:ld5 17 1 e4 e5 2 tllf3 tllc6 3 .ltc4 .ltc5 4 c3 tllf6 5 d3
�b3 tllb6 1 8 l:tac l ! ? tt:lxd4 19 tllxd4 �xd4 20 (D)
tlla5 with an edge for White. This kind of play
resembles our 1 0 �b3 tt:lce7 main line above.
13 \\WxdS 14 .l:!.acl �d8
•••
Also 14 ... .lte6 is sensible, to blockade on d5 ;

e.g., 15 .l:!.c5 \\Wd8 ! ( 1 5 ... �xa2 16 �xa2 .ltxa2
17 b3) 16 .l:!.el .lid5.
15 tt:le4 (D)
1 5 tllb 3 intending .l:!.fe 1 and tt:lc5 is a more
complex route. Then play might go 1 5 ...tt:le7 1 6
.l:i.fel tlld5 1 7 tllc5 c 6 with unclear prospects.
This is the kind of slow move that White typ

ically makes in order to play it safe and engage
Black in a battle of positional skills. His ideas
by setting up this structure are:
a) to protect his e4-pawn;
b) to cover the d4-square against intrusions
by Black's pieces (in particular ... tt:ld4); and
c) to hold off on the move d4 until his pieces
are more developed, thereby avoiding the forc
ing variation that we saw in the main variation
15 ...tt:le7! 16 !Ifel tt:ld5 after 5 d4 exd4 6 cxd4 .ltb4+, which was soon
The opening is over and the chances appear followed by the centre-clearing move ... d5.
about equal. In spite of White's significant lead On the other hand, Black now has much
in development, Black's grip on d5 and absolute more freedom to develop his pieces. Without
GIUOCO PIANO 99
fear of d4 he can do so actively and should se advantage of the bishop-pair with no conces
cure equality. But don't expect the play to be sion on his part. Note that 7 ... tba5 right away
easy for either side. would have allowed 8 .ltb5 ! a6 9 i.a4 b5 10
Now I'm going to show one game out of the .ltc2, which saves White's bishop from ex
many that have been played, with the goal of in change and threatens b4. So we can see another
cluding some general ideas that will be applica advantage of White' s move c3. After the text
ble to similar positions. move, the positional threat of 8 ...tba5 is real, so
play can continue as follows:
Karpov - Korchnoi 8 .ltb3 .lta7 9 h3!? (D)
Merana Wch (8) 1981
5 d6
...
Black secures the e5-pawn against threats

such as b4-b5. Here is a general warning for
Black: you shouldn't be in too great a rush to
play the tempting ... d5, because your centre can
become too vulnerable; for instance, 5 ... d5? ! 6
exd5 tbxd5 7 b4 (7 �3 is also dangerous)
7 ... i.b6 8 b5 tba5 9 tbxe5. Notice that the pin
on the knight by 9 ...'ilie7? ! means nothing after
10 0-0 ! , because 10 ...'ilixe5?? loses to 1 1 .l:.el .
Also weak would be 5 ...0-0 6 0-0 d5? ! 7 exd5
tbxd5 8 b4! followed by 9 b5 . These lines show
one of the benefits that White gets by playing
c3. Karpov's move. It prevents ... .ltg4 in some
6 0-0 0-0 situations, but mainly it prepares l:.el without
6... .ltg4 is also possible, with more or less having to worry about ... tbg4. White is now
the same kind of position that we shall discuss ready to reorganize by moving his knight to c4
in Chapter 6 when we look at 4 d3 in the Two and e3, or to play l:.el followed by tbfl , in tum
Knights Defence. followed by tbg3 or tbe3. Perhaps you're fa
7 tiJbd2 (D) miliar with this sequence of moves, but if not, it
must look rather odd. In fact, the knight ma
noeuvre tbbd2-fl -g3/e3 is standard practice. I
won't go into detail at this point, but the princi
ple here is that if the centre is stable, players
may be able to embark upon long trips with
their pieces without being punished. From g3,
White's knight lusts after the wonderful square
f5 and protects e4; and after tbe3, the knight
sets its eyes upon both d5 and f5 (at the cost of
blocking off his queen's bishop). We shall see a
lot of this manoeuvre tbbd2-fl -g3/e3 in the
Ruy Lopez chapter, and it's good to be intro
duced to it now.
9 .lte6
.•.
9 . . . h6 is a good option. You will see a lot of

7 ... a6!? these 'little moves' in variations with d3 and
Advancing the a-pawn so modestly has two ... d6. The idea is to prevent .ltg5 after White's
ideas: to put the bishop on the safe square a7 knight moves. After 10 l:.el , there can follow
and to be able to play ...tba5 and capture the 10 ... tbh5 ! . Compare the game and comments
c4-bishop. That exchange would gain Black the below.
9... d5 !? is also playable at this point, al This isn't the only move by any means, but
though 10 .l:!.e l is curiously solid for White and it follows a 'mini-rule' that can apply to any
asks Black what he's going to do next. opening in which there are pawns on e4 and e5:
10 .ltc2 d5!? 1 1 .l:!.el dxe4 if White plays h3, the move ...tt:lh5 should be
Ripperger gives the fascinating line 11.. .d4 strongly considered. The reasoning is that after
12 tt:lc4 dxc3 13 bxc3 .ltxc4 14 dxc4, when ...tt:lf4, the knight can't be kicked out by g3
White's pawn-structure is thoroughly damaged, since ...tt:lxh3 will follow. But if White's bishop
but he has the bishop-pair and play down the b captures that knight (i.xf4), he will have ceded
and d-files. the bishop-pair; that is hardly disastrous but
12 dxe4 tllh5! (D) usually not a good thing for White (remember
how important it is to possess the two bishops).
Notice that in the note above about 9... h6, the
d-file was closed. This time we're about to get
an exchange of queens.
Of course this sort of technical guidance
only fits in certain situations, but it can also ap
ply to the Ruy Lopez and Philidor Defences,
and the same idea quite frequently occurs in the
King's Indian Defence, a very different open
ing indeed!
13 tllfl \\Wxdl 14 !Ixdl l:tad8 15 i.e3 f6 16
.ltxa7 tt:lxa7 17 tlle3 tllf4 18 h4 .ltf7 19 tllel
At this point Polugaevsky suggests 19...tt:le6
20 .ltb3 tt:lc5, which looks equal.
6 Two Knights Defence
1 e4 e5 2 lLif3 lbc6 3 .Jlc4 lLif6 (D) with 2 lbc3 lbf6 3 .Jlc4 lbc6 4 lbf3 (to avoid 4
d3 .Jlb4!?):
This is the Two Knights Defence. Its main

lines are definitely more ambitious and tactical Black can opt for 4....Jlc5, of course, proba
than those after 3 ....Jlc5. I shall focus on the two bly heading back to lines below; but he can also
main continuations, 4 lbg5 and 4 d4, and we'll play more decisively by 4...lbxe4!? 5 lbxe4 (5
also examine 4 d3 at some length due to its pop 0-0 is the Boden-Kiezeritsky Gambit again; as
ularity and its instructive nature. usual, Black's centre is more important than his
But first, let's briefly glance at a few rare king position after 5 .Jlxf7+? 'it>xf7 6 lbxe4 d5!
continuations: 7 lbeg5+ 'it>g8 with ...h6 coming next) 5...d5,
a) 4 c3?, as in the Giuoco Piano, is mistimed and we enter some fun and unresolved territory:
here due to 4 ... lbxe4 and White won't even get c1) 6 .Jlb5? dxe4 7 lbxe5 �g5! is a standard
his pawn back without severe disadvantage; for tactical trick of the kind that we also see in the
example, 5 �e2 d5 6 .Jlb5 f6! 7 d4 �d6 or Ruy Lopez. In this particular case White is in
7 ... .Jlg4. big trouble because of the attack on g2 and un
b) 4 0-0 lbxe4 5 lbc3!? is the maverick fortunate placement of the knight and bishop
Boden-Kiezeritsky Gambit (some inferior moves along the same rank. There follows 8 lbxc6 (8
are: 51:.el d5 6 .Jlb51'.c5!, 5 d4 d5 and 5 �e2 d4 �xg2 9 .l:lfl a6 101'.xc6+ bxc6 will win for
d5 6 Slb51'.g4! 7 d3 .Jlxf3 8 gxf3 lLif6). A tra- Black) 8 ...�xb5 9 lbd4 �e5 10 lbe2 .Jlf5 and
ditional main line goes 5...lbxc3 (Black can Black stands very well.
also spoil White's fun by 5...lbd6 6 Sld5 1'.e7 c2) 61'.d3! dxe4 7 1'.xe4 (D).
or 5... lbf6 6 .l:lel Sle7 7 lbxe5 lbxe5 8 .l:lxe5 d6 Now:
9 .l:lel d5 10 Si.f l 0-0, with equality in both c21) The traditional 7 ...Sld6 81'.xc6+ (8 d4
cases) 6 dxc3 f6 (interesting is 6 ...�e7!? 7 lbg5 exd4 91'.xc6+ bxc6 transposes) 8... bxc6 9 d4!
lbd8) 7 lbh4 g6 8 f4 �e7 (threatening ...�c5+) exd4 10 �xd4 0-0 11 0-0 is complicated with
9 'it>hl d6, when Black has a solid game and is an unclear imbalance, perhaps favouring White
still a pawn up, but some players would relish slightly; e.g., following 11...c5 12 �c3 1'.b7
the challenge facing White! with two bishops versus better structure.
c) 4 lbc3 (D) can be unique, especially be c22) 7 ...lbe7!? (Black plays ambitiously,
cause it can transpose from the Vienna Game threatening to win a piece by 8 .. .f5 and at the
have built up playing experience with 1 e4 e5

would even consider such a trade for White, but
those just starting out are often attracted to this
CDg5/CDxf7 idea (which appears in many 1 e4
e5 openings, such as the Giuoco Piano, Goring
Gambit and Ruy Lopez). It's important to know
that in most chess openings, two pieces are
better than rook and pawn, and usually the
equal or better of a rook and two pawns, until
there arises an ending or a considerably simpli
fied position. Of course, that claim contradicts
simple point-count chess (White has 6 or 7
points versus Black's 6). The explanation is that
the minor pieces enter the action earlier and co
same time sidestepping i.xc6+) 8 i.d3 (not 8 ordinate better in attack and defence, especially
'Dxe5?? °iVd4) 8 ...i.g4 9 h3 i.h5 10 i.b5+! on a crowded board. Keep in mind that rooks
(denying Black's knight access to the key c6- tend to get developed later and, more impor
square; 10 0-0?! °iVd5! and 10 g4?! i.g6 are tantly, to get blocked off if there are too many
definitely worse) 10... c6 11 i.c4 °iVd6 12 'iVe2 pawns and pieces around. There are few excep
f6 13 0-0 0-0-0 with equality. tions to this. Thus 5 'Dg5 and 6 'Dxf7 are mis
takes.
The Calm 4 d3 However, you should know that in an end
game with a rook and pawns versus bishop and
4 d3 (D) knight, the latter will often have trouble defend
This move quietly protects the e-pawn with a ing each other at the same time as they attempt
minimum of risk. to hold off the pawns. The bishop and knight
may do reasonably well if pawns are on the
same side of the board, but if the rook is escort
ing a pawn or two far from the opponent's king,
the minor pieces will normally have a very hard
time of it.
4 .tcs s 'Dc3
...
The position after 5 c3 is examined in Chap

ter 5 as part of the Giuoco Piano.
5 d6 (D)
...
We have arrived at a completely symmetrical

position.
It is not dangerous for either side but White's

modest pawn-push brings up some important
positional points. Instead of doing a systematic
analysis I want to emphasize a few characteris
tic types of positions that one should know to
understand this variation.
Before I even get to that, very inexperienced
players might want take a look at 4 ...i.e7 (a
slow move, but not a bad one) 5 'Dg5?! 0-0 6
'Dxf7? .l:.xf7 7 i.xf7+ Wxf7. Few players who
Two KNIGHTS DEFENCE 103
It's surprising how much chess content there when White has the choice of exchanging on
can be in such a simple position. We'll now e5, when his remaining e-pawns are doubled
look at a number of instructive continuations and isolated, or advancing to d5, which ham
and themes: pers his own pieces and does nothing positive.
6 tba4 can be met by 6.....tb6 7 tbxb6 axb6, There are a number of versions of this ex
which grants Black a solid game and an open change with varying results: sometimes the ad
a-file; but that may not be what he wants. vantages of the doubled pawns will outweigh
There is another way to give up the bishop-pair: their disadvantages, but just as often the reverse
6...�e7 7 tbxc5 dxc5. This sequence changes will be true. What counts is to be aware of the
the pawn-structure, and along with it the char issues.
acter of the game. In return for the bishop-pair It's very important to know when the move
Black gets an open d-file and freedom of devel ..tg5 (or ... .i.g4) is useful and when it is detri
opment. White's wished-for move d4 will be mental. Although that's a very complex question
next to impossible to organize. This kind of ex here are two types of positions that frequently
change varies from position to position, and anse:
crops up in the King's Gambit Declined (1 e4 In Case 1, White's bishop pins Black's knight
e5 2 f4 .i.c5 3 tbf3 d6 4 .i.c4 followed by d3 before Black castles by 6 .i.g5 (D).
and tba4) and even the English Opening, via,
for instance, 1 c4 e5 2 tbc3 tbf6 3 tbf3 tbc6 4
g3 .i.c5 5 .i.g2 d6 6 0-0 0-0 7 d3 .i.g4 8 tba4,
etc. In these cases most experts would tend to
regard the trade as an equal one, giving no ex
ceptional advantage to either player.
6 .i.e3 .i.xe3 (of course, 6....i.b6 or 6 ... h6 is
also possible; in the latter case the exchange on
c5 is not particularly effective) 7 fxe3 (D).
Then the harassment of the bishop by 6...h6 7

.i.h4 g5 is successful in so far as 8 tbxg5? hxg5
9 .i.xg5 l:lg8 fails to give White compensation
(the best try, 10 h4, is answered by 10....i.g4!
and White's attack is at an end). So White plays
8 .i.g3, but then his bishop is a little uncomfort
able running into a brick wall of pawns. Black
can play 8....i.e6, for instance, and prepare to
castle queenside with a fine game.
We talked about this in Chapter 3. White Let's compare Case 2, in which White plays
gains two important advantages from this trade: 6 a3, a handy move so that the bishop can be
he has opened his f-file and prevented Black's tucked away on a2. On an average level of play
knight from hopping to d4 (normally a main Black might respond with 6 ...0-0?!. But now
theme of the opening). That means that White White has 7 ..tg5! (D).
might want to move his forces to the kingside; Then he threatens tbd5, and the pin is both
e.g., tbh4-f5 is a good idea. But White's centre ersome anyway. By analogy with Case 1, Black
pawns have also lost their ability to advance might try 7 ...h6 8 .i.h4 g5?, but this time 9
successfully; for example, 7... 0-0 8 d4 .i.g4. tbxg5! hxg5 10 .i.xg5 is a whole different story.
Then Black can simply let the pawn sit on d4, Black has to prevent tbd5, for example by
(and in this case most of White's pieces aren't).

For that reason, 4 tl:ig5 has been called a begin
ner's move. Nevertheless, there have been
thousands of master games with 4 tl:ig5 for over
a century, and several whole books have been
devoted to precisely this position (not to men
tion lengthy parts of other books and countless
articles). The main point is that, principles not
withstanding, Black has a difficult time defend
ing f7 without making some kind of concession.
White's philosophy is simple: if it works, play
it!
4 d5
...
Black cuts off White's bishop with tempo

10 ...i.e6, when 1 1 tl:id5 i.xd5 12 exd5 tl:ib8 13 while dramatically helping his central situation
'irVf3 lands him in big trouble. For instance, and freeing his c8-bishop for action. There are a
13 ...'lt>g7 14 h4 tl:ibd7 15 i.b5 (or 15 J::i.h3 J::i.g8 number of alternatives over the next few moves
16 'irVf5, etc.) 15...tl:ib6 16 .l:.h3 (there are plenty that I won't be considering. One is the chaotic
of options; e.g., 16 c4i.d4 17 .l:.h3) 16...tl:ibxd5 4...i.c5! ?, which has the idea 5 tl:ixf7 i.xf2+ !?.
17 .l:.g3 'lt>h8 18 'irVf5. This has been analysed in excruciating detail,
Without the opponent castling, this sacrifi often past 20 moves, by players and theoreti
cial idea doesn't work, so you can see why both cians. Several experts seem to feel that playing
sides tend to play h3 and ...h6 before castling! 5 i.xf7+ instead of 5 tl:ixf7 grants some advan
The old saw about not moving pawns in front of tage. We'll leave the whole mess to them. In
your king has many exceptions. In almost any spite of the fascinating play that stems from
opening, with 1 e4 or 1 d4, there are plenty of this and other highly tactical sidelines, I shall
cases where either h3 or g3 will frustrate your mainly devote my attention to the main lines
opponent's attack. The same applies to ...h6 or and in general the more strategic (and popular)
...g6, of course. continuations. Naturally the course of events
after, say, 4 tl:ig5 i.c5 or the wilder 4 d4 lines
are instructive in the broader sense, conspicu
White targets f7: 4 llJg5
ously so in the realm of attack. They are, how
4 tl:ig5 (D) ever, singular in their nature, and the purpose of
this book is not to pursue particularities of forc
ing play but rather to broaden understanding of
openings and tie them together wherever possi
ble.
5 exd5 tl:ia5! (D)
Black continues to gain time for development
by attacking the c4-bishop. He is willing to sac
rifice a pawn to that end. The disorderly 5...b5
and 5...tl:id4 (sometimes transposing) fall into
the same category as 4...i.c5. A more familiar
line to inexperienced players is 5...tl:ixd5, when
6 tl:ixf7!? 'lt>xf7 7 'irVf3+ 'lt>e6 is known as the
'Fried Liver Attack'. According to theory this
line, if properly played, can be defended by
Black. White's other try, 6 d4!, has the similar
With this sortie White immediately breaks idea 6 ...exd4 7 0-0 i.e7 8 tl:ixf7!, this time lead
the rules about moving a piece twice in the ing to an extremely strong attack, at least accord
opening before the other pieces are developed ing to the older theory. That's because White
TwO KNIGHTS DEFENCE 105
has more open lines. A lot of study will be re

quired of anyone interested in these variations.
pieces out. Black's first goal is to attack the

king, with the hope that the cramping role of his
e4-pawn will render White's defence difficult.
But we're going to look at 5...ll'la5 because If that pawn is exchanged, he will develop
most good players choose that move, and be pressure down the d- and e-files. His only real
cause the resulting variations are marked by nu problem is the wayward knight on a5, which
merous defined strategic and tactical themes he hopes to reposition by ...lt'ib7 followed by
that can teach us about the Open Games. ...lt'ic5 or ...lt'id6.
6 .i.bS+ For his part, White wants to eliminate the e4-
This is White's point: he will stay a pawn pawn. If he can't do that he can bypass the pawn
ahead, having no pawn weaknesses himself. and put his pieces on more active posts, for in
His knight may look a little funny out there on stance by playing d4, .i.e3, lt'ic3 (or lt'id2), and
g5, but so does Black's on a5. Black has two perhaps 'iVd2. In a real game all of these plans
moves here: 6...c6 and 6....i.d7. Be forewarned conflict. You can only get a feel for the ideas by
that what follows is not a complete overview of looking at examples. Because of Black's open
the latest theory, but examples that will hope lines and pressure down the d-file, White will
fully illuminate the issues involved. probably have to keep his pieces on passive
squares while he unwinds.
Interposition with the Pawn
Estrin - Levenfish
6 c6!?
••• Leningrad 1 949
Black sacrifices a pawn, but he gains another
tempo by attacking White's bishop and thereby 11 f4!?
takes the initiative. This pawn advance has a poor reputation be
7 dxc6 bxc6 8 .i.e2 h6 9 ll'lf3 e4 10 lt'ieS cause it weakens White's kingside, but the re
.i.d6 (D) sulting play is fairly balanced. One advantage is
Of course there are legitimate alternatives that White keeps his d-pawn; compare 11 d4 in
for both sides along the way. For instance, White the games that follow this one.
could have played Steinitz's 9 ll'lh3 or, instead 1 1 exf3
••.
of 10....i.d6, both 10...'iVd4 and 10....i.c5 have It's not strictly necessary to make this cap
fairly good reputations. ture, but Black craves space and open lines in
But the position after 10....i.d6 arises more return for his pawn.
frequently than any other. Black wants to use 12 ll'lxf3 0-0 13 0-0 'iVc7 14 d4 c5!
his space advantage and develop quickly by at We've arrived at a position that can arise from
tacking the e5-knight. Whatever happens, he's other move-orders. Black wants to break up
a pawn down and has to keep making active White's centre and bring his rooks to the centre
and/or forcing moves before White gets his files as fast as possible. White simply needs to
get his pieces out, secure his position, and prove 18 'irVd3 tl:ig4 19 h3 is obviously risky, yet
that the extra pawn means something in the long plausible. Then 19 ... c4 20 'irVd4 ! i.c5? 21 i.f4!
run. Both sides have won their share of points. is good for White. This variation is in general
15 tl:ic3 a6 (D) double-edged, and neither side can afford to sit
passively by.
18...tl:ig4
Attacking h2 but focusing upon the weak
ness on e3.
19 h3 tl:ie3 20 i.xe3 J::i.xe3 21 .l:!.bl W/e7
Or 21.....if4!?. Black is putting extra pres
sure on the dark squares and limiting White's
plans. He certainly has enough for his pawn by
virtue of his bishop-pair and activity.
22 'irVd2? i.f4 23 'irVdl .l:Ib6 24 tl:id2!
Having messed up once, White finds the
right way to reorganize his pieces.
24 i.c7 25 J::i.f3 Wies 26 tt::in J::i.xf3 27 i.xf3
..•
(D)
A move designed to prevent tl:ib5.

16 dS!?
This pawn can become a target or it can pro
vide cover for White's pieces. Against other
moves Black will most likely play ...i.b7 and
.. J1ad8.
16 J::i.e8
•••
16 ...i.b7 17 Whl J::i.ad8 is an alternative,

hoping to put pressure on the d-pawn by direct
means.
17 'lt>hl!?
17 h3 would prevent Black's plan. Again,
17...i.b7 and ....l:.ad8 would probably follow
and White might answer in the same manner as
he employs in the game. By a clever reorganization White has de
17 . ..J::i.b8 (D) fended his d-pawn and has some control over
most key squares. He's not out of the woods
yet, but things are looking better. It's interesting
that in this game Black never achieved a full
central liquidation.
27 ...tl:ic4 28 tl:ia4 J::i.b S?!
28....l:!.f6! is better. It's important to keep
some pressure on White's king.
29 Wld3 tl:id6 30 'irVe3 'irVd4!?
The next fe w moves don't work out but Black
is in trouble anyway.
31 b3! i.fS 32 J::i.d l 'irVf6? 33 c4 l:tb7 34
tl:ixcS i.b6 35 b4
White is two powerful pawns up. The last
tactic 35 ... J::i.e7? can be met by 36 tl:id7!. Estrin
went on to win, but of course Black's opening
18 a3!? was not the cause.
Although White had success in that game he

was under significant pressure, in part because
11 f4 created an internal weakness on the sensi
tive e3-square. Most players would prefer to
have no weaknesses, even if it means having no
centre pawns!
11 d4
This is the most popular continuation, get
ting White's pieces out as fast as possible.
11. exd3
••
As was the case with 11 f4, Black doesn't

have to capture, but again he needs open lines to
pursue his attack, so why not create them now?
12 lbxd3 �c7 (D)
13 cS
.••
A double-purpose move that plans ...c4 and

prepares to bring Black's bishop to b7 along a
strong diagonal. Black has other strategies as
well:
a) 13....i.f5 14 .i.b2 0-0-0!? 15lLid2 �he8 is
a distinctive plan - maximum activity! Of course
Black's king won't be much better-placed than
White's because it lacks pawn-cover: 16 @fl
(16 h3 .i.xd3 17 cxd3 .i.e5 18 .i.xe5 �xe5 19
@fl lLid5 with an attack worth at least a pawn)
16...@b8 17 b4 (17 lbf3 tZ:le4!?) l 7...lbb7 18 a3
(18 h3) 18....i.xh2 and, having regained his
pawn, Black prospects aren't that bad, Short
Van der Sterren, Wijk aan Zee 1987.
A key position. Note that this is the 'vanish b) The aggressive 13...0-0 14 .i.b2 lbe4 was
ing centre' that we talked about in the introduc tried in Morozevich-Nenashev, Alushta 1994:
tory chapters. The Open Games (1 e4 e5) have a 15 lLic3 f5 16 h3 .i.a6!? 17 0-0 .l:lad8 18 �el c5
number of these because the move d4 is so ba gave Black some initiative.
sic to White's play, as is the move ... d5 to 14 .i.f3 �b8 15 c4!? 0-0 16 .i.b2 .l:le8+ 17
Black's. Obviously that results in a greater like @fl lbe4 18 .i.xe4 .l:lxe4 19 lbc3 �e6 (D)
lihood that the entire centre will be eliminated.
Such a position is naturally characterized by
open lines and tactical play. In this case the
tactics don't usually arise for a while as both
players jockey to achieve their most effective
formations. Then the action starts.
At this juncture we'll look at two games.
Beshukov - Malaniuk
Kstovo 1997
13 b3 (D)
The fianchetto is widely approved although
there are many options here. Getting a piece out
certainly feels right. Nevertheless, White was
more successful with 13 h3 in the next example. 20 lbbS?
Although this looks foolproof, 20 tl:id5 ! was Black's two bishops and attack are more
the way to go. than enough compensation for the exchange.
20 �d7 21 'irVc2 (D)
••• No better is 23 Wgl c4!, when 24 'irVc3!? is met
The alternative 21 'irVhS i.a6 might lead to by 24...J::i.g6.
22 i.c3 (22 tl:ixd6?! 'irVxd6 23 tl:ie5 tl:ixc4! 24 23 i.a6 24 h4
.•.
'irVxf7+ Wh8 25 tl:ixc4 i.xc4+ 26 bxc4 J::i.xb2 Playing for J::i. h3.
gives Black a meaningful advantage) 22... i.xb5 24 c4! 25 bxc4 tl:ixc4
••.
23 i.xa5 i.c6 with the idea ...J::i.g6. The oppo All of Black's pieces are participating in the
site-coloured bishops favour Black, who is the attack now. White's h l-rook is a tempo short of
attacker. getting into the action.
26 i.c3 'irVf5 27 'lt>gl i.b7 28 tl:ib4 i.e4
Or 28 ... �g4 29 tl:id5 J::i.g6.
29 'irVe2 J::i.g6 30 'irVn 0-1
A. Sokolov Timmermans
-
Paris open Ch 1 999
13 h3 (D)
21. .. J::i.xb5!
This move changes the whole equation. Now
White's interior weakness on d3 is exposed and
Black's two bishops finally are freed for attack.
21...i.b7 isn't as effective after 22 J::i.d l !, when
Black's attack is petering out.
22 cxb5
If 22 'irVc3, then 22...J::i.g6 23 cxb5 'irVxb5 24
.l:Idl i.b7 keeps the attack going. 13 i.f5
••.
22 �xb5 (D)
••. Maybe Black should just castle and hold
back on developing the bishop. It may want to
go to b7.
14 tl:ic3 0-0
14... 0-0-0!? would be like Short-Van der
Sterren in the notes to the last game.
15 0-0 .l:Iad8 16 l:tel a6
Black wants to prevent tl:ib5 in preparation
for ... c5, but it's not necessary. Instead, 16... c5
17 tl:ib5 i.h2+ 18 'lt>hl �b8! (D) would keep
the attack going.
Notice that after ...c5, Black can swing the
knight back to c6 and perhaps d4. This plan,
however you assess it, is the best try. From now
on White gains control of the position and one
is left wondering why anyone would sacrifice
23 J::i.d l that pawn in the first place!
TwO KNIGHTS DEFENCE 109
Interposition with the Bishop
Another continuation that gains compensation

for the pawn is 6...ii..d7. Here's a sample en
counter:
Bianchi - Escobar
corr. 1985
6 ii..d7 (D)
•••
17 .tn cs 18 'iVf3 lbc6 19 ii..e3

19 'iVxf5?? lbd4 traps the queen. Trying to
make something out of it by 20 1Wxf6 gxf6 21
ii..xh6 lbxc2 is futile.
19 ii..c8 20 lbe4 lbxe4 21 'iVxe4 lbd4
•••
Black tries to mix things up, since White is

completing his development with no problems.
22 ii..xd4 ii..b7 23 'iVh4 cxd4 24 �e2! 'iVaS?!
He may lack the firepower but it would be a
good idea to try 24...f5 and see how White re-
sponds. After the text-move, Black is not only a
pawn down but also has the worse position. This continuation is less common than 6 ...c6,
25 a3 1Wf5 26 .l:lael 'it>h8 27 'iVg4 'iVf6 28 but has a very good theoretical reputation.
llies ii..b8 29 f4 g6 30 lLid3 (D) What are the advantages of 6...ii..d7? For one
thing, it's a developing move, and developing
quickly is one of Black's most important goals
in this line. Moreover, 6...ii..d7 doesn't neces-
sarily lose a pawn (as 6...c6 does), because in
some variations Black may recapture White's
d-pawn. For the time being, White has an extra
pawn, of course, and that provides consolation
for his troubles. Furthermore, he has no weak
nesses. These imbalances will almost always
lead to interesting play.
7 'iVe2
7 ii..xd7+ 'i¥xd7 gives away any chances to
gain an advantage because Black regains the d
pawn.
7 ii..e7
••.
Protecting everything. Black tries to mark Black also plays 7 ...ii..d6 and defends his
time. e5-pawn. Then his queen is more cut off from
30 hS 31 'iVg3 ii..c6 32 lbeS ii..b7 33 h4!
••. d5, so he probably won't recover his pawn (af
@g7 34 'iVgS ii..a7 35 lLid3 'iVd6? 36 lle7 .l:.d7 ter White protects it with, for instance, 8 lbc3).
37 �le6! °iVdS? But when Black's kingside pawns get rolling
But 37... 'i¥xe6 38 �xe6 fxe6 39 lbeS is hope his bishop will become more effective. It's a
less for Black. trade-off that in practice has worked rather well
38 .l:.xg6+ 1-0 for Black.
8 lllc3
White defends his most important asset, the
d-pawn. Watch out for the trick 8 0-0? lllxd5! 9
�xd7+ 'irVxd7 10 d3 lllc6. To assess this posi
tion, just look at Black's central control.
8 0-0 9 0-0 c6! (D)
•..
You can see the results of the opening. White

is on the verge of getting massacred.
19 llle l f4 20 f3 �c5+ 21 d4
A sample of Black's attack would be 21 Whl
J::i. h6 22 h3 J::i.f5 23 a3 J::i.fh5 24 �b4 �b6 25 a4
�d7 and ... �xh3! next.
21...exd4 22 Whl J::i.h6 23 h3 .l:Ie8 24 a4 a6
Now it's a real gambit, one idea of which is 25 .l:.d2 'irVg5 (D)
to get that inactive knight off a5.
10 dxc6 lllxc6 11 �xc6
White has to win time to get organized. In
stead, multiple exchanges merely clarify Black's
central superiority: 11 lllf3 llld4! 12 lllxd4 exd4
13 �xd7 �xd7 14 llle4 .l:Iac8 15 lllxf6+ �xf6
16 'irVd3 (16 d3 'irVc7 and Black wins the c
pawn) 16...'irVc7 17 c3 J::i.fd8!. This clamps down
on White's development and leaves him strug
gling, Hendriks-Den Hamer, corr. 1985.
ll ...�xc6 12 d3 llld5
Here Black's two bishops, control of d4 and
mobile kingside pawns give him enough com
pensation for a pawn.
13 lllxd5?!
Too cooperative. Maybe White should risk 26 �xd4 �xd4 27 .l:Ixd4 'irVg3 0 1 -
winning another pawn by 13 'irVxe5. Then Black The threat is ....l:Ixh3+ and ...J::i.x el followed
has various dangerous moves such as 13 ... lllb4 by ... �xf3+ and there's nothing good to do
and 13...�f6, but White is two pawns ahead and about it. White's best idea is 28 J::i.a dl J::i.x el 29
will only have to give back one as he develops. J::i.x el J::i.xh3+ 30 gxh3 �xf3+ 31 'irVxf3 'irVxf3+
Another possibility is 13 lllce4. You shouldn't 32 Wgl 'irVg3+ 33 @fl �xh3+ 34 Wgl f3, but
get the impression that White has to sit back White would be materially and positionally
and get bowled over in this line. lost.
13 'irVxd5 (D)
•.•
Threatening checkmate. 13...�xd5 should Centra l Play: 4 d4

also be considered. In either case, Black plans
to play .. .f5 with a kingside attack. 4 d4 exd4 5 e5 (D)
14 lllf3 �d6 15 �d2 J::i.ae8 16 �c3 f5 17 It may seem odd to devote time to this con
l:.fdl J::i.e6 18 'irVfl J::i.g6 tinuation instead of its more famous alternative
5 0-0, yet the motivation for doing so is strong.

Apart from its popularity among top contempo
rary players (it is called the 'Modem Line'), 5
e5 produces positions with notable positional
features, at least before it degenerates into dis
array like the rest of the Two Knights! All three
of Black's replies are of interest.
An obvious alternative is 5 ll'lg5 d5! (as
usual, ...d5 frees Black's pieces if it doesn't fail
tactically; see the main line) 6 exd5 'ife7+ 7
�fl ll'le5 8 'it'xd4 (8 �b5+ c6 9 dxc6 bxc6)
8...ll'lxc4 9 'it'xc4 'it'c5 with equality.
move that combines themes of using the open

f-file in conjunction with an unopposed bishop
at b7) 9 �g5 (9 .l:.dl d5 10 ll'lxd4 ll'lxd4 11
.l:.xd4 c5 12 .l:.dl �e7 and here we have a good
French Defence!) 9...�e7 10 �xe7 'it'xe7 11
ll'lbd2 0-0 12 ll'lb3 .l:.f4 13 .l:.adl b6 14 ll'lbxd4
.i.b7 and ....l:.af8 is coming. This is based upon
analysis by Renet.
a2) 6 0-0 d5 7 exd6 ll'lxd6 8 .i.d5 .i.e7 9
.i.xc6+ bxc6 10 ll'lxd4 'it'd?!? 11 'ii'f 3 (or 11
ll'lb3 0-0 12 ll'lc5 'ii'f5) 1 l ...�b7 12 ll'lb3 c5 with
tactical complications in which the bishop-pair
will hold its own.
Now for a game: b) 5 ... ll'lg4 also seems to work out well
enough but needs to be tested a lot more before
Wendland - Groeber players will fully accept it. An obvious line is 6
corr. 1997 'ii'e2 'ii'e7 7 .i.f4, when Black plays the surpris
ing 7 ...d6! and White naturally replies with 8
5 ...d5 exd6 (D).
As a rule, Black should make this move
'when he can' in the double e-pawn openings,
and indeed, White has no way to avoid a loss of
tempo without concessions. On the other hand, B
one can argue that e5 itself costs White a move,
so barring a tactical disaster other responses
may be playable. Indeed, White hasn't estab
lished an advantage against the following two
rare replies, although he has a lot of leeway for
improvements. At any rate, both moves contain
useful positional ideas. I'll pick out a couple of
characteristic lines:
a) 5...ll'le4 (D).
al) 6 'ii'e2 was originally thought to be the
problem with 5 ...ll'le4, since 6 ...d5 7 exd6 is no
fun for Black. But after 6...ll'lc5, Black heads After the queens come off, White expects to
for the ideal blockading square on e6: 7 0-0 recover his pawn on d4 and secure the better
ll'le6 8 .i.xe6 (8 .l:.dl d5) 8...fxe6!? (a sharp middle- and endgame by virtue of Black's
remaining weak isolated d-pawn on an open 9 0-0

file. But Black has a clever trick that neutral As always, there are move-order issues for
izes those plans: 8...'iVxe2+ 9 ..i.xe2 ..i.xd6 10 both sides but that's more a matter of theory
..i.xd6 cxd6 11 tba3 tbge5 12 tbb5 (12 0-0-0 than understanding. As a case in point, delay
d3! 13 cxd3..i.e6 is equal) 12 ... d3! 13tbxe5 (13 ing 0-0 at this juncture by 9 ..i.e3 tips White's
cxd3 <J;;e7 with symmetry and equality again) hand. Black can then do without the .....i.c5
13...dxe5 14..i.xd3 <J;;e7! with equality, Fernan idea; e.g., 9.....i.e7 10 tbd2 c5 11 tb4b3 tbxd2
dez Garcia-Ivkov, Corunha 1990. 12 'iVxd2 d4 13 ..i.f4 ..i.b5!?.
We return to 5...d5 (D): 9 ..i.c5
•••
It may be that 9 .....i.e7 is playable, but it al

lows a dangerous pawn-roller that represents
Black's biggest nightmare in many double e
pawn openings. Look at this continuation: 10 f3
tbg5 (10. ..tbc5 11 f4) 11 f4 tbe4 12 f5 c5
(12.....i.c5!? 13 tbc3!?) 13 tbe2 ..i.b5 14 tba3
..i.c6 15 c4 d4 16 tbf4 ..i.g5 17 tbd3!, Svesh
nikov-Fercec, Nova Gorica 1996. At first this
seems all right for Black. Yet White's knight is
the ideal blockader of the d-pawn and targets
Black's weak doubled pawn on c5. This frees
White's pieces to roam the board, in particular
towards the kingside.
10..i.e3!? (D)
White modestly protects his centre before
6..i.b5 advancing pawns and exposing his own posi
6 exf6?! dxc4 gives Black space, free devel tion, but it may be too slow.
opment, the bishop-pair, and for the moment an White also has the aforementioned f3-f4-f5,
extra pawn. although with Black's bishop on c5 that may
6 tbe4 7 tbxd4 ..i.d7
••• not be easy to implement. For example, White
7 .....i.c5!? leads to complete anarchy in any can trade off kingside expansion for reduced
number of lines, the most absurd-looking idea central control by 10 f3 tbg5 11 ..i.e3, when
for White consisting of 8 tbxc6 ..i.xf2+ 9 <J;;f l 11...'iVe7 12 f4 tbe4 13tbd2 or 11...0-0 12tbc3
'iVh4 10 tbd4+ c6 11 tbf3 tbg3+ 1 2 <J;;x f2 is probably about equal.
tbe4++ 13 <J;;e3 'iVf2+ 14 <J;;d3 ..i.f5 15 tbd4
..i.g6 16 .l:!.fl and deep analysis has revealed var
ious forced draws. I'll refer you to specialists.
8..i.xc6 bxc6 (D) B
What is going on in this position? As usual,

Black is counting upon his two bishops and ac
tivity to compensate for his positional problems.
He would like to move his bishop to b6 and then 13 f6!

••.
successfully achieve the advance ...c5. Given The last of Black's dynamic ideas: to break
time, White would take advantage of the pawn down the centre. The modest 13...lbe6 is also
structure by a combination of moves such as equal.
(in some order) f3, 'it'd2, lbc3-a4 and/or lbb3, 14 lbc3
dominating the board from c5 and rendering 14 �xg5 fxg5 cedes Black the f-file, after
the bishops passive. That takes a few moves! which White can do little about ...�b6 and
10 'ii'e7!
••• ...c5.
The side with the bishops often depends upon 14 �b6!?
•..
tactical niceties to avoid disadvantages. Now 14.. .fxe5! 15 �xg5 l'kd6 was a tactical op
11 f3 can be answered by 1 l ...lbd6!, since the portunity which, however, arose logically from
e3-bishop hangs. That would be followed up by Black's positional play. Then 16 lbce2 �b6!?
...lbf5 (or ...lbc4) with active counterplay. 17 c3 exd4 18 cxd4 c5 keeps the initiative.
11 .l:.el At any rate, after 14...�b6, White stumbled:
Obviously 'it'd2 isn't on the cards, so White 15 lbce2? (D)
prepares f3 another way. But there's quite a dif A serious oversight. 15 lba4 is much better.
ference, in that f4-f5 won't be supported by a
rook on the f-file.
1 1 0-0 12 f3 lbg5 (D)
...
15 lbh3+!! 16 gxh3 fxe5 17 lbb3 .l:.xf3! 18

•.•
�xb6 cxb6 19 lbg3 .ihf8 20 .l:.fl �xh3 21

.l:.xf3 .l:.xf3
13 l'kd2 Black's mass of pawns gives him a distinct
As a case in point, 13 f4 lbe6! prevents f5 advantage.
due to exchanges followed by ...�xf5, and in This opening is a good illustration of posi
the meantime Black plans to get his centre roll tional trade-offs; the static features were as im
ing by means of ... �b6 and ...c5. portant as the dynamic ones.
7 Phi l idor Defence
1 e4 e5 2 tl:if3 d6 (D) and the next one. They believe (and theory
seems to verify) that the queenless rniddlegame
after 4 dxe5 dxe5 5 'iVxd8+ �xd8 is perfectly
fine for Black, who has the strategy .....i.e6 and
...tl:ibd7. White's main way to strive for an ad
vantage is 6 ..i.c4, when Black can accept dou
bled pawn in order to cover central squares:
6.....i.e6!? 7 ..i.xe6 fxe6 (D) with the idea .....i.d6
(or .....i.c5 first), ...tl:ibd7 and ... �e7. The posi
tion is considered to be equal.
The Philidor Defence has one virtue that few

e4 e5 openings have: Black decides what
opening is played! The underlying ideas of sur
render of the centre in the Philidor were men
tioned in Chapter 2; we'll explore them more
thoroughly and even look at a wild counterat
tacking scheme. Then we'll tum to a version of
the Philidor that uses a strongpoint approach in
one of its purest forms. The characteristic ideas
behind this not-so-old-fashioned opening are The reason that this move-order is consid
extremely instructive and applicable to many ered to be superior to (or at least less difficult
other openings. The Philidor is not a frequent than) 1 e4 e5 2 tl:if3 d6 3 d4 is that in the latter
visitor to master chess but has a remarkable fol case 3 .. tl:if6 4 dxe5 tl:ixe4 5 'iVd5 is awkward
lowing of contemporary players who have used for Black. See the note to 3 ...tl:id7 below (in
the defence extensively through the years. These the 'strongpoint' section). By playing 3 ...tl:id7,
include quite a few grandmasters, and even Ad Black can avoid this problem but runs into the
ams and Azmaiparashvili have dabbled in the possibility of 4 ..i.c4!, as we see below, not to
Philidor. Reaching back a few generations, mention the move-order 3 ..i.c4 examined in the
Tigran Petrosian was probably the last World next note. If all that is difficult to absorb, it will
Champion who tried it out. mean a lot more if you decide to take up the
It should be said, however, that most grand Philidor as Black or are faced with it as White.
masters who want to play the 'strongpoint' ver The most interesting aspect of this overview
sion of the Philidor Defence now use the order is that some extremely highly-rated grandmas
1 e4 d6 2 d4 tl:if6 3 tl:ic3 (thus far a Pirc De ters have been willing to play the Philidor De
fence) 3 ...e5. The idea is that after 4 tl:if3 tl:ibd7, fence via any move-order! After all, for many
Black has got into the main line of the Philidor years the Philidor was considered to be an anti
while avoiding the problems associated with quated and inferior opening for Black. Let's see
other move-orders that will be listed in this note what ideas have reinvigorated it.
PHIL/DOR DEFENCE 115
3 d4 (D) to gain time by ...tt:'ic6 without being pinned by

3 i.c4 is often overlooked with respect to i.b5 has a respectable record; 4...tt:'ic6?! 5 i.b5
move-order issues. Then 3 ...tt:'if6!? 4 tt:'ig5 d5 5 is the original continuation that made 4 'V//ixd4
exd5 seems bothersome, although a serious ex popular in the first place - after 5...i.d7 6 i.xc6
amination reveals that Black has equality or i.xc6 7 i.g5 with 0-0-0 soon to follow, White
stands only marginally worse after 5 ...h6 6 tt:'if3 achieves considerable pressure) 5 tt:'ic3 i.e7 6
e4!; for example, 7 tt:'ie5 (7 'it'e2 can be met by i.g5 0-0 7 i.c4 tt:'ic6 with equality.
7 ...i.e7 8 tt:'ie5 0-0 {e.g., 9 0-0?! i.c5! } or 4 tt:'if6 5 tt:'ic3 (D)
•••
7 ... i.b4!? 8 a3 !? 0-0) 7 ...i.d6 8 d4 exd3 9 tt:'ixd3

'V//ie7+ with the idea 10 i.e3 'it'e4!, equalizing.
Nevertheless, Black normally plays 3 ...i.e7 4
d4 exd4 (4...tt:'id7? fails to 5 dxe5 lt:Jxe5 {5...dxe5
6 'V//id5! } 6 tt:'ixe5 dxe5 7 'ii'h5 g6 8 'V//ixe5) 5
tt:'ixd4 tt:'if6 6 tt:'ic3 0-0. However, this transposes
into a 3 d4 exd4 line, which means that White
has successfully pre-empted Black's strong
point approach, that is, one in which Black
plays ...tt:'ibd7 without ...exd4 (see the 'Strong
point' section below). Thus Black may want to
look into 3 ...tt:'if6. Otherwise, 3 i.c4 makes an
other argument for the move-order 1 e4 d6 2 d4
tt:'if6 3 tt:'ic3 e5.
5 i.e7 (D)
...
The alternate strategy of activity and poten

tial attack begins with 5...g6, when White's
most aggressive set-up is 6 f3 i.g7 7 i.e3 0-0 8
'V//id2, as in a Sicilian Dragon. There usually fol
lows 8...tt:'ic6 9 g4 i.e6 10 0-0-0 tt:'ixd4 11
i.xd4. This is a position from which White has
won many games (and thus discouraged 5...g6).
Black could certainly use an open c-file, as in
the Dragon. He does succeed in throwing his
queenside pawns forward after l l ...c5 12 i..e3
'V/lia5 13 i.h6 i.. xh6 14 'V//ixh6 b5! 15 i..xb5
.l:.ab8 16 a4 a6, yet 17 .l:.xd6! axb5 18 e5 gives a
ferocious attack that has won several games for
The first and obvious point is that Black has White. All this is difficult to improve upon.
allowed White the greater share of the centre There have been scads of other attempts by
and blocked his own f8-bishop behind the d6- Black but he still seems to be in search of a sat
pawn, an unfortunate by-product of ...d6 but no isfactory solution. At any rate I shall concen
terrible thing in itself. There are now two basic trate upon lines with ...i.e7.
strategies that Black can pursue: surrender of With 5 ...i..e7, Black introduces a strategy
the centre or making e5 a strongpoint. that was explicated by Nimzowitsch: ...0-0,
....l:.e8, ...i..f8 and ...tt:'ibd7, both restraining and
Surrender of the Centre putting pressure on White's e-pawn. White has
more space and can thwart a direct attack but
3 exd4 4 tt:'ixd4
... still has to find a way to break down Black's
White has a reasonable alternative in 4 'V//ixd4, defences without allowing his pieces to spring to
although this hasn't scored as well as it did in life. This situation might remind you of similar
the 19th century after 4 ...tt:'if6 (4 ...a6 intending positions in the King's Indian Defence, such as
and increases his space advantage, but fails to

support the e-pawn (as f3 does). This position
has arisen in many games; here's one in which
Black takes the slow approach:
Restraint
lsanbaev - Sizykh
Novokuznetsk 1999
8 .i.f8 9 .i.f3 tllbd7 10 .i:tel c6 (D)

•.•
I d4 tllf6 2 c4 g6 3 tllc3 .i.g7 4 e4 d6 5 tllf3 0-0

6 i..e2 e5 7 0-0 exd4 8 tllxd4. In the case of the
Philidor, White's c-pawn is on c2 (rather than
c4, as in the King's Indian) and Black's bishop
is on f8 (rather than g7). You could argue that
in the King's Indian, White is more exposed in
the centre (d4 is unsupported by pawns); but in
the Philidor, Black's counterattacking chances
are limited by his passively placed bishop on
e7. Check out what happens in the second game
below!
6 .i.e2
White decides to go for a safe space advan The strategies are set. Black has insufficient
tage. He has an active alternative in 6 .i.c4 0-0 7 forces to attack and has to play with the back
0-0 leading to lines such as 7...l:te8 8 l:tel .i.f8 9 ward d-pawn that we also see in the King's In
a3 tllbd7 10 .i.a2 tllc5 (10...a6!?) 11 f3. It takes dian Defence. But d6 is well-protected, which
a born defender (with an opportunistic streak) gives Black the leeway to turn his attention to
to embrace this kind of position for Black, yet it the queenside; his main positive idea consists
is relatively solid. of attack on that wing based upon ...b5, with the
6 0-0 7 0-0 l:.e8 8 f4 (D)
... idea that White's forces are tied to protecting
against the freeing move ..d5.
.
For his part, White will develop, double on

the d-file, and slowly increase the pressure. He
B may prepare a pawn-break via e5 or a general
advance by g4.
11 .i.e3 tllc5 12 .i.f2 tlle6 13 't\Vd2 tllxd4 14
.i.xd4 .i.e6 15 !tadl tlld7
Directed against e5.
16 b3 f6 17 �hl .i.f7
Black has a passive but playable position.
White stands somewhat better but will need
time to organize a breakthrough (perhaps the
plan g4-g5 should be considered). In the event,
the game was quickly drawn.
With this move White commits to a pawn And now for something completely differ
structure in which he restricts Black's pieces ent:
PHIL/DOR DEFENCE 11 7
Counterattack White's first decision is important: where to

put the knight? From e2 it has prospects of as
Renet - Fressinet sisting on the kingside but has no particular
Clichy (rapid) 2001 square to go to yet. 10 ll'lde2 also allows one of
Black's pieces to settle on g4.
8 �f8 9 �f3 c5!? (D)
... The most common choice has been 10 ll'lb3,
which keeps White's pieces freer to move and
the g4-square covered, but from b3 the knight
doesn't have anywhere special to go to either.
There result some fascinating ideas following
10...ll'lc6 (D):
This bold move has been tried by at least two

very strong grandmasters and in at least 20
games! Black doesn't feel like defending pas
sively, so he aims at the central dark squares
(with ...ll'lc6 next) and stays true to the basic
idea of restraining White's centre. That by itself Here White has tried various moves to crack
might not make up for his pawn-structure but Black's strange-looking set-up:
Black also wants to advance his queenside a) 11 'it>h l a5!? (11....l:.b8 and ll ...a6 ap
pawns and attack White's pieces on that wing. pear more natural) 12 a4 �e6 13 ll'ld5 l:Ic8?!
To that end he will have support from a bishop (13 ...ll'lb4! is equal) 14 �d2 gave White some
on d7 and rook on b8. The obvious drawback is what better pieces in Brodsky-G.Kuzmin, Pula
his backward d-pawn on an open file. But as we ECC 1994.
see in several variations of the Sicilian De b) 11 .l:.el a5?! (this plan seems to appeal to
fence, such a pawn isn't necessarily an issue. players, but l l ...l:i.b8 looks considerably better)
There are lines like this in the Fianchetto 12 a4 d5? (12...�e6) 13 e5 (or 13 ll'lxd5!)
Variation of the King's Indian Defence. In that 13 ...d4 14 ll'lb5 ll'ld7 15 c3! dxc3 16 bxc3,
opening White's bishop is on g2, which is obvi Yurtaev-Payen, Calcutta 2000. Black is at a loss
ously analogous to a bishop on f3 in the Phili for moves here.
dor. In the position before us, however, Black is c) 11 �e3 d5!? (the craziest move of all!) 12
missing the powerful bishop on g7 that charac exd5 .l:.xe3 13 dxc6 'it'b6 14 'i¥d2?! (14 ll'ld2!?) .
terizes the KID, a condition that seems to be a Black's general strategy is a little hard to believe
serious drawback. Nevertheless, from f8 the in, but at this point he uncorks 14 ....l:.xf3! 15
bishop protects Black's only weakness on d6! l:i.xf3 c4+ 16 ll'ld4 �c5 17 'it>h l (17 l:Id l �g4)
How should White react? Obviously he will 17 ...�xd4 18 .l:.d l �g4 19 'i¥xd4 �xf3 20
have to restrain Black's expansion (presumably 'i¥xb6 axb6 21 gxf3 bxc6 with equality, Smirin
by a4). And he must eventually expand in the G.Kuzmin, USSR Ch (Leningrad) 1990. Note
centre or on the kingside. The move g4 sug that this was a high-level grandmaster game.
gests itself, although it must be properly timed 10 ll'lc6 (D)
...
so as not to weaken his king position. Now the advance e5 is prohibited for some
10 ll'lde2 time, and placing a knight on d5 is harmless or
23 J::!.d2+ 24 �bl 'iVxf3+ 25 l:.xf3 tl:id5!

•••
Now Black has the better ending.

26 tl:ixb5
26 tl:ige4 .i:td4 27 tl:ixd5 l:.xd5 28 tl:ic3 l:.d2
29 tl:ixb5 l:!.xc2 30 l:.c3 .i:tf2!. The active rook
and bishop-versus-knight favour Black.
26 l:!.xc2 27 l:!.dl l:.b8! 28 a4 tl:ib6 29 tl:ia3
•.•
l:.b2 30 l:!.cl tl:ixa4 31 tl:ic4 (D)
worse. You may recognize this kind of position

from the Sicilian Defence.
11 h3
11 f5?! .i:tb8 (11...tl:ie5!) 12 ..i.g5 ..i.e7 13
..i.f4 b5! 14 'iVd2 b4 15 tl:id l ..i.a6! 16 l:!.f2 ..i.f8
and White was totally disorganized in Scholl
Lutikov, Amsterdam 1968.
11 .....i.d7 12 g4!? h6! 13 tl:ig3 tl:id4
Black's ideal square.
14 ..i.g2 b5!? 15 a3..i.c6 31 l:.2b3
.••
This uses up the best retreat-square for 31...l:!.a2! is better still.

Black's d4-knight, but it does put pressure on 32 l:.xb3 l:.xb3 33 �g2 .l:tb4 34 f5? tl:ib2! 35
e4. l 5 ...l:!.b8 is safer and fully equal. tl:ixa5 tl:id3
16..i.e3 'iVb6 17 b4!? Here Black is clearly winning the endgame.
17 g5!? is interesting, now that Black's queen Very instructive.
has abandoned the kingside.
17 a5! 18 bxc5?!
..•
The e5 Strongpoint
White aims for e5 but he activates Black's
pieces instead of his own. 3...tt:Jd7 (D)
18 dxc5 19 e5 ..i.xg2 20 'it>xg2 'iVc6+ 21
..• Although it doesn't overlap with the gen
�h2 l:.ad8 22..i.xd4 l:!.xd4 23 'iVf3?! (D) eral themes that we're presenting, you should
23 'iVe2 l:.ed8 24 l:.adl tl:ie8 intending ...tl:ic7 be aware that another move-order issue arises
is probably better for Black, but manageable. after 3 ... tl:if6 4 dxe5 (4 tl:ic3 tl:ibd7 is the main
line) 4 ...tl:ixe4 5 'iVd5 tl:ic5 6 ..i.g5 'iVd7!? (after
6.....i.e7 7 exd6 'iVxd6 8 tl:ic3 0-0, White is for
choice) 7 exd6 ..i.xd6 8 tl:ic3; for example,
8...0-0 9 0-0-0 tl:ic6 10 tl:ib5 'iVf5 11 tl:ixd6 cxd6
12 'iVxf5 ..i.xf5 13 ..i.e3 tl:ib4! 14 tl:id4, Shur
Maliutin, Moscow 1997; now Black should
play 14 .....i.g6, when White has some advan
tage, although the position is still complex.
4 tl:ic3
4 ..i.c4 causes its own set of problems for
Black: 4 ...c6 (4 .....i.e7? 5 dxe5 tt:Jxe5 {5...dxe5??
6 'iVd5 } 6 tl:ixe5 dxe5 7 'iVh5 g6 8 'iVxe5) 5 tl:ic3
..i.e7 6 dxe5 dxe5 7 tl:ig5! ..i.xg5 8 'iVh5 with a
two-bishop position for White. In this situation
PHIUDOR DEFENCE 119
overprotection of the e5-pawn by means of

...'V/lic7.
8 a4 (D)
This move stops ...b5, which would win
much-needed space with tempo. The only good
way to do without it is to play d5, intending to
meet ...b5 with ...dxc6, a theme described be
low. But in this position Black could merely
work around the pawn by ... a5 and ...lt:Jc5, since
dxc6 is comfortably answered by ...bxc6, con-
trolling d5.
Black has decent counterplay if he moves

quickly: 8...'ii'e7! 9 i.xg5 ll'lgf6 10 'i!kh4 b5 11
i.b3 ll'lc5 and, for instance, 12 0-0-0 0-0 13
i.xf6 'ii'xf6 14 l'kxf6 gxf6 15 f3 a5. White must
stand better in this variation as a whole, but not
by much. Thus 3 ...ll'ld7 appears to hold to-
gether better than 3 ...ctJf6. Again, the lines of
the last two notes are a matter of practical play,
not of understanding, but they are important if
you decide to play the traditional Philidor
move-order 1 e4 e5 2 ll'lf3 d6 instead of 1 e4 d6
2 d4 ll'lf6 3 ll'lc3 e5.
4 ll'lgf6 5 i.c4 i.e7 6 0-0 0-0
.•• What are Black's goals now? He will gener
This time we see Black fortifying e5 as he ally follow up with ...'V/lic7 and/or ....l:.e8, to
does in so many lines of the Ruy Lopez. bolster e5 while keeping a careful eye on the
7 l:Iel (D) d6-pawn. Then we come to the point at which
White frequently plays the set-up with 7 he needs to develop his queen's bishop. This
'ife2 c6 8 a4 'V/lic7 9 l:tdl , when Black can do as may be prefaced by the safe moves ...h6, ...l:Ie8
prescribed in the note to 8 a4. and ...i.f8 (or even ...ctJf8), or Black may com
mence immediately. If he is allowed to com
plete the following plan he will usually have
solved his problems: Black places his pawns on
b6 and a6, his bishop on b7, and then advances
with ...b5. With completed development and
queenside play, he should stand well, especially
since his ideas of ...b4 and central attack are by
no means trivial to defend against.
Then what is White to do? There are a num
ber of answers depending upon one's style of
play and the specifics of the position. He has the
challenge of breaking down Black's defences,
and this time there is no open file or backward
pawn to focus upon. However, at the point that
Black plays ...b6 (and before ...i.b7) he is vul
7...c6 nerable to the move d5, since capturing will
A move necessary in order to get a little ma leave White in possession of the key outpost on
noeuvring room, and also to continue with the d5. If Black already has ....i.b7 in before White
plays d5, then Black has better chances of mak

ing a favourable mass-exchange upon that
square. The game will hang upon whether ex
changes and simplification leave White any
thing at all, or whether he can stifle Black's
counterplay by other means. There are three
standard alternatives to d5:
a) b3 and ..i.b2 or ..i.a3;
b) a5, to hamper Black's queenside plans;
and
c) tl:ih4-f5.
Incidentally, this kind of analysis suggests
that Black's plan would be even more effective
were White's bishop on e2 or fl, where it is of
ten placed. 10 d5!? ..i.b7 11 dxc6 ..i.xc6 12 ..i.b2 tl:ic5 13
We shall see these counter-strategies in the tl:id2 .i:tac8! provides piece-pressure to com
following sample game itself and in the note to pensate for the d5-square and White's potential
White's 9th move. along the d-file.
b) One of White's main ideas is to try to get
Vehi Bach - Cifuentes a knight to f5; for instance, 9 ..i.g5 h6! (9...b6 10
Platje d'Aro Barcino 1994 'V/J/.d2 ..i.b7 11 tl:ih4! { 11 dxe5 tl:ixe5} 11...exd4
12 tl:if5 gives White the better game; whenever
8 'V/J/.c7 (D)
... something like this can't be stopped, the plan of
A battle of heavyweights, lvanchuk-Azmai ...l:.e8 and .....i.f8 looks best) 10 ..i.e3 l:te8 (D).
parashvili, Montecatini Terme 2000, illustrates
Black's loss of the d5-square and his reaction to
it: 8 ...l:.e8 9 a5 (9 tl:ig5 .i:tf8 isn't helpful) 9.....i.f8
10 d5 b5!? (10...fil.c7 looks more natural but
Black doesn't want to be squeezed to death) 11
..i.b3! (11 axb6 tl:ixb6 12 ..i.b3 cxd5 13 exd5 is a
kind of position that we look at in several open
ings, where Black's potentially mobile kingside
majority is theoretically superior to White's on
the queenside; e.g., ...g6, .....i.g7, ...tl:ih5 and
...f5 might eventually follow; granted, the spe
cifics of the position will outweigh that factor
for some time, but I think that Black stands
well) 1 l...cxd5 12 tl:ixd5 h6 (versus ..i.g5) 13
c3!? (13 'V/J/.e2! a6 14 ..i.e3 with a small but defi
nite advantage), and here instead of 13...a6 14 11 tl:ih4!? (White should stand somewhat
..i.e3 with a grip on b6, Black should have better in such positions, although it's not clear
played 13.....i.b7 14 a6 ..i.c6 when he has the what he should play; maybe 11 a5) 11...exd4!
backward d-pawn, but it is well-defended (as in (this is normally a good response to tl:ih4, which
the Sicilian Defence). Then White has only a weakens control over d4) 12 ..i.xd4 tl:ie5 13
formal superiority. ..i.b3 ..i.g4! 14 f3 tl:ifd7 15 tl:if5 ..i.xf5 16 exf5
9 h3 d5. This position is difficult to assess, since
White has other thematic continuations. Not both sides have advantages.
all of them have been put into practice against c) 9 a5 is a natural alternative to hamper
challenging opposition. Black's queenside plans. Then one standard
a) White sometimes develops with 9 b3, hav idea for Black is to continue to batten down the
ing either ..i.a3 or ..i.b2 in mind. Then 9 ...b6!? hatches by 9...h6 (versus tl:ig5) 10 b3 l:.e8 11
PHIL/DOR DEFENCE 121
.i.b2 and now 11....i.f8 or 11...tt:'if8!? 12 h3 12...dxe5?! 13 tt:'ih4! and tt:'if5.

.i.e6. 13 tt:'ixe5 dxe5 14 :adl b5
These are just sketches of various set-ups. Once this move is in, everything is OK. No
In the maj ority of cases White will probably tice how neither White's knight nor bishop
retain some advantage with proper play, but have any forward square to go to.
not enough to invalidate Black's opening. Inci 15 .i.b3 h6 16 .i.h4 l:tad8 17 axb5 axb5 18
dentally, this kind of analysis again suggests l:txd8 l:txd8 19 l:tdl (D)
that Black's strategy would be even more ef
fective were White's bishop on e2 or f l , where
it is often placed.
9...b6 10 .i.g5 a6!
Neutralizing the idea of a5 and at the same
time contemplating expansion by ...b5.
1 1 iVe2 .i.b7 (D)
19 .i.cS!?
••.
A good idea, rerouting the bishop to a more

active position. Since White's e4-pawn is still
a concern and his bishop is away from the cen
tre at h4, the move 19...l:td4! was probably even
better. In general, Black has achieved excellent
activity, creating some problems that White
12 dxe5 needn't have allowed in his rush to simplify.
12 l:tadl b5 13 .i.b3 yields a standard pawn 20 l:txd8+ iVxdS
structure (also arising in the Old Indian and The opening is over and Black has at least
King's Indian Defences, and sometimes in the equality and perhaps more, since White's h4-
Ruy Lopez). Black has sufficient counterplay. bishop isn' t participating but the exchange .i.xf6
12 tt:'ixe5
•.. would cede the two bishops.
8 Ruy Lopez
1 e4 e5 2 lllf3 lllc6 3 ..llb5 (D) for Black. But what if Black imitates his re
These moves constitute the Ruy Lopez, aptly sponse to 3 ..ll c4 by playing as follows?
called the 'King of Openings'. It has domi 3. .. a6 4 ..lla4 b5 5 ..llb3 (D)
nated 1 e4 e5 chess for more than 100 years
and is considered the best chance for White to
gain the advantage in the play that follows
2 ...lllc6. Thereupon hangs the popularity of 1
e4 itself, no small burden for a single move to
bear.
After all, White is on the same diagonal as

after 3 ..ll c4 and Black has a couple of extra
moves in ... a6 and ...b5 that may help his posi
tion or at least not hurt it.
The answer is that the move ...d5 can no lon
ger be played with tempo, which negatively im
What's the point of 3 ..ll b5? One's first in pacts both of Black's normal defences after 3
stinct is that it threatens 4 ..ll xc6 followed by 5 ..ll c4. That consideration overrides all others, as
lllxe5, but Black's most popular answer 3... a6 we can see from the following discussion (see
shows that not to be the case, at least not imme Chapters 5 and 6 on 3 ..ll c4 if you need to). First,
diately. Then of course White wants to castle compare the old line 3 ..llc4 lllf6 4 d4 (notice that
quickly. But then why do most players use 3 after 4 lllc3, Black has the excellent response
..llb5 instead of the more aggressive-looking 3 4... lllx e4! 5 lllxe4 d5, whereas this would be a
..ll c4, which hits Black's weak f7-pawn? The blunder with White's bishop on b3) 4...exd4 5
answer is that 3 ..ll b5 is a prophylactic move e5, when 5 ...d5 ! gains a critical tempo. Not so
that works to squelch the opponent's opportuni with a bishop on b3 instead of c4. Then, look at
ties. If you look at the main lines after 3 ..ll c4, other main-line defence to 3 ..ll c4, i.e. 3 .....ll c5 4
for instance, it turns out that Black's key defen c3 lllf6 5 d4 exd4 6 cxd4 (6 e5 d5 !) 6.....llb4+ 7
sive/counterattacking move in a maj ority of ..lld2 ..ll xd2+ 8 lllbxd2 d5 ! 9 exd5 lll xd5 with
cases is ... d5 (as in most variations after 1 e4 equality; the Ruy Lopez prevents such solutions,
e5), attacking the bishop and establishing him as I'll show below for clarity's sake.
self in the centre. But putting a bishop on b5 ei 5. .. ..llc5
ther prevents or discourages that move. Let's The equivalent of the Two Knights Defence
see how this works in a few simple cases. Obvi would go 5 ...lllf6 6 d4! (D) (6 lll g5 is probably
ously, the immediate 3... d5? is bad due to 4 no improvement upon the main 3 ..ll c4 lllf6
exd5 'i1Vxd5 5 lll c3 with a terrible loss of time lines, but this is strong):
RUY LOPEZ 123
White's move serves not j ust to attack the al

most-trapped e4-knight, but also to stop ...d5.
That's the consistent theme involved in an early
..ll b3.
9 f5
...
Black can try the somewhat cheap win of

three pawns for a piece by 9 ...lll xf2?! 10 'it>xf2
dxc3+, but after 11 'it>g3! White connects rooks
and threatens 12 ..ll xf7+ 'it>xf7 13 'i1Vd5+. Thus
Black is stuck with 1 l...cxb2 12 ..ll xb2 0-0 13
lllc3! (or 13 h4) 13.....ll b7 14 lll e4 ..lle7 15 'i1Vd2
and White dominates the board. His king is un
touchable, and in general three pawns aren't
worth a minor piece this early in the game (un
6 ... exd4 (after 6 ... lll xe4 7 dxe5 White threat less two or three of them are passed and moder
ens 'i1Vd5 and ..ll xf7+, so Black needs to play ately advanced). When you factor in White's
7 . ..lllc5, when 8 ..\ld5 retains the bishop and es activity and far superior development, the as
tablishes the superior position - the e5-pawn sessment is clear.
cramps Black's game) 7 e5 lll g4 (there's no 10 cxd4 ..llb4+ ll lllbd2
...d5 response!) 8 0-0 and White threatens h3 Black can't even castle, but White will play
and .l:!.el. A sample line might be 8 .....llb4 (pre 0-0 and gai n a very large advantage.
venting .l:!.el ; 8 ...lllgxe5?? loses to 9 lllxe5 lllxe5 For the record, Black can try to j ustify 1 e4
10 .l:!. e l d6 11 f4) 9 c3 (or 9 ..lld5 threatening h3) e5 2 lllf 3 lllc6 3 ..ll b5 a6 4 ..ll a4 b5 5 ..ll b3 with
9 ... dxc3 10 bxc3 ..ll c5 11 'i1Vd5! 'i1Ve7 12 ..ll g5 the odd move 5.. .llla 5! ? (the 'Norwegian Varia
'i1Vf8 13 h3 lll x f2 14 .l:!.xf2 ..ll x f2+ 15 @xf2 and tion'), based upon 6 lll xe5 lllxb3 7 axb3 'i1Vg5,
White's pieces dominate the board. Notice how etc. But simply 6 0-0 d6 7 d4 is thought to give
without the move ...d5 Black wasn't able to White the advantage.
contest the centre.
6 c3 lllf6 7 d4 exd4 8 e5! Returning to 3 ..ll b5, we have seen one local
Were White's bishop on c4, Black would ized reason for preferring 3 ..ll b5 over 3 ..llc4.
have the resource ...d5! at this point. Instead the But what characterizes the Ruy Lopez itself?
knight has to move and lose time. For instance: Since each opening variation that begins on
8 llle4
.•• move 3 is so different in attributes, we cannot
8 ...lll g 4 9 cxd4 ..ll b4+ 10 lll c 3 and White speak of the ' nature' of the Ruy Lopez without
c hases the knights with h3 followed in some referring to specific systems. The most interest
cases by d5. ing way to approach the subject is to take a
9 ..lld5! (D) somewhat impressionistic historical look. In
the early days of the Ruy Lopez we saw some
understandable experimentation with moves
such as 3...llld4 and 3...f5, both still playable
today but on the very margins of legitimacy. As
positional concepts solidified, the great masters
of the late 19th and early 20th centuries drifted
towards 1 e4 e5 2 lllf 3 lllc6 3 ..ll b5 d6, which is
featured in the games of Steinitz, Lasker, Capa
blanca and many others. Without taking the
time to examine that variation (an exercise that
is well worth it), I can't demonstrate its draw
backs; but the crucial thing to remember is that
Black will immediately or eventually be forced
to surrender the centre by means of ...exd4 in
order to avoid complete passivity. Tarrasch is target the centre. The Chigorin Defence and
famous for helping to demonstrate this fact related lines were supplemented by systems
(and in fact he proposed the more dynamic which did without ... c5 entirely in order to at-
Open V ariation of the Ruy Lopez as an alterna tack e4 by means of ... i.b7 and ...�e8, with the
tive to the ...d6 lines). intention of pawn exchanges and even the free
Then came the so-called 'Closed' variations. ing advance ... d5. Thus, for example, the devel
The maj ority of players ultimately grew dis opment of the dynamic Zaitsev Variation and
content with having to live in the cramped situ lively advances in the formerly stodgy Breyer
ations that 3...d6 and ...exd4 usually imposed. Defence. Of late there have appeared new-found
Without dismissing options such as the re ways of opening lines in particular positions
cently-revived order 3 .. .lt::i f6 (the Berlin De based upon White's mode of development.
fence) 4 0-0 lt::i xe4 5 d4 lt::id6 6 i.xc6 dxc6 7 Within the ...e5/...c5 structures of the Chigorin
dxe5 lt::i f5, we find that the preponderance of Defence, for example, Black has skipped ..."Viil c7
masters turned to the more subtle move-order in favour of immediately exchanging centre
3 ... a6 4 i.a4 lt::if6 followed by ...i.e7, ...b5 and pawns, and in other cases the move ...exd4
...d6. The resulting variations tended to prevent alone has been used to establish a queenside
White from gaining the degree of space he majority accompanied by active piece deploy
commanded in the old ... exd4 lines. These for ments. Most interesting has been the complete
mations, arguably the most consistently im liquidation of the centre by means of the two
portant in all of chess history, are collectively exchanges ... cxd4 and ...exd4. Finally, confron-
named the 'Closed Ruy Lopez'. They are char tation by ...d5 is on the increase.
acterized by well-defended pawns on d6 and e5 What is White trying to do in the Closed
that form a bulwark against White's advances. variations? The first thing to realize is that there
Black generally achieves smooth development are very few variations in which he launches a
that targets each central square. To the extent mating attack or acts particularly aggressively
that White prevents Black's freeing moves, so within the first ten moves. In the main varia
Black stops White from redeploying his pieces tions, his idea continues to be prophylactic, i.e.
without risking the escape of his opponent's he tries to restrict Black's moves to those that
pieces from their cramped quarters. In particu are somewhat passive and fail to free his game.
lar, the moves ...d5 and ... exd4 carry with them The idea is that his space advantage in the cen
the potential for dynamism that can take ad tre (by no means a substantial one) allows him
vantage of White's relatively defensive minor to keep the game under control. When Black
pieces. In the meantime, his strongpoint of e5 does get frisky and tries to go tactical, White
and pawn on d6 give him a 4th-rank anchor that has attempted to arrange it that he will come out
is usually lacking in other e4 openings such as on top in any melee. In the meantime White
the Sicilian, Caro-Kann, Pirc, Alekhine, etc. slowly builds up his position and puts pressure
Arguably only the French Defence routinely on at least one area of the board and often two.
maintains a 4th-rank strongpoint, and that at the A queenside attack beginning with a4 is com
cost of a passive light-squared bishop. In the mon because it is not so easy for Black to de
Ruy Lopez too, there is generally a passive fend b5 without compromising his position.
piece in the form of the bishop behind the lines But over time White can also mount a kingside
on e7. However, that bishop is always devel attack. In that regard, notice the direction in
oped past the first rank and can theoretically in which White's bishops aim in the Ruy Lopez,
fluence both sides of the board. and they can be reinforced by knights on f5 (af
Such was the broad story of the Ruy Lopez ter the exotic-looking but now routine lt::id2-f l
until the past two decades. After playing strong g3/e3) while the other knight can head towards
point positions for so long, Black began to look g5 or, for example, to g4 via h2. If lt::if5 is pre
for more dynamic possibilities. First, without vented by ...g6 White sometimes plays i.h6
entirely jettisoning the idea of maintaining a (nudging the rook away from the sensitive f7-
pawn on e5 in the initial stages of the opening, square), the move "Viilf 3, and so forth. Ideally
top players increasingly used piece-play to (from White's point of view), Black will have to
RUY LOPEZ 125
play defensively until he can't protect against we devote a section to that variation below)
every breakthrough on both wings. This game 7 ... d6. Black has his bishop outside his pawn
program is what's glibly referred to as the chain and stands solidly. Of course, there's
'Spanish Torture'. We shall see how White's much more that can be said about 5 d3, but in
plans evolve when we inspect the individual general White would rather wait a move or two
Closed variations below. In its general con until he sees what his opponent is up to.
tours, by the way, the above description also 5 i.e7
•.•
applies to the Open Ruy Lopez: White tries to The first major decision about how Black will
keep Black's dynamism under control and then set his position up. After 5...b5 6 i.b3, 6 ... i.c5
switches to a gradual augmentation of his posi is the Meller Variation, examined in depth later;
tional advantages. and 6... i.b7 is called the Arkhangelsk, a varia
tion which I won't be investigating. The main
Let's look at the moves that introduce the alternative is 5...lllxe4, the important Open Vari
Closed Ruy Lopez: ation, which reaches its standard position after
1 e4 e5 2 lllf3 lllc6 3 i. b5 a6 4 i. a4 6 d4 b5 7 i.b3 d5. This will be discussed in de
We'll see the Exchange Variation with 4 tail in its own section.
i.xc6 dxc6 later on. Note that after 5 lll xe5 6 .l:!.el
'i1Vd4 Black recovers his pawn. Thus if White's 6 d4 is a sideline that might not be very inter
e-pawn becomes protected, the capture on e5 esting had we not seen something like it in the
may become a threat. introduction to the Ruy Lopez above, but with
4 lllf6 (D)
•.• the moves i.b3 and ... b5 included. There White
We saw the move 4 ... b5 above. gained the advantage, but here the presence of
the bishop on a4 makes equalizing relatively
easy. Two brief examples after 6...exd4 (D):
With this flexible continuation, Black threat

ens White's important e-pawn before deciding
upon the development of his other pieces. a) 7 e5 llle4 8 lll xd4 (8 .l:!.el lll c5 empha
5 0-0 sizes the bishop's poor position on a4) 8 ...0-0 9
For example, the passive 5 d3 allows Black lllf 5 d5 10 exd6 (10 lll xe7+ lll xe7 11 c3 lllc5
to become more aggressive without much risk: 12 i.c2 i.f5 is equal) l0... i.xf5 11 dxe7 lll xe7
5... b5 (notice that since e4 is covered, i.xc6 and 12 i.b3 lll c5 13 lll c 3 lll xb3 14 cxb3 'ilVxdl 15
lll xe5 has become a threat; however, 5...i.c5 is .l:!.xd l .l:!.ad8 16 i.f4 11z_ 11z Kramnik-Adams, Cap
another legitimate move-order, since 6 i.xc6 d' Agde (rapid) 2003. Neither side has any at
dxc6 7 lll xe5? loses to 7... 'i1Vd4 - both f2 and the tack or structural weaknesses.
e5-knight hang) 6 i.b3 i. c5 7 0-0 (7 lll xe5 b) 7 .l:!.el b5 8 e5!? lllxe5 9 .l:!.xe5 d6 10 .l:!.el
lll xe5 8 d4 i.xd4 { 8 ...lll xe4!? 9 dxc5 i.b7 10 (the initially attractive 10 .l:!.xe7+ 'i1Vxe7 11 i.b3
0-0 'i1Vf6 is also possible} 9 'i1Vxd4 d6 leaves invites 11...c5! ) l0 ... bxa4 11 lllxd4 i.d7 12
Black a full tempo up on the Meller Variation; 'i1Vf3 0-0 13 lllc 6 i.xc6 14 'i1Vxc6 llld7! ? 15
�3, Zapata-Anand, Manila OL 1 992. The sim is the old variation, perhaps not as good; at any
plest is now 1 5 .....llf6! intending ... a3. rate, Black wants to recapture with a bishop or
6 b5 7 ..llb3 (D)
••• queen on c6 to keep some control of d5) 1 2 h3
..lld7! 1 3 lt:Jbd2 c6 with a complex battle ahead.
The alternative 10 ..ll e 3 can lead almost any
where; e.g., 10...exd4 1 1 cxd4 d5 (or 1 1 ...lt:Ja5
and ... c5) 1 2 e5 lt:Je4, but Black should avoid
10...lt:Jxe4? 1 1 ..lld5 �d7 1 2 ..ll xe4 d5 1 3 ..ll c2!
e4 1 4 h3 ..ll h5 1 5 lt:Je5!.
7 d6
••.
This is actually a very important decision

that is sometimes misunderstood. As always, it
involves move-orders. If Black plays 7...0-0 at
this point, he can answer 8 c3 with the famous
Marshall Attack 8 ...d5, as discussed later. To
avoid that White will often play the Anti White prepares to play d4 next move. This
Marshall 8 a4 (as popularized by Kasparov) or is the starting-point of countless great battles,
8 h3 with similar intent (in that case to allow including games in the world championships
White to capture the pawn safely after 8 ...d5 9 between Kasparov and Karpov, Fischer and
exd5 lt:Jxd5 10 lt:Jxe5). However, after 7...d6, Spassky, and Smyslov versus Botvinnik. If
the move 8 a4 is no longer very effective be you look at the games between leading grand
cause e5 is defended and Black can develop masters today, they continue to contest this
smoothly by 8 .....lld7, 8 ... b4, 8 .....ll b7, or even same position and add new ideas.
8 ..lt:Ja5!?; see the section on the Marshall At
. We shall now discuss the Closed variations
tack for details. After 8 h3, Black can play themselves.
8 ..0-0 (or 8 .....ll b7, or 8 ... lt:Ja5!), when 9 c3 re
.
turns us to the main line.

To summarise: after 7 ...0-0, White can play Chigorin Defence
the Anti-Marshall 8 a4 or allow the Marshall by
8 c3 d5. By choosing 7...d6 instead, Black fore 1 e4 e5 2 lt:Jf3 lt:Jc6 3 i.b5 a6 4 i.a4 lt:Jf6 5 0-0
goes the Marshall but takes the sting out of the ..lle7 6 l:tel b5 7 i.b3 d6 8 c3 0-0 9 h3 lt:Ja5
Anti-Marshall's a4 move. Black makes the positional threat to ex
8 c3 0-0 9 h3 (D) change White's b3-bishop. This forces his re
The immediate 9 d4 enjoys periodic popu sponse, since you cannot afford to cede the
larity but you'll have to do the real work your bishop-pair in such a position without consider
self to discover its secrets. Since the point of 9 able compensation.
h3 was to prevent the pin on his knight, Black 10 ..llc2 c5 11 d4 �c7 (D)
will take immediate advantage of the chance to This is the Classical Chigori n Defence to
fight for d4 by 9 .....ll g4. This gives White the the Ruy Lopez, distinguished from the Modem
choice of 10 d5, when Black will try to break Chigorin by the move 1 l ...�c7. So far Black's
up White's pawn-chain by a timely ... c6; e.g., idea is clear: he has kicked the powerful Lopez
IO ...lt:Ja5 1 1 ..ll c2 �c8!? ( l l ...c6 1 2 dxc6 �c7 bishop off its best diagonal, secured some space
RUY LOPEZ 12 7
lvanchuk - Graf
Merida 2004
12 lllbd2 (D)
with ... c5, and then adopted a strongpoint policy

by defending e5. He feels that an immediate ex
change on d4 would amount to a surrender of the
centre and puts that idea on hold with l 1 ...'i1Vc7.
However, White must constantly watch over po
tential central exchanges, a situation that Black 12...:ds
hopes will limit his opponent's free develop This is a flexible move. It discourages White
ment. After l 1 ...'i1Vc7, Black will generally try to from playing dxe5 and leaves the bishop on the
bring his pieces out slowly before taking any c8-h3 diagonal for now in anticipation of d5.
drastic action in the centre. We shall see 1 2... cxd4 in the next game, with a
There are some drawbacks to this strategy. note on 1 2...lllc 6.
The first has to do with finding a useful, posi 13 b3!
tive plan. Exerting pressure down the c-file is Also flexible: White keeps Black's knight
natural but generally White can defend the crit out of c4 and would like to make simple moves
ical squares. Often Black will have to bring such as i.b2 and :c 1 . Although it seems ob
enough pieces to bear that a capture or two on scure at this point, b3 can also work with the
d4 will make White's centre vulnerable. At that moves a4 and i.d3, which are designed to tar
point White can implement his own ideas. He get b5 - watch for this theme in other games
can exchange pawns on c5 and try to exploit the with the Closed Ruy Lopez.
d5-square, or he can play d5 and then attack on 13... i.d7
the wings, sometimes by means of a4 and some Black sometimes plays ... i.b7 instead of
times by piece-play on the kingside. In general ...i.d7, but in the former case he should ex
White has the choice of developing his pieces change in the centre first, because of 1 3 ...i.b7
or closing the centre. 1 4 d5! (D).
Black's biggest problem tends to be his This gives White almost everything that he
knight on a5. He can return it to c6, of course, could want from advancing his pawn, a com
but that consumes time and can provoke a mittal decision that sometimes releases the
timely d5. Furthermore, White's d5 advance in pressure on Black's game. Let's consider this
and of itself can keep the a5-knight out of play. position. Black's bishop is badly placed on b7
At that point ... lllc4-b6 isn't bad, but it shows because its scope is limited by White's pawn
up an underlying problem with ... c5 combined chain and unfortunately the move ...f5 is no
with ... e5: a pawn on d5 can't be undermined where in sight. Thus Black will play ... i.c8 and
by ... c6. probably ... i.d7 with loss of time. What about
We shall come back to the idea of omitting that knight on a5? Right now it has no moves
...'i1Vc7. For the moment, here are two sample whatsoever because of White's pawns on b3
games which illustrate the classic Chigorin po and d5; as a rule if Black permits White to play
sition: d5 it's a good idea to have the move ...lll c4 in
hand. Then even if the knight is driven away, Sometimes White simply doubles or even
b6 is a good place from which to keep an eye on triples on the a-file in this kind of position.
White's a4 break, and Black reserves pros 20 :rs
•••
pects of ...lllbd7. However, in the diagram (af Or 20...h5!? 2 1 g5 lllh7 22 h4 f6 23 'i1Vd2.

ter 13 .....llb7 1 4 d5), Black will have to move 21 lllg3
his bishop in order to reroute the knight to the White has come out of the opening with a
uninspiring b7. From that square, alas, it is large advantage. Black simply has to avoid these
blocked from moving by the pawns on d6 and static positions unless he has already gained po
c5. Notice that if Black plays ...c4 and White sitional concessions.
plays b4, the situation is even worse. All right, 21 ..llc8 22 'i1Vd2 ..lld8 23 lllf5 llle8 (D)
•.•
it's a closed position and perhaps the knight can Or 23 .....llxf5 24 gxf5 'i1Ve7 25 '1t>h2 ..llb6 26
make just one more move from b7 to become :gl .
useful, i.e . ...llld8. But again, it is completely
restricted, this time by White's d5-pawn! Even
in a closed position, all this reorganization to
little effect gives White plenty of time to pre
pare and launch an attack. The moral of the
story is that with a bishop on b7, Black should
almost always play ...cxd4 and perhaps even
...exd4 once the restrictive move b3 is in. Apart
from that, both sides need to develop a feeling
about whether to play/allow d5 if c4 is still
available to the knight and/or Black's bishop is
placed on d7 in support of the queenside. These
decisions are terribly difficult and greatly as
sisted by playing experience with the opening.
14 lllfl lllc6?!
Black gets into trouble after this. 14 ...cxd4 This position deserves a diagram. Notice
1 5 cxd4 :ac8 looks better. Black's first rank. And the rook only recently
15 d5 lllb8 16 a4! .l:ta7 17 b4!? c4? (D) left a8! Aesthetics aside, we shall become very
This kind of position is nearly always much used to one feature of the Closed Ruy Lopez:
better for White, who has more space and all the regardless of who stands better, there are un
time in the world to build up. Black should have commonly few exchanges. Here we are on move
played 1 7...cxb4 18 cxb4 :cs 19 ..lle3 :b7 20 23 and there have been no pieces exchanged,
�c I bxa4 2 1 ..llxa4 'i1Vd8, just to keep some lines and only one pair of pawns.
open. Of course, he would still stand poorly. 24 @bl llld7 25 lllg 5 lll b6 26 f4! exf4 27
18 ..lle3 .l:!.b7 19 axb5 axb5 20 g4!? ..llxf4 ..llxf5 28 exf5 lllf6 29 ..llg3!?
RUY LOPEZ 129
White could consolidate by means of 29

llle4 ! lll xe4 30 ..ll xe4 i.M 31 :e2.
29...h6 30 lllf3?! (D) B
Again, 30 llle4! was quite strong.
In the game Black should now have played

22...lll c5 ! 23 lll g4 lll c6 with at least equality.
13 cxd4 i.d7 14 lllfl (D)
30 i.e7 31 :as lllbxd5 32 h4 lllxc3?!

••.
32...'i1Vc6 was Black's last chance to have a

say in things.
33 'i1Vxc3 lllxg4 34 :a6 'i1Vd7 35 'i1Vd4!
with a big advantage. White went on to win.
J. Polgar Acs
-
Hoogeveen 2002
12 lllbd2 cxd4
Black opens up the position to get some
breathing room.
1 2 ...lll c6 is really asking for White to play
dxc5, a Fischer favourite which intends lllf l This is the standard Ruy Lopez manoeuvre
e3-d5. If practice i s any guide, this general plan that has been popular ever since Steinitz started
causes little trouble for Black. Even in this fa playing it in variations with d3 instead of d4.
vourable form for White (because the c6-knight White's knight will either go to e3, eyeing d5
is exposed to a recapture on d5), Black can ap and f5 (while protecting c2), or to g3 where it
parently hold the balance: 1 3 dxc5 dxc5 14 lllf l covers f5 and protects the e-pawn (this discour
i.e6 (not a move that Black would like t o make ages ... exd4), while leaving the c l-bishop a
but he has to rush to cover d5) 15 llle 3 :ad8 1 6 good view of the kingside. Such meanderings
'i1Ve2 c4 1 7 lllf 5 ( 1 7 lllg 5 looks attractive but are ordinarily only possible in a closed position
1 7... h6 ! 18 lll xe6 fxe6 gives equality - another or in one with a stable centre.
case of the doubled e-pawns!) 1 7....l:!.fe8 ! 1 8 14 .l:!.ac8 15 llle3 lllc6 16 ..llb 3!?
•••
i.g5 llld7 1 9 i.xe7 lll xe7 20 lll g 5 h6!? (or Other players have preferred 1 6 d5 lllb4 1 7
20...lllf 8!) 21 lll xe6 fxe6 22 llle 3 (D), Fischer i.bl a5 1 8 a3 lll a6. Now 1 9 b4! should keep
O'Kelly, Buenos Aires 1970. the advantage because after l 9...axb4 20 axb4
This is an interesting position of the type dis lll xb4? 2 1 i.d2 White wins the knight. Black of
cussed in Chapter 3. Black's doubled pawns course hopes that the new weakness of c4 may
guard important squares and his knights have provide him compensation. Whether or not 1 6
good prospects, so the apparent weaknesses are d5 i s good, White opts here for activating the
not meaningful. light-squared bishop and keeping lines open.
This is a typical choice that the Lopez player

faces, and sometimes depends upon the style of Modern Chigorin
the player. Polgar is by any definition an at
tacker. Let's return to the position after 1 e 4 e 5 2 tbf3
16 tba5 17 tbd5 tbxd5 18 .llxd5 lbc4
••. tbc6 3 .llb5 a6 4 .lla4 tbf6 5 0-0 .lle7 6 :el b5 7
l 8 . . ..lle 6 would eliminate the powerful d5- .llb3 d6 8 c3 0-0 9 h3 tba5 IO .llc2 c5 l l d4 (D):
bishop; White maintains just a small edge with
19 a4.
19 i.g5! .llxg5
l 9 . . . lbxb2? fails to 20 'i1Ve2 .llxg5 2 1 tbxg5
lDc4 22 'i1Vh5, winning.
20 tbxg5 h6?!
Not best, but White still has the advantage
after 20 . . . .llc6 2 1 b3 ! lbb6 22 .l:!.c I .
21 tbxf7! .l:!.xf7 (D)
ll . ..cxd4
Here we have what I call a Modem Chigorin
Defence, in which Black skips ... 'i1Vc7.
12 cxd4 exd4!?
Rejection of the strongpoint approach! Black
shamelessly liquidates (i.e., surrenders) the cen
tre. With the recognition that the weakness on d6
isn't really serious (sometimes the pawn can
even go to d5), this radical policy has become an
22 :cl?! accepted one in just a few years.
The right move-order to implement White 's The alternative 1 2. . . .llb7 ! ? (D) hits the cen
idea was 22 .llx f7+! 'it>xf7 23 .l:!.c l . Black fails tre immediately so as to save time by compari
to take advantage of this slip. son.
22 'i1Vb8?!
••.
Correct was 22 ... .llc6 ! 23 .llxf7+ 'i1Vxf7 24

b3 lbb6.
23 b3 lbb6 24 .l:!.xc8+ .llxc8 25 .llxf7+ @xf7
26 dxe5 'it>e7
26 ...dxe5 27 'i1Vd8 'i1Vb7 28 .l:!.c l doesn't im
prove the situation.
27 exd6+
Still better is 27 lii'h5 ! , although that's not
clear without lengthy and complicated analy
sis, so the text-move is the practical decision.
White is winning in any case.
27 lii'xd6 28 'i1Vc2 .llb7 29 .l:!.dl lii'c6 30
..•
'i1Vd2 tbd7 31 .l:!.cl lii'f6 32 a4 .llc6 33 'i1Va5

bxa4 34 bxa4 tbe5 35 'i1Vc7+ .lld7 36 .l:!.dl 'i1Ve6
37 'i1Vc5+ 'it>e8 38 .l:!.d6 'i1Ve7 39 'i1Vc7! a5 40 :a6 Then 1 3 lbbd2 exd4 14 tbxd4 transposes to
1-0 the main line. However, White can also play 1 3
RUY LOPEZ 131
d5, which returns us to relatively normal chan lbc4) 1 9 . . . tbac6 20 axb5 axb5 2 1 .l:!.xa8 .l:!.xa8
nels and challenges Black to make something 22 tbfl tbe5 1h-1h Leko-Morozevich, Wijk aan
out of foregoing ..."Viilc 7. Instead he found him Zee 2002.
self in a familiar pattern in Morozevich-Pono 14 tbd2 .l:!.e8 15 b3 ..llf8 16 ..llb2 g6 (D)
mariov, Moscow 2001 : 1 3 ... .l:!.c8? (Black should
prefer 1 3 . . . lbc4 14 b3 lbb6 or 1 3 ... ..llc 8 1 4
lbbd2 ..lld7) 14 b 3 ! with problems similar to
those seen in the note about l 3 ... ..llb7 in the
Ivanchuk-Graf game above. Black ' s knight has
no return path and even the b7-bishop can't yet
get back to c8 ! Ponomariov understood these
issues and went for tactics by 1 4... "Viilc7 1 5 ..lld3
tbxe4 ! ? 1 6 ..llxe4 f5, but they fell short follow
ing 1 7 ..lld3 e4 1 8 ..llg5 ! ..llf6 ( 1 8 ... .l:!.fe8 1 9
..llxe7 .l:!.xe7 2 0 b4 tbc4 2 1 ..llxc4 "Viilxc4 22
"Viild2 ! ) 19 ..llxf6 .ihf6 20 ..lle2 exf3 2 1 ..llxf3 b4
22 tbd2 l:.ff8 23 a3 ! (threatening to win the
knight) 23 ... "Viilb6 24 axb4 "Viilxb4 25 l:.a4 "Viilc 3
26 l:.e3 ! "Viilb2 27 tbfl .l:!.c5 28 "Viile l ! and White
won the knight that Black marooned so early 17 "Viilf3
on. Black's dynamic possibilities were demon
We now return to 12 ... exd4 ! ? (D): strated by an inhuman following 1 7 .l:!.e2?! ..llg7
1 8 "Viile l .l:!.c8 1 9 .l:!.dl tbh5 ! (knights on the rim !)
20 ..llb l lbf4 2 1 .l:!.e3 "Viilf6 22 tb2f3 tbc6 23 Wh2
tbe5 24 g3?! tbd5 ! 25 exd5 tbxf3+ 26 tbxf3
.l:!.xe3 27 ..llxf6 .l:!.xel 28 tbxel ..llxf6 and Black
had the bishop-pair and a clear advantage in
Leko-Fritz 6, Frankfurt (rapid) 1 999 .
17 ..llg7 18 .l:!.adl .l:!.c8 19 ..llbl (D)
•.•
With 12 ... exd4, Black ignores his weakness

on d6 for the sake of activity, in the style of
modem openings from the Sicilian to the King's
Indian. In the following game we have an ex
ample of good strategy by both players.
Sorokin Ramesh
-
Sangli 2000 19 tbd7!?

.•.
Unveiling the g7-bishop and eyeing e5. Other

13 tbxd4 ..llb7 moves have been played including 1 9 . . . b4 and
The same idea is expressed by 1 3 ... .l:!.e8 1 4 1 9 ... tbc6 ! 20 tbfl tbe5 2 1 "Viile3 tbed7 22 "V/iif4
..llg 5 ! ? ( 1 4 tbd2 ..llb7 transposes to the main d5 23 tbg3, when 23 . . . tbxe4 (23 . . . "Viilc7 24 e5 ! ?
game) 14 ... h6 1 5 ..llh4 tbd5 ! ? 1 6 ..llxe7 tbxe7 was played in J.Polgar-Milos, Buenos Aires
17 tbd2 ..llb7 1 8 a4 "Viilb6 1 9 lb4b3 ( 1 9 tb2f3 2000, a marginally sound sacrifice but Polgar
brought home the point) 24 'bxe4 dxe4 25 Moves like this make the Ruy Lopez one of
i.xe4 i.xe4 26 l:he4 'bc5 is equal. the most fascinating openings in strategic terms.
20 'bfl (D) Black develops a piece backwards and cuts off
his own c8-bishop, at the same time taking his
eye off the vital d5-square ! But he is intent
upon forcing a resolution of the central dark
B squares, so l l . ..CDd7 serves the double purpose
of protecting e5 and clearing a square for the
bishop on f6 after pawn exchanges. Black also
recognizes that his queen might go to b6 in
stead of c7 in some lines, and even .. .f5 might
come into play. Let's look at two games:
Damljanovic - Ponomariov
Plovdiv Echt 2003
12 'bbd2
1 2 d5 ? ! releases the pressure just when
20 b4
•.• Black's pieces are best situated to destroy the
20 .. .f5 ! is also interesting and probably equal, centre: 1 2. . . CDb6 1 3 g4? ! (trying to anticipate
because White cannot exploit the a2-g8 diago the .. .f5 break, which would probably lead to
nal. the loss of White's important d-pawn) 1 3 . . . h5
21 'be3 14 CDh2 hxg4 1 5 hxg4 3'.g5 and Black already
The game has proceeded logically to this had much the better game in Fischer-Keres,
point and instead of the ambitious 2 1 . ..iVg5? ! Cura�ao Ct 1 962.
Black had 2 1 . ..'bf6 ! with equality. This mod 12 exd4
•••
em-style system seems to be fully playable. It 1 2 . . . cxd4 is the old continuation, and not
represents a dynamic treatment of even this necessarily worse. This move-order has some
most staid of openings. thing very specific in mind.
13 cxd4 'bc6 14 d5 'bce5 (D)
Keres Defence
Another way to bolster e5 has received renewed
attention from some of the world's top players.
It was first promoted by Paul Keres:
11 Cbd.7!? (D)
...
Black's idea is that he has achieved a Mod

em Benoni position with . . . b5 already in ! In
fact, White 's position is one that he might have
arrived at via the Modem h3/ i.d3 version of the
Benoni. Black would be thrilled to complete the
analogy by . . . �f6, so White has to act quickly:
RUY LOPEZ 133
15 'bxe5! 'bxe5 16 f4 'bg6 17 'bf3 3'.h4! 14 'bfl 'bc4 15 'b3h2

The simple idea is to play ...3'.g3 and force 1 5 b3 'bd6 centralizes the knight and pre
White into playing f5. Black is also ready to play vents 'be3.
.. .f5 himself. There were several games with 15 'bf6 16 iVf3 l1Vc7 17 'bg3
•.•
1 7 .. .f5 1 8 e5 dxe5 1 9 fxe5 3'.b7 before this one, White can get his pieces out by 1 7 'be3
but White finally got the better of the debate. 'bxe3 1 8 3'.xe3. Then 1 8 ... l::tad8 is at least equal
18 'bxh4 for Black.
1 8 l::tf l 3'.g3 19 f5 'be5 gives Black the dark 17 'bd6 18 'bhfl 'bd7 19 'be3 g6 (D)
•.•
squares that he needs.

18 iVxh4 19 f5?!
•••
19 l::tf l ! 3'.xh3 ! ? 20 gxh3 l1Vg3+ draws, al

though Black might simply bring a rook to the
e-file and see what develops.
19 'be5
•••
Black is at least equal, in part because White's

c2-bishop is so bad.
20 l:i.fl i.d7 21 3'.f4 l1Ve7 22 iVel f6 23 l1Vg3
l::tfe8 24 b3 a5
Black has secured the key e5-square and be
gun to attack.
Petrovic N . Davies
-
corr. 2003
At first sight we have a typical Ruy Lopez
12 dxc5 dxc5 13 'bbd2 (D) situation in which White has a kingside attack
and Black is trying to create queenside or cen
tral play. The problem for White is that Black's
kingside position is almost impossible to get at.
Thus Black has a significant advantage out of
the opening, and wins quickly when White over
reaches.
20 l1Ve2 c4 21 'bg4 h5 22 ctJh6+ @g7 23
ctJhf5+ gxf5 24 'bxh5+ @h8 25 exf5 iVc6 26
CDf4 CDf6 27 l1Vxe5 l::tg8 28 f3 l::txg2+ 0-1
B reyer Defence
1 e4 e5 2 'bf3 'bc6 3 3'.b5 a6 4 3'.a4 CDf6 5 0-0
3'.e7 6 l::te l b5 7 3'.b3 d6 8 c3 0-0 9 h3 CDb8!?
White intends to play the customary se (D)
quence CDfl-e3-d5. These days, players aren't As time went by, some players grew either
impressed by this single-minded attempt to get tired of the Chigorin Defence and/or suspi
a knight to the outpost. cious of its merits. Attention turned to this
13 3'.b7!
••• rather amazing retreat, the product of early
Black walked into it by playing 1 3 ... iVc7? 1 4 20th-century player Gyula Breyer's imagina
'bfl 'bb6 1 5 'be3 l::td8 1 6 l1Ve2 3'.e6 1 7 'bd5 ! tion.
CDxd5 1 8 exd5 i.xd5 1 9 'bxe5 in the famous 10 d4
game Fischer-Keres, Cura�ao Ct 1 962. White White sometimes holds off on this move,
has a terrific attack and is probably already hoping to exploit some subtle issues relating to
winning. tempi, but it really hasn't helped his cause.
compelled to play dxc6, will grant Black a

central majority. The second idea of capturing
Black' s e-pawn and swinging a knight to d5 is
not only fairly useless, as mentioned above, but
hard to implement. And that leads to White' s
third normal plan and in this case the most
promising: ctJfl -e3/g3. However, we see that
1 2 ctJfl ? drops the e-pawn. How exactly will
White get the reorganization he wants?
12 �c2
Remarkably, White abandons his favourite
diagonal without being chased away ! For the
record, the move 1 0 d3 that we mentioned
above had the point of ctJbd2-fl -e3 without
Indeed, the main alternative 1 0 d3 ctJbd7 1 1 needing to retreat by �c2. But in that variation
t0bd2 �b7 1 2 ctJfl ctJc5 has been analysed to Black could get . . . ctJc5 in, which in turn is the
more than 20 moves with a verdict of equality. target of attack by d4, and so forth - this is all
10 tLlbd7 1 1 ctJbd2 �b7 (D)
... much too obscure for our purposes. After 1 2
Note for the unwary: 1 1 ...l::te8 ?? allows 1 2 �c2, the onus i s on Black to make some sort of
�xf7+! with the idea 1 2 . . .@xf7 1 3 ctJg5+ 'it'g8 useful move as White pursues his knight tour.
14 lLJe6. Hence:
12 :es (D)
•••
To reach the position in the diagram, Black

has wasted two moves getting re-developed and The basic plan now is ... �f8, with . . . g6,
his pawns do not fight for control of d4 as in the ... �g7 and ... h6. This maximizes Black's mi
Chigorin and Keres variations (with . . . c5). Nor nor-piece pressure upon e4 and d4 such that
has he chased White' s bishop off the ideal a2- White needs to take into account the moves
g8 diagonal. In fact, he has a position that re ... exd4 and even . . . d5 at every turn. The move
sembles a Philidor Defence (as does the �xf7+ ... c6 is essential in many lines (again, compare
tactic). So what's the point? First of all, Black the Philidor Defence), but given the opportu
has no weakness on d5 and can expel any piece nity, Black might even be able to sneak in our
that lands there with ... c6. Then there's the ele old Chigorin-style ... c5, when for starters his
mentary fact that White has to search for a plan. d7-knight sure beats a knight on a5 ! It's all ex
Consider his three main approaches against the tremely complicated and the stuff of high strat
Chigorin and Keres Defences. White sometimes egy rather than mating attacks (at least for the
played d5, a pawn that is now subject to under foreseeable future). I'll feature a first-class en
mining by ... c6 with the danger that White, if counter.
RUY LOPEZ 135
Ponomariov Gyimesi
- Here's the standard Breyer problem: Black
Moscow 2005 has no obvious targets and no positive plan.
This was illustrated by 27 . . . l::txd3 28 l1Vxd3 ti:Jf6
13 tt:Jn 29 ti:Je3 3'.b7 30 ltig2 iVb8 3 1 3'.b2 l1Vc7 32
At this juncture White has played some fun 3'.b3 ! l::te7 33 iVd8 'iVxd8 34 l::txd8 ti:J6d7 35
damentally different ideas such as 1 3 b4, 1 3 b3 ti:Jh4 ! 'it'h7 36 ti:Jhf5 ! gxf5 37 ti:Jxf5 i.f6 38
and 13 a4. In the last case, for example, 13 ... i.f8 ti:Jxe7 i.xe7 39 l::te 8 and White soon won in
can be answered by 14 i.d3 aiming at the Shirov-Leko, Ljubljana 1 995.
queenside. In fact, White's main advantage in 13.....ifS 14 ti:Jg3 g6 15 3'.g5!?
these lines is that if he can prevent any radical At this point White has done extremely well
central action by Black, he can exert pressure at the highest levels with 15 b3 intending c4,
on both wings. Generally White gains a limited when 1 5 . . . 3'.g7 1 6 d5 or 1 5 ... c6 1 6 i.g5 ! has
edge if that happens, but nothing that allows given Black fits. But Malcolm Pein and Andrew
him to exceed a normal percentage score; e.g., Martin have done a thorough analysis to show
1 4 ... c6 1 5 b3 g6 1 6 l1Vc2 (often you'll see Black that 1 5 . . . d5 ! works: 1 6 3'.g5 h6 1 7 3'.h4 ! ? (D).
break out successfully; e.g., 1 6 3'.b2 3'.g7 1 7
ii.fl l1Vc7 1 8 l::ta2 d5 ! 1 9 axb5 cxb5 2 0 exd5
ti:Jxd5 with the initiative, Tseshkovsky-Dorf
man, Erevan Z 1 982) l 6 . . . 3'.g7 1 7 3'.b2 ti:Jh5 1 8
ii.fl iVb6 1 9 b4 ti:Jf4 20 dxe5 tt:Jxe5 2 1 tt:Jxe5
dxe5 22 c4 with the kind of typical slight pres
sure White often gets, Karpov-Beliavsky, Biel
1 992.
But the absolute key for Black is not to allow
an inflexible, passive structure, even if it is the
oretically defensible. This happens in our main
game, and in the following impressive contest
in which White played 14 b4 (instead of 1 4
i.d3) 14...ti:Jb6 1 5 a 5 ti:Jbd7 1 6 3'.b2 l::tb 8 1 7
l::tb l h6 1 8 ii.al 3'.a8 1 9 l::te3 ! g 6 20 iYe2 c 6 2 1
c 4 3'.g7? (never allow White a quasi-permanent In this position Judit Polgar destroyed Boris
bind unless forced to; now was the time to play Spassky in the 8th game of their Budapest match
the typical central counterattack with 2 1 . ..exd4! in 1 993 following 17 . . .dxe4 1 8 ti:Jxe4 g5 1 9
22 3'.xd4 bxc4 ! 23 ti:Jxc4 and then 23 . . . d5 with dxe5 ti:Jxe4 20 3'.xe4 i.xe4 2 1 l::txe4 gxh4 22
equality, or Shirov's idea 23 . . . c5) 22 dxe5 dxe5 l::td4 l::te7 23 e6 ! fxe6 24 ti:Je5. Pein and Martin
23 c5 ti:Jh5 24 g3 l1Vc7 25 l::td3 l::tbd8 26 l::td l tore into the line 1 7 . . . g5 ! 1 8 ti:Jxg5 hxg5 1 9
ti:Jf8 27 ti:Jf! (D). 3'.xg5 exd4! with huge complications ultimately
favouring Black. Notice how the modern Breyer
seems to do best when Black can successfully
implement the 'blow it all to bits' strategy as
opposed to the 'cleverly shift around' method
of defence. That said, the latter can also be sat
isfactory with very accurate play.
15...h6 16 �d2 �g7 17 a4 c5!? (D)
We've switched back to Keres-style play!
17 . . . c6 is the positional option which has in
mind an eventual ... d5, and may be preferable.
In spite of hundreds of games by the chess elite,
you' ll normally see the same set of basic struc
tures and approaches. One wonders about a
. . . d5 break instead, as in the previous note.
3'.f7 41 �d4+ @e6 42 ti:Jf5 ! i.f8 43 �xf4 @d7

44 ti:Jd4 �el + 45 @g2 i.d5+ 46 3'.e4 3'.xe4+
47 tt:Jxe4 3'.e7 48 ti:Jxb5 ti:Jf8 49 ti:Jbxd6 ti:Je6
50 �e5 1-0 Fischer-Spassky, Sveti Stefan/Bel
grade ( 1 ) 1 992.
20 3'.xb3 ti:Jc5 21 c4
21 3'.c2 ti:Jfd7 is easy for White.
21 bxc4?!
•.•
This gives up key squares. 2 1 . ..�d7 ! ended

in a draw in another game. In fact, theory goes
much further than this in some Breyer lines,
which is pretty amazing considering that the
play is so unforced.
22 3'.xc4 �c7 23 �e2 l:.ebS
lS d5 c4 Black has to get his rotten bishop back to a
The point: Black gets a knight to c5. Other decent diagonal so he brings the rook into activ
wise shutting in his own bishop and not having ity first. The one on g7 isn't looking so great ei
the ... c6 option would be the worst of both ther.
worlds. 24 a5 ii.cs 25 3'.e3 ti:Jfd7 26 J:.ecl @h7
19 b4 cxb3 You could argue that only now are we truly
Black certainly doesn't want to get squeezed at the end of the opening. As so often Black
to death, although he made that huge misjudge stands very solidly but is at a loss for a plan.
ment in a famous encounter: 1 9 . . . ti:Jh7? 20 3'.e3 27 ti:Jel i.f6 2S ti:Jf3 �dS 29 �d2! 3'.g7 (D)
h5 21 �d2 l:.f8 22 l:.a3 ! ti:Jdf6 23 l:.eal �d7 24
l::t l a2 ! l::tfc8 25 �c l i.f8 26 �al �e8 (Black
can only wait around and defend against White's
threats on the a-file) 27 ti:Jfl 3'.e7 28 ti:J l d2 @g7
29 ti:Jbl ! (D).
30 h4!
There you have it: once the opponent is tied
down to passive defence, you open up another
front. This is classic chess strategy.
30...hS?!
29 ... ti:Jxe4 ! ? (desperation because there was Maybe Black should make his stand on the
nothing to do about White's threat; for exam g6-square instead by something like 30 . . .l:.a7 3 1
ple, 29 . . . ti:Jd7 30 axb5 axb5 3 1 l::txa8 l::txa8 32 h5 l::tab7. It's easier to defend third-rank pawns
l::txa8 �xa8 33 �xa8 �xa8 34 ti:Ja3; the rest of than to surrender outposts and try to survive.
the game is pretty, so I'll give the moves) 30 31 ti:Jg5+ @gs 32 l::ta3!
3'.xe4 f5 3 1 3'.c2 3'.xd5 32 axb5 axb5 33 l::t a7 The kingside beckons, and in any case this is
@f6 34 ti:Jbd2 l::txa7 35 l::txa7 l::ta8 36 g4 hxg4 a useful move.
37 hxg4 l::txa7 38 �xa7 f4 39 3'.xf4 exf4 40 ti:Jh4 32 �e7 33 �dl
.•.
RUY LOPEZ 13 7
You can see the tactics coming now. All

White needs is one more piece, and he doesn't
fail to realize that.
33 l:!.a7 34 l:!.ac3 l:!.c7 35 3'.e2! l:!.b4 36
•••
'bxh5! (D)
system. The game can easily tum extremely

tactical and because it gives lively play is a
great favourite among today' s players on both
sides of the board. Some of the attacking ideas
associated with this variation have been among
the most beautiful of modem chess. From a
36 gxhS 37 3'.xh5 f6 38 i.f7+ <3o>f8 39
••• practical point of view, however, the fun and
'be6+ rtlxf7 entertaining main lines cannot be worked out
The attack continues successfully for White over the board and if your goal is opening mas
in lines like 39 ... 'bxe6 40 l:!.xc7 'bxc7 4 1 l:!.xc7 tery then they simply must be memorized. I
<3o>xf7 42 l:!.xc8 'bf8 43 h5 ! l:!.xe4 44 h6 and have primarily tried to indicate the general con
wins. tours of play, and for that purpose will present
40 'bxg7 <3o>xg7 41 �h5 l:!.xe4 42 3'.h6+ some dated but fantastic world championship
<3o>h8 43 l:!.g3 1-0 games, along with a couple of more recent ex
The rook got over there on the last move of amples.
the game! 1 1 'bbd2
From Black's point of view, the pure Zaitsev
If you look at a lot of games with the Breyer can only be used when a draw is acceptable, be
Defence you'll find that Black needs to fight for cause 1 1 CDg5 l::tf8 1 2 'bf3 repeats the position.
his own space (often by ... d5) and/or liquidate Some degree of bluff is involved. Of course
pawns; otherwise he can suffer through a long Black can deviate at that point and play another
period of inactivity with little room to ma defence to the Lopez, such as 1 2 . . . h6 intending
noeuvre. Even in the latter case most of the po to enter a very similar but less immediately ag
sitions are defensible with perfect play, but they gressive system by 1 3 . . . l:!.e8 and 1 4 . . . i.f8. That
are very difficult to handle in practice. Thus we sequence is sometimes named after Smyslov.
can look forward to the fighting methods as As for White, he can play 1 1 CDg5 l::tf8 1 2 f4,
holding the real key to the long-term success of which originally was thought to deter Black
the Breyer. from Zaitsev's move-order, but this is now con
sidered fine for Black after 1 2 ... exf4, and enter
prising players will most likely prefer l 2 . . . exd4
Zaitsev Variation 1 3 cxd4 d5 1 4 e5 'be4!, in view of 1 5 'bxe4?!
dxe4 1 6 l:!.xe4 'ba5, when Black will follow up
1 e4 e5 2 'bf3 'bc6 3 3'.b5 a6 4 i.a4 'bf6 5 0-0 by . . . 'bxb3 and . . . c5 with two bishops and ter
3'.e7 6 l:!.el b5 7 3'.b3 d6 8 c3 0-0 9 h3 3'.b7 10 rific pressure.
d4 l::te8 (D) 1 1 i.fS (D)
...
Zaitsev gets credit for developing this set White must now make an important decision
up with 9 . . . 3'.b7 and 1 0 ... l:!.e8 into a complete between 1 2 a4, which keeps lines open, and 1 2
d.5 .a more restrained approach with which he White prepares to shift his pieces to the king
hopes to cramp Black' s game. side, his only real area of strength. What fol
1 2 a3 stops Black's main . . .ti:Jb4 idea, but it's lows is more a demonstration of attacking and
slow. One interesting reply is 12 . . .�d7; e.g., 1 3 defensive skill than understanding, but the lat
d5 t:De7 1 4 ti:Jfl ti:Jg6 ! ? 1 5 .llc 2 c6 (usually the ter is still important:
sign of equality) 1 6 dxc6 .llxc6 1 7 .llg5 ti:Jh5 1 8 17 fS!?
..•
� ti:Jgf4 1 9 �g4 �xg4 20 hxg4 ti:Je6 ! with When Karpov played this no one really un
good counterplay, Bacrot-1.Sokolov, Reykjavik derstood how perilous it was. The idea is that a
2003 . central takeover would tend to be of more value
than a flank attack, but that has no real validity
Kasparov - Karpov as a principle of play. Over the years more play
New York/Lyons Wch (22) 1990 ers have drifted towards l 7 . . . c4, although that is
by no means easy either. Anand-Adams, San
12 a4 Luis Wch 2005 shows how White can target
This simple move threatens to pile up on the Black's king with a dangerous attack: 1 8 axb5
b-pawn and practically compels Black to un axb5 1 9 ti:Jd4 �b6 20 ti:Jf5 ti:Je5 2 1 l::tg 3 g6 22
dertake something active. ti:Jf3 ! ti:Jed3 (D).
12 h6 13 .llc2
•.•
Again, as in the Breyer, White' s knight can't

continue its journey to fl without this support
for the e-pawn.
13 exd4 14 cxd4 ti:Jb4! 15 .llb l c5
•••
A dynamic plan with all kinds of conse

quences. We have a Benoni structure in which
Black has already made considerable queen
side progress, but after White's next move the
b7-bishop will be shut off and White's pieces
are aimed at Black's king.
16 d5 (D)
16 ti:Jd7
•••
The whole point of . . . ti:Jb4 resides in this

move. which both prepares . . . ti:Je5 with ... c4
and . . . ti:Jbd3 to follow, but also contemplates Now it looks as though Black will beat off the
the risky .. .f5 to destroy White' s centre. Aban attack by eliminating White's bishops. First, he
doning the protection of Black's king is not threatens . . .�xf2+. Unfortunately, all of this
without danger, of course. was theory, and Anand had prepared 23 �d2 !
17 l:.a3! .llx d.5? (very tempting, but Black had to be
RUY LOPEZ 139
greedy and find his way through the 'only 20 axb5 d5 21 'bc3 l::txel+ 22 'bxel d4 23
moves' : 23 ... 'bxe l ! 24 'bxe l 'bxd5 ! 25 'bxh6+ 'ba2
i.xh6 26 'iixh6 l::ta l ! 27 l::txg6+ fxg6 28 'i:Vxg6+ Other involved lines begin with 23 'be4! ?
with a draw) 24 'bxh6+ ! 3'.xh6 25 'iixh6 'iixf2+ axb5 24 f6 llxa3 25 bxa3 CDd5 26 fxg7 i.xg7
26 'it'h2 'bxe l 27 CDh4 ! CDed3 (it' s hopeless at and 23 3'.e4 dxc3 (23 . . .lla7 24 'be2 CDf6 25
this point) 28 'bxg6 'i:Vxg3+ 29 'it'xg3 fxg6 30 i.f3 d3) 24 3'.xa8 'i:Vxa8 25 'i:Vxd7 'iie4.
'i:Vxg6+ @f8 3 1 'iif6+ ltig8 32 3'.h6 1 -0. 23 'bxa2 24 i.xa2 c4! 25 l::txa6 'bc5! (D)
.•.
It's now customary for Black to bring a

knight to d3 and White to swing his rook to g3
in this line; what counts are the specific tactics
and one's skill in carrying them out. Contrary to
the impression given by this game, there's a fair
amount of room for original play in even these
critical Zaitsev lines, and the odds of the aver
age player or even a master reaching something
this theoretical are extremely low. Which is to
say that the Zaitsev is still a fun system, both in
the tactical variations and in the positional con
tinuations given above.
18 exf5
A more famous and exciting contest from the
same match went 1 8 l::tae3 'bf6 1 9 ctJh2 ! ? 'it'h8
20 b3 ! (White feels that the attack requires only 26 l::txa8 'i:Vxa8 27 ii.bl d3 28 3'.e3 'i:Va5 29
one more piece and wants the bishop on the b3! 'bxb3 30 'bxd3! cxd3 31 i.xd3 'bes 32
long diagonal) 20... bxa4 (20... fxe4 21 'bxe4 ii.fl 'i:Vc7 33 'i:Vg4 'it'h7
'bfxd5 ! ? 22 llf3 ! CDf6 23 l::txf6 gxf6 24 CDg4 is 3 3 ... h5 ! ? is answered by 34 'iVd4 with equal
typically complicated) 21 bxa4 c4 22 3'.b2 fxe4 ity.
23 CDxe4 'bfxd5 24 l::tg 3 l::te6 ! 25 CDg4 ! 'i:Ve8? (a 34 3'.c4 3'.xc4
beautiful line is 25 . . . CDf4? 26 'bxh6! l::txh6 27 The tempting continuation 34 . . . 3'.e8? allows
CDg5 'iic7 28 'be6! 'bxe6 29 l::txe6 l::th4 30 l::tg4 35 3'.xh6 ! with the idea 35 ... @xh6 36 'i:Vh4+
llxg4 3 1 'i:Vxg4 CDd3 32 l::th6+ ltig8 33 'i:Ve6+ i.h5 37 g4.
'iif7 34 l::th 8+ ! ; best is 25 ...'bd3 ! 26 i.xd3 cxd3 35 'i:Vxc4 'i:Ve5 36 'iif7 3'.d6 37 g3 'i:Ve7 38
27 l::txd3 'i:Va5, which is unclear) 26 'bxh6! c3 'i:Vg6+ 'it'h8 39 3'.d4 3'.e5! 40 3'.xc5 'i:Vxc5 41
27 CDf5 ! cxb2 28 'i:Vg4 3'.c8 (28 . . . 'bc3 loses to 'i:Ve8+ 'it'h7 42 'i:Vg6+ 'it'h8 43 'i:Ve8+ •h-112
the pretty 29 CDf6 ! l::txe 1 + 30 'it'h2; and 28 ... g6
29 'it'h2! is a similar theme, threatening 'i:Vh4+ L. Dominguez - Morovic
and CDg5: 29 . . .'i:Vd7 30 ctJh4! 3'.c8 3 1 'bxg6+ Havana 2002
llxg6 32 'i:Vxg6, winning) 29 'i:Vh4+ l::th6 30
'bxh6 gxh6 31 'it'h2 ! 'i:Ve5 (3 1 . . .3'.g7 32 'bxd6 12 d5 (D)
'i:Vxe l 33 'i:Vxh6+! ) 32 CDg5 'iif6 33 lle8 i.f5 34 This changes the entire character of the game.
'iix h6+! 'i:Vxh6 35 CDf7+ 'it'h7 36 i.xf5+ 'i:Vg6 Notice how, as in the Breyer Defence, Black
37 i.xg6+ 'it'g7 38 llxa8 3'.e7 39 l::tb 8 a5 40 retains the option of playing . . . c6 to break up
i.e4+ 'it'xf7 41 i.xd5+ 1 -0 Kasparov-Karpov, White's centre. Speaking in general terms,
New York/Lyons Wch (20) 1 990. White will usually answer by dxc6, after which
18 i.xdS!? 19 'be4 i.f7!?
••. he has been fairly successful in keeping Black
Still another game between these giants con from achieving . . . d5 . The problem is that he
tinued 19 ...CDf6 20 'bxf6+ 'iixf6 2 1 i.d2 ! 'i:Vxb2 must devote all his resources to this effort and
22 i.xb4 i.f7 ! 23 lle6! 'i:Vxb4! 24 llb3 ! 'i:Vxa4 allow other equalizing methods. Here are a
25 i.c2 with an unholy mess. That game, Kas couple of ways in which this dilemma plays
parov-Karpov, New York/Lyons Wch (4) 1 990, out:
was eventually drawn. 12 'be7
.••
A more conventional approach is 1 2 . . . 'bb8 Black has full compensation and, remark
1 3 lLlfl lt:\bd7 14 CD3h2 (DJ. ably, 28 l1Vxc4 fails to 28 . . .d5 29 iVe2 i.a6 and
. .. CDd3 with ... i.c5+ to follow.
The Zaitsev is a wonderful opening whose

results are determined by both positional and
combinative skills. Much main-line theory (in
the attacking lines) has been worked out and
should be memorized if you're facing top-notch
competition. On the other hand, both White and
Black have alternatives at an early stage.
M011er Defence
1 e4 e5 2 'bf3 lt:\c6 3 i.b5 a6 4 i.a4 'bf6 5 0-0
5 d3 is slow and there are several established
l 4 . . .lt:\c5 1 5 i.c2 c6 (Black is using a typical solutions. But in our context it allows Black to
device from the Breyer Defence) 1 6 b4 CDcd7 get his bishop in front of the pawn-chain with
1 7 dxc6 i.xc6 1 8 i.g5 (the fight to stop . . . d5 out punishment: 5 ... b5 6 i.b3 i.c5. As we said
begins) 1 8 . . . h6 1 9 i.xf6 'bxf6 20 CDg4 ! 'bxg4 in the introduction to the Closed Ruy, there can
2 1 'ii'xg4 i.d7 22 iVf3 l::tc 8 23 i.b3 i.e6 ! 24 follow 7 0-0 (7 'bxe5 'bxe5 8 d4 i.xd4 ! 9
i.xe6 (sadly, the moves thus far are all theory) 'iVxd4 d6 threatening ... c5 - compare 5 . . . i.c5 6
24 .. .fxe6 25 a4 'iUd7 26 axb5 axb5 27 l::ted I 'bxe5 below) 7 ... d6 with equality. The follow
'ii'c6 28 l:.d3 l::ta8 29 l::tad I l::ta7 30 CDg3 l::tf7 3 1 ing variation can be directly compared to our
'ii'g4 l:.f4 32 l1Vg6 l::tf6 3 3 l1Vg4 l::tf4 with equal M0ller analysis: 8 a4 l::tb 8 ! ? (8 . . . i.b7) 9 axb5
ity. Pelletier-Bacrot, Biel 2004. A perfectly bal axb5 10 c3 0-0. Now White can play 1 1 d4
anced game. i.b6, but he's a tempo down on a note to our
13 'bfl h6 14 CD3h2!? c6 15 'bg4 'bxg4 16 main game.
hxg4 cxd5 17 exd5 iVd7 18 'bg3 a5 19 a3 a4 5 b5
•••
20 3'.a2 J:.ac8 21 'be4 5 ... i.c5 is also played at this juncture, nor
Here Black uncorked a beautiful exchange mally transposing; for instance, 6 'bxe5 'bxe5
sacrifice: 7 d4 b5 8 i.b3 transposes to the note to White's
21...l::tc4! (DJ 7th move below.
22 i.xc4 bxc4 23 i.xh6! 'bxd5! 24 i.d2 6 i.b3 i.c5 (DJ
tik.7 25 'bg3 l1Vc6 26 f3 'be6 27 iVe2 'bes The M0ller Variation. If you think about it,
l/2 - lfi this is a real test of the entire Ruy Lopez concept:
RUY LOPEZ 141
For years everyone assumed that this was a

grave drawback to 6 . . . i.c5. It gives White the
two bishops and the greater share of the centre.
It turns out, however, that Black's remaining
pieces have great scope and are very well-placed
to attack the key e4 point. First, ... c5-c4 is
threatened, winning the bishop, and that threat
gives Black time to develop his pieces aggres
sively. Here are just a couple of lines:
a) 10 c3 c5 (or 10 ... i.b7) 1 1 iVe3 0-0 12 ctJd2
l::te8 1 3 f3 i.b7, as in Kholmov-Lomineishvili,
Moscow 1 997, illustrates how Black can use
his active pieces to take the initiative. Among
other things he threatens . . .c4 and . . . d5.
if Black manages to create a successful strong b) 1 0 f4 tLlc6 1 1 iVc3 i.b7 1 2 e5 (this posi
point defence by ... d6 with his bishop outside tion arose a few times at the top levels until
the pawn-chain, then he has the best of both White became satisfied that he had no advan
worlds. The M0ller has enjoyed a great revival tage) 1 2 . . . ctJe4 1 3 iVe3 ctJa5 1 4 ctJd2 ctJxb3 1 5
among the world' s best players over the past cxb3 ! ? ctJxd2 1 6 i.xd2 0-0 1 7 i.c3 iVM 1 8
ten years or so. You can imagine how liberating l::ta el l::tfe8 1 9 iVf2 iVxf2+ 20 l::txf2 dxe5 1h-1h
it feels to live for once without that passive Anand-Topalov, Linares 1 997.
bishop on e7 ! But along with his advantages, 7...l::tbS 8 c3 d6 9 d4 i.b6
Black is presented with a few challenges. Con We've arrived at the main line.
cretely, White has the fork trick 7 ctJxe5 CLixe5 10 CLia3!
8 d4. Then, on a positional level, White's i.g5 1 0 axb5 axb5 1 1 ctJa3 ! usually transposes.
can pin the f6-knight and it can't be unpinned 10...0-0!?
by . . . i.e7. It also turns out that Black' s queen Black gambits a pawn for activity and pres
side is difficult to protect, much as in the sure on the centre. In fact, there isn't a lot of
Closed variations but more awkwardly because choice. But the recommended order 1 0. . . exd4 !
the bishop gets in the way. Perhaps most impor might eliminate some later issues: 1 1 cxd4 ( 1 1
tantly, Black has to be careful that, if his attack axb5 axb5 1 2 ctJxd4 ! ? is another method that
ing ambitions are frustrated, he isn't left with a probably isn't any better but deserves attention)
forlorn bishop cut off from the action on b6. l l . ..0-0 1 2 axb5 axb5.
7 a4! 11 axb5 axb5 12 tLlxbS (D)
White can also play the critical variation 7
CLixe5 CLixe5 8 d4, forcing Black into 8 ... i.xd4 9
iVxd4 d6 (D).
For his pawn Black has pressure on both the

e4- and d4-pawns; in particular, . . . i.g4 will be a
bothersome move. Moreover, White's pieces A. Ivanov - Zilberstein

on the b-file are loose. USA Ch (San Diego) 2004
12 exd4
..•
This move-order bypasses one of White's op 14 l:Ia4

tions. The game Adams-Leko, Miskolc (rapid) White plays one of the four or five moves that
(3) 2005 showed a clever new way for White to are available in this position. 1 4 l:Ia4 has been
play following 1 2 ... .Jlg4 1 3 .Jlc2 exd4 ( 1 3 ... d5 ! ? used with success, but this game shows its risky
i s an ambitious way to mix it up; as so often in side. Anand's 14 .Jle3 in the next game concedes
the Lopez, White is well-placed to meet early the return of a pawn but to good effect.
pawn-breaks and it seems he has some advan 14...l:Ie8 15 .Jlc2!?
tage after 14 h3 .Jlxf3 15 �xf3 exd4 16 exd5 15 .Jlg5 may well be better.
�xd5 1 7 �xd5 lt:Jxd5 1 8 .Jle4, as in L.Domin 15 �d7!
•.•
guez-Rodriguez, Buenos Aires 2005, but there We begin to see what Black has for the pawn.
may be improvements for Black) 14 lt:Jbxd4 ! This move indirectly attacks White's knight
(up to this game, 14 cxd4 was normally played) and, because White has to defend his centre, it's
14 . . .lt:Jxd4 15 cxd4 .Jlxf3 16 gxf3 lt:Jh5 1 7 Wh 1 . difficult to stop the queen from penetrating.
White has arranged a solid defence, and Black 16 lt:Jc3 .Jlxf3 17 gxf3
lacks the pieces to conduct a convincing king Naturally 1 7 �xf3 allows 1 7 . . . lt:Jxd4.
side attack: 17 . . . �f6 1 8 .Jle3 lt:Jf4 19 l:Ia4 ! l:Ia8 17 �h3 (D)
•.•
20 l:Ib4 (20 b3 !1 lt:Je6 2 1 d5 yields a small

advantage) 20. . . lt:Je6 (the crazy continuation
20 . . . l:Ial ! 2 1 �xal lt:Jd5 ! 22 .Jld l ! lt:Jxb4 23
�a4 lLid3 24 �c2 lt:Jf4 25 �d2 lt:Je6 26 d5 was
suggested, with White keeping the edge) 2 1
.l:i. g I l:Ifb8 2 2 f4 and White keeps the pawn and
the better game. 1 2 . . . exd4 avoids all this con
fusion.
13 cxd4
1 3 lt:Jbxd4 ! ? should again be considered, but
it leads to a new set of complicated options that
I'll have to leave to theory and practice.
13 .Jlg4 (D)
•..
This is the logical result of this variation

when Black's ideas have succeeded. He had
just enough pressure on d4 to cripple White's
f-pawns and at the same time keep enough
pieces on the board to make threats. This leads
to a nice tactical game, to which I shall give
only a few notes:
18 .Jle3 l:Ie5! 19 l:Iel ?! l:Ih5 20 .Jlf4 l:Ih4! 21
.Jlg3 lt:Jh5!
With the idea 22 .Jlxh4 lt:Jf4 and mate.
22 �e2 lt:Je5! 23 l:Id2
23 dxe5 lt:Jxg3 and mate next move.
23 lt:Jf4 24 .Jlxf4
.•.
The basic position. You can see how White's 24 �fl lt:Jxf3+ 25 Wh l �xh2+ 26 .Jlxh2
centre is under pressure and his b5-knight is l:Ixh2#.
loose. But is it enough? We' ll look at two con 24 lt:Jxf3+ 25 �xf3 �xf3
•..
trasting games: Black has a decisive material advantage.

RUY LOPEZ 143
Anand Shirov
- We are at the end of the opening stage, and
Groningen FIDE KO 1997 Anand has won it. Black's b6-bishop is left
with no good moves.
14 .Jle3 (D) 18 �c8 19 c4!?
•.•
This is double-edged because it makes the

light-squared bishop a bad one.
19 .Jlf5 20 .l:Ie2!
..•
White threatens 21 c5, which if played imme

diately would have been answered by ... .Jla5.
20 lt:Ja5 21 .Jla2 c5! 22 d5?!
••.
Now the a5-knight has no decent moves. On

the other hand it exerts nice pressure on the
queenside and White's a2-bishop is at least as
bad. White should have preferred 22 .Jlf4.
22 .Jld8! 23 .Jld2 �a6?!
•..
The nice idea 23 ... .Jlf6 ! 24 .Jlxa5 �a6 ! pro

vides equality, since 25 .Jld2? loses to 25 . . . .Jlxal
26 �xal .l:Ibl + .
24 �a4 .l:Ia8 2 5 .Jlc3 lt:Jb7 2 6 �dl! .Jla5? 27
14...�e8?! .Jlb2 .Jlb4?
This is Black's standard idea, to threaten The bishop should be back on the kingside
. . .lt:Jxe4 without losing material after .Jld5, but for defence. The rest of the game demonstrates
it doesn't appear to work out. Also bad is what happens when there are no pieces over
1 4. . . �d7? 1 5 .Jla4! lt:Jxe4 1 6 lt:Ja3 ! . So the dar there.
ing capture 14 . . . lt:Jxe4 ! ? is probably best; for 28 lt:Jh4! .Jlg6 29 f4 �a4
example, 1 5 �c2 ( 1 5 .Jld5 �e8) 15 . . . lt:Ja5 ! 16 29 . . .f6 30 lt:Jxg6 hxg6 3 1 �d3 f5 32 �g3 is
.Jla4 d5. killing.
15 h3!? 30 �xa4! .l:Ixa4 31 f5 .l:Ifa8 32 .l:Ie7! .Jlh5
Not bad, but 1 5 .Jla4 ! is very strong, with the If 32 . . . lt:Ja5, 33 fxg6 wins.
idea 15 . . .�xe4 1 6 lt:Jc3 . 33 g4! f6 34 gxh5 .l:Ixa2 35 .l:Ixa2 .l:Ixa2 36
15 .Jld7
.•. h6! (D)
Not 1 5 . . . .Jlxf3? 16 �xf3 lt:Jxe4?? losing a Attacks with reduced material are always
piece after 17 .Jld5. fun to watch.
16 lt:Jc3!
White gives back material but ends up with
the better pieces and a superior structure.
16 lt:Jxe4 17 .l:Iel lt:Jxc3 18 bxc3 (D)
•..
36 .l:Ixb2
•.•
A pretty line is 36 . . . gxh6 37 .Jlxf6 lt:Jd8 38

lt:Jg6 ! .
37 l:.xg7+ Wh8 38 .l:Ixb7 .Jlc3 39 .l:Id7 Wg8

40 l:.d8+ Wf7 41 .l:Ih8! .Jld4+ 42 Wfl 1-0
Even with White's mistakes you can see how
his strategy challenges Black to find sufficient
counterplay. The M0ller is a fascinating and
unresolved variation.
Open Va riation
I e4 e5 2 lt:Jf3 lt:Jc6 3 .Jlb5 a6 4 .Jla4 lt:Jf6 5 0-0
lbxe4 (D)
1 2 .l:Ia6 �d7 1 3 .l:Ixc6! �xc6 1 4 lt:Jxe5 �c5

( 1 4. . .�d6 1 5 .Jlf4) 1 5 lt:Jxf7 ! Wxf7 1 6 �f3+
We7 1 7 .Jlxd5 .l:Ib6 1 8 .Jlg5+ Wd7 1 9 .Jlxe6+
.l:Ixe6 20 �f7+ and wins.
6 d4
6 .l:Ie 1 provides another reason why delaying
... b5 until after ... lt:Jxe4 is helpful: 6 ... lt:Jc5 7
lt:Jc3 .Jle7 and the a4-bishop is attacked. How
ever, Black should steer clear of 7 ... lt:Jxa4 8
lt:Jxe5 lt:Jxe5?? 9 l:.xe5+ .Jle7 10 lt:Jd5.
6 b5 7 .Jlb3 d5 8 dxe5 .Jle6 (D)
•..
The starting position of the Open Variation

of the Ruy Lopez. Now we're leaving the realm
of Black's 1 st-3rd rank manoeuvring in favour
of staking a full claim to the centre. Perhaps be
cause of this assertive posture, the Open Ruy
has been the playground for some of the sharp
est tacticians in history.
As always, you'll have to be careful about
the move-orders, which we'll cover in the next
few notes. For instance, the inverted moves
5 ... b5 6 .Jlb3 lt:Jxe4?! can run into 7 a4! (D).
Instead, 7 d4 d5 transposes to the main line,
and 7 l:.e l d5 8 lt:Jc3 lt:Jxc3 9 dxc3 .Jle6 1 0 a4
b4 1 1 a5 ! ? is a recurring tactical idea: White With these moves we have reached the prin
threatens .Jla4 and then lt:Jxe5. This is some cipal variation of the Open Ruy Lopez. Black
what unclear but difficult for Black. announces that he is playing dynamically and
We've seen the power of a4 throughout the will steer clear of those protracted positional
Ruy Lopez, and it especially applies to the struggles that we have seen above (often with
Open Variation. After 7 a4, the play might go: no exchanges in the first 20 moves). Neverthe
a) 7 . . .b4 8 .l:Iel d5 9 d3 lt:Jf6 10 a5 ! . less, the Open Ruy has a great number of con
b ) 7 . . ..Jlb7 8 .l:Iel lt:Ja5 9 .Jla2 and White has sistent positional features, more so than the
ideas of d3 or d4 and lt:Jg5. average attacking system. Already the funda
c) 7 . . . l:.b8 8 axb5 axb5 9 .l:Iel d5 10 lt:Jc3 ! , mental question arises: tactics apart, what is
and now, fo r example, 1 0 ... lt:Jxc3 1 1 dxc3 .Jle6 each side playing for? In the positional phase,
RUY LOPEZ 145
we have an answer that comes close to being This queen move has always been hanging
universal: control of the d4-, e5- and c5-squares. around in the margins. White's usual idea is .l:Idl
Assuming that the e5-pawn isn't captured or followed by c4, although he may just play lt:Jbd2
liquidated, the real battle tends to be around d4 depending upon Black's course of action.
and c5. That may seem too broad a statement, 9 .ii..e7
•.•
yet if you study this opening you'll be surprised For example, 9 ... .ii..c5 is met by 1 0 lt:Jbd2.
to see that games consistently come down to 10 .l:Idl 0-0 11 c4! bxc4 12 .ii..xc4
this theme, whether directly or in the back We have reached a well-known position.
ground. If White can prevent Black from suc Black now enters a forcing sequence to salvage
cessfully playing the moves . . . c5 and . . . d4, he his d-pawn by means of a counterattack.
will generally have the upper hand. If Black 12 .ii..cS 13 .ii..e3 .ii..xe3 14 �xe3 �b8! (D)
•.•
gets one of those moves in without negative Moving out of the pin and hitting b2. 14 .. .f6 ! ?
consequences, he'll usually equalize or better. is Black's normal source of counterplay when
The reasons are relatively simple. From White's pressured in the centre. Theory doesn't like
point of view, securing an outpost on c5 can Black's chances in the tactics that follow, but
completely tie down his opponent and fix his they seem to work for him; e.g., 1 5 exf6 ( 1 5
backward pawn on c7 or c6. As for Black's �xe4?! dxe4 1 6 .ii.. xe6+ Wh8 1 7 .l:Ixd8 .l:Iaxd8
prospects, you can imagine the effects of the 1 8 lt:Jfd2 lt:Jxe5) 1 5 . . .�xf6 1 6 .l:Ixd5 ? ! �xb2 1 7
move . . . d4: freeing his e6-bishop, activating �xe4 .ii.. xd5 1 8 .ii..xd5+ Wh8 1 9 .ii.. xc6 .l:Iad8 ! .
his c6-knight, and cramping White' s pieces Unfortunately, simply 1 5 lt:Jbd2 ! forces some
(or, in the case of cxd4, opening up the d-file). kind of simplification with a small but definite
Since the opponents are usually very well edge for White.
aware how crucial these factors are, we'll of
ten see one of them switch to an attacking or
tactical mode if it appears they are losing the
d4/c5 struggle.
From the diagrammed position on the pre
vious page, I'll present game material with a
series of different 9th moves. It will at least
give you a start towards understanding how
the Open Ruy should and should not be played
by both sides.
Keres - Euwe
The Hague/Moscow Wch 1 948
9 �e2 (D)
15 .ii..b3 lt:Ja5 16 lt:Jbd2 lt:Jxd2?!
A single piece deserts the fight for c5 and
right away new problems appear. Later it was
found that 1 6 ... �a7 ! was the best way to fight
for c5 and the dark squares, as shown by 1 7
lt:Jd4 lt:Jxd2 (now this is all right) 1 8 �xd2
�b6 ! 1 9 .ii..c2 c5 ! 20 lt:Jf5 .ii..xf5 2 1 .ii.. xf5 .l:Iad8
22 b3 .l:Ife8 23 .l:Iel c4 24 �g5 �c7 ! with equal
ity, Kavalek-Karpov, Montreal 1 979.
17 .l:Ixd2 lt:Jxb3 18 axb3 .l:Ic8?!
Black doesn't recognize how utterly decisive
the control of c5 and d4 will prove. He should
aim for both squares by 1 8 . . . �b6; e.g., 1 9
.l:Ic2 ! ? ( 1 9 �xb6 cxb6 20 b4 i s also interesting)
19...'ifxe3 20 fxe3 .l:Ifc8 21 .l:Iac1 .l:Iab8 22 lt:Jd4. A blunder. It's worth showing how White
This looks good for White but his kingside wins anyway, due to his attack on two fronts:
pawns lack mobility and he may need a second 33...Wh7 34 .l:Ic8 �b7 35 e6! f6 36 .l:Id8 �g7 37
theatre of action. �c8! and the idea of .l:Id7 closes things out.
19 l:.cl 34 lt:Jc6 1-0
Here it is: White controls d4 and c5 and is
ready to double rooks (or triple pieces) down Ponomariov - Korchnoi
the c-file. Euwe doesn't want to be squeezed to Donetsk (3) 2001
death, so he tries to rid himself of the backward
pawn. 9 .i.e3 (D)
19 cS?!
•••
Last chance for 19...�b6, although this time

it fails to free Black's game after 20 .l:Ic5.
20 l:.xc5 .l:Ixc5 21 �xc5 �xb3 22 lt:Jd4! (D)
By this formerly-neglected but now popular

move, White targets the key d4- and c5-squares
right away. On the negative side he doesn't
challenge the e4-knight (as 9 lt:Jbd2 does), and
White has painted the ideal picture of dark potentially the bishop interferes with the pro
square control contrasting with Black's weak tection of White's e5-pawn by a rook on e l .
nesses. Note that Black's bad bishop has never 9 lt:Jc5
.•.
moved from e6. In the broader sense the rest is Black can take up the gauntlet by 9....i.c5! ?,
•just technique', but it turns out to be instructive daring White to win dark squares. A nice game,
indeed. by no means decisive for theory, went 10 �e2
22 'ifb7 23 h3 .l:Id8 24 Wh2
••. (10 .i.xc5 !?) 10 . .. .i.xe3 11 �xe3 lt:Ja5 12 lt:Jc3!
Preparing f4-f5. lt:Jxc3 13 �xc3 lt:Jc4 (the a5-knight was hang
24 g6 25 f4!
•.• ing and as usual the exchange 13 ... lt:Jxb3? 14
Even if you have wonderfully-placed pieces cxb3! would prepare to double on the c-file,
that are attacking weaknesses in the opponent's play lt:Jd4, and even indulge in f4-f5 in some
position, you usually need to have threats on cases) 14 .i.xc4 bxc4?! ( 14... dxc4 15 �e3 { 15
both sides of the board to break down his .l:Iadl �e7 16 lt:Jd4 0-0 17 �f3 } 15 ...0-0 16
defences. �c5 �b8 17 lt:Jd4 is bothersome but not too
25 hS•.. bad) 15 b4! (it's coming down to d4 and c5
Versus g4. again) 15 ...0-0 (Black wisely keeps the files
26 l:.d3 �d7 27 �b6 .l:Ia8 28 .l:Ia3 �a7 29 closed; 15 ...cxb3? allows White a big advan-
1i'b4 tage for the usual reasons after 16 cxb3 or 1 6
29 'ifxa7 will ultimately win, of course, but axb3) 1 6 lt:Jd4 �d7 ( 1 6...�e7) 17 a4 (D).
White doesn't want any technical problems. 17 ....l:Ife8 1 8 .l:Ife1 l:.ab8 19 h3 (an escape
29 �d7 30 �as .trs 31 .l:Ic3 .l:Ia7 32 .l:Ic5
••• square, a second front, or both?) 19 ....l:Ib6? (us
j_e4 33 �c3 �e7?? ing a valuable tempo, although it's not clear
RUY LOPEZ 147
bishops is unclear; however, they help White to

attack on the kingside. Compare the similar po
sition in the Komeev game above.
16 lt:Jed4 lt:Jxd4
Black's tactics are always dangerous in the
Open Ruy, and White had to anticipate that
1 6 . . . �xh3 ? ! fails to 1 7 .l:Ixc6! �g4 1 8 lt:Jh4!
�xh4 19 .l:Ixf6 �xf6 20 lt:Jf3 �h5 21 gxh3.
17 �xd4 .l:If5 18 lt:Je5 �c8 19 lt:Jc6
Again White has command of the c-file and
the d4-square, yet he must deal with Black's ac
tivity.
19...�d6 20 �c5! �d7 21 �xd6 cxd6
To cover e5 and c5. Now White shifts gears
what was better) 20 a5 .l:Ibb8 2 1 �d2 .l:Iec8 22 to make progress.
.l:Ia3 ! c5 23 bxc5 .l:Ixc5 24 ,:g3 (the point: be 22 lt:Jd4 .l:Ie5 (D)
cause of the knight on d4, the attack will crash
through) 24 ...�f5 ? ! (24 ... Wh8 25 �g5 .l:Ig8 26
f4) 25 �f4 ! �e6 (Motwani demonstrates a win
after 25 ... �g6 26 e6 ! �c7 27 exf7+ �xf7 28
.l:Ixg7+! Wxg7 29 lLif5+ - all these moves are
found instantly by a computer) 26 �h6 g6 27
lt:Jf3 ! .l:If8 28 lt:Jg5 f6 29 lt:Jxh7! 1 -0 Komeev
Martinez Lizarraga, Madrid 2000.
I O lt:Jc3! lt:Jxb3 11 cxb3 �e7 12 .l:Icl �d7
1 2... 0-0 1 3 lt:Jxb5 axb5 14 .l:Ixc6 .l:Ixa2 1 5
�c I ! and the familiar lt:Jd4 i s corning.
13 h3!? 0-0 14 lt:Je2 f6 (D)
This looks effective and it is certainly a move
with which Korchnoi has won many games, but
it has to be followed up precisely. l 4 . . . .l:Ifc8
may be better. 23 .l:Ic3 b4 24 .l:Ig3
As above, White needs both sides of the
board to break through.
24...a5 25 Wh2 �f7 26 lt:Jf3 .l:If5 27 �d4! g6
28 �d2! (D)
Black's bishop is awful.
15 exf6 .l:Ixf6?!
I 5 . . .�xf6 1 6 lt:Jed4 �xd4 17 lt:Jxd4 lt:Jxd4
1 8 �xd4 would cement White's bind on the po
sition. The influence of the opposite-coloured
28 .l:i.e8 29 lt:Jd4 .l:IfeS 30 f4! .l:Ie4 31 fS Wh8

••• give White a big advantage; both White and
32 'ifh6! "V/J/e7 33 lt:Jf3! "V/Jff8 34 "V/J/gS "V/J/g7 35 f6 Black should be aware of this idea) 1 4 lt:Jg3
'iff8 36 .l:Icl ! .Jlg6 15 h4 d4 ! (D).
Once more to the queenside.
36 h6 37 "V/J/d2 gS 38 .l:Ic7 .l:Ie2 39 "V/J/cl .l:I2e6
...
40 h4! .l:i.xf6 41 hxgS .l:Ig6 42 "V/J/f4 .Jle6

42 . . . .l:i.e7 43 lt:Jh4 ! .l:Ige6 44 .l:Ixe7 "V/J/xe7 45
g6! .
43 lt:Jh4! .l:Ig7 4 4 "V/J/d4 Wh7 4 5 "V/J/d3+ 1-0
A beautiful game, and another dream posi
tion for White.
It's time to see how Black can make his re

sources fully count. The themes in the notes
complement the main game.
Naiditsch - Korchnoi
Zurich 2002
1 6 .Jlg5 "V/J/d7 17 cxd4 lt:Jcxd4 1 8 lt:Jxd4 and
9 lt:Jbd2 lt:JcS 10 c3 .Jlg4 here I think that 1 8 ... .Jlxd4 ! would have equal
This standard move opens up the possibility ized or better. The point is that Black will get
of freeing Black's game by . . . d4. his cherished . . . c5 in.
11 iic2 (D) 12 .l:Iel 0-0
Korchnoi has also played 12 ... "V/J/d7 with vary
ing success. The idea is not just to overprotect
the d-pawn by . . . .l:Id8, but also to have the c6-
B knight protected in case White plays .Jle4 in re
ply to ... d4. Htibner-Korchnoi, Tilburg 1 986
went 13 h3 .Jlh5 1 4 lt:Jfl .l:Id8 1 5 lt:Jg3 .Jlg6
(ready for . . .d4) 1 6 lt:Jd4 0-0 17 .Jlf5 ! lt:Je6 (this
knight is pinned but it's also a superb block
ader) 1 8 .Jlg4 lt:Jcxd4 1 9 cxd4 c5 (D).
11 iie7
..•
A fascinating set-up is 1 1 .. .lt:Je6 ! ? 1 2 .l:Ie l

.i.c5. because Black has made two moves with
his light-squared bishop and then four with his
!ting's knight. Nevertheless, this is the ultimate
and consistent attempt to control the d4-square
- all of Black's minor pieces are devoted to it!
A critical continuation was seen in Kariakin
Flear. Hastings 2002/3 : 1 3 lt:Jfl ( 1 3 lt:Jb3 .Jla7 is
also important; it will be hard for White to hold Here's the key break, not necessarily optimal
off . . . d4 and/or . . . c5 forever) 1 3 . . . .Jlh5 ( 1 3 ... d4 because White will get an isolated pawn to
looks perfect until you see 1 4 .Jle4!, when the work against and some weak squares on the
tempo and pin on Black's knight turn out to queenside. Nevertheless, it's a good trade-off,
RUY LOPEZ 149
because d4 is what counts in this position. The

game continued 20 lt:Jf5 ! ? �a7 ! ? 2 1 lt:Jxe7+? !
�xe7 2 2 �e3 cxd4 2 3 �xd4 .l:Ic8 (suddenly
White has the bad bishop and Black has the
queenside advantage) 24 �d2 (24 .l:Ic l ?? .l:Ixc l
25 �xc l lt:Jxd4) 24 . . ..l:Ic2 25 �e3 � 26 �xe6
fxe6 27 f3 l:tfc8. Black's rooks are becoming
dominant, and we again have opposite-coloured
bishops. This time it's in Black' s favour: 28
.l:Iadl h6 29 a3 �e7 30 .l:Ic l �4 3 1 .l:Ixc2 .l:Ixc2
32 .l:Ifl �f5 33 .l:If2 .l:Ic4 ! 34 f4 (34 �b6 d4 35
�d2 d3 - always the same theme: unleash the
d-pawn if you can!) 34. . . �5 35 @h2 �di 36
.l:Id2 � 1 37 �c3 .l:Ie4 38 �f2 .l:Iel 39 .l:Ie2
.l:Ih l + 40 @g3 �xh3! 4 1 gxh3 �g6+ 42 @h4 17 f4 18 �d2 cS! 19 �dl
.•.
�f5 ! 0- 1 . A superb positional game. 19 dxc5? loses a piece to l 9 ...�f5 20 �c3 b4.
13 lt:Jb3 lt:Je6 (D) 19 c4 20 �c3 b4 21 �cl �xf3 22 �xf3
.•.
This idea again: play . . . d4 or bust! Or in any lt:Jxd4 23 �xf4 lt:Jxf3+ 24 gxf3 .l:Ic8 25 e6 d4
event threaten it. Black has also tried I 3 . ...l:Ie8 The key move again. White could resign.
and 1 3 . . . lt:Je4. 26 �es .l:IfS 27 �d2 d3 28 .l:Iadl c3 29 bxc3
.l:IxeS 30 .l:IxeS bxc3 31 �f4 c2 0-1
Svidler - Anand
Wijk aan Zee 2004
Here and in the game excerpts we see a more

balanced fight with each side utilizing their ad
vantages.
9 lt:Jbd2
We haven't seen a game yet in which the ex
change of e-pawn for f-pawn on f6 gives Black
compensation for White's greater command of
central squares. Here's a short excerpt in which
that's the case: 9 c3 �c5 I 0 �d3 0-0 1 1 �e3 f5
1 2 exf6 �xf6 1 3 lt:Jbd2 �xe3 14 �xe3 lt:Jxd2
14 lt:Jbd4?! 15 �xd2 .l:Iad8 1 6 .l:Ifel @h8 1 7 .l:Ie3 �g8 1 8
Now Black gets what he wants: the . . . c5 .l:Idl (D).
break. Better is 14 �d3 g6 15 �h6 .l:Ie8 1 6
.l:Iad l �f5 1 7 �d2 �xc2 1 8 �xc2, Geller
Unzicker, Bad Worishofen seniors Wch 199 1 .
14...lt:Jcxd4 15 cxd4 g6! ? B
Versus �d3. 1 5 . . . c5 is also possible.
16 �e3
1 6 �h6 .l:Ie8 1 7 �e3 f5 ! is a typical idea, tak
ing advantage of the fact that 1 8 exf6 �xf6 puts
so much pressure on the d-pawn.
16 fS ! (D)
•.•
17 �d3! ?
1 7 �b3 f4 and 1 7 h 3 �xf3 1 8 gxf3 f4 1 9
�c l c5! are as bad o r worse. There's really no
salvation.
1 8 . . .d4 ! 1 9 .l:Ieel ( 19 cxd4? lt:Jxd4) 1 9 . . . dxc3

20 'ifxc3 �xc3 2 1 bxc3 lt:Ja5 22 .Jlxg8 Wxg8
23 lt:Jg5 lt:Jc4 with equality, Kamsky-Anand,
Las Palmas PCA Ct (4) 1 995.
9 .Jle7
.•.
Perhaps the most disputed variation of the

Open Ruy Lopez begins 9 . . . lt:Jc5 1 0 c3 (D).
fxe3 �xe3+ 24 Wh l .Jld3, after which Black

has the better of it) 15 . . .�xd3 1 6 .Jlxd3 0-0-0
1 7 .Jle4 ( 1 7 .Jle2? ! lt:Jxe5 ! 1 8 lt:Jxe5 .Jlxe2)
1 7 . . . .Jlxf3 1 8 gxf3 lt:Jxe5 1 9 a4?! ( 19 b4! .Jld6
20 a4 was suggested; watch out for the bish
ops ! ) 1 9 ... b4 20 .Jlg5 f6 2 1 cxb4 .Jlxb4 22 .Jle3
lt:Jc4?! (22 . . .g6 and ... f5 would be strong) 23
.Jlf4 .Jld2 24 b3 lt:Jd6 25 .Jlxd6 .l:Ixd6 1h-1h
The first thing to note is that 10 ... lt:Jd3 1 1 Leko-Anand, Tilburg 1 998.
'ffe 2 lt:Jxc I is simply too slow and abandons the 10 c3 �d7 11 .l:Iel!? lt:Jc5 12 .Jlc2 .trs
queenside; for instance, 1 2 .l:Ifxc l .Jle7 1 3 a4! Entering into a less complicated position
or 1 2 1:.axc 1 .Jle7 1 3 lt:Jd4 ! . Here is an overview than we're used to. The simplification seems to
of Black's other options: help White somewhat.
a) 1 0. . .d4 1 1 lt:Jg5 ! �xg5 12 �f3 0-0-0 13 13 .Jlxf5 �xf5 14 lt:Jb3 .l:Id8 15 lt:Jxc5 .Jlxc5
.Jlxe6+ fxe6 1 4 �xc6 is an ultra-ultra-theoreti 16 .Jle3 .Jle7 17 lt:Jd4 lt:Jxd4 18 cxd4 (D)
cal variation that has been played and analysed
for many years following Karpov's use of 1 1
lt:Jg5 versus Korchnoi in the 1 0th game of their
1 978 World Championship match in Baguio
City. Anyone wishing to study this has to hit the
book s. You also have to consider whether you
want to go into 25-30 moves of tactical theory
following things like 1 4. . . �xe5 1 5 b4 �d5 1 6
'ffxd5 exd5 1 7 bxc5 dxc3 1 8 lt:Jb3 d4 1 9 .Jla3 g6
20 .Jlb4 .Jlg7 2 1 a4 d3, etc.
b) A more thematic yet unusual game went
l O . . . .Jlg4 1 1 .Jlc2 d4 ! ? (it looks awfully early to
advance in this manner, but Anand has played
. . .d4 on many early moves; we'll look at the
game and just a fraction of the theory) 1 2 lt:Jb3
d3 1 3 3lbl �d5 14 lt:Jxc5 ( 14 h3?! .Jlxf3 1 5 18 cS
.•.
'ifxf3 �xf3 1 6 gxf3 0-0-0 looks pretty bad for This is a necessity before White plays .l:Ic 1 ,
White) 14 . . . .Jlxc5 (D). and indeed it frees the d-pawn.
15 �xd3 (a line given by Mikhalevski is typ 19 dxc5 d4 20 .Jlxd4!
ically dynamic: 15 .Jlxd3 0-0-0! 1 6 .Jle2 �e4 ! ? The problem is that White is now two pawns
1 7 �el lt:Jxe5 1 8 lt:Jxe5 .Jlxe2 ! 1 9 lt:Jxf7 .l:Idl 20 up !
'ffx dl .Jlxdl 21 lt:Jxh8 .Jlc2 ! 22 .Jle3 ! .Jlxe3 23 20 0-0
•.•
RUY LOPEZ 151
20 ... .Jlxc5 21 .Jlxc5 .l:Ixdl 22 .l:Iaxdl puts 0-0 .Jle7 6 .Jlxc6!? dxc6 is the 'Delayed Ex
Black in great danger; e.g., 22 . . . h5 ! 23 .l:Id5 ! change Variation ' . Oddly, White takes two
.l:Ih6 24 .l:Iedl �c8 25 f4 with the idea f5. moves to capture the knight on c6 when he
21 c6! .l:Id5! 22 .l:Icl .l:Ic8 could have taken it straightaway on move 4. In
Black will get one of his two pawns back. fact White gets a couple of options that he
After that happens, bad bishop or not, White doesn' t get in the Exchange Variation. For in
can still play for a win. stance, 7 �e 1 ! ? gets out of the potential pin by
23 g3!? . .. .Jlg4. After Black defends his e5-pawn by
23 f3 may be a tad more accurate in view of 7 . . . lt:Jd7, White wants to develop by, for exam
23 . . .�e6 24 �d3 .l:Ixc6 25 �e4 ! f5 ! ? 26 exf6 ! ple, 8 b3 0-0 9 .Jlb2 .Jld6 10 d3, when Plaskett
�xe4 27 fxe4 .l:Ixc l 28 .l:Ixc l .l:Ixd4 29 fxe7. Davies, British League (4NCL) 2004/5 saw an
23...�e6 24 �d3 .l:Ixc6 (D) effective set-up for Black by 1 0. . . .l:Ie8 1 1 lt:Jbd2
lt:Jf8 ! 12 �e3 c5 1 3 lt:Jc4 lt:Jg6 with no prob
lems. There are other ways for Black to play, of
course. Losing a tempo means something even
in a positional opening.
25 .l:Ixc6
25 �e4 ! is still good, but in that case 25 .. .f5 !
26 �e3 (26 exf6?? �xe4) 26 . . . .l:Ixc l 27 .l:Ixc l
.l:Id7 at least forms a fairly solid blockade.
25 ... �xc6 26 �e4 4...dxc6
We'll stop here. White tried to press his ad Instead, 4 . . . bxc6 captures towards centre, but
vantage for many moves and after mutual inac as in many openings the capture away from the
curacies the game was eventually drawn. centre is better. Instead of freeing Black's queen
and queen's bishop, 4 . . . bxc6 slows Black's de
You can see that the Open Variation has a velopment and puts no obstacles in the way of
large number of variations to choose from. White's d4 (compare 5 d4 below). If Black
More significantly, both sides have options on could play an effective . . . d5 at some point he
so many moves that very little has been defini might have some justification, but White will
tively worked out. This is an ideal system for normally be able either to prevent that or to re
the average player, both from a practical and spond with e5 to good effect. Play can continue
educational point of view. 5 0-0 (5 d4 is also good, forcing the surrender
of the centre - remember that the side that sur
renders the centre needs quick development in
Excha nge Va riation order to compensate for that by means of piece
play) 5 . . . d6 6 d4 exd4 7 lt:Jxd4 .Jld7 8 lt:Jc3 lt:Jf6
1 e4 e5 2 lt:Jf3 lt:Jc6 3 .tbs a6 4 .Jlxc6 (D) (D).
The Exchange Variation of the Ruy Lopez is Black is actually a tempo down on 1 e4 e5 2
probably best known for its use by World Cham lt:Jf3 lt:Jc6 3 .Jlb5 d6 4 0-0 lt:Jf6 5 d4 exd4 6
pions Lasker and Fischer. Instead, 4 .Jla4 lLif6 5 lt:Jxd4 .Jld7 7 .Jlxc6 bxc6, a position that is
the same thing. Consequently, White will tend

to win the vast majority of pure king-and-pawn
endings.
However, there's a lot more going on here.
First of all, Black possesses the bishop-pair,
which can be in and of itself compensation for a
weakness. Then I think there's a rule of thumb
in his favour: usually the earlier in the game
that one side establishes an 'advantage in the
long run' such as doubled, backward, or iso
lated pawns, the less likely it is to last into the
endgame, or cause harm if it does persist that
far. In large part this is due to the fact that the
opponent has more time to adjust to the prob
favourable to White anyway ! Here's an exam lem and solve it directly or find counterplay.
ple without variations: 9 i.f4 i.e7 1 0 e5 dxe5 With that in mind, one can imagine that having
1 1 i.xe5 0-0 1 2 'i*'f3 c5 1 3 lbc6 i.xc6 1 4 'i*'xc6 more pieces on the board favours the side with
(Black's c-pawns are a disaster, and White is the weaknesses. And that's where a hypotheti
about to establish a large lead in development) cal problem arises: it may not seem vital at
14 . . .'ii'd7 1 5 'i*'xd7 ctJxd7 1 6 i.xc7 i.f6 1 7 i.a5 first, but White has a lead in development. This
l:.ab8 18 b3 l:tfe8 1 9 l:tadl lbf8 20 lbd5 i.b2 2 1 means that he can sometimes control the dis
:fe l :xel + 22 l:txel lbe6 23 lbc7 lbxc7 24 position of forces and arrive at the kind of po
ii.xc7 l:tc8 25 l:tdl i.f6 26 i.f4 c4 27 g3 cxb3 sition in which Black will be compelled to
28 cxb3 l:tc2 29 Ik l ! and White easily con exchange pieces. If the pawn-structure isn't
verted his extra pawn into victory in Illescas changed thereby, White comes closer to the
Gueneau, French Cht 1 99 1 . sort of endgame that he would prefer. In my ob
We now return to 4. . . dxc6 (D): servation, however, the 4:3 endgame advantage
very seldom arises in games between strong
players. In reality it is just as likely that Black's
bishops and active play will effect some struc
tural change along the way. However, barring
favourable exchanges (and it takes a lot of them
before a true endgame will come into view),
White may still be able to use his lead in devel
opment and in some cases his greater control of
territory to build up his forces and break through
in the centre before Black is ready for it. That
seems to be the more common way in which
White makes progress. Conversely, the varia
tions in which Black successfully restrains
White's central pawns or the ones in which he
develops rapidly have proven the most effective
5 0-0 in equalizing.
The first basic idea of the Exchange Variation This brings us to the difference between
is that White has the superior pawn-structure, White's more modem move 5 0-0 and the tradi
and that he will sooner or later exchange his d tional 5 d4. The latter move has a certain logic,
pawn for Black's e-pawn, establishing a 4:3 because White needs to disturb his opponent's
pawn-majority on the kingside. He hopes to game before Black can secure his position and
win a simplified position by using that majority find roles for his bishops. But after 5 d4 exd4 6
to create a passed pawn, whereas Black's 4:3 'i*'xd4 (6 lbxd4 c5 is easy for Black, because af
majority is 'crippled' and incapable of doing ter the exchange of queens, the two bishops can
RUY LOPEZ 153
develop quickly in coordination with harassing

White's king) 6 . . .'ii'xd4 7 ctJxd4 (D), the situa
tion has changed.
three that hold their own, and follow a few

games (with a number of imbedded excerpts) in
order to get a close feel for the ideas.
White's special advantage of having more M ilu - Vajda

pieces in play has disappeared. Given that cir Bucharest 1995
cumstance, we're down to the effectiveness of
the two bishops versus the potential advantages 5 i.d6
•••
of White's pawn-structure. Let's see: 7 . . . i.d7! This modest and logical development bol
(the idea is to get castled quickly, bringing the sters e5 and retains options for the knight and
rook to the open d-file, and perhaps play . . . c5 c8-bishop. It keeps the game interesting but is
and ... i.c6; 7 ... c5 is a good alternative; on the also non-forcing; thus it offers White more op
other hand 7 . . . i.d6 ! ? commits Black to a par- portunities to create trouble for his opponent
ticular development; then 8 ctJc3 ctJe7 9 0-0 0-0 than the other two moves under consideration.
1 0 f4 Ile8 1 1 lbb3 f6 1 2 f5 ! ? b6 1 3 i.f4 is the 6 d4 (D)
famous game Lasker-Capablanca, St Petersburg White should develop as rapidly as possible.
1 9 14, in which Black was probably not worse as explained above, and he also wants Black to
but he had to defend accurately and lost) 8 i.e3 play . . . exd4 to establish his 4 : 3 kingside major
0-0-0 9 lbd2 (9 lbc3 i.b4) 9 . . . lbe7 (9 . . . c5 1 0 ity. Black's bishops would find the time to de
lbe2 b 6 sets up a structure that Black normally velop smoothly after 6 d3 ctJe7; for example. 7
likes, because it is sound and makes room on c6 i.e3 0-0 8 lbbd2 (8 c3 ! ?) 8 . . . f6 9 a3 (probably
for a bishop or knight; for instance, 1 1 0-0-0 White would be better off with 9 c3 or 9 l0c4.
lbe7 1 2 ctJc4 lbc6 with equality) 1 0 0-0-0 lbg6 although in the latter case Black might cause
( 1 0 .. .f6 1 1 f3 ctJg6 1 2 h4 h5 1 3 ctJc4 c5 1 4 ctJf5 the same kind of problems by 9 . . . i.g4) 9 . . . c5 l 0
i.e6 is solid and equal, if uninspiring, Miles lbc4 i.g4 l l b4? cxb4 1 2 lbxd6 cxd6 1 3 h3
Karpov, Biel 1 992) 1 1 h3 Ile8 1 2 .l:the l i.d6 1 3 ( 13 axb4 f5 ! and the f3-knight is in trouble)
lbe2 f5 ! ? 1 4 exf5 ctJh4 1 5 ctJc4? ( 1 5 g4 ctJg2 1 6 1 3 . . . i.xf3 14 'iVxf3 bxa3 1 5 .l:tfbl b5, Ungu
Ilgl lbxe3 1 7 fxe3 .l:txe3 is to Black's advan reanu-Flear, Lenk 1 992. Black is a clear pawn
tage due to his bishops) 1 5 . . . lbxg2 1 6 Ilgl ahead.
lbxe3 17 fxe3 i.c5 with a big advantage for 6 ... exd4 7 'iVxd4!
Black, Peterson-Alekhine, brebro 1 935. 7 lbxd4 is slow: 7 . . . lbe7 (7 . . . 'iVh4 ! ?) 8 i.e3
You can see how easy Black's play is after 5 0-0 intends . . . f5, a double-edged move that is
d4 and why 5 0-0 (D), to which we now return, good in a position like this because it opens
is generally preferred. lines; e.g., 9 lbc3 f5 1 0 exf5 lbxf5 1 1 lbxf5
After 5 0-0 Black can choose among a wide i.xf5 with free and easy development. You can
array of defences, but most of them offer White see that White's kingside majority is no longer
good prospects for advantage. We'll focus on a relevant factor.
a) 8 e5 fxe5 9 ti'Jxe5 lets Black catch up in

development again : 9 ...ti'Jf6 (or 9...'iVf6 I O Ilel
ti'Je7 with equality) I O Ilel 0-0 1 1 i..g5 ( 1 1
'iVc4+ ti'Jd5 1 2 ti'Jc3 'iVf6 !) l l . ..'iVe8 1 2 ti'Jd2 c5
1 3 'iVc4+ i..e6 with full equality, Ungure-Lane,
Cappelle la Grande 1 995 . Black's pawn-struc
ture is the equal of White's. The moral for
Black is to get on with his development and
force the pace.
b) The most common move by quite a mar
gin is 8 ti'Jbd2. Black has two sound options,
both with a mind to watching over e5 :
b l ) 8 . . . ti'Jh6 ! ? 9 ti'Jc4 ti'Jf7 is a relatively old
but noteworthy idea: I 0 b3? 0-0 1 1 h3 b5 ! 1 2
7 f6 (D)
... ti'Jxd6 cxd6 1 3 i..f4 i..b 7 ! 1 4 Ilad l c5 1 5 'iVd3
An unfortunate necessity versus e5 which Ile8 1 6 .l:tfel Ile6 1 7 c4 'iVe7, Karaklajic-Gligo
puts White even further ahead in development. ric, Manila 1 975. Surrender of the centre for ac
However, if Black gets just a few moves to con tivity !
solidate by . . . ti'Je7-g6 and ... i..e6, he'll control b2) 8 ... i..e6 9 b3 (not 9 ti'Jc4? losing a pawn
e5 and stand well positionally. after 9 . . . i..x h2+! I O 'it>xh2 'iVxd4 1 1 ti'Jxd4
i..xc4 - thus 8 ... i..e6 has a preventative func
tion) 9 ... ti'Je7 I O i..b 2 ( 1 0 ti'Jc4 i..b 4! 1 1 ti'Je3 c5
1 2 'iVxd8+ .l:txd8 13 i..b2 0-0 and White hasn't
achieved what he needed to in terms of either
pawn-structure or neutralizing the black bish
ops, Schiissler-Westerinen, Copenhagen 1 979)
I0 . . . 0-0 1 1 .l:tadl (D).
We looked briefly at this position in Chap

ter 2 when discussing the vanished centre. As
explained there, static factors are temporarily
more important than dynamic ones, although
that might change at any moment. White would
like to make inroads before Black can stabilize
the position. Given time, the bishop-pair might
begin to assert itself. White's other option is to l I .. .'iVe8 !? (a frequent theme: the queen will
eliminate one of the bishops, probably the one reinforce Black's bishops by . . .'iVf7 or run off to
on d6; he can hardly stand worse in that case attack White's king; nevertheless, a good and
but the time it takes to achieve this will usually probably superior alternative was l l . .. c5 ! fol
let Black equalize or come very close to doing lowed by ... 'iVe8) 1 2 e5? ! (White should try 1 2
so. ti'Jc4! with good prospects; Black i s then under
8 i..e3 pressure to respond and e5 can wait until later)
A flexible move that develops without com 1 2 .. .fxe5 1 3 ti'Jxe5 c5 1 4 'iVe4 i..d 5. Here some
mitting the b l -knight. Alternatively: thing has definitely gone wrong for White:
RUY LOPEZ 155
Black's bishop-pair is too effective. The game ilb4 - but then it would have been better to
Lutikov-Westerinen, Jurmala 1 978 continued play the ... tt:.lg6/... 'iVe7 plan earlier) 1 4 .l:tf3 ?!
1 5 'iVg4 h5 ! ? 1 6 �3 ! tt:.lg6 1 7 tt:.lxg6 'iVxg6 1 8 f5 ! (the point; now White' s bishop will look
c4 ilc6 1 9 f3 ? ( 1 9 Iife l ! improves consider bad too) 1 5 e5 tt:.ld5 1 6 �3 ilf8 1 7 tt:.lfl g5 ! 1 8
ably, although 1 9 ... Ilad8 would have the idea fxg5 'iVxe5 1 9 ilf2 f4 with the initiative.
of ... ..tf4 with continuing pressure) 1 9 ... .l:tae8 10 0-0?!
•••
20 tt:.le4 ilxe4 2 1 fxe4 l:txfl + 22 .l:txfl Ilxe4 23 Once again it's better to keep same-coloured
'iVf3 'it>h7, and Black was not only a pawn up bishops on by 10. .. tt:.lg6 1 1 tt:.lc4 ilxc4 1 2 'iVxc4
but had the more active pieces. 'iVe7. Now we'll see how White can exploit the
8 tt:.le7 9 tt:.lbd2 .ie6
.•. structural advantage that he has so carefully
Now Black intends ... tt:.lg6, ...'iVe7, ... c5 and maintained:
. . . 0-0-0. In response, White finds a good plan to 1 1 tt:.\c4 ilb4 12 a3! 'iVxd4 13 tt:.lxd4 ilxc4
take advantage of his centralized pieces. 14 axb4
10 .l:tfdl This pawn quashes potential queenside play
A typical trick is 1 0 tt:.lc4? .ixh2+ 1 1 'it>xh2 by Black (not that the pawns were really going
�xd4 1 2 tt:.lxd4 ilxc4. anywhere). Now that White has the position he
Apart from the text-move, 10 'iVc3 ! makes a wants, he 's in no hurry.
lot of sense, preparing tt:.lc4 without allowing 14 tt:.lg6 15 f3! Ilad8 16 @f2 Ild7 17 .l:td2
.•.
the . . . ..txh2+ trick. Then Black can still play .l:tfd8 18 .l:tadl tt:.le5 19 h4! (D)
1 0. . . tt:.\g6 1 1 tt:.lc4 'iVe7, but in Webb-Hanley,
British League (4NCL) 2005/6 he tried to get
. . . 0-0-0 in faster by 1 0 ...'iVd7 ! ? 1 1 tt:.ld4 0-0-0
1 2 tt:.lxe6 'iVxe6 (D).
White begins to seize space; the f-pawn can

wait for the right moment since its advance can
weaken adjoining squares.
19.....tf7 20 b3 b6 21 tt:.le2 Ilxd2 22 .l:txd2
Only White can stand better here, although l:txd2 23 ilxd2
he will find it difficult to sustain a meaningful Minor pieces are superior to rooks when
advantage. It's interesting that in this and simi you're trying to win these characteristic Ex
lar positions, the same-coloured bishops help change positions; knights are best of all.
Black. Normally you'd think that it would be 23 @f8 24 g4! c5? !
.•.
nice for Black to have the 'good' bishop on e6, But White was eventually going to advance
but in practice you'll see that it's easier for his pawns with a winning game.
White to implement his planned expansion on 25 bxc5 bxcS 26 ile3 c4 27 b4! tt:.\c6 28 f4
the kingside under those circumstances. The (D)
game continued 1 3 f4?! (a little impulsive ; 28 'it>e8 29 c3 'it>d7 30 tt:.lg3 tt:.ld8 31 g5 tt:.lb7
...
White can always delay this and keep an edge) 32 .id4 fxg5 33 hxg5 g6 34 'it>e3 tt:.ld6 35 f5!
1 3 . . . .l:the8 ( 1 3 . . . tt:.\g6 ! ? is tactically playable - The advance is inexorable. 'it>f4 is on the
14 f5 'iVe5 ! or 1 4 e5 fxe5 1 5 f5 'iVe7 ! 1 6 fxg6 cards next.
35 gxfS 36 exfS i.dS 37 i.eS! i.f7

•.• Volokitin - Akopian
Also hopeless are 37 . . . tt:.lb5 38 tt:.le4 and Sochi 2004
37 . . . tt:.\f7 38 'it>d4 ! .
3 8 'it>d4 'it>c6 39 i.xd6 'it>xd6 40 tt:.le4+ 'it>e7 6 h3 hS!
1-0 Really, this is forced if Black wants to equal
A beautiful display of White's 'ideal' goal ize. He can't give up the bishop-pair for noth
when playing 4 i.xc6. However, I think that ing.
this is the exception and not the rule, and that 7 d3
the most important lesson of these examples as White needs to get some pieces out before he
a whole is that Black usually has the where can contemplate capturing the bishop. You can
withal either to change the pawn-structure or easily work out the consequence of doing so; at
otherwise to create counterplay in return for his the very least White will have to return the
doubled c-pawns. piece, since 7 hxg4? hxg4 8 tt:.lh2? 'iVh4 is aw
ful .
As indicated above, the way that White usu 7...'iVf6 (D)
ally gets the edge in practice is by exploiting his
development and space to create some other
type of advantage, even if it means straighten
ing out Black's pawns. White can often succeed
in doing this and give himself real chances;
whether he can achieve enough to win the game
is another matter. Also, some of Black's early
options deserve attention.
5...i.g4 (D)
This is the most radical move and the one
that was originally thought to be the most seri
ous problem with 5 0-0. Now 6 d4 loses a
pawn, and slow moves allow, for example,
. . .'iVf6 and . . . 0-0-0 and/or . . . i.c5. Thus White
prefers to attack the black bishop right away. It 8 i.e3
would take us too far afield to examine the Somehow the games keep petering out in
many tactical lines following 5 . . . i.g4 and it's this line although there are plenty of ideas:
another one of those moves that takes a lot of a) The first point is that, again, 8 hxg4?
study and memorization. I' m only going to hxg4 wins Black's piece back with advantage,
quote a few games to draw an outline of the since the knight can't move without allowing
play. ... 'iVh4.
RUY LOPEZ 157
b) 8 tt:.lbd2 has been the main line, but Black agreed in Kindermann-Dorfman, Jenbach 2003.
has done reasonably well. There are hundreds Kindermann analyses 1 8 i.. e3 g5 1 9 c3 !Ihf8 20
of games here; I'll just list a few excerpts after f3 f5 2 1 !lg 1 when 2 1 . ..c5 looks equal. Plenty
8 . . . tt:.le7: of ideas, but dubious results in terms of advan
b l ) 9 hxg4 isn't played because of 9 . . . hxg4 tages for either side.
1 0 g3 ! gxf3 1 1 'iVxf3 'iVe6 ! ? (or l l .. .'iVh6 1 2 8 ... i..xf3 9 'iVxf3 'iVxf3 10 gxf3 i.. d6 1 1
!Iel , when 1 2 . . .'iVh3, 1 2. . .tt:.lg6 and 1 2 . . . c5 are tt:.ld2 tt:.le7 (D)
all at least equal) 1 2 tt:.lc4 c5 1 3 i..e3 tt:.lc6 1 4
'iVf5 'iVxf5 1 5 exf5 f6 with equality, Deviatkin
Fressinet, Internet 2004.
b2) Some big names have been involved in
games after 9 Ilel tt:.lg6 10 d4 ( 10 hxg4? hxg4 1 1
tt:.lh2? i..c5 !) 1 0... tt:.\f4 1 1 dxe5 ( 1 1 hxg4! ? hxg4
1 2 g3 gxf3 1 3 'iVxf3 tt:.le6 1 4 dxe5 'iVh6 1 5 tt:.lb3
g5 ! 1 6 i..e3 'iVh3 ! 1 7 'iVg2 'iVh5 with equality,
Macieja-Adams, Rethymnon ECC 2003; you
see how crazy and specific this all is!) l 1 . . .'iVg6
12 tt:.lh4 i..xdl 1 3 tt:.lxg6 tt:.lxg6 1 4 !Ixd l .i:Id8
( 1 4...0-0-0 is also equal) 1 5 !Iel tt:.lxe5 1 6 tt:.lf3
tt:.lxf3+ 1 7 gxf3 i..b4 1h-1h Nisipeanu-Kasim
dzhanov, Bundesliga 2005/6.
b3) 9 tt:.lc4 (this is positionally the most in
teresting move) 9 . . . i..xf3 1 0 'iVxf3 'iVxf3 1 1 This time Black just wants to play ... c5 and
gxf3 tt:.lg6 1 2 i..e3 i..e7 ! (D). . . . tt:.lc6. These kinds of positions are equal and
don't say much for White' s winning chances
after 5 ... i..g4. On the other hand that can change
with one new discovery or reassessment.
12 .i:Ifdl
Nothing much happened in this game either:
1 2 .i:Ifb l ! ? c5 ! 1 3 @fl a5 1 4 a4 tt:.lc6 1 5 c3 f6 1 6
@e2 b6 1 7 !lg 1 @f7 1 8 tt:.lc4 !Iad8 1 9 !Ig2 h4
20 !Iag l g5 with equality, de la Villa-Delchev,
La Roda 2004.
12 c5 13 tt:.lc4 tt:.lc6 14 c3 @e7 15 @fl f6 16
•••
a3 a5 17 a4 g6 18 @e2 @e6 19 !Igl !Ihg8 20

Ilg2 !Iad8 21 !Iagl @f7 112-1/2
Thus 5 . . . i.. g4 gives every indication of being

a complete solution for Black. Here's one more
White has apparently lost his wished-for ad method of play that looks perfectly fine for
vantage in a possible pawn ending ! But that's him:
not very relevant, since the game won't get that
far in most cases. Black can reorganize by ... f6 Hector - Beliavsky
and . . . tt:.\f8-e6, or plunge ahead with ... 0-0-0 Copenhagen 2004
followed by . . .!Ihf8 and .. .f5, as suggested by
Kindermann. White usually plays for either f4 5 ...'iVf6 (D)
or d4: 1 3 !Ifd l (or 1 3 @h l i..f6 14 a4 0-0-0 1 5 This is a simple way to defend e5 . Black
a5 tt:.lh4 with equality, Hort-Spassky, Reykjavik prepares an early . . . 0-0-0. Up to this point
Ct ( l 6) 1 977) 1 3 . . . 0-0-0 14 @fl f6 ( 14... !Ihf8 ! ? White hasn' t found any way to gain the upper
intending .. .f5 is Kindermann' s idea) 1 5 @e2 hand.
tt:.\f8 1 6 f4 exf4 17 i..xf4 tt:.le6 and a draw was 6 d4!
8 ....llg4
Or 8 ... 'iVxd4 9 tt:.lxd4 .lld7 and the idea .. .f6
brings equality; or 8 . . ..ll d7.
9 'iVe5+ tt:.le7 10 .llxe7 'iVxe7 11 tt:.lbd2 0-0-0
12 'iVf4 h5 13 h3 .lle6 14 tt:.\g5 g6 15 'iVe3? .llh6
16 f4 'it>b8 17 tt:.lxe6? fxe6!
Black threatens ... e5.
18 e5 g5!
with a substantial advantage. At the time of
writing, the ball's in White's court.
Marshall Attack
Versus the slow 6 d3, 6 ... .ll g4 and 6 ... .llc5 1 e4 e5 2 tt:.\f3 tt:.lc6 3 .llb5 a6 4 .lla4 tt:.lf6 5 0-0
are good aggressive moves. .lle7 6 Ilel b5 7 .llb3 0-0
6...exd4 7 .llg5 I ' m going to emphasize and expand upon a
Or: point that I made in the Closed Lopez section
a) 7 e5 'iVg6 gives Black nice scope for his about move-orders. 7 ... d6 is a way to circum
bishops, as shown by 8 tt:.lxd4 .ll h3 9 'iVf3 .ll g4 vent the problems associated with White's
I O 'iVg3 0-0-0. 'Anti-Marshall' lines, which go 7 . . . 0-0 8 a4 and
b) 7 'iVxd4 and now 7 ....ll g4 ! ? puts White on 7 . . . 0-0 8 h3 (see below). 7 ... d6 will usually lead
the spot; e.g., 8 'iVe5 + ! ? 'iVxe5 9 tt:.lxe5 .lle 6 to the normal Closed Lopez after 8 c3 0-0 9 h3,
with equality. Black can also play 7 ... 'iVxd4 8 etc. After 7 ... d6, 8 a4 (D) is no longer very ef
tt:.lxd4 .lld7 or even 7 ... .lld7, yielding the stan fective, in part because e5 is defended.
dard type of equal position that we saw after 5
d4.
c) 7 tt:.lxd4?! .ll d7 8 .lle3 0-0-0 gives Black
everything he wants.
7 'iVd6 (D)
••.
Here Black has equalized easily following

each of the moves 8 ... .lld7, 8 . . .b4 and 8 . . ..ll b7 ;
compare 7 ... 0-0 8 a4. For those with more im
mediate ambitions, there's 8 ... tt:.\a5 ! ? 9 .lla2 (9
axb5 tt:.lxb3 10 cxb3 .llb7 1 1 bxa6 .llxa6 gives
8 'iVxd4 plenty of compensation: two bishops, activity,
8 tt:.lxd4 .ll d7 9 tt:.lc3 .lle7 IO .llxe7 tt:.lxe7 1 1 and those awful b-pawns) 9 . . .b4 I O c3 ! ? c5 1 1
tt:.lb3 is Magem-Morozevich, Pamplona 1 994/5. d4 cxd4 1 2 cxd4 ( 1 2 cxb4 works out reasonably
Easiest now is 1 1 . . .'iVxd l 1 2 .l:taxdl b6 intend well for Black in the various complications fol
ing . . . c5, denying the b3-knight good squares, lowing 1 2 . . . tt:.lc6 { 1 2 . . . tt:.lb7 ! ? } 1 3 b5 and now
followed by ... 0-0-0. 1 3 ... tt:.lb4 1 4 .llxf7+ 'it>xf7 1 5 'iVb3+ tt:.lbd5 ! ? 1 6
RUY LOPEZ 159
exd5 axb5 or 13 ... axb5 14 axb5 .ie6 15 bxc6

.ixa2) 1 2... 0-0! 1 3 tt:.lbd2 "VJJ/c 7, which is equal
according to Ivanchuk.
We now return to 7 ...0-0 (D):
tt:.lxg5 ! .ixg3 28 tt:.\f6+ 'it>h8 (or 28 . . . 'it>g7 29

tt:.le6+! ) 29 tt:.lxf7+ ! 'it>g7 30 "VJJ/xh6+ 'it>xf7 3 1
"VJJ/h7+ 'it>xf6 3 2 .ig5+! 'it>xg5 3 3 "VJJ/g7+ 'it>h5 34
.idl + .ig4 35 .ixg4+ 'it>h4 36 fxg3+ 'it>xg3 37
.if5+ 'it>f4 38 "VJJ/h6+ 'it>g3 39 "VJJ/g5#. Instead, the
8 c3 game continued 25 ... tt:.lbd7 26 .ixh6 .ixh4 27
At this juncture, White has got good mileage Ilg4 .ie7 28 .ig5 .ixg5 29 tt:.lxg5 f5 (29 ... "VJJ/b 2
out of two of the 'Anti-Marshall' variations: 30 Ilg3 Ile8 31 'it>h2 !) 30 exf5 gxh5 31 Ilg3
a) Kasparov has caused his opponents con tt:.lf6 32 tt:.le6+? (Kasparov falters in time trou
siderable difficulties with 8 a4, threatening ble - 3 2 tt:.le4+! wins; for example, 3 2. . . tt:.\g4 33
simply 9 axb5. At the moment Black's reme Ilxg4+ hxg4 34 "VJJ/xg4+ @f7 35 "VJJ/g6+ 'it>e7 36
dies are holding up well, but this version of the "VJJ/g7+ .l:tf7 37 f6+ 'it>e8 38 tt:.\xd6+! cxd6 39
Anti-Marshall is still a legitimate weapon and "VJJ/g 8+ .l:tf8 40 .ig6+ 'it>d8 41 "VJJ/xf8+ 'it>c7 42
leaves plenty of play on the board. Here's a "VJJ/e7+ tt:.ld7 43 f7) 32 ... @f7 33 Ilg7+ 'it>e8 34
classic example between the former and cur tt:.lxc7+ 'it>d8 35 tt:.le6+ 'it>e8 36 tt:.lc7+ 'it>d8 37
rent World Champions, with notes at critical tt:.le6+ 'it>e8 38 tt:.lc7+ 1h-1h Kasparov-Topalov,
junctures: 8 . . . .ib7 (arguably the best defence) Linares 2004.
9 d3 d6 10 tt:.lbd2 tt:.ld7 1 1 c3 tt:.lc5 1 2 axb5 axb5 b) The latest rage (and it again shows the re
1 3 .l:txa8 "VJJ/xa8 14 .ic2 b4 15 d4 bxc3 16 bxc3 spect that players have for the Marshall Attack)
tt:.\d7 17 tt:.lfl .if6 18 d5 (the position looks like is 8 h3, a move that leading grandmasters have
a main-line Closed Lopez and White faces a turned to with some (but not overwhelming)
similar decision with respect to the centre; success. Then 8 . . . d6 9 c3 transposes into the
both this move and 1 8 tt:.le3 have been played) Closed Ruy Lopez. And 8 . . .d5, the Marshall
l 8 . . . tt:.lcb8 ! 19 h4 tt:.lc5 20 tt:.\g3 .ic8 2 1 tt:.\g5 ! ? idea, comes up a little short after 9 exd5 tt:.\xd5
h6 22 tt:.lh5 ! .ie7 ( 22 . . . hxg5? fails to 23 tt:.lxf6+ 10 tt:.lxe5 tt:.lxe5 1 1 Ilxe5. This gives White a
gxf6 24 hxg5 fxg5 25 "VJJ/h5 f6 26 "VJJ/g 6+ 'it>h8 better grip on the kingside than he gets in the
27 'it>h2 followed by .l:t h l ; this is a position that Marshall Attack. Furthermore, White will fol
illustrates perfectly Kasparov's technique of low up with rapid development by tt:.lc3, surely
' cutting the board in two' - Black has a numer an improvement over having a pawn on c3. For
ical superiority in pieces, but four of them are all that, White shouldn't feel overconfident:
sitting helplessly on the queenside, cut off years ago, Blatny tried 1 1 . ..tt:.lb6 followed by
from defence of their king) 23 tt:.lh3 ! "VJJ/a 2! ... c5 and . . . .id6; this deserves some attention.
(23 ... .ixh4? 24 tt:.lxg7 ! 'it>xg7 25 "VJJ/h 5) 24 Ile3 Of course, Black needn't gambit; he usually
g6? (24. . . .ixh4! leads to balanced complica plays 8 ... .ib7 9 d3 (D).
tions) 25 Ilg3 (D). After this modest protection of the e-pawn,
Here 25 . . . .ixh4? allows a beautiful combi White has several methods of arranging his
nation: 26 "VJJ/d 2! g5 (26 ....ixg3? 27 tt:.\f6+) 27 pieces. For example, he can divorce himself
option, with the idea of ... .if8 and staying cen

tralized; and there are a number of other sensi
ble moves) 1 1 .ia2 c5 1 2 ctJbd2 ( 1 2 lbc3 lbc6
1 3 ctJd5 lbxd5 1 4 .ixd5 'iVc7 1 5 c3 ctJb8 1 6
.ixb7 'iVxb7 has also led to equality; this isn't
written in stone, of course) 12 ... lbc6 13 ctJfl
.ic8 ! ? (Black plays the familiar rerouting move
with his bishop even though there ' s no pawn on
d5 - his idea is to play the also-familiar . . . .ie6
and challenge White to exchange; early success
often leads to repetition and I suspect that other
moves will become more popular here) 1 4 c3
( 1 4 ctJe3 .ie6 15 .id5 ! , as in Sutovsky-Beliav
sky, Gothenburg Echt 2005, is interesting and
entirely from the Closed Ruy lines by lbc3, per perhaps even favourable for White; for the mo
haps to be followed by lbd5. He can play c3, in ment these h3/d3 set-ups are still producing
more or less traditional fashion (ctJd2-fl might some original positions) 1 4... .ie6 1 5 .ixe6 fxe6
follow). Or he has .ie3 followed by ctJd2, con (D).
templating d4. What has become the main line
goes 9 . . . d6 (even the active 9 . . . .ic5 has been
played here, reacting to White' s slow develop
ment, but Black ' s main alternative is 9 ... Ile8;
e.g., 1 0 lbc3 .ib4 ! ? 1 1 ctJg5 .l:tf8 1 2 a3 .ixc3 1 3
bxc3 lba5 1 4 .ia2 c5 1 5 f4 exf4 1 6 e5 ctJd5 1 7
.ixd5 'iVxg5 1 8 .ixb7 1/z-1/z Kramnik-Leko,
Brissago Wch (2) 2004) IO a3 (D).
It's amusing that in a very similar position

from this h3/d3 variation, the move . . . .ie6 is al
most never answered by .ixe6, whereas here
the world's best players have done so repeat
edly. As explained in the introductory chapters,
there is no rule about when to double Black's
pawns. Sometimes the lack of central mobility
after ...fxe6 is debilitating; and other times the
The minute that Black defends his e-pawn, extra squares that are covered by the e6-pawn
he is free to play . . .lba5 and rid himself of make it worth it for Black. From this point the
White's active bishop. With 1 0 a3, White gives game Topalov-Kasimdzhanov, San Luis Wch
the bishop a square to drop back to. This con 2005 continued 1 6 b4! ? (previously Kasparov
trasts with the customary set-up with c3 and had achieved nothing from 1 6 lbg3 ctJd7 1 7
.ic2. Now Black has been playing 1 O . . . lba5 .ie3 d5 1 8 exd5 exd5 1 9 a4 Ilb8 2 0 axb5 axb5
(although world-class players are playing this 2 1 b3 .l:ta8 1/z-1/z Kasparov-Topalov, Linares
way, I really wonder about the wisdom of set 2005) 1 6 ...'iVd7 ( 1 6 ... lbh5 1 7 ctJ l h2 ctJf4 1 8
ting up a Chigorin structure with . . . c5 and ced .ixf4 .l:txf4 1 9 'iVb3 'iVd7 20 a4! favoured White
ing the d5-square; the logical 1 0. . . Ile8 is a good in Adams-Kasimdzhanov, Linares 2005) 1 7
RUY LOPEZ 161
�3 .l:tfb8 18 tt:.l l h2 with an unclear position. A ( 10 tt:.\g5 tt:.la5 1 1 tt:.lxe4 tt:.lxe4 1 2 .l:txe4 i..b7 1 3
terrifically complicated game ensued. It' s hard d4 tt:.lxb3 14 axb3 �xd5 gives Black plenty of
to assess whether this relatively new approach play for a pawn: two bishops, superior develop
will prove to be a durable weapon for White. ment, and attacking chances) 10 ... exf3 1 1 d4
8 d5 (D)
... ( 1 1 g3 !?; 1 1 �xf3 i..g4 12 �g3 Ile8 13 f3
�d3 ! ? 14 fxg4 i..c 5+ 15 Ile3 Ilad8 16 tt:.la3 ? ! -
here's a good point to look for white improve
ments - 16 ... tt:.\e4 17 �f3 tt:.lxd2 1 8 i..xd2 �xd2
1 9 i..xf7+ 'it>h8 20 i..xe8 i..x e3+ 2 1 'it>hl Ilxe8
with approximate equality) 1 l . . .fxg2 1 2 �f3
( 1 2 i..g5 a5 ! ?) 12 . . . a5 1 3 i..g5 ( 1 3 a4 ! ?) 13. . .a4
14 i..c 2 b4 15 �xg2 Ila5 ! with an unclear at
tack. Who knows? 9 ... e4 is wide open to inves
tigation.
10 tt:.lxe5 tt:.lxe5 11 Ilxe5 c6 (D)
The Marshall Attack. In this horribly over

analysed (but highly instructive) variation,
Black sacrifices a pawn in return for a kingside
attack and active play.
9 exd5
Alternatives such as 9 d4 and 9 d3 are con
sidered harmless, although the former makes
good study material.
9 tt:.lxd5
•.•
The attempt to complicate by 9 ... e4 (D) has

been underestimated and might be a good alter The starting position for the main lines of the
native to the Marshall Attack proper. Marshall Attack. Essentially, the idea for Black
is to move pieces to the kingside and check
mate, whereas White wants to prevent that and
remain a pawn ahead ! It's not quite that simple,
of course. For example, Black will usually gain
some advantage in the centre as well, so that
even if White beats back the attack and remains
a pawn ahead, the game will often be drawn.
Black's initial attack is based upon exploiting
the light-square weaknesses that White will
have to create in order to fend off mate. And
White's defence will often consist of counterat
tacks that involve the sacrifice of material. At
the very least he will try to open queenside lines
with a4 and axb5, hoping for Ila6 or Ila7. Still,
the action is mainly on the kingside.
Not surprisingly, it is very risky. Analysis When I opened a book on the Marshall At
from diverse sources (see the article by Bucker tack and looked at the first paragraph of the first
in the Bibliography) includes these extremely chapter, I learned that for the 'old main line'
abbreviated lines, with suggestions: 10 dxc6 (which is still extremely popular), "the real
struggle begins around move 30" ! And in fact, . . . .tg4. He would also like to play l:th4 and
correspondence games sometimes take it a step perhaps even begin his own attack.
further, with one side playing a new move as Not to be contemplated is 1 5 tbd2? .tg4 1 6
the endgame begins ! Just to make it worse, the f3? ( 1 6 tbf3 Wkh5 1 7 'it>g2 f5 i s scary, to say the
majority of these analyses end in drawn posi least!) l 6 ... .txg3 ! and wins. The other impor
tions. In fact, this ultimate drawishness, in tant move is 1 5 .te3, to be seen in the next
combination with the tiresome theory, has dis game.
couraged numerous players from trying the 15 gS!
•••
Marshall. However, on a practical level, such Black stops 1 6 l:th4. He is able to do so be
considerations may not be relevant. At any cause of the tactic 1 6 .txg5? Wkf5 ! .
rate. even the very best players have discov 1 6 Wkfl !?
ered ways to create opportunities over the This move was discredited in the contest be
board, as in the game that follows. fore you but was of course revived later.
An incredibly beautiful game followed 1 6
Kramnik - Leko Wke2 f5 1 7 .txd5+ ( 1 7 l:te6 ! ?) 1 7 ... cxd5 1 8 l:te6
Brissago Wch (8) 2004 f4! ! 1 9 l:txd6 .tg4 20 Wkfl (D).
12 d4
The main line. Although 1 2 .txd5 is some
times played, by far the most important alterna
tive is 1 2 d3, as in our final game.
12 �d6 13 l:tel
•••
There is a good deal of theory on 1 3 l:te2

9'h4 14 g3 Wkh3, when a typical line is 1 5 tbd2
.i.f5 ( 1 5 . . . .tg4 1 6 f3 .tf5 is also played, with
good chances, although notice that the e2-rook
participates in 2nd-rank defence, which was the
main point of White's 1 3th move) 1 6 .tc2 ( 1 6
a4 l:tae8) 1 6. . ..txc2 1 7 Wkxc2 f5 1 8 c 4 Wkg4 ! .
13 Wkh4 1 4 g3 Wkh3 (D)
..•
20... Wkxfl +! 2 1 'it>xfl l:!.ae8 (does Black really

have enough for a piece here?) 22 .td2 .th3+
23 'it>gl fxg3 24 hxg3 l:!.e2 25 .te3 l:txe3 ! 26
fxe3 l:tfl + 27 'it>h2 g4! (a final twist! Black is
threatening perpetual check and there's nothing
to do about it) 28 l:txd5 1/z-1/z Ponomariov
Anand, Linares 2002. Amazing. But (sigh) an
other draw.
16 WkhS
..•
The position after 1 6 Wkfl came up again in a

different world championship encounter be
tween Anand and Svidler. Who knows what ei
ther had in store for the other? The game con
tinued 1 6 ...Wkxf l + 1 7 'it>xfl .tf5 1 8 f3 h6 1 9
A few thousand master games have reached tbd2 ! ? (a new move, at least among top players;
this position. Black has several attacking ideas, in general I am probably slighting correspon
the main ones being . . ..tg4, ... l:tae8 and .. .f5-f4. dence games, in which everything seems to
15 l:te4 have been played - at any rate, 1 9 l:te 1 is the
This is one of the more 'modem' moves (al older move) 1 9 ... .txe4 20 fxe4 (for the ex
though it's very old). First, White prevents change, White has a pawn, the bishop-pair, and
RUY LOPEZ 163
a big centre) 20 ... tbc7! 21 'it>g2 c5 22 e5 !fi.e7 25 Wkd3!!

•.•
23 tbe4 cxd4 24 cxd4 a5 25 !fi.e3 a4 26 !fi.d l White had probably calculated 25 . . . !fi.b8 26
tbd5 27 !fi.f2 l:!.ac8 28 l:!.b 1 f6 29 exf6 !fi.xf6 30 a7 !fi.xa7 27 l:!.xa7 with an excellent game. Or
tbd6 l:!.c6? ! 3 1 lbxb5 l:!.b6 32 !fi.xa4 with a nice he missed the beautiful move in the next note.
advantage, although naturally Black held on to 26 'it>f2
draw in Anand-Svidler, San Luis Wch 2005 . 26 a7 Wke3+ 27 'it>g2 !fi.xf3+! 28 tbxf3 Wke2+
11 tZ'id2 Ji.rs 18 f3! tZ'if6! 29 'it>gl tbg4 ! ! 30 a8WV+ 'it>g7 31 Wkxc6 Wkf2+
1 8 ... !fi.xe4? 1 9 fxe4 tbe3 20 Wkf3 tbg4 2 1 32 'it>hl Wkfl + 33 tbgl tbf2#.
tbfl and White is cleaning up. 26 !fi.xf3! 27 tbxf3 tbe4+ 28 'it>el tbxc3!
•••
19 l:!.el l:!.ae8 20 l:!.xe8 l:!.xe8 21 a4! Wkg6! This wins.

(D) 29 bxc3 Wkxc3+ 30 'it>f2 Wkxal 31 a7 h6! 32
h4 g4 0-1
Leko - Kasimdzhanov
Linares 2005
12 d4 !fi.d6 13 l:!.el Wkh4 14 g3 Wkh3 15 !fi.e3

(D)
22 axb5
22 tbe4 tbxe4 23 fxe4 !fi.xe4 24 !fi.xg5 ! was
suggested, and this is perhaps why Anand ex
changed queens in the game above.
22 !fi.d3 23 Wkf2?
•.•
Falling for an insidious trap. 23 Wkd l !fi.e2!

24 Wkc2 !fi.d3 25 Wkd l draws.
23...l:!.e2 24 Wkxe2 The position after 15 !fi.e3 is still a point of
This was Kramnik's point. At first it looks controversy after decades of research. Again,
extremely promising for White. Black wants to use those unsubtle ideas . . . i.g4,
24 !fi.xe2 25 bxa6 (D)
•.• .. .l:!.ae8-e6, and . . . f5-f4. Here is some utterly in
complete study material, finishing with some
up-to-date happenings.
15...!fi.g4 16 Wkd3 l:i.ae8
l 6 .. .f5 intends to blast open White's king
side; it can transpose to other lines, although
Black skips the move in our main game. In fact,
I'm jumping over all kinds of move-order is
sues as I go along. Here's one of hundreds of
games: 17 f4 ! 'it>h8 ! ? (considered best by most
analysts) 1 8 !fi.xd5 cxd5 1 9 tbd2 g5 ? ! (consis
tent, but the brute-force method comes up short,
so other moves have to be looked at here) 20
Wkfl Wkh5 21 a4 bxa4 22 fxg5 f4 23 !fi.xf4 l:i.xf4
24 gxf4 l:i.f8 25 l:!.e5 ! !fi.xe5 26 dxe5 h6 27 Wkxa6
and White is winning because Black' s position 20 l:tfeS

•.•
is so loose, Sax-Ehlvest, Skelleftea 1 989. We're getting to something more contempo

17 tbd2 l:te6 (D) rary. Another recent test was 20 ....th3 2 1 .td l
1 7 . . . f5 1 8 f4 ! g5 ! ? is the so-called ' Pawn Wkf5 22 Wke2 g6?! (22 ... c5) 23 Wkf3 Wkd3 24 .tb3
Push' variation, analysed by H.de Jongh in un l:txe3 (the usual exchange sacrifice, although
godly detail. I interpret him as concluding that sometimes White makes his first!) 25 l:txe3
Black is a bit worse in some endgame but Wkxd2 26 .txd5 cxd5 27 g4 Wkxb2 28 l:tael b4,
should draw ! lvanchuk-Grishchuk, Sochi 2005, and now 29
cxb4 ! Wkxd4 30 Wkxh3 Wkxb4 3 1 l:tdl is strong.
Is White getting the better of things in this line?
It's too early to tell.
21 .txd5 Wkxd5 22 h3 .trs 23 Wkg2 Wkxg2+
24 'it>xg2
With an extra pawn, even facing the bishop
pair, White has chances to win this position. He
came very close but only drew.
J. Polgar - Svidler
Wijk aan Zee 2005
12 l:tel .td6 13 d3
Here we have the other main system, super
ficially more modest for White but also full of
18 a4 poison.
The standard counterattack in this and most 13 Wkh4 14 g3 Wkh3 15 l:te4 (D)
..•
Marshall Attack systems. A bizarre tactic is 1 8

c4? .i.f4! ! of G.Kuzmin-Malinin, Sudak 2002,
threatening ...l:th6.
18 JihS
••
This time 1 8 ... .tf4?? loses to 1 9 i.xd5 ! .

1 9 axb5 axb5 20 Wkfl (D)
A famous game Tal-Spassky, Tbilisi Ct ( 1 )
1 965 went 20 c4 bxc4 2 1 tbxc4 .tb4 22 l:tec 1
.i.e2 23 .td l Wkxh2+ ! ? 24 'it>xh2 .txd3 25 tbe5
.ib5 26 .tb3 l:td8 27 l:ta7 f6 28 tbxc6! .txc6 29
:a6 �8 30 l:taxc6 l:txc6 3 1 l:txc6 tbxe3 32
fxe3 .i.d2 with equality.
B 15...tbf6
1 5 ... g5? 1 6 .txg5 Wkf5 was the trick in the
last game, but here the rook is protected.
16 l:th4 WVrs 11 tbd2! l:!.eS
1 7 ...Wkxd3 ?? 1 8 l:td4.
18 tbe4 tbxe4 19 l:txe4 l:txe4 20 dxe4 Wkxe4
21 .tc2
This doesn't look like much, but White has a
certain initiative and Black's weak pawn on c6
is a bother.
21. .Wke7 22 .tgS! f6! ?
.
RUY LOPEZ 165
22 ... Wkc7 ! 23 Wkd3 g6 24 l:td l .tf8 also fa 25 l:tdl l:td8 26 .te4!? (D)
vours White, but not by much. Even better is 26 .tb6 ! .tg4 27 Wkd3 .txd l
23 .te3 .te6 24 Wkf3! (D) 28 Wkxh7+ 'it>f8 29 .txd l .tc7 30 .tc5+ .td6
3 1 .te3 ! c5 (to cut off .tb3+) 32 Wkh8+ 'it>f7 33
.th5+ 'it>e6 34 Wkh7, when Black can hardly de
fend.
24 Wkd7?
.•.
24 ... l:tcS ! is an improvement. Nevertheless,

White would maintain her advantage after 25
l:te l (or 25 .te4 Wkd7 ! 26 l:td l .tg4 27 l:txd6 26 .txa2?
••.
.txf3 28 .txh7+ 'it>xh7 29 l:txd7 l:te8, but But it's already pretty bad in view of the
Black still has to fight for a draw) 25 . . . .te5 ! lines 26 ... .tg4? 27 l:txd6 and 26 ...l:tcS ! ? 27
(avoiding the trappy 25 . . .Wkf7 ! ? 26 .tc5 ! .te5 Wke2 l:td8 28 .tb6 l:te8 29 Wkd3, etc.
{ 26 ... .txc5 27 l:txe6 ! } 27 Wke4 g6 28 f4 .tf5 27 .tb6 .tb3 28 l:td4! c5 29 .txc5 Wke6 30
29 Wkxf5 gxf5 30 .tb3 .tb8 3 1 l:te7 ! Wkxb3 32 c4! 1-0
axb3 , when Black is terribly tied down) 26 The forced finish would be 30....txc4 3 1
.td4 Wkf7 27 Wke4 g6 28 .txe5 fxe5 29 Wkxe5 .tb6! (3 1 l:txc4? ! .txg3 3 2 l:td4 .txh2+) 3 1 . . .f5
.txa2 30 h4 with initiative to White, according (3 l . ..l:tc8 32 l:txc4! ) 32 l:txc4 fxe4 33 l:txe4
to Polgar. Wkd7 34 .txd8 Wkxd8 35 l:td4.
9 Ki ng's G a m bit
I e4 e5 2 f4 (D) nicely with quick development by means of the

moves tbf3, .tc4 and 0-0. In the best of worlds,
White might even get d4 and .txf4 in, estab
lishing the elementary picture of ideal piece
placement.
c) The traditionally weak f7-square (which
is guarded only by the king) is a target both
from a bishop on c4 and the rook on the newly
opened f-file.
Of course, Black has something to say about

these grandiose plans. In the King's Gambit
Declined with 2 ....tc5 we see that the a7-gl di
agonal has been ceded, making castling diffi
cult. We shall look at that in detail, because it
expresses some common ideas and illustrates
With the move 2 f4 we come to the King's the dynamic imbalance that the King's Gambit
Gambit, opening of the great romantics of the can still give rise to.
1 9th century. It is associated with wild attacks Varied problems occur after the most fre
and sacrifices of pieces, with each side focused quently played move 2 . . . exf4, called the King's
firmly upon their opponent's king. In modem Gambit Accepted. It's interesting that when he
times, however, it has become commonplace to accepts the pawn, Black's defences all seem to
describe the King's Gambit as an opening that involve one or both of two moves:
has taken on a simplifying character and leans a) The advance ... g5 . This protects the f4-
towards the endgame. Neither of these descrip pawn and claims a material advantage, with the
tions is very relevant to today's play, because additional benefit of blocking off the aforemen
most if not all of the great attacking lines have tioned f-file. The g-pawn can also advance fur
been neutralized and the early transition into ther to g4 (or be forced to advance), when it
endgames is a relatively unusual occurrence, may win time by attacking a knight on f3 and
given publicity by just a few older games involv has other possibilities including the common
ing well-known players. Although the King's idea of ... f3, disturbing White's pawn-structure
Gambit has no fixed disposition, modem play and introducing some tactical ideas if White
ers interpret it primarily in a positional manner, opens lines by gxf3 and exposes his king.
with sudden outbreaks of irrationality. b) As might be expected, ... d5 is an ideal
freeing move (as in almost all double e-pawn
Why would White play 2 f4? For a few fun openings). In particular, after White plays exd5,
damental reasons: this allows Black to place his knight on f6 with
a) It tries to exchange a flank pawn for a out being harassed by e5. It also frees the c8-
central pawn, thereby giving White a central bishop, gives the queen room, opens the often
majority. This is no small achievement, as we useful e-file, and gives Black a comfortable
see in numerous openings ranging from the square for his king's bishop on d6, protecting
Queen's Gambit to the Sicilian Defence. the gambit pawn. That's quite a bit for one
b) After either one of the moves ... exf4 or move, but naturally things don't go as smoothly
fxe5, White gains the open f-file. This dovetails as Black would have it either.
KING 'S GAMBIT 167
Now we'll look at two illustrative variations tbxe5?? 5 tbxe5 WVh4+ 6 g3 Wkxe4+ 7 Wke2
out of the many that have been thought up by Wkxhl 8 tbg6+ tbe7 9 tbxh8 and White will
both sides over the years. One is the main line win.
of the King's Gambit Declined, the other the After 3 ...d6, White has two basic options, 4
'Modem Defence' to the King's Gambit Ac tbc3 and 4 c3:
cepted.
Piece-Play
King's Gambit Declined
4 tbc3 tbf6
1 e4 e5 2 f4 i.cS (D) 4... tbc6? ! is an inaccurate move-order as it
If one wants to decline the King's Gambit, allows 5 i.b5 ! , when Black's centre is under
2 ... i.c5 has to be the most logical way, taking pressure.
over the critical g 1 -a7 diagonal and preventing 5 i.c4
White from castling. It certainly leads to com White doesn't get mated after 5 fxe5 dxe5 6
plicated and challenging play. tbxe5?! Wkd4! 7 tbd3 i.b6, but Black has a bind
A couple of other ways to forego acceptance and very quick development for the pawn; e.g. ,
are 2 . . . tbc6 3 tbf3 f5 ! ? and 2 ... Wkh4+ 3 g3 Wke7, 8 Wkf3 tbc6 9 i.e2 i.g4 1 0 Wke3 Wkd7 1 1 Wkg3
both plausible and requiring some preparation. .txe2 1 2 'it>xe2 0-0.
Note that 2 ... tbf6?! 3 fxe5 tbxe4 4 tbf3 leaves 5 tbc6 6 d3 .tg4 (D)
•••
Black's knight stranded in the middle of the White still can't castle! But Black has to
board, as well as securing a central majority. In watch out too. For instance, an unfavourable
one game Black made the best of a bad situation pawn-structure follows 6 . . . 0-0? 7 f5 ! with the
by 4... tbg5 ! 5 d4 (5 c3 tbxf3+ 6 Wkxf3; 5 i.c4 ?? idea .tg5 or in some cases g4-g5.
tbxf3+ 6 Wkxf3 Wkh4+ and ... Wkxc4) 5 ... tbxf3+ 6
Wkxf3 Wkh4+ 7 Wkf2 (to protect d4) 7 ... Wkxf2+ 8
@xf2 d6, when 9 exd6 ! i.xd6 1 0 tbc3 c6 (ver
sus tZ'ib5) 1 1 tbe4 i.c7 1 2 .td3 would have se
cured White a small but persistent edge.
However, after the main continuation 6 ....tg4,

7 f5? ! is a mistake because 7 ... tbh5 ! threatens
... tbd4, and there is hardly a good way to re
spond. White has at least two other candidates.
I'll try to present the main ideas without even
3 tZ'if3 dreaming of covering the complicated theory
Almost always played. For one thing, Black associated with this position.
was threatening 3 . . . .txgl 4 l:txgl Wkh4+ 5 g3
Wkxh2, and 3 fxe5?? Wkh4+ is a blunder of major Chigorin - Pillsbury
proportions. Hastings I 895
3... d6
This time White was threatening tbxe5, but 7 h3
3 . . . tbc6? doesn't protect the pawn due to 4 fxe5 A note on 7 tba4 follows the game.
7....txf3 8 Wkxf3 tbd4

Actually, 8 . . .exf4 9 .txf4 tbd4 seems fine,
but I'm not trying to rewrite theory.
9 Wkg3 (D)
14 <bc8 1S .tgS l:!.f8 16 Wke6+ <bb8 17 .th6

••.
l:!.e8 18 WVxeS tbd7 19 WkhS tbb6 20 .tdS a6 2 1

<bd2 tbxdS 2 2 tbxdS l:!.g8 23 g4 .tb4+ 24
tbxb4 Wkd4 2S tbc2 tbxc2 26 <bxc2 l:!.g6 27
.td2 l:!.d6 28 l:!.f3 Wka4+ 29 <bcl Wkxa2 30 .tc3
This is an infamous sacrifice. Pillsbury had l:!.c6 31 Wkxh7 bS 32 Wke7 Wkb3 33 <bd2 as 34
previously declared the move unsound! Since l:!.fS <bb7 3S l:!.cS l:!.aa6 36 gS files 37 WVxcS
this classic game has been critiqued many times, l:!.c6 38 WkdS Wka4 39 g6 b4 40 g7 bxc3+ 41
ru just add a note or two relevant to the open bxc3 Wkal 42 g8Wk Wkxc3+ 43 <be2 Wkc2+ 44
ing: <bf3 Wkdl+ 4S <bg3 Wkgl+ 46 <bh4 Wkf2+ 47
9 tbxc2+! ?
.•• <bhS Wkf3+ 48 Wkg4 Wkf6 49 WkgfS Wkh6+ SO
The alternative 9. . .0-0! ? i s totally unclear: <bg4 Wkg7+ Sl WkgS 1-0
10 fxe5 ( 1 0 <bd l exf4 1 1 .txf4 tbh5 1 2 Wkg5
tbxf4 1 3 Wkxf4 c6, and with White's king in the Eliminating the bishop by 7 tba4 is also
centre one would rather play Black) l 0... dxe5 very complicated. Generally Black retreats his
1 1 <bd l ( 1 1 .tb3 Wkd6 1 2 1:!.fl c6 1 3 .tg5 tbd7 bishop to b6 but there seems to be another pos
14 0-0-0 a5 gives Black the attack, according to sible formation: 7 . . . 0-0 8 tbxc5 dxc5 (D).
Renet) l l .. .Wkd6 1 2 1:!.fl c6 1 3 a4; perhaps dy
namic equality is the fairest assessment in this
situation.
10 <bdl tbxal
It seems to me that l 0... tbh5 1 1 Wkf3 tbxal
12 Wkxh5 Wkd7 intending ...0-0-0, or simply
12 . . . 0-0 may cast doubt upon the whole idea. It
probably isn't so easy.
11 Wkxg7 (D)
1 1 <bd7!
..•
l 1 . . .l:tf8 1 2 fxe5 dxe5 1 3 j_g5 j_e7 1 4 l:!.fl

looks like a winning attack in view of the long
line 14 . . .Wkd4 1 5 .txf6 0-0-0 16 Wkg4+ ( 1 6
.txe7? Wkxc4) 1 6 . . .<bb8 1 7 .txe7 Wkxc4 1 8
'it>c l ! , which I ' ll truncate at this point.
12 fxeS dxeS 13 l:tfl ! .te7 14 Wkxf7? ! Here's a structure that you' ll see in the
According to the analysts, 14 .tg5 ! was win Giuoco Piano and the Vienna Game as well.
ning. That's enough for the opening, so I'll just Black has good control of the centre; his dou
let you enjoy the rest of this titanic struggle un bled c-pawns secure d4 and open up the d-file.
perturbed: Black can even get rid of White's c4-bishop in
KING 'S GAMBIT 169
most cases if he needs to. In return, White tem stuck) 8 i.b5 tbge7 9 tbc3 f5 1 0 h3 ( 1 0 e5 ?
porarily has the two bishops, a good pawn dxe5 1 1 fxe5 0-0 and the e5-pawn hangs)
structure and prospects of a kingside attack. It' s 1 0 . . . i.xf3 1 1 'i!Vxf3 fxe4 1 2 'i!Vxe4 d5. Black
probably about even, but certainly worth a look has achieved the central dissolution that he
by both sides. Renet offers the following line, was aiming for and the fight is just starting.
full of many options: 9 0-0 tbh5 1 0 h3 i.xf3 1 1 This is all analysis.
'iixf3 tbxf4 1 2 i.xf4 tbd4 ! ? 1 3 �5 ! ? ( 1 3 'i!Vdl 5 fxe5
exf4 14 lhf4 b5 15 i.b3 tbxb3 1 6 axb3 'i!Vd4+ 5 d4 exd4 6 cxd4 i.b6 7 e5 (7 i.d3 i.g4;
1 7 'it>hl 'i!Vxb2; the pawn is real) 1 3 . . . exf4 1 4 Black has to work fast to compromise White's
l:hf4 g 6 l 5 l:tg4 ! ? b5 l 6 l:txg6+ hxg6 1 7 'i!Vxg6+ centre or it will dominate the position) 7 ... dxe5
with perpetual check. This could be a fascinat 8 fxe5 tbd5 9 i.c4 i.e6 (or 9 . . . tbc6 ! ?) with
ing variation to look into. equality; White's space is balanced by Black's
outpost.
Central Expansion 5 dxe5 (D)
.•.
4 c3
White simply goes for d4. This is instructive,
as it illustrates themes of the ideal centre.
4 tbf6
•••
The whole game revolves around whether

White's centre can be compromised. Because
of this the alternative 4 ... i.b6 ! ? would be in
triguing. The idea is to make a sort of prophy
lactic semi-waiting move, because d4 won't
come with a tempo on the bishop: 5 d4 (White
still has to get castled, so this is necessary; 5
i.c4 tbc6 doesn't help) 5 ... exd4 6 cxd4 i.g4!
(D).
6 d4
6 tbxe5 'i!Ve7 ! ? 7 d4 i.d6 recovers the pawn
and allows Black to work against an isolated e
pawn in return for the bishop-pair: 8 i.c4 i.xe5
9 dxe5 'i!Vxe5 1 0 0-0 tbc6 is equal (Black threat
ens ... 'i!Vc5+).
6 exd4 7 cxd4 i.b4+ 8 i.d2 i.xd2+ 9
..•
lbbxd2 0-0 10 i.d3

White has maintained his centre up to this
point, but it gets attacked right away:
10 ... tbc6 1 1 d5 tZ'ib4 12 Ji.bl
1 2 i.e2 l:te8 1 3 a3 tba6 leaves e4 weak, and
Black can be happy with his position.
12 ... c6 13 a3 tZ'ibxdS! ? 14 exdS l:te8+ 15
Black' s scheme reminds one of the Modern 'it>fl tbxd5 .
Defence in which the bishops precede the with an exciting and unclear attack.
knights in attacking the same e4/d4/f4 centre
( 1 e4 g6 2 d4 i.g7 3 tbc3 d6 4 f4 c6 5 tbf3 King's Gambit Accepted
.tg4, etc.). 7 i.e3 tbc6 (or 7 . . . d5 ! ? 8 e5 tbe7
intending . . . tbf5, a fascinating position with 1 e4 e5 2 f4 exf4
double-edged chances; Black's pieces will be Black takes up the challenge and plays a
well-placed but the b6-bishop could end up move that has been studied for over 1 50 years.
1 70 MASTERING THE CHESS OPENINGS
At this point we have an important alterna

tive, as well as our usual move-order discus
sion. It begins with the move 2 . . . d5 ! (or ' ? ' if, as
Black, you would like to get to the Kieseritzky
Gambit below) 3 exd5, and now 3 ... e4 (D) is the
Falkbeer Counter-Gambit.
lt:Jc3 i.a6 keeps White from castling and em

ploys the two bishops effectively; for example,
8 d3 Wkxf6 9 li:Jxd5 Wke6+ 1 0 Wke2 <:J;;d7 ! and ev
erything is covered: l l li:Jxf4 Wkxe2+ 1 2 lt:Jgxe2
lt:Jc6. Black has some advantage because for
one thing he can reorganize by . . . b6 and . . . i.b7
with powerful bishops.
Black's idea in the Falkbeer is to cramp b) 4 lt:Jc3 c6!. In this position White has
White's development. White's extra pawn on d5 tried nearly every move, but after . . . d5, he loses
is not impressively-placed anyway. The prob a key tempo. Later, when White plays d4 and
lem is that the powerful pawn on e4 has trouble captures with a piece on f4, he is left with a seri
staying there after 4 d3 li:Jf6 5 dxe4 lt:Jxe4 6 ous internal weakness on e3. You can check the
li:Jf3, the old main line going 6 ... i.c5 7 Wke2 i.f5 theory (critically, please ! ), but I don't believe
8 li:Jc3 Wke7 9 i.e3 ! , when Black has never found that White ever gets full equality.
a route to complete equality. The main idea is 3 d5
...
9 . . . i.xe3 10 Wkxe3 lt:Jxc3 1 1 Wkxe7+ <:J;;xe7 1 2 The 'modem' way of treating the King's
bxc3 i.xc2 1 3 <:J;;d2. This position has been ana Gambit. But in fact, most contemporary players
lysed for some years and seems to favour White. use 3 . . . g5 (D), the venerable Kieseritzky De
But after 3 exd5, Black can also play 3 . . . exf4 ! fence, to try to refute the King's Gambit.
4 li:Jf3 li:Jf6, when we have transposed into the
Modem Line that Black may be hoping for (it is
the variation analysed in this section). So 2 . . . d5
might be reasonable after all. Notice that this
order avoids 1 e4 e5 2 f4 exf4 3 i.c4 in the next
note.
3 li:Jf3
There must be at least a thousand master
games with 3 i.c4 (D), the Bishop's Gambit.
It has been subject to lengthy analyses for
well over a century. The old main line was
3 ... Wkh4+ 4 <:J;;f l , which offers White intriguing
attacking chances beginning with 5 li:Jf3, and
Black also enjoyed some brilliant attacks on
White's vulnerable king. But 3 ... lt:Jf6 is a big
problem for White : Everything gets very tactical and we' ll pri
a) 4 e5 d5 ! is our familiar device in e-pawn marily concern ourselves with 3 . . . d5, but I'll
openings. Then 5 i.b5+ i.d7 ! 6 exf6 i.xb5 7 mention two noteworthy continuations after
KING 'S GAMBIT 1 71
3 . . . g5 (again, a variation with thousands of queenside (even after the forward d-pawn dis
games to its credit) : appears). What's more, White ' s move 5 .tb5+
a) 4 .tc4 g4 5 0-0 gxf3 6 Wkxf3 is the time could further weaken Black' s pawns while rid
honoured Muzio Gambit, in which White sac ding himself of his only weak pawn. His natural
rifices a whole piece for a dangerous attack plan will be to plunge forward with d4 and c4,
against Black's exposed king. A line subject to securing free development. At the same time
much analysis goes 6 ... Wkf6 7 e5 Wkxe5 8 .txf7+ Black' s majority on the kingside is crippled,
'it>xf7 9 d4 with the idea 9 . . .Wkxd4+ 1 0 .te3 (D). and his f4-pawn is subject to attack along an
open file. He has no prospects of creating a
passed pawn on that side of the board.
But Black has one major advantage. White
will have to (and want to) move his d-pawn at
some point, but this creates an internal weakness
on e3. If Black manages to keep his f-pawn, he
can use that square to threaten White's position
by, for example, ... l:te8 and ... tbg4. Even if
White manages to win the f-pawn by .txf4, the
exchange of that bishop only worsens the situa
tion with respect to e3. Furthermore, White's
only real chance for advantage (or even equal
ity) is to advance his pawn to d4, since d3 ren
ders his game too passive. The problem then is
that the e4-square also becomes a weakness,
I show this merely to indicate how the old making moves like ... .tf5 and . . .tbe4 particu
style King's Gambit was played. In a book pur larly attractive. It's anybody 's guess which
porting to promote general understanding of side's advantages will be more important than
openings, this picture of anarchy has to be re the other's. Let's look at a game with sample
ferred to the specialists! lines in the notes:
b) A fairly important line seems to be 4 h4
g4 5 tbe5 tbf6 6 d4 d6 7 tbd3 tbxe4 8 .txf4 M . Ginzburg - Zarnicki
Wke7 9 .te2 tbc6 1 0 c3 .tf5 which hovers be Villa Martelli 2002
tween equal and somewhat better for Black.
4 exd5 tbf6 (D) s .tbs+
This is the only continuation that really tests
both sides. The others show why White should
be in a bit of a hurry:
a) 5 .tc4 tbxd5 6 .txd5 (or 6 0-0 .te6)
6 . . . Wkxd5 7 tbc3 Wke6+ 8 'it>f2 Wkb6+ 9 d4 .te6
is probably already better for Black, Fedorov
Godena, Batumi Ech 2002. For years, Fedorov
was the leading King's Gambit player among
grandmasters.
b) 5 c4? ! leads to typical developmental and
positional problems after 5 . . .c6! 6 dxc6 (6 d4
cxd5 7 .txf4 .tb4+ and White's interior central
squares are vulnerable; e.g., 8 tZ'ibd2 0-0 9 .te2
dxc4 I 0 0-0 b5 1 1 .tg5 { versus ... tbd5-e3 }
1 1 . . . .tb7 and Black can be happy) 6 ... tbxc6
One of the main lines of the King's Gambit. (D).
Structurally it looks good for White, at least at We see this sort of position in several open
first glance. He has a majority in the centre and ings. With Black's development and control of
It's even material at the moment. White is

banking upon his potentially powerful central
pawns (the one on d4 is passed). Black has
thwarted the development of White's c l -bishop
and has weaknesses on e4 and e3 to exploit.
8 0-0
8 �e2+ .ie6! 9 tt::lg5 0-0 ! . Black sacrifices a
pawn, but look at his terrific development after
1 0 tt::lxe6 fxe6 1 1 i.xc6 bxc6 1 2 0-0 ( 1 2 �xe6+
'it>h8 1 3 0-0 .l:te8 14 �h3 �b6 with an attack)
I 2 . . . i.c7 ! ? 1 3 c3 tt::ld5 with a great game. Kauf
man offers 1 4 �xe6+ 'it>h8 1 5 �xc6 .l:tf6 ! 1 6
�c5 f3 ! and Black's attack i s almost decisive
already.
d4. he must be better. Moves like ... i.c5, ...0-0, 8 0-0 9 tt::lbd2 i.g4! 10 c3 l':te8 (D)
•.•
. . ..l:te8 and . . . i.g4 are too strong, but if White

fights for the centre by 7 d4 (7 tt::lc 3 i.c5), he
runs into 7 ... i.b4+ 8 tt::lc 3 i.g4 9 i.xf4 i.xf3 10
'6'xf3 tt::lxd4 1 1 �e3+ tt::le 6 with a substantial
edge.
c) 5 tt::lc 3 tt::lxd5 6 tt::lxd5 �xd5 7 d4 i.e7 is
very easy for Black. Notice White's troubles
with his interior weaknesses: 8 c4 (probably
not best) 8 ... �e4+ 9 i.e2 tt::lc 6 l 0 0-0 i.g4 1 1
i.d3 i.xf3 1 2 i.xe4 i.xdl 1 3 .l:txdl g5, re
maining a pawn ahead.
5 c6 6 dxc6 tt::lxc6 7 d4 i.d6 (D)
•••
11 tt::lc4 i.c7 12 i.d2

If White is reduced to this, he's in trouble.
12 �d5 13 tt::la3 tt::le4
•••
Black stands better. The weaknesses on the

e-file are hurting White.
The 'Modem Defence' to the King's Gambit

looks appealing for Black. Of course, there's al
ways more to the story. It would be nice if
White could figure out a way to bypass such
technicalities and return to gambit play in the
romantic spirit.
1 0 I ntrod uction to the Sem i - O pen
Ga mes
The Semi-Open Games are a disparate group of between apparently similar openings before
openings with few characteristics in common turning to the practical material. First, it should
except that they immediately unbalance the play. be clear that the Caro-Kann would be a better
Another unifying factor is that they all prepare defence than the French if the disadvantages
to counter in some manner White's plan to cre mentioned above were their only problem. Af
ate a classic pawn-centre with 2 d4. It has been ter all, bringing a bishop out freely as Black
said that each of the openings under the 'Semi does in the Caro-Kann contrasts dramatically
Open Games' rubric has to 'give something up' with Black' s imprisoned bishop on c8 in the
in order to fulfil its mission. The French De French. It overshadows any other developmen
fence ( 1 e4 e6 2 d4 d5), for example, blocks in tal problem. In the Caro-Kann, limiting the op
the c8-bishop. The Caro-Kann Defence ( 1 e4 tions of the b8-knight (i.e., preventing it from
c6 2 d4 d5) takes c6 away from Black's knight. occupying c6) doesn' t seem that severe a pen
The Alekhine Defence ( 1 e4 tt::lf6) loses a tempo alty. In the abstract, a hypothetical knight might
to 2 e5 and fails to contest the centre. The Pirc be best off on c6; but in this particular opening
Defence ( 1 e4 d6 2 d4 tt::lf6) gives White an that piece will generally be happy on d7, con
ideal centre, and the Sicilian Defence ( 1 e4 c5) trolling e5 and defending f6 in key situations.
doesn't open lines along which Black's pieces And in the main line with . . .i.f5 the knight
can develop. doesn't even temporarily block Black's light
None of this applies to l ... e5, so one could ar squared bishop from developing. Of course, the
gue that in some sense that is the 'best' defence availability of c6 for a knight in the French De
to 1 e4. But l . ..e5 makes its own concession in fence shouldn' t be underestimated, not only be
that Black's e-pawn becomes an unprotected cause a knight there attacks d4 and e5, but
target of attack. Furthermore, if we look at the because d7 is left free for a bishop or for a
other Semi-Open Games listed above, all of knight retreating to d7 after White plays e5 (a
them except one attack White's centre pawn at major sequence in the French). So the trade-offs
e4, as can be seen from 1 e4 e6 2 d4 d5, 1 e4 c6 between the two openings aren't completely
2 d4 d5, 1 e4 tt::lf6, 1 e4 d5 and 1 e4 d6 2 d4 tt::lf6. one-sided, but if forced to compare, you feel
In the case of the Alekhine ( l e4 tt::lf6) and the that Black gets the better bargain by playing the
Scandinavian ( 1 e4 d5), the fact that White will Caro-Kann. However, what evens the scales is
effectively gain a tempo by 2 e5 and 2 exd5 , re the respective central situations of the two open
spectively, gives the counterattack an ambigu ings. Regardless of whether you play . . . dxe4 (as
ous character, but Black nevertheless creates an in the main lines of the Caro-Kann) or maintain
imbalance that he fails to bring about by play your pawn on d5 (as is the case in most varia
ing 1 . . .e5. The Sicilian Defence goes its own tions of the French), it' s unlikely that you'll be
way, as usual, neither developing nor attacking. able to attack White' s d4-pawn by means of
What an irony that it's the favourite move of the . . . e5 ; White can put a knight on f3, a bishop on
best players in the world! f4, a queen or rook on the e-file, etc. So the re
Since the chapter introductions cover the ba maining way to attack the centre and free one's
sics strategies (and the games much more so), I pieces is ... c5. Black plays that move in most
won't repeat what's said there. Still, it might be variations of the French Defence, truly threat
interesting to make a few general comparisons ening to liquidate White' s centre. But in the
Caro-Kann Defence, playing . . . c5 would cost Pirc player' s dream! Well, of course it's not at
Black a full tempo (i.e . . . . c6-c5). For that rea all clear whether Black can undermine White' s
son, the Caro-Kann defender will generally e-pawn Pirc-style. But the point is that h e has an
delay or forego . . . c5 and count upon the re extended centre to attack whereas in the Pirc
straining influence of his pawn on c6. That is Black is waiting for that opportunity. Again,
the right decision ( ...c5 on an early move is usu there are advantages and disadvantages to each
ally impractical anyway); nevertheless, it' s al approach. Most strong players would probably
most always better to break up the opponent' s worry more about their space disadvantage were
centre than to surrender the centre yourself. In they to play one of these openings.
the Caro-Kann, Black does gain counterplay Black's side of the Scandinavian variation 1
against White' s centre along the open d-file, but e4 d5 2 exd5 °iVxd5 with 3 tt::lc 3 °iVa5 might be
that is relatively easy to fend off. So both open compared with the Centre Game for White af
ings have their appealing and unappealing sides. ter 1 e4 e5 2 d4 exd4 3 °iVxd4 tt::lc 6 4 'iYa4.
The Alekhine Defence makes a funny con White has the extra move e4 in, but as with
trast with the Pirc Defence. In the Pirc, Black most reversed openings you have to decide
plays ... d6 first, to restrain White's centre, and whether that move is good or bad. The e4-
then ...tt::lf6 to attack it. In the Alekhine, Black pawn can be a target down an open file, fol
reverses this order, playing ... tt::lf6 first, losing a lowing, for instance, . . . tt::lf6, . . . i.c5 (or . . . g6
tempo, and then within a few moves after e5 he and ... i.g7), ... 0-0 and ... .l:te8. On the other
plays ... d6. It's as though Black had allowed hand, the e4-pawn does guard d5 in classical
White to play e5 successfully against the Pirc, an fashion. Perhaps this one's a toss-up as well.
advance that is Black's top priority to prevent! One can make this kind of comparison be
Thus one's first instinct is that the Pirc is a supe tween any number of positions in opening the
rior opening. It may or may not be, but the flaw ory, and it's a useful exercise to do so. The
in this argument can be stated more or less as reader might want to think about other funda
follows: in the Pirc, White often shouldn 't play mental properties of Semi-Open Games and
e5 because Black will either capture once and how they offer advantages or disadvantages.
then retreat the knight, or retreat without cap You will find that every defence has a balance
ture, in both cases undermining the centre by between negatives and positives whose sum
... c5 or if appropriate ... f6. In the case of the can't be too different from that of other open
Alekhine, Black has got White to commit his ings. Otherwise, some defences wouldn't be
pawn to e5 from where he is already in a position played at all, and others would find no willing
to undermine it. So in a way, he has achieved the opponents !
1 1 S ici l i a n Defence
l e4 c5 (D) Sicilian Defence played by masters, the Naj

dorf Variation:
2 tt::lf3 d6 3 d4 cxd4 4 tt::lxd4 tt::lf6 5 tt::lc3 a6
(D)
The Sicilian Defence has been the most

popular opening in top-level chess for the past
several decades and continues to be so today,
accounting for about 1 7% of all contests be What's going on? Can Black really get away
tween grandmasters, and an astonishing 25% of with this 4th pawn move, when it's not even a
the games in a database of Informators. Since centre pawn? Let's make some more sample
young players and aspiring masters show such moves:
enthusiasm for the Sicilian, it's hard to see 6 i.gS e6 7 f4 i.e7 8 °iVf3 'iYc7 9 0-0-0 tt::lbd7
those figures diminishing much. 10 i.d3 bS 11 l':thel (D)
What's so special about this opening? First
of all, 1 . . .c5 effectively prevents 2 d4, the pri
mary goal of a defence to 1 e4. To be more spe
cific, 2 d4 cxd4 3 'iVxd4 loses a tempo and
already risks disadvantage after 3 . . . tt::lc 6. If, in
stead, White plays 3 c3 and sacrifices a pawn
(the Morra Gambit), we have many years of
experience and analysis to show that Black at
the very least should have no problems equal
izing and almost certainly should gain an ad
vantage with accurate play. Of course other
openings also discourage d4 or prepare to meet
it effectively, so we have to look for more rea
sons to choose specifically the Sicilian De
fence. Since the vast majority of games are
contested in the Open Sicilian, i.e. 2 tt::lf3 and 3 White has all seven pieces developed, Black
d4, let's see what we can learn from the result just four, having made six pawn moves thus far.
ing positions. We need a concrete example to Furthermore, Black has only one piece beyond
think about, so let's start with the most popular the second rank; White has five. And of course
White commands more space. Before comment to developing by . . .b6 and . . . i.b7, thereby put
ing upon all this, let' s compare it with the Sozin ting another pawn on the third rank.
Variation of the Najdorf. After 1 e4 c5 2 tt::lf3 d6 Any experienced player knows that Black
3 d4 cxd4 4 tt::lxd4 tt::lf6 5 tt::lc 3 a6, White contin stands reasonably well in these positions. Again,
ues 6 i.c4 e6 7 i.b3 i.e7 8 0-0 b5 9 i.e3. In can we say why? The real key to the Open Sicil
that case, White has five pieces out to Black' s ian is that Black has a central majority. A cen
two. tral majority is a basic positional advantage that
A traditional Scheveningen/Najdorf line goes should never be underestimated and can com
6 i.e3 e6 7 i.e2 i.e7 8 0-0 (D). In this case, pensate for other problems in the position. If we
White has five pieces out to Black's two, and consider the centre with . . . d6 and ... e6, Black's
out of seven moves Black has played five with d- and e-pawns protect against incursions by
pawns. white pieces, thus giving Black time to catch up
in development. Next, every central majority
threatens to advance and this one is no differ
ent: once Black's pieces begin to get developed,
the move ... d5 will expand the scope of some of
them (for example, a bishop on e7, queen on c7
and a rook on d8 or e8), and create good posts
for others (e.g., a knight on e4 or d5). Because
White has to be constantly on the lookout for
this move (as well as ... e5 followed by ... d5 on
the next move) he has to devote forces to its pre
vention. That brings us to another important ad
vantage for Black in almost all Open Sicilians:
White's e-pawn is a target. It can be attacked by
a knight on f6, a bishop on b7, and perhaps an
other knight on c5. White can defend his e-pawn
Just to drive the point home, Black some with his light-squared bishop, but where should
times plays an even more extreme version of he put it? If the bishop goes to d3, it is blocked
these ideas, namely, 1 e4 c5 2 tt::lf3 e6 3 d4 cxd4 by its own pawn, and if it goes to g2 or f3, then
4 tt::lxd4 a6 5 tt::lc 3 d6 6 i.e3 b5, which adds up the bishop will also be passively defending. In
to six straight pawn moves and not even a fact, in both of these positions Black may at
piece out! You should notice another negative some point be able to play ... e5 and fix the e
aspect of every one of these lines, as if they pawn, preventing that bishop from getting out.
need it: Black's centre pawns on e6 and d6 are What does this all translate to? White's light
in passive positions blocking his own pieces, as squared bishop is by definition a bad bishop, be
third-rank pawns are known to do. Thus his cause his centre pawn is on a light square. I once
pieces have few prospects of being as active as heard the great Larsen say that after 3 d4 cxd4
White's. As we shall see, the same thing is true White was positionally lost! Tongue-in-cheek or
of most other Sicilian systems. not, he was undoubtedly referring to Black's
Back to our question: why then would any central majority, and the diversion of White's re
one, much less the world's elite, play the Sicil sources to the defence of his e-pawn.
ian Defence? Well, Black has an open c-file. What can White do in the face of these
But wait! White has an open d-file attacking a problems? He generally doesn't want to wait
weak pawn on d6. That should be even more around for an ending without changing the
effective. How about Black's minority attack pawn-structure, lest Black's central majority
with the pawn advance . . .b5? OK, at least that's and queenside minority attack become too in
a real plus (unfortunately at the cost of more fluential in that stage of play. In order to make
time taken away from development); but if he progress, White has to exploit his space advan
wants to, White has the time to stop that move tage (he almost always controls four ranks to
by playing a4. In many cases this reduces Black Black's three, with the other disputed). Hence
SICILIAN DEFENCE 1 77
you will see that many encounters feature White has plenty of options, but one is to oc
White' s advance e5, activating his bishop and cupy that square immediately. Then after 10
other pieces to gain serious and sometimes un tt::ld5 i.xd5 ! ? ( 1 0. . . tt::lxe4? 1 1 i.b6 and 12 lbc7)
stoppable attacking chances. A potential prob 1 1 exd5, White has two bishops but has lost his
lem in that case is that the e5-pawn will become outpost. More importantly, Black now has a
weak, so this decision has to be undertaken cau mobile central pawn-mass. One encounter pro
tiously. He can also turn to the advance f4-f5, ceeded 1 1 . ..tt::lbd7 1 2 c4 a5 1 3 tt::ld2 tt::le8 14 f3
hoping to force ...e5; sometimes, however, that i.g5 1 5 i.f2 f5 and Black' s 4:3 kingside major
cedes the e5-square to Black' s pieces - as al ity (a variant of the one we see in so many open
ways, proper timing is the key. Another attack ings) establishes itself. Black appears to have
ing option at White' s disposal is g4-g5, perhaps achieved a game with equal chances. Note that
in conjunction with h4, risking kingside expo White' s bishop on e2 is still bad. Obviously
sure in order to drive back Black's pieces. That one of the most complicated openings in chess
has been an increasingly popular and successful can't be boiled down to a couple of generalities,
strategy over the last decade. Finally, in addition but such themes will appear along with a multi
to all those ideas, White can try to take direct ad tude of others that directly or indirectly stem
vantage of Black's slow development and re from the basic properties of the opening.
fined pawn-play to sacrifice material and blow Before moving on to concrete variations, let
open the enemy position. His knights on c3 and me refer again to the very abbreviated descrip
d4 may be restricted by Black's pawns, but those tion that I gave in the introductory chapters re
same knights are habitually sacrificed on the garding the evolution of the Sicilian Defence.
squares d5, f5, b5 and e6. I'll expand upon it in certain particulars, but the
A completely different structure arises when point is the same. In the second half of the 1 9th
Black has a pawn on d6 and one on e5. That century, players met the Sicilian with 2 tt::lc 3
would seem worse than the . . . d6/. . . e6 systems, more than any individual variation (2 f4 was
because he gives up the d5-square to White's also a big favourite). When White played a line
pieces (and the f5-square can be handy for a of the Open Sicilian (i.e., 2 tt::lf3 and 3 d4),
knight). Let's see the best-known early version Black responded primarily with the Pin Varia
of this structure: 1 e4 c5 2 tt::lf3 d6 3 d4 cxd4 4 tion (2 tt::lf3 e6 3 d4 cxd4 4 tt::lxd4 tt::lf6 5 tt::lc 3
tt::lxd4 tt::lf6 5 tt::lc 3 a6 6 i.e2 (6 i.e3 can also be i.b4) or the similar Four Knights Variation (2
answered by 6 . . . e5, but that's another story) tt::lf3 e6 3 d4 cxd4 4 tt::lxd4 tt::lf6 5 tt::lc 3 tt::lc6), as
6 . . . e5 7 tt::lb 3 i.e7 8 0-0 0-0 9 i.e3. Both sides suming that White let him get that far. Notice
have played natural-looking moves and now that both of those variations have the primary
Black illustrates a basic idea: 9 . . . i.e6 (D) . goal of rapid development, and bear little resem
blance to the modem set-ups . . .d6/ . . . e6/ . . . a6 or
... d6/ . . . e5, with pieces generally constrained to
the second and third ranks. A few players ex
perimented with such systems, such as Louis
Paulsen. Among top players, he had to be the
most devoted Sicilian player of his time and his
games included everything from the Schev
eningen to ... the Paulsen! The latter variation is
truly hypermodem: 1 e4 c5 2 tt::lf3 e6 3 d4 cxd4
4 tt::lxd4 a6.
In the first part of the 20th century, leading
players began to investigate more Sicilian De
fences with limited success, and in particular we
see more Open Sicilians, including the Dragon
Variation and to a limited extent the Schev
This standard move contains the notion of eningen Variation. But the Sicilian only began
playing ... d5 soon, totally freeing Black's game. to be truly accepted as a leading defence in the
1 930s, and it took off in the 1 940s. The popu

larity of 1 . . .c5 hasn't stopped growing since,
turning ever more modern as it evolved. V aria
tions featuring fast development for Black are
now proportionally rare.
1 e4 c5 2 tt::lf3 (D)
.tb5+. It gets the bishop out of the way to

speed up White' s development, especially cas
tling. If the bishop is exchanged, that may be of
benefit to White because, as we have seen, his
light-squared bishop can be a problem in the Si
cilian Defence; formally speaking, it' s a bad
bishop whether or not White follows up with c4
(a common Moscow Variation theme). This is
By bringing out his knight White contests all true and indeed 3 .tb5+ can hardly be a bad
d4, but he also prevents . . . e5 , a move that move. Certain specialists have done well with it
might come in handy for Black. 2 tt::lf 3 intro at the highest levels. Yet the large majority of
duces White's Open Sicilian variations, those players prefer to use the Open variations with 3
in which he plays 3 d4, and after Black's re d4. The achievement of positive prospects is
sponse 3 . . . cxd4, recaptures with 4 tt::lxd4. These the main reason behind their decision. In the
variations constitute about 90% of master games Open variations of the Sicilian Defence, White
with 1 . . .c5. I shall group them according to tends to get a healthy lead in development and
Black's second move in so far as it is useful to space. But after 3 .tb5+, an exchange of this
do so. bishop on d7 will bring out another of Black' s
pieces and let him begin to catch up in devel
opment. If Black plays 3 . . . tt::lc 6, White' s only
I ntroduction to Systems threat is to cede his bishop-pair. That said, a
number of Black' s set-ups offer White good
with 2 d6 . . .
chances for advantage, so the defender should
know his theory and/or be a good intuitive
1 e4 c5 2 tt::lf3 d6 3 d4 player. We won't analyse the Moscow in depth
As so often, the less frequently played varia because there is so much to explore elsewhere.
tions say something about the main lines, i.e. In the broadest possible terms, and glossing
why they are the main lines. Here are some al over many options, the most frequently-played
ternatives to 3 d4 and a few move-order issues lines and ideas are as follows:
to think about. a l ) 3 ... .td7 4 .txd7+ �xd7 (after 4. . . tt::lxd7,
a) 3 .tb5+ (D) is called the Moscow Varia 5 d4 gives White some useful space, or he can
tion. It has its followers, in part because some play 5 c4, although neither course guarantees
players don't want to enter into all the compli an advantage) 5 c4 (D).
cations that arise from the Dragon, Najdorf and The idea is to set up a Maroczy Bind with
Classical Variations that we shall be looking at. out White's light-squared bishop, which in the
If one simply judges by appearances, there original Maroczy Bind (see the Accelerated
doesn't seem to be anything wrong with 3 Fianchetto Sicilian) tends to be a bad piece
transpires that Black controls the d4-square.

and without being able to play d4 White has lit
tle chance of making progress; Black simply
castles and exchanges off the forward knight)
4 ... e6 with the idea of . . . .i.e7, . . . 0-0, . . . tt::lc6 and
perhaps . . . d5 at a later stage. From White's
point of view the idea is to go for plain develop
ment by d3, 0-0 and perhaps a3 to hide the
bishop away on a2. Other ideas are a combina
tion of tt::lc 3 and .i.g5, fighting for control of d5.
As a whole, White will have difficulty making
progress.
c) A tricky alternative is 3 c3 tt::lf6 ! ( now that
White's knight cannot go to c 3 to protect the e
imprisoned behind its own centre pawns. Cas pawn) 4 .i.e2 ! ? (or 4 .i.d3) 4 . . . g6 (4 . . . tt::lxe4?? 5
tling kingside, d4 and b3 are all in the mix with 'iVa4+) 5 0-0 .i.g7 and with accurate play Black
a number of tricky move-orders. Needless to will find himself free from trouble.
say, Black can generate counterplay in the cen All of these Jines have their own theory that
tre either by targeting d4 with . . . g6 and . . . .i.g7, can be researched in books and databases.
or by playing . . . tt::lc6, . . .tt::lf6 and ... e6. He can 3 cxd4 4 tt::lxd4
•..
also play . . . a6 and aim for . . . b5, much as in the A fascinating and inviting variation is 4 'ii'xd4
Accelerated Fianchetto lines. The very absence tt::lc6 (this knight will be pinned, minimizing
of White' s light-squared bishop will make this White's loss of time; alternatively, Black can
easier to achieve. guarantee the win of a tempo against White· s
a2) 3 . . . tt::ld7 4 d4 cxd4 5 'iVxd4 leaves White queen by playing 4 . . . a6, when White also 'gains·
better developed and well centralized. The a move to play 5 c4 if he wants to - this is a sort
moves tt::lc 3 and .i.g5 can follow. Black has dif of Maroczy Bind position that we shall be dis
ficulty getting developed without allowing a fa cussing in various contexts) 5 .i.b5 i.d7 (to re
vourable e5 at some point, so he usually plays new the threat on the queen; after 5 ... 'ii'a5 + 6
. . . e5 and brings his pieces out via . . . tt::lf6 and tt::lc 3 'iVxb5 7 tt::lxb5 tt::lxd4 8 tt::lfxd4 'it>d8 9 c4 or
. . . .i.e7, probably followed by . . . tt::lc 5. This is 9 ile3, White has space and some development
very 'Najdorf-like' , absolutely legitimate, and edge, whereas Black has no weaknesses and the
not easy to talk about without specific exam bishop-pair - theory assesses this position fa
ples. vourably for White, perhaps optimistically so)
a3) 3 . . . tt::lc6 4 0-0 (4 d4 cxd4 5 'iVxd4 trans 6 ilxc6 ilxc6 (6 . ..bxc6 7 c4 is interesting) 7
poses to 2 . . . d6 3 d4 cxd4 4 'iVxd4 tt::lc6 5 ilb5, tt::lc 3 tt::lf6 8 .i.g5 e6 9 0-0-0 .i.e7 J O l':the l 0-0
thus giving that intriguing system added signif (D).
icance; we'll look at it immediately below via
the latter order) 4. . . .i.d7 5 .l:tel tt::lf6 6 c3 a6 7
.ifl .i.g4 8 d3 and White plans tt::lbd2 and h3.
In general Black is equal as long as he is able to
respond to d4 aggressively.
b) White can always play a move such as 3
.i.c4. Ordinarily anything of this nature can be
met by . . . tt::lc6, . . . g6 and . . . .i.g7. Then, because
of the bishop's position, . . . e6, ... tt::lge7 with ... d5
in short order will gain space and time; . . . a6 is
also a useful move. However, Black may not
like that structure and can set up by 3 . . . tt::lf6 4 d3
(4 e5 dxe5 5 tt::lxe5 is the sort of thing that wor-
ries less experienced players but after 5 ... e6 it
A thought-provoking situation has arisen, pit his pawn back with his bishops still on a ram
ting White' s knights, space, and superior devel page, Svidler-Kasparov, Linares 1 999.
opment against Black' s bishop-pair and central 4 tt::lf6 (D)
•••
majority. The general rule with knights is that

their owner needs to hurry to achieve an attack
or gain outposts before his opponent consoli
dates. Otherwise there will come a counterat
tack on the queenside (in this situation by . . . b5)
and/or in the centre (by ...d5 or ... e5). Theory
isn't particularly extensive on these lines and
they afford a lot of scope for creativity. One ex
ample that panned out well for Black was 1 1
@bl h6 1 2 i.h4 :es (Black must always be
wary of some variant of the trick 1 2. . .°iVa5 1 3
°iVd2 :acS?! 1 4 tt::ld5 ! °iVdS 1 5 tt::lxe7+ 'iYxe7 1 6
tt:Jd4 with space and a simple advantage, be
cause l 6 ... g5? 1 7 i.xg5 hxg5 l S 'iVxg5+ is kill
ing) 1 3 i.g3 (this is probably attempting to
avoid the standard idea 1 3 °iVd2 tt::lxe4 1 4 tt::lxe4 5 lbc3
i.xh4 1 5 tt::lxd6 when Black can play 1 5 ... :rs or Notice that now the move 5 i.b5+ simply
l 5... i.xf3 1 6 gxf3 :rs; but here we have a good helps Black to bring his pieces out by 5 ... i.d7,
example of the unexplored nature of this varia especially since White has forfeited the idea of
tion; White could just play 1 3 °iVd3 ! ?, when, for c4. The alternative 5 f3 ! ? has been played infre
instance, 1 3 ... d5 1 4 i.xf6 i.xf6 1 5 e5 i.e7 1 6 quently and yet without disappearing over the
lbd4 :cs 1 7 f4 i s promising; also possible i s 1 3 years. White's idea is to avoid blocking his c
h3 ! ? with the idea g4, as was actually played in pawn by 5 tt::lc 3 and thus be able to play c4, be
one game) 1 3 . . . d5 ! 1 4 e5 (compare the last note fore or after i.b5+. Black's principled response
- here White's bishop on g3 is just bad) 1 4... lbe4 to this plan is 5 . . .e5 ! (D).
1 5 tt::lxe4 dxe4 1 6 °iVxdS :exdS 1 7 tt::ld4 (D) .
This is our first example of the ...d6/. . . e5

1 7 ... i.eS ! (two bishops will be worth more structure. If Black makes conventional devel
than a pawn, even in this relatively simplified oping moves he can be tied down by c4 again;
position) l S c3 ( I S :xe4 i.c6 1 9 :eel i.xg2) that's not the end of the world but not what
l S ...:acS l 9 'it>c2 b5 ! 20 :xe4 b4 21 :e3 a5 22 most players want. Here are two instructive
tt::le 2 i.c6 23 f3 :xdl 24 'it>xdl i.c5 25 :d3 variations:
i.b5 26 :d2 i.e3 27 .l:td6 bxc3 2S tt::lxc3 (2S a) 6 tt::lb3 (not 6 tt::lf 5?! d5 !) 6 ... i.e6 (aiming
bxc3 i.c4 29 a4 :bS) 2S ... i.fl and Black wins for ... d5) 7 c4 a5 S i.e3 a4 9 tt::l3 d2 °iVa5 1 0 i.e2
SICILIAN DEFENCE 181
ii..e7 1 1 0-0 tt::lc 6 1 2 tt::la3 0-0, Rublevsky out, especially the long-ranging bishop on g7.
Ki.Georgiev, Yugoslav Cht (Budva) 1 996. Both and can castle early without creating any pawn
sides have plenty of things to do. weaknesses. His next moves are traditionally
b) 6 ii..b 5+ tt::lbd7 7 tt::lf5 d5 ! 8 exd5 a6! 9 ... ii..g 7, ... 0-0, ...tt::lc 6 and perhaps ... ii..d7 (or
ii..xd7+ li'xd7 (9 ... ii..xd7 10 tt::le3 ii..c5 1 1 tt::lc3 ... ii..e6) with ... .l:tc8 next. Thus Black activates
0-0 1 2 0-0 ii..f5 with active play, Malakhov his pieces rather quickly, especially by compar
Nisipeanu, Holon jr Ech 1 995) 10 tt::le3 b5 1 1 ison with most other Sicilian Defences. He can
tt::lc 3 ii..b7 1 2 0-0 b4 1 3 tt::le4 tt::lxd5 and Black attack on the queenside by, say, ... a5-a4, or
has more than his share of the centre. ... a6/... b5, in part because the g7-bishop exerts
We now return to 5 tt::lc3 (D): so much pressure in that direction.
What are the problems in Black's position?
Perhaps pawn-structure should be the first topic
of discussion. It's true that Black's important
central pawn on d6 is well defended by its
neighbour on e7, unlike the queen's pawn in the
...d6/... e6 structures which distinguish so many
Sicilian systems. We might also compare varia
tions beginning with 1 e4 c5 2 tt::lf3 e6 3 d4 cxd4
4 tt::lxd4, when there is an immediate weakness
down the open file on d6, whether or not it is
occupied by a pawn. At least at first sight that
difference favours the Dragon, and should be
worth something. But as in so many openings,
every advantage carries with it some disadvan
tage. In this situation a white knight can land on
After 5 tt::lc 3, we have finally arrived at a great d5 at the right moment and disturb Black's
dividing point in Sicilian Defence theory, and game. For instance, if White is attacking Black·s
will proceed to the main variations with 2 ... d6. king on g8, the move tt::ld5 might eliminate the
king's best defender. Or in a more positional
setting, White's tt::ld5 might force an exchange
Dragon Variation on that square that results in open lines for
White's pieces. Then there's the question of
1 e4 c5 2 tt::lf3 d6 3 d4 cxd4 4 tt::lxd4 tt::lf6 5 tt::lc3 what Black can do with his central majority.
g6 (D) normally his biggest asset in the Sicilian De
fence. Obviously ... e6 is risky, because the d6-
pawn could be very weak, in contrast to the nor
mal Sicilian lines where Black's bishop de
fends it from e7. And ... e5, the other typical
Sicilian advance, will block the g7-bishop if
Black isn't careful. Naturally both of those
pawn advances can be played under the right
circumstances, but they certainly aren't major
themes. This means that Black's main central
break is . . . d5, which White will do his utmost to
prevent. Assuming that White is successful in
doing so, Black will be using pieces more than
pawns to achieve his goals. Indeed, once we see
the typical positions from either the Classical
or Yugoslav Dragon we shall focus on piece-
The Sicilian Dragon is one of the oldest play on the queenside such as ... .l:tc8, ... ii..e6-c4,
forms of the Open Sicilian. Black gets his pieces ... tt::le5-c4, ... �a5, ...tt::ld 7-c5, ....l:txc3, etc. In
the most frequently played system, the Yugo loosening; White needs to get his pieces out first
slav Attack, the players castle on opposite sides by, for example, 1 2 'iVd2) 1 2 . . .tbe5 1 3 tbd2 ! ?
of the board and it's interesting to see the prior (White wants to prevent the exchange sacrifice
ity that White ' s pawns take in the attack (g4, on c3, but in vain; unfortunately, the attacking
h4-h5, f4-f5, etc.). This can be compared to move . . . .tc6 was coming regardless, and 1 3 g5
Black' s queenside pawns, which often stay at would be met by 1 3 ... �xc3 ! 14 bxc3 { 1 4 gxf6
home until his attack is complete. Finally, before �xe3 1 5 fxg7 @xg7 } 14 . . .tbxe4 with a domi
leaving the subject of pawn-structure, there' s nant position; compare the game) 1 3 ... �xc3 !
the simplest factor o f all: the g6-pawn offers a 14 bxc3 .tc6 (D).
target for attack, in particular by h4-h5 . In other
Sicilian Defence variations, White may achieve
an attacking advance such as g4-g5 (or a posi
tional one like a4-a5) but there' s no specific
pawn target. As usual, these various structural
issues tend to balance out; if they didn't, no one
would play the Dragon! I won't indulge in any
more generalities, considering that the Dragon
quickly breaks up into numerous variations that
superficially have little in common with each
other. It' s better to glean the ideas from the play
itself.
Classica l Dragon
1 e4 cS 2 tbf3 d6 3 d4 cxd4 4 tbxd4 tbf6 5 tbc3 Although the e4-pawn is protected, White's
g6 6 .te2 (D) weaknesses (pawns on e4 and c3, and the in
ternal weakness on e3) make it impossible to
keep things under control: 1 5 .tf3 tbxf3+ 1 6
'iVxf3 d5 ! (White's kingside is exposed and
Black has an active bishop-pair) 1 7 .td4 ( 1 7 e5
d4 1 8 'iVe2 dxe3 1 9 exf6 .txf6 20 'iVxe3 'iVd5 ! )
1 7 . . . dxe4 1 8 'iVh3 tbxg4 ! 1 9 .txg7 (White is
paying the price for f4-f5 and g4; 19 'iVxg4
.txd4+ 20 cxd4 'iVxd4+ leaves no defence)
I 9 . . . @xg7 20 tbxe4 (20 lbb3 tbe5 !) 20 . . .'iVb6+
2 1 tbf2 gxf5 . Black is not just threatening
. . . �g8 with a mating attack; he has regained
his material. This kind of thing has happened
to White a lot, and perhaps explains why the
Yugoslav with 0-0-0 is so popular: White can
push all of his kingside pawns without his king
White develops and announces the likeli being exposed.
hood of his castling kingside. 7 0-0 (D)
...
6....tg7 7 0-0 8 tt:Jb3

7 .te3 tbc6 8 0-0 0-0 is probably the most This strange-looking move is almost always
common alternative, but we're going to consider played in the Classical Dragon, in part out of
that set-up after the move 9 @h l . A famous necessity. First and foremost it prevents ... d5
contest that is often cited versus this move-order (Black's most important freeing advance),
was Daniliuk-Malakhov, Russian Ch (Elista) which is difficult to stop otherwise. It also cov
1 995: 9 lbb3 .te6 I 0 f4 �c8 1 1 f5 ! ? (winning a ers the a5-square (often used by Black's queen
tempo but ceding e5) I l .. ..td7 1 2 g4?! (too or knight) and supports the advance a4-a5 .
kingside. This position appears with opposite

colours in the ultra-popular 'Reversed Dragon'
variation of the English Opening. The latter is
used more by grandmasters as a reply to the
English than any other single system! That adds
weight to our coverage of the ideas and strate
gies here.
9 .ll gS
It seems that most Dragon experts consider
this the most interesting system. It brings the
bishop to the most active square and prepares
f4-f5, often followed by g4-g5 or a well-timed
e5. This carries with it two problems: Black
may get pressure along the a8-hl diagonal (us
Furthermore, 8 lbb3 protects the knight from ing the exchange sacrifice ... �xc3 and . . ..llc 6,
tricky ideas involving ... tbg4 or . . .�6. Natu for example, as we saw above), and White's ad
rally, there are also drawbacks to this retreat. vance f5 gives away the critical e5-square. Both
The main one by far is that on d4 the knight sides have chances. White can also play more
reaches more squares and is more effective for safely with @hl and f4, perhaps with .llf3, or
positive purposes. In fact, after securing his po .lld 3, or �el and .llf l , to protect the vulnerable
sition against . . .d5, White will often return the e4-pawn.
knight to d4. Beyond that consideration, on b3 These instructive options and equally popu
the knight is vulnerable to . . .a5-a4 and if White lar alternatives to 9 .llg5 show a plethora of
blocks this advance by a4, the move . . . .ll xb3 standard Sicilian themes:
can be productive in some (but certain! y not all) a) 9 @hl .lle 6 1 0 f4 and then:
positions. al ) 1 0. . . �c8 (this is a fundamental position)
Note that after 8 .lle3 tbc6, 9 lbb3 will gen 1 1 .ll f3 .llc4! ? (D).
erally transpose, but 9 f4? ! allows the tricky
9 . . .'iVb6 !, when Black's threats of . . . tbxe4 and
. . . 'iVxb2 tum out to be difficult to meet. Al
ways watch out for ...'iVb6 in Dragon positions
whether you're playing White or Black.
8 tt:Jc6 (D)
•••
1 2 �f2 (the best idea in almost all these posi

tions because the rook stays on the f-file and
can also swing over to the d-file if desired;
Black's centre would be unleashed after 1 2 �e 1 ?
e5 ! 1 3 f5 gxf5 ! 1 4 exf5 d5) 1 2 . . . e5 ! 1 3 .lle3 b5
14 fxe5 tbxe5 15 a3 'iVe7, Cabrilo-Chatalba
We'll spend a lot of time from this position shev, Cacak 1 99 1 . The trade-offs are visible:
because it shows so many ideas fundamental to White has the d5-outpost and possibilities of
any Sicilian Defence in which White castles putting pressure on the d6-pawn with all three
major pieces; Black has control of e5, pressure

on the queenside, and the pleasure of watching
White's horrible Sicilian bishop on f3. It' s prob
ably about equal.
a2) Black can also play for the key c4-square
by 10. .. tt:Ja5 1 1 f5 .tc4 1 2 tbxa5 .txe2 1 3
'iVxe2 'iVxa5 1 4 g4?! tbd7 (heading for e5, the
key to Black's defence) 1 5 �f3 e6 ( 1 6 tbd5 was
becoming a problem) 1 6 .td2 'iVd8?! (antici
pating White's idea of 'iVh4, for one thing, but
l 6 ...�ac8 must be better) 1 7 �h3 �e8 1 8 �fl
a6 19 'iVf2 (D).
the introductory chapters, doubled rook' s pawns

are usually not a problem until the endgame.
b) 9 . . .a6 is a popular choice; for instance, 1 0
f4 b5 1 1 .tf3 b4 1 2 tt:Jd5 ( 1 2 tba4 ! has ideas of
e5 that are hard to stop; then access to c5 and b6
could prove critical) 1 2 ...tbxd5 1 3 exd5 tba5.
This position has done well for Black. One ex
ample is 14 �bl ? ! tbc4 1 5 'iVe2 'iVc7 ! 1 6 'lt>hl
.tf5 1 7 g4 .td7 18 f5? tbe5 ! 19 'iVe4 .tb5 20
.te2 .txe2 2 1 'iVxe2 �fc8 22 �f2 a5 ! with the
idea of ... a4, Zapata-Miles, Thessaloniki OL
1 9 84. Compare Black's pieces with White's.
10 'it>hl (D)
We have a primitive yet powerful attack, White makes a somewhat slow move, but a
since all of White's pieces can be directed to major tactical and positional theme is that
wards the king. This is a picture of what Black White's immediate 1 0 f4 opens up his king to
shouldn't allow. Bednorz-Selig, Porz 1 989 con attack along the g l -a7 diagonal and allows
tinued 1 9 ... �f8? ( 1 9 ... 'iVe7 had to be tried; Black 10 ... b5 ! 1 1 ..txb5 ? ! ( 1 1 a3 a5 ! ; 1 1 'iVd3 .tc4 is
must remember to defend along the second equal) 1 1 .. .tbxe4 1 2 .txc6 ( 1 2 tbxe4 'iVb6+ 1 3
rank) 20 g5 ..txc3 2 1 .txc3 exf5 22 �xh7 ! f6 23 'lt>hl 'iVxb5) 1 2 . . .�6+ 1 3 'lt>h l 'iVxc6 -14 tt:Ja5
'iVh4 'iVe8 24 exf5 gxf5 25 �h8+, winning. tbxc3 1 5 tt:Jxc6 tbxdl 1 6 tbxe7+ 'lt>h8 1 7 �axdl
b) 9 �e 1 ! ? is a calm move, giving extra sup ..txb2 and White's queenside weaknesses will
port to the e-pawn if the e2-bishop decides to cause him some trouble, although it's close to
move. Although it's not obvious, a lot of Black's equal.
counterplay will have to do with putting pres 10 tt:Jas
...
sure on White's e-pawn, so this is a sensible Another established plan is 1 0 ... 'iVc8 1 1 f4
precaution. �d8, hoping for ...d5: 1 2 ..tf3 ..tc4 ( 1 2 ... a5 1 3
We now return to the position after 9 .tg5 tbd5 ! ) 1 3 �f2 ! e6? ! 1 4 �d2 ! 'iVc7 1 5 'iVel h6 1 6
(D): .th4 �d7 1 7 �ad 1 e5 1 8 ..txf6 ! ..txf6 1 9 .tg4 !
9 .te6
••• exf4? ! (but Black saw the alternative 19 . . . �dd8
Black develops simply, with an eye on c4 but 20 .te2 ! , when White wins the d5-square and
not abandoning the idea of ... d5. has the upper hand) 20 ..txd7 ! 'iVxd7 2 1 �xd6
a) Another idea is 9 ... b6 10 f4 .tb7 1 1 .tf3 'iVe7 22 �d7 'i¥e5 23 tbd2! ..te6 24 tbf3, con
tba5 ! . This illustrates a common and important solidating his material, Karpov-Miles, Bad
idea: if White doubles Black's a-pawns via Lauterberg 1 977. When Karpov was an e-pawn
tbxa5, Black will exert unpleasant pressure player, he was one of the greatest interpreters of
down the b-file. Remember that, as discussed in ..te2 systems against the Sicilian.
1 1 f4 �c8!? (D) Here we have a picture of both sides consis

Another typical tactical idea is l l ...tbc4! 1 2 tently following their plans in what seems like
f5 ! ? lbxb2 1 3 'i¥c l ..tc4 1 4 'i¥xb2 ..txe2 1 5 an idealized form.
tbxe2 tbxe4 1 6 f6 tbxf6 with three pawns for 14 'i¥d2 b4
the piece and prospects against White's weak At this point 15 tbe2 leads to a balanced po
ened queenside. sition with intriguing opportunities. Instead
White blundered with 15 tbdl ?? tbxe4 ! 1 6
..txe4 i n Onoprienko-Karr, Paris 1 996, and
now 1 6 . . . ..txfl ! would have given White al
most nothing for the exchange.
Yugoslav Attack
1 e4 cS 2 tt:Jf3 d6 3 d4 cxd4 4 tbxd4 tbf6 S tbc3
g6 6 ..te3
This is the usual move-order to introduce the
Yugoslav Attack.
6.....tg7 (D)
Not 6 ... tbg4?? 7 ii.. b5+, when White wins
material since 7 ... ..td7 loses a piece after 8
'iVxg4.
12 rs
The surprising thrust 1 2 e5 ! is a tactical
theme to watch out for when a knight is on a5.
Since 12 . . . dxe5?? loses a piece after 1 3 'i¥xd8
and tbxa5, Black has to calculate in advance
whether he can afford to play 1 2 . . . lbxb3 (not
1 2 ... tbe8? 1 3 tbxa5 'iVxa5 14 ilxe7 dxe5 1 5
i.xf8 ii..xf8, when Black has some compensa
tion for the exchange but not enough) 1 3 axb3
( 1 3 exf6 exf6; this kind of position isn't al
ways satisfactory for Black, although here it is)
l 3 ... dxe5 14 fxe5 tt:Jd5 (14 ... tbd7 15 �xa7 'iVb6
16 �a4) 15 �xa7 tbxc3 1 6 'iVxd8 �fxd8 1 7
bxc3 ii..xe5 1 8 c4 with ilf3 to come and a small
advantage.
12 ... ii..c4 13 i.d3 bS (D) 7 f3 (D)
Here is the Yugoslav pawn-structure. At this 10 f4 f6. Then everything is fine for Black un
point 7 �d2 can be met by 7 ... tbg4, when 8 less White plays the critical 1 1 h4! ?, leading to
.tbs+ .td7 achieves little, and 8 .tgS h6 9 .th4 an atypical attack: 1 1 .. .fxeS 1 2 fxeS .txeS 1 3
tbc6 1 0 tbxc6 bxc6 1 1 f3 �b6 1 2 tbd 1 gS 1 3 0-0-0! tbf6 1 4 tbf3 .txc3 l S �xc3 tbc6 1 6
.tg3 tbeS i s obscure. This idea becomes more .th6 l:.e8 ( 1 6 ... l:.f7 1 7 tbgS) 1 7 hS ! .tfS 1 8
relevant after 7 .tc4 tbg4 8 .tbs+ ! ? (8 0-0 hxg6 .txg6 1 9 l:.xdS lbb4 20 �xb4 tbxdS 2 1
tbxe3 9 fxe3 is bad, as might be expected; �d4 tbf6 2 2 .tc4+ @h8 2 3 �f4 intending tbeS
Black has a permanent outpost on eS that White and Black's not in very good shape.
simply won't be able to get around; e.g., 9 ... 0-0
1 0 �f3 e6 1 1 l:.adl tbd7 1 2 lbdbS tbeS 1 3 �e2
tbxc4 1 4 �xc4 .teS, etc.) 8 . . . @f8 9 0-0 (9 .tgS
h6 1 0 .th4 gS 1 1 .tg3 �b6 !) 9 . . . .teS ! 1 0 h3
tbxe3 1 1 fxe3 @g7 1 2 �f3 l:.f8 and Black' s
control of the eS-square gives him the better po
sition. Hence those wishing to play a system
with h3 and .tc4 should do it by the move
order 7 h3 tbc6 8 .tc4.
After 8 . . .tbc6, we have formally arrived at

the Yugoslav Attack. Since 9 . . . dS is a huge po
sitional threat, White has only three major
moves. I'll focus on the traditional main-line
move :
9 .tc4
a) 9 0-0-0 has less theory than 9 .tc4, which
means only a few books' worth! Castling doesn't
stop 9 . . .d5 (indeed, it's the main line), even
7 0-0
... though that move allows White, after 1 0 exdS
7 . . . tbc6 generally won't make much differ- tbxd5 1 1 tbxc6 bxc6, to win a pawn by 1 2 tbxd5
ence (unless you're a 'Dragdorf' player who cxdS 1 3 �xdS . Fortunately, Black can then
puts his knight on d7; this odd hybrid system play the clever 1 3 . . .�c7! with good compensa
has been moderately popular of late). If White tion, the first point being that after 14 �xa8
prefers to play 8 .tc4 at this point (delaying .tfS (threatening the queen and checkmate on
�d2), then 8 ... �b6 should be answered by 9 c2) l S �xf8+ @xf8 Black recovers most of the
.tbs threatening tbfS , which probably gives material with an ongoing attack. Thus White
White a small edge. More fun is 9 tbfS �xb2 usually plays 1 2 .td4, when the most combat
1 0 tbxg7+ @f8 1 1 tbdS tbxdS 1 2 .txdS with ive continuation is 1 2 . . .eS 1 3 .tcS (D).
the dark squares and the bishop-pair in return Now 1 3 ... .te6! 1 4 tbe4! l:.e8 l S h4! h6 1 6 g4
for a pawn. leads to all kinds of complications. Notice that
8 �d2 tbc6 (D) White didn't take the rook by 1 4 .txf8 ; it turns
Since the freeing move . . . dS is so vital, it's out that 14 . . .�xf8 (threatening ... .th6) gives
instructive to see what White might do if it is Black wonderful play for the exchange due to
played right away, something that most players his dark-square control and attack via ... l:.b8 and
don't even consider. After 8 . . . dS ? ! it seems nec ... fS, sometimes mixed with ... �b4 or . . . lbb4.
essary for White to respond aggressively if he is This is a typical case of a bishop being worth as
to gain the advantage, beginning with 9 eS tbe8 much or more than a rook until the players
We now return to 9 .llc4 (D):
reach a simplified position, assuming that White

makes it that far.
This is j ust the very briefest of introductions We have arrived at the starting-point of one
to 9 0-0-0. It's up to you to plunge into that ter of the most analysed opening variations in
ritory if you get the inclination. chess, and quite possibly the most analysed. In
b) 9 g4 is played much less frequently. The 1 975 many of us believed that the Yugoslav
idea is 9 . . . d5? 1 0 g5, winning a pawn. Few Dragon was beginning to get 'analysed out'.
Dragon aficionados use this line as White, how but decades later theory is continuing to ex
ever, primarily because of 9 ... Cl'ixd4 1 0 .llxd4 pand, with perhaps 200 times as much serious
.ll e6 or the immediate 9 . . . .lle6 (D). material having been played and analysed. Since
the main variations are so tactical and so criti
cal, you simply need to study in detail those that
you have chosen. Authors of the many Dragon
books and CDs are fond of saying that you can
play even the main lines of this opening armed
only with a firm grasp on the general ideas. In
fact that would only be possible on a low level
of play where you are more or less guaranteed
that whomsoever you play will not know much
theory. The simple fact is that the player who is
familiar with a Dragon variation and knows it
by heart will almost always beat the opponent
who doesn't. For one thing, it took untold hours
of home study and computer analysis to work
out most of the Dragon positions that are now
The point of the latter move is that 1 0 Cl'ixe6 part of theory, so the knowledgeable player will
fxe6 covers the key d5-square and opens the f benefit from the specific results of that work.
file against the weak f-pawn. Black's basic ideas On top of that, many of the best Dragon moves
of attack along the c-file in conjunction with the are counterintuitive and not the choice that
g7-bishop are essentially the same, and in some you would make under time constraints. Con
cases it's convenient to have an escape-square sequently the most practical solution for those
on f7. After the natural 1 1 .llc4, Black can play who want to play the Dragon as Black or use the
either 1 l . . �c8 followed by . . . Cl'ie5 or l l . . . d5 ! ?
. Yugoslav Attack as White is to find lines in
1 2 exd5 Cl'ie5 1 3 .lle 2 Cl'ixd5 1 4 Cl'ixd5 exd5 1 5 which to specialize and/or require less work. At
0-0-0 e6. This i s a line you might want to look any rate, this book is not intended as a theoreti
into as White and should definitely be aware of cal tome so I ' ll just present games that show a
if you're playing Black. number of themes for both sides.
In many ways the Dragon has simpler basic be happy to enter into it. He can use his queen
ideas than most other Sicilian Defences, which side attack by . . . b5-b4 to chase White' s pieces
contributes to its appeal. The Classical varia from defence, and it is quite common to sacri
tions feature standard kingside set-ups for White fice that b-pawn in order to open queenside
and a limited group of queenside attacking lines for the attack. There are numerous other
schemes for Black. I have devoted space to it ideas - too many, in fact, to explore here.
partly as a matter of practicality for the average
player. In the Yugoslav Attack we find a set of Returning to 9 ..tc4, what specifically does it
fairly straightforward themes to become famil do? It puts the bishop on an aggressive diago
iar with. For White, one such motif is the primi nal, yes, but also stops ... d5. For this purpose
tive h4-h5 to open the h-file, followed by ..th6, White subjects himself to a time-consuming re
exchanging the bishop that defends the vulner treat in the face of c-file pressure and ...tbe5 or
able dark squares around the king. Then White . . . tba5, hoping that the bishop's defensive role
proceeds to checkmate or otherwise overwhelm on b3 (guarding a2, protecting the king from b
Black by hook or by crook, using thematic file attack) will justify its exposure, even to the
moves such as tbd5 and tbxf6, g4-g5 or what longer-term idea . . . a5-a4. There is no way to
ever is at hand. Such is the barbaric stuff of tens explore all of the intricate theory of the entire
of thousands of games. On a much less frequent attack, of course, so I'll show a few games and
but arguably more sophisticated level, White game excerpts .
plays centralizing and prophylactic moves such
as �bl and �he l , perhaps in conjunction with Stefansson - Ward
lDb3, ..td4 and either e5 or tbd5 . Alternatively, Reykjavik 1 998
playing Wbl and lbb3 by themselves is a way to
batten down the hatches against Black's queen 9 ..td7
•••
side attack. That may come in conjunction with Black simply develops. His idea is to put a
the simplifying tbd5 . ..tg5 is a common move rook on c8 and play ... tt:Je5, sometimes directly
in many variations, increasingly popular as the by . . . �ac8 but often with the order ... 'i¥a5,
years have gone by, and particularly against the ...�fc8 and then ... tbe5, as in this contest.
...h5 lines. This serves the purpose of threaten 10 0-0-0
ing ..txf6 and tbd5 at some point, but also has The immediate 10 h4 will often transpose af
the idea that strategies involving f4 and e5 have ter 1 0... 'i¥a5 and l l . ..�fc8, but this order is said
more chance of success. The bishop move can to discourage the 'Chinese Dragon' which uses
be beneficial in that if Black plays his standard the scheme ...�b8 and ... b5 ; I won't go into any
...lbe5-c4 manoeuvre, White may be able to details of that still-controversial notion, but it's
slide the queen away, perhaps to e2, because worth considering. A variation with a long and
capture by ...tbxe3 is no longer possible. As a independent history is l 0... �c8 1 1 ..tb3 tbe5 12
general rule, neither side can lose their dark 0-0-0 (12 g4 can now run into 1 2 . . . a5 ! ? 13 a4
squared bishop without putting their position in h5 when one can argue that the weakening of
peril, unless of course that happens via sacrifice White's queenside favours Black by compari
or other forcing sequence. son with other ... h5 lines) 1 2 ...tbc4 ( 1 2 . . . h5
Obviously you have to play this variation for transposes into a form of the Soltis Defence,
quite a while to understand or be helped by that which we shall see later on) 1 3 ..txc4 �xc4 14
characterization. What about Black? Instead of h5 ! ? ( 1 4 g4 b5 ! ? 15 h5 has also been analysed
pawn-pushes such as h4-h5, he has the two in depth) 14 ...tbxh5 (D).
seemingly unavoidable ideas of . . . tbe5-c4, to 15 g4 (seemingly small variations in move
rid White of one of his bishops, and ... �xc3. order can make all the difference in the Dragon;
The latter exchange sacrifice can be played as e.g., 15 tbde2 'i¥a5 1 6 ..th6?! { a typically crazy
part of a mating attack, or to set the stage for an continuation is 16 g4 ! tbg3 ! ? 1 7 tbxg3 ..txc3
all-out assault, or simply to weaken White's 18 bxc3 'i¥a3+ 19 Wbl ..te6 20 'i¥h2 ! h5 2 1
structure such that if the right endgame or tbf5 ! ! �b4+ 2 2 cxb4 ..txa2+ 23 Wal ..tb3+ 24
queenless middlegame comes along, Black will Wbl ..ta2+ with a draw } 1 6... ..txc 3 ! 1 7 tbxc3
match) 24 ... .ll xd5 (24. . . dxe5 25 tbxf6+ exf6

26 lbh5 and mates) 25 exf6 exf6 26 'i¥xh7+
@f8 27 'ifu8+ 1 -0 Karpov-Korchnoi, Moscow
Ct ( 2) 1 974.
10 �aS
...
After 1 0 ... tt:Je5 1 1 .llb 3 l:.c8 1 2 Wbl , the

modem move 1 2 ... l:.e8 ! ? (D) has been surpris
ingly successful:
l:.fc8 and it' s hard to stop ... l:.xc3 without com

promising White's position) 1 5 ...tt:Jf6 1 6 tt:Jde2!
(the classic game that follows shocked the chess
world for its simplicity: instead of launching all
his pawns and pieces into the kingside attack,
White guards the c3-knight with both the other
knight and a rook, and then proceeds to attack
undisturbed) 1 6...�a5 ( l 6 ...l:.e8 ! has been played
since this time with decent chances, salvaging
the bishop in the case of 1 7 .llh 6 .llh8 ! ; in the The idea is that ...l:.e8 gives the critically im
meantime bold ideas such as 1 7 e5 tbxg4! 1 8 portant g7-bishop a chance to save itself from
fxg4 .llxg4 with dynamic compensation be exchange ( 1 3 .llh 6 .llh 8), and it also guards the
came commonplace) 1 7 .llh 6 .llxh6? (offering e-pawn in some lines with tbd5 (thus preparing
the exchange with the retreat 1 7 ... .llh8 ! ? is a . . . �a5). Finally, . . . l:.e8 is a key element in many
better try) 1 8 �xh6 l:.fc8 I 9 l:.d3 ! (now White' s of the variations in which Black defends by
e2-knight will come to the aid of the attack) . . .h5 , so it also serves a purpose against a king
1 9 . . .l:.4c5 (D). side pawn avalanche. An amazing amount of
good from such a nondescript move ! White can
proceed 1 3 .llh6 .llh8 14 h4 tbc4 1 5 ..txc4
l:.xc4 1 6 tbde2 b5 1 7 h5 b4 ! 1 8 tbd5 tbxd5 1 9
hxg6 hxg6 20 �xd5 .lle6 2 1 �d3? (2 1 �5)
21 ... �a5 (Black is already on the verge of win
ning) 22 b3 l:.ec8 ! 23 .Ile 1 (23 bxc4 .llxc4 24
�e3 �xa2+ 25 @c l .ll x e2! 26 �xe2 .llc 3 and
mate next move) 23 ... .ll g7 24 l:.d2 �e5 (this is
one way to win, just lining up along the power
ful diagonal; although 24... l:.8c5 threatening
25 . . . �xa2+ ! would have ended things quickly)
25 c3 bxc3 26 l:.c2 l:.b4 27 .llh6 .llxb3 ! 28 axb3
l:.xb3+ 29 @c 1 .llxh6+ 30 l:.xh6 �g5+ 0- 1
R.Perez-Y.Gonzalez, Holguin City 2002.
We now return to the position after 1 0... �a5
20 g5 ! l:.xg5 (20 . . . lbh5 2 1 tbf4 ! ) 2 1 l:.d5 ! (D):
l:.xd5 22 tbxd5 l:.e8 23 tbef4 ! (23 tbxf6+? 11 h4
exf6 24 �xh7+ @f8 and there is no mate) Thi s is the most principled move for the at
23 . . . .llc 6 24 e5 ! (these are wonderful tactics in tacker: waste no time and go for the kill! These
what was effectively a world championship days h4-h5 is normally played without the
16 .td4
Here is the kind of centralized defence that
we haven't seen yet: @bl , ClJb3 and .td4; these
moves secure White's king and prevent all those
.. .l:.xc3 sacrifices, at least for now. 16 h5 allows
the predictable 1 6... l:.xc 3 ! 1 7 'iVxc3 'iVxc3 1 8
bxc3 Cl'ixh5 and even with the queens off Black
has more than enough play, with moves like
. . . a5-a4, . . ..te6 and ... l:.c8 to come.
16 .te6 17 hS aS
.••
Black in response charges forward with his

pawns, also not the main strategy that we have
seen him employ.
18 a4 bS! ! (D)
support of g4 if . . . h5 hasn't been played, since
the g-pawn advance costs a crucial tempo and
weakens f3. B ut against 10 . . .l:.c8 instead of
l O . . .'iVa5 there is a great body of theory about
both approaches. White has other related op
tions after 10 . . .'iVa5, notably 1 1 @b l , but also
1 1 .tb3 with 1 2 .tg5 and l:.he l in mind.
11...Cl'ieS 12 .tb3 l:.fc8 13 @bl
This patient move introduces a plan com
bining defence with attack. The more aggres
sive 1 3 h5 Cl'ixh5 gives Black a free view down
the long diagonal. Then we have more stan
dard themes, such as in this encounter from the
old days: 14 @bl (this looks similar to 1 3
@b 1 , but falls into the usual exchange sacri
fice) 14 . . . l:.xc3 ! ? 1 5 'iVxc3 ( 1 5 bxc3 l:.c8 1 6 Apparently first used in over-the-board GM
.th6 Cl'ic4 1 7 .txc4 l:.xc4 with a positionally practice in this game. Black insists upon open
winning game for Black) 1 5 . . . 'iVxc3 16 bxc3 ing lines with the maximum speed.
l:.c8 17 @b2 ( 1 7 .tg5 ! ? ; 1 7 Cl'ie2 .tb5) l 7 . . . a5 19 Cl'ixbS 'iVb8 20 Cl'ic3
18 a3 ClJf6 19 .tf4 Cl'ie8 ! ? ( 1 9 . . . b5) 20 .tg5 a4 Ward analyses 20 h6 .th8 2 1 e5? ! dxe5 22
2 1 .ta2 Cl'ic6 with equality, Spassky-Stein, .txe5? 'iVxe5 23 'iVd8+ Cl'ie8 ! and Black wins.
Russia-Ukraine (Uzhgorod) 1967. 20 l:.b4! 21 hxg6 hxg6 22 .txf6?!
••.
13 Cl'ic4 14 .txc4 l:.xc4 I S Cl'ib3 'iVc7 (D)

.•• Trying to exchange some pieces. 22 l:.h4 ! was
played in several other contests, when the play is
dynamic and unclear; for example, 22 ... .txb3
(or 22...'iVb7) 23 cxb3 �xb3 24 CDb5 l:.b4 25
l:.dhl (25 'iVc2 'iVb7) 25 . . .l:.xa4 26 Cl'ic3 l:.xd4 !
27 'iVxd4 ClJh5 28 'iVd2 a4 threatening ... a3, and
Black has a real attack, Mallee-Mikhailov, corr.
Wch 1 977-83.
22 .txf6 23 ClJdS .txdS 24 'iVxdS l:.a6!
•••
Preparing to triple the rooks and queen on

the b-file, and also to play ... e6.
2S f4 e6 26 'iVd3 l:.ab6 27 l:.h3 l:.xa4
Intending . . . �ab4 and ... a4.
28 fS dS! 29 fxe6 'iVeS
The point: White's queenside is collapsing.
30 exf7+ @rs 31 c3 fxg4 tbf6 1 9 l:.h4 l:.fd8 ! 20 l:.ghl tbexg4 and

Best but depressing is 3 1 �c3 �xc3 32 bxc3 Black has extra material and all the key squares,
l:.xe4 33 Wa2 l:.e2. Valeriani-Raty, corr. 1 985 .
31 ... l:.xb3 32 exd5 l:.ab4! 33 l:.d2 �el+ 34 13 ... tbc4 14 .txc4 l:.xc4 (D)
@a2 �cl 0-1
A cute finish would be 35 �e2 l:.a3+ ! .
Soltis Variation
1 e4 cS 2 tt:Jf3 d6 3 d4 cxd4 4 tbxd4 tbf6 5 tbc3

g6 6 .te3 .tg7 7 f3 0-0 8 �d2 tbc6 9 .tc4 .td7
10 0-0-0 l:.c8 11 .tb3 tbeS 12 h4 hS (D)
15 tbde2
The best attribute of . . .h5 is that White has to
prepare so long to play an effective g4; for in
stance, 1 5 g4?! hxg4 1 6 h5 tbxh5 1 7 l:.dgl
�c8 ! 18 fxg4 .txg4 1 9 tbd5 l:.e8 20 l:.h4 e6 ! 2 1
tbc3 f5 ! 2 2 tt:Jdb5 �c6 and Black had a clear
material and positional advantage in Hardicsay
Hemdl, Oberwart 1 984.
This is the Soltis Variation, the most fre Another option for White is 15 lbb3 �c7 1 6
quently played line of the Yugoslav Attack. .td4 (D) , the plan that w e saw in the last sec
Black simply stops White's pawn advance and tion, with Black a 'tempo ' ahead due to playing
dares him to break down Black's own defences ... �c7 in one move rather than two ( . . .�a5-c7).
before getting overrun on the queenside. The
typical Dragon themes that we showed in the
first game still apply, so we'll discuss a few ad
ditional ideas as we go along. Remember that
this is a non-technical inquiry that undertakes
to instruct by example.
Anand - Kasparov
New York Wch (1 1 ) 1995
13 Wbl
A rather slow move, although White pre
pares to meet 1 3 .. .�a5 by 14 tbd5 ! .
Instead, 1 3 g4?! lets Black break up White's
centre way before his king feels any danger:
1 3 . . .hxg4 14 h5 tbxh5 15 .th6 e6 (cutting off Play is fairly balanced; e.g., 1 6 . . . .tc6 17
White's bishop and opening up a diagonal for g4 ! ? ( 1 7 �e3 tbd7 ! ? 1 8 .txg7 Wxg7 1 9 g4!
Black's queen) 1 6 l:.dgl �f6 1 7 .txg7 ( 1 7 hxg4 20 h5 tbf6 21 tbd4 .td7 22 hxg6 fxg6 23
fxg4? .txh6 1 8 �xh6 �f4+ 1 9 �xf4 tbxf4 20 �h6+ is a little scary but probably all right for
l:.fl g5 is positionally killing) l 7 . . .�xg7 1 8 Black) 1 7. . . e5 ! 1 8 .te3 hxg4 1 9 h5 gxf3 ! 20 h6
lLixe4 ! 2 1 lLixe4 ! .ixe4 22 hxg7 .l:i.xc2! 23 this stuff) 1 6 @b 1 .l:i.c8 (or l 6 . . . b5) 17 lLie2 a5
gxf8'i:V+ @xf8 24 @al .l:i.xd2 25 lLixd2 .id5 26 1 8 .l:i.d4 ( 1 8 lLif4 ! is an excellent alternative).
lLib l .ie6, Pieretti-Perilli, corr. 1 985. Probably Now in Cabanas Bravo-Semprun, 2004, Black
Black's pawns should outweigh all those pieces ! found the nice idea 1 8 ... a4 ! 19 .ixa4 .ixa4 20
15 ... b5 16 .ih6 .l:i.xa4 �6+ 2 1 .l:i.b4 'i:Vf2 with complications
Short but tension-filled was 1 6 e5 ! ? dxe5 17 generally in his favour. I ' ll just give the raw
.ig5 .l:i.c7 1 8 .ixf6 exf6 1 9 g4 'i:Ve8 20 gxh5 moves: 22 lLif4 .l:i.xc3 23 .l:i.c l .l:i.xc2 24 lLih3?
.ie6 21 .l:i.dgl b4 22 lLie4 f5 23 h6 fxe4 24 hxg7 (24 .l:i.xb7) 24 ... .l:i.xc l + 25 'i:Vxcl 'i:Vxg2 26 lLig5,
@xg7 25 h5 .ixa2+ ! 26 @xa2 'i:Va4+ 27 @bl and now Black had 26 . . . l2Jd3 ! 27 'i:Vc8+ @g7
.l:i.d8 28 'i:Vxd8 'i:Vxc2+ 29 @al 'i:Va4+ 1h-1h Li 28 'i:Vxb7 lLixb4 29 'i:Vxb4 'i:Vh l + 30 @c2 'i:Vxh4,
berzon-Miles, Haifa OL 1 976. winning.
16 'i:Va5 17 .ixg7 @xg7 18 lLif4 .l:i.fc8 19
•.. 14 .ixc4 .l:i.xc4 15 .ixg7 @xg7 16 @bl
lLicd5 'i:Vxd2 20 .l:i.xd2 lLixd5 21 lLixd5 @rs 16 lLid5 e5 ! (D) is strangely logical, in spite
The game is equal. of giving up d5 as a permanent outpost and ex
posing the d6-pawn to attack !
To wrap up the Dragon section, we'll ex
plore two games, each one featuring a move by
White's dark-squared bishop.
Short - Fleck
Bundesliga 198617
13 .ih6 (D)
For years this natural continuation was con
sidered the real test of 12 ... h5, and it arguably
did more for the Soltis Variation than anything
else because of the great games it produced.
The conventional wisdom is that Black, if well
prepared, has nothing to fear.
Since Black has traded off his g7-bishop, he

can place his central pawns on dark squares.
Ridding himself of the well-placed knight on
d4 hurts White's queenside defensive prospects
and is worth a pawn if necessary; for instance,
l 7 lLie2 lLixd5 1 8 'i:Vxd5 .ie6! 1 9 'i:Vxd6 'i:Va5 20
a3 .l:i.fc8 2 1 c3 .l:t4c6 22 'i:Vb4 'i:Va6 23 .l:i.d2 .l:tb6
with a powerful attack, Westerinen-H.Mi.iller,
Germany tt 1 989/90.
16 'i:Va5 17 lLib3 'i:Vc7
•..
1 7 ... 'i:Ve5 ! is probably better.

18 g4!? hxg4 19 h5 gxf3 20 .l:i.dgl! l:.g8! 21
hxg6 fxg6 22 lLid5 'i:Vd8 23 lLid4 e6 24 lLif4
and White has a dangerous attack.
13... l2Jc4
The popular alternative 1 3 ... .ixh6 14 'i:Vxh6 lvanchuk - Topalov
l:l:xc3 1 5 bxc3 features your customary ex Belgrade 1995
change sacrifice again, which is risky but has a
respectable standing. For example, 15 ... 'i:Vc7 13 .ig5 (D)
( 15 . . .'i:Va5 and 15 . . .'i:Vc8 are also possible - refer We already discussed the virtues of this con
to the books for pages of games and analysis on tinuation in the introductory remarks to the
Yugoslav Attack. I should note that, along with 1 5 h5 appears to let Black get through on
the idea of being able to sidestep .ixc4 after the queenside before White can do the same on
. . . l2Jc4, there is a similar idea connected with the other wing, but it's a close call and could
the move . . . l2Jg4. As a response to f4 (or even change with one new move. The play can de
as the second move of the sacrificial device generate into a primitive slugfest; for example:
... .ixg4), ... l2Jg4 will not gain a crucial tempo 1 5 ... lLixh5 1 6 lLid5 .l:i.e8 ( 1 6 . . . .l:i.xd5 ! 17 .ixd5
on the dark-squared bishop because it has gone 'i:Vb6 is another course) 17 f4 lLic4 1 8 'i:Vf2 ! ? b5
to g5. 1 9 f5 a5 20 'i:Vh4 .l:txd5 ! 21 exd5 lLixb2 ! 22 fxg6
At present 13 .ig5 is considered the main fxg6 23 lLic6 'i:Vb6 24 .l:i.del a4 25 .ie3? 'i:Vc7 26
line of the Soltis 1 2 h4 h5 variation and the 'i:Vg5 axb3 27 axb3 l2Jd3+! 28 cxd3 .ixc6 29
themes are useful to study. @dl .l:i.a8 30 .if4 e5 ! 31 'i:Vxg6 .ixd5 32 .l:i.xh5
13 .l:i.cS!
•.• .if3+ 33 @d2 .l:i.a2+ 34 @e3 exf4+ 35 @xf4
The move that salvaged Black's cause in the 'i:Vf7+ 36 'i:Vxf7+ @xf7 37 .l:i.f5+ .if6 38 .l:i.c l
Soltis Variation. It is useful in several ways: @g6 0- 1 Kravtsov-Soloviov, St Petersburg
a) It protects the 4th rank against advances 1999. Fun stuff.
by f4 and e5. 15 lLic4 16 'i:Ve2 (D)
•••
b) It opens up the possibility of a sacrifice on A popular move at the time of writing, but 1 6
g5 to eliminate the crucially-important bishop. 'i:Vd3 has hundreds o f games and truckloads of
c) The rook helps to defend b5, sometimes analysis to its credit.
as a preliminary to ... b5.
d) Black prepares to double rooks on the c
file.
White has several options and there are count
less games from this position connected with
intricate analysis by many strong masters. At
this point if Black plays 1 3 . . . l2Jc4, White can
take advantage of the absence of his bishop
from e3 to play 14 'i:Ve2. Then 14 . . .l2Ja5 15 @bl
a6? illustrates how one slight error in these
lines can land you in terrible trouble: 16 g4 ! e5
(D).
1 7 gxh5 ! ! exd4 1 8 lLid5 lLixb3 19 h6 ! ! .ib5
20 �h2 d3 2 1 cxb3 lLixd5 22 hxg7 .l:i.c2 23
.ixd8 .l:i.xd8 24 .l:i.d2 1 -0 Nunn-Mestel, London
1 986. I like these old games; they seem so inno 16 'i:VcS
..•
cent and refreshing ! 1 6 ...ltJaS ! ? 17 e5 lLixb3+ 1 8 lLixb3 .l:i.xc3 1 9

14 g4 hxg4 15 f4! bxc3 .ic6 ! i s a wild line that appears to be
dynamically equal, the high-level stem game

going 20 l:.hfl lL'ie4 2 1 �c4 d5 22 �xe4 ! ? dxe4
23 l:.xd8 l:.xd8 24 i.. xe7 l:.d7 25 i..g5 i..f8 26
lL'id4 i.. c5 27 f5 i..xd4 28 cxd4 gxf5 29 l:.xf5
l:.xd4 30 i.. e3 l:.a4 3 1 l:.g5+ �h7 32 l:.h5+ �g8
33 l:.g5+ 1/z- 1/z Smirin-Ivanchuk, Paris 1 994.
17 i.. xf6
Typical tactics arise from 17 f5 lL'ixb2! 18
�xb2 l:.xc3! 1 9 fxg6! l:.xb3+ ! 20 axb3 fxg6 21
l:.hel �c5 and things are still unclear, Fogar
asi-Palkovi, Budapest 1 996.
17 i.. xf6 18 lL'idS l:.xd5!?
...
Another exchange sacrifice ! This one elimi

nates White's best piece and allows Black's
mighty bishop to survive. There are also games 20 gS! 21 fxg5 i.. xg5+ 22 �bl f5! 23 l:.d3
•..
with 1 8 ... b5 ! ?, allowing 1 9 lL'ixf6+ (paradoxi f4 24 i.. xc4 �xc4 0-1

cally, letting the bishop live by 1 9 h5 g5 ! 20 Ivanchuk is known for resigning early. Nev
i..xc4 bxc4 2 1 �e3 l:.xd5 22 exd5 is White's ertheless, in the hands of a player like Topalov
best try) 19 ... exf6 20 h5 g5 (DJ. the pawns and bishops will definitely win in the
end. This is another relatively old game be
tween world-class players that shows how para
doxical and counterintuitive the best play in the
Dragon can be. Don't think that you can depend
upon this section as reliably up-to-date theory,
because that is always changing. Instead, it is
intended to be a set of noteworthy schemes and
tactics.
Najdorf Variation
1 e4 c5 2 lL'if3 d6 3 d4 cxd4 4 lL'ixd4 lL'if6 5 lL'ic3
a6 (DJ
In modem chess you basically play what

works ! You'd think that giving up your most
important piece in the middle of getting at
tacked would be suicidal, but the specifics of
the resulting odd-looking pawn-structure actu
ally hold up. White can't seem to make prog
ress; e.g., 2 1 �f2 �d8 ! 22 h6 �e7 23 l:.del
gxf4 24 �xf4 �e5 25 �f2 g3 ! 26 �gl �f4+
27 �bl lL'id2+ 28 �al lL'ixe4 and Black has a
winning game, Kasarova-Krasilnikova, Ekater
inburg 1 997 .
19 exd5 b5 20 h5? (DJ
Even though Black's had plenty of options in
the notes, White could show that he's still on
the right track by playing 20 i..x c4! bxc4 2 1 Whereas the Dragon Variation may have the
c 3 ! , when he stays material u p and may well be most appeal to the average player, the Naj
able to consolidate and/or keep up the attack. dorf Sicilian has been the favourite opening of
top-level players for many years now. Part of 6 . . .e5 7 lL'if3 lL'ibd7 8 a4 (versus . . . b5) 8 . . . .ie7 9
this was certainly the influence of World Cham .id3 0-0 1 0 0-0 (D).
pions Fischer and Kasparov, consistent devotees
of the variation. There is also the inherent com
plexity and diversity of the Najdorf concepts
and themes, to some extent in contrast to the
relatively straightforward ideas of the Sicilian
Dragon. White has a large variety of absolutely
independent systems available for choosing,
and Black can respond with varying basic
structures. The Najdorf has an especially fluid
character: again in contrast to the Dragon, we
see more central breaks to go along with flank
attacks, and in most variations the centre ulti
mately plays as large a role as the attacking for
mations on either side of the board.
What is that magical little move 5 . . . a6 all
about? First of all, flexibility, which is perhaps Now with 10 . . .exf4!? Black strives to control
the most valuable asset in modem openings. As the e5-square in return for d5 . This is the quint
the move 4 ... a6 does in the Paulsen Sicilian ( 1 essential Sicilian strategy, since it provides an
e4 c 5 2 lL'if3 e6 3 d4 cxd4 4 lL'ixd4 a6), so outpost on e5 at the same time that it opens up
Black's 5 . . . a6 in the Najdorf makes an implicit the e-file and the h8-al diagonal i.n support of
challenge to his opponent. White has played that square. But White gets something from the
five unexceptionable moves (e4, lL'if3, d4, lL'ixd4 deal too: he gains the key d4-square (usually for
and lL'ic3), which essentially tell Black nothing a knight), gets an open f-file, and may be able to
about what he is up to. But now it is time for put pressure on d6 more effectively because of
White to commit one of his bishops, which by a bishop that occupies f4. These trade-offs have
defining the play will allow Black to respond to be constantly evaluated when Black considers
accordingly. The development of White's light whether to play - and White considers whether
squared bishop is particularly meaningful in to allow - the capture . . . exf4. Of course Black
that regard. If it goes to e2, then Black might has other moves which we won't elaborate
play . . . e5, which would not be highly recom upon here, notably 10 . . . 'i:Vc7 and 1 0. . . lbc5. Af
mended in the Dragon or Taimanov Sicilians, ter 1 0 . . . exf4, White plays 1 1 @hl ! (after 1 1
for instance. If White's bishop ends up on c4, .ixf4 Black takes the pawn and lives to tell the
Black can block the bishop by . . .e6, and so tale: 1 1 . . .'i:Vb6+ 1 2 @h l 'i:Vxb2) 1 1 . ..lL'ie5 1 2
forth. Similarly, a dark-squared bishop on e3 or .ixf4 'i:Vc7 1 3 'i:Vd2 .ie6 14 lL'id4 .l:i.fe8 ! ? and
g5 will require different strategies from Black. now:
For all that, 5 . . . a6 is fundamentally slow and a) 15 lL'if5 ! ? could be answered by 15 . . . .ixf5
simply invites White to go on an offensive. 16 exf5 d5 ! with active play in return for the
Thus the Najdorf is a risky system in which the bishop-pair; nevertheless, this line is unclear.
slightest inaccuracy can spell disaster. But as 15 ... .if8 16 .ig5 lL'ifd7 is also possible, but then
Kasparov says, 'High risks mean high rewards', 1 7 a5 ! (to keep a knight out ofb6) 1 7 . . .f6 1 8 .if4
adding that with the Najdorf, Black will usually .l:i.ac8 19 lL'ie3 ! gets a knight to d5 with some ad
get a chance to seize the initiative at some vantage. The odyssey of the knight from f3 to d5
point. But he cautions that any generalizations in four moves brings to mind lL'ibd2-fl -e3-d5 in
about strategy need to be supported by thor the Closed Ruy Lopez. It also shows that giving
ough homework. up d4 to a centralized piece can have more than
We shall investigate 6 .ig5, 6 .ic4, 6 .ie2 the obvious consequences.
and 6 .ie3. b) 15 .ixe5 dxe5 16 lL'ixe6 fxe6 (D).
The continuation 6 f4 is rare these days. One We see this structure in several Sicilian lines,
idea that demonstrates a basic Sicilian theme is and also in other openings where the move
.ixe6 or ... .ixe3 occurs, or even l2Jg5 and lLixe6. repeat themselves you will have a good head-
Black' s pawns protect the key central squares start.
d5 and d4, as well as f5 and f4. In such positions There are of course many ways in which the
the key questions are whether the pawns can be play can develop. Most of them have White ei
attacked (they are unsupported by other pawns) ther attacking on the kingside or in the centre.
and who has the better pieces. Mainly because In both cases he will resort to piece sacrifices
of the relative strengths of the bishops, I prefer whenever they are useful or necessary, because
Black. Of course this is just one example, not a the pawns alone won't generally be enough to
verdict! break down Black's position. Black has some
interesting counterattacking ideas on the king
side, but will usually proceed with a basic plan
The i.g5 Attack
of development followed by central and
6 .ig5 queenside attacks. Or he can leap into action by
White's predominant choice for years was playing the so-called Poisoned Pawn Variation
this direct attacking line, and it is still the choice and grabbing material. I'll outline these possi
of many experts. White wants to play f4 fol bilities in a few games.
lowed by pawn-breaks like e5 or f5 if allowed.
Since Black usually prevents those, White's Sulskis - Pelletier
main set-up begins with 'i:Vf3 and 0-0-0. We shall Warsaw Ech 2005
see the associated themes as we go along. In the
meantime Black plays ...e6 followed by what are 7 .te7 8 'i:Vf3 'i:Vc7 9 0-0-0 lLibd7 10 g4
•••
the most typical Najdorf moves ... l2Jbd7, ... 'i:Vc7, Or:
. . . b5 and ... .ib7, usually but not always with an a) After 1 0 'i:Vg3 Black has a key defensive
early ....ie7. Again, the standard set-ups will ap manoeuvre that comes up again and again:
pear with examples. 10 . . .h6 1 1 .ih4 g5 ! (D).
6 ... e6 7 f4 (D) 12 fxg5 lLih5 ( 1 2 ....l:i.g8 has also equalized,
This is the most direct attacking scheme quickly recovering the pawn) 1 3 'i:Ve3 'i:Vc5 ! (this
that you will see in the Najdorf and has led to attacks g5 for the third time) 14 @b 1 ( 1 4 'i:Vd2
crazy sacrificial brilliancies for both sides for .ixg5 1 5 .ixg5 'i:Vxg5 1 6 .ie2 lLihf6 17 lLif3
years. Three games will follow, and since the 'i:Vxd2+ 1 8 .l:i.xd2 @e7 is equal) 14 ... hxg5 1 5 .if2
variations are so tactical and diverse, I shall l2Je5 . Here is the main point of ... g5 : Black
lean towards recent examples and stay at least counts upon this knight to hold everything to
within shooting distance of current theory. gether. Kengis-Vitolins, Jurmala 1 983 continued
Again, only specific study of concrete varia 1 6 'i:Vd2 'i:Vc7 17 lLif3 b5 ! 1 8 .ie3 ( 1 8 lLixg5
tions will let you truly master 6 .ig5, whether gives Black good queenside play after l 8 ... b4 19
White or Black. It should be said, however, lLia4 .ib7 20 lLib6 .l:i.b8) 1 8 ...g4 1 9 lLixe5 dxe5
that if you can pick up some of the ideas that 20 .id3 lLif4 2 1 'i:Vf2 .ib7 with equality.
The idea is that occupation of d5 is worth a piece

if you add to it an attack along an open e-file and
dangerous kingside pawns.
For decades 1 3 f5 ! ? has been the principal
continuation ( 1 3 a3 is supposed to be met by
1 3 ... .l:i.b8 and ... b4). I ' ll give some illustral:iYe
lines (and not necessarily best play) : 1 3. . .�
( l 3 ... .ixg5+ has also been tested for years md
it seems that White has more than enough for
his temporary pawn loss; to emphasize the de
gree of specific study surrounding these lines.
I'll note John Emms's reference to the case ol
one grandmaster losing to another because ol a
novelty on move 28 that produced a brillianl de
b) The same idea can be introduced by 10 fence on move 3 1 ! ) and here are two mcwa:
.id3 h6 1 1 .ih4 g5 12 fxg5 lL'ie5 13 'i:Ve2 lL'ifg4. a) 14 g6!? is a thematic break that is f�
This time the queen is better placed on f3 than tured all over the Sicilian landscape: 14...bxg6
on g3, so the play is less clear: 1 4 lL'if3 ! hxg5 1 5 15 fxg6 fxg6 1 6 b4 ! ? lL'ia4 17 lbxa4 bxa4 18
.ig3 .id7 1 6 h3 ! ? ( 1 6 .l:i.dfl would b e more like e5 ! ? (aggressive, but that doesn't necessarily
a real test, because Black has to protect his f7- mean good!) 1 8 ... dxe5 ? (D) ( 1 8 ... d5 feels rigbl;
pawn before he can castle; e.g., 1 6. . . lL'ixd3+ 1 7 Black may even get time for ... .i.xb4 ).
'i:Vxd3 0-0-0! ? 1 8 h 3 lL'ie5? 1 9 lbxe5 dxe5 20
.l:i.xf7 .ie8 2 1 .l:txe7 !) 1 6...lL'ixf3 1 7 gxf3 lbe5
1 8 f4 gxf4 19 .ixf4 0-0-0. This pawn-structure
is fine for Black, who can now become active
with ... .ic6 and perhaps . . . f5 .
10 b5 l l .ixf6 lL'ixf6 12 g5 lbd7 (D)
•.•
1 9 .id3 ! ( 1 9 'i:Vxa8 exd4 20 .ixa6 0-0 was

Black's idea, after which 2 1 .l:i.hfl is unclear)
1 9 ... .ig5+ 20 @bl .if4 ! 2 1 'i:Vxa8 exd4 22
.ixg6+ (White should play 22 .ixa6 !, and this
time 22 ... 0-0 23 .l:i.xd4 doesn't seem to cut it for
Black) 22 ... @e7 23 .l:i.hfl ! ? e5 24 'i:Ve4 •c4 25
A line contested in untold numbers of games .l:i.xf4! 'i:Vxb4+ 1h-1h Markzon-de Firmian. New
throughout the years. We' ll outline a few ideas York Open 1 99 1 .
while we follow the main game. b) 1 4 f6 gxf6 1 5 gxf6 .if8 1 6 J:l:g l 1'.d7
13 lL'ifS!? ( 1 6. . . h5 ! ? 1 7 .l:i.g7 b4 1 8 lL'id5 ! exd5 1 9 exd5 is a
Hardly the main move; I'll promote it because typical tactic; White has cleared out the e-file
it's refreshing, and also so that we have some and captured the c6-square - whether that's suf
thing current to mull over. In the Sicilian De ficient for a piece has to be decided upon a
fence we see knight sacrifices on b5, d5, e6 and case-by-case basis) 1 7 .l:i.g7 b4 1 8 lL'id5 ! exd5
f5, all hoping to break down Black's defences. 19 exd5 (D).
Offering a rook.
19....l:i.ae8!
After 1 9 . . . .ixh 1 ? 20 f6+ @d8 2 1 �xh8+
@c7 22 �xh7 White wins three pawns and has
an ongoing attack for the knight.
20 .l:i.d3! @d8 21 .l:i.hdl .l:i.hg8 22 �c3?!
22 �f6+! is better.
22 ... .l:i.e2 23 .ig4! .l:i.f2 24 �f6+ @c8 25
�xd6 �c7 26 .l:i.c3 .ic4 27 b3 .l:i.xf4 28 h3
�xd6 29 .l:i.xd6 @c7 30 .l:i.f6 .l:i.xg5
The game is equal and was eventually drawn.
When Black delays ... .ie7, another set of

tactics can arise. A couple of these are repre
The same idea. This time Black can escape sented in the course of examining another slug
the e-file checks by castling, but he's still under fest:
attack; for example, 1 9 ... 0-0-0 20 .l:i.xf7 .ih6+
2 1 @bl .l:i.df8 22 .l:i.xf8+ .l:i.xf8 23 lL'ie6 ! 1 -0 Kosten - Kr. Georgiev
Shmuter-Kaspi, Tel Aviv 1 996; Black could Saint Afrique 2005
play on, but 23 . . .lL'ixe6 24 dxe6 .ixe6 25 .ih3 !
.ixh3 (25. . .�d7 26 �a8+) 26 �xh3+ �d7 27 7...lbbd7 8 �f3 �c7 9 0-0-0 b5 (D)
�xh6 is pretty hopeless.
13...exfS
1 3 . . .b4 ! ? may be better but this is more illu-
minating.
14 lL'idS �b7!?
A typical line given by Kosten is 14 . . .�c5 1 5
exf5 .ib7 1 6 f6 ! gxf6 1 7 lL'ixf6+! .ixf6 1 8
�xb7 .l:i.c8 1 9 .id3 .ig7 2 0 .l:i.hel + with an un
clear attack.
15 �c3!
1 5 exf5 ? lL'ib6 exchanges White's key piece.
15...lL'ib6 16 lL'ixe7!?
Or 16 �xg7 ! .l:i.f8 17 lbxe7 �xe7 1 8 �d4
.l:i.b8 1 9 .ig2.
16...@xe7!? 17 �xg7 .ie6 18 exf5 .ids 19
.ih3! (D) 10 .id3
Here we see the centralization strafegy: White
ignores the idea of kingside attac Kby g4 in fa
vour of .l:i.hel and potential advances and/or
sacrifices in the middle of the board.
10... .tb7
After 1 0... b4?, we get that sacrifice 1 1 lL'id5 !
again, but this time White is simply better after
l l . . .exd5 1 2 .l:i.hel ! .ib7 1 3 exd5+ @d8 1 4
lL'ic6+ .ixc6 1 5 dxc6.
1 1 .l:i.hel �b6!?
The older move l l . . . .ie7 runs into another
lL'id5 idea: 1 2 �g3 ! b4 1 3 lL'id5 exd5 1 4 exd5
(threatening lL'if5) 1 4 . . . @d8 ( 1 4... g6 1 5 �h4!)
1 5 lL'ic6+ ! .ixc6 16 dxc6. This position has been
played and analysed extensively by Thomas b) Even nicer is 1 2 . . . exd5? 1 3 lLic6 ! ! (look
Luther. I ' ll j ust follow a recent game: 16 . . . l2Jc5 for this in similar positions !) 1 3 . . . .ixc6 1 4
17 .ih4! .l:i.g8 18 .ixh7 ! .l:i.h8 19 'fifxg7 .l:i.xh7 exd5+ .ie7 1 5 dxc6 lLic5 1 6 .ixf6 gxf6 1 7
(D). .if5 'fifc7 18 b4! lLie6 1 9 'fifh5 lLig7 20 .id7+
�f8 2 1 'fifh6, Chiburdanidze-Dvoirys, Tallinn
1 980.
13 .ixf6 gxf6 14 .txb5! 'fifes 15 tillf6+
�d8 16 lLixd7 'fifxb5 17 lLixf8 .l:i.xf8 18 9a3
.l:i.c8! 19 'fifxd6+ �e8 20 .l:i.e3 .l:i.g8?
A fatal mistake. Nevertheless, 20 . . .9c6 2 1
'fifd2 leaves White with three pawns and a nice
attack for the bishop.
21 .l:i.c3 .ic6 22 f5! .l:i.xg2 23 fxe6 :tl? 24
.l:i.c5
White is winning.
Poisoned Pawn Variation

1 e4 c5 2 lLif3 d6 3 d4 cxd4 4 ltJxd4 lLif6 5 lLJc.J
20 'fifxf6 ! .l:i.xh4 2 1 'fifxf7 .l:i.h8 22 .l:i.e5 ! .l:i.f8 ! ? a6 6 .ig5 e6 7 f4 'fifb6 (D)
2 3 'fifg 7 .l:i.a7? (23 . . .b 3 ! ? is better) 24 .l:i.xc5 'fifb6
(24. . ..l:i.xf4 25 .l:i.c4! .l:i.xc4? 26 'fifg8+) 25 'fife5 !
and White had enough attack to convert to a
winning position in B.Vuckovic-Tadic, Herceg
Novi 2005. Of course you needn't study specif
ics to play the Najdorf, just rely upon general
ideas. Right.
12 lLidS!? (D)
This is an astonishing move that those raised

with classical chess principles would simply re
ject as a typical beginner's mistake. Black goes
running after a pawn when he is undeveloped
and already under attack. What's worse. he
does so with the queen, which you're not sup
posed to bring out too early because it will lose
time.
Again! There have been a lot of games be 8 'fifd2 'fifxb2 9 .l:i.bl
tween leading grandmasters with other moves, Sometimes White plays 9 lLib3 instead, but
but this move is the scariest for Black. we'll stick with the overwhelming favourite.
12 'fifxd4!
..• 9 'fifa3 (D)
.•.
Or: Now it's White's move and he has perfect at

a) One simple but very pretty variation is tacking squares for his pieces; in addition, after
1 2 ... lLixd5? 13 exd5 'fifxd4 14 .l:i.xe6+ ! fxe6 1 5 White castles he will very likely put his rook on
'fifh5+ g 6 1 6 'fifxg6+ hxg6 17 .ixg6#. an open f-file after the normal advances e5 or
and unprotected by a pawn, and the central

square e3 also qualifies. In addition, the fourth
rank squares c4 and e4 tum out to be vulnera
ble, especially significant since a white bishop
on c4 would be loose. Even d4, although poten
tially able to be protected by a pawn on c3, can
be shaky in practice. In this variation, weak
nesses tend to mean loose pieces and potential
outposts for the enemy.
c) Central pawn-majority. It cannot be over
stressed what Black's strongest weapon is in
the Sicilian Defence: his extra central pawn,
which in the main line of the Poisoned Pawn
Variation sometimes becomes a central pawn
f5 . His other rook is already on an open file and mass capable of giving exceptional protection
after l:l'.b3 (with tempo ! ) it can swing along the to Black's king and pieces.
third rank and attack where needed - this is a Having said that, the most important point to
standard theme in many openings, by the way. remember is something that Kasparov eternally
Thus every white piece will be participating in stresses: this variation depends upon specific
an attack against an opponent with almost no tricks and tactics for both sides, and there is no
pieces out and no safe place for his king. For overriding reason that White's attack shouldn 't
Black, this is a sure recipe for disaster. win, nor that Black's defence shouldn 't prevail;
Or is it? In fact, the Poisoned Pawn Variation to a large extent the result is just the way things
has been taken seriously for well over four de work out.
cades now and has survived countless attempts We' ll examine one game and a bundle of
to refute it. In the meantime, Black's outrageous notes from the key position after 9 ... 'i:Va3.
pawn theft has played a considerable role in
revolutionizing chess theory and practice. Under Thinius - Kersten
the leadership of World Champions Fischer and Bad Zwesten 2006
Kasparov, players began to realize that Black
could play this and similar positions with every 10 f5!
expectation of success. Why? There are several The modem continuation. White doesn't
general answers, but three stand out: fully bum his bridges as he does in the old and
a) The queen on a3, although subject to fur extremely natural line 1 0 e5 dxe5 1 1 fxe5 t2Jfd7
ther attack, is also an attacking piece, able to tie 1 2 .ic4 (D), in which White is blasting open so
White down to protecting his own position and many lines and developing so quickly that it's
prevent him from straying at will. Older theory amazing Black can survive. But Fis � r and
would say (at least when the queen sortie is be others demonstrated that he does so and then
ing contemplated) that most pawn raids with some.
the queen would have to be accompanied by Now 1 2 . . .l2Jxe5 ? goes too far after 1 3 l2Jxe6,
other retreating moves by her to get back to but Black has no fewer than three satisfactory
safety. But now there are plenty of situations in moves, at least two apparently leading to an ad
openings where a queen retains her position in vantage for Black in a position that at first was
the enemy camp, saves time, serves a useful considered close to a forced win for White !
function, and says ' Show me' . Computer anal a) Fischer and others used 1 2 . . .'i:Va5 with
ysis has assisted in finding new examples. success; its theory has advanced considerably
b) Black has no weaknesses! White, on the and the verdict seems to be dynamic equality.
other hand, has a problem that we often refer b) One classic line goes 1 2 . . ..tb4 1 3 l:l'.b3
to in this book: internal weaknesses, espe 'i:Va5 14 0-0 0-0 15 .if6 ! ? ( 1 5 l2Jxe6? ! fxe6 1 6
cially those on the third rank. The main one .ixe6+ �h8 1 7 l:l'.xf8+ .ixf8 1 8 'i:Vf4 l2Jc6 ! 1 9
here is on the c3-square, adjacent to the centre 'i:Vf7 'i:Vc5+ 20 �h l l2Jf6 ! 2 1 .ixc8 l2Jxe5 22
S!CIUAN DEFENCE 201
Now:
a) 1 6 . . . .ic5+? ! 1 7 cj;;h l .l:!.f8 1 8 c4 .l:!.xfl + l 9
.l:!.xfl .ib7 20 �c2 ! (not 20 .ig4? dxc4 and
Black defended by . . .�d3 and won easily in
Fischer-Geller, Monte Carlo 1 967) 20... e4 2 1
.ig4 .ie7 22 �f2 winning for White, Tal-Bog
danovic, Budva 1 967.
b) 16 ... .l:!.a7 ! 17 c4 �c5+ 18 cj;;h l d4 1 9
.ih5+ g 6 2 0 .id 1 ! with a powerful attack that
keeps Black's king running around in the cen
tre; e.g., 20... .ie7 2 1 .ia4+ cj;;d8 22 .l:!.f7 (22
.ixe7+ .l:!.xe7 23 �g5 cj;;c7 24 l:Jel is unclear)
22 . . . h6 23 .ixh6 e4 24 .ie3 e5 25 .ig5 e3 26
.ixe3 and Black is struggling, Grij alva-B.Gon
�e6 lL'ieg4 0- 1 Tringov-Fischer, Havana 1 965) zalez, Internet ICC 2000.
1 5 ...lL'ixf6 1 6 exf6 (formerly considered a draw) 14 .ixf6 gxf6 15 lL'ie4 .ie7
1 6. . . .l:!.d8 ! 1 7 .l:!.xb4 �xb4 1 8 �g5 g6 1 9 lL'ixe6 Maybe 1 5 ... �xa2 1 6 .l:!.d l .ie7 1 7 .ie2 0-0 is
( 1 9 �h6 �f8) 1 9. . . .ixe6 20 .ixe6 �xc3 ! 2 1 also adequate. After some 1 5 more moves of
.ixf7+ cj;;xf7 2 2 �h6 lL'ic6 23 �xh7+ cj;;e6 24 analysis and game tests, it apparently draws no
�xg6 �d4+ 25 cj;;h l .l:!.f8 ! 26 .l:!.el + cj;;d6 27 matter which of several attacking methods
�g3+ cj;;c5 28 c3 �xf6 0- 1 Ballester-Monteau, White uses !
French Cht 2002. 16 .ie2 (D)
c) According to modem theory, 1 2 . . . �c5 !
is probably best of all, attacking those weak
nesses that we discussed and pretty much forc
ing 1 3 .ixe6 fxe6 14 lL'ixe6 �xe5+ 1 5 �e3
.id6 ! , when White is coming up short.
10 ... lL'ic6 1 1 fxe6 fxe6 12 lL'ixc6 bxc6 13 eS!
White had better strike fast in order to open
lines and weaken Black's effective central de
fenders.
13...dxeS
The most popular line, establishing a central
pawn-mass with which to defend the king. That
said, there is a long history behind 1 3 . . . lL'id5.
One line out of hundreds goes 14 lL'ixd5 cxd5
1 5 .ie2 dxe5 1 6 0-0 (D).
16...hS
This stops .ih5+. Gipslis-Korchnoi, USSR
Ch (Leningrad) 1 963 shows how delicate
Black's situation is: 1 6 . . . 0-0? 1 7 l:.b3 �a4 1 8
c4 cj;; h8 1 9 0-0 l:.a7 20 �h6 f5 2 1 l:.g3 .ib4 22
lL'if6 1 -0. Mate is unstoppable.
17 .l:!.fl !?
17 l:.b3 has its own lengthy theory, as does
1 7 0-0 f5 and now 1 8 .l:!.f3 or 1 8 .if3. In both
cases Black seems to survive, with draws being
the customary result.
17 ...fS (D)
1 7 . . . �xa2 is risky: 1 8 .l:.d l �d5 1 9 �e3 with
a strong attack, Radjabov-Ye Jiangchuan, Calvia
OL 2004. The new generation is still finding

new ideas in this mess!
The name 'Sozin' is connected with ..ltc4 in

both the Najdorf and the Classical lines, so I'll
designate 6 ..ltc4 as the Najdorf Sozin. Although
18 l::tf3! 'fixa2 19 l::tfb3 'fia4! this direct bishop development has never been
The variations are almost infinite; for exam as popular as 6 ..ltg5, and in these days not as
ple, 1 9 .. .fxe4 20 'fic3 (intending l::ta l) 20... ..itd8 ! popular as 6 ..lte3, it is still used successfully by
may hold on, although that is shaky. loyal adherents. The play after 6 ..ltc4 divides
20 lLid6+ ..ltxd6 21 'fixd6 into a set of positional and primarily tactical
Threatening l::tb 7. lines, so I'll treat it that way. It's probably fair to
21...'fia5+ say that the slower lines tend to end up in equal
This position had already been played! In ity or even in Black's favour because of his
Fernandez Siles-Gamundi Salamanca, Albacete long-term advantages, but the more numerous
2004, Black misplaced his queen and lost dynamic lines are much harder to assess, with
quickly by 2 1 . ..'fie4? 22 l::tb7 'fih4+ 23 g3 'fid8 beautiful tactics seemingly the rule rather than
24 'fixe5 l::th6 25 'iig7 1 -0. the exception.
22 @fl @f7?! Black almost always plays . . . e6 (usually
This attempt to evacuate the king loses. 6 . . . e6) in order to restrict the scope of the c4-
22 . . . l::t a7 ! is Kosten's suggestion, using sec bishop. After that White has to be careful about
ond-rank defence. This move is the key to the move . . .d5 or . . .lbxe4 followed by a . . .d5
many lines. fork, so he will retreat his bishop to b3. That's
23 l::tb 7+! @g6 24 'fie7! the basic position from which strategies are
Improving upon yet another game, where the formed, as we shall see.
inferior 24 l::tc 7? had been played. 6 e6
•.•
/
24 ..ltxb7 25 'fixe6+ @g5 26 'fie7+ @g6 27
..• 6 . . .lbxe4? walks into 7 'fih5 ! wiili multiple
'iid6+ @g5 28 h4+ @f4 29 '>t>f2! 1-0 threats, when the best that Black can do is
Black is helpless in the face of 30 g3+ or 30 7 . . .d5 ! 8 ..ltxd5 lLid6. Then, however, White
l::tb4+. plays 9 0-0 and Black has trouble getting his
pieces out, since . . . e6 is met by a capture on that
I 'm sure that all this back-and-forth activity square and . . . g6 by 'fie5 .
will persist for years to come. The theoretical Players often wonder why Black doesn't
result is probably a draw, but the practical out simply attack the bishop right away with the
come depends heavily upon one's preparation. useful move 6 . . .b5. One problem is that the b
pawn advance is committal; since . . .e6 will
Najdorf Sozin Attack doubtless be played anyway, Black may not
want White to know on what basis he will set up
1 e4 c5 2 lbf3 d6 3 d4 cxd4 4 lbxd4 lbf6 5 lbc3 his attack. Velimirovic-Mrdja, Yugoslavia 1 984
a6 6 ..ltc4 (D) went 7 ..ltb3 (7 ..itd5 ! ? is also very interesting
because 7 . . . lbxd5 8 exd5 yields a structure that primarily positional and revolved around these
is almost always favourable to White, so Black factors.
might try 7 ... 'g.a7 ! ?, and if 8 ..lte3, then 8...'g.c7 ! ?) As in many Najdorf variations, if White plays
7 . . . ..ltb7 (7 . . . e6 transposes to a main line) 8 f4 and e5 (instead of f5) the game will often
..lte3 ! ? (or 8 'fie2; or 8 0-0 b4 9 lLid5 lLixe4? 1 0 tum very tactical, and White may have to shift
'g.el lbc5 1 1 ..ltg5 ! - White has too many pieces his strategy to piece sacrifices before his ad
out) 8 ...lbbd7 9 f4 lLic5 1 0 0-0! ( 1 0 e5 dxe5 1 1 vanced central pawn falls. Those lines are very
fxe5 lbxb3 1 2 axb3 ..ltxg2 !). White's 1 0th move position-specific and exciting. Alternatively,
introduces a type of e-pawn sacrifice that has White sometimes foregoes f4 altogether and
dozens of variants and forms. Sometimes it simply brings his pieces out. This has become a
works and sometimes it doesn't. The positional very popular strategy, although it contradicts
basis c 9115 ists of a lead in development, the what for years was the conventional wisdom,
opening of White's e-file, and Black's difficul i.e. that the b3-bishop ran into a brick wall at
ties in proceeding with his own development. e6 and that it took pawn advances to remedy
This particular game continued 1 0. . .lbfxe4 1 1 that.
lbxe4 ..ltxe4 ( l l .. .lbxe4 1 2 f5) 1 2 f5 ! (stopping We' ll follow various games from this posi
both . . . e6 and . . . g6) 12 . . . lbxb3 1 3 axb3 'iid7 1 4 tion. I ' ll show a lot of tactical ideas which are
'iig4 ! ..itd5 ( 1 4 . . .d 5 1 5 c4) 1 5 'g.f2 g6 1 6 c4 ! (it fairly universal in their character and apply
seems that in almost every game with this kind elsewhere, but there will also be some purely
of attack White needs to open up another unique and creative combining for your enjoy
front) 1 6 . . . bxc4 1 7 bxc4 gxf5 1 8 lLixf5 ..ltb7 1 9 ment. Dynamic attacking play is what has al
..ltd4 e5 20 ..ltb6 f6 2 1 'g. d l d5 22 cxd5 'fif7 23 ways drawn the average player to the Najdorf
l::tc2 'iig6 24 'fia4+ @f7 25 'iid 7+ @g8 26 'g.d3 Sozin.
1 -0.
7 ..itb3 (D) Morozevich - Agrest
St Petersburg Z 1993
7 ..lte7 8 f4
•.•
Other common continuations are 8 0-0 and 8

..lte3.
8 b5! ?
.•.
This natural move allows a typical tactical se

quence although Black is used to such things in
the Najdorf. His main alternative is 8 ... 0-0, when
9 0-0 is usual; a fairly obscure continuation is 9
f5 ! ? exf5 (9 ... e5 1 0 lbde2 and White will have
an easier time of controlling d5 with Black hav
ing castled and he not having done so) 1 0 exf5
d5 1 I 0-0 lbc6 1 2 @h l ! with a quite interesting
isolated d-pawn position. White doesn't have the
The starting position for most variations. If usual restraint on Black's centre, but his ad
White plays f4-f5 and Black responds with ... e5, vanced pawn interferes with Black's customary
the argument revolves around occupation of the IQP activity.
d5-square. Should White succeed in exploiting 9 e5! dxe5 10 fxe5 lLifd7 1 1 ..ltxe6! ? (D)
it as a pure outpost, he will probably stand This thematic sacrifice pervades the 6 ..ltc4
better. When Black can prevent a piece from lines, and also occurs in the Classical Sicilian
establishing itself there or gain compensating and even in the English Attack (usually via g4-
advantages, his natural Sicilian attack on the g5, ..lth3 and ..ltxe6). When it works, it is the ul
queenside will usually come into play. The timate triumph of the bishop over its nemesis
variations that top players such as Fischer en on e6. In this situation, objectively, maybe White
tered into (he played both sides of 6 ..ltc4) were should prefer 1 1 'iig4 with the idea l 1 .. .'iVc7 I 2
'i'Jxg7 'i'Jxe5+ 1 3 'i'Jxe5 lbxe5 1 4 ..ltf4 and White Reutsky - Shtyrenkov

enjoys a pleasant advantage. Noiabrsk 2003
7...lbbd7
This development has been popular for some
years, especially after Kasparov used it versus
Short in their world championship match. The
knight temporarily prevents e5, but generally it
goes to c5 next, from which post it can protect
e6 against f4-f5 and eliminate the b3-bishop
when Black chooses to do so. White is chal
lenged to find a way to attack Black's solid
structure.
8 f4 lbc5 (D)
1 1 lbxe5!
••.
We get to see two of the major themes of the

..ltc4 Najdorf: destruction of Black's centre by
..ltxe6 and of White's by . . . lbxe5. White obtains
an overwhelming attack after 1 l .. .fxe6?? 1 2
lbxe6 'i'Ja5 ( 1 2 . . .'i'Jb6 1 3 lbd5 ! ) 1 3 lLixg7+ @f8
( 1 3 . . .@d8 1 4 lbe6+ '>t>e8 1 5 'i'Jh5#) 14 0-0+
lLif6 ( 1 4. . . @xg7 1 5 'i'Jg4+) 15 exf6 ..ltc5+ 1 6
@hl , and Black is getting slaughtered.
12 ..ltxc8
1 2 ..itd5 ! ? is another idea.
12 ...'i'Jxc8 13 lLid5 ..ltc5!
Black gets busy defending his dark-square 9 'i'Jf3
weaknesses. White was threatening lLib6 and White can return to traditional positional play
'i'Je2, with secondary ideas of ..ltf4 and 0-0. by 9 f5 ..lte7 1 0 'i'Jf3 ! ( 1 0 fxe6 was played in
14 b4!? ..lta7 many games following Short's example, but
After 14 . . . ..ltxd4 1 5 'i'Jxd4 lbbc6 1 6 'i'Jc5, then White turned to this developing move,
Black is tied up and White can develop by ..itf4 which reserves the idea of exchange on e6, and
or ..ltb2 with ideas of castling on either side of also prepares g4-g5) 1 0...0-0 1 1 ..lte3 (otherwise
the board. it's hard to develop) l l . ..e5 1 2 lLide2 lLixb3 1 3
15 ..ltf4 'i'Jd7 axb3 b5 1 4 g4 (we see the difference between
1 5 . . . 'i'Jc4 1 6 lLif5 ! 'i'Je4+ 1 7 'i'Je2 is nicely the early days of ..ltc4 with f4-f5 and today's
symmetric. If Black plays 1 7 . . . 'i'Jxd5, 1 8 .i:i.d 1 version ! The advance of the g-pawn changes the
wins, but 1 7 . . .'i'Jxf5 1 8 g4 ! drives the queen entire dynamic of the position) 1 4 . . .b4 (Black
away from protecting e5 . certainly can't wait around for g5 and lbd5) 1 5
16 ..ltxe5 'i'Jxd5 17 ..ltxg7 'i'Jxg2 18 'i'Je2+ lba4 ..ltb7 1 6 lbg3 'ilc7 ! ? (the aggressive 1 6 . .d5
.
'i'Jxe2+ 19 @xe2 ..ltxd4 has also been played, when the battle begins be
Probably 1 9 . . . l:i.g8 20 lLif5 lLic6 improves for tween White's rapid development and Black's
Black. central play) 1 7 0-0-0 .i:i.ac8 1 8 .i:i.d2 d5 ! (White
20 ..ltxd4 l:i.g8 21 a4 lbc6 22 ..ltc5 was again ready for g5 followed by f6 and a
Although this position is probably within kingside attack) 19 g5 d4?! ( 1 9 ...lbxe4! 20 lbxe4
Black's drawing range, White's bishop proved dxe4 2 1 'i'Jg2 ! f6 22 .i:i.hdl is difficult to assess
decisively superior to Black's knight in the long but White has notions of lbb6-d5) 20 gxf6 dxe3
run. 21 'i'Jxe3 ..ltxf6 22 lLih5 (D).
Black had probably missed 24. . . @xf7 25

'ilih7+ ..ltg7 26 .i:i.fl + @e7 27 lbxe6 ! ! (instead of
27 'ilixg7+? @d8) 27 . . .@xe6 28 'ilixg7 with
mating threats and a quick win.
25 l:i.xg7+! @xg7 26 'ilih6+ @gs 27 'ilixg6+
@hS 2S lbxe6 lbde5 29 ..itd4 'ilih7 30 li'f6+
@gs 31 ..ltxe5 lbxe5 32 lD2d4 'ilif7 33 .i:i.fl 1-0
Resignation seems premature but there fol
lows 33 ... 'ilixf6 34 l:i.xf6 ! threatening g6-g7 and
Black can ' t do much about it.
Finally, we get to Black's main move:

7 b5 (D)
•..
22 . ..'ilie7 23 l:i.g l (23 l:i.d7 ..ltg5 !) 23 . . .l:i.c7

(everything seems to be holding together, but
now comes the overloading move) 24 lbc5 !
with a nice attack based upon either lbd7 or
lbxb7 and .i:i.d5, Vega-Lopez Gomez, corr. 1 995.
9 ..lte7 10 0-0 0-0 11 ..lte3 'ilic7 12 .i:i.ael
..•
l::teS 13 g4 b5 14 g5 lLifd7 15 f5
No subtlety here: White goes for the kill but
Black gets the wonder square e5. Again posi
tional factors determine the tactical possibilities.
15 ...lbe5 16 'ilih5 g6 17 'ilih4 ..itfS
Or l 7 . . . lbxb3 1 8 cxb3 b4. Now White can
try 1 8 fxe6 ! ? and lbd5.
lS fxg6 hxg6 19 ..itd5!? ..ltb7 20 ..ltxb7 'ilixb7
21 b4 lbcd7 22 .i:i.f2 .i:i.acS 23 lbce2 (D) Here are three games with two fundamen
tally different strategies.
Kristjansson - Tukmakov
Reykjavik 1972
s f4
This is the traditional pawn attack. White
wants to play for f5 and force a response that
gives him control of d5. Options that emphasize
piece attack are given in the next game.
S ... ..ltb7 9 f5 e5 10 lbde2 ..lte7
Playing 1 0 . . . lbbd7 first may be the most pre
cise order; for instance, 1 1 ..ltg5 ..lte7 1 2 lbg3
.i:i.c8 ! (Black tries to counter White's appropria
tion of d5 with queenside action) 1 3 0-0 ( 1 3
23...lbc4? ..ltxf6 ! ? lbxf6 1 4 0-0 { 1 4 lbh5 } 1 4 . . . h5 ! threat
Black will get punished for moving this key ens to win the e-pawn after . . . h4) 1 3 . . . h5 ! (D).
defender; it's almost impossible to break down A fantastic move that directly stops White's
such an ideally-placed piece when it's supported only real threat, which was to bring the knight
by a bishop and another knight. He would stand to h5 in order to eliminate another defender of
very well with either 23 . . . ..ltg7 or 23 . . . lbb6. d5. Now White went rapidly downhill: 14 h4?
24 .i:i.xf7! ..ltg7 b4 1 5 ..ltxf6 ..ltxf6 ( 1 5 . . .lbxf6 is also good) 1 6
In general White won't get much advantage

if he has to capture on d5 with a pawn instead of
a piece. Here he stands considerably worse.
17 'iib6 18 a3 a5 19 axb4 axb4 20 l::txa8+
.•.
..ltxa8 21 'iif2 'iia 5! 22 0-0 ..ltxd5 23 c4 ..ltc6 24

l::te l h3 25 g3 l::th5 26 lbd4?
Better, but still depressing, would be 26 .ltd 1
l::t g5 27 lbc 1 ..itd7 ! .
26...exd4 0-1
Christiansen - Wojtkiewicz
USA Ch (San Diego) 2006
8 'iif3 (D)
lbd5 ..ltxh4 1 7 lbxh5 'itig5 (the contest is al
ready over!) 1 8 f6 g6?! ( 1 8 . . . filh5 ! 1 9 l:H5
..itf2+ ! wins right away due to 20 l::txf2 'iih6) 1 9
lbg7+ '>t>d8 20 l::tf3 ..ltg3 2 1 'iid3 ..ith2+ 22 '>t>fl
lbc5 23 l::th3 l::th4 24 'iif3 lbxb3 25 axb3 l::txh3
26 'iixh3 ..ltxd5 27 exd5 'iixf6+ 28 @e 1 'iif4
0- 1 R.Byme-Fischer, Sousse IZ 1 967. A game
that went a long way toward discrediting 8 f4.
11 lbg3!
This move improves upon 1 1 ..ltg5, which as
we just saw only assists Black's attack.
1 1 ...h5! ?
The same idea, but without ..ltg5 in, maybe
Black is asking for too much. Instead, 1 1 . . .lbbd7
is natural and probably best.
12 'iif3?! Originally no one liked this idea but over
White could take over d5 directly by 1 2 the years it has assumed the mantel of ' Main
..itd5 ! lbxd5 1 3 lbxd5 h4 1 4 lbh5 . Line ' . 8 'iif3 is less weakening and develops
12...lbbd7 13 ..ltg5? the pieces more quickly than 8 f4.
Losing the thread. He should have developed 8 ...'iic7
by 1 3 0-0. 8 . . . 'iib6 9 ..lte3 'iib7 is the other conven-
13 h4 14 ..ltxf6 lLixf6 15 lbge2 b4 16 lbd5
••. tional defence, slow but perhaps playable.
lbxd5 17 exd5 (D) 9 ..ltg5 lbbd7 10 0-0-0
White's moves are very natural but rarely
used until recently. This was probably due to
Fischer's example; he consistently employed
the idea of f4-f5 to break down Black's e6/f7
structure. The logic was that the e6-pawn ren
dered White's b3-bishop ineffectual, so it had
to be eliminated. However, that strategy simply
didn' t succeed versus accurate play, so White
finally turned to a different concept. Pieces can
precede pawns in an attack as long as the two ul
timately cooperate. The great Tai always seemed
to bring his pieces out to active squares before
organizing pawn-breaks, if indeed his opponent
survived up to that point.
10 i.e7 11 e5! (D)

••• Or 1 9 . . . 'tl¥e8 20 lbxf8 'ti¥xf8 2 1 l:i.d7+ @g8
and one nice win is 22 lbe4 ! h6 (22. . .i.xe4 23
'tl¥b3+) 23 lbxf6+ 't\Yxf6 24 .i:i.ee7 ! 't\Yg5+ 25
't\Yxg5 hxg5 26 l:i.xg7+ @f8 27 .i:i.df7+ @e8 28
.i:i.xb7, etc.
20 lbg5+! @g8 (D)
20. . . i.xg5+ 2 1 'tl¥xg5 is resignable.
The introduction to a fantastic pawn sacri

fice. Before White's idea had always been 't\Yg3.
11 i.b7 12 't\Yg3! lbxe5
..•
1 2. . . dxe5 1 3 i.xe6 fxe6 14 lbxe6 'tl¥c6 1 5

lbxg7+ @f7 has also been tried.
13 i.xe6! fxe6
1 3 ... 0-0 would bail out. As always, it's very 21 .i:i.xf6 gxf6 22 lbge4+
hard to assess things. One line might be 14 i.xf6 Or 22 l:i.e7 ! lbg6 23 lbe6 and l:i.g7# follows.
..ltxf6 1 5 lbd5 'ti¥d8 1 6 lLixf6+ 't\Yxf6 17 i.f5. 22 lbg6 23 fxg6 1-0
•••
14 f4!
The attack peters out after 14 lbxe6 'ti¥d7 ! 1 5 A similar and wild example of putting devel
lbxg7+ @f7. opment first is seen in the following game:
14 lbg6? !
•.•
14 ...lbc4 is the main test, when 1 5 lbxe6 Michalek - Fedorchuk

't\Ya5 ! ? l 6 lbxg7 + @f7 1 7 .i:i.he 1 needs help from Plzen 2003
a combination of computers and imagination.
15 lbxe6 'ti¥d7 16 .i:i.hel! @f7 17 f5! (D) 8 ..ltg5 i.e7 9 't\Yf3 't\Yc7 10 0-0-0
Now we have the same position as in the
Christiansen game, but with a bishop on e7 in
stead of a knight on d7.
10 ... b4!? 11 e5! (D)
White has just a pawn for his piece, but

Black can only watch as his position collapses.
17...lLifS 18 ..itxf6! ..itxf6 19 l:i.xd6 'tl¥c8
1 1•••..ltb7?!
a) Typical tactics follow l l .. .bxc3? 1 2 exf6 !
.Itb7 1 3 lllxe6 ! fxe6 14 li'h5+ g6 1 5 li'h3 !
cxb2+ 1 6 @xb2 and Black is getting killed.
b) But sacrificing an exchange by 1 l . . . dxe5 !
is also typical. For instance, 1 2 li'xa8 ( 1 2 ..ltxf6
might improve) 1 2 . . .exd4 1 3 l:.xd4 ! bxc3 1 4
l:.c4 cxb2+ 1 5 @bl ..ltc5 1 6 ..ltf4 e5 1 7 l:i.el ( 1 7
..lta4+ may be better) 1 7 . . .0-0 1 8 l:.xe5 ..ltb7 1 9
l:i.exc5 li'xf4 ! 20 l:.xf4 ..ltxa8 2 1 f3 ll'lbd7 22
l:.c7 ll'lb6 and White has only a minimal advan
tage.
12 exd6 ..ltxd6 13 li'h3 0-0
1 3 . .. bxc3 14 lllxe6 fxe6?! 1 5 li'xe6+ li'e7 1 6
l:.xd6 ! . consistently with great success was Karpov, yet
14 ..itxf6 bxc3 15 li'g4 (D) practically every major player has been on one
or both sides of it. Kasparov played it at least
four times versus major players with an idea
that will be seen below. White' s concept is sim
B ple, at least at first sight. He wants to develop
and get castled without exposing his pieces to
the tempo-gaining attacks that 6 ..ltg5, 6 ..ltc4
and 6 ..lte3 are often hit with. 6 ..lte2 also covers
the g4-square against an invading knight and
thus prepares to put a bishop on e3 . While 6
..lte2 is almost always associated with f4, the
advance g4 has increasingly been used in con
junction with it in order to drive away the f6-
knight and prevent . . . d5 before undertaking
more aggressive action.
The negative side of ..lte2 is fairly obvious: it
15 ..ltf4+
.•• is passive and creates no threats. Nor does the
Or 1 5 . . . g6 1 6 ll'lxe6. bishop protect the critical e-pawn, which indi
16 @bl ..lth6 17 lllxe6! cates that it will most likely end up on f3 or d3
Unleashing a devastating series of tactics. at some point. Consequently, White's bishop
17 fxe6 18 li'xe6+ @h8 19 ..lte5 li'a5 20
••• will often take two moves to get to a relatively
..ltxc3 li'c5 21 ..itd4 li'c6 22 li'e7 l:.e8 passive square.
Just as bad are 22 . . .l:.c8 23 ..ltxg7+ ..ltxg7 24 6...e5 (D)
l:.d8+ and 22 . . .ll'ld7 23 l:.he 1 ! . Although Black can play 6 . . .e6 and trans
23 ..ltxg7+! ..ltxg7 24 .i:i.d8 ll'ld7 25 li'xe8+ pose into another variation, this is the original
lllf8 26 li'f7! 1-0 'point' of 5 . . . a6. On the move before, 5 . . . e5
would have been met by 6 ..ltb5+, creating some
awkwardness on the light squares; for example,
Classical 6 i.e2 System
6 . . ...ltd7 7 ..ltxd7+ li'xd7 8 ll'lf5, after which the
1 e4 c5 2 ll'lf3 d6 3 d4 cxd4 4 ll'lxd4 ll'lf6 5 lllc3 knight will head for e3 in many situations, al
a6 6 ..lte2 (D) ready with a complete grip on d5. None of this
As various systems have come in and out of can occur once . . .a6 is in. The move 6 . . . e5 sets
fashion, this solid and unpretentious develop up one of the archetypal Sicilian structures.
ment has always been there as a sensible alterna Black's idea will be to threaten . . . d5 as soon as
tive to the heavily theoretical attacking sys possible and force White to react in a way that
tems. The last world-class player to play it is otherwise unfavourable. The analogous idea
SICIUAN DEFENCE 209
is Boleslavsky' s innovative 6 . . .e5 after 5 . . .lbc6 b) 7 lbf3 is played reasonably often. White
6 ..lte2 e5, a move that at first shocked the chess sometimes follows with the sequence of moves
world because it gave up an outpost on the cru a4, 0-0 and lLid2-c4-e3, to reinforce control of
cial d5-square and also created a backward d5, but that is obviously very slow. An exciting if
pawn on d6. Boleslavsky' s move is analysed in speculative game continued 7 ... h6 (a good solu
the section 'Sozin Attack (and the Classical Si tion is 7 . . . ..lte7 ! 8 ..ltg5 lbbd7 9 a4 0-0 1 0 0-0 h6
cilian)' below. Note, by the way, that after 6 1 1 ..ltxf6 lbxf6 1 2 ..ltc4 ..lte6, Van der Wiel
..ltg5, 6 ... e5? would be a self-pin; and after 6 Beliavsky, Wijk aan Zee 1 985) 8 ..ltc4 ! ? (D) .
..ltc4, 6 . . .e5 fails to block off White ' s dangerous
bishop. On the other hand 6 ..lte3 e5 is very
common.
Now:
b l ) The natural 8 . . .b5? ! 9 ..itd5 lbxd5 1 0
lbxd5 ..ltb7 runs into 1 1 a4 ! . One of the first
I'll fit the variation 6 ..lte2 e5 into one game; things to know about the ..lte2 system is that
please forgive the dense notes, which attempt to Black must be careful about . . . b5, which can be
encompass the major ideas of the variation. a weakening move. Obviously that doesn't ap
ply to other Najdorf systems in which White
Geller - Fischer castles queenside.
Cura(:aO Ct 1962 b2) 8 ... ..lte6 ! ? 9 ..ltxe6 fxe6 (we've arrived at
that central doubled-pawn structure again - it
7 lbb3 covers all the central squares but generally
Although knights on b3 are often poorly lacks mobility; this would be equal except for
placed in the Sicilian, this retreat leaves the White's tactical ideas) 10 lbh4 ! ( 1 0 0-0 lbc6)
move f4 available to attack White's centre and 10 . . . lbc6?! ( 1 0 . . .�f7) 1 1 lbg6 ( 1 1 f4 ! was an
kingside. It also supports the idea a4-a5, and opportunity missed) l l . ..l:i.g8 1 2 0-0 �f7 1 3
has a defensive function by keeping an eye on lbxf8 l:i.xf8 1 4 f4 �g8 1 5 ..lte3 ( 1 5 f5 d5 !)
c5 and potentially exchanging a knight on that 1 5 . . .exf4 1 6 l:i.xf4 'fic7 17 'fie2 lLie5 (all at once
square. We shall see that White's action in the 6 Black has the piece placement he wants: e5 for
..lte2 N ajdorf is very often on the queenside, in his knight and no outpost on d5 for White's) 1 8
contrast to his main 6th-move alternatives. ..ltd4 l:i.f7 1 9 l:i.d 1 .i:i.af8 20 �h 1 'fic4 ! 2 1 'iid2
a) 7 lbf5 d5 ! exploits the white knight's b5 22 a3 'fic6 ! 23 ..ltxe5 dxe5 24 .i:i.f3 lbxe4 25
hanging position to achieve Black's favourite lbxe4 'fixe4 and in Van der Wiel-Portisch, Til
freeing move. White can develop quickly and burg 1984, the passed extra pawn was enough
control d5 by 8 ..ltg5, but 8 . . . d4 9 ..ltxf6 'fixf6 1 0 to win.
lbd5 'iid 8 gives Black a space advantage with 7 ..lte7
..•
easy development for Black's bishop-pair. He Black will sometimes aim for an immediate
also has a handy break with . . .g6 and . . .f5 in . . . e5 by means of 7 . . . ..lte6, but that is asking for
store. f4-f5 ; e.g., 8 f4 'fic7 (the difference between
this and normal lines is that White is able to an

swer 8 . . .exf4 with 9 ..itxf4 in one move, as op
posed to having to play ..lte3 first) and now:
a) It's increasingly popular to push the g
pawn in all Sicilian variations but here's an
older example: 9 g4! ? h6 (9 . . . exf4 1 0 g5 ! lLifd7
1 1 ..ltxf4 lbc6 1 2 'iid2 ..lte7 1 3 0-0-0 lbce5 1 4
lbd4 with a big advantage; White i s already set
up for lLif5) 10 g5 hxg5 1 1 fxg5 lLifd7 (D).
lbd5 ..ltg5 1 2 0-0 lbe7 !, when the pieces are

coming off while Black retains his two bishops
and an advantage, Amason-Kasparov, Dort
mund jr Wch 1 980.
8 ...0-0
8 . . . ..lte6 9 f4 'fic7 1 0 a4 ( 1 0 f5 ! ?) 1 0. . .lbbd7
1 1 ..lte3 0-0 1 2 'lt>h l exf4 1 3 l:i.xf4 (White tries a
different idea; he's not too worried about . . .lbe5
and would rather aim the e3-bishop at the
queenside, where a5 and lbd5 may be influen
1 2 ..ltg4! l:i.h4 1 3 ..ltxe6 fxe6 14 ..lte3 ..lte7 1 5 tial) 1 3 . . . lbe5 (now we'll get a particularly in
'fif3 lbc6 1 6 'iig 3 .i:i.h8 1 7 'iig4. Black i s tied structive game, especially with regard to piece
down, and h4-h5 can follow, D.Gurevich-Bala placement in typical pawn-structures) 1 4 lbd5
shov, USSR 1 974. ..ltxd5 15 exd5 lbfd7 1 6 l:i.b4 l:.fe8 ! 17 a5 ..itf6
b) 9 0-0 lbbd7 1 0 f5 ..ltc4 1 1 a4 ! (prevent (D).
ing . . . b5, and planning a5 in order to restrict
Black's queenside) 1 1 . . ...lte7 1 2 ..lte3 0-0 1 3 a5
b5 14 axb6 lbxb6 (fine, but now Black has an
isolated a-pawn in one of those exceptional po
sitions where he has insufficient counterplay
down the b-file) 1 5 @h l .i:i.fc8 1 6 ..ltxb6! 'fixb6
1 7 ..ltxc4 l:i.xc4 1 8 'fie2 .i:i.b4 1 9 l:i.a2 (White's
manoeuvre has given him control of d5 and a
useful open a-file; note that this rook protects
b2) 1 9 . . .h6 ( 1 9 . . .'fib7 20 .i:i.e l ) 20 .i:i.fal ..itf8 2 1
l:i.a4 ! (21 l:i.xa6?! l:i.xa6 2 2 l:i.xa6 'iib7 { hitting
e4 } 23 lba5 'fic7 is equal) 2 1 .. .l:i.c8 22 l:i.xb4
'fixb4 23 'fixa6. White is a clear pawn ahead,
Karpov-Bronstein, Moscow 1 97 1 . A model
treatment.
We now return to 7 . . . ..lte7 (D): It's very hard to break down a structure like
8 0-0 Black's when there's a pawn on d5 and when
Again 8 g4 has been played, as well as 8 ..lte3 Black is able to use his strongpoint on e5 as a
..lte6 9 lbd5. But Black will have plenty of stepping stone. In this game, White drifts and
counterplay if White rushes to exchange his Black takes over the initiative: 1 8 ..ltgl ..ltg5 1 9
dark-squared bishop for the sake of controlling lbd2 lbf6 20 lLifl g6 2 1 l:i.d4?! l:i.e7 2 2 c4 l:i.ae8
d5 : 8 ..ltg5 ..lte6 9 ..ltxf6 ! ? ..ltxf6 1 0 'iid3 lbc6 1 1 23 b4, Hulak-Portisch, Indonesia 1 983, and now
S!CIUAN DEFENCE 211
the thematic 23 . . .lbed7 ! controls all the key 1 7 lLil a2 lbc5 1 8 lbb4 'ilie8 1 9 g3 ! l:i.c7 20 ..ltg2
squares. .i:i.dc8 21 b3 ..lte6 22 lbcd5 lbxd5 23 lbxd5
9 ..lte3 ..ltxd5 24 .i:i.xd5. White has control of d5 and the
9 @hl has been played by Kasparov on oc two bishops, Karpov-Nunn, Amsterdam 1 985.
casion. It's a move that White will want to make 10 a4 ..lte6 1 1 a5 lbbd7 12 lbd5 lbxd5! ?
anyway, and then wait to see how Black is com Black i s trying to save the bishop-pair.
mitting his pieces, but that may not be too help 13 exd5 ..itf5 14 c4 ..ltg6 15 .i:i.cl lbc5?!
ful: White has the advantage in any case, but
a) 9 . . . b6 ! ? (this is the accepted solution, 15 .. .f5 1 6 c5 ! ? (or 1 6 f4) 1 6 .. .f4 1 7 cxd6 'ilixd6
avoiding 9 . . .b5 1 0 a4 !) 1 0 ..lte3 ..ltb7 1 1 f3 b5 ! 1 8 ..ltc5 lbxc5 l 9 lbxc5 ..ltf7 ! 20 ..itf3 .i:i.fb8 ! and
1 2 a4 b4 1 3 lbd5 lbxd5 1 4 exd5 lbd7 1 5 c3 . . . b5 doesn't look too bad.
bxc3 16 bxc3 ..ltg5 ! 1 7 ..ltgl 'ilic7 1 8 c4 a5 1 9 16 lbxc5 dxc5 17 b4! (D)
lbd2 f5 1h-1h Anand-Gelfand, Dos Hermanas
1 997 . Black has secured the c5 outpost and he
already has his kingside majority.
b) 9 ... lbc6 1 0 f3 ..lte6 is also fine; for exam
ple, 1 1 lLid5 a5 1 2 ..lte3 a4 1 3 lbc l ..ltxd5 ! ? 1 4
exd5 lbd4 ! (D).
Geller's opening strategy has resulted in a

textbook position.
17....i:i.acS
The idea is 1 7 ... cxb4 1 8 ..ltb6 with c5 next.
18 'ilib3 ..itd6 19 .i:i.fdl
White can also secure two passed pawns by
This pawn sacrifice turns Black's f6-bishop 1 9 bxc5 ..ltxc5 20 ..ltxc5 'ilixc5 2 1 'ilixb7 'ilixa5
into a powerful piece while White's on e2 re 22 'ilib2 ! .
mains passive: 1 5 ..ltxd4 exd4 1 6 'ilixd4 'ilia5 1 7 19...'ilie7 20 bxc5 ..ltxc5 2 1 ..ltxc5 l::txc5 22
.i:i.d l lbd7 1 8 l::tb l .i:i.fe8 with plenty of play for l::tal ! .i:i.d8 23 l:i.a4 ..trs 24 l:i.b4 ..ltcs (D)
the pawn, which may very well have to be re
turned anyway, Adams-Kariakin, Wijk aan Zee
2006.
9...'ilic7
9 . . . ..lte6 1 0 'ilid2 lLibd7 1 1 a4 l::tc 8 ( l l . ..lLib6
generated active play following 1 2 a5 lbc4 1 3
.txc4 ..ltxc4 1 4 .i:i.fd l l::tc 8 1 5 lLic l d5 ! 1 6 ..ltb6
{ 1 6 exd5 ..ltb4! } 1 6 ...'ilie8 1 7 exd5 ..ltb4 1 8 d6
'iVd7 1 9 lbd3 'ilixd6 20 lbxb4 'ilixb4 21 lbe4
'ilixd2 22 lbxf6+ gxf6 23 .i:i.xd2 ..lte6, with equal
ity, in Leko-Shirov, Dortmund (2) 2002) 1 2 a5
'ilic7 1 3 .i:i.fd l .i:i.fd8 14 'iliel 'ilic6 1 5 ..itf3 ..ltc4
1 6 lLic l ! (heading for b4; Karpov typically con
centrates his pieces on the weak point) 1 6 . . . h6
25 �b6! �d6 allowing . . . lbg4. However, Black does have

25 . . .�xa5 26 d6 1Wd7 27 ii.f3 leaves Black the move 6 . . .1Wh6 ! ? preventing 7 ii.e3 due to
short of reasonable moves. 7 . . . 1Wxb2. This has the same idea as ... 1Wh6 in
26 1Wb4 1Wc7 27 �xd6 1Wxd6 28 �bl 1Wc7 29 the Classical Sicilian, namely, to force the
1Wa4! ii.d7 30 1Wa3 �xa5 31 �xb7! 1Wxb7 32 knight back to b3 even at the cost of a tempo
1Wxa5 g6 33 h3 ( . . . 1Wh6-c7). Then if the knight returns to its
White is getting ready to push the passed 'best' square d4, it' s Black who has gained the
pawns. tempo. But with the knight remaining on b3 it' s
33 1Whl+ 34 '1t>h2 ii.rs 35 1Wc3 1We4 36
.•• not s o easy for Black; e.g., 7 lbb3 e 6 8 g4! ? ( 8
ii.f3 1Wd4 37 1Wxd4 exd4 38 g4 ii.cs 39 c5 a5 1We2 intending ii.e3 i s also played; perhaps 8
40 c6 'lt>f8 41 d6 1-0 a4 is also good, since a5 can't be prevented)
The game might finish with 4 l . . .'lt>e8 42 8 . . . lbc6 (Judit Polgar has played both 8 . . . 1Wc7
ii.d i ii.a6 43 g5 ilb5 44 c7 ii.d7 45 ii.a4, etc. 9 ii.e3 b5 and 8 . . . tt:Jfd7; these both look like
Geller was one of the great 6 ii.e2 players, and better ways to go) 9 'iie2 1Wc7 I O ii.e3 b5 l l
of course Fischer was the premier Najdorf 0-0-0 with advantage. In view of White ' s many
player of his time. options against 6 . . .1Wb6, it looks as though 6 f3
is safe enough and avoids the ... lbg4 lines men
English Attack tioned in the next note.
6 e6
•••
1 e4 c5 2 lbf3 d6 3 d4 cxd4 4 tt:Jxd4 lbf6 5 lbc3 Or:

a6 6 ii.e3 (D) a) 6 . . . tt:Jg4 7 ilg5 h6 8 ii.h4 g5 9 ilg3 ilg7
(D) has been the subject of many grandmaster
games, notably Kasparov's.
This move in conjunction with 7 f3 is known

as the English Attack, which can be used against
systems with or without . . .lbc6. Here we look at The idea is to take White's dark-squared
Najdorf variations, primarily those that use ... e6 bishop away from its most effective diagonal
and skip . . . lbc6 in favour of moves like . . . b5, and use the e5-square productively. Still, Black
... ii.b7, ... ii.e7 and ... 'iic 7. This is a hot line in has weakened his kingside and the variation
contemporary chess and full of analysis going seems to have fallen out of favour, so we won't
20 moves or more, so my coverage will be lim be looking into it here.
ited. Nevertheless, the English Attack lines are b) 6 . . . e5 is the typical Najdorf solution that
full of interesting and original positional ideas we saw under 6 ii.e2 e5. A unique idea is 7
that express a new way of playing the Sicilian lbf3 ! ? ii.e7 8 ii.c4, which resembles 6 ii.e2 e5
Defence for both sides. These positional con except for three things:
siderations make it a good topic of study. 1 ) White generally would like to play ilg5
The move-order 6 f3 with 7 ii.e3 is a way of in lines with lbf3, so as to weaken Black's con
transposing to the main English Attack without trol of d5. But here White has already moved
the bishop has to e3, so it would be a loss of

tempo to bring it to g5 .
2) White has gained a move by playing ii.c4 B
in one jump (instead of ii.e2-c4). Of course
White may not want the bishop to be exposed
so early to . . . b5, but that doesn't seem to be
much of a problem.
3) On a less important note, White's some
times-useful manoeuvre of lbd2-c4-e3 (after
a4) is no longer possible because both c4 and e3
are occupied.
At any rate, after 8 ii.c4 play can continue
8 . . . 0-0 9 0-0 ii.e6 ! ? 1 0 ii.b3 lbc6 1 1 ii.g5 (D).
time to get all of three moves in, that is, . . .b5,

... lbfd7 and . . .lbb6. Now 8 ... lbb6 9 g5 lbfd7
blocks ... b5, when both 1 0 a4 lbe5 1 1 f4 lbec4
1 2 ii.c l and 1 0 f4 (preventing . . . lbe5) yield
considerable advantages. After 8 g4, therefore,
Black may as well play 8 . . . h6 9 h4 (9 "fiid2 b5 !
gives Black the extra time he needs for a pro
ductive transfer to the queenside and in fact
transposes into the 'Main Line' of the English
Attack, but 9 "fiie2 ! ? and 0-0-0 is definitely
worth looking into) 9 . . .b5 10 �gl ( 1 0 a4 ! is
strong, with the idea 1 0 . . .b4 1 1 lbc6 "fiic7 1 2
tt:Jxb4 d5 1 3 lbd3) 1 0 . . . lbb6 (so Black has
gained his tempo but at the cost ofloosening his
White has excellent control of d5 now, but position on both wings) 1 1 g5 lbfd7 (Wed
having that one square at your disposal isn't suf berg-Akesson, brebro 2000) and again 1 2 a4 !
ficient for a significant advantage in the Sicilian looks strong, intending 1 2. . .lbc4 1 3 ii.c l ! , when
Defence. Anand-Leko, Wijk aan Zee 2006 con Black has to do something about axb5, and
tinued l l . . .lbd7 l (Black gets rid of his bad White can respond to 1 3 . . . "fiia5 with 1 4 g6!
bishop) 1 2 ii.xe7 "fiixe7 1 3 tt:Jd5 "fiid 8 14 c3 lbde5 15 gxf7+ lbxf7 16 axb5 ! "fiix al 1 7 ii.xc4.
lba5 ! 1 5 �el �c8 1 6 h3 (this prepares �e2-d2 In this entire subvariation we see the problems
without allowing . . .ilg4) 1 6 . . .lbb6 17 lbxb6 with playing . . . lbbd7 and blocking Black's re
"fiixb6 1 8 ii.xe6 fxe6 1 9 �e2! �c6 20 "fiid3 "fiic7 treat . . . tt:Jfd7 .
2 1 �dl . White has some pressure, but Black b) 7 . . . h5 ! ? (D) is a positional theme to re
isn't yet in serious trouble. member, since it prevents g4, which is White's
7 f3 b5 main idea in the English Attack.
Little-played alternatives are usually reveal Of course this comes at a cost, both in terms
ing, and here we have a couple of ideas to con of weakening the kingside and time, but in
sider: some Sicilian positions that's worth it (notably
a) 7 ... lbbd7? ! is the most natural continua the Classical lines with . . . lbf6 and . . . lbc6).
tion for Najdorf players but they should under Here it's rather unclear: 8 "fiid 2 lbbd7 9 ii.c4
stand that it gets in the way when White pursues lbe5 ! ? (not 9. . . b5? 1 0 ii.xe6; but a plausible
his normal English Attack: 8 g4! (D). move is 9 . . .lbc5 ! ?, intending . . .b5 or even . . . d5)
Black would have liked to play ... b5 and 1 0 ii.b3 b5 1 1 0-0-0 ii.b7 1 2 ii.g5 ! (taking ad
. . .lbfd7-b6 (compare the main line below for vantage of . . . h5) l 2. . ."fiia5 ! ? ( 1 2. . . ii.e7 may be
an 'explanation' of this bizarre idea). Unfortu better) 1 3 'lt>bl (Khalifman-Van Wely, Wijk aan
nately, 7 . . . lbbd7 means that he doesn't have Zee 2002), and Fedorowicz suggests 1 3 . . . �c8,
development in this manner? The answers are

several. Consider that Black saves the move
. . . h6 (played in the main line), thereby 'gain
ing' a tempo and, crucially, keeping his king
side without weaknesses. Thus White will have
to play a lot more moves (such as g5, h4, h5 and
g6) in order to make contact with the enemy
king (which normally castles kingside). Sec
ondly , if attacked by . . . b4, White will not be
able to play ll:ia4 as he does in many lines. This
can speed up Black' s attack, especially since
ll:ice2 can be met by . . . ll:ic4. So White is re
duced to ll:ib 1 in most cases. To be fair, ll:ib 1 is
a good enough answer in most cases but it' s not
limiting White to a moderate advantage. Even White's first choice. Finally, the . . . ll:ifd7-b6
if . . . h5 doesn't appeal to you in this exact posi manoeuvre allows for a very harmonious de
tion, you should be aware of it (as both Black velopment by Black involving . . . ii.b7, . . .ll:ibd7
and White) when playing or confronted with and . . . �c8. This is also one of the few lines in
the many different versions of the f3/g4 attack. which an early . . .d5 is feasible, because the usual
8 g4 problems of g5 and e5 (with tempo) are not
Although it's a complicated issue, it's proba present.
bly better to toss in g4 first, before 'Yi'd2, because On the flip side, White has five aggressively
after 8 'Yi'd2 ll:ibd7 Black has more options, placed pieces and a large space advantage, both
whereas 8 g4 ll:ibd7 ?! 9 g5 drives the knight in the centre and on the kingside. His e4-pawn,
away. usually a target of attack in the Sicilian, is dou
8... h6 bly supported and not yet threatened. Imagine
For the reason given in the last note, this is being unhappy with that!
needed if Black wants to play . . . ll:ibd7. But Anyway, from the diagram, we have:
8 . . . ll:ifd7 ! ? 9 'Yi'd2 ll:ib6 (D) is an important and a) 1 0 'Yi'f2 and 1 0 f4 ! ? are both feasible.
still viable alternative. b) 10 a4 bxa4 1 1 ll:ixa4 ll:ixa4 1 2 �xa4 (D)
is an important yet funny line.
As with so many Sicilian ideas, this is ini

tially hard to believe: Black has made nine Now Black has no space, no pieces out after
moves to get one piece out! And yet 8 . . . ll:ifd7 1 2 moves, and his a-pawn is isolated ! This posi
and 9 . . . ll:ib6 has been played by many of the tion is a tribute to the central pawn-majority and
world's leading players including Kasparov. the . . . d6/. . . e6 structure. It also provides evidence
How can Black ignore the classical rules of for a recurrent idea: that an isolated a-pawn on
an open file normally isn ' t a serious problem

until the ending. 1 2 . . . ii.e7 and now:
b l ) 1 3 ii.e2 0-0 14 0-0 ii.b7 1 5 Ilfal lbd7
1 6 lbb3 (really stacking up on the queenside in
order to win that a-pawn; White finally wins it
but all his forces are so diverted by that task that
he leaves the rest of the board open) 1 6 . . . l:rb8 ! ?
1 7 ii.a7 l:rc8 1 8 tt:Ja5 ii.a8 1 9 ii.xa6 lbe5 ! 20
ii.e2 f5 ! with obvious compensation for Black,
Anand-Kasparov, Kopavogur (rapid) 2000.
b2) Anand tried to improve with 1 3 g5 ver
sus Topalov in Wijk aan Zee 2004, when Dear
ing suggests 1 3 . . . ii.b7 ( 1 3 . . . 0-0 1 4 h4 ! was
played, when h5 followed by g6 is a problem)
14 ii.e2 d5 1 5 e5 ltJd7 1 6 f4 lbb6 1 7 l:ra2 lbc4. From this point we'll follow a relatively re
Black appears to stand reasonably well. cent game.
c) 1 0 0-0-0 ltJ8d7 1 1 °Yi'f2 ( 1 1 lbcxb5 ! ? axb5
1 2 lbxb5 is a wild sacrifice that is currently un Anand - Kasimdzhanov
der a cloud) l 1 . . .ii.b7 1 2 ii.d3 l:rc8 1 3 'lt>bl ? ! Leon (rapid) 2005
( 1 3 lbce2 i s the main line, when Black can de
lay castling to get something going in the centre 9 tt:Jbd7 10 0-0-0 ii.b7 11 h4 b4 12 lba4
•••
by 1 3 . . . d5, 1 3 . . . "Yi'c7 1 4 'lt>bl d5, or Kasparov's °Yi'a5 (D)

1 3 . . . lbc5) 1 3 . . .l:rxc3 ! (another instance of the After this comes a long sequence of theoreti
positional exchange sacrifice . . . l:rxc3; it is cal moves. 1 2. . . d5 ! ? is a fascinating but very
played in other Sicilians, notably the Dragon) risky alternative.
1 4 bxc3 (D).
13 b3 ltJcS 14 a3 l:rc8 15 °Yi'xb4

Black's compensation is obvious with moves 15 axb4 lbxb3+ 1 6 lbxb3 "Yi'xa4 is the Main,
like . . . lba4, . . . °Yi'a5 or . . . "Yi'c7, . . . tt:Je5, etc., in the Main Line! The games and analysis are fasci
air. What's worse, White can't undertake any nating, but extend beyond 30 moves at points
thing useful, since as so often his rooks will be and are decided by details that don't have much
fairly useless until an ending, which probably to do with chess understanding. So I'll go with
isn ' t going to happen ! 1 4. . . °Yi'c7 (or 14 . . . lba4 ! ) something cleaner:
1 5 lbe2 ii.e7 1 6 g 5 0-0 1 7 h4 lba4 and Black's 15 °Yi'c7 16 'lt>bl tt:Jfd7 17 lbb2 d5 18 °Yi'd2
•..
attack was too powerful (even . . . d5 followed) in dxe4

Movsesian-Kasparov, Sarajevo 2000. 1 8 . . . tt:Je5 ! ? has also been played.
9 °Yi'd2 (D) 19 f4 lbf6 (D)
20 ii.e2! not have meant much, but ilb5 systems are in
Credit this ' ! ' to Anand. He also mentions 20 creasingly popular, and 2 tt:Jf3 lbc6 3 ii.b5 has
l:tgl . Dearing has analysed 20 ii.h3 lbd5 21 M ! even driven top-level grandmasters to change
out to a wonderful position, although my ana their preferred variations or at least their move
lytical engines produce absurd-looking things orders. Another benefit has to do with the f8-
like 2 1 . ..ii.d6 ! ? 22 bxc5 1Wxc5 with the idea 23 bishop, which after 3 d4 cxd4 4 tt:Jxd4 is now
c4 •xa3 24 lbc2 �3 25 lbal lbc3+, leading free to go to various positions such as c5 and
to a repetition. M; both carry the prospect of more confronta
20 ltJdS 21 lbc4 tt:Jd7?!
••• tional chess than, say, 2 . . . d6 offers. We also see
Anand claims a small advantage for White a lot of early queen moves; for example, to c7
after 2 1 ...ile7 22 g5, and leaves 2 1 .. .�d8 with and b6 without first playing . . . d6.
out comment. The opening is well past, so let's Needless to say, 2 . . . e6 comes with some
just visually enjoy the rest. negatives. On a smaller scale, Black has less
22 gS! tt:Jxe3 23 1Wxe3 ii.dS 24 �hfl ii.cs 2S flexibility in meeting the moves 3 c3 and 3 d3.
'ji'c3? hxgS 26 ltJfS! ii.xc4? 27 lbxg7+ @e7 28 It should be added that these moves pose no
�xc4 �hg8 29 hxgS e3 30 rs ltJeS 31 fxe6 serious threat; however, Black may not get to
l:txg7 32 �d7+! lbxd7 33 1Wxg7 1-0 choose the variation with which he is most
comfortable (see below). And ... e6 does weaken
the d6-square, which is a drawback in a num
I ntroduction to Systems ber of lines, especially those in which Black
delays . . . d6. Moves such as lbb5 and ii.f4 can
with 2 e6 . . .
be problems, and in general White's move e5
can have more force in many positions since it
1 e4 c5 2 lbf3 e6 (D) can't be captured by a pawn.
This advance of the e-pawn caught the atten Oddly enough, the fact that 2 ... e6 cuts off the
tion of many early practitioners of the Sicilian path of the c8-bishop isn't of great consequence.
Defence. Black threatens to challenge, if not Normally that bishop will attempt to go to b7 or
take over, the centre by playing ... d5 next or if necessary take its place on d7, and these are
within a few moves. The game as a whole takes the usual squares in other Sicilian variations as
on a different character with 2 . . . e6 as opposed well. Taken as a whole, 2... e6 is neither better
to 2 . . . d6 or 2 . . . lbc6. Naturally, it can transpose nor worse than the alternatives, as can be seen
to the same lines and structures if an early . . . d6 from its percentage scores in various lines.
follows; but if not, Black has new options with 3 d4
respect to his development and overall strategy. The alternatives are not threatening but both
One noteworthy difference with 2. . . e6 is that sides might want to look into 3 c3 and 3 d3.
White has no ii.b5 option, as he does after These moves are good study material in any case
2 . . . lbc6 and 2. . . d6. A few years back that might because the positions are of a standard nature:
a) After 3 c3, Black has to decide which down the b- and ct-files. In this position he can
anti-c3 method to choose. It's important to know also liquidate the weaknesses and gain activity;
something about the move-orders, especially for example, 1 1 0-0 ii.e7 1 2 tt'ic3 0-0 1 3 ii.c2
when compared to 1 e4 c5 2 c3, which is cov ii.a6 14 �el c5 1 5 dxc5 °Yi'xd l 1 6 �xdl ii.xc5,
ered in the Alapin section of this chapter. A big Blatny-Shaked, Kona 1 998.
difference is that after 2 c3, Black can play a2) The other obvious response to 3 c3 is
2 . . . d5 3 exd5 °Yi'xd5 4 d4 tt'if6 5 tt'if3 ilg4, a 3 . . . d5, when 4 exd5 can lead to two unrelated
move that is no longer available when he plays set-ups:
2 tt'if3 e6 3 c3 d5 4 exd5 °Yi'xd5 5 d4. Further a2 1 ) Upon 4 . . . °Yi'xd5, we might get 5 d4 tt'if6
more, in the main lines after 2 c3 tt'if6 3 e5 tt'id5 6 ii.d3 (6 ii.e2 tt'ic6 7 ii.e3 cxd4 8 cxd4 ii.e7 9
4 d4 cxd4 5 cxd4, Black retains the option of tt'ic3 °Yi'd6 transposes to one of the lines stem
. . . d6 without . . . e6. That isn't true after 2 tt'if3 e6 ming from 2 c3; it is considered harmless)
3 c3 tt'if6 4 e5 tt'id5. Thus Black needs to oper 6. . . tt'ic6 7 ii.e3 cxd4 8 cxd4 ii.e7 9 tt'ic3 °Yi'd6
ate within a narrower range of systems, which with analogous ideas to lines versus 2 c3; again,
have to be studied if one is to gain real under refer to the Alapin section. Of course there are
standing. I ' ll pursue just a few themes out of options on every move.
many: a22) Black can also play 4. . . exd5 5 d4 tt'ic6,
a l ) Several basic structures can arise from when an isolated queen's pawn position can
3 . . . tt'if6 4 e5 tt'id5 5 d4 cxd4 6 cxd4 d6, which easily follow:
Black has played with adequate results for many a22 l ) 6 ilb5 ii.d6 7 dxc5 ii.xc5 8 0-0 tt'ige7
years. One idea is that he can forego the devel 9 tt'ibd2 0-0 10 tt'ib3 ii.d6 (D).
opment of his queen's knight until White's for
mation is clear; e.g., 7 ilc4 tt'ib6 and now 8
ii.d3 ! ? dxe5 9 dxe5 tt'ia6 ! ? 1 0 0-0 tt'ic5 1 1 ii.c2
°Yi'xd l 1 2 �xdl ii.d7 1 3 tt'ic3 l::k 8 with comfort
able development, Shaw-Short, Catalan Bay
2003. If White plays 8 ii.b3 instead, Black has
8 . . . dxe5 9 tt'ixe5 (9 dxe5 °Yi'xd 1 + 1 0 ii.xd 1 tt'ia6 !
and . . . tt'ic5 or . . . tt'ib4, another case in which de
laying the development of the queen's knight is
beneficial) 9 . . . tt'ic6 1 0 tt'ixc6 bxc6 (D).
We've transposed to the French Defence

variation 1 e4 e6 2 d4 d5 3 tt'id2 c5 4 exd5 exd5
5 ilb5+ tt'ic6 6 tt'igf3 ii.d6 7 dxc5 ii.xc5 8 tt'ib3
ii.d6 9 0-0 tt'ie7, except that White might not
want to play 1 0 c3 in that case; that is, in the Si
cilian 2 . . . e6 3 c3 version he has committed to
making that move before he might want to. This
is a rather sophisticated thing to worry about
for all but very experienced players; neverthe
less, it makes the position easier for Black to
Here we have a standard position from sev play than it usually would be, and might give
eral openings, with the backward c-pawn ver less-advanced players a feel for the consider
sus the isolated queen's pawn. Even if White ations that go into top-level opening play. At
were better developed than he is here, Black any rate, all the themes of isolated queen's
would have enough play by combining pressure pawns apply to the diagrammed position; for
instance, White blockades the d-pawn and seeks 2) Black will need to use an extra tempo if
appropriate simplification while Black uses his he wants to play . . . e5 . The implication is that
active pieces and freedom of movement to com Black would find . . .e5 a desirable move to make,
promise White' s position. Typical moves for which can be the case in lines with . . . g6, . . . ii.g7,
White are �e l , ii.g5-h4-g3, tt'ibd4, iVc2 and . . . tt'ige7 and . . . 0-0. This . . . d6/. . . e5 formation
ii.d3. Typical moves for Black are . . . ilg4, (called the Botvinnik structure) discourages
. . . tt'if5, . . . �e8, and . . .'iVb6 or . . .iVf6. Whether some practitioners of the King's Indian Attack.
you want to play this position for either colour In more specific terms, most players would
is a matter of taste. rather face the 'French Defence' set-up of . . .e6,
a222) You sometimes see the line 6 ii.e3, . . . d5 , . . . tt'ic6, . . . tt'if6 and . . . ii.e7 than others
when apart from 6 . . . cxd4 7 ii.xd4, Black has which do not involve the move . . . e6. However,
the interesting move 6 . . . c4. This is particularly the issues that I raise regarding reversed open
appropriate so as not to allow dxc5 and justify ings apply here. Those who are familiar with
the passive position of White's bishop on e3. the King's Indian Defence (which is the King's
White can't yet bring his bishop into active play Indian Attack with colours reversed) know that
on d3, and Black can develop effortlessly by some of the moves that Black might play in a
. . . ii.d6 and . . . tt'ige7 unless White does some King's Indian Defence don't work out as well in
thing right away. So there usually follows 7 b3 ! the King's Indian Attack, because Black hasn't
cxb3 8 axb3 ii.d6 9 ii.d3 tt'ige7 (D). committed to the position which makes those
moves effective. Here's an example: 3 . . . tt'ic6 4
g3 g6 5 ilg2 (in a paradoxical turnabout that
characterizes the flexibility of chess positions,
White can seek a radical change in the course of
the game by 5 d4 ! ?, moving his pawn a second
time but hoping to exploit of the weaknesses
created by . . . e6 and . . . g6; it turns out that there
are several good answers, including 5 . . . d5 ! ?
and 5 . . .cxd4 6 tt'ixd4 ii.g7 7 tt'ib5 d5 ! ?, a pro
ductive pawn sacrifice) 5 . . . ii.g7 6 0-0 tt'ige7 7
tt'ibd2 (the typical King's Indian Attack move)
7 . . . 0-0 8 �el d6 (or 8 . . . e5 ! ? 9 tt'ic4 d6) 9 c3 e5 !
(D).
Since the procedure c4 and tt'ic3 won' t really

break down Black's centre ( . . . ii.e6 or even
... tt'ib4 should do well to protect the d5-pawn),
Black can be happy with both his pawn-structure
and development. After I 0 iVc2 ( 1 0 0-0 ii.f5),
Adams-Nunn, Hastings I 996n continued I 0 ...h6
1 1 0-0 0-0 with equality. Nunn suggests the
more interesting sequence 1 O . . . ii.g4 1 1 tt'ibd2
�c8 1 2 iVbl ii.h5 intending . . . ii.g6 to exchange
White's good bishop on d3.
b) Some players believe that 3 d3 with a
King's Indian Attack set-up (g3, .ltg2 and 0-0)
is more appropriate against 2 tt'if3 e6 than
against either 2 tt'if3 tt'ic6 or 2 tt'if3 d6. There In some ways White's rook is misplaced on
are at least two ideas behind this assertion: el because it doesn't support the pawn-break f4
I ) Black's queen's bishop can't get out to an and is generally not useful against the Botvin
aggressive square. nik structure, which consists of . . . c5, . . . d6 and
... eS. Of course if the rook returns to fl , White variations, mostly old-fashioned and out of fa
is actually a tempo down on a King ' s Indian vour, so it is instructive to see how the players'
Defence position! Therefore White may well considerations differ from those in the conven
tum to the idea of queenside attack by a3 and b4 tional lines. The Four Knights itself is still play
with an interesting struggle ahead. able but I should warn you that it probably falls
Let's return to 2 lbf3 e6 3 d4: short of equality after 6 lbxc6 bxc6 7 eS lbd5 8
3...cxd4 4 lbxd4 (D) lbe4, a highly tactical line that has been thor
oughly tested in recent years. It makes sense
that a variation that targets the weak d6-square
would be a good weapon against this opening.
However, we'll follow another variation that
produces games with fundamental conflicts be
tween positional and tactical ideas.
Buchenthal - Rosen
German Cht 197819
6 lbdb5 ii.b4 (D)

At this point 6 ... d6 7 ii.f4 eS 8 ii.gs trans
poses into the Sveshnikov Variation, which is
more commonly arrived at via 1 e4 cS 2 lbf3
lbc6 3 d4 cxd4 4 tt:Jxd4 lbf6 S lbc3 eS 6 ltJdbS
After 4 lbxd4, Black has a number of op d6 7 ii.gs. With the first move-order, both sides
tions, from which I shall choose two basic strat have made an extra move due to . . . e6-eS and
egies: the Sicilian Four Knights Variation, and ilf4-gS. The Sveshnikov contains all kinds of
the Paulsenffaimanov complex. positional and tactical themes, but I felt it less
instructive over a broad range of strengths than
other variations of the Sicilian, so have not dis
Sic i l ian Fou r Kn ights cussed it in depth in this book.
4 tt:Jf6 5 tt:Jc3
.•.
Not S eS? 'Yi'a5+ and 6. . . 'Yi'xeS.

5...lbc6 (D)
7 tt:Jd6+
7 ilf4 leads to crazy tactics and lengthy the
ory after 7 . . .lbxe4 8 'Yi'f3 ! d5 9 lbc7+ 'lt>f8 1 0
0-0-0 ii.xc3 1 1 bxc3 gS, which ultimately yields
The Four Knights is a perfect example of a equal play according to the books and comput
Sicilian line that emphasizes development over ers. By contrast, a notoriously dull line for both
structure. That is true of only a couple of Sicilian sides is 7 a3 ii.xc3+ 8 tt:Jxc3 dS 9 exdS exdS 10
ii.d3 0-0 1 1 0-0. In spite of White' s two bish Budapest 1 995. Black has equality and perhaps
ops, Black is supposed to be able to reach more.
equality. Unfortunately, he may have to play 10 exd5 iVxd5 11 0-0 ii.xc3
some thankless defence in order to demonstrate l 1 . ..iVh5 ? ! 1 2 iVxh5 lbxh5 was played in
that. several old games, with activity and quick de
7...@e7 velopment pitted against the bishops. Maybe
The king may be subject to some attack White is a bit better, but not necessarily so, be
here, but it would definitely be a mistake to cause he still has to neutralize Black's posi
give up the dark squares by 7 . . . ii.xd6?. As it is, tional threats; e.g., 1 3 ii.d2 lbf6 (or 1 3 . . . �hd8;
Black ends up with a significant lead in devel the king is useful on e7) 14 a3 ii.d6 15 lbe4
opment. lbxe4 1 6 ii.xe4 f5 ! 1 7 ii.d3 lbe5 with equality,
8 lbxc8+ �xc8 9 ii.d3 (D) Keres-Trifunovic, Moscow 1 947 .
A case in point of how Black's development 12 bxc3 (D)
can outweigh other factors is 9 ii.d2 d5 ! (or
even 9 . . . ii.xc3 1 0 ii.xc3 lbxe4 1 1 ii.xg7 �g8)
10 exd5 tt:Jxd5 1 1 lbxd5+? ! ( 1 1 iVg4 ii.xc3 1 2
bxc3 iVd6) l l .. .iVxd5 1 2 ii.xb4+? ( 1 2 c 3 ilc5 B
favours Black) 1 2. . .lbxb4 1 3 iVxd5 lbxc2+ 1 4
@d2 exd5 1 5 �c l lbb4 and Black was a pawn
ahead in Sanz Calzada-Jordan Garcia, Cata
lunya Club 1 999.
A stark picture of knights versus bishops in

which it seems as though Black is swarming all
over his opponent's position. But White can
catch up quickly with the moves �bl , c4, and
ii.a3+ or ii.b2, so there is some urgency to act.
12 ...�hd8
12 . . . iVa5 ! looks more accurate, preventing
ii.a3+ and attacking c3.
A common decision in chess now arises: 13 �bl
does Black double the c-pawns and then protect 1 3 c4! ? iVa5 might lead to 14 ilb2 @f8 ! 1 5
his position by . . . d6, slowly exploiting the ii.xf6 gxf6, when White's weaknesses are more
weaknesses, as in the Nimzo-Indian Defence? important than Black's; for instance, 1 6 iVg4
Or does he emphasize space and rapid develop lbe5 1 7 iVh5 lbxc4 1 8 iVxh7 @e7 and Black's
ment, using his lead in those departments to king is completely safe. White might do best
force concessions from his opponent? to activate his pieces by the slightly odd ma
9 d5!
••• noeuvre 1 3 ii.a3+ @e8 1 4 iVbl ! .
Here the open-lines approach is more strik 13...�d7
ing. Nevertheless, 9 . . . .ltxc3+ 10 bxc3 �e8 1 1 Also possible is simply 1 3 . . . b6, with equal
.lta3+ d6 is also legitimate: 1 2 0-0?! ( 1 2 �bl ! ity.
'iVc7 1 3 0-0 �ed8 14 f4 is better; e.g., 1 4 . . . e5 1 5 14 ii.a3+ @e8 15 iVcl a6! ? 16 c4 iVh5 17
fxe5 lbxe5 1 6 �f5 iVxc3 with complications) f4? !
1 2 . . .@f8 ! ? (or 1 2 . . .iVa5 ! 1 3 ii.b4 iVc7) 1 3 iVe2 1 7 iVf4 looks better, with a highly unclear
@g8 14 �abl iVc7 1 5 �fd l d5, Major-Binder, situation.
17...ltJaS 18 c5 'Yi'd5 19 f5 lbc4 20 fxe6 fxe6

21 �dl (D)
prevented lbb5, lbf5 and lbc3-d5, he can wait

to see how White develops and then react ac
cordingly. Among other plans are expansion on
In this position Raetsky points out that the queenside by ...b5 and . . . ii.b7, queen devel
2 1 . . .'Yi'd4+ ! 22 'lt>h l lbe4 (perhaps 22 ... 'Yi'h4! is opment to b6 or c7 (again awaiting events). ac
even better) is fine. At any rate, this example of tive piece-play by ... ii.b4 and/or ... ll'lge7-g6.
Black's unusually rapid development in the Si and return to a conventional formation with
cilian shows that he can achieve equal chances ... d6 and . . .lbf6. His king ' s bishop in particular
in this traditional variation. can go to e7, d6, c5 or b4; it even finds its way
to g7 in some lines, with the move ... g6 creating
Apart from the Sicilian Four Knights, Black the travesty of four dark-square holes on Black's
has various means of setting up a structure that third rank.
includes ... a6 within the next few moves, but What about White? Let's think about those
delays . . . d6. The immediate 4 ... a6 (without a Sicilian knights in their customary positions on
very early ... lbc6 or . . . d6) is the Paulsen Sys c3 and d4. This is as good a place as any to talk
tem, also called the Kan Variation, whereas about their positive role in positional as well as
4 ... lbc6 followed by ... a6 on one of the next two attacking terms. Granted, these white knights
moves is usually referred to as the Taimanov are ideally restricted by Black's pawns on ab
Variation. Some of Black's ideas in these lines and e6 (the one on e6 being rock solid versus
are typical of the other Sicilian systems but direct fire by f4-f5). And if there were a knight
many are unique to the . . . e6/. . . a6 structure. on b3 it would merely aim at the well-protected
squares c5 and a5. But the knight is generally
preferable on d4 in working together with the
Pau lsen System one on c3 because their effect is prophylactic.
i.e. they prevent Black from making desired
4 ..a6 (D)
. freeing moves. Thus if Black plays ... e5. the
It's curious that this was one of the first Sicil knights are well-placed to land on d5 and f5.
ian Defence lines that was taken seriously by And if Black plays . . . d5, then the c3-knight
Louis Paulsen, and therefore by many of his plays a role by attacking the pawn. Moreover. if
successors. Alekhine, for instance, had trouble White responds to ... d5 by exd5, then after
deciding upon how to meet the Sicilian, feeling . . . exd5 the knight on d4 becomes an ideal
that the early ... a6 idea took precedence. The blockader. In the same situation, if White is
reason that this strikes us oddly is that the able to respond to . . . d5 with e5, the d4-knight
Paulsen is so modem in spirit: Black fails to de will be powerfully placed and can support f4-f5
velop a piece and creates dark-square weak as well. Thus White's knights are restricted. but
nesses on b6 and d6. His play is extremely so are Black's centre pawns, so we might call
flexible, and that is one of its points. Having this a situation of mutual prophylaxis. Notice
that this state of affairs also applies to the Tai This is a specialized instance of what can be
manov Variation and to a lesser extent, every a beneficial thinking tool. It's often useful to
Sicilian line with pawns on a6 and e6. The imagine yourself having an extra move when
modest difference in the case with ... d6 and you're playing an opening variation. What
. . .e6 in is that the e6-pawn is easier to attack. would you do? Can you use the move produc
After 4 ... a6, White's first decision is whether tively? This is a very good exercise that will
to: sometimes give you greater understanding of
a) put a pawn on c4 and emulate the Mar an opening than detailed and time-consuming
oczy Bind; study might.
b) play for normal development by 5 lbc3 ; 7 .id3
or Black does well after 7 e5 ! ? lbe4 8 �g4
c) wait to decide by playing 5 .id3. lbxc3 9 a3 .if8 ! 1 0 bxc3 �a5 1 1 �g3 d6 !, a
book line that has remained unchallenged for
Playing Maroczy-Style years.
7 ... lbc6 8 lbxc6 dxc6! (D)
5 c4 lDf6 6 lbc3 .ib4!? (D)
Another 'tempo-loss' paradox: if you com

What are the ideas here? With the move 5 c4, pare the equivalent Taimanov line again, White
White is doing his best to prevent Black from has gained the move .id3 in return for the inef
even thinking about ... d5 and . . . b5, his tradi fectual move ... a6. Surely that must improve
tional freeing moves. And Black's development White's prospects? But without .id3 having
is rather strange. At this point he appears to been played, the sequence 7 lbxc6 dxc6 ! ? in the
have lost a tempo on the analogous Taimanov Taimanov allows 8 �xd8+, as we shall discuss
line, which (as shown in the next section) goes in that section. This is probably playable for
I e4 c5 2 lDf3 e6 3 d4 cxd4 4 lDxd4 lbc6 5 c4 Black, but leaves him with a different set of
lDf6 6 lbc3 .ib4. After all, in the Paulsen line, problems. As it stands in the Paulsen after
Black's a6-pawn is pretty irrelevant compared 8 ... dxc6, 9 �xd8+ isn't possible. So White has
to having a knight on c6 (as he would have in gained in development but lost in opportunity.
the Taimanov Variation - see also below). That 9 e5
may on balance be true but there is also the 9 0-0 e5 ! frees the c8-bishop and wins an
typical paradox of modern Sicilian lines that outpost on d4. In that situation as well, it's good
being a move behind will sometimes result in for Black to have queens on the board.
the better position! In the Taimanov version 9...�a5
above, White's best move is probably 7 lDxc6, Now the play gets forced:
whereas in the Paulsen White doesn't have that 10 exf6 .ixc3+ 11 bxc3 �xc3+ 12 .id2
option because there's no knight on c6 to cap �xd3 13 fxg7 .l:.g8 14 .ih6 �c3+ 15 @fl �f6
ture. 16 �cl e5 17 c5
1 7 .l:.bl ?! .ie6 ! 1 8 .l:.xb7 0-0-0 with ... .l:.d4

next.
17 .ie6
.•.
Up to here we have theory. If Black is happy

with this position then 5 c4 doesn't pose a prob
lem for him. Otherwise Black should consider a
positional approach, such as 6 . . .'V//lc 7. The point
is that you have to be ready for concrete lines
but also understand positions like the one after
9 0-0 e5.
Conventional Development
5 lDc3 'V//lc7 (D)

al) Take a look at 19 .ie3 exf4 20 ii.xf4
lbe5 21 lDf5 .if6. We've seen this ideal set-up
for Black before; ... lbc4 or ... lbg6 will come
next. If White had any chance of equalling the
effect of Black's domination of e5 and his
threats to the e-pawn, he would have to have
some pieces ready to come to d5, which is not
realistic at the moment.
a2) 1 9 .ic3 d5 ! (the standard freeing move;
White's position is falling apart) 20 @h l (20
.ia5 .ic5+ 21 @h2 .ib6 22 .ixb6 lDxb6)
20 ...dxe4 21 .ia5 (21 fxe5 b4 with the idea 22
.ixb4 e3) 2 1 . ..lDb6 ! 22 'Villc 3 'Villx c3 23 bxc3
exf4 24 lbxe4 .ixe4 25 .ixe4 .ic5 26 .id3 g5
27 .l:.ael .l:.c8 and Black is winning, Fontaine
6 .i d3 Svidler, French Cht 2003.
The most popular of several continuations, b) 6 f4 b5 (the early fianchetto is a trade
at least in club-level chess. mark of the Paulsen; 6 ... lbc6 is another option
a) I'll present just one example of 6 g3. in Taimanov-style) 7 .id3 .ib7 8 'Ville 2 lbc6 9
Black could then transpose into a Taimanov or lbxc6 'Villx c6 10 .id2 (the beginning of a medio
other Sicilian by 6 ... lbc6 or 6 ... d6, but he has a cre plan; more interesting is 10 a3 .ic5 1 1 .ie3
unique and effective move in 6 . . . .ib4 ! : 7 lDe2 or 10 0-0 .ic5+ 1 1 '.th l lDe7 1 2 e5 ! ?) 10 ... .ic5
lbf6 8 .ig2 .ie7 (it's as if Black played ... .ie7 (D).
and White had his knight transferred to e2 with
out using any time) 9 0-0 0-0 10 h3 d6 (D).
This position should be equal, since Black
has his normal queenside expansion themes
and White can't do much on the long diagonal.
Generally White will tum his attention king
side: 1 1 .ie3 lbc6 1 2 g4 (the beginning of an
ambitious advance) 1 2...b5 1 3 f4 lbd7 14 lbg3
.l:.e8 15 'Villd2 .ib7 1 6 lDce2 (this is a sign that
White doesn' t know how to continue with his
kingside attack and indeed there doesn't seem
to be a good plan in that area of the board)
1 6 ... .l:.ad8 l 7 lDd4 lDxd4 1 8 .ixd4 e5 ! (the
standard manoeuvre) and now:
1 1 0-0-0 ! ? (when Black's only action is on This is a solid choice which results in fixing
the queenside, this seems strange, especially White's pawns. Naturally, 6. . .lbf6 is playable.
since White has no real prospect of attacking on Instead, a nice attack followed 6. . .b5 ! ? (risky) 7
the kingside, where Black stands so solidly; 0-0 .ib7 8 l:.el d6 9 .ig5 (creating the same
still, White isn't in any trouble at this point) problem for Black as he had in the last note:
1 1 . . .lbe7 (whether you are White or Black, be White prepares his sacrifice by cutting off es
aware of 1 l . . .b4 12 lbd5 !) 12 �h5 ! l:.c8 1 3 cape-squares from Black's king) 9 . . . lbd7 ? !
@bl b4 14 lbe2 a5 1 5 f5 .ia6 ! ? 1 6 l:.hfl .ixd3 ( 9. . . lDf6) 10 a4 ! b4 1 1 lbd5 ! exd5 1 2 exd5+
17 cxd3 exf5 1 8 exf5 lbd5 19 l:.c 1 0-0 ! and the .ie7 1 3 lDf5 lbe5 14 lbxe7 lbxe7 15 .ixe7
position is unclear, Meister-Poluliakhov, Kras @xe7 1 6 f4 �c5+ 1 7 @h l �xd5 1 8 �g4 h5 1 9
nodar 200 1 . 20 d4? ! would be answered by �h3 �d4 20 fxe5 �g4 2 1 exd6+ @xd6 22
20 . . . lDf6. �xg4 hxg4 23 l:.adl ! and White should win,
c) There are of course countless games with Ghinda-Kirov, Timisoara 1987.
6 .ie2, when one can return to a Taimanov with 7 lbxc6 dxc6 8 0-0 lDf6 9 f4 e5! (D)
6 . . . lbc6 followed by ...lDf6 or to a Scheven
ingen set-up with 6...lDf6 and 7 ... d6. But the
Paulsen faithful like to play 6. . .b5 7 0-0 .ib7 in
every position. Here it looks wrong after 8
l:.el ! . That's a useful move in any case but in
quite a few variations of the Sicilian it prepares
some form of the sacrifice 8 . . . b4? ! 9 lbct5 ! (D).
Reaching a type of position that we see in

other Sicilian variations. Black has two excel
lent bishops and active pieces so it's up to
White to use his superior development quickly.
We'll follow a game with all the customary
ideas:
9 ...exd5 10 exd5 @d8. Now White has vari Lanka - Santo-Roman

ous ways to pursue the attack, and chooses a Prague 2000
good one: 1 1 .if3 ! d6 (this begs for a check on
c6, but it's not easy to get one's pieces out in the 10 f5!
face of ideas such as �e2 and lDf5, with .ie3, Other moves:
c3 and l:.c 1 if needed) 12 .if4 lbd7 13 lbc6+ a) 10 fxe5 ? ! .ic5+ 1 1 @h l lbg4 12 �f3
.ixc6 14 dxc6 lbc5 1 5 �d5 ! lDf6 16 �xc5 ! 0-0 ! ? ( 1 2 . . . .ie6 1 3 .if4 lbxe5 14 �g3 f6 1 5
dxc5 17 l:.ad 1 + @c8 1 8 .ig4+ @b8 19 l:.d8+ .ixe5 fxe5 gives Black the two bishops; or
@a7 20 l:.xa8+ @xa8 21 .ixc7 lbxg4 22 l:.e8+ 1 2 . . . lbxe5 with equality) 1 3 .if4 lDxe5 14 �g3
@a7 and my database game gives 23 l:.b8 1 -0 .id6 ! 15 l:.ad 1 f6. Black has his outpost on e5 in
Yang Xian-Ramos, Moscow OL 1 994. I sus front of an isolated pawn again, and this time he
pect that this is a typo for 23 .ib8+ ! @b6 24 c7 doesn' t have to worry about lbd5 ideas or a
with an instlmt win. Otherwise 23 l:.b8? lbf6 weak pawn on d6.
would definitely be worth playing on. b) 10 @h l and now 10 ... .ic5 intends ... h5,
6... lbc6 while queenside castling via 1 O ... .id7 is another
option ( 1 0... h5 ! ? could also be played immedi 1 3 a4 .ib7 1 4 .ig5 tlJg4 1 5 iLidl !
ately). White covers his weak squares; the knight
10 .icS+ 11 'it>hl h5! (D)
••• wasn't getting to d5 anyway.
A characteristic move ofthis system. Now the 15 ... .ie7 16 .ixe7
game is double-edged. Another encounter went Or 1 6 .id2 ! ? with equality.
l 1 . ..h6 ! ? 1 2 a4 l:tb8 13 �f3 b5 14 �g3 'it>f8 1 5 16...�xe7 17 iLif2
�f3 .ib7 and Black did well in Qin Kanying-Ye Black is well off with 1 7 iLie3 �h4 1 8 h3
Jiangchuan, Shanghai 2000; his king doesn't iLif6 ! .
seem comfortable on the kingside, however. 17...�h4 18 h3
Although Black went on to win after 1 8 ... 0-0,
he should prefer either 1 8 . . . iLixf2+ or 1 8 ... l:td8,
with equality in either case.
The Non-Committal Line
5 .id3 (D)
By comparison with the analogous Taimanov
Variation (4 ... tt:Jc6) White is glad to be able to
post his bishop on d3 without first having to de
fend, retreat, or exchange his d4-knight. Impor
tantly, he retains the option of playing c4.
12 �f3
The main alternative is 1 2 .ig5 ! ? ; for exam
ple, 1 2 . . . tlJg4 1 3 �e2 .ie7 ! ? ( 1 3 . . . b5 ! ?) 1 4
.id2 ( 1 4 .ixe7 �xe7 1 5 �f3) 1 4 . . . .ic5 1 5 h3
( 1 5 iLid l ) 1 5 . . .�e7 16 iLia4 .ia7 1 7 b4 ( 1 7
.ia5 �h4) 1 7 . . .b 5 1 8 tt:Jc5 ? ! .ixc5 1 9 bxc5
�xc5 20 a4 .ib7 2 1 l:.abl l:.d8, Tiviakov
Cacho Reigadas, Arco 1 998. White went on to
win, but Black's position looks quite healthy.
12... bS (D)
Black plays to prevent iLia4; an alternative is
1 2. . .tlJg4 1 3 �g3 .id7 ! ? 1 4 .ie2 0-0-0! 1 5
.ixg4 hxg4 1 6 �xg4 g6 ! . 5...iLif6
Black's position is ultra-flexible, with seem
ingly infinite room for creativity. At this point he
has moves such as 5 ...iLie7, 5 ... �c7, 5 . . . .ic5 (6
iLlb3 .ia7 or 6 ... .ie7), 5 ... �b6 (with the idea of
misplacing the knight and then playing . . . �c7;
we discuss that ploy elsewhere in this chapter),
or 5 . . . g6 (D), which deserves a diagram.
All of Black's pieces are on the back rank and
his position is the definition of holes ! Too bad
there isn't a pawn left over to put on c6. Yet
plenty of grandmasters have played 5 ... g6 and at
least one of them, a leading Paulsen theoretician,
thinks that Black can equalize from this position
with two different set-ups. To me, the most
My feeling is that one of these many 5th

move alternatives might be more rewarding than
5 . . .tDf6, which allows White to set up a common
and generally effective formation.
6 0-0 d6
6 ... 'iVc7 7 'iVe2 d5 ! ? (uh-oh, this again!) 8
exd5 tDxd5 9 .ii..c4 tDf6 10 .ii.. g5 .ii..e7 1 1 tDc3
0-0 1 2 l:.adl b5 13 .ii..d3 .ii..b7 14 'iVe3 tDbd7 1 5
tDe4 l:.fe8 1 6 tDxf6+ 1h-1h Akopian-Svidler,
Moscow 2004. Average rating of the players?
Over 2700.
7 c4! b6 8 tDc3
8 b3 .ii.b
. 7 9 'iVe2 tDbd7 10 tDc3 g6! (duelling
fianchettoes are common in this line, and the one
plausible idea is ... .ii.. g7, ... tDe7 and ... d5; but if on g7 hits the loose d4-square; perhaps White is
he can't get ... d5 in, Black can settle for ... d6, a little better but that has to be demonstrated) 1 1
. . . tDbc6, etc., when he has done reasonably well. ilb2 .ii.. g7 1 2 l:.adl 0-0 1 3 f4 e5 ! 1 4 fxe5 tDxe5
That's not all. Are you ready for the outra 1 5 ii.bl l:.e8 and Black won his e5-square in
geous 5 ... d5? In the few tests of this move thus Seitaj-Gheorghiu, Thessaloniki OL 1984. For
far, no one seems to have come close to refuting his part, White's got a wonderful d-file to use, so
it. Most games have gone 6 exd5 'iVxd5 7 0-0 he might claim equality. Then again, there's that
tDf6 (7 ... .ii..d7 ! ? in some sense gains a tempo, awful bishop on bl which needs attention, so
because now 8 tDc3?? allows 8 . . .'iVxd4 and maybe Black has the better of it after all.
there's no check; of course you may not want 8 .ii.. b7 (D)
.•.
your bishop on d7) 8 tDc3 and now 8 . . . 'iVd8, but

Black could also play 8 ...'iVd6 (D):
This is a normal position, from which we'll

follow a model game.
Then Black has another of those 4:3 kingside
pawn-majorities that we talk about in so many P. Popovic - Pikula
openings, including the French Defence Tar Banja Koviljaca 2002
rasch line that this resembles so strongly. Com
pare that variation: 1 e4 e6 2 d4 d5 3 tDd2 c5 4 9 f4 .ii..e7 10 'iVe2
exd5 'iVxd5 5 tDgf3 cxd4 6 .ii.. c4 'iVd6 7 0-0 tDf6 This set-up introduces a strategy with which
8 tDb3 tDc6 9 tDbxd4 tDxd4 1 0 tDxd4. It would White has won many games.
be a confirmation of the Paulsen's remarkable 10 ... 0-0 1 1 .ii.. d2 tDbd7?
flexibility if Black could actually get away with A fundamental mistake. l 1 . . .tDc6 is much
(and equalize following) 5 . . . d5. better, although still not problem-free.
12 l:.ael (D) White doubles, triples, occupies the outpost,

and wins.
16...'it>g7 17 l:.efl lDh7 18 'iVf2 'iVe8 19 lDd5!
.ixd5 20 cxd5 .ig5 21 .ic3 .id8 22 'iVe2! b5
Otherwise White simply takes the a-pawn
and attacks on the queenside too.
23 lDaS lbg5 24 l:.g3 l:.h8 25 lDc6 .ib6+ 26
'it>hl l:.h5?
But Black won't like 26 ...f6 27 .id2! or
26 . . . lbh7 27 .id2 lDb8 28 lDxb8 l:.xb8 29 l:.h3.
27 'iVxh5! gxh5 28 l:.xg5+ @rs 29 l:.xh5
'it>g8 30 l:.g5+ 'it>f8 31 .id2 f6 32 l:.g3 'it>f7 33
.ie2! 'iVh8 34 l:.h3 'iVg7 35 .ih6 'iVh7 36 .ih5+
'it>g8 37 .ig4 tt:Jrs 38 .ixf8 'iVxe4 39 .ie6+
'it>xf8 40 l:.xf6+ 'it>g7 41 l:.f7+ 'it>g6 42 lDe7+
'it>g5 1-0
This is a great piece-formation for White. White stood much better all the way. This is
It's not that the attack is so powerful yet, but a good piece-formation to remember.
that Black hasn't a shred of queenside or central
counterplay.
12 g6
.•. Tai manov Variation
Played to prevent a breakthrough by e5 ; it
looks necessary. 1 e4 c5 2 lDf3 e6 3 d4 cxd4 4 lDxd4 lbc6 (D)
13 f5!
Now you can see why that knight was better
off going to c6.
13...e5
1 3 . . . gxf5 1 4 exf5 e5 1 5 lDc2 l:.e8 1 6 lbb4 !
b5 l 7 lDbd5 doesn't look so bad at first, but af
ter inevitable exchanges on d5 Black will be
positionally lost. A good position to study;
White will subsequently get space and two
bishops, a deadly combination.
14 lbb3 'it>h8 15 fxg6 hxg6? (D)
But 1 5 . . . fxg6 1 6 .ih6 is pretty bad.
By deploying the knight to c6, Black breaks

with the noncommittal Paulsen approach. He
decides early upon the position of the queen's
knight rather than keeping open the option of
. . . lDd7. He also allows White to play lDb5. In
return, he has developed a piece, and his c6-
knight limits White's options (for instance, the
anti-Paulsen move 5 .id3 simply loses a piece
here) . We' ll briefly examine White's three ma
jor lines: 5 c4, 5 lDb5 and 5 lDc3.
5 g3 allows the freeing advance 5 ... d5. Then 6
.ig2 can be met by 6... .ic5 ! 7 lDb3 .ib6 8 exd5
16 l:.f3! exd5, a convincing pawn sacrifice for Black; for
example, 9 .ixd5 (9 lDc3 lDge7 1 0 lbxd5 lbxd5

1 1 'iVxd5 'iVxd5 1 2 .ixd5 lbb4) 9... 'iVe7+ 1 0
'iVe2 'iVxe2+ 1 1 @xe2 lbb4. Instead, the some
what dull 6... dxe4 ! ? has been used in practice,
achieving equality.
'Maroczy Lite'
5 c4 (D)
This advance is somewhat rare but leads to
material that is potentially useful. White tries to
set up a sort of Maroczy Bind. This is slow in
the face of the rapid development that 2 ...e6 and
4 . . . lbc6 makes possible, yet both sides must
play accurately. You don' t see this kind of freeing move very
often in the Sicilian because, behind the scenes,
White makes his moves so that there is a spe
cific drawback to . . . d5, such as a multiple cap
ture or e5. He is normally successful in doing
this, and that' s why you seldom see an effective
early ... d5 in any well-played Sicilian, including
the Najdorf, Rauzer, Scheveningen, Dragon or
for that matter Taimanov. This is obvious to a
player accustomed to the Sicilian, but perhaps
not to a newcomer who sees many games with
. . . d6 and . . . e6 and assumes that Black just pre
fers to play with less active pieces. In the dia
grammed position White can't even maintain
equality, as a short analysis will show you.
Returning to 7 lbxc6 (D), Black has two re
5 lDf6
..• captures.
5 ...'iVh4 ! ? 6 lbb5 ! ? (6 lbc3 .ib4) 6 . . .'iVxe4+
7 .ie2 'iVe5 could get wild and woolly; if Black
can get away with an extravagant move like
5 ... 'iVh4, it shows that the loss of time involved
with 5 c4 is meaningful.
6 lDc3 .ib4 7 lbxc6
White exchanges this so as to play .id3 and
protect the e-pawn (we discussed this in the
Paulsen section). It's important to see that Black
is not committed to setting up a prepared forma
tion with, say, ... 'iVc7 and ... a6. The Taimanov
move . . . lbc6 goes well with quick development.
For instance, White can't simply make Maroczy
Bind moves such as 7 f3 ? ! 0-0 8 .ie3, because
8 ... d5 ! (D) is precisely the type of pawn-break
that Black wants to make, and White needs to
prevent, in any Sicilian Defence. Muzychuk - Gershon
Sometimes students are so intent upon set Dresden 2003
ting up some restricted Sicilian position with
. . . d6 and ... e6 that they forget about the basics. 7 bxc6
..•
This is the usual move, strengthening Black's

centre. He can also play 7 ... dxc6 ! ? 8 �xd8+
@xd8, which is awkward but probably OK so
long as Black is able to achieve ... e5, the ideal
move that he needs, in order both to get his
light-squared bishop out and to secure an out
post on d4. For example, 9 f3 ! ? (9 e5 ! is proba
bly better, interfering with Black's plans; then
9 ...lbe4 I O a3 ! ? .ixc3+ 1 1 bxc3 b6 should be
looked at - White has no worries, but on the
other hand it's hard to see how he will make
progress) 9 ... e5 I O .ie3 @c7 (D).
9 0-0 .ic5
It can be advantageous to delay castling for
reasons that will be seen, and it won't hurt to in
crease Black's control of d4. But 9 ....ixc3 I O
bxc3 d6 has also been played.
10 .ig5?! (D)
This looks natural enough but turns out badly.
White has the interesting option of I O 'ii'f3 � .
flirting with �g 3 but also preparing .ig5 i f it
makes positional sense.
1 1 l:.c I lDd7 ! 12 @f2 ! ? .ic5 (removing a de

fender of d4 while getting rid of White's good
bishop) 1 3 lba4 ! (preventing . . . lbc5 and ... a5
after the exchange of bishops) l 3 ... .ixe3+ 1 4
@xe3 lbf8 !, intending . . .lbe6 and . . .lbd4, Lau-
tier-Ivanchuk, Tilburg 1 992. Now White has 1 5
c5 ! , to get his bishop to c4 (thus the reasoning
behind l:.c I and lba4). But Black can still play
1 5 ... .ie6 1 6 b3 lbg6 ! 1 7 .ic4 .ixc4 1 8 l:.xc4
lDf4 ! 19 g3 lbe6 20 l:.dl l:.hd8, etc., with the
same advantageous outpost on d4.
For those who are familiar with the King's
Indian Defence, notice that we have here the 10...h6 1 1 .ih4 d6 12 l:.bl g5 13 .ig3 h5! 14
same central pawn-structure, same weakness, h3 h4 15 .ih2 g4!
and same manoeuvres by Black as appear in the A tactical ploy to remember.
Exchange Variation of that opening ! Of course 16 hxg4 lbxg4 17 b4 .id4 18 lDe2 .ib6 19
White didn't put up much resistance to this c5
plan. Otherwise l 9 . . . h3 or l 9 . . .�f6 comes.
We now return to 7 ... bxc6 (D): 19...h3! 20 .ig3 dxc5 21 bxc5 .ixc5 22 'ii'c2
8 .id3 �d6 23 l:.bcl
After 8 e5 comes 8 ...lbe4 9 �d4 �a5 ! . In this position, the easiest path to advantage
8 e5
•.• lay in 23 ... hxg2 24 @xg2 .ib6 25 �xc6+ �xc6
Or 8 ... 0-0, or 8 . . . d5 ! ?, but in the latter case 26 l:.xc6 .ib7 27 l:.c2 f5, when White's posi
watch out for 9 cxd5 cxd5 ?? I O �a4+. tion is declining.
Hedgehog �d3 ( 1 9 �f3 l:.xf2+ ! 20 'it>xf2 �c5+) 1 9 . . .l:.g4

with the upper hand. The variations are much
5 tt'lbS d6 6 c4 more complicated than that, but the verdict re
With this move White sets up another sort of mains the same.
Maroczy Bind and Black generally plays in b) 9 'Lld2 ! (D) improves, albeit not enough
what is called 'Hedgehog' fashion: pieces and for White to get excited about:
pawns curled up on the first three ranks waiting
for the chance to burst out into activity. This
particular form of the Hedgehog has done rea
sonably well over the years, although at the B
very top levels Black still seems to run into
problems from time to time.
Before entering into that discussion, a varia
tion with a colourful history begins 6 .if4 e5 7
.ie3. Now Black can play 7 . . . a6, but the main
line goes 7 . . . 'Llf6 8 .ig5 ! ? (the bishop moves
for the third time in a row ! This move protects
e4, of course, and also strengthens White's con
trol over d5) 8 . . . .ie6 (D).
9 . . . .ie7 10 .ixf6 .ixf6 l l 'Llc4 0-0! 12 �xd6

�c8 ! . White has invested a lot of time and the
bishop-pair to win one pawn. There have been
quite a few games from this position demon
strating full compensation for the pawn. Black
has very active pieces and White's are subject
to attack. Also, Black may play ... a6 followed
by ... tt'ld4 and dominate the board from that
square. One illustration: 1 3 c3 l:.d8 1 4 �c7
.ie7 1 5 �xc8 l:.axc8 (Black's initiative persists
even without queens) 1 6 'Llba3 ctJd4 ! 1 7 cxd4
.ib4+ 1 8 'it>e2 .ixc4+ 19 'Llxc4 l:.xc4 20 'it>f3
l:.cxd4 2 1 .ie2 l:.8d6! and Black eventually
Without going into too much detail, Black's won an opposite-coloured bishop ending in
development has again been quite rapid, and Borisek-Navara, Balatonlelle 2003.
White can't keep a grip on the position. There 6...'Llf6 7 'Lllc3 a6 8 'Lla3 (D)
are two options here:
a) The old main line 9 tt'l l c 3 a6 10 .ixf6
gxf6 1 1 'Lla3 (threatening 'Llc4-e3, or 'Lld5, or
in some cases �h5) was solved in style by
l l . ..d5 ! 1 2 exd5 .ixa3 1 3 bxa3 �a5 14 �d2
0-0-0. It becomes clear that White won't win
the piece, and his development is slow while his
extra pawn on the a-file is hardly useful: 1 5
.ic4 l:.hg8 1 6 l:.d l , and various analysts have
looked at 16 . . . l:.xg2 ! ( 1 6 . . ..if5 ! ? 1 7 .id3 .ixd3
1 8 �xd3 'Lld4 1 9 0-0 'it>b8 yielded equality in
the famous game Fischer-Petrosian, Buenos
Aires Ct ( 1 ) 1 97 1 ) 1 7 �e3 ( 1 7 'Lle4 �b6)
17 ...tt'ld4 1 8 @fl ! 'Llxc 2 ! ? (or 1 8 . . . �c7) 1 9
At this juncture we look at two games. The loosening his position. Although strong mas
first will illustrate White's set-up with the ag ters have had success neutralizing this strategy.
gressive f4. The second serves to represent the it is still a valid approach and in any case quite
overall main line with f3. instructive.
Before that I should mention Kasparov's fa 8 .ie7 9 .ie2 0-0 10 0-0 b6 1 1 .ie3 lDe5
.•.
mous gambit in the 1 985 world championship l 1 . . ..ib7 is a more accurate choice if Black
match against Karpov, which went 8 ... d5 ! ? 9 wants to prevent f4 from being effective, be
exd5 exd5 10 cxd5 lbb4 (D). Contrary to the cause his knight has not used up time on ... lt:Je5-
general opinion, this is still unresolved. d7. Then the immediate 1 2 f4 gives Black some
easy ways to counteract White's structure, in
cluding 1 2 . . . l:.c8 1 3 l:.c l l:.e8. Development of
the rook to e8 supports . . . d5, because the ex
change of White's e-pawn will bring the rook
into a position facing the vulnerable e3-bishop.
Nevertheless, White can play 1 2 l:.c l , hoping
for 1 2. . . lbe5 1 3 f4.
12 f4 (D)
In the critical encounter Karpov played 1 1

.ie2 .ic5? ! 1 2 0-0? .if5 and got into great dif
ficulties, losing a brilliant game. Later Karpov
played 1 2 .ie3 ! with advantage, the point being
1 2 ... .ixe3 1 3 �a4+ and 1 4 �xb4. The best
move after 1 1 .ie2 is l l . . . lDfxd5, which has
been heavily analysed down to a promising
piece sacrifice for White, as has 1 1 .ic4 .ig4
( l l . . .b5 1 2 0-0! bxc4 1 3 l:.e l + .ie7 1 4 d6) 12 The majority of masters have used a forma
�d4 b5 1 3 lbcxb5 ! with a big mess. The point tion with f3 in this variation, as in the next
of mentioning all this is twofold: main game below. Those positions are very
a) If you play moves such as 8 . . . d5 you sim well-known and fairly easy to play because of
ply have to memorize a lot of material. the limited set of piece placements that they
b) If the move 8 . . . d5 works it invalidates 5 logically allow for. Although his chances of
lbb5, because if White can't prevent . . . d5 in the gaining an advantage are slim if Black plays
Sicilian by direct means it is extremely unlikely accurately, White has more opportunities for
that there will be any way to gain a positional original play after f4.
advantage thereafter. 12 lDed7 13 .if3 .ib7 14 �e2
...
Black implemented a positionally effective

White Plays f4 plan with . . . h6 in Briiggemann-Lutz, Erfurt
2004: 1 4 'it>hl h6 (D).
Nunn P. Cramling
- Black's point is to answer 15 g4? ! with
Zurich 1984 15 . . . lbh7 ! and ... g5 next. Then Black has essen
tially made White's f3-bishop a bad one, since
I'll use this game without much analysis to e5 can't be played and he has neutralized any
demonstrate an ambitious plan with f4 that af pawn advances at the same time. The game pro
fords White attacking chances but at the cost of ceeded 1 5 l:.c l �c7 1 6 lbabl l:.ac8 1 7 b4 l:.fe8
\t>h l 'ifd5 28 a3 'ifc4 29 f5 !? 'ifxc2 (29 . . .ctJg5 ! ?

might be worth a try) 30 'ifxc2 l:!.xc2 3 1 fxe6
l:!.c6 32 a4 1/z-1/z Tseshkovsky-Kasparov, USSR
Ch (Minsk) 1 979.
15 l:!.fdl 'ifc7 16 l:!.acl l:!.ac8
This was the time for 1 6 ... h6 ! , to answer 1 7
g4 with l 7 . . .ctJh7 ! ; compare what happens next.
17 g4! h6
Kasparov's idea 1 7 ... ltJc5 1 8 'ifg2 d5? fails
now that Black's queen is on c7: 1 9 cxd5 exd5
20 e5 ctJfe4 21 ltJxd5.
18 h4!
This is a different story, because ... g5 is pre
vented.
1 8 a3 'iVb8 1 9 ctJd2 (a standard reorganization, 18 ctJh7 19 'ifh2
.•.
but over the years it' s become clear that a A good move, and the natural 1 9 g5 also
knight on d2 versus the Hedgehog is primarily looks promising; e.g., 1 9 ... hxg5 20 hxg5 e5 2 1
defensive and limits positive operations; tradi ctJd5 i..xd5 2 2 cxd5 'ifb7 2 3 l:!.xc8 l:!.xc8 24
tionally the knight belongs on d4) 1 9 ... i..f8 20 i..g4.
We l .ia8 21 'iff2 i..c6 22 l:!.fe l i..e7 and nei 19 ltJc5 (D)
.•.
ther side was doing much.

We now return to 14 'ife2 (D):
20 'ifh3!
Now 20 g5 can be answered by 20. . .f5 ! ? 2 1
14 l:!.eS
••• i..h5 ltJxe4 2 2 ltJxe4 i..xe4 2 3 i..xe8 l:!.xe8 and
The very young Kasparov played 1 4 ... 'ifc7 Black possesses the terrible a8-h l diagonal.
1 5 l:tac l l:!.ac8 16 g4 ltJc5 and the game demon The text-move prevents ... f5.
strates that Black needn't play ... h6 if he has 20 i.. f6
.•.
other central prospects: 1 7 'ifg2 d5 ! 1 8 e5 ctJfe4 20 . . . g5?1 is met by 2 1 hxg5 hxg5 22 l:!.c2 !
1 9 cxd5 exd5 20 b4 (20 l:!.fd l ) 20... ltJxc3 21 and l:!.h2.
:Xc3 d4 ! (a typically tactical solution) 22 i..xd4 21 ltJabl g6 22 l:!.c2 i.. g7 23 l:!.cd2 i..f8 24
Wd7 23 ctJc2 i..xf3 24 l:!.cxf3 ! (24 l:!.fxf3? ltJe6! g5 h5 25 i..f2 i.. c6 26 ltJa3! (D)
and White's pawns and pieces are loose; 24 Terrific ! The knight heads towards its right
Wxf3 ltJa4 25 l:!.xc8 .l:.xc8 also costs White ma ful square on d4. White has shown admirable
terial) 24 ... ltJe6 25 i..e 3 ! (25 i..xb6 'ifc6 hits patience throughout this manoeuvring stage.
two pieces, so 26 i..c5 ltJxc5 27 bxc5 i..x c5+ 26 ltJd7 27 ltJc2 ltJc5 28 ctJd4 i.. b7 29 f5!
•••
could follow; Black's activity provides plenty Finally! White transforms his space advan
of compensation) 25 ... f5 ! ? 26 exf6 i..xf6 27 tage into concrete gains.
29...exfS 30 exf5 ll'le4 31 fxg6 fxg6 32 ll'lxe4 doing reasonably well if you compare this wilb
i..xe4 33 ll'le6 �b7 34 i..xe4 �xe4 35 li:'ixf8 a 'normal' Hedgehog arising from the English
li:'ixf8 36 l:!.xd6 Opening. The knight on a3 can t possibly be su
'
White has a considerable advantage now, al perior to that on d4 and it has used four moves
though the mutually exposed kings make the to get to the edge of the board! On the olbc:r
position difficult for both sides. We are past the hand Black can't even think about ...b5. So
opening stage and I ' ll let the moves speak for what' s going on here? From Black' s point of
themselves. Towards the end White's king looks view it would be nice to do something positive
exposed but according to my chess engine before White catches up by rerouting his knight
Black never had any kind of perpetual check. and pushing his queenside pawns. But in this
36 l:!.xc4 37 l:!.el �a8 38 l:!.xe8 �xe8 39
.•. sort of Hedgehog formation Black famously
�b3 �f7 40 l:!.f6 �d5 41 l:!.xb6 \t>h8 42 l:!.f6 waits until the opportunity comes for . ..b5 or
l:!.cl+ 43 \t>h2 l:!.hl+ 44 \t>g3 �es+ 45 \t>g2 ... d5. What to do? There are two main stnre-
�h2+ 46 @f3 �h3+ 47 \t>e4 'ifg4+ 48 l:!.f4 gies. One is to play moves like . ..:es. . . .9b8
�e2+ 49 �e3 �c2+ 50 �d3 'ifc6+ 51 �d5 and ... i..f8, and then get serious about . ..d5. 1be
�c2+ 52 @f3 ll'le6 53 i.. d4+ lt:'ig7 54 l:!.fS+ 1-0 other is to embark upon the now-famous plan
of ... i..d8-c7 (with minor threats on the king
White Plays f3 side) followed by ... @h8, ... l:!.g8 and ... g5-g4.
generally with more serious threats. This is an
Anand - Illescas important strategy for both sides to know, if
Linares 1 992 only because White has been blown away by
the attack in so many games. There's another
8 ... b6 9 i..e2 i..b7 10 0-0 lt:'ib8!? rather silly-looking attack by ... h5-h4 (and. if
An old move-order that ends up transposing allowed, ... h3 to enhance the power of the b7-
to the main line. This knight normally travels to bishop). This has been tried several times in
e5 and then back to d7. I should also mention recent master practice without White having
that the moves ... i..e7, ...b6, ... i..b7, . . . 0-0, ... 'ifc7 found a convincing counterplan. Of these three
and . . . l:!.ac8 have been played in almost every ideas, the easiest for White to stop should be the
sequence. For once I'm going to ignore move first (a . . . d5 break) but he has to be careful. as
order issues and concentrate upon the basic po shown by 1 5 . . . l:!.ad8 (instead of 1 5 . . . l:tacS.
sition. which is probably objectively better) 1 6 �fl
1 1 f3 i..e7 12 i..e3 lt:'ibd7 13 'ifd2 0-0 14 'ikb8 1 7 ll'lc2?! l:!.fe8 1 8 \t>h l ? d5 ! 1 9 cxd5 exd5
l:!.fdl �c7 15 l:!.acl (D) 20 exd5 i..d6 2 1 g3 b5 22 a3 �a8 23 �g2
15...l:!.acS ll'le5 ! ? (23 . . . lt:'ic5 wins back the d-pawn) 24
We're roughly at what might be considered l:!.bl ? ! 1/z- 1/2 Morovic-Leitao, Sao Paulo 2002.
the main line; at any rate, several high-level Probably 24 ... lt:'ic4 favours Black; at any rate he
games have gone this way. Black should be can be satisfied if he achieves . . .d5 safely.
How about White? Taking the ... .itd8-c7, Black makes it difficult to protect b6 .
. . .l:tg8, . . . g5 idea first, White will first play 'lt>hl 22 l:tel llled7 23 a3 .ltb7 24 b4 l:tc7 25 'llb3
and .ltg l to guard h2, and then .itfl for potential .lta8 26 lt:Ja4! .ltc6 27 'llb 2!
second-rank defence by the queen. With that Very original ! c4 needs protection in a lot of
formation you can see that in our main game lines and 'lld3 could be useful at the right time.
those first moves of the plan, . . ..itd8-c7, can be 27 .lta8 28 .ltd4 l:tdc8 29 l:tedl .lte7 30
•••
difficult to implement. Furthermore, White's 'iff2 'i¥b7?

knight on a3 may not be badly placed to meet But 30 ... b5 3 1 llla5 ! is good for White.
Black's strategy. With the queen on b8, for in 31 'lla4 l:tb8 32 'llxb6! lllxb6 33 llla5 'ifa7
stance, the moves l:tbl and b4 will discourage (D)
... d5, when Black has to watch out for the move
c5, followed in some cases by lllc4. Alterna
tively, White can run his queenside pawns at
Black by 'llc2 and b4, a4 and a5. Right in the
middle of that process Black has to be able to
strike in the centre based upon the looseness of
White's queenside; whether he is able to do so
resolves the question of who stands better.
16 .itfl (D)
As explained, this clears the second rank,
and the bishop might have been a target along
the e-file anyway.
B 34 c5
A nic� combination. White wins his piece
back with more to come.
34 dxcS 35 bxc5 lllc8
•••
After 35 ... .ltxc5 36 .ltxc5 'llfd7 37 l:txd7

l:txd7 38 .ltxb6 White wins due to the back rank.
36 c6 l:tb6 37 l:tbl l:txbl 38 l:txbl 1-0
Conventional Development
Lukin - Taimanov
St Petersburg 1 995
16 ...l:tfe8
Fun with move-orders: 1 6 . . . 'ifb8 is right if 1 e4 c5 2 'llf3 e6 3 d4 cxd4 4 lt:Jxd4 'llc6 5 lllc3
Black's plan is to play . . . .itd8-c7, but it's very White chooses the simple path. If Black plays
unclear and involves a pawn sacrifice: 1 7 'lt>hl 5 . . . 'llf6 in this position, we return to the Sicil
(l 7 'llc2 .itd8 !? 1 8 'ifxd6? .ltc7; there are plenty ian Four Knights. In spite of many fascinating
of options here) 1 7 . . . .itd8 ! ? (or 17 ... l:tfe8) 1 8 struggles resulting from the most frequent con
.itf4 ( 1 8 'ifxd6 ! ? .ltc7 1 9 'ifd2 .itxh2 20 g4 tinuation 5 ... 'ifc7, I'm going to forego that and
'ifg3 ! ) 1 8 . . . llle5 1 9 'ifxd6 ! ? .ltc7 20 'ifd2 'llh5 explore only one set-up within the Taimanov
2 1 .lte3, and now 2 1 . . .'llg6 22 g3 ! ? f5 ! is scary. Sicilian proper.
Needless to say, this just scratches the surface. 5...a6 (D)
17 'lt>hl 'ifb8 18 'llc2 lt:Je5 This sequence can sometimes transpose into
This time 1 8 . . . .itd8? 1 9 'ifxd6 .ltc7 20 'ifd2 the . . . 'ifc7 lines. But Black often uses this
.ltxh2?? 2 1 g4 .lte5 22 g5 costs Black a piece. move-order because he wants to play 6 ... lllge7
19 b3 .lta8 20 .ltgl l:ted8? ! 21 'lld4 .itf8? ! next, the system that Taimanov himself loved
and promoted. That move prepares . . . ctJxd4 fol of the possibility of ... fxe4. Amason-Romm
lowed by . . . ltJc6, or ... ltJg6 with dark-square ishin, Lone Pine 1981 continued 14 lOd4 ( 1 4
control over e5 . Black's strategy provides yet ctJd5 ! .ltb7 ! { the idea was 14 . . .exd5 1 5 exf5
another demonstration of the flexibility associ ctJh8 16 f6! } 1 5 ltJxe7+ ltJxe7 1 6 ttJcs .i.c6 17
ated with . . . e6 and . . . a6. .ltd4 "fie8 ! with equality, Orlov-Taimaoov, St
6 .lte2 Petersburg 1 995) 1 4. . ..ltc5 15 exf5 i.xd4 16
There is also independent theory on 6th .itxd4 l:txf5 1 7 "iig4 .ltb7 with advantage for
move deviations after 5 . . . a6 showing ideas that Black. Compare the bishops and rooks !
do not also apply to the Paulsen: 7 ltJxd4 8 "fixd4 ltJc6 9 'ikd3 'ikc7
•••
a) After 6 ltJxc6 bxc6, White can play 7 e5 ! ? This is a normal Taimanov/Paulsen �

(not considered too dangerous for Black after which incidentally renders 'ikg3 useless.
7 . . ."fic7 8 f4 d6) or 7 .itd3. In the latter case 10 .ltgS
White has an aggressive posture but his knight White interferes with Black's developmenl.
might be better-placed on d2, from where it has and his bishop strengthens the effect of a poceo
the squares c4 and f3 within reach. In both tial knight sacrifice on d5 .
cases the move ... a6 tends to be wasted; on the 10....itd6! (D)
other hand these may not be the best plans for
White versus a Paulsenffaimanov structure.
b) Here's a really exotic opening idea for
those who've never seen it: 6 g3 ctJge7 7 ctJb3
d6 8 .ltg2 .itd7 9 0-0 ctJc8 ! ? (Black prepares to
transfer his knight to the queenside and in the
meantime lends extra support to d6) 10 f4 .lte7
1 1 .lte3 0-0 1 2 "fie2 b5 with the idea . . . ctJb6-c4.
This can apply to several positions.
6 ltJge7 7 0-0
.••
7 ctJb3 b5 8 0-0 ltJg6 9 f4 .lte7 10 .lte3 0-0 1 1

i.d3?! (this doesn't seem to work out; possibly
1 1 °fid2 is better) l l . . .ctJb4 12 °fih5 ctJxd3 13
cxd3 (often this pawn-structure is a pleasant
one for White, and his queen on h5 looks partic
ularly well situated) 1 3 . . .f5 ! (D). It's not unusual for the dark-squared bishop
One move turns everything around: White's to go to this square in the Paulsen and Taima
advance f5 is no longer a factor, his e3-bishop nov. In general (i.e., in a broader context than
has been restricted, and Black's b7-bishop will this specific line), ... .itd6 has several points:
have assistance with attacking the centre. White a) It sometimes develops a piece with tempo
even has to keep a watch on his f4-pawn because by threatening . . . .ltxh2+. Then if White replies
h3, he has weakened his kingside and failed to 1 1...ltJeS

contest the f4-square. But upon g3, Black's ad Not 1 1 ....ltxh2? 1 2 g3. However, 1 1 . ..f6 1 2
vance ... h5-h4 can be extremely annoying. Fur .lte3 b5 1 3 f4 .lte7 14 e5 i s perfectly good,
thermore, the bishop can switch to the g l -a7 Nijboer-Van Mil, Dutch Ch (Eindhoven) 1 993.
diagonal when called upon to do so. If White Sommerbauer then suggests that Black snatch
plays f4, a bishop on c5 may be very strong. the centre pawn and hold on to it by 1 4... fxe5 1 5
b) Black's bishop controls important dark fxe5 ltJxe5 1 6 °iid4 .ltb7 1 7 l:!.adl l:!.f8, a contin
squares from d6; in some lines it goes to e5 in uation admitting of some risk, of course.
advance of moving the d-pawn, or it may go to 12 °iid2 f6 13 .ith4 ltJg6 14 .ltg3 .ltxg3
f4 and trade bishops. The ending in that case is It might be more Taimanov-like for Taima
easier for Black than it looks provided he keeps nov to have continued 14 ...ctJf4 looking for a
his king in the centre. dark-square grip following ... g5, with . . . b6,
c) Black may also be able to delay f4, which . .. .ltb7 and a kingside attack to follow; for ex
is key to White's strategy. One theme in this re ample, 1 5 .ith5+ We7 1 6 l:!.adl .lte5 1 7 .itf3 g5,
gard is ... ltJe5-g6, perhaps in conjunction with etc.
... f6. Along those lines, it's worth noting that if 15 hxg3 b5 16 f4 .ltb7 17 .itd3 0-0
White had avoided the exchange of knights on The position is equal.
d4, then the same idea could be expressed by
... ltJge7-g6. By the time f4 is played (probably
supported by the preparatory move g3), then Sozin Attack (and the
Black will normally have ...b5 and ... .ltb7 in, so
that White has some weakness on the long di
Classica l Sici lian)
agonal which will discourage him from playing
e5. All this is rather exotic and clearly won't be 1 e4 c 5 2 ltJf3
achieved in one game, but it demonstrates the The Sozin Variations are characterized by
same sort of flexibility that we saw in the Paul the move .ltc4, and can arise from either 2 ... e6
sen Variation. or 2 ... d6. The following variation is known as
11 Whl (D) the 'Classical Sicilian' :
Here's another typical set-up for Black: 1 1 2 ... d6 3 d4 cxd4 4 ltJxd4 ctJf6 5 ltJc3 ltJc6
°iih3 0-0 1 2 l:!.adl f6 ! 1 3 .Itel b5 ! ? ( l 3 ... b6 (D)
would prevent the next move). White now has a
common tactic that must always be weighed by
both sides: 1 4 .ltxb5 ! ? axb5 15 ltJxb5 .ltxh2+
16 'iixh2 'iixh2+ 17 Wxh2 l:Ixa2 1 8 l:!.fe l g5 !
1 9 f3 ltJe5 (a pseudo-outpost) 20 Wgl .ltb7 2 1
ltJc3 l:Iaa8 2 2 Wf2 .ltc6 2 3 l:Ih l 1h-1h V.Miiki-
0st Hansen, Gj0vik 1 985 .
This can include a variety of lines but the

most important ones are the Richter-Rauzer
Attack (6 .ltg5), and the Sozin Attack (6 .ltc4),
which is the subject of this section. These
moves both strongly discourage Black from
playing . . . e5.
Instead, 6 .lte2 e5 (D) is the Boleslavsky For one thing, this capture no longer loses a
Variation, one of the original . . . e5 Sicilians that piece and Black retains his bishops.
still discourages players from 6 .lte2. 13 exd5 .itf5 14 c4 0-0 15 'ifa4!? a5
And here the moves ... .l:.c8 and ... ctJb8 help
to set up a blockade on c5 .
16 a3 .itd7 17 'ifdl a4 lS b4 axb3 19 'irxb3
ltJa6 20 .lte3 'ifc7 21 a4 ltJcS! 22 .ltxc5 'irxc5
23 'ifxb7 .l:.c7 (D)
When compared to the Najdorf with 5 . . . a6 6

Sl..e2 e5, it turns out that ... ltJc6 is usually more
useful than . . . a6. An example with typical cen
tral themes:
Apicella - Kramnik Black has the benefit of the bishop-pair. con

Moscow OL 1994 trols the dark squares, and can play against the
weaknesses on a4 and c4. Meanwhile White's
1 e4 c5 2 ctJf3 ltJc6 3 d4 cxd4 4 ltJxd4 ctJf6 5 fl -bishop is pathetically bad. All for a pawn.
ltJc3 d6 6 .lte2 e5 7 ttJf3 h6 24 'ifh3 :as 25 ltJd2 rs
This is played to prevent 8 .ltg5, which would The central majority is a weapon in any line
strengthen White's control of d5, although Black where a knight capture on d5 has been met by
has done well enough with 7 . . . .lte7 too. exd5.
S h3 .lte6 9 0-0 .lte7 10 .:!.el .l:.cS 11 .itfl 26 ctJbl .lth4 27 g3 .itf6 2S ltJc3 e4 29 :al
liJbS! ? (D) 'ifaS
Delaying castling has certain positive ef 29 ... .ltxc3 30 'ifxc3 .l:.xa4 is already bener
fects. for Black but Krarnnik wants more.
30 .:!.cl .lteS 31 .l:.cc2 lk5 32 ctJbS 'ifi>h8
Now even a pawn-storm by ... g5 and ... f4 be
comes a possibility.
33 'ife3 'ifb4 34 ctJd4 .l:.ccS 35 ctJe6 .lha4 36
.l:.xa4 'ifxa4 37 .l:.d2 'ifal! 3S \t>g2 l:tbS 39 tDf4
.:!.bl 40 'ife2 .:!.el 41 .l:.a2 .l:.xe2 42 :xal lhf2+
43 \t>xf2 .ltxal 44 \t>e3 @gS 45 ctJe6 g6 46 c5
.lteS 47 g4 .ltxe6! 48 dxe6 d5
And so forth - Black has three passed pawns!
Another significant difference between 2.. . e6

and the Classical order (with 2 ... d6, 4 ...ctJf6 and
5 ... ctJc6, for example) is that after 6 .ltc4 in the
latter instance, Black has the option of 6... 'iWb6
(the 'Benko Variation' ) rather than transposing
12 ctJdS ltJxdS to a Sozin by 6 ... e6. His idea is to disturb the
d4-knight. White can respond in a number of ( 1 3 . . . .itd7 develops simply and sensibly) 1 4
ways, but by far the most common one is 7 b3 ! (to stop . . . ltJc4) 14 . . . l:tb8 1 5 l:tadl ( a posi
ctJb3, in order to protect the b-pawn and play tional mistake is 1 5 f5 ? ctJc6 ! 16 fxe6 fxe6 1 7
.lte3; for example, 7 ... e6 8 0-0 .lte7 9 .lte3 .ltg5 .lte7 1 8 l:tad 1 ltJe5, Hossain-Goloshcha
"fic7. As I describe elsewhere, this creates a sit pov, Dhaka 2003 ; in this position a maj or
uation in which Black seems to have wasted an plus for Black is that d5 and f5 are unavail
important tempo by . . . "iib6-c7, but White may able to White' s pieces) 1 5 . . . ltJc6 (or 1 5 . . . .itd7
make up for that by playing ctJb3-d4. Lines with equality) 16 .itf2 (heading for g3 or h4)
with . . ."iib6 have become more popular over 16 . . . ctJd7? ! 1 7 .ltg3 ctJxd4 1 8 °fixd4 b5 1 9
time for this reason. axb5 axb5 20 b4 g6 2 1 e5 ! d5 22 f5 ! gxf5 23
ctJxd5 ! "fic4 (23 . . . exd5 ? 24 e6) 24 °fid2 h6 25
1 e4 c5 2 ctJf3 e6 3 d4 cxd4 4 ltJxd4 ctJf6 h3 exd5 26 .itxd5 °fixb4 27 c3 "fic5 28 l:txf5
Here both 4 . . .ltJc6 5 ltJc3 d6 and the Najdorf l:te6 29 l:txf7 ! ctJb6 (29 . . . @xf7 30 °fif4+ \t>e8
move-order 2 . . . d6 3 d4 cxd4 4 ctJxd4 ctJf6 5 31 .ltxe6 is decisive) 30 l:tdfl ctJxd5 31 l:txf8+
ltJc3 a6 circumvent the Keres Attack described °fixf8 32 l:txf8+ \t>xf8 33 °fixd5 and White
in the next note, but of course they have their went on to win easily in Adams-Topalov,
own peculiarities. Wij k aan Zee 2006.
5 ltJc3 d6 (D) 6...ltJc6 7 .ltc4 (D)
6 .lte3 We have arrived at the Sozin Attack. It is sim

6 g4 is the Keres Attack, which has a high ilar to the Najdorf 6 .ltc4 variation but Black's
reputation among players, and accounts for the knight is on c6. That implies the possibility of
fact that the Scheveningen with . . . e6, . . . d6 and earlier simplification by . . . ctJxd4, which ren
...ctJf6 (before other moves) is not played as ders White's idea f4-f5 less effective. Black is
much these days - Scheveningen lines more of also unlikely to have to worry about the sacri
ten arise by transposition from the lines men fice .ltxe6 that was a hallmark of the Najdorf
tioned in the previous note. line. And even the possibility of f4 followed by
However, I am using this sequence of moves e5 can lose force because Black has a natural
in order to lay out some move-order issues and retreating and counterattacking square for his
transpositions. Instead of 6 .lte3, for instance, 6 attacked knight on d7 .
.ltc4 a6 transposes to the 6 .ltc4 Najdorf. White has ample resources, the nature of
The traditional 6 .lte2 can also transpose which are completely dependent upon his choice
into other variations such as the Najdorf with of piece deployments, especially that of the
. . . ltJbd7 ; but if . . . ltJc6 is played soon the varia queen. If she goes to f3, for example, Black will
tions take on their own character. One line is struggle to achieve ... b5 unless he exchanges on
6 . . . a6 7 0-0 .lte7 8 f4 ltJc6 9 .lte3 0-0 10 a4 d4. B ut . . .ctJxd4 brings another piece to the cen
"fic7 1 1 \t>h l l:te8 1 2 .itf3 .itf8 1 3 °fid2 ltJa5 ! ? tre, normally White's bishop, which then aims
at Black' s king. In the absence of that exchange, 10 f4 t?Jxd4 11 i.xd4

the move f5 can still be effective. If White cas 1 1 'ifxd4 runs into l 1 . ..t?Jg4 ! .
tles kingside he will almost certainly play f4 1 1...bS! 1 2 a3
and aim for e5. But the most compelling varia This is the slow approach. The main line for
tions arise when White castles queenside and years, and indeed what still may be the main
plays his remaining attacking weapon, namely line of the entire Classical Sozin, goes 12 e5
g4-g5. It is no coincidence that the g-pawn ad dxe5 1 3 fxe5 tiJd7 1 4 t?Je4 (there's not enough
vance established itself in this variation some force behind 1 4 'ifg4 t?Jc5 1 5 :ad l �b7 1 6
years back and presaged the flood of g4 attacks \t>h l 'ifc7, when the game i s equal) 1 4...�b7
in the Sicilian and other openings. Somehow (D).
modem theory keeps settling upon that move as
the most effective one in long-disputed attack
ing variations.
7...i.e7
Here we'll show three games illustrating the
...e6/...t?Jc6 positions: one in which White cas
tles kingside, another in which his king stays in
the middle of the board, and a modem one in
which he plays 0-0-0.
Sozin with Kingside Cast ling

Fischer - Spassky
Reykjavik Wch (4) 1972
8 0-0 Here we have a quintessential old-style Si

This introduces a traditional and still impor cilian scenario: White's forces aim at the king
tant line. side, including his knight, queen, both of his
8 0-0 9 i.b3
••• bishops, and his rook on the open file. He
White has to watch out for 9 ... d5, and also would like to play 'ifg4 and tiJf6 (even as a sac
for the trick 9 ... t?Jxe4 ! 1 0 t?Jxe4 d5, opening up rifice), and l:tad l , whereas the sacrifice llxf7
the centre and freeing Black's game. might easily enter into the picture.
9...a6 (D) For his part, Black will be sure to target
White's unsupported weakness on e5. Black's
kingside is generally solid and his e6-pawn ne
gates the pressure from White's b3-bishop. He
would also like to simplify, beginning by ex
changing off the bothersome knight on e4. be
fore White can cause him tactical difficulties.
In the meantime Black has his usual control of
the c-file, supporting desirable moves such as
... t?Jc5 at the right moment. Play usually contin
ues 1 5 tiJd6 ( 1 5 'ifg4 i.xe4 1 6 'ifxe4 t?Jc5 is
equal) 15 . . .i.xd6 1 6 exd6 'ifg5 17 l:tf2 ! (this
protects the 2nd rank and prepares 'ifd2; after
any exchange of queens the two bishops will be
a major advantage; 1 7 'ife2 e5 1 8 i.c3 'ifg6 has
been analysed thoroughly following the Short
This position could also have come from Kasparov world championship match, leading
2 ... d6 via the Najdorf 6 i.c4 Variation, as it did to equality with best play) 1 7... a5 ! 1 8 'ife2 :a6 !
in the main game that we are looking at. (White was threatening i.xe6) 1 9 i.c3 (or 1 9
.ixe6 l:!.xd6 with equality) with unclear play; In return for a pawn, Black's bishop-pair
perhaps 19 . . .b4 20 .itd2 'iic5 2 1 .itf4 is best. A rakes the kingside and White's e5-pawn is
great study line! weak. The opening has ended successfully for
12....ltb7 13 °iid3 Spassky. I'll skimp on the notes as we proceed
1 3 'iie l a5 ! with ... b4 next; then 1 4 ctJxb5 ? ! through the rniddlegame:
ltJxe4 (or 14 . . . a4 and then . . .ltJxe4) conquers 19 'iie2
the centre and eliminates White's attacking Here 1 9 'iig3 ! 'iixg3 20 hxg3 improves. Then
chances. 20 . . . .lta6 2 1 a4 .ltxb5 22 axb5 .ltd4 is only
13 aS! (D)
•.• nominally better for Black.
19 l:!.ad8 20 l:!.adl l:!.xdl 21 l:!.xdl h5!?
••.
2 1 ... .lte3 ! 22 ctJd6 .ltc6 is better, when a cute

line is 23 l:!.fl .ltf4 24 °iif2 'iig4 ! with the idea
of ... °iih3 ! .
2 2 ltJd6?! .lta8 23 .ltc4
Not 23 l:!.fl ?! h4 24 ltJxf7 ? h3 ! 25 ltJxg5
hxg2+ 26 'iixg2 l:!.xfl#.
23 h4 24 h3 .lte3 25 'iig4 'fixes
•••
As the centre pawn falls, so does White's

ability to control the enemy pieces and keep his
own out of trouble.
26 °iixh4 g5! 27 'iig4 .ltc5! 28 ctJbS \t>g7
Now ... l:!.h8-h4 looms.
29 liJd4 (D)
This pawn sacrifice diverts White from the

centre by threatening . . . b4.
14 e5?! B
The right spirit, but weakening. A better try
is the aggressive 14 f5 !?, but then Black can
counter by 14 ...b4 1 5 axb4 axb4 with approxi
mate equality after 1 6 °iib 5.
14 dxeS 15 fxe5 ltJd7 16 ltJxbS
.••
1 6 ltJe4 .ltxe4 ! 1 7 'iixe4 ltJc5 1 8 .ltxc5

.ixc5+ 19 \t>h l °iid4; that pawn on e5 is a struc
tural problem, so White has to be wary of too
much simplification.
16 ltJcS 17 J.xc5 J.xc5+ 18 \t>hl 'iig5 (D)
••.
29 l:!.h8?!
•••
But now 29 ... l:!.d8 ! 30 c3 °iie3 ! was terribly

strong. Spassky's model use of the bishop-pair
falls short only for tactical reasons.
30 ctJf3 .itxf3 31 °iixf3 .itd6?
3 1 . . .l:!.h4 ! may still have been winning, one
line being 32 l:!.fl l:!.f4 33 °iie 2 l:!.xfl + 34 °iixfl
.itd6 (34 ... °iixb2) 35 \t>gl °iih2+ 36 \t>f2 .ltc5+
37 @el 'iie5+ 38 °iie2 °iixb2 and the opposite
coloured bishops are still helping Black.
32 °iic3! 'iixc3 33 bxc3 .lteS 34 l:!.d7 �6 35
\t>gl
and the game was drawn shortly thereafter.
Short - Kasparov 1 6....ltb7) 1 7 exd6 bxc3 18 dxe7 cxb2+ 19 �b2

London PCA Wch (12) 1993 'ifxe7 with the idea of ... a5, . . . .ltb7 and ... l:.fc8.
15....ltb7 16 ltJe4 dxe5!
8 .ltb3 0-0 9 f4 a6 10 'iff3 Threatening . . . 'ifd4 among other moves.
This attacking move used to be popular, un 17 l:!.gl g6 18 l:!.dl .ltxe4 19 'ifxe4 'irxh4+
derstandably, since it gets ready to castle queen With an attack. At the very least Black can
side and blow the opponent off the board. But get three pawns for the piece, but White bas
the queen on f3 is subject to harassment on the some activity, so an assessment of 'equal' seems
long diagonal, especially in the game line. fair. Many similar tactical themes appear in the
10 ltJxd4 11 .ltxd4 b5 (D)
.•• positions with . . . e6 and . . .b5 .
Ve limirovic Attack
Boto - Buntic
Bosnia 2001
8 'ife2 (D)
12 .itxf6!
12 e5 dxe5 hits the d4-bishop, almost forcing
1 3 .ltxe5 ( 1 3 'ifxa8 'ifxd4 1 4 'iff3 exf4 already
gives Black two pawns, the bishop-pair and a
load of weaknesses to work on, all for an ex
change) 1 3 ... l:!.a7 ! 1 4 l:!.dl l:!.d7 1 5 0-0 .ltb7
with the better pawn-structure and position.
12....itxf6! 13 e5 .lth4+ 14 g3 l:!.b8! (D) This move, together with queenside castling,
characterizes the Velimirovic Attack. Within
hundreds of brilliancies that have been played
by both sides of this opening, we find certain
themes that are fundamental to attacking in the
Sicilian Defence. Many of them were first
played in games with this variation, or at least
brought to prominence by their use in them. I'll
try to show a few of these essential building
blocks of Sicilian attacks.
Looking over the older games by Veli�
vie himself, you see the tactical philosophy ex
pressed by Kasparov, who stresses 'cutting the
board in two' , resulting in attractive-looking
pieces uselessly stranded from defence of the
king.
15 gxh4 8 ...a6
The superiority of Black's pawn-structure 8 . . . 0-0 9 0-0-0 'ifa5 was played in the famous
shows in lines like 1 5 l:!.fl .lte7 1 6 0-0-0 b4 ! (or encounter Fischer-Geller, Skopje/Krusevo/Ohrid
1967. It isn't too stunning by Velimirovic At

tack standards, but since most of the fun in this
section will be White's, I ' ll show how Black
fights back when apparently lost: 10 .ib3 ctJxd4
l l .ixd4 .id7 1 2 @b l .ic6 1 3 f4 .l:.ad8 14 .l:.hfl
b5 15 f5 b4 16 fxe6 bxc3 17 exf7+ @h8 18 .l:.f5
1iib4 19 'iiVf l lbxe4 20 a3? (20 1iif4 !) 20... 1iib7
2 1 1iif4 .ia4 ! ! (a brilliant move that turns the
tide) 22 1iig4 .if6 23 .l:.xf6 .ixb3 0- 1 .
9 0-0-0 (D)
.id4 ctJg4 24 1iif3 .l:.d7 25 1iih3 ! (wonderful ge

ometry) 25 . . . ctJe5 26 f4 .ih4 27 .l:.e2 .l:.e7 28
fxe5 (finally recovering his material) 28 . . . dxe5
29 .ic5 ! .ig5+ 30 @b1 f6 3 1 1iih5+ .l:.g6 32 h4
1iic 8 33 i..xe7 1iixf5 34 .ib4 1iif4 35 1iixh7
.l:.h6? 36 'iiVe7# ( l -0) Velimirovic-Popovic, Novi
Sad 1976.
10 .ib3 1iic7 11 .l:.hgl
Let's look at another Velimirovic tour de
force and representative of the themes that he
brought to the fore: 1 1 g4 ctJd7 12 ctJf5 ! exf5 13
In this position Velimirovic's first idea was ctJd5 1iid8 14 gxf5 ctJa5 15 lbxe7+ 1iixe7 16
the uninhibited g4-g5 followed by whatever .id5 @h8 17 .l:.hgl ctJf6 (D).
was necessary to get at Black's king. Then at
tention focused mainly upon g4 and �hg l , with
precisely the same strategy but differently exe
cuted. Sometimes White has also succeeded
after f4 and either f5 or e5, but that hasn't estab
lished itself as well as the other two.
9...0-0
Here is one of Velimirovic's games in the
first days of the Attack. The thing that shocked
people about this and games in the next notes
was not that sacrifices like ctJf5 and ctJd5 were
being made but how slow they seemed to be and
how little material was needed to make the at
tacks work: 9 ... 1iic7 1 0 .ib3 ctJa5 1 1 g4 b5 1 2
g 5 ctJxb3+ 13 axb3 ctJd7 14 ctJf5 ! ! (D).
The ' ! ! ' comes from annotators at the time of 1 8 1iif3 ! (these relatively slow moves char
the game, deservedly so for the attack's origi acterize the Attack to this day) 1 8 ... ctJxd5 19
nality; these days the idea is second nature, but .l:.xd5 ctJc4 20 f6 ! ! 1iixf6 2 1 1iixf6 gxf6 22 .id4
the specific tactics and White's sustained attack lbe5 23 f4 ctJd7 24 .l:.xd6 .l:.g8 25 �dl �e8 26 f5
are still mind-boggling. 14 . . .exf5 15 ctJd5 1iid8 �xe4 27 .l:.gl h5 28 .l:.g5 ! .l:.g4 29 .l:.xf6 ! .l:.gl +
1 6 exf5 .ib7 17 f6 gxf6 1 8 .l:.hel .ixd5 (Black's (29. . .@h7 3 0 �xh5+ @g8 3 1 .l:.h8+ !) 3 0 @d2
exchanging all the pieces off - this must be .l:.g2+ 3 1 @e3 1-0 Velimirovic-Bukal, Yugosla
right) 19 .l:.xd5 .l:.g8 20 gxf6 ctJxf6 21 .l:.f5 (still via 1 97 1 .
a full piece down) 2 1 .. ..l:.b8 22 .ia7 .l:.b7 23 1 1 ctJd7
•••
Now we're getting around to some real fire 23 ctJc7 ctJc5 24 lbxa8 .ie6 25 1iie2 1iixa8 26
works. Try this out: l l . ..b5 1 2 g4 b4 1 3 lbxc6 1iih5 @g8 27 .ixc5 dxc5 28 f5 .id5 29 f6 (a gor
1iixc6 1 4 ctJd5 exd5 1 5 g5 dxe4 1 6 gxf6 .ixf6 geous attack, begun 1 5 moves before! ) 29...l:!.d8
1 7 .id5 1iia4 1 8 1iih5 ! .ie6 (D). 30 f7+ @h8 3 1 1iih4 a5 32 �gel a4 33 ,.xd8
1iixd8 34 �e8 1iig5+ 35 @bl 1iixg6 36 l:!.xf8+
@h7 37 �h8+ @xh8 38 f81iV+ .ig8 39 l:!.d8 l -0
Velimirovic-B.I vanovic, Niksic 1 978.
Ultimately you could say that it's White's
positional advantages (space, occupation of
d5, harmonious piece placement, and so on)
that allow these attacks to succeed, as indi
cated by their duration and the absence of di
rect tactics for so many moves after the
sacrifice.
12 g4 ctJxd4
Now let's allow Shirov to show his amazing
talent from the black side of the board: 12 . . . ltJc5
1 3 ctJf5 b5 ! 14 .id5 (D).
19 �xg7+ ! ! .ixg7 20 �gl �fc8 (20 ...�ac8

2 1 �xg7+ @xg7 22 .id4+ f6 23 1iig5+ @f7 24
1iixf6+ @e8 25 1iixe6+, etc.) 2 1 �xg7+ @xg7
(2 1 . . .@f8 22 �xf7+!) 22 1iih6+ @g8 23 .ixe4
b3 (23 ... �c4 24 .ixh7+ @h8 25 .ig5 �f4 26
.if5+) 24 .ixh7 + @h8 25 .if5+ @g8 26 1iih7+
@f8 27 .ih6+ @e8 28 1iig8+ @e7 29 .ig5+
@d7 30 1iixf7+ @c6 3 1 .ixe6 @b6 32 .ie3+
@a5 33 .ixc8 �xc8 34 1iif5+ �c5 35 .ixc5 1iib5
(35 ... bxa2 36 b4+ @b5 37 'iiVd7+) 36 .ib4+ @xb4
37 a3+ @c4 38 1iixb5+ axb5 39 cxb3+ @d3 40
@d l 1-0 Ostapenko-Yartsev, USSR 1 969.
But let's enjoy another of those classic games
by the master: l l . . .ctJa5 1 2 g4 b5 1 3 g5 ctJxb3+
1 4 axb3 ctJd7 15 f4 b4 (D). 14 . . . .ib7 ! (he'll j ust continue to leave every
thing hanging ! ) 1 5 g5 �fc8 16 �g3 .if8 ! 17
1iih5 g6 1 8 ctJh6+ @h8 1 9 1iih4 b4 20 .ixc6
bxc3 21 .ixc5 (I'm ignoring the mistakes; ob
viously any game like this can't be flawless )
2 1 . . .cxb2+ 22 @bl .ixc6 23 �xd6 ! .ia4 ! ! 24
�c3 .ixd6 ! 25 .id4+ e5 26 �xc7 l:!.xc7 27
.ixb2 �b8 ! 28 @al �xb2! 29 lbxf7+ �xf7 30
@xb2 �f3 3 1 1iig4! .ia3+ 32 @al ! �xf2 33
1iih3 ! (it goes on and on ! ) 33 ... .ie7 34 1iic 8+
�f8 35 1iixa6 .ixc2 36 h4 .ic5 37 @b2 .ixe4
38 1iie6 �b8+ 39 @c3 .id4+ 40 @c4 .if5 41
1iif7 �c8+ 42 @d5 .ih3 43 h5 gxh5 44 ,.f3 !
.ig4 45 1iif6+ @g8 46 g6 hxg6 47 1iixg6+ @f8
48 1iif6+ @e8 49 1iig6+ @e7 50 1iig7+ @d8 5 1
1iif6+ @c7 5 2 1iic6+ @b8 5 3 1iib5+ @a7 54
16 ctJf5 ! exf5 17 ctJd5 1iid8 1 8 exf5 �e8 1 9 1iia4+ @b7 55 'iiVb4+ .ib6 56 @xe5 �c5+ and
g6 ! fxg6 20 fxg6 h6 2 1 1iic4 @h8 22 .id4 .if8 Black won in Onishchuk-Shirov, Bundesliga
MASTERING THE CHESS OPENINGS
.,.,, _,_... .U of these games will give

_ After all those ideas, 20 ... exd4 21 °Yi'h6! looks
J1119 - s lbe best possible lessons in Sicil pedestrian.
ian tactics and combinations. 21 gxf7 .if6 22 .l:.xg7! .ixg7
13 .i.xd4 bS 14 gS b4 15 °Yi'hS ctJeS (D) Or: 22... .ixf5 23 .l:.xh7+ ! ; 22... .ie6 23 .ixe6
After 1 5 ...bxc3, there's nothing fancy: 1 6 .ixg7 24 f6 .l:.xf7 25 .ixf7 °Yi'xf7 26 fxg7+
l:t.d3! and wins. °Yi'xg7 27 .ixc3.
23 f6 °Yi'd8 24 .l:.gl 1-0
The above presentation may have been self

indulgent, but those attacks are the Velimirovic
Attack, and to understand them is to understand
the variation. More significantly, the same at
tacking themes quickly spread to the practice of
the Sicilian Defence in general and appear in
multitudinous variations today.
Accelerated Fianchetto
1 e4 cS 2 ctJf3 ctJc6 3 d4 cxd4 4 ctJxd4 g6 (D)
16 f4 lbg6 17 fS! lbf4

A wonderful variation that you shouldn't
miss: 17 ... bxc3 1 8 .l:.dfl ! cxb2+ 1 9 @b 1 lbe5
20 .l:.f4 f6? 2 1 .l:.h4 lbc4 22 .ixc4 "Yi'xc4 23
°Yi'xh7+ @f7 24 °Yi'g6+ @g8 25 °Yi'xg7+! @xg7
26 gxf6++ @f7 27 .l:.g7+ @e8 28 .l:.xe7+ @d8
29 .ib6+ °Yi'c7 30 .ixc7#.
18 °Yi'f3 eS 19 g6! ! (D)
This system is similar to the Dragon (into

which it often transposes), so my treatment will
mainly concern White's most important chal
lenge to Black's move-order, one that is not
available in the Dragon:
5 c4
This variation is known as the 'Maroczy
Bind', and indeed the same name is often ap
plied to White's pawn-structure when it arises
19 bxc3
.•• in other openings.
The variations are amazing: l 9 . . .hxg6 20 Before moving on to it, let me point out a few
°Yi'xf4! exf4 2 1 .l:.xg6 and Black is helpless; unique features in the Accelerated Fianchetto
l 9 . . ..if6 20 °Yi'xf4! exf4 (20 ...exd4 21 °Yi'h6! !) after the normal-looking 5 lbc3 .ig7, which is
2 1 .ixf6 @h8 22 .ixg7+ @xg7 23 gxf7+ @h6 loaded with tricks and positional traps:
24 .l:.d3. a) After 6 ctJb3, 6 ... ctJf6 7 .ie2 0-0 8 0-0 d6
20 °Yi'xf4! @h8 leads us back to a Classical Dragon. Instead,
the Accelerated Fianchetto move-order gives b4 1 ) At this point the move 7 ...,.a5 has
Black another option in 6... .ixc3+ ! ? 7 bxc3 some other tricks associated with it. For exam
ctJf6, trying to exploit White's doubled c-pawns ple, 8 1iid2? lbxe4 ! 9 lbxc6 1iixc3 ! !, or 8 f3?
at the cost of losing the important dark-squared 1iib4 ! 9 .ib3 ctJxe4 ! . White should simply play
bishop; see the section of Chapter 3 devoted to 8 0-0 0-0 9 ctJb3 1iic7 10 f4 d6 1 1 .ie2 with a
doubled c-pawns for a short discussion of pre sort of Classical Dragon in which the queen is
cisely this position. arguably a little misplaced on c7. This line has
b) 6 .ie3 ctJf6 (D) and then: proven a disincentive for those who are consid
ering playing 7 . . . 1iia 5.
b42) Black usually plays 7. .. 0-0 8 .ib3! (an
other trick is 8 f3 'iiVb6 ! with the ideas of fue4
...
and ... lbg4 as well as the direct .....xb2) 8 . .d6 .
(Black has speculative options such as 8 . a5 !?.

. .
an extremely complicated line; however, top

masters who have specialized in that move have
usually abandoned it) 9 f3 .id? 10 ..d2. This
gives Black one extra opportunity to steer clear
of the main lines: 10 ... lbxd4 ! ? 1 1 .i.xd4 b5
with a complicated game that seems to favour
White slightly. On the flip side, many players
and theoreticians feel that 10 ...l::tc8 1 1 h4! saves
a critical tempo over 1 1 0-0-0 and leads to an
advantage for White. This is all in the books (or
bl) 7 f3? ! 0-0 8 1iid2, in order to get into a at least most of it is), and will certainly repay
Yugoslav Dragon, allows Black to free his game study.
immediately in the classic fashion: 8 ... d5 ! and We now return to 5 c4 (D):
White should probably simplify by 9 exd5 ctJxd5
1 0 lbxc6 bxc6 1 1 ctJxd5, with rough equality,
before he stands worse due to the weaknesses
that f3 has created; see how Black has saved a
tempo by playing ... d5 rather than ... d6 and ... d5.
b2) Likewise, the Classical moves 7 .ie2
0-0 8 0-0 can be answered by 8 ... d5 ! .
b3) Therefore White might want to play in
the style of the Reversed Dragon by 7 ctJb3 0-0
8 .ie2, if slower play is his inclination.
b4) 7 .ic4 (D) and then:
After 5 c4, White has a large space advan

tage that will usually dominate the centre for as
long as White maintains the c4/e4 structure. On
the negative side, he has a weakness on d4
(much as in a queen's pawn opening where
White plays d4, c4 and e4). White's plan is to
use his superiority in space to expand and throt
tle Black's position. All three areas of the board
are available, but he will normally use the centre
and queenside. Many endgames favour White.
and in particular Black has to be sure that with position) 1 5 ... °Yi'xd2 1 6 i..xd2 tZ'ixd5 1 7 exd5
his queen on a5 and White' s on d2, the move ( 1 7 cxd5 i..d4+ 1 8 @hl i..d7 with equality)
tZ'id5 won' t be effective. l 7 ... i..d4+ 1 8 @h l i..d7 and Black has no prob
Black would like to achieve the break . . . b5 in lems, Uribe-Perelshteyn, Oropesa del Mar U-
order to chip away at White's centre; obviously 1 8 Wch 1 998.
this usually involves . . . a6. Sometimes he can 6 i..e3
play .. .f5 for the same purpose, but that is un 6 tbc2 ! ? tZ'if6 7 tZ'ic3 can be a very irritating
common until later in the game. Finally, he sequence for Black because it prevents ex
would like to work on the dark squares, espe changes and increases White's control over d5.
cially in view of the unprotected state of d4. The succeeding play is rather technical, but
Some specifics follow: White will continue i..e2, 0-0, and aim to gain
more space by b4, while Black will play . . . 0-0,
Bareev - Pavlovic . . . d6, perhaps with . . . a5 and ... i..e6 depending
Plovdiv Echt 2003 upon what White does. The analogous Rubin
stein Variation of the English Opening goes 1
5 i..g7
.•. c4 c5 2 tbc3 tbf6 3 g3 d5 4 cxd5 tZ'ixd5 5 i..g2
This is the traditional main line of the Mar tZ'ic7 6 tZ'if3 tZ'ic6 7 0-0 e5. Even a tempo down,
oczy Bind. Black has reasonably good prospects.
5 . . .tZ'if6 6 tbc3 d6 (6...tZ'ixd4 7 °Yi'xd4 d6 is Incidentally, if Black likes one of the options
another well-known idea, when White has vari with an early . . . tZ'ixd4, that exchange will pre
ous ways to proceed, including 8 i..g5 i..g7 9 vent the tZ'ic2 variation.
°Yi'd2) 7 i..e2 tbxd4 8 °Yi'xd4 i..g7 9 i..e3 0-0 1 0 6 tZ'if6 7 tZ'ic3 0-0 8 i..e2 d6 9 0-0 i..d7
••.
°Yi'd2 i..e6 1 1 0-0 °Yi'a5 1 2 .l:.ac l ( 1 2 .l:.fc 1 puts 9 ...tbxd4 10 i..xd4 i..e6 is a long-studied line
both rooks on the queenside, which seems a which has lost some of its popularity. Needless
good idea; White would like to play f3 and in to say, that may be only a temporary situation.
some cases .l:.abl and/or tZ'id5; for instance, 10 °Yi'd2 tZ'ixd4 11 i..xd4 i..c6 12 f3 (D)
12 ... .l:.fc8 1 3 f3 ! with the idea 1 3 . . . i..xc4? 1 4
tZ'id5) 1 2 . . JHc8 (D).
White's development has been natural and

normal. He still controls more space and is
We have reached a standard position. Here's ready to attack in the centre and on the queen
an example of a successful anti-Maroczy Bind side.
idea for Black: 13 b3 a6 1 4 f3 ( 1 4 f4 b5 1 5 f5 is 12 aS
...
a highly-charged line with plenty of theory to Preparing to take over the dark squares.
study) 1 4 . . . b5 ! 1 5 tZ'id5 (after 1 5 cxb5 axb5 1 6 13 b3 tZ'id7!
tZ'ixb5 { 1 6 i..xb5 .l:.xc3 1 7 "Yi'xc3 °Yi'xb5 and This is the point of Black's system: he wants
Black has won material } 1 6 ... °Yi'xd2 1 7 .l:.xc8+ to end up with a wonderful knight on c5 oppos
i..xc8 1 8 i..xd2 .l:.xa2 Black has the more active ing a restricted light-squared bishop.
14 i.e3! 20...1iih8 (D)

14 i.xg7 @xg7 used to be played, but White
wants to keep his good bishop in order to sup
port his queenside play. Otherwise he has noth
ing to challenge the c5-knight with.
14 lbcs (D)
••.
The purpose behind . ..1iib6, ... �fc8, ... h5,

. . . 1iid8 and ... h5 ! Black tries to maximize his
pressure on the long diagonal. But White still
has more space and central control, so he can't
be too worried. The opening is finally over and
15 �abl both sides have followed their plans. Bareev
The basic idea is simple: �fc 1, a3 and b4. The proceeded to win the battle of ideas, at least this
execution turns out to be more complicated. time:
15 ...1iib6 21 ctJf4 b6 22 @b l
More dark-square control. He wants to re 22 ctJd5 also leads to some advantage after
strain b4, and also to connect rooks. 22 . . . i.xd5 23 exd5 and b4 to follow after due
16 �fcl �fc8 17 �c2! preparation. White's bishops are aiming the
17 a3? lbxb3 ! 18 i.xb6 ctJxd2 19 �b2 lbxc4 ! right way.
20 i.xc4 i.d7 turns out to be good for Black. 22...@g8?! 23 b4 axb4 24 axb4 lbd7 25
Now White is ready for a3. ctJd5 i.xd5 26 cxd5 .l::txc2 27 1iixc2 .id4 28
17...1iid 8! 18 i.fl 1iic 6!
After Black's last move, 1 8 a3 a4 ! 19 b4 ctJb3 White has a large advantage that he con
is unclear. verted in good order.
18...hS ! ?
A wild idea: Black isn't doing anything seri Clearly, one of the key questions surrounding
ous on the kingside, as we shall see. He just the Accelerated Fianchetto is whether Black
wants to redeploy. can actually gain anything significant from the
19 a3 'Accelerated' aspect of it, by comparison with
Another game proved the strength of White's the standard Dragon. If not, then why allow
queenside pressure: 19 'iiVe l i.e5 20 i::td l 1iib6? White the extra option of the Maroczy Bind.
21 lbb5 ! i.xb5 22 cxb5 'iiVa7 23 i.c4! (now that which at the very least reduces Black's chances
all of Black's pieces are on the other wing, of playing for a win? After 1 e4 c5 2 ctJf3 ctJc6 3
White turns to the kingside) 23 . . . 1iib8 24 f4 d4 cxd4 4 ctJxd4 g6 5 lbc3 i.g7 6 i.e3, etc., you
.tf6 25 e5 ! dxe5 26 i.xc5 .l::tx c5 27 i.xf7+! will see that even in this best of worlds, Black
@xf7 28 �xc5 and White was winning in may not achieve all that he wants. That is, Black
Agrest-Brynell, Nordic Ch (Bergen) 200 1 . uses a bag of tricks in order to lure White away
1 9...@h7 20 lbe2!? from the Yugoslav Attack, but a moderately
White is heading for d4 or f4. The knight knowledgeable opponent will know how to
was also a target in some lines in which White avoid the pitfalls and return play to the main
played b4. line Yugoslav channels. Black indeed gains in
some respects by limiting White's options: spe The main responses are 2 ...d5 and 2 ... tZ'if6.
cifically, White has to commit his bishop to c4 We' ll have only a partial look at those but with
and has lost the opportunity to play the popular relevant details. Other fairly respectable con
antidote to the Dragon involving 9 0-0-0. But tinuations include 2 . .. b6, 2 . . .d6 (and perhaps
Black needs to be clear that he can only avoid even 2 . . .1iia5 ! ?), but I'll skip those and talk
the 9 .ic4 version of the Yugoslav Attack by briefly about some alternatives that are better
playing sidelines such as 8 . .. a5, which are un known:
likely to equalize fully. a) 2 ... e6 is discussed under the order 2 tZ'if3
None of this should discourage a player who e6 3 c3 in the 'Introduction to 2 ... e6' , except for
is below master level, of course. There will al the line 3 d4 d5 4 exd5 (4 e5 is the Advance
ways be plenty of winning chances against av French, an important transposition to bear in
erage competition, even with the Maroczy Bind. mind) 4 ... exd5 (4 . . . 1iixd5 will usually trans-
I do think, however, that you probably won't pose) 5 .ie3 when 5 ... cxd4 6 .ixd4 tZ'ic6 7 .ib5
want to play this system for life. a6 is considered equal. The plan discussed via
the 2 tZ'if3 move-order of 5 ... c4 still has the ef
fect of making .ie3 look like an unnecessarily
Alapin Variation passive move and with care to bring his pieces
out quickly Black should stand satisfactorily.
l e4 c5 2 c3 (D) b) 2 ... g6 3 d4 cxd4 4 cxd4 d5 will often
transpose to the variation 1 e4 c5 2 tZ'if3 g6 3 c3
.ig7 4 d4 cxd4 5 cxd4 d5 . The main lines are 5
exd5 (5 e5 .ig7 is similar to the transposition
mentioned; then 6 tZ'if3 .ig4 7 .ib5+ tZ'id7 has
been played, among others) 5 .. . tZ'if6 6 tbc3 (6
tbf3 tZ'ixd5 7 tZ'ic3 .ig7 comes directly from that
line) 6 ... .ig7 7 .ic4, when Black chooses his
method of regaining the pawn: ... a6, or ... tZ'ibd7
and . . . tZ'ib6. According to current knowledge,
he is able to get a quite playable game and come
close to achieving full equality.
Counterattack with ... dS
2 d5 3 exd5 1iixd5 4 d4 tZ'if6

•••
Known as the 'Alapin Variation' , 2 c3 partic 4 ... tZ'ic6 5 tZ'if3 .ig4 is also played, when one
ularly appeals to those turned off by the mas ambitious line for White is 6 .ie2 cxd4 7 cxd4
sive theory associated with the Open Sicilian. e6 8 tZ'ic3 'iiVa5 9 h3 .ih5 1 0 d5 ! ? but 1 0 . . . exd5
On the positive side, White tries to build up a 1 1 tZ'id4 tZ'ixd4 ( l 1 . ...ixe2) 1 2 .ixh5 tZ'ic6 was
centre with minimal risk. Generally, however, 2 easy enough (and equal) for Black in Nayer
c3 lacks punch and might not appeal to the at Lautier, Khanty-Mansiisk FIDE WCup 2005 .
tacking player. 5 tZ'if3 (D)
My stated philosophy in this book is to exam 5 .tg4
•••
ine the most 'important' openings, especially 5 ... tbc6 is an important alternative for those
older and established ones that have played a who are unhappy with some aspect of 5 ... .ig4,
leading role for many years. For the purposes of perhaps 6 dxc5 in the next note. Play usually
presenting a Sicilian variation that isn't in that continues 6 .ie2 cxd4 (or 6 ... e6 7 0-0 cxd4) 7
mould, I've chosen the Alapin Variation instead cxd4 e6 8 tZ'ic3 1iid6 9 0-0 .ie7 . Black wants to
of, for instance, the Closed Sicilian, because it play . . . 0-0, ... b6 and . . . .ib7, with ....l:.fd8 in
has some universal ideas that are applicable to some cases. White can build up by .ie3, 1iid2
other lines in this book and opening study in and .l:.fd 1, but his position would contain little
general. dynamic potential. Therefore White sometimes
line goes 6 . . .1iixc5 ( 6 ...'iiVxd 1 + 7 @xd 1 e5 8 b4

e4 9 h3 has been tested and argued about for
some years now; most players seem to shun it
as Black) 7 ctJa3 (7 .ie3 1iic 7 8 h3 .ih5 9 ctJbd2
ctJbd7) 7 ... ctJbd7 8 h3 .ih5 9 .ie3 1iic 8 ! and
Black should have a satisfactory position. That
assessment is not shared by everyone, however.
6...e6 7 .ie3
7 c4 1iid7 only serves to expose White's cen
tre.
7 cxd4
•••
Now that 8 dxc5 is a threat (in some cases c4

is as well), Black exchanges. But by delaying
he has committed White's bishop to the rather
tries to force the pace: 10 ctJb5 'iiVd8 (10 ...1iib8 !? passive post on e3.
1 1 g3 ctJd5 1 2 .ic4 a6 1 3 .ixd5 axb5 14 .ie4 8 cxd4 ctJc6 9 ctJc3 1iid6 (D)
favours White's active pieces) 1 1 ctJe5 ! ? ( 1 1
.if4 ctJd5 1 2 .ig3 0-0 { or 12 ... a6 } 1 3 .ic4 a6
14 .ixd5 exd5 1 5 ctJc7 .l:.b8 with equality)
1 1 . ..0-0 ( l l ....id7 1 2 ctJxd7 'iiVxd7 1 3 .ie3 0-0
is also fine: d5 is permanently blockaded) 1 2
lbxc6 bxc6 1 3 ctJc3 .l:.b8 (D) .
This retreat is better than another (for exam

ple, 9 ... 1iid 8) for two reasons: it stops the active
move .if4 and allows Black to increase the
pressure on White's d-pawn after ....l:.d8 at some
point. The dark-squared bishop belongs on e7
anyway.
We see this kind of position in several open 10 0-0 .ie7
ings. As long as Black can develop quickly and We've reached a standard isolated pawn posi
use the b-file, his isolated c-pawn is not a prob tion in which White will pit his activity against
lem. White's d-pawn is just as exposed and is Black's more static advantages, primarily pres
obviously not going anywhere if Black doesn' t sure on the isolated queen's pawn and well
exchange it. Neither side can claim much, if placed pieces. White may nudge Black's bishop
any, advantage, but either side can play for a to h5 by h3 and reserve the move g4 for later.
win. He sometimes builds up by means of 'iiVb 3 fol
6 .ie2 lowed by bringing a rook to d l . We saw many
6 dxc5 was brought to the forefront about a examples of this type of position in the intro
decade ago and has enjoyed a steady popular ductory chapters. This particular one appears
ity. That may say less about the move's merits unpromising for White, since his pieces are less
than it does about White's difficulties in getting active than in those positions and d5 will be ex
an advantage with 2 c3. In any case, the main tremely hard to achieve.
The 2 ... lbf6 Variation
2 lbf6 3 es liJdS (D)

••.
6 ctJb6
.•.
If Black plays 6 ... e6 7 cxd4 d6, we again

have the line referred to in the note to 5 ... ctJc6.
7 .ib3 d5
This set-up resembles an Alekhine Defence, Capturing the offered pawn by 7 ...dxc3 8
and Black would very much like to have . . . c5 in lbxc3 is risky.
if he were playing that opening ! But it's not so 8 exd6 'Jilixd6 (D)
simple, because White is not intending to kick
the d5-knight around with pawns when he can
develop his pieces instead. The Alapin with
2 . . .ctJf6 can lead to rather theory-heavy play.
Here's an outline of some of the wide range of
continuations:
4 d4
4 ctJf3 lbc6 5 ii.c4 ctJb6 6 ii.b3 c4 7 ii.c2 g5 ! ?
is another eccentric line; current theory has it as
equal, whatever that means in such an unbal
anced position.
4...cxd4
The usual starting-point. Only a 2 c3 special
ist would know if White has any way to squeeze
something from the position.
5 ctJf3 Black is generally thought to have equality
White can also play the direct 5 cxd4 d6 here, although as usual there are lengthy analy
(5 ... e6 6 ctJc3 lbxc3 7 bxc3 'Ji/ic7 8 .id2 b6 has ses of variations which, fortunately, your aver
been a popular system in the past) 6 ctJf3 lbc6, age opponent will never have heard of.
when a traditional line is 7 .ic4 (or 7 ctJc3 9 0-0
lbxc3 8 bxc3 e6) 7 ... ctJb6 (7 ... e6) 8 .ib5 dxe5 9 Among many other established lines is 9
lbxe5 .id? I O .ixc6 .ixc6 l l ctJxc6 bxc6, giv ctJa3 a6 1 0 0-0 .ie6 1 1 .ixe6 'Ji/ixe6 with equal
ing us the standard structure discussed above in ity.
the 2 . . . d5 line. The game is equal. 9 .ie6 10 .ixe6 'Ji/ixe6 1 1 ctJxd4 ctJxd4 12
•.•
5... ctJc6 (D) 'Jilixd4 .i::td8 13 'Ji/ih4 'Ji/ie2 14 ctJd2

5 ... d6 6 cxd4 e6 leads to a line discussed in Now Black can play 14 . . . h5 ! ? or 1 4 ... g6,
the introduction to the 2 ctJf3 e6 section. both of which have been satisfactory for him.
6 .ic4 Nevertheless, White may be interested in play
6 cxd4 d6 7 ctJc3 can be met by 7 ... e6 8 ing such a position because it's double-edged
ctJxd5 exd5 with equality. enough to be interesting.
1 2 Ca ro - Ka n n Defence
l e4 c6 (D)
for White for a couple of reasons. For one thing

it' s a target of a bishop on f5 and sometimes a
This is the official starting-point for the knight on b4; but more significantly it's not up
Caro-Kann. Now 2 d4 is played in a clear ma on c4 to make the centre more mobile, nor on
jority of games, though naturally White can try c3 where it would protect White's d4-pawn.
to react to the idea of 2 ... d5 in other ways. As Furthermore, you should note that without a c
always, these early deviations can be very edu pawn, White has the handy move 1iib3 if Black's
cational, and one of them sets up a respectable c8-bishop strays from the queenside.
structure that is played on a regular basis: Thus, instead of 4 . . . 1iix d5, Black almost al
ways plays 4 . . . ctJf6 5 ctJc3 (5 i.b5+ will even
Caro-Kann Two Knights tually lead to White losing back his d-pawn
after either 5 . . . i.d7 or 5 . . . ctJbd7; in the latter
2 ctJf3 case, . . . a6 and either . . . b5 or . . . ctJb6 can fol-
2 c4 is another important alternative to the low) 5 . . . lbxd5 6 d4 and we're back to the iso-
main lines: 2 . . . d5 (2 . . . e5 3 ctJf3 d6 4 d4 is some lated queen's pawn position that characterizes
sort of Old Indian Defence that most Caro the Panov Attack.
Kann players won't be comfortable with) 3 The text-move (2 ctJf3) is easily White's most
exd5 cxd5 usually transposes to the Panov At promising independent try and deserves a look
tack by means of 4 d4 - see later in this chapter. for those who want a somewhat less-travelled
White can also try 4 cxd5, when 4. . . 1iixd5 loses path.
a tempo after 5 ctJc3. However serious that may 2 d5 3 ctJc3 (D)
...
or may not be, White will follow up with d4 and This sequence tries to use piece-play and
some advantage; e.g., 5 . . . 1iia5 6 d4 ctJf6 7 ctJf3 quick development to cause Black discomfort.
(D). For instance, line 'a' in the next note is a good
This compares well for White with a Scandi example of this.
navian Defence ( 1 e4 d5 2 exd5 1iixd5 3 ctJc3 3...i. g4
1iia5 4 d4 ctJf6 5 ctJf3), because Black cannot re This is most players' choice. Otherwise:
strain White's centre by the useful . . . c6. In re a) If Black plays 3 . . .dxe4 4 lbxe4 i.f5?!,
turn, White has no pawn on c2. But in the Scan White shows the benefit of his quick develop
dinavian, the c2-pawn can be a disadvantage ment by harassing the bishop with 5 ctJg3 i.g6
Many readers are aware that chess-players

all over the world fell in love with World Cham
pion Mikhail Tal's brilliant and romantic at
tacks, which have influenced all of us since.
What they may not know is that Tal also de
lighted chess fans with his astonishing origi
nality and piquant humour. One of the most
wonderful moves in his career reflected both of
these qualities: 5 gxf3 ! ! ? (D).
6 h4 threatening h5 . Compare this with the

standard line 2 d4 d5 3 lt:Jc3 dxe4 4 lt:Jxe4 j.J5 5
lt:Jg3 .ll g 6. In our current position with 2 lt:Jf3
d5 3 lt:Jc3, the f3-knight is ready to spring to e5 .
White will at the very least win the two bishops
and remain with good development (keep in
mind that winning the bishop-pair often comes
at the cost of development). After 6 h4, play
goes 6 ... h6 (6 ...lt:Jf6 7 h5 .lle4 8 lt:Jxe4 lt:Jxe4 9
d4 e6 10 .ll d3 gives White two bishops and
good development) 7 lt:Je5 �d6 (7 . . . .llh 7??, to To play such an absurd move in a casual
preserve the bishop, loses to 8 �f3 ! lt:Jf6 9 �b3 game is one thing; in an international tourna
with a double attack on f7 and b7) 8 lt:Jxg6 ment, another. But in the refined atmosphere of
�xg6 9 d4 and White will soon play .lld 3 forc the World Championship one doesn't do such
ing the queen to move again. things ! Nevertheless, in Tal-Botvinnik, Mos
b) 3 . . . lt:Jf6?! 4 e5 lt:Jfd7 5 e6 ! fxe6 6 d4 fa cow Wch (3) 1 960, White shocked everyone
vours White, who would like to play .ll d3 with (and, I hope, made them laugh) by recapturing
lt:Jg5 or, if Black plays ...lt:Jf6, then lt:Je5 paraly with the g-pawn. I think that only recently have
ses him. we begun to see a growth in players' receptivity
c) The move . . . c6 doesn't go very well with towards apparently unprincipled moves in the
3 . . . d4 4 lt:Je2; for instance, 4 . . . c5 5 c3 and opening. Tal would have been pleased by this.
Black's centre can't be held after 5 . . . lt:Jc6 6 At any rate, he promptly got an inferior game
cxd4 cxd4 7 �a4 d3 . but recovered and fought his way to a draw. In
4 h3 .llxf3 spite of the condemnation that 5 gxf3 received,
This is the standard choice. Instead, 4 ... .llh5 Tal's mistaken follow-up is easily improved
is very risky in view of 5 exd5 cxd5 6 .llb 5+ upon (by his own suggestions, for starters), and
lt:Jc6 7 g4 .llg 6 8 lt:Je5. This position is sup it's a little disappointing that so few players
posed to be playable for Black, although it is have risked their precious ratings just once to
dangerous for him in view of h4-h5 or simply give the move a try. Kudos to Chris Depasquale,
d4 and .ll f4. Several books touch upon it, per who has two games out of the 28 with 5 gxf3 in
haps not thoroughly enough; if I were White I Megabase 2006.
would look at (and as Black I would worry We now return to 5 �xf3 (D).
about) 8 .. Jk8 9 h4 ! , intending 9 ... d4 (Black is This move, on the other hand, is represented
suffering after 9 .. .f6 10 lt:Jxg6 hxg6 1 1 d4) 10 by about 2 1 00 Megabase games, still somewhat
h5 .llxc2 1 1 �xc2 dxc3 12 .l::th 3 ! (or 12 �b3 e6 less than 2% of all Caro-Kanns.
1 3 dxc3) 12 ... e6 13 lt:Jxc6 bxc6 1 4 .l::tx c3, etc. White has gained the two bishops in return
5 �xf3 for somewhat reduced central control. Black is
CARO-KANN DEFENCE 253
Black intentionally surrenders the centre.

8 lt:Jxe4 lt:Jxe4
8 ...�xd4 9 c3 �d8 1 0 0-0 is risky because
White's pieces come out so quickly, but it's
hardly clear.
9 �xe4 lt:Jd7 10 c3 lt:Jf6 1 1 �e2 .i.d6 (D)
happy to have exchanged his light-squared

bishop, since he will be setting up his pawns on
light squares. Play can develop along several
lines.
5 ...e6
Another set-up begins with 5 ... lt:Jf6 6 d3 e6.
White's bishops don't have any exceptionally
good squares, and while White gets reorga Black sets up the kind of restraint structure
nized Black will get all his pieces out and play that is seen in the Slav, Scandinavian and other
... e5, trading space and active pieces for the two defences: pawns on light squares to comple
bishops. A good piece organization for Black to ment the dark-squared bishop while restraining
achieve that is . . . lt:Jd7, ... g6 and ... .i.g7. White's centre. As in those openings, one idea
6 d4 (D) is to get developed quickly and play either ... c5
White can also play 6 d3 lt:Jd7 7 .i.e2 (7 .i.d2 or ... e5. See the Chapter 3 on structures for
.i.d6 8 d4 a6 9 0-0-0 b5 10 .i.d3 lt:Je7 1 1 h4 �b6 some examples.
is also equal, Planinc-Petrosian, Yugoslavia
USSR (Ohrid) 1972) 7 ... g6 8 0-0 .i.g7 9 �g3 Let's now tum to the main lines.
�b6 10 <.t>hl lt:Je7 with equality, Anand-Karpov, 1 e4 c6 2 d4 d5 (D)
Brussels Ct (3) 1 99 1 .
The Caro-Kann resembles the French De

6 lt:Jf6
••. fence in that Black places a pawn on d5 on the
6 ... dxe4 is also possible. second move and forces White to decide what to
7 .i.d3 dxe4 do with his e4-pawn: advance, exchange, defend
or gambit. Some of the resulting positions are The c3 Systems

quite similar. It has also been said that the Caro
Kann resembles the Slav because l ...c6 takes 4 i.d3 tl:ic6 5 c3
away the 'best' square c6 from the knight, but This more conservative development isn't
keeps an open view for his light-squared bishop. supposed to promise White much, but it has
Not surprisingly, however, the characters of the some sting and the pawn-structure is particu
positions arising from 1 e4 and 1 d4 turn out to larly thought-provoking.
be radically different: there isn't a Slav Advance 5 tl:if6 (D)
.••
Variation, and e4 by White is a rarity in the Slav. Now we'll run into some familiar ideas.
In any event, 2 . . . d5 compels a response. I'll
be looking at 3 exd5 (both the Exchange Varia
tion and Panov Attack) and 3 e5, the Advance
Variation. I think that those variations are the
most useful and consistent in terms of this
book's organization. I won't be dealing with the
3 tl:ic3 main lines, although naturally they're
full of wonderful ideas. I'm also leaving out the
'Fantasy Variation' 3 f3 , although it has its
points of interest. A curious positional compar
ison arises after 3 f3 e6 ! ? 4 tl:ic3 i.b4 (these are
not the only moves, of course), when 5 e5(?) c5
is actually a good version of the French De
fence because White's ' extra' tempo due to
. . .c6-c5 has been used for the awful move f3,
which not only loosens White's kingside but 6 i.f4
takes away f3 for the knight and cuts off White's This is the customary approach, and most
queen from the kingside. likely to achieve something tangible.
However, let's say that White plays 6 tl:if3,
Exchange Variation Black answers with the natural 6 . . . i.g4, and
there follows 7 0-0 e6 8 tt:'ibd2 i.d6 9 .l::t e l 0-0.
3 exd5 cxd5 (D) All very logical. Then White might want to re
spond to the presence of Black's bishop on g4:
10 tl:ifl °V/iic 7 1 1 i.g5 tt:'id7 12 tl:ig3 (D).
An enormously instructive variation, the

study of which will benefit all chess-players.
We'll look at White's slow build-up with c3, What do we have here? A classical Queen's
and then tum to the more aggressive Panov At Gambit Declined Exchange Variation (Carls
tack with c4. bad) with colours reversed ! In this position,
especially since one of the standard QGD plans (see the pawns on d5 and e6). Of course, 7 lt:Je2
. . .lt:Je4 is not available (that' s lt:Je5 in our Caro is also possible.
Kann case), Black might want to play the minor 7 �d7 8 lt:Jd2 e6 9 lt:Jgf3 .lld 6!
••.
ity attack 1 2 ... l:tab8 1 3 h3 .llxf3 14 �xf3. Then This discovery hurt the popularity of the
it' s as though White has played four of Black's set-up with c3, .lld3 and .llf4.
common Queen's Gambit moves ( ... .lle7, the re 10 .llxd6 �xd6 1 1 0-0
capture ... .llxf6 and the repositioning ... .Ile7-d6) After 1 1 �xb7 l:tb8 1 2 �a6 0-0! (and not
in one move ! Of course neither side has played 1 2 ... l:txb2? 1 3 .llb 5), Black is considered to
according to a conventional Caro-Kann plan, have at least enough play for his pawn; e.g., 1 3
but it's interesting to see how the same pawn b 3 ? (to protect b2) 1 3 . . . l:tb6 14 �a4 e5 � 1 5
structure in both an e-pawn opening and a d dxe5 lt:Jxe5 with the e-file and threats against
pawn opening leads to the theme of minority at White's king. Then Black stands considerably
tack vs kingside attack. Naturally we could better.
have obtained an exactly reversed position by 4 ll ....llxf3 12 lt:Jxf3 0-0 13 W.el �b8 14
lL:\f3 lt:Jc6 5 c3 .llg4 6 .lle2 e6 7 0-0 .lld6 8 lt:Je5 b5 15 a3 a5 (D)
lt:Jbd2 lt:Jf6 (or 8 . . .lt:Jge7 ! ?), etc., but those are
not the most pointed moves, especially for
White.
Let's return to the Caro-Kann line after 6
.llf4. We' ll follow the young Kasparov.
Lanka - Kasparov
Leningradjr 1977
6 .llg4 (D)
.••
Black gets out in front of his pawn-chain, an

advantage in both the Caro-Kann and the
Queen's Gambit Exchange Variation.
The minority attack in pure form.

16 l:te3 l:tfc8 17 �dl b4 18 axb4 axb4 19
l:tfel
1 9 f4 bxc3 20 bxc3 gives White some attack
and may well be better.
19 bxc3 20 bxc3 �d8 21 l:th3?!
•••
Passive. The most interesting move is 2 1

lt:Jg4! .
2 1 g6 22 �d2 lt:Jxe5 23 dxe5 lt:Jd7 24 ifh6
•••
lt:Jf8 25 .llfl l:tb3 26 l:tcl �a5 27 �e3 (D)

Kasparov has achieved the isolation of the
c-pawn but his f8-knight is much worse than
the fl -bishop, so he needs to make concrete
7 �b3 progress.
7 f3 .llh5 has the idea of . . . .ll g6 with the ex 27 �a3 28 l:tf3!? l:tc7 29 �f4 �b2
.•.
change of White's good bishop. This is a com Black has had chances for a while to play
mon theme even with a knight on f3; in that .. .d4; e.g., 29 ... d4 ! 30 c4 l:txf3 3 1 gxf3 lL:\d7 32
case, after lt:Je5 and ... .llg 6, Black isn't afraid of l:tbl �c3 33 l:tdl l:tc5.
lt:Jxg6 because knights are strong in these posi 30 h4 h5 31 @h2 d4
tions and he gets a useful h-file. Moreover, the This is still good.
light-squared bishop is still Black's bad bishop 32 .lld3! ? dxc3!
On 32 ... .l::tb xc3, 33 .l::td l stops Black in his This is the starting position of the Panov At
tracks. tack, also known as the 'Panov-Botvinnik At
33 .l::tc2 �a3? tack' because of the former World Champion's
Perhaps missing White's next move. Instead contributions to it.
33 ... .l::tb 4! was very strong and would probably 5 ...e6
produce an eventual win. Now White will gain 5 . . .g6 gives a Griinfeld-like position in which
the advantage. Black usually lets White temporarily win the d
34 .llc4 .l::tb 4 35 .l::tfxc3 �a4 36 �cl �a7 37 pawn and then tries to win it back via ... tt:'ibd7-
f4! ? b6 or . . .a6/... b5. I'll forego that line here.
This i s loosening. 3 7 �d2 ! �a4 3 8 �e2 The main alternative is 5 ... tl:ic6, which intro
would retain White's superior position, although duces a different set of problems and structures
making progress will be difficult. that I shall try to outline in broad-brush fashion.
37 ... �d4! 38 g3 'ilg7 39 'ilh3 tl:id7 40 .llfl The presentation will most likely be inaccurate
.l::txc3 41 .l::txc3 tt:'ib6 42 �e3 �dl 43 �f3 �al from an advanced theoretical point of view but
44 .llg2 tl:id5 should be helpful for the student:
The game is equal. a) 6 .llg5 (D) threatens 7 .llxf6 followed by
45 .l::td3 .l::tb 2 46 f5! ? gxf5 47 �xh5 �cl 48 8 cxd5 .
.l::td l �c2 49 �g5+ 1/z-1/z
Panov Attack
4 c4
This introduces one of the classic variations
of the Semi-Open Games, one that has attracted
many great players through the years. The at
tack on d5 poses some problems for Black, be
cause if he captures on c4, White's bishop gets
out to an active square in one jump, whereas in
many isolated queen's pawn positions like the
one soon to occur, White has to take two moves
(usually .lld3 and then .llxc4) to get there, or he
has to play a less desirable extra 'waiting' move
such as a3, .l::tc l or .llg 5. Compare isolated Black can respond by 6 . . . e6, or by various re
queen's pawn positions in the Queen's Gambit freshing set-ups that begin with 6 ... �a5 and
Declined or Nimzo-Indian, for instance, and 6 ... .lle 6 ! ?, the popularity of the latter illustrat
see the further comments below. ing the flexibility and pragmatism of chess
4 ... tt:'if6 5 tt:'ic3 (D) ideas. Although I won't be looking into those, I
should mention that 6 ... e6 7 lt:Jf3 i.e7 leads to

one of those positions that we were talking
about in which White may not want to lose a B
tempo after 8 i.d3 dxc4 9 i.xc4, and thus con
siders moves such as 8 l:tc 1 and 8 a3. The inde
pendent idea of 8 c5 is also possible. These
positions will repay study, and in fact you might
want to sit down and try to work out the details
without recourse to books or other sources.
b) 6 lL:\f3 (D) is the main move.
his bishops, queenside pressure, and the move

d5.
But 9 ... lt:Jb6 is not fully worked out. It allows
White attacking sequences beginning with 1 0
d5 ! (or 1 0 i.e3 e 6 1 1 0-0-0, which may be best
defended by l l . . .i.e7 1 2 d5 exd5 1 3 i.xb6
�xb6 14 �xb6 axb6 1 5 lt:Jxd5 l:txa2 1 6 �bl
l:ta5 with equality) 1 0. . . lt:Jd4 1 1 i.b5+ lDd7 1 2
�a4 lt:Jxb5 ( 1 2 . . .e5? 1 3 dxe6 lt:Jxe6 14 i.g5 !
with the idea 14 . .. lt:Jxg5 1 5 0-0-0) 1 3 't!kxb5 g6
14 0-0 (D).
Then:
b l ) Sometimes 6 . . . i.e6 is played to threaten
. . . dxc4 and protect the centre at the same time.
Anand-Miles, Wijk aan Zee 1 989 shows one
way to convert the pawn-structure (pretty much
by force) and then build up a superior position:
7 c5 g6 8 i.b5 i.g7 9 lt:Je5 i.d7 10 i.xc6! bxc6
1 1 0-0 0-0 1 2 J::i.e l i.e8 1 3 h3 (or 1 3 �e2)
1 3 . . .@h8 14 i.f4 lt:Jg8 1 5 b4 f6 1 6 lt:Jf3 �d7 1 7
a4. White has a moderate but certain advantage.
It's difficult for Black to find anything positive
to do, and Anand won rather easily.
b2) The most important reply is 6 ... i.g4,
because it carries with it the positional threat of
7 ... i.xf3 and gives White few serious options.
The most common and well-analysed one is 7 White intends l:tel and/or i.g5. This is all
cxd5 lt:Jxd5 8 �b3 i.xf3 9 gxf3 (D). theory, one nice idea being 14 . . . i.g7 15 i.g5
White has implemented the mini-rule that ( 15 J::i.e l 0-0 16 i.g5 is supposed to be some
when one side brings their queen's bishop out what better for White) 1 5 ... h6? 1 6 i.xe7 ! �xe7
early, the opponent should strongly consider 1 7 �b4+ @e8 1 8 J::i.ael + i.e5 1 9 f4 �h4 20
bringing his queen to the queenside, in this case �e4 ! and White is virtually winning with extra
to b3, because the bishop has abandoned de material and terribly strong pawns. Such tactics
fence of that wing. Here White has carried out stem from open lines and rapid development.
that idea at the cost of doubled f-pawns. Now Returning to 9 gxf3, the main continuation is
9 ... lt:Jxd4?? loses to 10 i.b5+, and 9 ... lt:Jxc3 1 0 9 .. . e6, when 10 �xb7 leads to the line 10 ... lt:Jxd4
bxc3 i s considered good for White because of 1 1 i.b5+ lt:Jxb5 1 2 �c6+ @e7 1 3 �xb5 't!kd7
14 lt:Jxd5+ �xd5 1 5 J.g5+ f6 1 6 �xd5 exd5 1 7 Velimirovic - Benko

J.e3 (D). Vrnjacka Banja 1973
6 J.e7 7 cxd5 exd5 (D)

•••
Again, we have bookloads of theory on this

fascinating and educational ending. White's
horrible doubled f-pawns are compensated for I think it's important to understand that this
or outweighed by his open files (providing some recapture, while safe-looking, can give White
bothersome threats against the king), his more some real prospects.
effective bishop, and Black's own two weak 8 J.b5+ lt:Jc6
pawns. Probably the result with perfect play is a 8 ... J.d7 9 J.xd7+ lL:lbxd7 10 0-0 0-0 1 1 �b3
draw, but Black has to play more accurately lt:Jb6 1 2 .l::te l .l::te8 13 J.g5 with a definite advan
than White does, which probably explains a lot tage, Petronijevic-Nikolic, Belgrade 1 997. Here
of players' inclination towards the more com we see one danger in the pawn-structure after 7
mon move 5 . .. e6 (D), to which we return now. cxd5 exd5, which is that the d-pawns are iso
lated. Normally when such a pawn is masked by
another (and thus not on an open file), it poses no
problems. But the fact that 1 1 �3 targeted the
d5-pawn caused Black's knight to go to a miser
able position on b6, which became the reason for
White's advantage.
9 lt:Je5 J.d7 10 0-0 0-0 11 l:lel .l::tc8 12 J.g5
J.e6 13 J.xc6 bxc6 14 lt:Ja4 h6 15 J.xf6 J.xf6
16 lt:Jc5 (D)
6 lt:Jf3
Here we'll look at a few of the thousands of
games that have been played from this position.
Fortunately there's quite a bit of material on iso
lated queen's pawns throughout this book (for a
lengthy introduction to the subject, see Chapter
3), so this lesson will not stand on its own. We'll
see three different 6th moves for Black.
White stands well, although 1 6 .l:k 1 was pos makes it almost certain that a white rook will
sibly more accurate. reach the c-file before Black' s: 13 ... �c8 14 i.f4
�7 1 5 �c6 ! .l::tab8 1 6 .l::tfc l lt:Jb4 1 7 �d7 ! lt:Ja6
6 ... i.e7 7 cxd5 lt:Jxd5 is a main-line IQP po 1 8 .l::tc 3 .llf6 1 9 �f5 ! .l::tfe8 ( 1 9 ....ll x e5 20 i.xe5
sition and is similar to 6 ... i.b4 below. If this .l::tbc8 21 �f6 ! would be a pretty finish) 20 .l:.h3
were a book on theory I'd have to be more spe h6 2 1 .ll x h6! �c8 (2 1 .. ..ll xe5 22 i.g5 !) 22
cific, but this really is one of those instances in lt:Jd7 .l::te 6 23 i.xg7 ! 1 -0 Larsen-Pomar, Span
which it's fair to say that the ideas are much ish Cht (Centelles) 1978.
more important than the details. Therefore I'll 8 lt:Jc6 9 0-0 0-0 10 .l::tel (D)
...
limit myself to one main game, although an

other very attractive contest is embedded in the
note to White's 8th move.
Matveeva - Anand
Frunze 1987
6 i.e7 7 cxd5 lt:Jxd5

•••
This transposes to a Queen's Gambit Ac

cepted (since i.c4 is played next), but just as
often it arises from a Panov move-order. The
themes are like those after 6 ... i.b4 below, and
they can in fact transpose if White plays i.d2-
g5 while Black plays ... i.b4-e7.
8 i.c4
8 i.d3 0-0 9 0-0 (9 h4! ? has also been tried) We have reached a standard position. Al
and we're in another standard IQP position. though the theory of the line was to change
Watch out if you're trying to transpose into a later, the game shows a beautiful defensive ef
formation with ... b6, ....ll b7 and ... lt:Jbd7, which fort and a model for Black:
is standard in isolated queen's pawn positions. 10 a6 11 .llb3 lt:Jxc3 12 bxc3 b5 13 't!kd3
•••
That plan often doesn't fit if you don't have a Black can hold White off after the thematic
knight on f6; for example, 9 . . . b6? ! 10 lt:Jxd5 ! 1 3 d5 : 1 3 . . .lt:Ja5 ! 1 4 dxe6 i.xe6 15 i.xe6 't!kxd l
exd5 ( 1 0... �xd5? 1 1 �c2 ! hits h7 and threat 16 .ll xf7+ .l::txf7 1 7 .l::tx dl .llf6.
ens .lle4) 1 1 lt:Je5 i.a6?! 1 2 i.xa6! lL:lxa6 1 3 13 i.b7 14 .llc2 g6 15 .llh6 .l::te8 (D)
•••
�a4 (D).
16 .l::tadl
Look at those wonderful light-square targets. Here was a chance for 1 6 a4 ! b4 1 7 c4 with a
Just as importantly, the attack on the a6-knight small advantage, according to G.Kuzmin.
16 .l::tc8 17 h4! ? �d5!

•••
A beautiful trap is 1 7 ... .ixh4? 1 8 d5 ! lt:Ja5

1 9 d6 ! .ixf3 20 d7 ! ! .ixd l 2 1 dxe8�+ �xe8
22 �d4 f6 23 .ixd 1 ! , winning.
18 .ib3 �h5! (D)
8 h6! 9 .if4
•.•
After 9 .ixf6 .ixf6 1 0 .ib5 0-0 1 1 0-0,

l l . . .lt:Je7 ! 12 b4 b6 illustrates a way to stop
White's queenside roller: 1 3 �d2 bxc5 1 4 bxc5
.id7 with equality, Timman-Kramnik, Amster
dam 1 996.
19 �e3 lt:Ja5! 20 .ig5 lt:Jxb3 21 axb3 .ixg5 9 lt:Je4 10 .ib5
.•.
22 hxg5 .ixf3! 23 gxf3 .l::ted8 Perhaps 10 .l::tc 1 ! ? .

White's position has been shredded, and it's 10 lt:Jxc3 ll bxc3 .id7 12 0-0 0-0 13 .l::tc l !
.•.
hard to see what actually went wrong. .l::te 8!

24 d5 .l::txd5 25 .l::txd5 exd5 26 @g2! h6! 27 13 ... b6 runs into 14 c4 ! , which was the point
gxh6 @h7 28 �e7?! �f5 29 �e3 �f6! 30 .l::tc l of 13 .l::tc l .
At this point, although 30 ... aS ? kept some 1 4 .l::tel .if6 (D)
advantage and Anand eventually won, the stron
gest way was 30 ... b4 ! 3 1 c4 (3 1 cxb4 d4 32
�d2 .l::tc 3 !) 3 l . ..dxc4 32 l:txc4 .l::txc4 33 bxc4
a5. Then the connected pawns would have been
too strong.
Kasparov - Anand
Amsterdam 1996
6 lt:Jc6
•.•
This gives White the chance to play a scheme

that isn't available (or effective) in most other
positions:
7 .ig5 .ie7 8 c5! ? (D)
There are a couple of ideas behind this move.
One is to launch a queenside attack by b4-b5 15 .l::tbl
and drive Black's pieces back. The other is to Kasparov mentions 15 .id3 ! ? b6 16 cxb6
control e5 by whatever means possible with axb6 17 .ibl with a small edge. The opening is
out White having to attend to his d-pawn after over and White has the initiative.
... dxc4. White's key moves in this process are 15 b6 16 .ia6 .ic8
•••
.ib5, 0-0, .if4 and .l::te l , followed by tt:Je5 itself. 1 6. . . bxcS 17 .ib7 .
This most famous game with 8 c5 illustrates 17 .ib5 .id7 18 .ia6 .ic8 19 .id3!? bxc5
both a good solution for Black and White's at 1 9 ....id7 ! is best, when White still has to
tacking possibilities. demonstrate how to get through.
20 tt:'ie5 .i.d7 .i.g5 would have transposed to a normal IQP

20... tt:'ixe5 ! ? 2 1 dxe5 .i.g5 22 .i.xg5 'ti'xg5 23 position.
.i.b5 ! �d8 24 .i.c6. 8 tt:'ic6 9 a3 .i.e7
.•.
2l �b7 9 ... .i.a5 I O .i.d3 ! ? tt:'ixc3 1 1 bxc3 tt:'ixd4 12

Now things go downhill for Black. White's tt:'ixd4 'ti'xd4 13 .i.b5+ .i.d7 1 4 0-0 is an unclear
opening strategy has been a major success. gambit.
21 .i.xe5 22 dxe5 �b8
••. 10 .i.d3 .i.f6?!
Or 22 ... .i.c8 23 'ti'g4 !. Kasparov doesn't let This just doesn't seem to work out. I O.. tt:'if6 .
up in what follows. is a normal isolated queen's pawn position,

23 �xb8 'ti'xb8 24 'ti'g4 �f8 25 �e3 'ti'd8 26 when White has a lead in development and could
h4! 'ti'a5 27 l:tg3 �e7 28 'ti'xg7 �d8 29 'ti'xf7 gambit a pawn but probably just plays 1 1 .i.e3.
'ti'xc3 30 .i.b5 'ti'a5 31 l:tg7 tt:'ie7 32 .i.xd7 1 1 0-0 tt:'ixc3
�xd7 33 'ti'f6 d4 34 .i.xh6 c4 35 .i.g5 'ti'c5 36 Not 1 l .. .tt:'ixd4 ?? 1 2 tt:'ixd4 .i.xd4 1 3 •a4+;
�xe7+ 1-0 nor 1 l . . ..i.xd4? 12 tt:'ixd5 exd5 1 3 .i.b5.
The finish would be 36 ... l:txe7 37 'ti'xe7+ 12 bxc3 (D)
'ti'xe7 38 .i.xe7 �xe7 39 �fl .
Fedorowicz Enkhbat-
USA Ch (Seattle) 2003 B
6....i.b4 (D)
12... h6
In order to get castled.
13 'ti'e2! 'ti'd5
He still can't castle due to 1 3 . . . 0-0 14 •e4.
Something has already gone wrong.
14 l:tbl a6 15 c4 'ti'h5 16 'ti'e4 �f8
7 cxd5 tt:'ixd5 A terrible concession.
For those of you wondering, 7 ... exd5 is a re 17 l:tel tt:'ie7 18 .i.d2 'ti'f5 19 'ti'e3 'ti'h5 20
spectable alternative, although seldom played. 'ti'f4! tt:'ig6 21 'ti'c7 @gs 22 .i.e4 �h7 23 'ti'xf7
Black may have the most problems with 8 .i.b5+ Now it's really over. Black only lasted a few
again; compare 6 . . . .i.e7 7 cxd5 exd5 above. more moves.
8 'ti'c2 23...l:tdS 24 .i.e3 l:td7 25 'ti'e8 l:te7 26 'ii'a4
More often 8 .i.d2 is played here. A game in .i.d7 27 'ti'dl ! .i.e8?
which White wasn't ambitious enough went But 27 . . . @g8 28 l:txb7 is awful.
8 . . . tt:'ic6 9 .i.d3 tt:'if6 I O 0-0 0-0 1 1 .i.g5 h6 !? 1 2 28 tt:'ig5+ 1-0
.i.e3 ( 1 2 .i.h4 ! ?) 1 2 . . ..i.d6 1 3 l:tc 1 ! ? ( 1 3 l:t e 1 )
1 3 ... e5 ! 1 4 h 3 ( 1 4 tt:'ixe5 tt:'ixe5 15 dxe5 .i.xe5 is Advance Variation
equal) 14. . . .i.e6 15 'ti'd2 'ti'a5 16 .i.xh6 exd4 !
17 tt:'ib5 'ti'xd2 1 8 .i.xd2 .i.b8 with equality, 1 e4 c6 2 d4 d5 3 e5 (D)
J.Polgar-Karpov, Dos Hermanas 1 999. Instead, This extremely popular move has led to re
8 . . . 0-0 9 .i.d3 .i.e7 I O 0-0 tt:'ic6 1 1 a3 tt:'if6 1 2 markably exciting, creative play. There have
for Black in the French. The flip side of the

question is whether this basic white pawn
B structure, advocated by Nimzowitsch, really
can be played for an advantage. According to
conventional theory, White is supposed to over
protect the e5-pawn against assault by .. .f6,
such that any capture with .. .fxe5 will leave him
with a permanent outpost on d5 from which his
pieces can't be driven away. That is not so eas
ily done. Alternatively, White can use his extra
tempo simply to hold on to the c5-pawn and re
main a pawn up ! It' s worth examining these
two strategies, which can be introduced in a
few ways:
been more discoveries here than in any other a) 5 .ltd3 tt:'ic6 (5 ....ltxc5? 6 'ti'g4 forces
variation of the Caro-Kann, and indeed more Black into an awkward defence of his g-pawn;
than most openings. he either has to move his king or make the very
3 .ltf5
... weakening move ... g6) 6 tt:'if3 (this is the same
Strong players over the years have tried to at as 5 tt:'if3 .ltxc5 6 .ltd3) 6 . . ..ltxc5 7 0-0 tt:'ige7 !
tack White's pawn-chain at the base by 3 ... c5, (7 .. .f6 8 'ti'e2 ! - strongpoint - 8 ...fxe5 9 tt:'ixe5
the idea being that 4 c3 might allow Black's c8- tt:'if6 10 .ltf4 0-0 1 1 tt:'id2 tt:'ixe5 1 2 .ltxe5 'ti'b6
bishop to develop outside Black's own pawns; 1 3 tt:'ib3 .ltd6 14 .ltxd6 'ti'xd6 15 l:tae l ) 8 'ti'e2
e.g., 4 . . .tt:'ic6 5 tt:'if3 cxd4 6 cxd4 .ltg4 or some tbg6 9 c3 0-0. Black appears to be doing fine in
such. But this takes Black two moves with his these positions.
c-pawn, and 4 dxc5 changes the structure dra b) Also by analogy with the French De
matically. After 4 ... e6 (D), you may recognize fence, the apparently untried 5- 'ti'g4! would be
the similarity to the French Defence Advance very interesting, tying Black's bishop to f8 and
Variation, i.e. 1 e4 e6 2 d4 d5 3 e5 c5 4 dxc5, but preparing tt:'if3, .ltd3, etc. The queen is ready to
then it's Black's move! overprotect e5 from g3, as shown by 5 . . . tt:'ic6 6
tt:'if3 f5 (6 ... 'ti'c7 7 .ltb5) 7 'ti'g3 'ti'c7 8 tt:'ic3 (or
8 .lte3) 8 ... .ltxc5 9 tt:'ib5 'ti'd7 1 0 .lte3 !? .ltxe3
1 1 tt:'id6+ @f8 1 2 fxe3 (D).
That's because of the tempo loss ... c6-c5. In

the French Defence move-order. 4 dxc5 is harm
less or worse because it weakens e5 and brings
Black's pieces out quickly. The first question, The strong point lives, and before d6 can be
then, is whether having lost a whole tempo, this challenged White will have played 0-0-0 and c4.
position can still be played for Black. That is at c) Black has yet another difficulty if White
least possible, since it is generally favourable uses his extra tempo to hang on to the c5-pawn:
5 .i.e3 (which is the preference in practice)

5 ... tt:'ih6 6 c3 tt:'if5 7 .i.d4 tt:'ic6 8 tt:'if3 "iic7 9
.i.b5 with a solid advantage. Probably Black
can play better but he looks short of equality in
any event.
Although playable, it appears that after 3 . . . c5
4 dxc5, White's extra tempo somewhat out
weighs Black's pleasant pawn-structure. The
move 5 °iig4 ! is particularly worrisome.
Thus the normal move 3 . . ..i.f5 (D), to which
we now return, is critical:
16 . . . tt:'ixe5 ! 17 °iie2 tt:'ixc4 1 8 .i.xd5 ! °iixb5 ! ?

1 9 .i.xc4 °iib6 2 0 fxe6? ! 0-0-0 ! 21 exf7 tt:'ie7 22
"iie6+ @b8 23 .i.f4+! ? @a8 24 °iixb6 axb6 25
.i.e5 l:!.hf8! and so forth, Shirov-Nisipeanu, Las
Vegas FIDE KO 1 999, a game in which half of
the moves are worth a page of analysis each.
Or, more recently, 6 ...f6 7 h4 ! fxe5 8 h5 .i.f7
9 dxe5 tt:'id7 I O f4 °iib6 1 1 tt:'id4 ! 0-0-0 1 2 .i.h3 !
tt:'ie7 1 3 a4 ! c5? 14 a5 °iia6 1 5 .i.fl c4 1 6 b4 ! b6
17 .i.e3 bxa5 18 J:!.xa5 °iib6 19 tt:'if5 tt:'ixf5 20
.i.xb6 lllxb6 21 gxf5 .i.xb4 22 °iid4 1 -0 Nai
ditsch-Dautov, French Cht 2005 .
White has an amazing number of valid op The 4 lllc 3 and 5 g4 variation is rich in ideas
tions in the position after 3 ... .ltf5, expressing and recommended to the tactically inclined, but
diverse and creative approaches. I've chosen to not explicable in organized fashion. We have
look at two modern variations (4 tt:'if3 and in more useful ground to cover in the sense of un
particular 4 .i.e3) in most detail because they derstanding chess in general.
represent a mix of approaches, from the purely b) 4 h4 is the sort of exotic move provoked
positional to tactical. by the bishop's placement on f5, i .e. g4 would
However, some of the alternatives are them now drive it off the h7-bl diagonal. Again, the
selves main lines and hence deserve close at play will be based upon specifics, but there's a
tention: wonderful line from older days that goes 4 . . . h6
a) 4 tt:'ic3 e6 5 g4 .i.g6 6 tt:'ige2 has led to (4 . . . e6?? loses a piece to 5 g4, but, among oth
great attacking chess and been a favourite for ers, 4 . . . h5 and 4 . . .°iib6 5 g4 .i.d7 are played) 5
over a decade now. Unfortunately, there isn't g4 .ltd7 (5 . .. .lth7 6 e6! fxe6 7 .i.d3 has ideas
much to say about the line in a short space ex like .i.xh7, °iid 3 and tllf3-e5 in mind; this is a
cept that it generally leads to random-looking standard idea in several openings) 6 h5 e6 7 f4
chaos! The resulting melees are completely de c5 8 c3 lllc 6 (D).
pendent upon the precise tactics of individual This is a classic picture of space versus the
positions (and the preparation put into them). kind of rapid development that goes with an in
While there are naturally consistent themes and cipient central attack. White hasn't moved a
even buried positional indicators, I can't begin piece yet but he threatens to squeeze Black to
to clarify what goes on. Consider, for example, death. The latter must develop as fast as possi
making sense of this: 6 ... c5 7 h4 h6 8 f4 ! ? .i.e7 ble and open lines to counteract that. 9 tt:'if3
9 .i.g2 ! .i.xh4+ I O @fl .i.e7 ! 1 1 f5 ! .i.h7 1 2 °iib6 and now:
tt:'if4 °iid7 1 3 tt:'ih5 ! ? .i.f8 1 4 dxc5 lllc6 1 5 lllb5 ! b l ) Tal-Pachman, Bled 1961 continued I O
.i.xc5 16 c4 ! ! (D). tt:'ia3 cxd4 1 1 cxd4 0-0-0 1 2 tt:'ic2 @b8 1 3 .i.d3
Ci'Jce7 and who knows what's happening! But

this example only emphasizes both the poten
tial of cramping pawns to shut down counter
play completely and the consequent necessity
of immediate action by the other side.
c) 4 Ci'Je2 is another variation that can eas
ily become tactical, sometimes right off the
blocks. One line that resembles 4 Ci'Jc3 e6 5 g4
is 4 . .. e6 5 Ci'Jf4 c5 6 g4 ! ? Si..e4 7 f3 Wih4+ 8 @e2
and so forth you can imagine how important
-
home analysis is in such a line!

d) 4 Si..d3 was eliminated from general use
by the manoeuvre ... Si..x d3 and ... Wia5+ and
. . . Wia6; e.g., 4 . . . Si..x d3 5 Wixd3 e6 (or 5 . . .Wia5+ 6
Ci'Jge7 14 l:tb 1 and White was ready to play b4- Si..d2 Wia6) 6 f4 (6 Ci'Jc3 �6 7 Ci'Jge2 Wia6 8 Wih3
b5. In keeping with his open-lines approach, Ci'Jd7, Sax-Arlandi, Baden 1 999) 6 ...Wia5+ ! 7 c3
Black should play 14 ... f5 ! , when the positional Wia6 ! 8 Widl ? (this costs a tempo and gives
threat of . . .fxg4 encourages 15 g5 g6 ! and White Black the light squares and better piece place
has failed to close the kingside so Black will ment; although 8 Wixa6 Ci'Jxa6 leaves a good
have an attack there. bishop versus a bad one, White has space as a
b2) 1 0 @f2 ! ? (White continues with the compensating factor) 8... c5 9 Ci'Je2 Ci'Jc6 1 0 Si..e 3
anti-development theme) 1 0 . . .f6 1 1 @g3 0-0-0 cxd4 1 1 cxd4 Ci'Jge7 12 0-0 Ci'Jf5 13 Si..f2 h5 14
12 a3 (now b4 is the idea) 12 . . . c4 1 3 Ci'Jbd2 Ci'Jbc3 Si..e7 15 a3 l:tc8 16 @h l Ci'Ja5 17 Wia4+
(D ). l:tc6 1 8 b4 Ci'Jc4 (D) .
After 1 3 . . . @b8 14 b3 ! White launched a suc I n the introductory chapters we talked about
cessful assault on the queenside in Malaniuk colour complexes. Here everything goes to the
Psakhis, USSR 1 979: 14 . . . cxb3 15 Ci'Jxb3 Wic7 light squares. 19 b5 Wixa4 20 Ci'Jxa4 l:tc8 2 1 l:tfc l
16 Si..d3 Si.. c 8 1 7 a4 Ci'Ja5 1 8 Ci'Jxa5 Wixa5 19 0-0 22 �c3 Si..d 8 23 h3? h4 24 g4 hxg3 25 Ci'Jxg3
Wib3 Ci'Je7 20 Si..d2 Wic7 21 l:thc l f5 22 c4!, Ci'Jxg3+ 26 Si..x g3 Si..a5 27 l:td3 Ci'Jd2 28 Ci'Jc5 b6
breaking through. Instead, Black could have (or 28... Ci'Je4! 29 Ci'Jxe4 dxe4 30 l:te3 l:tc3) 29
opened lines by 1 3 ... f5 ! 14 gxf5 ( 14 g5 g6! 1 5 Ci'Jb3? (29 Ci'Jb7) 29 ... Ci'Jxb3 30 l:txb3 l:tc2 3 1 l:td3
hxg6 Ci'Jge7 1 6 Si..x c4 Ci'Jxg6 1 7 Si..d 3 Si.. g 7) l:tfc8 32 Si..e l Si..x el 33 l:txel l:tf2 34 a4 l:tcc2 35
14 . . . Ci'Jge7 ! , when the fight switches to control l:tal g6 36 a5 l:tb2 37 axb6 axb6 38 @gl l:txf4 39
of the kingside light squares, as illustrated by l:ta8+ @g7 0- 1 Wachweger-Schmitzer, Bergen
the sequence 1 5 fxe6 Si..xe6 ( 15 ... Ci'Jf5 + ! ?) 1 6 Enkheim seniors 1 997. White had an off-day but
Si..h 3 Si..x h3 1 7 @xh3 Ci'Jf5 1 8 b 3 g 5 19 hxg6 the point should be clear enough.
The Short Variation White has queenside pressure and the f5-
bishop is cut out of the action.
4 tt:'if3 (D) 18 l:ttbS 19 tt:'id2 @f8 (D)
•••
This move revolutionized the Advance Vari 20 h4! @es 21 g4 .i.h7 22 h5

ation by showing that White could opt for slow Short wins space on a second front that is to
and simple development with protection of the be opened later - a classic chess technique. Of
pawn-chain, usually by c3. This is in spite of ten you simply have to have more than one area
the fact that Black has a 'good' French Defence of attack to break down a well-fortified posi
due to the development of his bishop outside tion.
the pawn-chain. As it turns out, such an ab 22 tt:'idS 23 l:tbc3 tt:'ib6 24 tt:'ib3! ltJa4 25
.•.
stract theoretical view doesn' t mean much in l:k7 tt:'ixb2 26 tt:'ic5

practice and there are cases when the bishop He concludes by switching to that second
would be better-placed on d7. Short and others front and conducting a direct attack on the king.
won various nice games until theory caught up 26 b5 27 g5! (D)
•..
and roughly evened things up. Today the same

structure is widely seen, and the Short Variation
itself has evolved, often involving .i.e3 instead
of c3. Here's one of Short's original wins. It B
illustrates some of the underlying ideas and
some that have more to do with pawn-chains as
a whole.
Short - Seirawan
Manila IZ 1 990
4 e6 5 c3 c5 6 .i.e2 tt:'ic6 7 0-0 h6 8 .i.e3!

.•.
With the idea dxc5.

8 cxd4 9 cxd4 tt:'ige7 10 ttJc3 tbcs 11 l:tcl
••.
White has active development and the c-

file. 27 tt:'ic4
..•
11 a6 12 tt:'ia4 tt:'ib6 13 tt:'ic5 .i.xc5 14 l:txc5

•.. Or 27 ... hxg5 28 .i.xg5 .i.f5 29 l:te7+ @f8 30
Now Short has gained the two bishops. His tt:'id7+ @g8 3 1 l:te8+.
opening has been a success. Let's see how it 28 gxh6 gxh6 29 tt:'id7 tt:'ixe3 30 fxe3 .i.f5 31
plays out: @r2 z:tb1 32 tt:Jr6+ @rs 33 z:tgl! 1-0
14 0-0 15 'ti'b3! tt:'id7 16 l:tc3 'ti'b6 17 l:tfcl
..• There would follow 33 ...l:txc7 34 l:tg8+ @e7
'ti'xb3 18 l:txb3 35 l:te8#.
The Zviagintsev Variation . . .�6 has to be answered should Black choose

to play it within the next few moves. Let's re
4 Si.e3 (D) call again our idea about the early development
of White' s dark-squared bishop: whenever that
happens, Black should always consider target
ing the queenside dark squares. That normally
applies to Si.f4 or Si.g5, but there' s no particular
reason to reject the same thought after Si.e3 (al
though at least White needn't worry as much
about d4). The queen sortie to b6 also assists
with ... c5 . So it is probably best played right
away or early on, because given a little time
White can play tt:'ib3 or c3 and b4.
Black, having been informed that a bishop is
on e3, can also aim to put a knight on f5 or g4
and look for a favourable way to obtain the two
bishops. One might want to compare all this to
the Kupreichik line in the Advance French: 1 e4
This modern move (a typical case of bishop e6 2 d4 d5 3 e5 c5 4 c3 tt:'ic6 5 Si.e3. In that case,
before-knight development) has several points. too, Black will often play . . . tt:'ige7, aiming for
First, it helps to stop ... c5 , which after all is . .. tt:'if5 , or even more often ... tt:'ih6 with the dual
Black's goal once he has played ... Si.f5, and all ideas . . .tt:'if5 and . . .tt:'ih6. Of course there's much
the more so after ... e6. It also directly protects more in terms of strategy in this wonderfully
d4, the main target of Black's attack. Moreover, complex line. The players' positional under
a piece gets out that normally has difficulty standing will usually be put to the test, and
doing so in the Advance Variation. Now the therein lies the chance for a challenging and
queen's knight can follow by moving to d2 and competitive game.
not interfere with that bishop, leaving the pawn We look at two games after 4 Si.e3 :
moves c3 and c4 free to be played.
Here you might compare 4 tt:'ic3 above; one Shirov - Dreev
of the reasons that White must strike out with Poikovsky 2006
the early tactical move g4 in that line is that he
is no longer able to play c3 and protect d4, so 4 e6 (D)
...
his centre is faced with demolition by ... c5 and

... cxd4. After 4 Si.e3, however, the option of c3
exists, or White can defend with pieces follow
ing tt:'id2-b3 or he can counterattack by c4. On
the kingside we have a somewhat similar situa
tion, in that delaying tt:'if3 retains the option of
f4, while the g4/h4 ideas that we see after 4
tt:'ic3 are not ruled out. Eventually White will
probably play Si.e2 and 0-0, but he doesn't want
to waste a precious tempo on those moves until
it is necessary. As in so many openings today,
White's underlying philosophy is one of flexi
bility.
There are of course drawbacks to all this,
first and foremost that White is not granted two
moves for every one of Black's, and can only 5 tt:'id2
implement these strategies one at a time ! Fur 5 c3 can be slow and is perhaps out of touch
thermore, there is the concrete problem that with White's philosophy of flexibility: maybe
the pawn wants to go to c4, so don't decide yet. Si.g4 with an obvious advantage, Shirov-Eren
In Haba-Dautov, Bundesliga 2002/3 Black burg, Caleta 2005.
played 5 . . . CDd7 6 ctJd2 f6. Now 7 f4? ! looked a 8 h4
little loose after 7 . . .'ti'b6 ! - compare the Ad An innovation. 8 f4 had been played previ
vance French. The b2-pawn is attacked, but if ously.
White doesn't protect e5 again he can run into 8...f6 9 h5 Si.f7 10 ctJf3 ctJh6!?
. . . fxe5 or even the risky . . . g5 !?. In fact, the 1 0. . . lLie7 1 1 g4! covers f5 and is typical of
game continued 8 'ti'b3 g5 9 exf6 g4 ! 10 f7+ the unrestrained expansion in this variation.
@xf7 and Black already had a pleasant advan 1 1 Si.xh6!? gxh6 12 exf6 a5 13 a4 .i.b4+ 14
tage. @fl ?
5 ctJd7 6 Si.e2! ? (D)
••• 14 c3 Si.d6 destabilizes the b3-knight, tying
Utterly noncommittal ! There have been all down White's queen.
kinds of moves played here, especially 6 f4, 14... 'ti'd8 (D)
strengthening the centre and meeting 6 . . . c5 with
7 lLigf3. Again this seems loose after 7 . . . 'ti'b6,
when White went 8 Si.e2 ! ? in Morozevich-Bar
eev, Russian Cht (Sochi) 2004, sacrificing the
b-pawn based upon development and open lines.
The game went 8 . . .ctJh6 9 h3 'ti'xb2 1 0 c4 Si.c2 !
(to exchange queens and clear f5 for a knight)
1 1 'ti'c l 'ti'c3 ! ? ( 1 1 . . .'ti'xc l + ! 1 2 J:!.xc l Si.e4
should be fine) 12 @f2 ctJf5 13 ctJfl and Black
had to deal with White's centralized pieces and
some tactics based upon the c2-bishop. Never
theless, he stood well in the opening. Maybe f4
isn't such a great idea. The simplest option is 6
ltJgf3, as in the Short Variation.
15 'ti'cl 'ti'xf6 16 ctJh2!

Aiming at the h6-pawn. Shirov has also
played this move in the French Defence.
16...J:!.g8 17 lLig4 'ti'g5 18 g3!? Si.rs 19 ¢'g2
'ti'xcl 20 J:!.axcl
This position illustrates the two-sided nature
of 4 Si.e3: it can lead to tactics or positional
play. Here White controls e5 and can add f4 and
c4 into the mix.
20...J:!.g5?!
This sacrifices the exchange for insufficient
compensation (perhaps Black thought that he
could retain the two bishops), but 20. . . Si.g7 21
f4 ! was pretty bad.
6 ... 'ti'b6 7 lLib3 Si.g6 21 f4 l:!.xh5 22 J:!.hel! Si.e7
A natural alternative is 7 . . . f6. White's centre 22 ... l:!.f5 23 Si.d3 Si.g6 24 J:!.xe6+ wins for
is hard to assail, as shown by 8 ctJf3 lLie7 9 0-0 White.
Si.g6 1 0 c4 ! ? (pretty good, but the simple 1 0 23 lLie3 ctJf6 24 Si.xh5 Si.xh5 25 f5!
Si.f4 ! keeps a nice advantage) 1 0. . . a5 1 1 lLic5 ! ? Now it's a matter of technique, although
lLixc5 1 2 dxc5 'ti'xb2 13 ctJd4 ! ? ( 1 3 Si.d4! 'ti'c2 when you hear that phrase, remember that some
14 'ti'xc2 Si.xc2 1 5 exf6 is very strong, but players' technique is better than others.
Shirov is known for taking a few chances for 25 ... Si.d6 26 c4 Si.b4 27 J:!.fl Si.e2 28 �f2
fun) 1 3 . . . fxe5? ( 1 3 . . . @f7 ! ) 1 4 lLixe6 l:!.c8 1 5 Si.xc4 29 lLixc4 dxc4 30 l:!.xc4 lLie4 31 J:!.e2 exf5
32 tt:'ic5 0-0-0 33 tt:'ixe4 fxe4 34 J:!.xe4 @d7 35 while also targeting Black' s weakened king
l:!.h4 .i.f8 36 l:!.h5 l:!.a8 37 d5! .i.g7 38 b3 @d6 side.
39 l:!.g4 .i.f8 40 dxc6 bxc6 41 l:!.g8 @e6 42 6 c4!? (D)
J:!.xh6+ @f7 43 l:!.h8 J:!.a6 44 l:!.6xh7+ .i.g7 45 The strategy of opening queenside lines is
J:!.c8 @g6 46 l:!.h4 J:!.b6 47 l:!.g4+ @h6 48 J:!.c4 common with 4 .i.e3, and all the more so with a
1-0 queen on c 1 . There are many options in these
positions; for instance, 6 tt:'if3 c5 7 .i.d3 ! ? might
Grishchuk Anand - be a refreshing idea.
Mainz (rapid) (8) 2005
This closely-fought encounter is full of posi

tional niceties in the opening and early middle
game.
4 ... 'ti'b6 5 'ti'cl (D)
Avoiding dark-square weaknesses, and as
sisting White's forthcoming plan in this game.
6 dxc4!
.•.
Black has had some difficulty with the loss

of territory after these alternatives:
a) 6 ... tt:'ie7 ?! 7 c5 ! 'ti'a5+ 8 .i.d2 'ti'c7 9 tt:'ic3
tt:'id7 1 0 .i.e2 h6?! (after 1 0 .. .f6, White's flexi
ble strategy pays off with 1 1 f4 ! ; Kasparov sug
gested 1 0. .. .i.g6 ! ? 1 1 tt:'if3 .i.h5) 1 1 b4 g5 ? 1 2
g4 ! .i.g6 1 3 h4 and White will win at least a
5...e6 pawn, Kasparov-Shirov, Moscow (Russia-Row
5 . . .tt:'ih6 ! ? is frequently played. It has the rapid) 2002.
idea we mentioned above, aiming at g4. In one b) 6 ... .i.xbl ?! 7 J:!.xbl .i.b4+ 8 .i.d2 .i.xd2+
game White played simply 6 h3 and 7 tt:'if3, ask 9 'ti'xd2 and White had space and smooth de
ing where the knight is going, but that is rather velopment in Gelfand-Dreev, Moscow 2002.
slow. A more provocative course is 6 tt:'if3 e6 7 tt:'id2!?
(6 . . .tt:'ig4 7 .i.f4 e6 8 h3) 7 c4 ! (7 tt:'ibd2 c5 8 Simply 7 .i.xc4 tt:'ie7 8 tt:'ie2 'ti'd8 9 0-0 led
tt:'ib3 ! ?) 7 ... dxc4?! 8 tt:'ibd2 ! ? (or 8 .i.xc4 with a to some advantage in Shirov-Anand, Monaco
small but definite advantage) 8 ... .i.d3 and now: (blindfold) 2005. In that manner White keeps
a) 9 .i.xd3 ! ? (a little too fancy) 9 . . . cxd3 1 0 his space advantage, which is the key to the
.i.xh6 gxh6 1 1 0-0 tt:'id7 ? ! ( l l .. .tt:'ia6 !) 1 2 J:!.dl Short Variation and to 4 .i.e3 (its derivative).
'ti'a6 13 tt:'ie4 J:!.g8 14 tt:'ie 1 0-0-0 1 5 tt:'ixd3 with a 7 'ti'a5!
•.•
pleasant position, Anand-Khenkin, Bundesliga To stop tt:'ixc4.

2002/3. 8 .i.xc4 tt:'ie7 9 tt:'ie2 tt:'id7 10 0-0 tt:'id5 1 1
b) White should simply win the pawn back tt:'ig3 .i.g6 1 2 h4! h 6 1 3 h5
with tempo by 9 .i.xh6 ! gxh6 1 0 tt:'ixc4! .i.b4+ White's whole idea here, rightly or wrongly,
( 1 0... 'ti'b4+?! 1 1 tt:'ifd2 .i.xfl 1 2 J:!.xfl with a3 is to acquire space. Black has found a perfect
and tt:'ie4 to come) 1 1 @d l .i.xc4 1 2 .i.xc4. reorganization, however, and has equal play.
Then White is in the pleasant position of being 13 .i.h7 14 a3 'ti'd8 15 tt:'if3 .i.e7 16 J:!.dl
•••
able to claim the advantage after @e2 and J:!.d 1 , J:!.c8

16 ... tt:'i7b6 1 7 .i.e2 'ti'd7 would emphasize 22...tt:'ixf3+ 23 �n tt:'ih2+! 24 �el

the light squares, although I see no special plan Or 24 �gl tt:'ig4 25 .i.f5 ! 'ti'h4 ! .
for Black. 24...'ti'e7 25 'ti'e4 tt:'if6?
17 .i.d3!? (D) Anand falters. 25 . . . J:!.xh7 ! 26 'ti'xh7 tZ:'ixf4 27
'ti'xg7? 'ti'c5 ! is winning due to the idea of
... 'ti'b5.
26 'ti'a4+ �f8 27 J:!.acl (D)
What a decision ! Once again White opts to

exchange his good bishop and retain his bad
one. However, there is a compensating factor
here. When one has space, a bad bishop can be 27...l:!.a8
used to protect your weakest central point (d4) Avoiding Grishchuk's clever idea 27 .:.Xc l ..
while you advance on the wings. 28 J:!.xc l J:!.xh7 29 l:!.c8+ tt:'ie8 30 tt:'ig6+! fxg6 3 1
At any rate, Anand chooses to tum the game hxg6 l:!.h8 3 2 'ti'f4+ 'ti'f6 (32. . .�g8 3 3 9b8 ) 33
down tactical channels, spoiling White's fun: 'ti'd6+ 'ti'e7 34 'ti'f4+ with a draw. What follows
17 c5 18 dxc5 .i.xc5
•.• is unclear:
1 8 ...tt:'ixc5? 19 .i.xh7 tt:'ib3 20 'ti'bl tt:'ixal 2 1 28 .i.bl ! tt:'ihg4 29 'ti'b4 ! ? 'ii'xb4+ 30 ub4
.i.e4 tt:'ib3 2 2 .i.xd5 exd5 23 tt:'if5 ! �f8 24 'ti'a2 ! By this means White wins the seventh rank.
followed by J:!.xd5 is devastating. 30...e5 31 tt:'id5 tt:'ixh5 32 l:tc7 g6 33 0.el!?
19 .i.xh7! ? .i.xe3 20 'ti'bl .i.f4! 21 tt:'ie2 tt:'ixe3 34 txe3 �g7 35 l:tdd7 l:thf8 36 .A.al
tt:'ixe5 J:!.ac8?
Black is obviously better now. Much better and unclear is 36 .�6 ! 37 ..
22 tt:'ixf4? J:!.xf7+ J:!.xf7 38 J:!.xf7+ �g5.

Slightly crazy: White wants to confuse mat 37 J:!.xb7 J:!.cl + 38 �d2 J:!.fi 39 .:U7 tt:)g3
ters. Black would have an easy advantage fol 40 .i.d5! l:!.f2+ 41 �c3 tt:Jn 42 �d3 .:c12+ 43
lowing 22 .i.e4 tt:'ixf3+ 23 .i.xf3 .i.e5, although �e4 l:!.f2 44 b5 h5 45 J:!.xf7+ 1-0
in theory this is the better of evils. The b-pawn queens.
1 3 French Defence
1 e4 e6 the Queen's Gambit Declined ( 1 d4 d5 2 c4 e6),

Strictly speaking, this move defines the where in most of the traditional lines the move
French Defence. However, I shall pass over . . . e5 is needed to bring the c8-bishop into play.
White's second-move options, and get straight A significant exception in the Queen's Gambit
to the position that most players think of as the occurs when the bishop is freed by White' s vol
starting-off point. untary exchange on d5. The situation with the
2 d4 d5 (D) Semi-Slav ( 1 d4 d5 2 c4 c6 3 tt:'if3 tt:'if6 4 tt:'ic3
e6) is obviously worse still. In any event, re
turning to the matter of the French Defence, we
find that if White advances his pawn to e5, the
freeing move . . . e5 becomes unlikely in the short
term, so Black may try to develop his light
squared bishop via . . .b6 and . . . .i.a6. More often
it stays on c8 or d7 for a while, perhaps await
ing the move . . .f6, after which it plays a useful
defensive role guarding e6. The bishop may
later transfer to the kingside (g6 or h5) via e8.
It' s interesting that the Sicilian Defence varia
tions which include the moves ... d6 and . . . e5 are
a mirror image of certain Tarrasch French main
lines, right down to the role of the bad bishop as
protector of a backward d-pawn (in the Sicil
The French Defence ranks behind only the ian) or a backward e-pawn (in the French); see
Sicilian Defence and l .. . e5 as a reply to 1 e4. It the section on 3 tt:'id2 tt:'if6 for more about that
is hard to characterize in general terms since it remarkable comparison. Finally, Black' s light
combines highly tactical and ultra-positional squared bishop may go in the other direction to
types of play. But the French has one quality c6, b5 or a4. Where it ends up will reflect the
that few other openings have, and perhaps none pawn-structure and thus indicate the nature of
to this extent: a persistence of central struc the play.
ture. In the main lines (mostly characterized by What else is going on in the initial position?
White's move e5), the fundamental formation On the most basic level, Black's second move
of . . . e6 and ... d5 has a tendency to last for many of the French Defence attacks the e-pawn ! Ac
moves into the middlegame and fairly often cording to the Hypermodern theorists, White's
into an endgame. The exceptions consist of e4-pawn is too much of a target for 1 e4 to be a
variations with . . .dxe4, which are proportion good move, and in fact Black puts the question
ately infrequent, and lines in which Black to White, who has to choose between exchang
achieves the freeing move ... e5, something that ing the pawn, advancing it, protecting it, and
White usually denies his opponent until the gambiting it. We discuss this in the Introduc
middlegame. tion to the Semi-Open Games (Chapter 1 0).
This brings us right away to the main disad Looking over White's options against the
vantage of the French Defence, Black's light French, we find:
squared bishop. Whether that piece assumes a a) There is no realistic method of gambiting
useful role can determine the success of the White's e4-pawn that doesn't leave him strug
opening. We run into a similar phenomenon in gling for equality.
FRENCH DEFENCE 2 71
b) Exchanging the d-pawns by 3 exds exds

(D) immediately frees Black's queen's bishop Tarrasch Variation
and dissipates White's advantage. This is called
the Exchange Variation. 1 e4 e6 2 d4 dS 3 lbd2 (D)
In spite of the symmetrical result of this ex This move defines the Tarrasch Variation, for
change, a player on either side of the board who many years considered White's safest choice
seeks a double-edged game will find it easy to and a good way to get a small advantage with
do so. Not only are all the pieces on the board out taking many chances. Most contemporary
still present, but the only file down which rooks players have abandoned that point of view; it's
can penetrate is the e-file. However, the 5th, 6th now become obvious that White will have to
and 7th ranks are thoroughly covered. This ne risk something to gain something. However, as
gates the need to put the rooks on an open file at with all openings, White's rewards in these
all and allows them to support pawn advances riskier variations are greater than in the old
on either wing. See theoretical books and mas days, when he would end up in the superior po
ter practice to confirm this. sition but in some drawish ending with his op
c) The two most popular lines against the ponent having one weak pawn or a bad bishop.
French Defence protect the e-pawn: 3 lDd2 and 3 What's the basic idea for White? First of all.
lbc3. Those are what I'll be concentrating upon. convenience. White protects his e-pawn but
Both moves are exceptionally rich in strategic avoids the annoying pin that occurs after 3 lbc3
concepts. Black's responses to the Tarrasch De .i.b4. Then there's flexibility. White can still
fence (3 lbd2), for example, are diverse enough play either e5 or exd5 (or sometimes dxc5) and
to cover in depth major subjects such as the iso doesn't commit himself until he sees what Black
lated queen's pawn, the central majority, and is doing. In that sense he gains the advantage of
pawn-chains. And 3 lbc3 leads to some of the setting the agenda, at least in some main lines.
most complex play amongst the 1 e4 openings. If Black plays 3 . . .lbf6, for example, it's pretty
d) The Advance Variation (3 e5) has some much compulsory to play 4 e5 if one wants an
what narrower strategic scope, concentrated advantage, but after 4 . . .lbfd7, there's already a
mainly around pawn-chains. I talked about 3 e5 choice between 5 f4 and 5 .i.d3, and in the latter
at some length in Chapter 3. Since pawn-chains case White has another choice after 5 . . . c5 6 c3
are also part of the Tarrasch, Classical and lbc6, between 7 lbe2 and 7 lbgf3 . Naturally
Winawer Variations, I've not dealt with the Ad Black has a few options too, but if he commits
vance Variation in this chapter. As always, it's to 3 ...lbf6 they're not so bothersome during the
better to study some variations in depth rather first few moves. On the other hand, 3 ... c5 4
than all of them superficial!y, and I think the se exd5 gives Black two main options, 4 ... exd5
lected variations have the most to offer in terms and 4 . . . �xd5 . Ifhe so desires, White can play 4
of chess understanding. lbgf3 and avoid the 4 . . .�xd5 lines. Naturally
this comes at the cost of submitting to the ne isolani. Or consider 4 ... �xd5 5 lbgf3 cxd4 6
cessity of other moves, and so forth - there are j_c4 �d6 7 0-0 lDf6 8 lbb3, when the knight
always trade-offs. will capture on d4 with a centralized position.
We'll consider the most popular responses The old main line of 3 ...lDf6 4 e5 lDfd7 also
to 3 lbd2: the central counterattack 3 ... c5, and justifies the knight's placement on d2; for ex
the provocative 3 ... lbf6, a variation featuring ample, 5 j_d3 c5 6 c3 lbc6 7 lbe2 cxd4 8 cxd4
pawn-chains. I'll try to provide just enough de f6 9 exf6 lDxf6 10 lbf3 and White's pieces are
tail to communicate the primary ideas in each coordinated, centralized, and aiming at the king
branch. Those two moves are still the main side. Similarly, 5 f4 c5 6 c3 lbc6 7 lbdf3 shows
variations because they challenge the centre in the knight in a favourable light.
a way that forces White to concede something If Black's newly-popular moves 3 ...j_e7 and
and fix the structure. I'll concentrate on them 3 ... lbc6 have less positive effect on the game in
for that reason. Nevertheless, I should say that terms of forcing the play, they also make it dif
3 ... j_e7 has established itself as a main-line ficult for the d2-knight to do as much. More
anti-Tarrasch weapon, and at this point of time over, 3 ... j_e7 and 3 ...h6 (along with the mildly
3 ...lbc6 gives every indication of becoming an revived 3 ... a6) are the kind of useful waiting
alternative of equal worth to the others. Today, moves that we talk about in Chapter 2. Accord
in fact, for the first time since the Tarrasch was ingly, in spite of their own serious drawbacks
introduced, strong players with Black are con (mainly the fact that they don' t attack the centre
sistently playing the moves 3 ...lbc6 (the Gui as effectively in various situations), such third
mard Variation) and 3 ...j_e7 (D), whereas even move alternatives deserve attention. I shall make
3 ...h6 ! ? has been used with success by grand a comparison between 3 ... lbf6 4 e5 lDfd7 and
masters (although much less often). the Guimard line 3 ... lbc6 4 lbgf3 lDf6 5 e5
lbd7 below.
Tarrasch with 3 . . . cs
3 cs (D)
...
I think that there's a common idea here,

namely, that the knight isn't that well-placed on
d2! Can it really justify its position, blocking
off the c l -bishop and queen? Clearly it will
have to move again, and to a useful position.
Which leads to the question: why bail it out? Black challenges the centre immediately. The
Why give it a useful role? Black's traditional idea is twofold, depending upon what kind of
3rd moves do just that; for example, 3 ... c5 leads position he wants and what White does. After 4
to lines such as 4 exd5 exd5 5 lDgf3 lDc6 6 j_b5 exd5 by White, Black can accept an isolated
j_d6 7 dxc5 j_xc5 8 0-0 lbge7 9 lbb3 ; in that pawn by 4... exd5, or undertake to work with a
variation the knight on d2 has become a superb central/kingside majority by means of 4 ... �xd5.
one, gaining a tempo on Black's bishop and I'll illustrate those options using games.
covering the d4-square, right in front of the 4 exd5
FRENCH DEFENCE 1 73
a) One idea after 4 c3 is 4 . . . cxd4 (4 ...lbf6 5 4 ... exd5 is a classic, well-respected system
e5 lDfd7 is a transposition to 3 ...lbf6) 5 cxd4 that directly tests an isolated queen's pawn po
dxe4 6 lbxe4 ..ib4+ 7 lbc3 lDf6 8 lDf3 0-0, sition. White's next few moves have histori
when Black is a tempo up on some well-known cally been the choice of most grandmasters.
isolated queen's pawn positions from the Caro 5 ..ibS+ (D)
Kann and Nimzo-Indian. The more common move-order is 5 .!Ligf3
b) 4 lbgf3 (D) is a main option that I won't lbc6 (in spite of appearances, 5 ... c4 ! ? seems to
go into except to point out three unique, non be holding its own theoretically, but White is
transpositional lines: generally not put off by it) 6 ..ib5. This trans
poses, and is the usual route, to the main line.
Here an easy answer to 6 ..ie2 is 6 .. .!Lif6 (or
.
6... cxd4 7 0-0 ..id6) 7 0-0 ..id6 8 dxc5 �xc5 9

lbb3 ..ib6 ! ; compare the main lines below.
b l ) 4 ... lbf6 5 exd5 lbxd5 ! ? and, for exam

ple, 6 lDb3 lbd7 7 g3 ..ie7.
b2) 4 ... lbc6 5 ..ib5 (for 5 exd5 exd5 see be
low) 5 ... dxe4 (5 ... cxd4) 6 lbxe4 ..id7 7 0-0
lbxd4 8 ..ig5 f6 9 lbxd4 cxd4 10 ..ih4 ..ie7 l l 5 lbc6
.•.
c3 ..ixb5 1 2 �h5+ 'lt>f8 1 3 �xb5 �d5 with an 5 ... ..td7 6 �e2+ ! ? ..ie7 7 dxc5 .!Lif6 8 .!Lib3
excellent game. 0-0 intending ....l::i.e8 has always offered enough
b3) 4 ... cxd4 is a third choice. You can refer play to equalize. The more interesting chal
to theory for the details. lenge to ... ..td7 systems begins with 5 .!Ligf3
and goes 5 ... lbf6 ! ? 6 ..ib5+ ..id7 ! ? 7 �xd7+
Recapture with the Pawn lbbxd7 8 0-0 ..ie7 9 dxc5 lbxc5 . Some top
players use this for Black with the idea of get
4 exdS (D)
... ting rid of his bad bishop for White's good one.
The trade-off is that it is much easier for White
to maintain a d4 blockade after simplification.
The line may well be equal for Black with great
care, but it's very hard to get winning chances.
6 lbgf3 ..id6 7 dxc5
7 0-0 ! ? cxd4 at best gets to the same position
but gives Black more options, as in these sam
ples from the 1 974 Karpov-Korchnoi Candi
dates match after 8 lbb3 lbe7 9 lbbxd4 0-0 I O
c3 ..tg4 1 1 �a4 (D):
a) l l . ....ih5 12 ..ie3 �c7 1 3 h3 (. .. lbxd4
was threatened) 1 3 ... lba5 ! 14 ..id3 lbc4! 1 5
lbb5 �d7! 1 6 ..ixc4 dxc4 1 7 .l::i.f dl lDf5 ! (it
doesn't take much to get in trouble: l 7 ... a6? 1 8
Karpov - Uhlmann
Madrid 1973
10 .i.gS 0-0 11 .i.h4

This is straightforward positional chess:
White wants to exchange pieces via .i.g3, be
cause simplification helps to secure the static
disadvantages of the isolated pawn. B ut it's not
just any piece that White wants off the board:
it's Black's good bishop that might, for in
stance, have supported a freeing pawn-push to
d4.
11 ....i.g4 (D)
The right move, neutralizing a defender of
l:i.xd6 �xb5 1 9 �xb5 axb5 20 lbd4 and White' s d4. If he later captures on f3 and White recap
pieces dominate) 1 8 �xc4 .i.xf3 1 9 gxf3 lDxe3 tures with the queen, two fewer pieces will
20 fxe3 �xh3 21 lbxd6 �g3+ 22 'lt>fl �xf3+ protect that crucially important square. In the
23 'lt>e l �g3+ 1h- 1h Karpov-Korchnoi, Mos meantime the pin is awkward to meet.
cow Ct ( 1 2) 1 974.
b) 1 l ...�d7 12 .i.e3 a6 1 3 .i.e2 lbxd4 14
�xd4 lbc6 15 �d2 l:i.fe8 1 6 l:i.adl l:i.ad8 1 7 .i.b6
.i.c7 (it's OK to simplify because every piece of
Black's is active and White has to watch out for
...d4) 1 8 .i.xc7 �xc7 1 9 l:i.fel h6 20 h3 .i.f5 2 1
.i.fl l:i.xel ! 2 2 �xel �b6 (always a good square
for the queen, eyeing d4 and b2) 23 l:i.d2 .i.e4!
24 �e2 lba5 25 �dl �f6 with equality, Kar
pov-Korchnoi, Moscow Ct ( 1 6) 1 974.
7....i.xcS 8 0-0 lbe7 9 lbb3 .i.d6 (D)
9 . . . .i.b6 ! ? 10 l:i.el and .i.e3 has always been
j udged to be in White's favour and it probably
is; nevertheless, White doesn't have much af
ter the gambit 1 0 . . . 0-0 1 1 .i.e3 .i.g4 1 2 .i.xb6
�xb6 ! 1 3 .i.xc6 lbxc6 14 �xd5 lbb4. 12 .i.e2 .i.hS
This is a funny-looking retreat, but it makes
sense to be able to avoid more exchanges by
putting the bishop on g6. But these are difficult
positions, and later Uhlmann found a better
way to play it. See the next game.
13 l:i.el �b6
Again Black eyes d4. But Karpov will be
very careful not to let the isolated queen's pawn
advance.
14 lbfd4 .i.g6 15 c3 l:i.fe8 16 .i.fl .i.e4 17
.i.g3 .i.xg3 18 hxg3 (D)
Finally White rids himself of Black's good
bishop. It's amazing how he handles this posi
tion, since Black looks actively placed. The
next idea is exactly what Black usually does,
We'll follow three main games from the point but it seems to land him in more trouble:
after 9 ... .i.d6: 18 . as
. .
FRENCH DEFENCE 1 75
32 .l::i.c l + 33 'lt>f2 .l::i.c2+ 34 'lt>e3 .te4

•••
Or 34 ... J::i.e6+ 35 .l::i.xe6 fxe6 36 l:txb6.

35 .l::i.xf7 .l::i.g6 36 g5 'lt>h7 37 .l::i.fe7 lbb2 38
.te8 .l::i.b3+ 39 'lt>e2 .l::i.b2+ 40 'lt>el l:td6 41
.l::i.xg7+ 'lt>h8 42 .l::i.ge7 1-0
There could follow 42... .l::i.b l + 43 'lt>d2 l:tb2+
44 'lt>c3 .l::i.c2+ 45 'lt>b3 .l::i.c8 46 .id7 l:tf8 47 f5 !.
etc. An unassuming masterpiece.
The next game shows the good points of

having an isolated queen's pawn, namely. in
creased activity.
Vogt Uhlmann
-
Perhaps 18 ....l::i.ad8 ! ? should be tried, although East Gennan Ch (Potsdam) 1974

White still has the better game.
19 a4 lbxd4 20 lbxd4! 10 .tg5 0-0 11 .th4 .tg4 12 .ie2 (D)
Not 20 cxd4 lbc6 with equality.
20...lbc6
After 20 ... �xb2? 2 1 lbb5 White threatens
.l::i.e2 and lbc7.
21 .tb5 .l::i.ed8 22 g4! (D)
12 ....l::i.e8!
Uhlmann's improvement over the Karpov
game; it essentially gains a tempo for central
action. It seems that ... .th5 was too slow.
13 .l::i.el �b6!
The move of the game ! And the timing is This covers d4 and peeks at b2, because
perfect. White allows Black to rid himself of his ... a5-a4 will expose that square.
isolani; at first sight he seems to have only the 14 lbfd4
slightest of advantages. The point is that Black's 14 .txe7?! .l::i.xe7 ! leaves Black terribly ac
bishop will never get back for defence. How tive, with the idea of 15 �xd5? ! lbb4 ! .
this happens is worth seeing: 14 lbg6! (D)
•••
22...lbxd4 23 �xd4 �xd4 24 cxd4 .l::i.ac8 25 Suddenly White is in trouble because his
f3 .tg6 26 .l::i.e7 b6 27 .l::i.ael h6 pieces are too loose. Black threatens ... lbxh4.
No better is 27 .. .f6 (to get the bishop back to 15 lbxc6
f7) 28 .l::i. 1 e6 ! .l::i.c l + 29 'lt>h2 .l::i.b8 30 .l::i.d6 and White strengthens Black's centre, but it' s lhe
.l::i.dd7. only move. 15 .txg4? .l::i.x el + 16 �xe l lDxd4
28 .l::i.b7 .l::i.d6 29 .l::i.ee7 h5 30 gxh5 .txh5 31 threatens both c2 and h4, and there's no defence
g4! .tg6 32 f4 because after 1 7 lbxd4 �xd4 Black wins a
So simple. piece.
21 lbd4
One point of 20 . . .h5 is 2 1 �xh5 a4 22 lbd4
�xb2.
21. h4 22 .i.h2 .i.f6 23 l:tdl a4!
••
More space !
24 �c2 �c5 25 �d3 l:i.e8 26 b4! axb3 27
axb3 �b6 28 b4 (D)
White fights back, securing the c5-square as
a potential outpost for his knight.
15 l:i.xe2! 16 l:i.xe2 bxc6

.•.
There's no hurry to take the rook on e2 be

cause the f-pawn is pinned.
17 .i.g3 .i.e7!
Black surprisingly avoids the exchange of
bishops, seeing that his bishop will become
strong and White's subject to harassment. But to
lose time like this takes some courage. Actually,
1 7 ... .i.xg3 ! ? 1 8 hxg3 lDf8 ! wasn't bad either
because Black's knight would cover all the good
squares from e6. This is more or less the end of 28•••l:i.e4 29 .i.d6
the opening but it's worth seeing more because The logical 29 lbb3 is frustrated by 29 ... l:i.c4 !
the moves are wonderfully instructive. 30 l:i.c l �b5 !, preventing 3 1 lbc5?? due to
18 h3 .i.xe2 19 �xe2 a5! 3 1 . ..l:i.xc5 ! .
Threatening . . .a4 and ...�xb2. 29 .i.xd4 30 cxd4 �d8! 31 .i.c5 lbf4
•.•
20 c3 h5! (D) Now . . . h5-h4 is looking especially fore

sighted because White's kingside is vulnera
ble.
32 �f3 �g5 33 l:i.al
Trying to get the bishop back for defence
loses the d4-pawn: 33 .i.d6 lbe2+.
33 'lt>h7 34 'lt>h2 l:te6! 35 �g4 �xg4! 36
.•.
hxg4 l:i.e2! (D)
A great stratagem: Black puts White's bishop

into temporary oblivion. Perhaps even more
significant is Black's aggrandisement of space.
This by itself is a good thing, as long as you
don't give the opponent targets to attack by do
ing so.
FRENCH DEFENCE 177
Well anticipated. White can't hang on to ev

erything. And White's officially 'bad' bishop
(because of the dark-squared centre pawn) really
is bad!
37 .l:i.fl
White loses material anyway after 37 @gl
tlld3.
37 tlld3 38 f4 tllf2! 39 l:i.al tllxg4+!?
•.•
Instead, 39 ... @g6! ! would have been a bril

liant move to make just before the time-control,
based upon 40 ..if8 tllxg4+ 4 1 @h3 @h5 42
l:i.gl l:i.e3+ 43 g3 tllf2+ 44 @h2 h3 ! 45 ..ixg7
@g4 ! and Black will actually checkmate after
... l:i.e2. The rest of the game is also enjoyable:
40 @h3 tlle3 41 l:i.a6 l:i.xg2 42 @xh4 (D) 12 . tt:irs
. .
42 l:i.xc6? l:i.g3+ 43 @h2 tll g4+ 44 @h l h3 Black concentrates upon d4 as usual. He can
45 b5 tllf2+ 46 @h2 l:i.g2#. also think about a move like ... tllh4. Or, after
his dark-squared bishop vacates d6, the knight
can go there to great effect.
13 c3 �f6 14 ..tc2 l:i.d8 15 �d3 g6 (D)
16 g4 was threatened.
42 l:i.g4+! 43 @h3 l:i.xf4 44 l:i.xc6 g5! 45 b5

•••
g4+ 46 @h4
46 @h2 l:i.f2+ 47 @g3 l:i.f3+ 48 @h4 f6 ! with
... l:i.h3# next.
46 f6! 47 ..td6 l:i.f3 48 ..tg3 @g6 0-1
.•. 16 �d2
Rublevsky-Dolmatov, St Petersburg 1 998
Adams - Yusupov went 16 ..id2 a5 ! (with the idea ... b6 and ... ..ta6;
Port Barcares 2005 this also gains space, a key consideration for
both sides) 1 7 a4 b6 1 8 ..ie3 ..ta6 19 �d2 tllxe3
10 l:i.el 0-0 11 ..id3 (D) 20 �xe3 @g7 2 1 ..id3 l:i.e8 22 �d2 ..ic4 ! 23
This position has been considered the main tllbd4 tllxd4 24 tllxd4 ..ic5 with equality.
line for some time now. 1 1 ..id3 doesn't cover 16 ... ..tfS 17 �f4 �g7
d4, but prevents 1 l .....ig4? due to 1 2 ..ixh7+ 1 7 ... ..id6 1 8 �a4 ..id7 doesn't look so bad.
@xh7 1 3 tllg 5+. The d4-square never seemed but White could repeat by 1 8 �d2.
to be quite enough for White to win with any 18 ..td2
way. Adams-Lputian, Armenia-Row (Moscow)
l l.. h6 12 h3
. 2004 pitted a super-grandmaster against one of
From what we've seen, stopping ... ..tg4 is a the world's leading French Defence experts: 1 8
good idea. h 4 �d6 (with so much pressure on d4, Black
can afford an ending) 19 .i.d2 .i.e6 20 l:i.adl It's very risky at best to leave only the oppo
�xf4 21 .i.xf4 d4 ! (D). site-coloured bishops on the board. Generally
bishops of opposite colour favour the attacker,
and only in a simplified ending do they become
drawish.
24 lbxd4 lbxd4 25 .i.xd4 (D)
As always, the thematic break. Normally if it

works (as it does here) Black will equalize be
cause he liquidates his potentially weak IQP.
However, the main idea is to free Black's pieces,
in this case the rook and e6-bishop. The next Now we have opposite-coloured bishops with
few moves were 22 .i.xf5 .i.xb3 23 axb3 gxf5 Black's king a little weak. Generally this is
24 lbxd4 lbxd4 25 cxd4 .i.xd4 26 .i.xh6 .i.xb2, enough to make the attack work.
with equality. 25 l:i.e8 26 �c7! .i.xa2? 27 �c6!
•••
18 g5 19 �h2 b6? !
.•. Ouch. This hits a8, e8 and h6.
There have been two other suggestions here, 27 l:i.f8 28 �xh6 f6 29 l:i.e7 l:i.f7 30 l:i.del
•••
both reasonable-looking, but perhaps not fully l:i.af8 31 l:i.xf7 'lt>xf7 32 g4! 1-0
equal. Maybe l 7 ... .i.d6 was the real solution. If the queen goes to f4 or f3 in order to keep
a) McDonald offers 19 . . . .i.f8 ! 20 .l:i.adl .i.d6 f6 guarded, it's mate on h7. If 32 ... �g6, then 33
2 1 �h l . l:i.e7+ wins the queen.
b) 19 . . .lbd6 ! ? 20 l:i.adl lbc4 2 1 .i.c l .i.f8 22
lbbd4 .i.d6 is given by Pedersen. Recapture with the Queen
20 l:i.adl .i.a6?! 21 .i.xf5! �xf5 22 .i.e3
.i.c4!? 23 lDbd4! (D) 4 �xd5
••.
This recapture represents a very different ap

proach from that of 4 ... exd5 , as we'll see.
5 lDgf3 cxd4 6 .i.c4 �d6 (D)
6 . . .�d8 is also played from time to time. The
only unique variation of note in that case arises
when Black follows up with ... a6, . . .�c7 and
. . . .i.d6. That is quite rare but interesting. In any
event, we'll concentrate upon 6 . . .�d6.
We see that with 4 ... �xd5, Black is willing
to lose quite a lot of time to get to a Sicilian-like
structure with an extra central pawn (the one on
d4 is usually recovered by White) and a king
side majority (4:3). In doing so he braves many
attacking tries by White, whose lead in devel
opment and tactical tricks were probably the
23 .i.xd4?
••• reason that French Defence players were put
FRENCH DEFENCE 2 79
off this line for so many years. With more and lbxd4 .i.c5 ! is also reasonable: 1 2 lDb3 .i.e7 1 3
more willingness to defend difficult positions .i.f4 .i.d7 14 .i.e2 { White intends to exert pres
in openings, players who recognized the very sure from f3; this is the standard plan } 14 ...ltJd.5
real advantages in Black's pawn-structure and 15 .i.g3 h5 ! 16 h3 h4 17 .i.h2 0-0-0, Akopian
his smooth development began to try out the Shirov, Merida 2000) 1 1 .i.f4 °VJll.xd4 1 2 lDxd4
line. A standard plan goes ... lbc6, . . . lbf6, . . . a6, l:i.c8 (or 1 2 ....i.c5) 1 3 .i.e2 lbd5 14 .i.g3 h5 !
...°VJll.c 7 and ... b5 with ... .i.b7, depending upon (D).
how much is permitted him. These are typical
Sicilian ideas and, as in the Sicilian, players re
alized that the pawn on e6 in particular makes
Black's position hard to crack. White in turn
uses his lead in development to restrict Black's
own pieces from getting out by posting pieces
on support-points like e5, still looking for and
often finding attacking chances. For Black, it' s
all about structure: White doesn't have a centre
pawn to attack with, and his c- and f-pawns
tend to take a long time to enter the fray. About
1 5 years ago this turned into the most popular
line of the Tarrasch at the very top levels and it
is still leading to great wins for both sides.
We'll explore three games from this position.
Black even has some initiative ! Acs-Shaked,
Lastin - Bareev Budapest 1997 continued 15 h4 .i.c5 16 lDb3
Russian Cht (Sochi) 2004 .i.b6 17 c4 lbe7 1 8 .i.d3 f6 ! .
1 0...a6 l l .i.b3 °VJll.c7 1 2 °VJll.f3
7 0-0 lbf6 8 lbb3 lbc6 9 lbbxd4 lbxd4 10 A renowned line that has been analysed to
lbxd4 death is 1 2 l:i.e l .i.d6 1 3 lDf5 .i.xh2+ 14 Wh l
White has recovered his pawn. In the early 0-0 15 lbxg7 l:i.d8, eventually leading to a draw.
days of the variation, quite a few players tried 12 .i.d6 13 Whl
..•
10 °VJll.xd4 (D). White uses this defensive technique a lot; he

In spite of White's lead in development, refuses to weaken himself by h3 or g3.
Black has a solid position. If the game actually Another line runs 1 3 h3 0-0 14 .i.g5 lbd7 15
reached a simplified ending with no structural c3 b5 ! ? (D).
changes, Black's central majority would give This standard 'sacrifice' usually works be
him the advantage. In any case, Black's most cause Black gets active play regardless of the
popular continuation is 10 ... .i.d7 ( 1 0 ... °VJll.xd4 1 1 material: 16 .i.c2 (too greedy is 16 "VJl/.xa8 .i.b7
1 7 �xf8+ { not 1 7 �a7? .i.c5 } 1 7 ... lbxf8; al do so. This type of position has arisen repeat
though White has two rooks for a queen, Black' s edly. Who prevails is a question of specifics.
bishops rake the kingside and ... �c5 together 17 eS! 18 'lt>xh2!?
.•.
with ... lbg6-h4 can follow) 1 6... .i.b7 1 7 1lV115 Perhaps better is 18 lDf3 .i.g4 1 9 .i.g5 ( 19
g6 1 8 1lV114 and now: 'lt>xh2? 0-0) 1 9 ... h4 20 .i.xf6 gxf6 2 1 'lt>xh2
a) In one game Black got careless and fell hxg3++ 22 'lt>gl .
for a nice sacrificial piece attack: 1 8 ... e5 ? 1 9 1 8 h4 19 'lt>gl (D)
•••
lDf5 ! f6 ( 1 9 . . .gxf5 2 0 .i.xf5) 2 0 lbxd6 �xd6 2 1 This time 1 9 'lt>g2 ! ? is worth looking into .
.i.h6 l:i.fe8 22 l:i.ad l �e7 ? 2 3 .i.b3+ 'lt>h8 24 All these moves are hard to assess.
l:i.xd7 ! 1 -0 Azarov-Wiedenkeller, Saint Vincent
ECC 2005 .
b) 1 8 ... l:i.fe8 ! 19 l:i.fel lbb6 20 l:i.e2 j.,e7
(20 ... e5 ? 2 1 lDf5 ! gxf5 22 .i.f6) 2 1 l:i.ael .i.xg5
22 �xg5 lbd5 23 l:i.e4 f6 ! 24 �h4 e5 25 .i.b3
'lt>g7 26 .i.xd5 .i.xd5 with complicated play in
which Black seems to have the better of it,
Tiviakov-Lalic, Port Erin 2005 .
13 .tes
.•.
Another characteristic move to know about.

Black anticipates .i.g5 (attacking his f6-knight),
and also forces White to commit to a method of
defending his knight.
14 .i.e3
Logically developing and covering d4. In
stead, 1 4 c3 .i.d7 15 .i.g5 .i.xd4 ! ? ( 1 5 ... 0-0 is 19 0-0-0?
•••
solid) 1 6 cxd4 .i.c6 1 7 �e3 lbd5 is one of the Bareev gave the improvement 1 9 ... 'lt>f8 ! 20
ideas that originally made the ... .i.e5 idea popu lDf3 .i.g4 2 1 .i.g5 e4 22 .i.f4 exf3 23 �xa6 !
lar. Black blockades the IQP and equalizes. bxa6 24 .i.xc7 h3 25 'lt>h2 .i.f5 26 .i.b6, and the
14 .i.d7 15 l:i.adl h5! (D)
••• situation is still uncertain.
Yet another standard procedure ! From now 20 lDf3 hxg3 21 fxg3?
on White has to be careful about ... lbg4 ideas. 2 1 �c4 ! was practically winning, according
16 �e2! to Bareev. White is keeping the extra piece:
Not 16 h3?? lbg4 (threatening ... .i.xd4 and 2 l .. .gxf2+ 22 .i.xf2 .i.c6 23 l:i.xd8+ 'lt>xd8 24
...�h2#) 1 7 l:i.fel .i.xd4 1 8 .i.f4 �b6 and wins. �d3+ with .i.g3 to follow.
16 .i.xh2 17 g3
••. 21 e4 22 .i.f4
•.•
This is the idea behind 'lt>h 1 ; White wants to Perhaps 22 �c4 ! ? was still the move.
win the bishop, although it is obviously risky to 22 exf3 23 �f2 �c6! 24 l:i.d6? (D)
.•.
FRENCH DEFENCE 281
What's this move about? In some positions

we get from 4 . . . ii'xd5, it helps Black rearrange
by . . . ii'c7 and . . . .td6. But in this case he wants
to strike out with . . . b5 before anything else hap
pens. If White's bishop retreats, he will have
fewer worries about a sacrifice on e6, a problem
that is always present when White has rooks on
the open central files and a knight on d4. To
some extent the game is becoming a race.
10 0-0-0 b5 11 .td3 .tb7 (D)
24 tllg4!
•.•
Bareev had presumably seen this blow long

before.
25 .I:txc6+ .txc6 26 .txf7
The queen can't move because . . .f2+ forces
mate.
26...tllxf2 27 .te6+ .td7 28 .txd7+ .I:i.xd7
29 @xf2 .I:th2+ 30 @xf3 .I:txc2 31 .I:tel .I:tf7 0-1
Another important variation is:

7 ii'e2 (D)
I'll show two games from this position.
Oral - Khuzman
Batumi Echt 1999
12 tllbxd4 tllxd4 13 tllxd4 ii'd5 ! (D)
This is a line that goes directly for the kill.

White forgets about recovering his pawn on d4
for the moment and concentrates on quick de
velopment, normally including 0-0-0. The cur
rent main line continues:
7 ... tllf6 8 tllb3 tllc6
Black has to catch up in development and Black attacks g5 and a2 (and g2!).
anyway, the harder he can make it for White to 14 .txf6 gxf6 15 tllxb5!?
recover his pawn on d4, the better. But White 1 5 @bl ! is solid. Then 1 5 . . .0-0-0 1 6 tllf3
isn't going to slow down. could be followed by c4 or .te4 depending
9 .tg5! a6 upon Black's response.
15 ii'xa2 16 tllc7+ <Jile7 17 ii'h5 (D)

••.
Not 1 7 tllx a8?? .th6+ and White has to give

up his queen on e3.
1 5 . . . 0-0-0!? is obviously risky. White might

try 1 6 tllf3 @b8 1 7 tlle5 .I:thf8 18 .te3 ! ? .
16 l:.hel
White piles up on e6. This is one of those
17 ....th6+! openings where Black knows what's coming
A shot! The obvious 1 7 ... l:.d8? 1 8 ii'c5+ but can't always stop it.
l:.d6 fails to 1 9 ii'xd6+ ! @xd6 20 .tc4+ <Jilxc7 16...l:.fe8
21 .txa2. A typical tactic follows 1 6... l:.fd8 ! ? 1 7 tllf5 ! ?
18 ii'xh6 ii'al+ (this move seems to appear in every line ! But in
The point was to distract White's queen from this case it's probably not that great) 1 7 . . . exf5
a5. 1 8 ii'xe7 ii'xe7 1 9 .I:txe7 .txg2 20 .I:i.gl <Jiff8
19 @d2 ii'a5+ 20 @cl ii'xc7 (forced) 2 1 .txf6 gxf6 22 l:.eel .te4 and Black
Black wants to keep playing instead of ac should come out OK.
cepting 20. . . ii'al+ 2 1 <Jifd2 ii'a5+. 17 t2Jf3 h6
21 l:.hel!? .td5 22 .te4 .txe4 23 l:.xe4 ii'aS Better seems 1 7 ... l:.ac8 1 8 tll e5 l:.c7. The
11z_11z text-move is weakening and Black probably
Typical of the back-and-forth nature of this overlooked the reply.
variation. 18 tlle5! hxg5?! 19 hxg5 g6 20 gxf6 .txf6
21 f4! (D)
Rozentalis - Luther
Panormo ECC 2001
12 @bl
This looks slow but is interesting to compare
with Oral-Khuzman. There the key move . . . ii'd5
hit a2 as well as the kingside, while here Rozen
talis protects his a-pawn, at the cost of time.
12 ... .te7 13 tllbxd4 tllxd4 14 tllxd4 ii'c5
(D)
Black has also played simply 14 . . . 0-0 here,
but that's risky. The text-move gets the queen
off the d-file and gains a tempo by attacking
White's g5-bishop.
15 h4!
Or maybe he hasn' t gained a tempo, since Now how can Black get any play going?
the h-pawn will be handy in an attack! . . ..txe5 will leave the dark squares unbearably
15 ...0-0 weak.
FRENCH DEFENCE 283
21. .I:i.ad8 22 �g4 i..g7?

•• 5 c5 6 c3 tllc6
•.•
He had to try something like 22 . . . �f8. 6 . . . b6 intending ... i.. a6 looks logical. Never
23 tllxf 7! �xf7 theless, after 7 tllh3 (or 7 �g4) 7 ... i..a6 8 i..xa6,
Black was probably counting upon 23 ... .I:i.xd3 White misplaces Black's knight and can put his
24 .I:i.xd3, but it's not even close. Rozentalis space advantage to good use. If White is really
gives 24... �f2 25 tllh 6+! �f8 26 .I:t.xe6 .I:t.xe6 worried about this (and no one seems to be) he
27 .I:i.d8+ �e7 28 .I:i.e8+ �xe8 29 �xe6+ �d8 can play 5 c3.
30 tllf7+ �c7 31 �d6+. Now I'll examine the two main moves, 7
24 i..xg6+ �f8 25 .I:t.xd8 1-0 tlle2 and 7 tll gf3.
There follows 25 ....I:i.xd8 26 �xe6 with mates
threatened on both f7 and e8. The Traditional 7 tlle2
Tarrasch with 3 . . . ltJf6 7 tlle2

This continuation has dominated the practice
1 e4 e6 2 d4 d5 3 tlld2 tllf6 of 3 ... tllf6 since time immemorial, but recently
With this move Black challenges White to it's been sharing the spotlight with 7 tllgf3.
set up a pawn-chain and the game enters typi 7 cxd4 8 cxd4 f6 (D)
.•.
cal French territory. Since White's e-pawn is

threatened, there's nothing much else to do but
advance.
4 e5
4 exd5? ! exd5 already favours Black because
the knight on d2 is poorly placed for an Ex
change Variation, as you can confirm by experi
mentation.
4...tllfd7 (D)
9 exf6
Don't fall for 9 f4? ! fxe5 I O fxe5 tllx d4! l l
tllxd4 �h4+ 1 2 g3 �xd4.
A big-time alternative whose consequences
have never quite been solved is the tactical and
less common 9 tllf4. It's a real mess, and unfor
tunately very theoretical, in the sense that many
logical moves are losing and the forced nature
of the play doesn't admit of time-consuming
Now we have a position characterized by over-the-board reflection. I'll give a few im
pawn-chains. The traditional choices here are 5 portant moves, skipping most of the options:
i..d3 and 5 f4. 9 ... tllxd4 (9 . . . �e7 I O tllf3 fxe5 l l dxe5 tlldxe5
1 2 tllxe5 tllxe5 1 3 �h5+ tllf7 14 0-0 g6 1 5
Development by 5 i..d3 �e2 makes i t hard for Black to develop) l 0
�h5+ �e7 (D).
5 i..d3 l l exf6+ ! ( l l tllg6+ hxg6 1 2 exf6+ �xf6 ! ?
5 c3 is another move-order, intending 5 . . . c5 1 3 �xh8 �f7 is a very old line, but at the least
6 i..d 3. White might prefer that because of the Black can also transpose by l 2. . . tllxf6, so why
next note. give him an additional option?) l l . . . tllxf6 (now
Now if White plays conventionally by 1 1 0-0

( 1 1 .tbl ! .td6 12 ii'd3 is much harder to meet;
compare what follows) 1 l ....td6 1 2 .tbl 0-0! 13
ii'd3 Black replies 1 3 ....I:i.d8 ! 14 ii'h7+ 'it>f7 and
plays ... tllf8 next. Although gaining a tempo
with 1 1 .tbl poses Black a problem, it may be
amenable to solution.
10 tllf3 .td6 11 0-0 (D)
1 1 .tf4 ii'a5+ confuses White's pieces: 1 2
ii'd2 ( 1 2 'it>fl ii'c7 1 3 .txd6 ii'xd6 i s equal)
1 2. . . .tb4! 1 3 tllc 3 0-0 14 0-0 tlle4.
l l ... 'it>xf6? 1 2 'iV114+ g5 13 tllh5+ 'it>f7 14 ii'xd4

is killing) 1 2 tll g6+ hxg6 1 3 ii'xh8 'it>f7 14 ii'h4
(notice the trap 14 tllf 3?? .tb4+, winning the
queen; people have lost this way ! ) 14 . . . e5 15
tllf3 e4 ! ? (15 ... tllxf3+ 16 gxf3 .tf5 1 7 .txf5
gxf5 1 8 .tg5 also leads to deep analysis) 16
tllxd4 .tb4+ 1 7 .td2 .txd2+ 1 8 'it>xd2 ii'a5+
19 '>t>dl exd3 . These are mad positions; I refer
you to the books and databases.
9 tllxf6
...
It's a shame that this venerable line, which

has generated such great games and so many at
tractive and thought-provoking ideas, has be 1 1 ii'c7
...
come laden with theory in so many byways. Black aims at White's kingside, but even
Nevertheless, we'll take a look at the basic more significantly, he stops .tf4, which effec
complex of variations, so that you can get a tively exchanges Black's good bishop. This
start in understanding what's going on. comes at the cost of committing the queen
9 ... ii'xf6, keeping an eye upon the critical early on, which could be considered a relative
e5-square, is another idea that is moderately loss of time.
alive after some years of experimentation. The I'd guess that at least thousand pages of anal
essential idea can be seen after 10 tllf3 h6 (D) ysis (adding up all sources and annotations)
(to stop .tg5-h4; e.g., 1 0 . . . .td6 1 1 .tg5 ii'f7 1 2 have been devoted to the lines beginning with
.th4 0-0? 1 3 .txh7+). 1 1 . . .0-0 (D).
FRENCH DEFENCE 285
I'm going to stick with the queen develop 8 0-0 .td6 9 c4 0-0 10 c5 .te7 1 1 tt:Jb3 h6 12
ment instead. But I do think that it's intriguing .te3 e5! 13 dxe5 tt:Jdxe5 14 tLixeS tt:Jxe5 15
to compare this position with the Sicilian line 1 .te2 c6 (D)
e4 c5 2 tt:Jf3 d6 3 d4 cxd4 4 tt:Jxd4 tLlf6 5 tt:Jc3
tt:Jc6 6 .te2 e5 7 tLlb3 .te7 8 0-0 0-0 9 .te3 .te6
(D).
This is the relevant position. Black has made

his . . . e5 break and retains a healthy centre be
cause his d5-pawn is fully protected. Contrast
It's a mirror image ! Notice especially the this with the case of the 3 . . .tt:Jf6 main lines: be
roles of the bad bishops protecting backward cause of the insertion of the moves ...c5 and
pawns. The biggest difference in structure is . . . cxd4, Black almost always ends up with a
Black's open f-file in the French Defence. He weak isolated queen's pawn if he plays . . .e5.
also has the opportunity to attack White's vul 16 l:.cl °fiig6 17 .thS °fiih 7!?
nerable d-pawn. Both of these are comparative Or 17 . . . °fiif5 with the idea . . .tt:Jg4. Black may
advantages. But in the Sicilian Defence posi stand slightly better.
tion, Black has an important minority attack 18 l:.c3 .tfS 19 tt:Jd4 .tg6 20 .te2 lbe8 21
with . . . a6 and ... b5, by which he gains space, at tt:Je6 l:.f7 22 tt:Jf4 .te4 23 .thS g6 24 .te2 ..tf6
tacks the queenside, and helps with control of 25 .td4 °fiih 8! (D)
d5. In the French, Black has nothing of the sort;
as such, his strategy is more piece-based, with
moves like . . . "file? (supporting the idea of . . . e5),
. . . tt:Jh5 (or . . .tt:Jg4), intending to attack on the
king side.
And as long as we're digressing, a thought
provoking comparison also arises between this
sort of position and that of the Guimard De
fence, 3 tLld2 tt:Jc6. It turns out that the lack of a
c-pawn in our 3 tLld2 tLlf6 French Defence can
be a disadvantage in comparison with the Gui
mard ! Let's look at a fairly normal example:
Rasik - Cernousek
Ostrava 2005
26 l:.g3 @h7 27 .tc3 °fiif8!
1 e4 e6 2 d4 d5 3 tt:Jd2 tt:Jc6 4 tt:Jgf3 tt:Jf6 5 e5 Threatening both 28 ... "fiixc5 and 28 . . . .th4.
tt:Jd7 6 .td3 f6 7 exf6 °fiixf6 Suddenly Black is winning.
7 . . .tt:Jxf6 is also played in this kind of posi 28 tt:Jd3 tt:Jxd3 29 .txd3 "fiixc5 30 °fiig4?
tion. The same ideas apply. .txd3 0-1
There would follow 3 1 l:.xd3 i.xc3 32 l:.xc3

�xf2+ 33 l:.xf2 l:.el +.
Now let's return to the 3 ... tllf6 variation with

1 1 . . .0-0. As I said, I'm going to be following
l l .. .�c7, but here are a few notes on 1 1 . ..0-0,
partly for the sake of showing some typical
themes but mainly to demonstrate how crazy
and specific these lines have become ! After
1 1 . . . 0-0, White usually takes the opportunity to
exchange bishops by 1 2 i.f4 (which is why
l l ...�c7 is played) 1 2 ...i.xf4 1 3 tllxf4 tlle4.
The main line goes 14 tlle 2 l:.xf3 15 gxf3 tllg5
(D).
We now return to the position after l l . . . �c7
(D):
Now:
a) One serious option is 1 6 'it>h I e5 ! 17 dxe5
tllxe5 (right at this moment in time this seems Let's look at a real game:
better than 1 7 ... tllxf3 1 8 i.xh7 + ! 'it>h8 1 9 tllg l
tllc d4! 20 i.d3 !) 1 8 tllg l �f6 1 9 i.e2 i.d7 !? Biti - Gleizerov
(this has ideas of both ...i.c6 and the powerful Zadar 2005
... tlle6-f4! - always remember the power of a
knight on f5/f4 in front of doubled pawns; Black 12 i.gS
could try 1 9 . . . tlle 6 ! ? 20 �xd5 'it>h8 21 l:.fd I The main line, which has the logical idea of
i.d7) 20 �xd5+ 'it>h8 21 i.b5. White may have i.h4-g3 to exchange off that good bishop of
a small edge in the whole line; this is just to get Black's. Then White can start thinking about
you started! occupying the juicy outpost on e5 by means of
b) 1 6 f4 tllf3+ 1 7 'it>g2 (D). l:.el . If Black plays ... e5 first, he gets saddled
Now look at the deranged things some play with an isolated queen's pawn and remember
ers do: 17 ... �h4 ! ? 18 'it>xf3 �h3+ 19 tllg 3 e5 that they tend to be weaker after simplification,
20 'it>e3 ! exf4+ 21 'it>d2 fxg3 22 hxg3 �h6+ 23 especially the exchange of the better bishop.
f4 tllxd4 ! ? (23 ...�f6) 24 l:.hl �6 25 i.xh7+?! That's White's general strategy, but of course it
( ' obviously' better is 25 'it>c l . . . perhaps !) takes time. As explained after 1 1 . . . 0-0, Black
25 ... 'it>f8 26 'it>c I �c6+ 27 'it>d2 �b6 with a tends to rely on piece-play, so he'll start putting
draw ! Unless you adore theory (and making im everything close to the king, provoke weak
provements on moves 20-35 of an opening vari nesses and then strongly consider . . .e5 in order
ation), you might want to play something else. to bring the last pieces into the attack. 1 2 tllc 3,
FRENCH DEFENCE 287
12 g3 and 12 h3 are all interesting alternatives to determine if it' s good. It seems to be in this
which we won't go into. case, although theory hasn 't yet settled down.
12...0-0 13 .th4 tllh5 14 ii'c2 16 gxf3 .txh2+ 17 'it>hl tllf4! (D)
Since Black's doing all right here (I guess),
White might want to try 14 tllc3 a6 15 l:.c l g6 1 6
tlla4. Then best seems 1 6. . . .td7 ! 17 .tg3 ( 17
l:.el l:.xf3 !) of Mamedov-Hanley, Nakhchivan
jr Wch 2003 , when McDonald likes 17 . . . tllxg3
18 hxg3 ii'a5 !?.
14 h6 15 .tg6
...
McDonald also analyses 15 l:.ac l g5 ! (D).
18 tllg3 e5!?
Typical tactics for this variation. As usual,
I'm not going to go into much detail about such
a precise tactical variation. 1 8 . . .ii'd6 used to be
considered brilliant, but maybe not so much
these days. Check the books. 1 8 . . . ii'b6, how
ever, is at the moment theoretically satisfactory
after 19 l:.ad l tllxd4 20 ii'a4 tllxg6.
It's typical of the French Defence that such a 19 l:.fel
move can be good - it also works in several Or 19 'it>xh2 ii'd6 20 .ih7+ 'it>h8 21 dxe5? !
lines of both the Advance and the Winawer tllxe5 2 2 .tf5 .txf5 2 3 tllxf5 ii'e6 ! with a great
Variations. Aside from snatching space and attack, Ulybin-E.Berg, Santa Cruz de la Palma
launching an attack on the king, it gives Black's 2005.
pieces more room to move about safely and not It looks like 19 .th7+? 'it>h8 20 'it>xh2 g5 ! 2 1
get too cramped on the queenside. Black's pos .tf5 gxh4 2 2 .txc8 ii'xc8 2 3 tllf5 ii'd7 ! ! i s win
sible follow-ups include . . ..te8-g6 and . . . ii'g7. ning for Black, Can-E.Berg, Kusadasi 2006.
Black's knight is also better protected after It's enough to make your brain explode. On the
... tllh 5-f4. Specifically, this version of . . . g5 al other hand, these tactics are kind of amazing !
lows Black to exchange White's good bishop 19 .th3! (D)
•.•
for his knight, and threatens . . . g4, weakening

d4.
The analysis continues 1 6 .ig6 (trying to
disrupt Black's build-up; Pedersen analyses 1 6
.tg3 tllxg3 1 7 tllxg3 ( 1 7 hxg3 ii'g7 } 1 7. . .ii'g7
1 8 tllh5 ii'e7 ! intending 19 . . . g4, 19 . . . tllb4 or
19 ....td7) 1 6. . . tllf4 17 tllxf4 .txf4 (17 ... gxf4! ?)
1 8 .tg3 .txg3 19 hxg3 ( 19 fxg3 ! ?) 19 . . . ii'g7
with pressure on d4 after, e.g., . . . g4.
15 l:.xf3!
••.
Alas, we now enter into high theory again.

I'll reduce things to an outline, with few details.
The standard exchange sacrifice on f3 is hardly
surprising in the French, of course, but it's hard
20 .I:i.adl
A simple but beautiful idea is 20 'it>xh2?
tllxd4 ! 21 "fiixc7 (21 °fiid l tllxg6 22 'it>xh3 °fiid7+
23 'it>h2 tllf4 !) 2 1 . . .tllxf3+ 22 'it>hl il.. g2#.
20 il..xg3?
•••
Much better is 20 . . . tllxd4 21 l:.xd4 (a differ

ent version of the last note is 21 "fiix c7? il.. g2+
22 'it>xh2 tllxf3#) 2 1 . . .°fiixc2 22 il..xc2 exd4 23
'it>xh2 d3 24 ii.bl g5 .
21 il..xg3?
He should play 21 fxg3 ! .
21...il.. g2+ 22 'it>h2 il..xf3 2 3 l:.d2 e 4 (D)
(instead of tlld2-f3, White plays tllgf3 directly)

was worth something, and that he might have a
clearer path for his queen to the kingside than
with having two knights to jump over (on e2
and f3).
Still, in some respects Black calls the shots.
As long as Black doesn't commit to a radical
move right away, White needs a positive plan.
He can't play tllb3 due to . . . c4, and dxc5 gives
up the centre. That means that a slow move by
Black at this point could be the most effective
course, as in the following game.
24 il..x e4!? Zhang Pengxiang M. Gurevich

-
And here another brilliant line is 24 °fiib3 Hoogeveen 2004

°fiid7 ! 25 il..xf4 tlla5 ! ! 26 °fiib4 "fiig4 27 il..g3
°fiixg6 28 °fife? tllc4 ! 29 l:.c2 tlld6 ! 30 'it>gl (30 7 g6!?
•••
°fiixd6 °fiih 5+) 30 . . . tllf5 ! 3 1 °fiixb7 °fiih5 ! 32 Black has the strange-looking idea of play
"fiixa8+ 'it>h7, as given by McDonald. ing ... il.. g7 and ... f6, breaking up White's cen
24 dxe4 2S l:.xe4 il..xe4 26 "fiixe4 l:.f8 27 dS
••• tre. It's hard to stop !
"fifes 0-1 At this juncture an especially noteworthy al-
ternative is 7 . . . il..e7, preparing the now-routine
The Fashionable 7 tllgf3 Variation attack with ... g5 . Then the critical continuation
is 8 0-0 and Black has two main tries:
1 e4 e6 2 d4 dS 3 tlld2 tllf6 a) 8 ... g5 9 dxc5 ! (D) has done fantastically
There are two ways to transpose into the well and is an instructive positional device for
main line here: 3 . . . c5 4 tllgf3 tllf6 5 e5 tllfd7 6 White.
c3 tllc6 7 il..d3 and 3 . . . il..e7 4 tllgf3 tllf6 5 e5 By surrendering the centre he gives his king's
tllfd7 6 il..d3 c5 7 c3 tllc6. The latter is a varia knight the d4-square, which ruins Black's at
tion that could also arrive by the game's order tacking plan, and he can blast away at the sud
but with 7 ... il..e 7. There is a lot of theory on that denly poor-looking g-pawn by f4; for example,
position. 9... tlldxe5 (9 ... g4 10 tlld4 tlldxe5 1 1 il..b5 il..d7
4 es tllfd7 S il..d3 cS 6 c3 tllc6 7 tllgf3 (D) 1 2 tll2b3 gives White the advantage, Smirin
Developing the knight in this way was al Akobian, Minneapolis 2005) 10 tllxe5 tllxe5
ways regardedi:as second-best, because now the 1 1 tllb3 il..d7 1 2 f4 !, Hracek-Stellwagen, Solin
d2-knight has nowhere good to go. Then play gen 2005. White is ripping open the f-file and
ers began to feel that the tempo White 'gained' Black's king will have trouble getting away.
FRENCH DEFENCE 289
b) 8 ... a5 is a sort of prophylactic move, dis 8 h4!

couraging tllb3 at any point due to ... a4. One The critical move. 8 0-0 .tg7 9 .I:te 1 0-0 10
game went 9 l:.el (9 dxc5 tlldxe5 doesn't make tllf l has been played a lot but Black comes oul
much sense without the g-pawn as a target) well after I O ... cxd4 1 1 cxd4 °iVb6! (White con
9. . . cxd4 l O cxd4 g5 ! ? (D) (now that there's no trols the critical squares after l l ...f6 1 2 exf6
dxc5, Black can go ahead). tllxf6 1 3 .tb5 !) 1 2 .tc2 f6 1 3 exf6 tllxf6 14
.ta4 tlle4 1 5 .txc6 bxc6 16 tllg3 e5 ! (if Black
can play this in the French he's usually in good
shape) 1 7 .te3 exd4 1 8 .txd4 ( 1 8 tllxd4 tllxf2
1 9 tllxc6 "iVxc6 20 .txf2 l:.xf2 21 @xf2 •cs+
22 l:.e3 .ta6 is winning for Black) l 8 ... .txd4
1 9 'iVxd4 tllg5 ! , with clearly the better position
for Black.
8...h6!?
A strange-looking move, yet consistent with
the waiting policy. Now 9 0-0 looks inconsis
tent with h4, so White must try to make some
thing happen. 8 ....te7 9 'it>fl ! ? (D) is another
odd idea, but White wants to 'castle' while
leaving the rook on h l !
1 1 g4 ! ? (this radically prevents . . . g4; unfor

tunately, it exposes White's king; so does 1 1
h3 ! h5 1 2 tllf l g4 1 3 hxg4 hxg4 1 4 tll3 h2, but at
least then White wins the g-pawn! Black has to
break up White's centre while he still can:
14 . . . tllxd4 ! 1 5 'iVxg4 .tc5, and we reach an
other position that is hard to assess; it looks
about equal) l l . . .h5 12 h3 hxg4 1 3 hxg4 iVb6
14 iVa4 (Gormally-McDonald, London 200 1 )
and McDonald suggests 14 . . .tllf8 ( 1 4. . .f6! ? 1 5
tllf l !) 1 5 tllf l .td7, when White has t o respond
to the discovered threat: 1 6 .te3 ( 16 .tbs tllxe5 !
- remember that tactic; it's seen all over the
place) 1 6...iVxb2 1 7 l:.abl tllb4 . Black has some
advantage. For example, 9 . . . 0-0 (if 9 . . .iVb6, then 10
Let's return to 7 ... g6 (D): 'it>gl ! cxd4 1 1 cxd4 tllxd4 12 tllxd4 °iVxd4 13
tllf3 is a typical gambit in this line; Black has a

lot of weak squares) 10 'it>gl f6 1 1 exf6 .txf6
1 2 tllg5 ! ? .txg5? ( 1 2. . .ii'e7 has the idea of
playing ... e5) 1 3 hxg5 ii'e7 14 tllf l ! e5 1 5 tlle3
ii'f7 16 tllg4 'it>g7 17 dxc5 tllxc5 1 8 tllf6 and
White had a winning game in Sebag-V.Popov,
Cappelle la Grande 2006.
9 0-0
Another game with typical themes contin
ued 9 a3 ! ? ii'b6 10 0-0 g5 ! 1 1 hxg5 hxg5 1 2
tllxg5 cxd4 1 3 cxd4 ii'xd4! 1 4 tlldf3 ii'g4 1 5
tllxf7 ! l:.g8 ! 1 6 tll7g5 tlldxe5 17 .te2 tllxf3+
1 8 .txf3 ii'h4, Perunovic-E.Berg, Gothenburg
Echt 2005 ; here 1 9 .th5+ '>t>d8 20 \i'f3 is best,
with mutual chances after 20... .te7. White's somewhat better pieces. The opening
9... g5 has come out about even, as the game continua
You'd think that White had gained a tempo, tion shows.
but now his king is committed and ... g4 is a real 16...0-0 17 tllxd4 ii'f6!? 18 l:.cl l:.d8 19
threat. ii'd2 tllf 8!?
10 h5! ? Black is attempting to combine kingside de
After 10 hxg5 hxg5 1 1 g 4 .te7 1 2 l:.e l , fence with challenging d4 via ... tlle6. Instead,
Black might try 1 2... tllf8 ! ? with the idea ... tllg6 1 9 ... tllb6 ! is solid, protecting the d-pawn.
and ... tllf4 or ... tllh4. The position is closed 20 tllf3! .tf4 21 .txf4 gxf4 22 ii'b4
enough to justify these elaborate knight ma 22 .te2 ! ? .tg4.
noeuvres, and White can do the same by 1 3 22 .tg4 23 tlle5 .txh5 24 ii'xb7 f3 25 tllc6
•..
tllf l ! cxd4 14 cxd4 .td7 1 5 tllg3 ! (D). fxg2! 26 l:.e3! ?

Or 26 tllxd8 ii'h4 (26 . . . l:.xd8 27 l:.e3) 27
'it>xg2 ii'g4+ 28 'it>fl ii'f3 (28 . . . ii'h3+ 29 'it>g l
l:.xd8? 30 .tfl ) 29 'it>g l \i'g4+ with a draw.
26 l:.e8 27 tlle7+ 'it>h8 28 ii'xd5 (D)
...
I'll leave you to contemplate this picture.

10 ... ii'b6 11 c4!? cxd4 12 cxd5 exd5 13
tllb3 .tg7 14 tllfxd4!
A pseudo-sacrifice. White has to move rap
idly or his centre will fall as Black's king scur 28....tg4?!
ries to safety. White now threatens 1 5 tllf5 . 28 ... tlle6 29 ii'xh5 l:.xe7 30 l:.f3 \i'g5 3 1
14 ... tllxd4 15 .te3 .txe5 16 l:.el! (D) ii'xg5 hxg5 32 'it>xg2 tllf4+ 3 3 'it>g3 l:.d8 is
White (assisted by .tfl if needed) recovers roughly equal.
the piece with a strange position. Black's extra 29 l:.g3 l:.ad8 30 l:.xg4! l:.xe7 31 ii'xd8
pawn and some weakness are pitted against l:.el + 32 l:.xel ii'xd8 33 .tc4
FRENCH DEFENCE 291
Now it looks as though White stands better disarray of pieces and threats, with White try
but Black works his way out. ing to defend an exposed king against Black' s
33 lllg6 34 .ltxf7!? llle5 35 .:gs+ 'i¥'xg8 36
••• open lines and advanced centre. O f course, there
.ltxg8 lllf3+ 37 Wxg2 lllxel + 38 Wfl llld3 39 are two possible outcomes in White' s favour.
.ltb3 Wg7 40 .ltc2 lllxb2 41 Wg2 Wf6 42 .ltb3 Either the sacrifice isn 't possible, when White
lf2.lf2 will almost always enjoy a large, cramping
A fantastic back-and-forth battle ! space advantage and potential attacks on both
wings. Or Black' s sacrifice may prove insuffi
Seizing Space by 5 f4 cient for equality. Ensuring such a result takes a
lot of accuracy on White's part, however, and
1 e4 e6 2 d4 d5 3 llld2 lllf6 4 e5 lllfd7 5 f4 (D) many players seem to have grown tired of be
ing on the receiving end of brilliancies.
The normal and logical response to 5 f4 is to
attack the d-pawn with 5 ... c5. As a mini-rule.
we can generalize that attacking the front of a
double-winged pawn-chain like this with .. .f6 i s
best delayed until some of your other pieces
are out. You might compare the King's Indian
c4/d5/e4 double-wing, in which . . . c6 can be
very useful, but doesn ' t usually occur until
Black has castled. Nevertheless, you will see
that eventually ... f6 will be essential to open
counterattacking lines.
5 c5 6 c3
•••
I should mention that 5 c3 c5 6 f4 is another

move-order that reaches this posi tion.
White constructs a big centre, with the pawns 6 ... lllc6 7 llldf3
forming a wedge that extends into Black's posi This is played in the great majority of games.
tion. The advantage is obvious: it's now ex Otherwise the knight on d2 is only getting in
tremely hard to break down the front of the the way.
pawn-chain, which is always the essence of 7 'i¥'b6
.•.
Black's strategy in the .ltd3 lines. The moves Here Black has the option of a 'closed' sys
... f6 and .. .fxe5 can be answered by fxe5 (or in tem with 7 ... cxd4 8 cxd4 f5 (D) that is quite
fewer cases by dxe5, which cedes the c5-square playable.
to Black's pieces). I'll concentrate on this
d4/e5/f4 pawn-chain structure here and in the
Classical French because it's different from
most other pawn-chains in this book. The only
other major opening with similar properties is
the King's Indian Defence with pawns on e4,
d5 and c4, and Black's treatment in that open
ing is radically different from what we shall see
in the French Defence.
White's strategy has one major drawback: he
has to make so many pawn moves, not only
these first four but also c3 and usually g3 and/or
h4. Even a3 and b4 are part of a typical forma
tion. Because of this it turns out that Black can
almost inevitably sacrifice something in the
centre to open up attacking lines for his better The idea is . . . .lte7, ... 0-0 (this may be de-
developed pieces. The result is often a confused layed), ... lllb6, . . . a5-a4(-a3), ... .ltd7 and attack
on the queenside. White can never fully neu play 1 7 ... 'i¥'xb2 1 8 .lte2 'i¥'c3+ 1 9 'i¥'d3 I:!.xf3+ !
tralize this attack if Black is careful. His prob - somehow Black always has tactics in these
lem is that White plays for I:!.gl and g4 and lines - 20 Wxf3 'i¥'xd3+ 21 .ltxd3 .ltxd4 22
Black's king must be defended. The manoeuvre I:!.ae l , and probably White still has a tenuous
. . ..ltd7-e8-g6 can be useful in that respect. It's advantage) I S ....ltcS (or I S ... lt'ixeS ! ?) 16 ctJc2
an interesting system for positional players. lt'ixeS with a strong attack.
One example after 9 .ltd3, by transposition: 13...lt'ixd4! (D)
9 ... .lte7 ! I 0 lt'ie2 lt'ib6 1 1 h3 0-0 12 g4? ! ( 1 2 a3 !
a5 1 3 b3 a4 14 b4 lt'ia7 I S I:!.gl .ltd7 and now 1 6
lt'ic3 'i¥'e8 ! ? 1 7 g4 ctJbS or 1 6 g4 .ltbS) 1 2 . . .aS!
13 a4?! lt'ib4 1 4 .ltbl .ltd7 IS Wf2?! I:!.c8 16
I:!.g I Wh8 17 Wg2 .lte8 ! 18 Wh2? (I 8 Wh l ! ?)
18 ... .ltg6 19 lt'ic3 'i¥'e8 20 lt'iel 'i¥'f7 21 I:!.g2
fxg4 22 .ltxg6 'i¥'xg6 23 hxg4 'i¥'h6+ 24 Wgl
gS! and Black stood much better in Ye Jiang
chuan-Short, Lucerne Wcht 1989.
We now return to 7 ... 'i¥'b6 (D):
A theme that occurs again and again; you

need to know it whether you're playing White
or Black.
14 lt'ie2! lt'ixeS
This is good, but a creative suggestion by
Kalinichenko is more fun: 14 ... lt'ixf3 ! I S .ltxb6
ctJfxeS+ 1 6 Wg3 lt'ixb6. This looks overwhelm
ing. Black has only two pieces for the queen but
his minor pieces will slaughter White.
15 .ltxd4 lt'ig4+ 16 Wg3?!
We' ll look at three games with characteristic Better is 16 Wg I .Iles 17 .ltxh7+ ! ? Wh8 ! 1 8
tactical motifs that you should know. .ltxcS 'iY'xcS+ 1 9 lt'ied4 eS, but it's obviously
good for Black.
Saltaev - M . Gurevich 16 'i¥'d6+ 17 Wxg4 eS+ 18 Wg3 exd4+ 19
•..
Cappelle la Grande 2001 Wf2 .ltg4 20 I:!.cl (D)
8 h4 cxd4 9 cxd4 .ltb4+ 10 Wf2 f6 11 .lte3

fxeS! ?
Or l I ... 0-0, denying White squares such as
f4.
12 fxeS 0-0 13 .ltd3?
This move tends to interfere with White's
control over d4 and in general exposes the
bishop to later attack by Black's centre. 1 3 a3
.lte7 14 b4 is one course, and 1 3 lt'ie2 ! tries to
take advantage of Black's I I th: 1 3 . . . lt'idxeS ! ?
(McDonald's recommendation) 1 4 dxeS d4 l S
lt'iexd4 ( 1 S .ltxd4 ! lt'ixd4 1 6 lt'iexd4 .Iles 1 7
We3 looks solid enough; Black should probably
FRENCH DEFENCE 293
20....l:.ae8!? Krupkova - Gleizerov

20 . . . 'i:Ve5 ! is decisive, since ... 'i:Ve3+ can't be Mariehamn!Osteraker 1 997
stopped except by 2 1 ilxh7+ Wxh7 22 ll:iexd4
.l:.ae8, etc. 8 g3 cxd4 9 cxd4 ile7 10 ilh3
21 ll:iexd4? ! White follows a traditional plan in which he
2 1 'i:Vb3 .ltd2 ! . tries to force Black to defend his e-pawn. That's
21...ilaS! simply too slow, so Black will be forced to sac
Threatening . . . .ltb6. rifice a piece instead:
22 b4 'i:Vxb4 23 Wg3 hS 24 .l:.bl ilc7+ 0-1 10 ... 0-0 11 ll:ie2 r6! 12 .:n?
White has to be consistent and take the
Gufeld - H ummel pawns: 12 ilxe6+ ! Wh8 1 3 ilxd5 fxe5 14 fxe5
Las Vegas 2000 ll:idxe5 ! 15 dxe5 ll:ixe5 and Black has a power
ful attack, but with White a piece up it's hardly
8 g3 cxd4 9 cxd4 .ltb4+ 10 @r2 gS!? clear.
This is a theoretical line that should ulti 12 ...Wh8 13 ll:ic3!? rxeS 14 rxeS (D)
mately be equal. Neither of the games I'm giv
ing is best play, but show how each colour can
quickly get into trouble.
11 ile3?!
A win for White with a great finish went 1 1
fxg5 ll:idxe5 1 2 ll:ixe5 ll:ixe5 1 3 ile3 ll:ic6 14
ll:if3 ilf8 1 5 'i:Vd2 ilg7 1 6 ild3 ! ild7 1 7 .l:.ac l
ll:ixd4 1 8 'i:Vc3 e5 1 9 ilxd4 exd4 20 .l:.he1 + Wd8
2 1 'i:Va3? ! ilf8 22 ll:ie5 ! ile8 23 'i:Va4 ! h6?? 24
'i:Vd7+! 1 -0 Solak-Kozamemik, Ljubljana 2003
(24 ... ilxd7 25 ll:ixf7#).
1 l...g4! 12 ll:id2 r6! (D)
14....l:.xf3!
A sacrifice that's almost as old as the French
Defence itself. It's a little more difficult in this
situation to summon up the courage to do it. be
cause Black has to reorganize before he can
bring all his pieces into the attack. Generally.
however, . . . .i:.xf3 should become your first in
stinct as Black in these f4 positions, and Public
Enemy Number One for White !
15 'i:Vxr3 ll:ixd4 16 'i:VhS! 'i:Vd8!
White's back-rank threats are prevented and
Black's pieces get out to aggressive positions.
13 'i:Vxg4? 17 'i:Vdl?
White should have played 1 3 ll:ib3 ! fxe5 1 4 1 7 @d 1 ! is the best try, even if it is no fun to
dxe5 ilc5 1 5 ll:ixc5 ll:ixc5, which i s equal or defend: 1 7 . . . ll:ic6 ! 1 8 ilf4 ( 1 8 ilxe6? ! ll:idxe5
perhaps slightly better for Black. and Black is well on top) 1 8 . . . ll:ic5 intending
13...ilxd2 14 ilxd2 'i:Vxd4+ 15 @el 'i:Ve4+ . . . ild7. With the king on d 1 , these positions are
16 @r2 ll:idxe5! 17 'i:Vg7 .l:.f8! 18 ile2 'i:Vxhl 19 awfully hard to play for White. Black even has
ll:ir3 ll:ig4+! 20 'i:Vxg4 'i:Vxal 21 rs exrs 22 'i:Vr4 a pawn for the exchange.
ild7 23 'i:Vd6 'i:Vxb2 24 .tr4 .l:.r7 25 'i:VxdS @rs 17 ... ll:ic6 18 ilxe6 d4! (D)
0-1 19 ll:idS
Classica l Variation
1 e4 e6 2 d4 dS 3 tllc3 tllf6 (D)
Or 1 9 tlle2 tlldxe5 20 i..x c8 .l:.xc8 with moves

such as ... d3, . . . tllc4 and . . . tllg4 to come. Black
is practically winning already.
19 ... tlldxeS 20 i..xc8?
A blunder. But 20 tllxe7 i..xe6 21 tllxc6
tllxc6 is terrific for Black because of his tre The Classical lines of the French begin here.
mendous unopposed bishop and White's king To continue our discussion of the d4/e5/f4 cen
position. tre, we're going to examine the main line with �
20...'iY'xdS (D) that set-up.
4 e5
I won't be discussing the important alterna
tive 4 .ltg5, when the Maccutcheon, 4 . . . .ltb4,
can resemble the Winawer Variation.
4...tllfd7 (D)
Everything is falling apart for White.

21 i..xb7
Or 2 1 i..f5 .i:.f8 ! , which threatens ....l:.xf5 and
there's no defence.
21 ...i..b4+
2 1 .. .d3 ! 22 i..xc6 'i¥'xc6 23 i..e 3 'i¥'e4 wins 5 tllce2
for Black. This odd-looking move is designed to avoid a
22 Wf2 number of Black's options. For example, White
Now 22 ...'i¥'f7+ 23 Wgl 'i¥'xb7 wins. could play 5 f4 c5 6 tllf3 tll c6 7 tlle2 (7 .lte3 is
After the inconsistent 1 2 .l:.fl ? the opening one of the main lines of the Classical French,
was a disaster for White, but the objective as not covered in this book) 7 . . .'i¥'b6 8 c3, trans
sessment of the g3/.lth3 manoeuvre is anyone's posing to the variation that we are examining.
guess. But Black would have the choice of capturing
FRENCH DEFENCE 295
the pawn on d4 on moves 6 and 7, or of playing

a move other than ... 'iVb6 on move 7.
S...cS 6 c3
6 f4 leads to its own move-order deviations
like 6 ... ii.e7 7 lllf3 0-0 8 c3 f6 ! ? or 6 ... 'i¥'b6 7
lllf3 ii.e7; or even 6 ... 'i¥'a5 + ! ? 7 c3 b5 ! ?. All of
these deserve more trials, as they are seldom
seen in master play.
6 lllc6 7 f4 (D)
•••
... a4 in and let him try to infiltrate on the queen

side; even if he gets a piece to b3 it doesn't look
as though Black would get anything useful out
of it. In the meantime, that' s a big and danger
ous pawn-mass that White would have at his
disposal on the kingside.
12 'i¥'c7 13 lllegl! ? (D)
.•.
These pawn-chain lines look like the Tar

rasch lines with 5 f4 and sometimes transpose
into them, but in some ways White has a better
grip on the centre. For example, there are no
lines in which ... cxd4 followed by ... ii.b4+ both
ers him. We'll look at two games from this posi
tion, one in which White tries to maintain his
entire pawn-chain and another in which White
plants a piece on d4 and establishes himself
there:
Anand Shirov-
Teheran FIDE Wch (4) 2000 A surprising and clever retreat: White un
develops his pieces only in order to hold the
7 'i¥'b6 8 lllf3 f6 (D)
•.• centre together and anticipate all of Black's
The lines are formed for a classic battle: threats. On the other hand, although Anand's
White wants to batten down the hatches, avoid last few moves are ingenious and were praised
making any weaknesses in his own position, se by one and all, they're also slow. That's Black's
cure and increase his space advantage, and fi cue to throw everything he's got at the white
nally, drive back Black's pieces. For his part, centre.
Black wants to blast open the centre, sacrifi 13 a4!
•.•
cially if necessary. This loosens things up a bit before launching

9 a3 ii.e7 10 h4 0-0 1 1 .l:.h3! as 12 b3 into the complications. You'll see later how
All these pawn moves can be a little slow. useful this interpolation is. Anand recommends
White's got a lot of space on the kingside and 1 3 . . . b6, but is exchanging bishops useful? See
might want to use a move to secure it. With that my comments in the next game.
in mind, he could simply allow Black to get 14 b4 fxeS 15 fxeS llldxeS! (D)
so I'll just take it easy and retreat by 1 8 . . . 'iVc7

(D). Then I'll play . . . e5 (hitting h3) and . . .i.. f6."
That' s probably a good plan:
It's now or never. That's about all you need

to know about these lines when playing Black:
if you don't sacrifice at an early stage, you'll
probably never be able to sacrifice later! White
will just have too many pieces covering all the If you want to see Black's reward for sacri
key squares and then you'll die slowly, waiting ficing his piece, try to defend the diagrammed
around as he slowly advances on your cramped position for White. It may or may not be that he
position. can succeed in repulsing the attack, but few
If you're handling the white pieces, the sac players could do so in practice. One line would
rifices are also about all that you need to know ! be 19 i..g5 ! ( 19 'iVh5, to get out of the way of
If you can prevent those, the rest won't be diffi the bishop and play i..d 3 next, can be answered
cult. So try to set up your pieces for maximum by 1 9 . . . cxb4 20 axb4 g6 21 'iVh6 e5; for instance,
post-sacrifice defence, as Anand has tried to do 22 l:.g3 l:.xfl + ! 23 Wxfl 'iVc4+ 24 lt'ie2 .ltg4 !
here by playing l:.h3 and lt'ieg l , both designed 25 l:.xg4 'iVxg4 with a big attack) l 9 ... .ltxg5 20
to overprotect the f3-square, which is generally hxg5 e5 21 l:.f3 .ltf5 22 'iVf2 ! ? cxb4 23 cxb4 e4
the most vulnerable target. It pays off for him in 24 'iVg3 l:.ac8 ! 25 'iVxc7 l:.xc7 26 l:.f2 d4 and
this game. the pawns and open files make life pretty tough
16 dxeS lt'ixeS 17 lt'ixeS 'iVxeS+ (D) for White. Objectively it's hard to assess this as
favourable for either side.
19 Wdl (D)
Not 1 9 l:.xh4? 'iVg3+.
18 'iVe2 i..xh4+!?
Shirov pours more gasoline on the fire. He
could also say to himself, "I've got two mobile
centre pawns and tremendously active pieces, 19 'iVf6?!
•.•
FRENCH DEFENCE 29 7
After this White gets his pieces out too fast. 10 lt'ig3 was suggested, although then 1 O .. .f6
There was nothing wrong with 19 . . . 'i¥'xe2+! 20 keeps the pressure up. The move 10 b3 !?, as in
�xe2 (D). the Anand game, also looks slow because
Black's queen doesn't have to go to b6. On the
other hand, a rook on the second rank is one of
the best defensive pieces in almost any position
(don't forget that !), and one on a2 might come
in very handy later.
10 f6 1 1 lt'iegl? !
.•.
White plays like Anand in the Shirov game.

but without Black's queen on b6. Probably 1 1
.l:.h3 ! is best. After that move Anand's idea of
. . . b6 and . . .�a6 has been suggested. The prob
lem I have with this positional device is that if
White's good bishop is exchanged for Black ' s
bad one, all that does is give White a few pre
cious extra tempi to defend his massive centre.
Then he can begin an advance with his f-pawn
Don't forget how valuable centre pawns are ! that will free his other bishop. It seems much
It's instructive how they remain so after simpli better to pursue the usual sacrificial ideas in the
fication. Of course White has his chances too. centre.
A sample line would be 20 . . . �f2 (20 ... �f6 ! 1 1 cxd4 12 cxd4 'i¥'b6 13 �d3?! fxeS 14
•••
also has some good points; e.g., 2 1 bxc5 .l:.a5 22 fxeS lt'idxeS! (D)
�e3 �xc3 23 .l:.bl d4 or 23 ... e5) 21 �e3 ! e5 22
�xf2 .l:.xf2 23 .l:.g3 b6 24 Wel .l:.f6 25 .l:.e3 .l:.e6
and it's not clear who's better. Maybe the whole
ending is about equal. At least it's not boring !
20 ctJf3!
Finally White's pieces are active. Now Anand
isn't worried about the centre any more. The
rest is pretty easy.
20 'i¥'xc3!? 21 �b2 'i¥'b3+ 22 @cl es
.••
22 ... �f6 23 �xf6 .l:.xf6 24 lt'ie5 and the

queen is trapped.
23 .:Xh4 �rs 24 'iY'dl e4 2S 'i¥'xb3 axb3 26
lt'id2 e3 27 ctJf3 .l:.ae8 28 Wdl c4 29 �e2 �e4
30 Wcl .l:.e6 31 �c3
White is two pieces ahead for the blockaded
pawns. Anand went on to win easily. There it is again! You can see why these posi
tions are so difficult for White to defend, regard
Macieja - lvanchuk less of whether he's in satisfactory condition
Moscow FIDE KO 2001 according to theory.
lS dxeS lt'ixeS 16 �c2
7 �e7 8 ctJf3 0-0 9 a3 as
•.• l 6 lt'ixe5?? 'i¥'f2#.
Once again Black is not thrilled with allow 16 �d7!
•..
ing b4, although White needs to spend extra When you've got this kind of attack and
time doing so, and maybe 9 . . . f6 is good; for ex there are no immediate sacrifices, you can al
ample, 10 b4 ! ? cxd4 1 1 cxd4 fxe5 1 2 fxe5 ctJb6 ways bring up the reserves. The centre is your
would be an interesting positional solution. long-term compensation. Besides, . . . �b5 could
Even the bad c8-bishop would get out. be strong at some point. If you're White, the
10 h4 best thing to do is to try to simplify, and if that's
impossible, obscure the issue as much as you 22 . . . 'i¥'xe5 23 'i¥'xh7+ Wf8 24 'i¥'xh4 .l:.exc2 and
can. Black wins.
17 'i¥'e2 (D) c) In response to 19 Wd l there is a simple
but hard-to-see piece of geometry: 19 ... .lta4! !
20 .ltxa4 'i¥'d4+ 2 1 llld 3 'i¥'xa4+, mating in a
few moves.
18 ...Wxh7 19 'iY'xeS .ltd6 20 .lte3 'iVb3 21
llld2 (D)
2 1 'i¥'xd6 'i¥'xe3+ 22 llle2 .l:.c2 wins at once.
Now Ivanchuk finishes it off prettily:
17....l:.ac8!!
Now every piece is in the act. Of course
Ivanchuk has a few of them hanging, but he's
got it all worked out. Instead, l 7 . . .lllxf3+ ?! 1 8
lllxf3 .ltb5 looks attractive except for 1 9 .lte3 ! .
1 8 .ltxh7+!
Tough defence ! The variations after 1 8 lllxe5
.ltxh4+ ! (D) are fantastic: 21....l:.fl+! 22 Wxfl 'i¥'d3+ 23 Wf2 .ltxeS 24
lllgf3 .ltxb2 25 .l:.abl .l:.c2 26 .i:.hdl eS 27 g3
.ltg4 0-1
Winawer Va riation
1 e4 e6 2 d4 dS 3 lllc3 .ltb4 (D)
a) 19 .l:.xh4 .l:.xc2! (Black must avoid the

tempting 19 ... 'iY'xg 1 +? 20 Wd2 .l:.f2 because 2 1
.ltxh7+ Wf8 2 2 .l:.f4+ ! .l:.xf4 23 lllg6+ We8 24
lllxf4 refutes the attack; White's got a lot of
pieces, and sometimes they simply do the job)
20 'i¥'xc2 (20 'i¥'e3 .l:.xc l + 21 'iY'xc l 'iVxgl +)
20 ... 'iY'xgl + 21 Wd2 .l:.f2+ 22 c.t>c3 d4+ 23 Wb3
a4+ ! and the queen goes. This is the Winawer Variation. Black pins
b) 19 Wd2 'i¥'d4+ 20 'i¥'d3 .i:.f2+ 21 llle2 the c3-knight and, in the same way that he does
.:Xe2+ ! 22 Wdl (22 Wxe2 .ltb5 23 'i¥'xb5 .:Xc2+) when he plays 2 ... d5, puts the question to White:
FRENCH DEFENCE 299
exchange, gambit, protect, or advance? We are 5 .ltxc3+

•••
going to concentrate upon the main line, which Black cedes the bishop-pair to White in or
is marked by the advance. der to gain a tempo and inflict doubled pawns
4 e5 on his opponent. 5 . . ..lta5 is a respectable option
Easily the most ambitious move, restricting played by some specialists, but isn't nearly as
Black's development and staking out territory popular; we'll pass that by.
on the side of the board that Black's dark 6 bxc3 llle7 (D)
squared bishop has just abandoned. 6 . . .'iVc7 is also played, posing a different set
Various other lines may be found in the of problems. Those who enjoy play upon col
books, several of them involving the sacrifice our complexes may be attracted to lines such as
of White's e-pawn with subsequent recovery. 7 'iVg4 f5 8 'iVg3 cxd4 9 cxd4 llle7 1 0 .ltd2 0-0
For example: 1 1 .ltd3 b6 1 2 llle 2 .lta6, when White's con
a) 4 a3 .ltxc3+ 5 bxc3 dxe4 6 'iVg4 lllf6 7 centration upon dark squares (h4-h5-h6, lll f4-
'iVxg7 l:tg8. h5 and .ltb4 are typical ideas) contrasts with
b) 4 .ltd2 dxe4 5 'iVg4 lllf6 6 'iVxg7 l:tg8. Black's on the light squares (by . . ..l:.c8 and
c) 4 llle2 dxe4 5 a3 .Ile7 6 lll xe4. ... lllb8-c6-a5, for example).
And so forth. There are numerous options on It is my belief that the most instructive and
every move of these lines, with theory tending engrossing lines follow from the positions after
towards a verdict of equality with best play. As 6 ... llle7.
always, the reader may want to consult special
ized books to learn more.
4 cS (D)
...
Black decides to attack the base of the pawn

chain first. He will almost inevitably attack the
front of it later.
At this j uncture, White chooses between the

Positional variations, involving the moves lllf 3,
a4 and/or h4 in various orders, and the 'French
Poisoned Pawn Variation' 7 'iVg4.
Positional Variations
5 a3
White wants to force a decision by Black's In this section we'll look at lines in which
bishop; you'll have to check theoretical works White bypasses tactics for a while and tries to
and databases in order to find out about the alter establish a positional edge. In spite of initial ap
natives. Of these, 5 dxc5 and 5 .ltd2 are perhaps pearances, both players will use both sides of
the most interesting. If you are not inclined to the board to generate play. We' ll look at a series
play the main lines presented below, this may of games beginning with 7 h4 and 7 lllf 3.
be a good place to investigate potential weap 7 h4 (D)
ons for use. The defender, of course, should be With this move White charges forward to as
aware of and prepared for White's various 4th sault Black's position, not caring about piece
and 5th-move alternatives. development. He has several ideas, beginning
and tactical issues and I shall discuss as many

as possible in context.
7 tllbc6 8 hS 'iY'aS
.•.
Black attacks White's c-pawn and, inciden

tally, threatens . . . tllxd4.
9 ild2 (D)
9 'i¥'d2? cxd4 1 0 cxd4 'i¥'xd2+ and 1 1 . . . tllxd4
wins the d-pawn.
with h5-h6 to compromise Black' s kingside.

Black' s kingside dark squares are already weak
ened due to the loss of his f8-bishop, so if White
can establish holes on f6 and h6 it will not only
give him good squares for his pieces, but also
discourage Black from attacking on the king
side. White' s advantage in space is on the
kingside, and h4 only enhances that advan
tage.
Other matters of note, many of which apply There follow two games that stem from this
to the Winawer in general: strategically rich position.
a) White has the two bishops;
b) The pawn advance h4-h5 makes kingside Hector - Hillarp Persson
castling very difficult for Black, and almost York 1 999
compels . . . 0-0-0;
c) The rook's pawn advance doesn't block 9 ild7
...
off White's queen as lines with tllf3 do, so 'i¥'g4 Developing as quickly as possible; Black an
is always an issue; nounces his intention to castle queenside.
d) White has a potentially strong resource in 10 h6 gxh6 (D)
l:.h3-g3/f3 or l:.h4-g4/f4, the latter rook move
also introducing the possibility of dxc5 and
l:.hb4 for attack.
Black's main advantage is less subtle: a
growing lead in development. It's quite possi
ble that he'll have every piece except one of his
rooks in action when White still only has one
piece out! White also has weak doubled pawns
on his c-file, with the usual problem that if
Black exchanges pawns on d4 White gets rid of
his doubled pawns only to find that his remain
ing backward c-pawn on an open file can be as
least as much a problem as the doubled pawns.
In general, Black would like to exploit White's
queenside light-square weaknesses on c4 and
a4. Finally, Black can usually open files on the A funny position because Black's doubled
kingside, after which his rooks directly face h-pawns are so weak on an open file and White
White's king. There are many other positional apparently controls the kingside. But Black finds
FRENCH DEFENCE 301
a remarkable idea after which both players assessment which also applies to the opening
have the opportunity to play on both sides of in general.
the board! In fact that' s often the case in the 14 .l:!.fS! 15 'iVcl ! (D)
..•
Winawer. Although White has space on the

kingside, Black can counter with . . .f6 and open
lines for his pieces there. And Black's attack on
White ' s weak queenside squares can boomer
ang when White uses the b-file and dynamic
pawn moves on that side of the board including
c4, a4-a5, and sometimes dxc5.
11 Cllf3 0-0-0 12 i.d3 c4
Black closes the side of the board on which
he appears the strongest!
13 i.e2 Cllg8! ! (D)
15 f6 16 'i¥a3
•.•
This is White' s point: to activate his queen

on the precious dark squares, thereby freeing
his dark-squared bishop to help on the king
side. This would be positionally winning ex
cept that time is an element that can't be
discounted.
16 .l:!.f7 17 jg,f4 lllge7 18 exf6 .l:!.xf6 19
.•.
jg,xh6 .l:!.g8 20 @n
20 g3 Cllf5 2 1 i.d2 .l:!.fg6 transposes.
20 lllfS 21 jg,d2 .l:!.fg6 22 g3
•..
This retreat is Black's salvation, a move in Now White would like to secure his entire
vented by Uhlmann, the Hercules of the French position with jg,f4, but it's Black's move:
Defence. Instead of going to the obvious f5, the 22 eS! (D)
•.•
knight deters jg,xh6 (which would mean giving

up White's dark-squared bishop) and prepares
the key move ... f6. After that, Black's two files
on the kingside can cause trouble.
14 a4!?
Both sides' ideas begin to become clear af
ter 1 4 @fl f6 ! ? 1 5 'iVe l fxe5 1 6 lll xe5 ? ! ( 1 6
dxe5 ! .l:i.f8 1 7 g 3 'i¥c7 ! 1 8 jg,f4 Cll c e7 intending
. . . lllg6 is very complicated; when White main
tains a pawn on e5 he improves his defensive
prospects) 16 . . . Cllxe5 1 7 dxe5 Cll e7 ! 1 8 jg_xh6
.l:!.hg8 19 jLf3 ( 1 9 'iVd2 Cllf5 20 jg,g5 .l:!.xg5 21
'iVxg5 'iVxc3 ) 1 9 ... jg,e8 ! and Black has the su
perior game, Short-Psakhis, Isle of Man 1 999.
Black will play . . . jg,g6 and has a solid advan
tage. But these lines always have play for both 23 .l:!.xh7
sides and White eventually won the game. In A torrent of tactics follows 23 lllxe5 ! ? Cllxe5
view of White's improvement on his 1 6th move 24 dxe5 .l:!.xg3 ! ; for example, 25 'iVb4 'iVxb4 26
I think a verdict of dynamic equality is fair, an cxb4 .l:!.3g7 and it's hard for White to unravel
and counter . . .d4; e.g., 27 c3? Clld4 ! ! 28 cxd4 30 a5

jg,h3+ 29 We l .l:!.g l + 30 jg,fl .l:!.xh l . 30 'iVd6 .l:!.xc3+ 3 1 'iVxb6 axb6 is hopeless;
The other try is 23 dxe5, but 23 . . . .l:!.xg3 ! 24 Black will end up with four passed pawns.
fxg3 lllxg3+ 25 Wf2 'iVb6+ ! is also strong for 30 .l:!.g2+! 31 @xg2 'iVg6+ 0-1
•••
Black, a rook down, because his key move A typical game, of the kind that White can
. . . llle4+ will ruin any normal defence like 26 also win (and sometimes does) if he penetrates
Clld4 Cllxd4 27 cxd4 'iVxd4+ 28 'iVe3 ? (28 jg,e3 to the king before Black can drum up a sound
'iVxe5 with a huge attack and material to come) attack.
28 . . . llle4+.
23 e4! 24 llle5?
..• Hellers - Gulko
A resourceful try is 24 Cllh 4! Cllxh4 (not Biel IZ 1993
24... lllx g3+? 25 fxg3 .l:!.xg3 26 .l:!.xd7 ! Wxd7 27
jg,f4) 25 .l:!.xh4 Cllxd4 ! 26 cxd4 'iVxd2 and Black 9 cxd4 10 cxd4 'iVa4
..•
stands better but it's messy. Here we have an innocent-looking position in

24 lllxe5 25 dxe5 (D)
•.• which White had originally played 1 1 c3 or 1 1
jg,c3 with equality. Then, playing against Anand
in Linares 1 992, Kasparov found an ingenious
sacrifice to get his usual initiative for a pawn.
11 Cllf3! ? lllxd4 12 jg,d3 lllec6 13 Wfl (D)
25 e3!
.•.
Black maintains a constant initiative before

his king can get into trouble.
26 jg,xe3 lllxe3+ 27 fxe3 .l:!.xg3 28 Wf2
28 'iVe7 jg,c6 defends everything. White's idea is to use his two bishops on
28 'iVb6 29 .l:!.n .l:!.xe3! (D)
•.• newly-opened lines. In addition, his move h6
Accurate to the end. can potentially weaken Black's kingside dark
squares. Indeed, the game Kasparov-Anand
went 1 3 ... Cllxf3 ! ? 1 4 'iVxf3 b6? 1 5 h6 ! with a
powerful initiative. Later an instructive solu
tion was found that uses Black's pieces to maxi
mum efficiency:
13 lllf5 14 jLxf5 exf5 15 h6 .l:!.g8! 16 jg,g5!
•••
jg,e6 17 .l:!.h4!?
The obvious 17 hxg7 .l:!.xg7 1 8 jg,f6 runs into
1 8 ... 'iVc4+ 1 9 Wgl .l:!.g4 20 .l:!.xh7 Wd7 (D), when
Black has some advantage.
Every piece on light squares!
17 'iVa6+ 18 @gl gxh6 19 i.f6 .l:!.g4!
.•.
With opposite-coloured bishops, both sides

have rushed to exploit their respective strengths.
FRENCH DEFENCE 303
White is not short of opportunities for cre

ative play in these lines. We turn to 7 ctJf3:
Short - lvanchuk
Horgen 1995
7 ctJf3 (D)
20 .l:!.bl
Better, but still good for Black, is 20 .l:.xh6 ! ?
Wd7 2 1 .l:.xh7 .l:.ag8 22 .l:.h2.
20 .l:.xh4 21 .txh4?!
.•.
Perhaps 21 lllxh4 would improve.

21 .l:.cS 22 c4 �xc4?!
•••
Not a terrible move, but Black could play

22 ... dxc4 23 �d6 (23 ctJd4 ctJxe5 24 ctJb5 ctJd3
is even better for Black) 23 ... c3 24 ctJd4 �d3 ! White develops and guards the centre. Not
(D). surprisingly, this is his traditional move in the
Positional lines.
7 �c7 8 h4 .td7 9 h5 h6
..•
This time Black wants to hold the kingside

while he works on White's queenside weak
nesses.
10 .td3 .ta4 11 dxc5! (D)
Notice again the colour-complex motif for

both sides. Now Black simplifies into a win
ning position: 25 lllxe6 °iVxd6 26 ctJg7+ Wf8 27
exd6 Wxg7 28 d7 .l:.c7 ! 29 d8� lllxd8 30 .txd8
.l:.d7.
23 .l:.xb7 lllxe5 24 lllxe5 °iVxh4 25 .l:.xa7
�f4 26 llld3 �d4 27 .l:.a4 .l:.c4! 28 .l:.a8+ @e7 Tripling pawns may seem odd, especially
29 �1 .l:.a4? since White gives up protection of the e5-pawn
29...@f6 would still keep a considerable ad as well. Indeed, all of White's pawns will be
vantage. vulnerable, but Black can only take one at a
30 �b7+ Wf6 31 .l:.xa4 "iVxa4 32 °iVb2+ d4 time! In compensation, White gains d4 as a
33 �d2 @g7 112-1/2 transfer point for his pieces and the rooks can
spring into action along the 4th rank; for exam 20 ... 'iVxc3+! should have been tried, when
ple, by .l:!.h4 and .l:!.b l -b4. The bishop-pair can the trick 2 1 jg,d2 'iVxc5 22 .l:!.xb7 ! (with the idea
also become more effective with more room to jg,b4) can be answered by 22 . . . a5 .
manoeuvre in. Here and in other Winawer posi 2 1 @xdl .l:!.e8 22 .l:!.xf5 @g8 23 .l:!.g4 .l:!.e7 24
tions the move dxc5 is an important part of .l:!.g3 Wh7 25 jg,d4 .l:!.he8 26 .l:!.f4 tt:Jrs 27 .l:!.fg4
White's bag of tricks. (D)
11 llld7 12 .l:!.h4 'iVa5? !
•••
A mistake; Black's queen belongs o n c 7 in

these lines. Note that l 2 . . . lllx c5? is a blunder
due to 1 3 .l:!.xa4, but 1 2. . .jg,c6 ! is double-edged, B
when White has to attend to his pawns and a
complex battle will result.
13 jg,e3! .l:!.c8
Logically bringing another piece into play
along the open file, but 1 3 . . . jg, c6 was still cor
rect. Not 1 3 . . .'iVxc3+?! 14 Wfl and the queen is
in trouble.
14 .l:!.bl jg,c6 15 'iVd2 'iVxa3 16 .l:!.g4 Wf8 (D)
27...g5
Instead, occupying the natural blockading
square on e6 only lets White's knight in on the
ideal attacking square f5: 27 . . . llle6 28 lllh4
.l:!.g8 29 ctJf5 .l:!.d7 30 lt:Jd6 and the f-pawn rolls
forward.
28 hxg6+ lllxg6
28 ... fxg6 is met by 29 .l:!.f4 or 29 .l:!.h4.
29 .l:!.h3 jg,d7 30 lllg5+ @g8 31 e6! (D)
17 .l:!.bb4
Not a bad move, preparing to swing the
rook into action along the fifth rank. Perhaps
even better was Short's line 1 7 @fl ! ? lllx c5 1 8
jg,xc5 'iVxc5 1 9 ctJd4 with the idea 'iVf4 and a
strong attack. This illustrates White's use of
the tripled pawns to secure a support-point for
his knight.
17...lllf5
Practically a necessity in order to defend key
squares. The knight is Black's best piece, so
White will get rid of it. In the meantime, White
gets a situation with opposite-coloured bishops Finally this key breakthrough, opening the
that will aid his attack. fatal diagonal for White's unopposed bishop on
18 jg,xf5 exf5 19 .l:!.gf4! 'iVal+ 20 'iVdl ! d4. From now on White has a clearly winning
The advantage i s also clear i n a queenless advantage.
middlegame. 31...jg,xe6 32 lllxe6 .l:!.xe6 33 .l:!.xh6 .l:!.el+ 34
20...'iVxdl+?! @d2 .l:!.8e2+ 35 @d3 @rs 36 .l:!.h5 ! .l:!.e4
FRENCH DEFENCE 305
White wins the ending after 36 . . .llle5+ 37 The key move is .. .f6, which helps defensively
i.xe5 .l:.xe5 38 .l:.xe5 .l:!.xe5 39 .l:!.b4. but also gives him central threats and a very
37 .l:!.xe4 dxe4+ 38 @d2 .l:!.bl 39 c6 useful f-file. In any case, Black' s first decision
Or 39 g3 ! . is whether to gambit the pawn by 7 ... 'i¥c7, or
39 bxc6 40 i.xa7 @e7 4 1 i.d4 .l:!.b5 42
•.• play 7 . . . 0-0 and hang on to his material for a
.l:!.xb5 cxb5 43 c4! while. The latter is the preferred choice these
The finishing blow. Bishop vs knight with an days but it' s not clear that the former won't
extra passed pawn will win. come back into fashion.
43 bxc4 44 @c3 ctJf4 45 g3 llle6 46 @xc4
••• As in so many lines of the French Defence,
@d6 47 i.f6 @c6 48 g4 @d6 49 c3 @c6 50 one notices the persistence of the central pawn
i.e5 lllc5 51 @d4 lt:Jd3 52 i.g3 lllc5 53 i.f4 structure. This gives both the positional and
1-0 tactical themes a certain logical consistency, al
White manoeuvres his bishop to the centre: though it doesn't seem to limit their variety.
53 ... llld3 54 i.e3 lllc5 55 @e5 llla4 56 i.d4.
The Gambit
French Poisoned Pawn
7 'i¥c7 8 'iVxg7 .l:!.g8 9 'iVxh7
•••
1 e4 e6 2 d4 d5 3 lllc3 i.b4 4 e5 llle7 5 a3 Now in addition to the other advantages listed

i.xc3+ 6 bxc3 c5 7 'i¥g4 (D) above, White has a passed h-pawn. It is worth
remembering, however, that the advance of a
passed rook's pawn very seldom poses a threat
until far into the middlegarne. Their real strength
appears in simplified positions and, of course,
as outside passed pawns in an ending.
9 cxd4 10 llle2 (D)
•••
Other moves like 10 @dl are possible, yet

this is how they've played it for 50 years in the
vast majority of games.
The grand old flagship of the Winawer Vari

ation. White wants to exploit Black's lack of the
dark-squared bishop by direct means. He will
try to get Black to weaken himself or castle into
a potential attack. It makes sense to work on the
side of the board where he has space and to
wards which his bishops aim. In addition, the
e5-pawn cramps Black in that part of the board.
As is true in the positional lines, White would
love to get rid of the g-pawn so that his unop We' ll examine two of those encounters from
posed dark-squared bishop can have a field day this position, with general considerations dis
on squares like h6 and f6. cussed therein.
For his part Black's first goal is to attack
White's centre and queenside, where White al Karpov Agdestein
-
ready has serious weaknesses. Ironically, how Oslo 1984

ever, he usually ends up playing on the kingside
too, whether or not he castles in that direction. 10 lllbc6 1 1 f4
.•.
1 1 cxd4? lllxd4! (threatening c2) 1 2 ctJxd4? because if he waits too long for White to get
is bad in view of 1 2. . . 'iVc3+. his position organized (and especially to get
1 1...jg,d7 12 'iVd3 dxc3 13 'iVxc3 (D) his bishops out!) he will have little to counter
Simply recapturing the c3-pawn is very pop White's space, extra pawn, and bishop-pair.
ular, since it has both a cramping effect and at Fortunately, his knights at the moment are very
tacking strengths. On the other hand, the c-file well-placed and influential. It' s a kind of semi
is open for Black's rooks. There are many other closed position in which for now the knights
moves here such as 1 3 I:Ibl , 1 3 I:Igl and 1 3 are superior to bishops. He also has play down
lllx c3. the c-, g- and h-files, whereas right now White
isn ' t using his king' s bishop or rook at all. Is
that enough to make up for White's own advan
tages? Probably not, except that White also has
isolated c- and a-pawns, and hasn't castled.
Note too that there is an important interior
weaknesses on e3, a square that can be an
chored by . . . d4 and is particularly vulnerable to
a knight on f5 and a queen on b6. Generally
speaking, White' s advantages, if he can keep
them, are probably the better ones in the long
term, so you will usually find Black trying to
upset the equilibrium in the near term.
All in all one can see why players would be
willing to take both sides of this position.
13 tt:Jrs
•••
Let's take stock. White's advantages are A multipurpose move that holds off ctJd4 and
pretty obvious: he has cleared out the kingside, controls e3. This introduces the idea of . . .'iVb6
which for one thing means that the move .. .f6 at some point, which White promptly forestalls.
(to undermine his centre) would be unsupported 14 .i:Ibl (D)
by a pawn. His bishop-pair can be usefully
placed on d3 and e3 for both attack and de
fence, although the dark-squared bishop can be
a target in that case. A bishop on a3 would be
wonderful but generally it's just too hard to
achieve. In view of this his most potent posi
tional threat is to expand on the kingside by
means of g4 (preceded by either h3 or I:Ig 1 ).
That would deny Black's knight the powerful
f5-square. Alternatively, White can try to play
g3, jg,g2 and 0-0; that arrangement of pieces is
another trade-off because when the bishop
moves off its original diagonal White tends to
be subject to central and queenside attack.
Finally, we shouldn't forget that h-pawn. If
White can get bishops to d3 and f6, then Black 14 I:IcS !?
•••
will at the very least have to devote several 1 4 . . . 0-0-0 connects rooks and is considered
pieces in an attempt to stop it from marching up better. Then . . . @b8 clears the way for . . .I:Ic8,
the board and promoting. and also for the intriguing manoeuvre ... b6,
What is Black doing in the meantime? He ... jg,c8 and ... jg,a6 or ... jg,b7. For all the time
has sacrificed a pawn, apparently in order to get that takes, White cannot easily find his way into
at White's queenside and gain development. Black's position.
That lead in development is absolutely critical 15 jg,d2
FRENCH DEFENCE 307
White protects against 15 ...lllcd4, which was @g4! jg,xa8 loses to the double attack 29 jg,b5 ! ,
otherwise a strong move. threatening the queen and mate! Amazing.
15 d4 16 'iVd3 lllce7 17 lllxd4! lllxd4 18
•.• 21. ...l:!.c8 22 'iVb4
'i¥xd4 ctJf5 19 'i¥xa7 'iVxc2 (D) Now the situation looks bad for Black, in
spite of a few tactical details.
22 llld 4! 23 @f2 .l:!.g4?
•••
23 ... 'iVe4 ! 24 .l:!.el 'iVd5 was suggested, but 25

.l:!.c l lllc2 26 'iVd6 ! should do.
24 'iVxd4! 'iVxbl 25 .l:.gl ! 'iVa2 26 jg,e2 .l:!.c2
27 .l:!.dl ! (D)
This sort of position is not easy in spite of

White's extra pawns, because Black has all the
light squares and a superb piece placement ifhe
gets . . . jg,c6 in.
20 'iVxb7!
A good example of what we saw in the intro
ductory chapters: a pawn-raiding queen, as long Having given up the exchange, Karpov's
as it can't be trapped, often does better to stay bishops and three extra pawns reign. Notice
close to the enemy camp to bother his pieces in how the central structure has remained basi
stead of retreating to hypothetical safety. cally the same throughout the entire game. That
20 .l:.c7 (D)
.•• is typical not only of the Winawer but of the
French Defence in general.
27 .l:!.g8 28 g3 jg,c6 29 'iVd3 jg,d5 30 'iVb5+
••.
@rs 31 f5! exf5 32 'iVxd5! 'iVxd5 33 jg,b6+

@e7 34 .l:!.xd5 @e6 35 .l:!.d6+ @xe5 36 jg,f4+
@e4 37 .l:.d7 .l:!.a8 38 .l:!.e7+ 1-0
Fichtl - Golz
Dresden 1959
10 dxc3 11 f4 ctJbc6 12 jg,e3

.••
A slightly odd move-order by both sides. 1 2

'iVd3 jg,d7 1 3 jg,e3 lllf5 would transpose. White
is playing this way in order to keep the c-file
closed and develop quickly.
12 jg,d7 13 'iVd3 ctJf5 (D)
•••
21 'iVb8+ 14 jg,d4 ! ?
Karpov gives the remarkable variation 2 1 One o f White's ideas with jg,e3 was t o dis
'iVa8+ .l:!.c8 2 2 .l:!.b8 @e7 ! ? 2 3 jg,b4+? (23 'iVb7 ! ) courage ... d4 and this move flat-out prevents it.
2 3. . . .l:!.c5 ! 24 jg,xc5+ 'iVxc5 25 .l:!.xg8 'iVcl + 26 Nevertheless, the bishop can't be maintained
@e2 jg,b5+ 27 @f3, which should be a draw af on d4 without allowing exchanges that at least
ter 27 . . . 'iVdl + 28 @f2, because 27 ... jg,c6+? 28 equalize for Black. Other games have seen 14
This is a standard device for Black, intended

to destroy White's central structure at all costs.
18 0-0-0 l:ige8!
Now the idea is . . . e4.
19 g4 e4 20 �h3? ! ctJd6 21 ii.el
White has four passed pawns but is getting
crushed in the centre.
21. d3 22 ctJc3 ctJc4! (D)
••
ii.fl; e.g., 14 . . . 0-0-0 1 5 l:ib l ! ? ( 1 5 �xc3 returns

us to a more modem look) 1 5 . . . d4 (anyway) 16
lllg3 (headed for e4 if possible) 16 .. .f6 ! (a use
ful resource to remember) 17 ctJxf5 exf5 l 8
exf6 �xf4 with a nice advantage, Fuchs-Ohl
mann, Dresden 1959. White can't afford to
give up the centre in this line without compen
sation.
14...0-0-0 15 i.xc3?!
White wants to win material before Black ex Introducing the ideas of . . .�a5 and . . .�b6.
changes on d4. After 1 5 .l;i.bl ? ! f6 16 exf6 ctJfxd4 23 cxd3 �xf4+ 24 i.d2 lllxd2 25 l:ixd2
l 7 ctJxd4 �xf4, Black's advancing centre will ctJd4 26 �g2
dominate the board. Perhaps the best move was White can't stop . . .ctJb3+.
15 g3; for example, l 5 . . .@b8 1 6 .l;i.bl ?! ( 1 6 26 lllb3+ 27 @c2 lllxd2 28 �xd2 e3!
..•
i.g2 lll fxd4 1 7 ctJxd4 ctJa5 followed by . . .�c5 and Black won in short order.
is a commonly-occurring piece disposition that
gives Black at least equality) l 6 . . . ctJfxd4 1 7 The Contemporary 7 ... 0-0
lllxd4 lllxd4 1 8 �xd4 f6 ! 19 exf6 e5 ! with the
idea 20 fxe5 l:ig4! and . . .,l;i.e4+. 1 e4 e6 2 d4 d5 3 ctJc3 i.b4 4 e5 c5 5 a3 i.xc3+
15 ... d4 16 i.d2 6 bxc3 llle7 7 �g4 0-0 (D)
l 6 ctJxd4? loses to l 6 . . . lllcxd4 1 7 i.xd4
i.b5 ! 1 8 �xb5 ctJxd4.
16 ... f6! 17 exf6 e5! (D)
Instead of sacrificing the g-pawn, Black de

fends it and intends to confront White on the
FRENCH DEFENCE 309
kingside, almost always by means of ...f6 or 15 i.d2 Cllf5 16 'iVg5 exd4 17 cxd4 c3 18
... f5 . He retains the option of . . . c4 (keeping i.e3 lllcxd4 19 i.xd4 lllxd4 20 .l:.gl
White' s bishop away from d3) or ... cxd4 (at Most of these moves are forced. Now White
tacking the centre). seems to have an attack but everything is cov
The drawback of 7 ... 0-0 is that it subjects ered.
Black to a dangerous attack by White's pieces, 20 g6 21 'iVe5 'iVc5 22 'iVeS+ 'iVf8 23 'iVxf8+
•••
in particular the queen on g4, knight on f3, and Wxf8 (D)

one or both bishops. The attack can be supple
mented by h4-h5 and .l:.h3, or by f4, assuming
that White's knight has moved.
Now 8 i.d3, bringing the bishop to d3 before
deciding upon anything else, is the overwhelm
ing favourite. 8 Cllf3 is the most natural move,
but was put under a cloud by several games, in
cluding the following:
Roth - Kindermann
Vienna 1996
8 Cll f3 Cllbc6 9 i.d3 f5 10 exf6 .l:.xf6 1 1 i.g5

e5!
This is the characteristic pawn-break in the
7 ... 0-0 line. If Black gets ... e5 in, he'll usually Black is an exchange for a pawn down, yet
be in control of events. Of course, White will he's winning easily; look at White's five iso
sometimes prevail for tactical reasons. lated pawns and his rooks.
12 'iVg3 (D) 24 @dl i.f5 25 .l:.cl .l:.e8 26 .l:.el b5! 27
i.d3 i.xd3 28 .l:.xe8+ @xe8 29 cxd3 c2+ 30
@el a5
A pleasant finish. Kindermann will just
march his pawns homeward.
31 f4 b4 32 axb4 axb4 33 Wd2 Cllb3+ 34
@xc2 lllxcl 35 @xcl Wd7 0-1
We'll look at a few games after 8 i.d3, be

ginning with two very nice ones for White.
J. Polgar - U hlmann
Amsterdam 1990
8 i.d3 (D)
8 ...f5 9 exf6 .l:.xf6 10 i.g5 .l:.f7 11 'iVh5
12 .l:!.xf3!
..• White's unsophisticated strategy comes down
Here is another instance of the eternal ex to checkmate.
change sacrifice on f3 in the French Defence; it 1 1 h6
.•.
is only rivalled in frequency by the ... .l:!.xc3 sac 1 1 . . .g6 is also played. We won't cite theory
rifice in the Sicilian Dragon. here but follow the game instead.
13 gxf3 12 i.g6 .:.rs 13 lllf3 lll bc6 14 o-o 'iVc7!?
1 3 'iVxf3 e4. 1 4 ... i.d7 ! is probably the best move (and
13 c4 14 jg,e2 'iVa5!
•.• certainly the most practical one), leading to
Black attacks the c3-pawn, but also unpins complicated play.
his knight. 15 i.xe7 'iVxe7
15 . . .lllxe7 l 6 ctJe5 (D) is the very picture of a and here the easiest of several wins was 26
dominant knight versus a bad bishop: i.xf5 .l:!.xf5 27 'iVxh6 Wf7 28 'iVh7+ Wf6 29
llle 5.
Polzin - Giemsa
Bad Wiessee 2004
8 i.d3 'iVa5 9 i.d2 lllbc6 (D)
Black can try to get rid of the knight by

16 ... lllc 6 17 f4 cxd4, but 1 8 i.d3 ! threatens
•g6 with a mating attack, and 1 8 . . . lllxe5 1 9
fxe5 doesn't help because it opens the f-file.
16 .U.ael 'iVf6!?
Maybe 16 . . . i.d7 is better, but then 1 7 c4 is
tough to meet. Black tries to tie White down to defending
17 llle5! cxd4?! 18 f4! dxc3 his queenside.
A cute line is 1 8 ...lllxe5 19 fxe5 'iVg5 20 10 ctJf3
.:.Xf8+ Wxf8 21 'iVf3+ We7 22 'i¥f7+ Wd8 23 By the way: whether White or Black, always
•f8+ <Jl;c7 24 'iVd6#. watch out for i.xh7+.
19 g4! (D) 1 0 fS 11 exf6 .l:!.xf6 12 'iVh5 ctJf5 13 c4!
••.
White plays inventively, combining f4 with 1 3 g4 is also promising. Then 1 3 ... c4 leads to
g4. Polgar is devastating in such positions. a heavily-analysed sequence that shows pure
19 'iVe7 20 i.d3!
•.. attack vs defence at its best and is probably
20 g5 lllxe5 isn't clear. somewhat in White's favour, one line being 14
20 'iVeS 21 ctJg6 i.d7
••. gxf5 cxd3 1 5 .l:.gl i.d7 1 6 c4 'iVc7 1 7 i.h6
On 2 1 ....l:!.f7 comes 22 g5. i.e8 1 8 'i¥h4 filh6 19 'iVxh6 dxc4 20 cxd3 cxd3
22 g5! .l:!.f7 23 gxh6 gxh6 24 Whl! llle7 25 21 f6 i.g6, Shirov-Khalifman, Dos Hermanas
:g1 lllf5 2003; now 22 .l:!.xg6 ! ? hxg6 23 'i¥xg6 looks
FRENCH DEFENCE 311
promising: 23 . . .�f7 24 �xf7+ @xf7 25 fxg7

@xg7 (25 ... l:td8 26 l:tbl ) 26 @d2. As this is a
variation involving high theory, you'd do well
to check recent developments.
13 �a4 14 g4 dxc4 15 .te4 tZ:lcxd4 (D)
.•.
1 5 . . . tZ:lce7 1 6 .tg5 �a5+ 1 7 @fl g6 1 8 �h3

l:tf8 1 9 gxf5 exf5 20 .td3 ! cxd3 2 1 .txe7 and
White was on the verge of winning in Stell
wagen-Kim, Iraklion 2004.
1 4 .txf5 ! ? l:txf5 1 5 �xh6 tZ:le7 works out

nicely for Black because of his outpost on f5,
the potential for . . .e5 or ... �c7, and the specific
idea of . . . l:tf7-h7.
14 l:tfS 15 gxh6 e5 16 �g2 �e8!
•.•
1 6 ... @h7 17 tZ:lf3 e4 1 8 tZ:lg5+ @h8 may also

favour Black, although White would have more
active pieces than in the game.
17 dxc5
Even worse is 1 7 �xd5+? .te6 1 8 �g2
16 gxf5! tZ:lxc2+ 17 00! tZ:lxal 18 .tc3 exd4.
�dl+ 17 e4 18 .tb5 .td7 19 tZ:le2 tZ:le5 20 .txd7
.•.
1 8. . .e5 19 tZ:lxe5 .txf5 20 .txf5 l:txf5 2 1 �xd7 (D)

�xf5 �d l + 22 .te l �d5 23 l:tg l and there's no
counterattack. Now it's just a slaughter:
19 @g2 �d8 20 tZ:lg5 h6 21 l:tdl .td7 22
fxe6 �e8 23 exd7! �xh5 24 d8�+ l:tf8 25
�d5+ @h8 26 .txg7+ 1-0
For something more attractive from Black's

point of view, we'll look at a game in the same
variation with another result. Remember that
these are meant to be edifying games, not theo
retical ones, as shown by the date.
Aseev - Vladimirov
USSR Army Cht (Leningrad) 1989
8 .td3 f5 9 exf6 l:txf6 (D) 21 tZ:ld4!?

10 �h5!? h6 1 1 g4 tZ:lbc6! 12 g5 g6! 13 White wants Black to have to earn his attack
�h4 by straightening out White's centre pawns. Af
White cedes the initiative entirely after 1 3 ter the sequence 21 0-0? tZ:lf3+ 22 @h l tZ:l5h4
�xh6 l:tf7 1 4 .txg6 l:tg7 1 5 .td3 e5 ! ?, when it 23 �g3, 23 ... �b5 ! is a nice shot. Probably 2 1
is difficult for White to keep a balance (or Black @d l ! i s best, although naturally Black also has
could even force a draw by 15 . . . c4 16 .te2 l:th7 the upper hand after 2 1 . ..@h7.
17 �f6 l:tf7, etc.). 21 tZ:lxd4 22 cxd4 tZ:lf3+ 23 @dl @h7 24
•..
13 tZ:Jrs 14 �h3
..• l:tbl l:tac8 25 h4 b6! 26 l:th3?
Although hardly desirable, 26 h5 l:tg8 ! 27 9 �h5 tZ:lg6 10 tZ:lf3 �c7 1 1 h4!? cxd4 12
°iWh3 ! is best. @dl ! dxc3 13 tZ:lg5
26 tZ:lxd4! (D)
••. Black is fine after 1 3 l:th3 f6 ! 1 4 exf6 e5 1 5
fxg7 l:tf6 ! .
1 3...h6 14 f4 (D)
27 �g3?
Last chance for 27 .te3.
27 ... �a4 28 l:tc3 �a6! 29 .te3 �e2+ 30 14...hxgS??
�cl tZ:lf5 31 �g5 d4 32 .txd4 tZ:lxd4 0-1 You really can't afford to make a mistake in
Due to 33 �e7+ @xh6 34 �g5+ @h7 35 these lines ! The way to play it was 1 4...tZ:lce7 !
1i'e7+ @g8. 15 l:th3 tZ:lf5 with the idea 1 6 g4?? tZ:lxf4 1 7
.txf4 g6.
Finally, a short demonstration of primitive 15 hxg5 l:td8 16 a4!
attacking power: Probably what Black missed . .ta3 becomes
the decisive factor.
Guseinov - Riazantsev 16 ...aS 17 �h7+ @rs 18 .ta3+ tZ:lce7? 19
Moscow 1 997 �h8+ 1-0
8 .i.d3 tZ:lbc6 (D) There are hundreds and hundreds of wild at
tacking and counterattacking games in the
variations after 7 �g4 and I highly recom
mend that you take some time to study and en
joy them. But the most interesting feature of
these games is that there are dozens of consis
tently recurring tactical themes that stem from
the nature of the underlying position, that is,
from the pawn-structure. Thus the term ' char
acteristic tactics' applies to these and other
Winawer lines as much as to any other in chess.
The wonderful part is that so many basic types
of tactics mixed with creative thinking can
generate a near-infinite number of combina
tive possibilities.
1 4 P i rc Defence
1 e4 d6 2 d4 tZ:lf6 (D) By way of comparison with the King's In

dian Defence, White has omitted the move c4
in favour of tZ:lc3. How does this influence the
play? First of all, the d4-square is theoretically
stronger than in the King's Indian, because it
can be supported by c3. In reality, defending d4
still turns out to be a problem for White after
moves like ...e5 and/or . . .tZ:lc6 (in some cases
supported by . . . .tg4), because it's not so easy to
redirect the c3-knight without losing too much
time. Furthermore, if White plays d5 (say, in re
sponse to ... e5 or . . . c5), that pawn lacks the sup
port of White's c-pawn. Black also has some
queenside attacking ideas that may not be as ef
fective in the King's Indian; for example, . . .c6
and . . . a6, both intending ... b5.
This is the Pirc Defence, an opening plenti Let's continue with the comparison by look
ful in useful strategic ideas. With l . ..d6 Black ing at the positive side of White's position.
restrains White's e-pawn in preparation for First, 3 tZ:lc3 is a developing move, unlike c4 in
2 ... tZ:lf6, 3 ... g6 and 4 . . ..tg7, much as in the tra the King's Indian. Traditionally, development
ditional King's Indian Defence ( 1 d4 tZ:lf6 2 c4 of knights to c3 and f3 is the best way of arrang
g6 3 tZ:lc3 .tg7 4 e4 d6). The immediate differ ing your pieces when you have an ideal centre.
ence is that White doesn't have time for c4 in By playing 3 tZ:lc3, White also gives himself
the Pirc. leeway to try more ambitious moves after 3 ... g6
Before we move on to alternatives and move such as 4 f4. In the King's Indian Defence this
orders, let's examine some characteristics of the advance is playable and more menacing (4 e4
opening. In the great majority of cases, the first d6 5 f4 is the Four Pawns Attack), but it is also
moves are 1 e4 d6 2 d4 (setting up the ideal cen riskier because White has made so many pawn
tre) 2. . .tZ:lf6 3 tZ:lc3 (see alternatives below) 3 . . . g6 moves and he has a broader centre to defend. In
(D) when we have arrived at the basic position.
, the Pirc Defence, the main line with 4 tZ:lf3
yields a solid, classically centralized position.
White has aggressive piece deployments avail
able such as .tc4 or .te3 in combination with
�d2. Direct moves such as these are generally
easier to implement than in the King's Indian
because the Pirc centre is not quite as vulnera
ble to ... c5 and ... e5 moves (which is not to say
that those moves won't be played) .
Notice that the move-order 1 d4 d6 2 e4 tZ:lf6
also lands us in a Pirc Defence. White of course
has some good alternatives in that case, such as
2 c4 and 2 tZ:lf3, but 2 ...tZ:lf6 is perfectly viable
against those moves as well, possibly leading
into a version of the King's Indian Defence.
Black can go his own way with things like I d4 a4 'Lic5 ! 1 3 axb5 'Lid3+ 14 @fl 'Lixc I , etc .
d6 2 c4 e5 or I d4 d6 2 'Lif3 .tg4 ! ? . A great deal Chernin' s 9 ... h5 is also good.
of theory now exists on these and related posi But let's continue with a better and more re
tions. We've come a long way from the days of alistic way for White to develop: 6 'Lif3 'Lic6 7
a near-compulsory I e4 e5 and I d4 d5. 0-0 (7 'iVe2 'Lic5 8 .tc2 .tg4 ! ; a high-level en
3 'Lic3 counter Yusupov-Adams, Hastings 1 989/90
White plays the most natural and important went 7 'Libd2 'Lic5 8 'Lib3 ..ig4 { or 8 ... 'Lixd3+ }
move, defending the attacked pawn. The Pirc is 9 0-0 'iVd7 I O l:tel 0-0-0 with equality) 7 . . . 'Lic5
one of those openings in which the early op 8 .tc2 .tg4 ! 9 l:tel (9 'Libd2 .te7 or even
tions are instructive, so we'll look at a couple 9 . . . d4 ! ?) 9 . . . .te7 (D), shown in the diagram.
over the next few moves. Here 3 .td3 is some
times played, intending to enter into a system
with c3 and perhaps 'Lie2. There's nothing
wrong with that, of course, but apart from the
normal and good 3 . . . g6 it allows Black to strike
out in the centre immediately by 3 . . . e5. Then 4
c3 allows the cute idea 4 ... d5 ! . It's a case of "If
you don't want to take the centre, I shall !'' I've
discussed this position briefly in Chapter 3 (un
der 'Cross-Pollination'). Let me repeat and ex
pand upon that exposition. After 3 .td3 e5 4 c3
d5, we find that it's a great asset to know about
a wide variety of chess positions and themes.
Chemin did a fascinating analysis of 4 . . .d5,
which I've abbreviated and modestly revised.
The play will usually continue 5 dxe5 'Lixe4 This position is remarkably like a main line of
(D). the Open Variation of the Ruy Lopez, namely, I
e4 e5 2 'Lif3 'Lic6 3 .tb5 a6 4 .ta4 'Lif6 5 0-0
'Lixe4 6 d4 b5 7 .tb3 d5 8 dxe5 .te6 9 c3 'Lic5
I O ..ic2, and now (for example) IO ... .te7 1 1
l:te l (or 1 1 'iVe2 .tg4) I I . ...tg4! (D), which
helps to control d4 and in many cases is fol
lowed by ...d4 or even . . . 'Lie6 and ... d4.
In Chapter 3, we already looked at White's

option 6 .txe4 dxe4 7 'iVa4+ (7 'iVxd8+ @xd8
gives Black the two bishops and active play; e.g.,
8 ..if4 'Lid7 9 'Lid2 'Lic5) 7 ... .td7 8 'iVxe4 .tc6.
The position is similar to pawn sacrifices made
in various openings. Here 9 'iVg4 is forced, when
9 ... 'iVd7 ! is very strong: 1 0 'iVg3 ( I O 'iVxd7+ In fact, the only difference between the two
'Lixd7 1 1 'Lif3 .txf3 1 2 gxf3 'Lixe5 1 3 ©e2 lines is Black's insertion of . . . a6 and . . .b5 in the
0-0-0) I O... 'Lia6 ! ; for example, 1 1 'Lie2 .tb5 ! 1 2 Ruy Lopez version. Without entering into yet
PIRC DEFENCE 315
another digression, I'll just say that this has

both positive and negative features.
3 g6
•.•
Here there's an important transposition 3 . .. e5

4 ll'if3 lt'ibd7, when we're in a Philidor De
fence ! What's more, this is arguably the only
safe way to get to this version of the Philidor
because 1 e4 e5 2 lt'if3 d6 3 d4 ll'if6 allows 4
dxe5 ll'ixe4 5 °iYd5, which is not to everyone's
taste. Details about this and associated move
order issues can be found in Chapter 7.
We shall now move on to a discussion of the

main lines of the Pirc.
What about advancing right away? You won't
Austrian Attack find much in the books about this, and it's easy
to say that 5 e5 is too ambitious, especially
4 f4 (D) since it resembles other openings in which rash
pawn attacks are insufficiently supported and
expose the centre. But it's another thing to
show that. Let's do our own analysis and per
haps pick something up about how to study an
opening while we're at it. As we know, varia
tions that are not highly respected are often the
most instructive. Let's look at two answers to 5
e5 as representative of typical ideas in the Aus
trian Attack:
A: 5 ll'ifd7
•..
B: 5...dxe5
A)
5 lt'ifd7
•.•
This is a dynamic move in the spirit of the

As I've pointed out with many examples at Pirc, avoiding simplification and accepting the
the very beginning of Chapter 3 on pawn-struc challenge. The positions that result are little
tures, the first reaction to 'unusual' defences explored and instructive to investigate.
that cede the centre is generally to throw as 6 ll'if3 c5!
many pawns forward as possible and push the The recommended continuation for Black.
opponent off the board. The Pirc was infre It's consistent to destroy White's centre before
quently played and generally held in low regard he consolidates; of course, that depends upon
until the mid- l 960s, and indeed this response the outcome ! This line transposes into 5 ll'if3
dominated early theory. c5 6 e5 lt'ifd7 . By the 5 e5 move-order, White
4....tg7 has avoided a few of Black's options along the
Before moving on to the main line 5 lt'if3, way.
played in thousands of games, let's see if we 7 lt'ig5!?
can understand what the relation between this This odd sortie may well be the best move,
structure and pawn advances might be. although 7 .tc4 could use more investigation.
Positions with a knight on d7 and the possibil
Austrian Attack with 5 e5 ity of the move e6 are notoriously tactical, so
the move 7 lt'ig5 is likely to occur to you if
5 e5 (D) you've run across such positions before.
Black's hope was to see something along the central pawns makes up for his poor king posi
Jines of 7 exd6 0-0 8 dxc5 (8 dxe7 'iVxe7+ 9 tion.
'iVe2 'iVxe2+ 1 0 ll'ixe2 l:te8 1 1 @f2 ll'ic6 1 2 c3 9 exf7+ @f8
ll'if6 with an initiative for the pawn) 8 . . . 'iVa5 ! 9 And above all, not 9 . . . @d8?? 1 0 lt'ie6#.
..ie2 ..ixc3+ ! 1 0 bxc3 'iVxc3+ 1 1 ..id2 'iVxc5 1 2 10 lt'ie6+ @xf7 11 ll'ixd4 lt'ic5! ?
dxe7 l:te8 ! 1 3 l:tbl lt'ic6 (D). Now Black gets good piece activity. Play
might continue as follows:
12 ..ic4+ ..ie6 13 ll'ixe6! ..ixc3+ 14 @f2
ll'ixe6 15 bxc3 'iVc5+!? 16 'iVd4 l:tc8 17 ..ib3 (D)
This is the game B .Ivanovic-M.Gurevich,

Lucerne Wcht 1 989. White can't easily reorga
nize, and look at those exposed internal weak
nesses on e3 and e4 ! This way of destroying 17 lt'id7!?
•.•
White's centre is relatively common, especially Black still has difficulty freeing his king, and
the device of allowing White's capture dxe7 he isn't out of the woods after 1 7 . . .'iVf5 1 8 l:te l
and responding with . . . l:te8. ll'ic6 1 9 'iVd5 either.
7 cxd4
••• 18 l:tel ll'idf8 19 g4!
Worse is 7 . . . h6? ! 8 ll'ixf7 ! @xf7 9 e6+ @g8 Here White still has threats. Naturally both
10 exd7 lt'ixd7 1 1 ..ie3. sides have a lot of other options along the way,
8 e6! (D) and it's unlikely that White can actually force
an advantage in this line. But the characteristics
of the position are what count here, and they
can only be indicated by analysis (which in this
case is a lot of fun). Mainly, I wanted to show
that it's easy to dismiss 'premature' attacks on
principle without testing whether the principle
in question applies to a specific position (or
even whether it is valid at all).
8)
5 dxe5 6 fxe5! ? lt'id5 7 ll'if3 (D)
.•.
White retains his centre. Then we have

Black's bishop looking rather restricted on g7.
An apparently logical move is:
7 f6?
•.•
8 'iVa5!?
••. However, this tends to be dubious before
Not 8 . . . dxc3?? 9 exf7+ @f8 1 0 ll'ie6+. But a Black has castled and developed, and is ex
good move to counter White's attack might be tremely poor in this exact position. We shall
8 . . . lt'ic5 9 exf7+ @f8, when Black's mass of look at a better option after the end of this line.
PIRC DEFENCE 31 7
White has a massive advantage. This is a

poorly-played example, but it serves as a warn
ing as to the consequences of neglected devel
opment.
Obviously, Black is much better off accept
ing the limited problems that stem from 7 . . . 0-0
8 ..ic4 (D).
8 exf6
White makes his usual response to .. .f6. The
recapture 8 ... ..ixf6 would leave Black with a
weak isolated e-pawn on an open file; then 9
ctJe4, 9 ..ic4 and 9 ..ih6 are all good moves.
Therefore Black in principle would prefer:
8 exf6?! (D)
•••
In this position the defence that has actually

been played by grandmasters is 8 ... ..ie6, when
9 ..ixd5 ..ixd5 I 0 ct:Jxd5 W/Jxd5 1 1 Wife2 is a
modest line with an excellent record. Black
normally plays l l .. .b5 (to stop c4) 1 2 0-0; for
example, 1 2 ... ctJd7 1 3 c3 ( 1 3 b 3 ! , intending
1 3 . . .b4 1 4 a3 !) 1 3 . . . ctJb6 ( 1 3 ... a6) 1 4 b3! a5 1 5
..ia3 Wifd7 16 ..ic5 with the better game for
White, Unzicker-Chandler, Buenos Aires OL
1 978. At the end White is enjoying more space,
while Black's bishop is still hemmed in on g7.
Possibly Black should opt for 8 ... ctJxc3 9 bxc3
c5 10 0-0, or 8 ... ctJb6 9 ..ib3 ..ig4 IO 0-0 ctJc6
1 1 ..ie3 ct:Ja5 1 2 Wifd3 ! ? and White's centre and
Recapturing with the pawn might be OK in space may count for somewhat more than the
some positions but here it is much too early be bishop-pair, but that's open to argument.
cause Black's king is stuck in the centre and he After all that, I should add that after 5 . . . dxe5,
also suffers from a weakness on e6 and a bishop 6 dxe5 (D) is less instructive but may be even a
on g7 that is blocked off. better move (or at least an easier one to handle
9 ..ic4 W/Je7+ in practice).
Against 9 ... ..ie6, I O Wife2 simply wins. The Theory gives White a slight advantage after
only apparent try is I O... @f7, but after 1 1 0-0 6 ... Wifxdl + 7 ©xdl ct:Jg4 (7 . . . ctJh5 resembles the
there's no defence to both ctJg5+ and ctJe5+. main line 5 ctJf3 0-0 6 e5 dxe5 7 dxe5 ctJh5, but
Moreover, the supporting move 9 ... c6 fails to I O in our case you won't get the pin on g4 that hap
Wlie2+ ! @f7 1 1 0-0 l:te8 1 2 ct:Jg5+ @f8 1 3 Wiff3 ! . pens there - see below; one of several ways for
10 @f2! White to proceed is the calm 8 ..ie2! ..ih6 9
Threatening l:tel . ct:Jd5 @d8 I 0 g3 c6 I I ctJc3 ©c7 1 2 ..ie3 with
10 ct:Jxc3 1 1 bxc3 Wifd6 12 l:tel + @dS 13 a4
.•. better-placed pieces) 8 ctJd5 ! @d8 9 ©el c6 I O
..if5 14 ..ia3 ctJc3 f6 ! ? 1 1 h 3 ctJh6 1 2 ctJf3 .
Austrian with Conventional Development
B 5...0-0
From this basic position we'll look at some
games:
Ljubojevic - Timman
Bugojno 1980
6 e5
Again White attempts to run the opponent
over.
6...dxe5
6 . . .ll'ifd7 aims for . . .c5. Since after 7 ..ic4
In general, 5 . . . dxe5 may leave something to ll'ib6 8 ..ib3, 8 . . . ll'ic6, 8 . . .c5 and 8 ...ll'ia6 all give
be desired; it illustrates the dangers of early Black fairly easy equality, I'll mention 7 h4! ?.
simplification when the opponent commands White is intent upon checkmate. This leads to
more space. very long and supposedly worked-out varia
What have we learned by this exercise? For tions that you are invited to study in depth with
one thing, White seems to have more positive whatever resources are available. One main line
chances after 5 e5 than indicated by what little is 7 . . . c5 8 h5 cxd4 (D).
existing theory is devoted to it. More signifi
cantly, we see how one might go about investi
gating a position on one's own, and how helpful
it is to have general knowledge about structures
and their characteristic properties.
The Austrian Attack Main Line: 5 ll'if3
5 lt'if3 (D)
With this natural move we return to normal
theory.
9 �xd4 (9 hxg6 dxc3 10 gxf7+ l:txf7 1 1

..ic4 e6 12 lt'ig5 lt'ixe5 1 3 �h5 h6 !) 9 . . .dxe5 1 0
�f2 ! e4 ! 1 1 lt'ixe4 ll'if6 12 ll'ixf6+ exf6 1 3 hxg6
l:te8+ 14 ..ie3 hxg6 15 ..id3 �b6! ( 1 5 ... �a5+
16 c3 with an edge for White, Banas-Kinder
mann, Tmava 1 987) 1 6 @d2 �a5+ 1 7 @c 1 ( 1 7
c3 ..if5) 1 7 ... ll'ic6 1 8 ..id2 �d5 ( 1 8 . . . lt'ib4 ! 19
..ic4 ..if5) 1 9 lt'ih4? ( 1 9 �h4 ..if5 20 ..ic3 l:te6
is equal) 1 9 . . . ..ig4 20 �fl lt'id4 21 ..ic4 �c6
and Black was winning in Varadi-V.Ivanov,
Nyiregyhaza 2002. You get the idea: this is a
Now in the face of 6 e5 (or 6 .ltd3 and 7 e5), line demanding preparation from both sides. It
Black has two logical reactions, the natural de would be fun to look for an improvement in the
veloping move 5 . . . 0-0 and the central counter middle of the confusion from move 9 on.
attack 5 . . . c5. We' ll try to understand each. 7 dxe5!?
PIRC DEFENCE 319
If White wants to keep the pieces on, 7 fxe5 10 ..ixf3 11 gxf3 ct:Jc6 12 c3 l:tad8+ 13
•••
ct:Jd5 8 ..ic4 will generally transpose to the 5 e5 @c2

line; that's a position that's a little irritating for Remember this handy square for the king;
Black and makes you wonder about 6 . . . dxe5. you'll want to go there in queenless middle
7 'iVxdl+ S @xdl ctJh5! (D)
... games that arise from numerous openings.
Black sees a third option (other than 8 . . . ctJfd7 13 ..ih6 14 b4 e6
.•.
or 8 . . . ctJg4). With a knight on the rim Black can The game is equal.
try to force weaknesses. Instead, 8 . . . l:td8+ 9
..id3 ctJd5 10 ct:Jxd5 l:txd5 1 1 @e2 is probably Beliavsky - Anand
better for White, at least in practice. Munich 1991
6 ..ie3 (D)
This has been a successful move in many
games; White not only develops, but also dis
courages ... c5, which is Black's main source of
counterplay. Anand finds a way to challenge
White's centre that draws upon several of the
main ideas that the Pirc has to offer. Then he ap
plies a touch of ingenuity.
9 ..ic4!?
Allowing doubled pawns but gaining the
bishop-pair in return.
9... ..ig4!
As mentioned above, this pin wasn't avail
able in the line 5 e5 dxe5 6 dxe5 'iVxdl + 7
@xd l . White has to be very careful now.
10 ctJe2! (D)
Odd, but perhaps best. The natural 1 0 @e2
ct:Jc6 1 1 ..ie3 runs into 1 l . . ...ixf3+ 12 @xf3 ? 6 b6
.•.
..ixe5 ! ; and 10 ..ie3 ..ih6 ! shows another point Black prepares to play . . .c5 anyway, and he
of . . . ct:Jh5. hopes that the move . . . ..ib7 will come in handy
at some point. Notice that 6 ... ct:Jg4 is met by 7
..ig l followed by h3. That's why White didn't
castle first before playing ..ie3.
B 7 e5 ctJg4 8 ..igl c5
This is the standard picture of an undermin
ing process by Black .
9 h3 ctJh6
A standard Pirc manoeuvre: from here the
knight can go to f5 hitting g3 and d4. But while
this happens, White will form a huge centre.
10 d5 ! ..ib7
11 'iVd2
1 1 g4 keeps the knight trapped and unable to
move (another Pirc theme that has won White
many a game), but Black has anticipated that: Look at White's wonderful centre pawns !
l l ...dxe5 1 2 fxe5 e6 ! (there are no obvious Surely Black has gone wrong?
threats but suddenly all of Black's pieces will 19 c4!
.•.
be aimed at the centre) 13 ..ic4 ll:id7 ! 14 ..ih2 The star move, which of course Anand has
l:te8 ! (D). anticipated. First, he stops ..ic4 in the most rad
ical manner, and at the same time he prepares
the line-opening . . .c 3 ! . It's amazing that he can
do all this against White's well-protected ad
vanced pawns, but White is also slightly behind
in development.
20 ..ie2!? l:tc8 21 @bl!?
Probably 2 1 ..if4 was better, but 2 1 . . .ll:ie7!
22 d6 ll:id5 has the idea of ... c3 and also the
sneaky ...�d7-a4.
21... ..ih6! 22 ..if4 ..txf4 23 �xf4 ..ixd5 24
h4 (D)
On 24 g4 ll:ig7 ! , the knight will get to the
ideal blockading square e6.
'The threat is stronger than its execution' ! 15

0-0 exd5 16 ..ixd5 ..ixd5 17 ll:ixd5 ll:ixe5 1 8
ll:ixe5 ..ixe5 1 9 ..ixe5 l:txe5 20 ll:if6+ @g7 and
Black is a pawn ahead.
11 ll:if5 12 ..ih2 dxe5 13 fxe5 e6! 14 0-0-0
•••
14 g4? ll:ih4 takes away another central de

fender; 14 d6 is met by 14 ... ll:id7 and the long
diagonal adds to White's worries.
14 exd5 15 ll:ixd5 ll:ic6
..•
If Black gets a knight to d4 everything will

fall apart for White, so:
16 c3 ll:icd4!
Black plays it anyway !
17 ll:if6+! 24 c3 25 bxc3 l:txc3 26 h5 ll:ie3! 27 ll:ig5
•••
17 cxd4 �xd5 1 8 @bl l:tad8 and . . . ll:ixd4 �c7 28 ll:ixh7 l:tb3+! 0-1
will follow, picking up material. The end would be 29 axb3 �c2+ 30 @al
17 ..ixf6 18 cxd4 ..ig7 19 d5 (D)
..• �c3+ 31 @bl �xb3+ 32 @al ll:ic2#.
The Main Line with 6 ..id3
6 ..id3 (D)
Here White takes a breather from immediate
attacking mode. The initial idea is pretty obvi
ous: he wants to castle and decide later upon
which attack to pursue. Apart from e5 again, a
likely candidate for attack consists of the trans
fer of the queen to the kingside by �el -h4 fol
lowed by f5 and ..ih6. White can also build up
patiently by means of @hl and ..ie3.
We' ll follow two games with 6 . . . ll:ia6 and
6 . . . ll:ic6. A natural question arises: why not
PIRC DEFENCE 321
Since 8 ...cxd4 was a positional threat, and

since 8 dxc5 lllxc5 brings the knight into the
centre with an easy game, White tries to take
the c5-square away from the knight and cramp
Black's position at the same time. If he gets the
chance he will simply forge ahead with his cen
tral pawns by e5 and drive Black back, or he
might play for f5 in conjunction with moves
like Wke 1 -h4. This is all quite dangerous.
For the moment, however, it is premature for
White to play 8 e5? lllg4 ! 9 h3 cxd4 10 llle2
llle3 ! . See how the centre keeps collapsing in
these extended-centre lines? White should also
avoid 8 .ixa6? ! cxd4 ! 9 lllxd4 bxa6 (D).
play 6 ... .ig4 directly? A plausible answer is
that, without the possibility of ... lllg4, White can
play 7 .ie3, but then 7 ... e5 ! ? is rather compli
cated. Better is 7 h3 .ixf3 8 Wkxf3; for example,
8 ... e5 9 dxe5 ! dxe5 1 0 f5 ! with a straightforward
advantage, intending 10... lllc 6 1 1 g4 llld4 1 2
Wkf2.
J. Polgar - Svidler
Tilburg 1996
6...llla6 (D)
We talk about doubled a-pawns at several

points in this book. Generally speaking, the sur
render of one's light-squared bishop and ceding
of the b-file is a poor deal. Here White's sound
position should balance out those advantages;
e.g., 1 0 Wkd3 with the idea of Wh l and .ie3, or
perhaps .id2. But L.Barczay-Sandor, Hungary
1 968 showed how easily the active black pieces
can create threats: 10 lllb 3? a5 ! 1 1 Wkf3 .ib7 12
a4? (but 12 .ie3 a4 1 3 llld2 a3!) 12 ... Wkb6+ !
1 3 .ie3 Wkb4 (suddenly the e-pawn is falling)
14 .l:i.ael ( 14 .id4 lllxe4 !) 14 ....ixe4 1 5 lllxe4
By developing his knight to the rim, Black Wkxe4 16 Wkxe4 lllxe4 1 7 .id4 .ixd4+ 1 8 lllxd4
prepares ... c5, keeps a diagonal open for his f5 ! 19 lllc6 Wf7 20 lllxa5 .l:i.fc8 (White has re
c8-bishop and, non-trivially, stays out of the gained his pawn - the a-pawn - but Black's e4-
way of pawn attacks by White. The move's knight is a rock and he has two open queenside
main drawback, obviously, is that the knight is files) 2 1 .l:i.e2 .l:i.ab8 22 lllb 3 .l:i.c4 23 a5 a6 24
far from the centre and cannot be fully effective .l:i.d l lllc5 ! 25 .l:i.del .l:i.b7 26 lllxc5 .l:i.xc5 . White
in that area of the board. As an instructive les loses a pawn, and Black stands to win.
son in positional themes you couldn't do better 8 .ig4 (D)
...
than to study this variation. A position with some curious features has
7 0-0 c5 8 d5 arisen. It's often the case that if Black tries to
play a Benoni structure when White has a knight Wkxf8 1 6 .ie2 Wkf2 1 7 ti:Jd5 ! and Black has no
on c3 and no pawn on c4, he can't drum up attack) 1 5 . . . .l:i.xf3 1 6 gxf3 !Dexd3 1 7 cxd3
much play on either side of the board. The key !Dxd5 ! 1 8 exd5 WVh4 (ouch; the dark squares
move ... b5 is hard to get in before White's cen are falling) 1 9 Wg2 WVh5 20 .if4 Wkxd5 2 1 �3
tre starts rolling. But in this case the elimination Wkxb3 22 axb3 .ixb2 23 .l:i.a2 (23 .l:i.el d5 ! ?)
of White' s f3-knight not only takes the pressure 23 ... .ie5 24 .ixe5 dxe5 25 .l:i.e2 Wf8 and Black
off e5, but gets rid of the c8-bishop, which is of eventually won in Komeev-Marin, Spanish Cht
ten in the way (for example, it interferes with (Lanzarote) 2003.
the connection of Black' s rooks). Now Black' s b) Two younger giants played very accu
plan will be ... !Dc7, ... a6, ... .l:i.b8 and ... b5, sup- rately in Grishchuk-Ponomariov, Lausanne
ported by ...Wkd7 or ... !De8-c7 if necessary. Bar- 2000: 9 .ic4 !Dc7 1 0 h3 .ixf3 1 1 Wkxf3 a6 1 2
ring that, the move ... e6 may undermine White ' s a4 b6 1 3 Wkd3 ( 1 3 f5 ! ? ti:Jd7) 1 3 . . .Wkb8 ! 1 4 .ie3
centre. Of course, White will have the consid Wkb7 1 5 .l:i.abl (D).
erable advantage of the bishop-pair to compen
sate him for these troubles.
9 a3
A few other instructive excerpts:
a) 9 Whl e6 ! ? (not the only move, of course)
1 0 dxe6 fxe6 1 1 f5 ! (this is an excellent move
that does several things at once: it activates the
c I -bishop, short-circuits Black's planned ... d5
due to e5, and attacks the e6 point which, if it
falls, will give White the chance to control d5)
l l . ..ti:Jb4 (Black has to forget the ... b5 plan and
concentrate upon the centre) 1 2 h3 (also dan
gerous is 1 2 fxg6 hxg6 1 3 e5 dxe5 14 .ig5 !)
1 2 ... .ixf3 13 .l:i.xf3 (D).
1 3 ...ti:Jd7 ! (try not to forget this move! A
knight on the e5 outpost is the best defender 15 ... e6 ! (perfectly timed; Black is stuck on
and attacker in most Sicilians, King 's Indians, one front and takes the chance to hit the centre,
and Benonis, as well as in a variety of other based upon tactics) 1 6 b4 ! ( 1 6 dxe6 fxe6 1 7
openings) 1 4 fxe6 !De5 ! ? ( 14 ... .l:i.xf3 ! is a better Wkxd6 .l:i.fd8 ! 1 8 .ixe6+ Wh8 1 9 Wke7 .l:i.e8)
and more adventurous idea, because 1 5 Wkxf3 16 ... exd5 17 exd5 cxb4 1 8 .l:i.xb4 b5 ! 19 axb5
CLJe5 keeps White's pieces under check, and the axb5 20 .ib3 (White doesn't want to lose his
positionally superior 1 5 gxf3 !Dxd3 1 6 Wkxd3 d-pawn but now Black utilizes the a-file to
Wkh4 1 7 Wg2 .id4 1 8 .ie3 !De5 and ....l:i.f8 gives equalize) 20... .l:i.a5 2 1 f5 Wka6 22 fxg6 hxg6 23
Black attacking chances) 1 5 ti:Jd5 ? ! ( 1 5 .l:i.xf8+ ! !De4 .l:i.al 24 !Dxf6+ .ixf6 25 .l:i.bf4 .ie5 (that
PIRC DEFENCE 323
square again !) 26 l:i.4f3 l:i.xfl + 27 l:i.xfl llle8 28 17 c3

.i.d4 112-112. White' s remaining bishop is bad, so The b2-pawn needs protection, and neither
he doesn' t have the forces to do any damage. 1 7 .i.d4? .i.xd4 1 8 °iVxd4 'iYxc2 nor 1 7 l:i.abl b6
9 ... llld7 10 h3 .i.xf3 11 °iVxf3 l:i.c8! (D) is very inspiring.
l 7... lllc5 18 .i.xc5 l:i.xc5 19 l:i.ael °iVa5 20
°iVg3 °iVb6 21 l:i.f2 e6 22 dxe6 fxe6 23 .i.g4 l:i.f6
24 °iVe3 h5 25 .i.dl l:i.c8 26 °iVxb6 1/z-112
The opposite-coloured bishops ensure equal
ity. A fair result from a well-played game.
Hellers Ftacnik
-
Haninge 1 989
6... lllc6 (D)
Simple but also insightful. White will stop

. . .b5 and use his bishop-pair if given half a
chance, so Svidler decides to make room for his
pieces in a more aggressive way, based upon
some good calculation.
12 .i.e3
1 2 °iYe2 would prevent c4 temporarily but
1 2. . .°iVa5 1 3 llld l (not 1 3 .i.d2?? c4! 1 4 .i.xc4
'iVc5+) 1 3 ... lllc7 ! (threatening ... b5) 1 4 .i.d2
°iVb6 1 5 c4 e6 ! breaks up the centre just in time.
12 ...°iVa5 13 °iVf2 c4 14 .i.e2 lllac5 This is the most conventional move. It strikes
Now that the knights have access to c5 they're at the slightly shaky d4 point ( ... .i.g4 can fol
roughly as good as the bishops. Giving up the low) and Black contemplates ... e5. Having given
dark squares by 1 4 ... .i.xc3? 1 5 bxc3 'iVxc3 is 6 ... llla6 so much attention, I'm going to pick
not recommended in any case, but White even out only a few points of interest here.
has 1 6 .i.g4! (protecting c2) 1 6 ... l:i.c7 1 7 .i.d4 7 0-0
'iVa5 1 8 °iVh4 with a terrific initiative. White has one very dangerous alternative:
15 .i.f3 llla4 16 lllxa4 'iVxa4 (D) a) 7 .i.e3 lllg4 8 .i.gl e5 9 fxe5 dxe5 10 d5
llld4 is not clear, but probably Black stands sat
isfactorily.
b) 7 d5 lllb4 8 .i.'c4 (8 .i.e2 c5 ! ?) 8 ... c6 !
changes the central equation and should be OK
after 9 a3 cxd5 10 exd5 llla6 or 10 ...°iVa5 !?.
These ideas also show up in the main line with
4 lllf3 .
c) 7 e5 is not so easy to equalize against,
since Black has neither . . . c5 nor . . .e5 at his dis
posal; for example, 7 ... dxe5 8 fxe5 llld7 ! ? (D)
(with the idea of ... lllb4 and . . . c5, although that
may not achieve much; Black has the moves
8 ... lllg4 and 8 . . . lllh5 to look at, and the same
moves before exchanging - a key move against
every one of them is .i.e4, strengthening White' s

control over the centre; both sides should check
theory for details).
with a distinct edge, Hector-Ftacnik, Haninge

1 990.
b) 8 ... lllh5 ? 9 .i.e3 dxe5 10 dxe5 f6 1 1 exf6
.i.xf6 1 2 h3 and Black's got that awful e-pawn
Now: and e6-square to deal with.
c 1 ) 9 .i.e4 ! ? is very interesting, because 9 dxeS llldS 10 h3?!
Black needs a plan, and White does well after Better is 10 lllxd5 °iYxd5 1 1 °iYel ! .
9 ... lllb6 10 0-0 .i.g4 1 1 .i.e3 ! with the idea 1 0 lllxc3 1 1 bxc3 .i.fS 1 2 .i.e3
...
l 1 .. .lllc4 1 2 .i.f2 lllxb2? 1 3 °iYbl lllc4 1 4 'iYxb7 The g7-bishop is suffering, but we've al
lll4a5 1 5 °iYb5 .i.d7 16 l:i.abl l:i.b8 1 7 °iYc5, ready seen that ...f6 would come with problems
when Black's pieces are uncoordinated, espe and isn't worth it yet.
cially that knight on a5. 12 °iYd7
..•
c2) 9 llle4 lllb4 10 .i.c4 c5 1 1 c3 lllc6 1 2 0-0 Hellers-Ftacnik, Haninge 1989. It's about
cxd4 1 3 cxd4 lllb6 1 4 .i.b3. Here is the key equal. White should play 1 3 l:i.bl and llld4.
point. Since ...f6 isn't possible, Black needs to
put pressure on the d-pawn or eliminate some Austrian with 5 . . . c5
of White's pieces: 14 ... .i.g4 ( 14 ... llla5 15 .i.c2
.i.e6 ! ? looks initially promising, but 1 6 °iYel 1 e4 d6 2 d4 lllf6 3 lllc3 g6 4 f4 .i.g7 5 lllf3 cS
lllc6 1 7 °iYh4 threatens various attacks with (D)
.i.h6, lllfg5, l:i.f3-h3, lllc 5, etc., in whatever or
der works !) 1 5 llleg5 e6 16 h3 .i.xf3 1 7 lllxf3
°iYd7 1 8 .i.g5 with a small but definite advan
tage because of the dark squares and Black's
hemmed-in g7-bishop, Wang Zili-D.Gurevich,
Lucerne Wcht 1 9 89.
7....i.g4 8 eS (D)
It makes a lot of sense to make this move
when ... c5 is a long way off.
8 dxeS!
•••
There are two other possibilities that should

give you an idea of the strength of White's cen
tre:
a) 8 ... llld7 9 .i.e3 dxe5 (9 ...lllb4 10 .i.e4 ! )
10 dxe5 f6 ( 10 ...lllb 6) 1 1 exf6 exf6 (as s o often,
this turns out to leave Black's position a little Black plays to free his game immediately
airy) 1 2 h3 .i.e6 1 3 .i.b5 lllb6 1 4 'iYe2 l:i.e8 1 5 and avoid the cramped games that can arise af
l:i.ad l 'iVe7 16 l:i.fel °iYb4 1 7 llld4 .i.c4 1 8 °iYf3 ter 5 ... 0-0. The problem is that many lines here
PIRC DEFENCE 325
are simply tactical sequences of 'only' moves, somewhat, as ... e6 remains in the air) 1 9 ...l:i.f7
so I ' ll try to limit the quantity of material. Note 20 .i.xg7 Wxg7 2 1 'iVe3 e6 22 lllc 3 d5 ! (D).
that 6 e5 lllfd7 ! ? (not the only move) transposes
to the 5 e5 line.
Hermlin - Chipashvili
USSR 1976
6 .i.bS+
This is still the critical line. Black has held
his own for years after 6 dxc5 'iVa5 7 .i.d3
'iVxc5, but this is the variation that most resem
bles other openings in its positional themes,
and deserves a look. We' ll follow Kindermann
M.Gurevich, Haifa Echt 1989: 8 'iVe2 0-0 (if
Black wants to be sure of getting . . . .i.g4 in, he
can play it now) 9 .i.e3 'iVa5 J O 0-0 (D).
Black has dissolved his weakness and taken
over the initiative. He was never in trouble in the
opening. We' ll follow the game with minimal
notes: 23 °iVf3 fxe4 24 dxe4 d4 ! ? (24... l:i.af8 ! 25
exd5 l:i.xf4 26 'iVxf4 l:i.xf4 27 l:i.xf4 exd5 and ... d4
follows) 25 llle2 °iVxf3 26 l:i.xf3 e5 27 fxe5?
l:i.xf3 28 gxf3 d3 ! 29 lllc 3 (29 lllf4 lllxe5 30 l:i.dl
g5 !) 29 ... l:i.f8 30 Wg2 lllxe5 3 1 l:i.fl g5 32 l:i.f2
lllg6 33 l:i.fl lllf4+ 34 Wg3 Wf6 35 h4 We5 ! ? 36
hxg5 l:i.g8 37 Wh4 h6 ! ? 38 gxh6 l:i.g6 39 llld l
(39 h7! l:i.h6+ 40 Wg3 l:i.xh7 4 1 Wf2 l:i.h2+ 42
We3 l:i.xb2 43 l:i.bl !) 39 ... l:i.xh6+ 40 Wg3 l:i.g6+
4 1 Wh4 Wd4 0- 1 .
6 .i.d7 7 es
...
This is the main move, leading to complica

1 0. . . lllc 6 ( J O. . . lllb d7 ! ? is a Sicilianesque tions that any player of 5 ...c5 must know. 7
move that has been tried out, but most players .i.xd7+ is a more interesting move from a posi
don't want to be so cramped) 1 1 a3 .i.g4 (the tional point of view: 7 ... lllfxd7 (7 ... lllbxd7 8 d5
move-order has been a bit strange; normally isn't as easy for Black, in part because 8 ...b5
J O... .i.g4 comes first) 1 2 h3 .i.xf3 1 3 'iVxf3 can now run into 9 e5 and 8 ... 0-0 9 °iVe2 isn't
llld7 (Black is essentially playing a Sicilian comfortable) 8 d5 b5 ! ? 9 °iVe2 ! b4 JO llld l
Defence, where his knights are harmoniously lllb6 ! ? (to prevent e5 by hitting the d-pawn) 1 1
placed and he should have equality; for the mo 0-0 0-0 (or 1 l . ..°iVc8 ! ?), and now:
ment, ... .i.xc3 is threatened) 14 .i.d2 °iVb6+ 1 5 a) 12 c4 bxc3 1 3 lllx c3 'iVc8 ! intending
Whl lllc5 1 6 l:i.abl lllxd3 1 7 cxd3 f5 ! . An ex . . .°iVa6, Martinovic-Jansa, Lingen 1 988. End
cellent move. It creates a few weaknesses, but ings should be fine for Black: the combination
blocks off the fl -rook and especially the d2- of ... c4 and the b-file grant active play.
bishop from entering the game; the move f5 b) 1 2 lllf2 'iVc8 1 3 l:i.el ? ! a5 1/z - 1/z Shirov
would have freed them both. The game contin Beliavsky, Madrid 1 997. Once again ...'iVa6 is
ued 1 8 llld5 ( 1 8 g4 e6 19 gxf5 exf5 20 llld5 corning.
°iVb3 2 1 .i.c3 l:i.ae8 22 °iVg2 lll e7, Glek-Lobron, c) 12 f5 ! has been suggested, because the
Bundesliga 1 990/1 ; Black should have no prob knight can't get to e5 yet. Then 1 2 ... gxf5 1 3 c4
lems) 1 8 ... °iVb3 1 9 .i.c3 (the same position but yields surprising compensation. As usual, f5
without g4; the difference should favour Black frees the c l -bishop and f l -rook.
7 l2Jg4 (D)
.•. 1 1 lLixb5 �a5+ 1 2 c3 ( 1 2 �d2 .i.f2+ 1 3 @d l
lLie3+ 1 4 @e2 �xb5+ 1 5 @xf2 lLig4+ 1 6 @g3
lba6 ! turns out well) 12 . . ..i.f2+ 1 3 @d2 .i.e3+
1 4 @c2 �a4+, etc. , which is apparently equal.
b) 8 .i.xd7+ �xd7 9 d5 dxe5 10 h3 e4 ! 1 1
lLixe4 lLif6 is a trick worth remembering that
comes up again and again. Now 1 2 lLixf6+
.i.xf6 gives Black equality thanks to his power
ful bishop on f6, and the pawn-grab 1 2 lLixc5
can be met by 12 . . . �d6 ! 1 3 �d4 ( 1 3 lLixb7??
�b4+) 13 . . .0-0 1 4 lLie4 lLixe4 1 5 �xe4 lLid7
with compensation. There may be a way for
White to do better in this less-investigated line.
8 ...cxd4
Convoluted theory focuses upon 8 . . . .i.xb5 9
lLixb5 dxe5 ! 1 0 hxg4 �a5+ 1 1 .i.d2 ( 1 1 c3 e4;
8 h3!? 1 1 lbc3 exd4) 1 1 ...�xb5 1 2 dxe5 �xb2 ! 1 3
a) 8 e6 is White's most famous continuation: .l:i.bl �xa2 1 4 .l:i.xb7 �d5 1 5 �bl with an un
8 .. .fxe6 (8 ... .i.xb5?! leads to well-analysed com clear situation.
plications beginning 9 exf7+ @d7 {forced} 10 9 �xd4 lbh6 (D)
lLixb5 �a5+ 1 1 lbc3 cxd4 1 2 lLixd4 .i.xd4 13
�xd4 lbc6 14 �c4 �b6 1 5 �e2 h5 1 6 .i.d2
lLid4 1 7 �d3 lLif5 1 8 lLie4 .l:i.ac8 19 0-0-0 and
White comes out with somewhat the better
game) 9 lLig5 .i.xb5 (D) , and now we have more
theory :
10 g4
White shuts out the knight from re-entering
the game. 10 .i.xd7+ �xd7 1 1 g4 lbc6 1 2 �e4
0-0 is a standard position in which neither side
has made much progress, even though there's
plenty of play; e.g., 1 3 .i.d2 dxe5 14 fxe5 �e6? !
a l ) 1 0 lLixb5 �a5+ 1 1 c3 �xb5 12 �xg4 ( 1 4 .. .f5 ! 1 5 exf6 exf6 i s equal) 1 5 0-0-0 f5 16
cxd4 1 3 lLixe6 �c4! 14 lLixg7+ @f7 with mind exf6 �xe4 1 7 lLixe4 exf6 18 lLic5 ( 1 8 .i.c3)
boggling complications - you'll need to consult 1 8 ... .l:i.ae8 19 .l:i.he l ( 1 9 lLixb7) 19 . . ..l:i.xel 20
the books and databases for this one. .l:i.xel f5 21 g5 lLif7 22 .i.f4 b6 23 lbd3 .l:i.c8 24
a2) 10 �xg4 .i.c4 1 1 b3 .i.xd4 12 .i.d2 .i.d5 @d2 with equality, Thorhallsson-Gretarsson,
is another line that will require study. Hafnarfirdi 1 992.
a3) 1 0 lLixe6 and now Black has the famous 10....i.xbS 11 lLixbS �as+ 12 l2Jc3 l2Jc6 13
resource 1 0 ... .i.xd4 ! , with the point that 1 1 �e4 0-0-0 14 .i.d2 dxeS 15 fxeS fS!? 16 �c4?
lLixd8 .i.f2+ 1 2 @d2 .i.e3+, etc., is a draw. Best is 16 exf6 exf6 1 7 �e6+ @b8 1 8 0-0-0.
There are further well-worked-out tactics after 16 fxg4 17 hxg4 .l:i.hf8
•.•
PIRC DEFENCE 32 7
White's e-pawn is weak. Most of the lines 6 °iVe2 c6

in this whole variation (excluding 6 dxc5) have Black can't stop e5, but this stabilizes the
a positional basis but are also forcing. They centre.
should probably be learned by heart. 7 e5 dxe5 8 dxe5 llld5 9 .i.d2 .i.g4!
This gets rid of some pieces and puts real
The .tc4 Variation pressure on White's e-pawn at the same time.
10 h3
1 e4 d6 2 d4 lllf6 3 lllc3 g6 4 lllf3 Not 1 0 0-0-0?! e6 1 1 h3 .i.xf3 1 2 gxf3 llld7
4 .i.c4 could be the right move-order if you 1 3 f4 °iVh4. Then Black has real pressure on the
want to play this sy stem, depending upon what f4-pawn and White's pawn-structure is bad.
you think of 4 ... lllxe4 5 .i.xf7 + (or 5 lllxe4 d5 10 .i.xf3 11 gxf3 (D)
•..
6 'iVe2 ! ? dxe4 { 6 . . . dxc4 ?? 7 lllf6# } 7 'iYxe4) 1 1 'iVxf3 e6 1 2 °iVe2 llld7 1 3 f4 °iVh4+ 1 4

5 ... Wxf7 6 lllxe4 .i.g7, when a sample line is 7 °iVf2 'iVxf2+ 1 5 Wxf2 f6 1 6 exf6 .i.xf6 and
ll:if3 l:i.f8 8 c3. Black has some attack even with the queens off,
4... .i.g7 5 .i.c4 (D) Serrnek-Nogueiras, Moscow OL 1 994.
We'll take a quick look to see how both sides 1 1 ...e6

handle this potentially tactical line. Or 1 l ...lllxc3 12 .i.xc3 e6. Notice that we're
now in another of those ... e6/... c6 restraint struc
Rublevsky - Khalifman tures and White has no d-pawn. Wrute lacks a
St Petersburg 1999 good plan.
12 f4 llld7
5 ...0-0 Very solid. Black has at least equalized.
Now 5 ... lllxe4 ! ? 6 .i.xf7+ (6 lllxe4 ! ? is also 12 ...°iVh4 1 3 °iVg4 ! would gain a tempo because
possible) 6...Wxf7 7 lllxe4 l:i.f8 8 0-0 Wg8 might Black doesn't want to straighten out White's
be worth trying. Black has the bishop-pair and a pawns when he also has the advantage of two
central majority with a nice f-file. On the other bishops.
hand, White has a space advantage and Black's 13 h4!?
squares down the e-file are vulnerable, while at Trying to break things open a bit for Ills bish
the same time White's knights are nicely cen ops.
tralized. Probably it's one of those many chess 13...lllxc3 14 .i.xc3 lllb6 15 .i.b3 h5 16
positions in which, if the owner of the two bish .i.d2 a5 17 a3 llld5! 18 c4 llle7
ops (Black) can stabilize the position and avoid Black heads for the perfect outpost on f5 .
serious weaknesses, Ill s centre and bishop-pair 19 0-0-0 lllf5 20 .i.c3 'iYe7 21 .i.c2 l:i.fd8 22
will assert themselves in the long run. But White l:i.xd8+ °iVxd8 23 .i.xf5 exf5 24 °iVe3 'iVe7 25
looks ready to use his knights and major pieces l:i.dl .i.f8 26 °iVb6 l:i.e8! (D)
along the open e-file to prevent that. 27 e6!?
Black waits for h3, his bishop often has no

where useful to go), and discourages White
from pressing too hard in the centre.
a) You can get a feel for White's space ad-
vantage after 6 ...ll:Jc6?!, which allows 7 d5. A
funny line is 7 ...ll:Je5 8 ll:Jd4 ! ? (8 ll:Jxe5 ! dxe5 9
.i.e3 leaves Black looking for a plan) 8 ... c5 ! 9
dxc6 ll:Jxc6 1 0 .i.e3 with an exact transposition
to a Sicilian Dragon. White can meet 7 ... ll:Jb8
with 8 h3, preventing ... .i.g4, when Black is
pressed for space. Another good approach is 8
l:i.el e5 ! ? (8 ... .i.g4 9 .i.f4) 9 dxe6 ! .i.xe6 1 0 .i.f4
with a simple central advantage: 1 0...h6 1 1
ll:Jd4 .i.d7 1 2 °iYd2 Wh7 1 3 e5 ! dxe5 14 .i.xe5
White wants to attack. Instead, 27 .i.xa5 (D) .
.i.h6 28 .i.d2 °iYxh4 29 °iYxb7 .i.xf4 30 'iYxc6
l:i.xe5 is a mess - look at all those passed pawns!
But White's king isn't safe.
27... fxe6
Not 27 ... 'iVxe6? 28 °iYd4.
28 °iYxa5 .i.g7 29 l:i.gl .i.xc3 30 'iVxc3 Wf7
31 °iYg3 °iYf6 32 l:i.dl l:i.d8 11z_11z
A double-edged variation, but Black was
positionally better out of the opening.
Classical Variation
1 e4 d6 2 d4 ll:Jf6 3 ll:Jc3 g6 4 ll:Jf3 il.. g7 5 il..e2
0-0 6 0-0 (D)
White has a substantial advantage in this

game between two ex-World Champions, and it
surprisingly turned into a miniature following
14 ... ll:Je4? 1 5 ll:Jxe4 .i.xe5 1 6 ll:Jf3 .i.g7 1 7 l:i.adl
'iVc8 1 8 .i.c4 .i.e8 1 9 ll:Jeg5+ ! hxg5 ( 1 9 ... Wg8
20 ll:Je6 wins for White) 20 ll:Jxg5+ Wg8 2 1
°iYf4 ll:Jd7 22 l:i.xd7 ! .i.xd7 2 3 .i.xf7+ 1 -0 Tal
Petrosian, USSR Cht (Moscow) 1 974.
7 .i.e3
One disadvantage of 6 . . . .i.g4 is that it has
allowed this move without White having to
bother about ... ll:Jg4. Instead, 7 h3 .i.xf3 8 .i.xf3
has never given Black serious problems after
8 ... e5. White has the two bishops but his posi
6....i.g4 tion is hampered by the knight on c3, which al
This is Black's most logical and classically lows Black to sink his knight in on d4 and
motivated development. With this move he pre exchange a bishop, or otherwise play a well
pares to put pressure on d4 by ... .i.xf3 and a timed .. .f5. The tempo lost by h3 is meaningful;
combination of . . . ll:Jc6 and ... e5 , when his otherwise perhaps White could reorganize and
knights may well be a match for White's bish gain the advantage. Instead of 8 ... e5, Black also
ops. 6 ... .i.g4 also helps to clear his back rank (if has the more ambitious move 8 ... ll:Jc6, again
PIRC DEFENCE 329
taking advantage of his extra tempo. Then 9 ( 1 1 f4 ! i s probably good, but some players may
.i.e3 e5 1 0 dxe5 dxe5 l l lll d5 is well-answered find it too loosening; one line among many is
by l l . . .lllxd5 12 exd5 llld4, but of course there' s l l . . .llled7 12 dxc6 { or 12 .i.d4 } 12 ... bxc6 1 3
oodles of theory to look at. °iYd3 'iVb8 1 4 a3 with a small edge; play what
7 lllc6 (D)
... works ! ) l l . . .a5 ! ? ( l l . ..°iYa5 1 2 l:i.a3 aims for
l:i.b3, but 1 2 ... l:i.fc8 1 3 °iYd2 cxd5 14 exd5 lllc4
1 5 .i.xc4 l:i.xc4 1 6 l:i.b3 b6! is equal, and illus
trates a common plan for Black) 12 .i.d4 llled7
1 3 °iYd2 'iYc7 1 4 l:i.adl l:i.ac8 1 5 l:i.fe l l:i.fd8 and
White has difficulty playing for advantage be
cause Black' s pieces are so well-placed, Rozen
talis-Ftaenik, Manila OL 1 992.
We now return to 8 °iYd2 (D):
8 °iYd2
Logical: White connects rooks, contemplates
.i.h6, and challenges Black to advance in the
centre. The other important main line is 8 d5,
and then:
a) 8 . . . lllb 8 can lead to the characteristic ma
noeuvre 9 llld4 .i.xe2 1 0 'iYxe2 c5 l l lllf3 °iYb6
1 2 l:i.abl °iYa6! , either exchanging queens in a
position with no weaknesses or mobilizing the 8 e5
...
queenside pawns. This may be Black's best 8 . . . l:i.e8 used to be played, a useful move that
line. waits for White to commit before playing . . . e5.
b) 8 . . . .i.xf3 9 .i.xf3 llle5 1 0 .i.e2 c6! (D). But White has a more useful move in 9 l:i.fel !
a6 ! ? (to prevent lllb5 in view of the line 9 ... e5? !
1 0 d5 .i.xf3 1 1 .i.xf3 lll d4 12 .i.xd4 exd4 1 3
lllb5 ; note that 9 l:i.fel protected the e-pawn in
this variation) 10 l:i.adl (every white piece is
centralized) 10 ... e5 1 1 dxe5 dxe5 (1 l . . .lllxe5
1 2 lllxe5 dxe5 1 3 °iYc l °iYc8 14 .i.xg4 'iYxg4 1 5
f3 'iVe6 1 6 °iYd2 .i.f8 1 7 lll d5 and White cap
tures with pieces on d5, winning the d-file) 1 2
'iYc l 'iVe7 1 3 lll d5 lllxd5 14 exd5 lll d8 1 5 c4 f5
1 6 c5. White is in control of the game, Geller
Pi'ibyl, Sochi 1 9 84.
9 d5
The endgame 9 dxe5 dxe5 1 0 l:i.adl has given
Black problems but a good line is 1 0. . .°iYc8 ! 1 1
'iVc l l:i.d8 l 2 l:i.xd8+ lllxd8 ! followed by . . . llle6,
Black has to strike quickly before White aiming at d4 and f4; e.g., 1 3 l:i.d l llle6 1 4 h3
consolidates the two bishops. His plan includes .i.xf3 1 5 .i.xf3 c6 1 6 llle2 'iYc7 17 c3 a5 1 8 °iYc2
moves like . . . °iYa5, . . . cxd5, a rook to the c-file, .i.f8 ! intending ... .i.c5, Kaidanov-Wolff, USA
and perhaps . . .lllc4. A typical line goes 1 1 a4 ! ? 1 990. . . . .i.f8 is a nice move to remember, getting
rid of both White' s good bishop and Black's h4-h5, @g2 and l:i.hl , or he may want to support
bad one ! the pawn-push f4, or he can do what he does in
9 !De7 (D)
... the game) 1 2 .. .f5 1 3 .i.e2! ti:Jf6 14 f3 (that's the
end of Black's kingside attack) 1 4 ... °iYd7 1 5
.i.b5 'iVc8 l 6 l:i.f2 a6 1 7 .i.fl ti:Jh5 1 8 .i.h3 °iYe8
1 9 !De2 @h8 20 c4 (after all that we get two
bishops and a standard-looking queenside for
mation) 20... b6 2 1 l:i.dfl °iYf7 22 f4 and White
had too much firepower in Spassky-Parma, Ha
vana OL 1 966. An excellent positional demoli
tion.
11 ti:Jel b5
Some noteworthy play follows l l ...!Dg4 1 2
.i.xg4 .i.xg4 1 3 f3 .i.d7, because Black has the
two bishops but he isn't well organized to meet
14 f4 ! (D).
Now the question is whether White can make

something of his space advantage. The exam
ples seem to indicate that he can.
Donaldson - Felecan
Kona 1998
10 l:i.adl!
It's odd, but this position seems the least
promising that we have seen thus far for White.
He has made no progress on the queenside,
Black still has his bishops, and with the knight
on e7, the idea of . . .f5 has gained force. How
ever, White has achieved something that he This seems to grant Black an outpost on e5
hasn't been able to do in any other line. So far but he can't get to it, whereas White will win
Black has always successfully traded White's more than his share of the centre. For example,
d-pawn after ... c6, allowing him to take over 14 . . . .i.g4 1 5 ti:Jf3 f5 ( 1 5 ... °iYd7 1 6 fxe5 dxe5 1 7
good posts rapidly before White could reorga .i.c5 f5 1 8 °iYg5 .i.f6 1 9 °iYh6 l:i.f7 2 0 d6, Gligo
nize. With threats and piece activity, White's ric-Ptleger, Moscow Echt 1 977) 16 h3 ! .i.xf3
bishops didn't have time to find good positions. 1 7 l:i.xf3 with the ideas l:i.dfl and l:i.del .
But here Black doesn't get . . . c6 in because the 12 a3 aS 13 b4
d-pawn will fall. And .. .f5 is still to be dis The bottom line is that White has space and a
cussed, but has its problems. This means that better grip on the position. A different order is
White has time to put his pieces on the appro 1 3 tD<l3 °iYb8 14 f3 c6 ! ? ( 1 4 ... b4 1 5 !Dbl ! bxa3
priate squares and make a pawn-break, either 1 6 !Dxa3) 1 5 dxc6 .i.xc6 1 6 b4 d5 ( 1 6. . . axb4?
with f4 or c5. 1 7 !Dxb4 ! l:i.xa3 1 8 !Dxc6 !Dxc6 1 9 !Dxb5 l:i.a2
10 .i.d7!
.•• 20 .i.c4 and Black is in big trouble, Vogt-Ber
Instead of waiting around, Black embarks nard, Wildbad 1 990; l 6. . .l:i.c8 ! ?) 1 7 .i.c5 l:i.e8,
upon a new idea: queenside expansion. A game Kuczynski-Chemin, Budapest Z 1 993, and now
of Spassky's is a model of how White should White could play 1 8 exd5 !Dexd5 19 !Dxd5
handle the exchange on f3: 1 0....i.xf3 1 1 .i.xf3 !Dxd5 20 bxa5. It seems that White keeps the
ti:Jd7 12 g 3 ! ? (a little odd-looking, but the move advantage in a number of ways, which is a good
is very flexible; White may be interested in sign for 10 l:i.adl and his opening as a whole.
PIRC DEFENCE 331
13 axb4 14 axb4 �b8 lS r3 (D)

.•. 16 llld3 c6 17 dxc6 .llxc6 18 lllr 2!? l::td7
White still stands better after l 8 . . . d5 1 9
.llc 5 lllc 8 20 exd5 lllxd5 2 1 lllxd5 l::txd5 (not
2 1 ....llxd5?? 22 c4) 22 �e3 l::txd l 23 l::tx dl
�c7 24 llle4 !.
19 lllg 4!?
Going for the f-file. Perhaps White had a
better move, but he foresees the promising sac
rifice ahead.
19 lllxg4 20 rxg4 dS 21 exdS lllxdS 22
.•.
lllxdS l::txdS 23 �xdS! .llxdS 24 l::txdS (D)
1S...l::td8
1 5 . . . l::ta3 (D) has two good answers:
If the b-pawn falls, White gets two passed

pawns, and f7 is also a target. But his pieces are
loose on the kingside, so Black gets counter
play. Although one feels that White should have
a way to combine attack and defence, the posi
tion can probably be assessed as dynamically
equal. Since the opening is long over, I'll just
a) 1 6 lllb l l::ta8 1 7 c4! bxc4 1 8 .llxc4, and show the moves of this fascinating game:
now the only freeing move, 1 8 ... c6, would open 24 �cS 2S l::tcS �b7 26 .tr2 l::tc8 27 .tr3
..•
up the c4-bishop: 19 dxc6 lllxc6 20 lllc 2 ! . �d7 28 l::tdl �e6 29 l::td6 �e8 30 .llc6 �e7 31
b) 1 6 llld3 c6 1 7 dxc6 .llx c6 1 8 l2lc 1 ! (or .llxbS l::ta8 32 l::td l .llh6 33 h4 e4 34 gS e3 3S
1 8 l::ta l , or l 8 lllf2) l 8 . . .l::td8 19 lllb 3 d5. This is .ll g3 .llg7 36 .lle2 l::td8 37 l::tfl l::td2 38 .llc4
a transposition to Thorsteins-Kasparov, Saint .lld4 39 l::txr7 �xf7 40 .llxr7+ @xf7 41 l::tc7+
John blitz Wch 1 988. It looks as though White @e6 42 @n @rs 43 l::tc4 @e6 44 c3 .lla7 4S
will win a clear pawn after 20 .llc 5 ! l::td7 (or l::tc6+ @rs 46 l::ta6 @g4 47 l::txa7 @xg3 48 l::te7
20... lllc 8 21 exd5 lt:lxd5 22 llla5) 2 1 .llxe7 @r4 49 bS l::td l + SO @e2 l::td2+ Sl @el l::tc2 S2
l::txe7 22 exd5 ! .lle8 (22 . . . l::td7? 23 �c l �a7+ b6 l::txc3 S3 b7 l::tb3 S4 l::txh7 l::tb2 SS l::te7 @g4
24 @h l lllxd5 25 lllxb5 .llxb5 26 .llxb5 l::td6 27 S6 l::te4+ @hS S7 g3 l::txb7 SS l::txe3 l::tb2 S9
.llc4) 23 d6 and lllc 5. l::te2 l::tb3 60 @r2 filg3 112.112
I ndex of P layers
Numbers refer to pages. When a player's name appears in bold, that player had White. Other
wise the FIRST-NAMED PLAYER had White.
ACS - Polgar, J. 1 29 DAMUANOVIC - Ponomariov 1 32

ADAMS - Yusupov 277 DAVIES - Petrovic 1 3 3
AGDESTEIN - Karpov 305 DJURIC - Larsen 7 8
AGREST - Morozevich 203 DOMINGUEZ, L. - Morovic 1 39
AKOPIAN - Volokitin 1 56 DONALDSON - Felecan 328
ALEKSANDROV - Kupreichik 97 DREEV - Shirov 266
ANAND - Beliavsky 3 1 9; Grishchuk 268; ENKHBAT - Fedorowicz 261
Illescas 233; Karpov 53; Kasimdzhanov ESCOBAR - Bianchi 1 09
2 1 5 ; Kasparov 1 9 1 ; Kasparov 260; ESTRIN - Levenfish 1 05 ; Libov 1 6
Kramnik 43 ; Matveeva 259; Shirov EUWE - Keres 145; O'Kelly 95
143, 295 ; Svidler 1 49 FEDORCHUK - Michalek 207
ANIKAEV - Petrosian, A. 66 FEDOROWICZ - Enkhbat 26 1
APICELLA - Kramnik 237 FELECAN - Donaldson 328
ASEEV - V ladirnirov 3 1 1 FICHTL - Golz 307
AVTONOMOV - Spassky 4 1 FISCHER - Geller 209; Portisch 57;
BAREEV - Lastin 279; Pavlovic 246 Spassky 239
BAUER - Lasker, Em. 1 7 FLECK - Short 1 92
BELIAVSKY - Anand 3 1 9; Hector 1 57 FRESSINET - Renet 1 1 7
BENKO - Velimirovic 25 8 FTACNIK - Hellers 323
BERNSTEIN, 0. - Capablanca 69 GELLER - Fischer 209; Hamann 65 ;
BESHUKOV - Malaniuk 1 07 Smyslov 56
BEZGODOV - Morozevich 64 GEORGIEV, KR. - Kosten 1 9 8
BIANCHI - Escobar 1 09 GERSHON - Muzychuk 228
Gleizerov 286
BITI - GIEMSA - Polzin 3 1 0
BOGOLJUBOW - Kramer 77 GINZBURG, M . - Zarnicki 1 7 1
BOHM, H. - Korchnoi 65 GLEIZEROV - Biti 286; Krupkova 293
BOLOGAN - Rublevsky 63 GOLZ - Fichtl 307
BOTO - Buntic 24 1 GRAF - lvanchuk 1 27
BUCHENTHAL - Rosen 2 1 9 GRISHCHUK - Anand 268; Ponomariov 56
BUNTIC - Boto 24 1 GROEBER - Wendland 1 1 1
CAPABLANCA - Bernstein, 0. 69 GUFELD -Hummel 293
CERNOUSEK - Rasik 285 GULKO - Hellers 302; Lakdawala 75
CHIGORIN - Pillsbury 1 67 GUREVICH, M. - Saltaev 292; Zhang
CHIPASHVILI - Hermlin 325 Pengxiang 2 8 8
CHRISTIANSEN - Wojtkiewicz 206 GUREVICH, V. - Zakharov 3 1
CIFUENTES - Vehi Bach 1 20 GUSEINOV - Riazantsev 3 1 2
CRAMLING, P. - Nunn 23 1 GYIMESI - Ponomariov 1 35
INDEX OF PIAYERS 333
HAMANN - Geller 65 LOBRON - Yusupov 42

HECTOR - Beliavsky 157; Hillarp Persson LUKIN - Taimanov 234
300 LUTHER - Rozentalis 282
HELLERS - Ftacnik 323; Gulko 302 MACIEJA - Ivanchuk 297
HERMLIN - Chipashvili 325 MALANIUK - Beshukov 1 07
HILLARP PERSSON - Hector 300 MATANOVIC - Petrosian 78
HUMMEL - Gufeld 293 MATVEEVA - Anand 259
ILLESCAS - Anand 233 MICHALEK - Fedorchuk 207
IORDACHESCU - Wohl 80 MILU - Vajda 1 5 3
ISANBAEV - Sizykh 1 1 6 MOROVIC - Dominguez, L . 1 39
IVANCHUK - Graf 1 27 ; Madeja 297 ; Short MOROZEVICH - Agrest 203 ; Bezgodov 64
303; Topalov 192 MUZYCHUK - Gershon 228
IVANOV, A. - Zilberstein 142 NAIDITSCH - Korchnoi 148
KARPOV - Agdestein 305 ; Anand 53; NUNN - Cramling, P. 23 1
Kasparov 35; Kasparov 1 38; Korchnoi O' KELLY - Euwe 95
42, 69; Korchnoi 99; Lantier 44; ORAL - Khuzman 28 1
Taimanov 32; Uhlmann 274 PAVLOVIC - Bareev 246
KASIMDZHANOV - Anand 2 1 5 ; Leko 1 63 PELLETIER - Sulskis 1 96
KASPAROV - Anand 1 9 1 ; Anand 260; PETROSIAN, A - Anikaev 66
Karpov 35; Karpov 1 38; Lanka 255 ; PETROSIAN, T. - Matanovic 78; Spassky
Short 24 1 81
KERES - Euwe 1 45 PETROVIC - Davies 1 33
KERSTEN - Thinius 200 PIKULA - Popovic, P. 226
KHALIFMAN - Rublevsky 327 PILLSBURY - Chigorin 1 67
KHUZMAN - Oral 28 1 POLGAR, J. - Acs 1 29; Svidler 1 64, 32 1 ;
KINDERMANN - Roth 309 Uhlmann 309
KORCHNOI - Bohm, H. 65; Karpov 42, 69; POLZIN - Giemsa 3 1 0
Karpov 99 ; Naiditsch 148; Ponomariov PONOMARIOV - Damljanovic 132;
1 46 Grishchuk 56; Gyimesi 1 3 5 ; Korchnoi
KOSTEN - Georgiev, Kr. 1 98 1 46
KRAMER - Bogoljubow 77 POPOVIC, P. - Pikula 226
KRAMNIK - Anand 43; Apicella 237; Leko PORTISCH - Fischer 57
1 62 RAMESH - Sorokin 1 3 1
KRISTJANSSON - Tukmakov 205 RASIK - Cemousek 285
KRUPKOVA - Gleizerov 293 RENET - Fressinet 1 1 7
KUPREICHIK - Aleksandrov 97 REUTSKY - Shtyrenkov 204
LAKDAWALA - Gulko 75 RIAZANTSEV - Guseinov 3 1 2
LANKA - Kasparov 255; Santo-Roman 224 ROSEN - Buchenthal 2 1 9
LARSEN - Djuric 78 ROTH - Kindermann 309
LASKER, EM. - Bauer 17 ROZENTALIS - Luther 282
LASTIN - Bareev 279 RUBLEVSKY - Bologan 63; Khalifman 327
LAUTIER - Karpov 44 SALTAEV - Gurevich, M. 292
LEE, G. - Taulbut 7 8 SANTO-ROMAN - Lanka 224
LEKO - Kasimdzhanov 1 63 ; Kramnik 1 62 SEIRAWAN - Short 54, 265 ; Short 68
LEVENFISH - Estrin 1 05 SHIROV - Anand 143, 295; Dreev 266
LIBOV - Estrin 1 6 SHORT - Fleck 1 92; Ivanchuk 303 ; Kasparov
LJUBOJEVIC - Timman 3 1 8 24 1 ; Seirawan 54, 265; Seirawan 68
SHTYRENKOV - Reutsky 204 UHLMANN - Karpov 274; Polgar, J. 309 ;

SIZYKH - Isanbaev 1 1 6 Vogt 275
SMYSLOV - Geller 56 VAJDA - Milo 153
SOKOLOV, A - Timmermans 108 VEHI BACH - Cifuentes 1 20
SOROKIN - Ramesh 1 3 1 VELIMIROVIC - Benko 258
SPASSKY - Avtonomov 4 1 ; Fischer 239 ; VLADIMIROV - Aseev 3 1 1
Petrosian 8 1 VOGT - Uhlmann 275
STEFANSSON - Ward 1 8 8 VOLOKITIN - Akopian 156
SULSKIS - Pelletier 196 WARD - Stefansson 1 88
SVIDLER - Anand 1 49; Polgar, J. 1 64, 321 WENDLAND - Groeber 1 1 1
TAIMANOV - Karpov 32; Lukin 234 WOHL - Iordachescu 80
TAULBUT - Lee, G. 78 WOJTKIEWICZ - Christiansen 206
THINIUS - Kersten 200 YUSUPOV - Adams 277; Lobron 42
TIMMAN - Ljubojevic 3 1 8 ZAKHAROV - Gurevich, v. 3 1
TIMMERMANS - Sokolov, A. 108 ZARNICK! - Ginzburg, M. 1 7 1
TOPALOV - Ivanchuk 1 92 ZHANG PENGXIANG - Gurevich, M . 288
TUKMAKOV - Kristjansson 205 ZILBERSTEIN - Ivanov, A. 1 42
I ndex of O peni ngs
Numbers refer to pages. Codes are ECO codes.
English Opening 286; C07 272, 2 78, 279, 281 , 282; C08 273;
A20 85; A26 1 6; A34 65, 66; A36 60 C09 45, 274, 275, 2 77; C l O 74; C l 1 294, 295,
297; C l 5 298; C l 8 299, 300, 302, 305, 307,
Czech Benoni 308, 309, 310, 311, 312; C19 303
A56 55
1 e4 e5 without Ruy Lopez
Benko Gambit C21 25; C22 1 74; C30 1 66, 167, 1 69; C33 23;
A5 8 55 C36 1 69, 171; C41 27, 1 14, 1 15, 1 16, 1 1 7, 120;
C42 63; C45 63, 64; C54 16, 91 , 95, 97, 99;
Modem Benoni C55 101, 1 1 1 ; C56 26; C57 104; C5 8 109; C59
A65 48 105, 107, 108
Scandinavian Defence Ruy Lopez

BOl 78 C60 122; C62 123; C68 26, 61, 151, 153, 156,
157; en 125; c1s 140, 142, 143; cso 144,
Alekhine Defence 146, 149; C81 145; C82 148, 314; C88 158;
B03 56; B04 56, 80 C89 162, 163, 164; C92 137, 138, 139; C94
133, 135; C96 84, 130, 131, 132, 133; C97 126,
Pirc Defence 127; C99 129
B07 60, 83, 313; B08 37, 327, 328; B09 36,
315, 316, 318, 319, 320, 321 , 323, 324, 325 1 d4 d5 without 2 c4
DOI 77; D02 60
Caro-Kann Defence
B IO 251 ; B 1 2 53, 54, 251, 265, 266, 268; B 1 3 Queen's Gambit
254, 255, 256; B 1 4 258, 260, 261 ; B 1 7 78; B 1 9 D 1 6 47; D 1 8 75; D27 43; D28 41; D34 45; D41
27, 81 27, 259; D42 39; D43 47; D53 42; D56 1 1 ; D58
69; D63 49; D66 15
Sicilian Defence
B 20 1 75; B 22 46, 248, 250; B 27 1 78; B 30 46, Griinfeld Defence
62 ; B34 67, 244, 246; B40 216; B41 221 , 222; D91 74
B42 225, 226; B43 223, 224; B44 35, 227, 228,
230, 23 1, 233; B45 219; B46 234; B50 84; B54 Queen's Indian Defence
1 76; B56 1 78, 236; B58 237; B59 72, 285; B70 E 1 2 68
181, 182; B72 185; B73 37; B74 28; B78 12,
191, 192; B79 188; B 80 28; B 84 1 76; B 86 203, Nimzo-Indian Defence
204; B 87 205, 206, 207; B88 239, 241 ; B 89 E24 32, 56; E32 44; E34 72; E54 32; E57 42;
238, 241 ; B90 194, 202, 212, 215; B92 46, 177, E59 15
208, 209; B96 196, 198; B97 200; B99 196
King's Indian Defence
French Defence E69 57; E70 24; E76 19, 36; E90 37; E9 1 65;
COO 2 70; C02 48, 50; C03 271 ; C04 50, 285; E92 25; E94 31; E97 37, 51, 52; E99 19, 51
COS 53, 283, 288, 291, 292, 293; C06 50, 283,

Watson John 2006 Mastering The Chess Openings Vol 1 EGambit OCR

Uploaded by

Copyright:

Available Formats

Watson John 2006 Mastering The Chess Openings Vol 1 EGambit OCR

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Watson John 2006 Mastering The Chess Openings Vol 1 EGambit OCR

Uploaded by

Copyright:

Available Formats

Mastering the

Copyright © John Watson 2006

Gambit Publications Ltd, 99 Wallis Rd, LondonE9 5LN,England.

Edited by Graham Burgess

Gambit Publications Ltd

1 The Nature of Chess Openings: Fundamentals 11

2 Opening Ideas and Positional Features 23

3 The Significance of Structure 36

Section 1: Open Games

4 Introduction to 1 e4 and the Open Games 87

6 Two Knights Defence 101

7 Philidor Defence 114

8 Ruy Lopez 122

M011er Defence 140

9 King's Gambit 166

Section 2: Semi-Open Games

10 Introduction to the Semi-Open Games 173

11 Sicilian Defence 175

12 Caro-Kann Defence 251

13 French Defence 270

Winawer Variation 298

14 Pirc Defence 313

Index of Players 332

To Maura, the Light of My Life

CDs and DVDs

Books and Articles

Golubev, M.; The Sicilian Sozin; Gambit 2001

available in front of the 2nd rank and both

.ilf4 tllh6 (D)

To confuse things a bit more, the phrase

runs 4 .i.gS [indirectly controlling e4 and d5 by

[indirectly controlling d5 and e4 by unpinning

but doesn' t control a central square] 6 l2Jf3 [d4

At least Black has established some influence

tral squares that it controls (or helps to control)

e5] 3 i.c4 [d5]. Black typically responds with

and d5], the Two Knights.

and e5], and now 3 .i.bS attacks a piece that

Another backward move that allows White to

White's king i s sheltered and relatively safe.

One of the most important guidelines in chess

Space and Its Properties

The amount of territory that is under one's con­

the centre. You can see how positively that af­

By contrast, look at this example from one of

At this point Black controls every central

This is a real-world example: Black has an

3. Rooks like to have open files, preferably

4. Apart from wide-open games in which

lt:lf3 lllc6 3 .i.c4 .i.c5 4 b4 .i.xb4 5 c3 .i.a5 6

ble for White to attack aggressively. Of course

Several very important central formations will

also has the bishop-pair. White has a mobile

The vanished centre shows up in old gambit

1 e4 e5 2 ltJf3 ltJc6 3 j,_b5 a6 4 j,_xc6 dxc6 5

This is all theory, that is, published knowl­

White has a pleasant advantage because he

A surrender of the centre occurs in the old

White has won some famous battles from

The amount of territory that is under one's con

the centre. You can see how positively that af

This is all theory, that is, published knowl

There follow some definitions and short ex

here (see Chapter 1 for a description of the out

would put a very bad bishop on c3 whose po

Domination of a colour complex doesn't nec

In that case four of Black's pieces are devel

first goal is to maintain the blockade on d5, usu Spassky Avtonomov

reverse approach, exchanging rooks (if any Isolated c-Pawns

In this instance Black made only a half

great attack as shown by one game that contin

g4 (D) First Black defends against b4, in tum pre

important central squares, both the black out

What are the main characteristics of the po

Thus White has given Black a compact struc