the

two
knights
defence

by Jan Pinsky
EVERYMAN CHESS
Gloucester Publishers pic www.everymanchess.com
First published in 2004 by Gloucester Publishers plc (formerly Everyman Publishers
plc), Northburgh House, 10 Northburgh Street, London EC1V OAT

Copyright © 2004 Jan Pinski

The right of Jan Pinski to be identified as the author of this work has been asserted
in accordance with the Copyrights, Designs and Patents Act 1988.

All rights reserved. No part of this publication may be reproduced, stored in a retrieval
system or transmitted in any form or by any means, electronic, electrostatic, magnetic
tape, photocopying, recording or otherwise, without prior permission of the publisher.

British Library Cataloguing-in-Publication Data

A catalogue record for this book is available from the British library.

ISBN 1 857442830

Distributed in North America by The Globe Pequot Press, P.G Box 480,
246 Goose Lane, Guilford, CT 06437-0480.

All other sales enquiries should be directed to Everyman Chess, Northburgh House,
10 Northburgh Street, London EC1V OAT
tel: 02072537887 fax: 02074903708
email: [email protected]
website: www.everymanchess.com

Everyman is the registered trade mark of Random House Inc. and is used in this
work under license from Random House Inc.

EVERYMAN CHESS SERIES (formerly Cadogan Chess)

Chief advisor: Garry Kasparov
Commissioning editor: Byron Jacobs

Typeset and edited by First Rank Publishing, Brighton.

Cover design by Horatio Monteverde.
Production by Navigator Guides.
Printed and bound in Great Britain by Biddies Ltd.
 CONTENTS I

Bibliography 4
Introduction 5

1 e4 e5 2 et:Jf3 et:Jc6 3 ~c4 et:Jf6

4 et:Jg5
1 Introduction and 4 ... dS 5 exdS bS!? 7
2 Fritz Variation 4 ... dS 5 exdS ltJd4 19
3 4 ... dS 5 exdS ltJaS - Introduction 30
4 4 ... dS 5 exdS ltJaS - Main Line 42
5 Traxler Gambit 4 ... i.cS!? 60

4d4
6 4 ... exd4: Introduction 86
7 4 ... exd4 5 eS dS 100
8 4 ... exd4 5 0-0 113

4 Others
9 4d3 138

Index of Complete Games 159

I BIBLIOGRAPHY I

Books
Encyclopaedia of Chess Openings Volume C (Sahovski Informator 2000)
The Two Knights' Defence and Traxler Counter-Gambit, J6zsef Pilk6vi (Caissa 2001)
The Two Knights Defence, Yakov Estrin (Bats ford 1983)
The Two Knights Defence, A.Beliavsky & A.Mikhalchishin (Bats ford 2000)
Dreispringerspiel bis Kiinigsgambit, Paul Keres (Sportverlag Berlin 1974)
Play the Open Games as B.hck,John Emms (Gambit 2000)

Periodicals
Chess Informant #1-#87
New in Chess Yearbook #1-#68

Software
ChessBase 8.0
Fritz 8
I INTRODUCTION I

'You should speculate in two cases. When you have reasons to do so, or when you don't,'
said Mark Twain. It is similar with the Two Knights Defence. You can play it when you
have a reason for doing so, and when you don't. Why? Well, if nothing else, it leads to inter-
esting play. For a club player it is a good weapon even against very strong players and also
against weaker players. Black is just obliged to know what he is doing as soon as he has
played (1 e4 e5 2 tLlf3 tLlc6 3 Ji.c4) 3... tLlf6. Finally, but not lastly, the Two Knights Defence
15 a kind of intellectual relic in chess. Every chess player should know something about this

opening - otherwise he has no chess culture. 3... tLlf6 was already played in 15th century,
which makes the Two Knights Defence for chess what Rome is for the Western civilization!

A few notes about the opening

4 d3 looks very innocent, but is still played by many top grandmasters; for example, World
Champion Vladimir Kramnik, the 2002 European Champion Bartlomiej Macieja, Grand-
master Leonid Yudasin, and so on. The most important thing to remember is the following
plan in the main line (4 ... .te7 5 0-0 0-0 6 .l:!.e1 d6 7 c3): ... ~h8, ... tLlg8, and ... f7-f5, even if it
means sacrificing pawns.
Another unforcing line is 4 d4 exd4 5 e5. This variation has, in my opinion, the greatest
future of all those in this book. Here Black has two less well-known continuations (5 ... tLlg4
and 5... tLle4), which he should try out. Secondly, White has some sidelines which gives a
good basis for independent investigations.
The move 4 tLlg5 rose from the dead in the 1990's, woken to life by Grandmasters
Alexander Morozevich and Nigel Short. It is quite an interesting variation. If you are a hard-
working chess player you can seriously hope to get your work returned by good results here
for White. Almost all those who play this position as Black are not so well acquainted with
the actual theory. Many variations give White the chance to win the game from home
through dedicated preparation.
And then there is 4 tLl g 5 Ji.c5!? Now what is this? Leaving f7 unguarded like a beginner
falling for the fool's mate? Not really. These days nobody normal plays Traxler's attack regu-

5
Two Knights Defence

larly with Black. Still it is a tricky line that cannot be completely dismissed just on account of
opening theory.

I hope the Two Knights Defence will bring you pleasure, be you Black or White.

Jan Pinski,
Warsaw,
December 2003.

6
 CHAPTER ONE I
Introduction and
4 ttJg5 d5 5 exd5 b5

1 e4 e5 2 l2lf3 l2lc6 3 .ltc4 l2lf6 4 l2lg5 The most natural move. There is only one
All chess books teach us that we should alternative: 4 ... ..ltcS!? - the Traxler Gambit
not moves the same piece twice in opening, (see Chapter 5). 4 ... tLlxe4? has never been
and in most cases this is good advice. How- played by anyone good. Black hopes for 5
ever, 4 tLlgS is defInitely a correct move. tLlxf7?! (5 tLlxe4 dS) S...1lVh4 6 0-0 tLlxf2 7
Why? There is one important reason: the f7- l:rxf2 .ics with a strong initiative. But 5
pawn (like the f2-pawn) is Black's biggest ~xf7+ is the strongest move and after
weakness in the fIrst few moves of the open S.. .'Jile7 6 d4! h6 7 tLlxe4 ~xf7 8 dS White is
game (1 e4 eS). So White breaks the rule much better.
about moving the same piece, but in so doing 5 exd5
attacks Black's biggest weakness. That is the Here Black has three good options:
secret! S... tLlaS (Chapters 3 and 4), and two very
In this chapter we consider the position similar moves, S... bS (see below) and its twin
after 4 ... dS 5 exdS bS, which is a minor alter- brother S... tLld4 (the Fritz Variation) which
native to the main line S... tLlaS, but still inter- have a common main line covered in Chap-
esting. An important point is that S... bS 6 ter 2.
~f1! tLld4 leads to Chapter 2. Weak is S... tLlxdS? and now:

Game 1
Bahram-Hector
Stockholm 1998

1 e4 e5 2 l2lf3 l2lc6 3 .tc4 l2lf6

The living legend, grandmaster David
Bronstein, thinks that this opening should be
called Chigorin's Counter-Attack rather than
the Two Knights Defence! Bronstein be-
lieves fIrmly in the strength of Black's last
move.
4121g5 d5

7
Two Knights Defence

a) 6 ctJxf7!? gives White has a very strong commented: 'If someone can come up with
attack after 6.. .'it>xf7 7 'iVf3+ ~e6 8 ctJc3 such a new sound plan in a position which
ctJb4 (8 ...ctJe7?! 9 d4 c6 10 .igs h6 11 Jixe7 has been known for 500 years, it seems that
Jixe7 12 0-0-0 l:tf8 13 'ii'e4 .igs+ 14 ~bl chess is truly immortal!'
l:tf4 15 'iVxes+ ~f7 16 ctJxds cxds 17 .ixds+ 6 dxc6?!
~f8 18 .ib3 J:tfs 19 'iVe4 g6 20 h4 and After this move White is actually fighting
White has a completely won position accord- to say alive. The paradoxical 6 .ifl is the
ing to old maestros i\1ieses and Bardeleben) 9 main move here and will be explained in
'iWe4 c6 10 a3 ttJa6 11 d4 ttJac7 12 .if4 ~f7 Game 4. 6 Jixbs?! has also been played, but
13 .i.xes and the white attack is very strong it is weaker, as can be seen in Games 2 & 3.
according to grandmaster Reuben Fine. 13 6 ... bxc4
O-O-O!? is also interesting. Now White has the following possibilities:
But the piece sacrifice is not necessary for 7 tUc3
White to obtain the advantage ... Best, but insufficient for equality.
b) 6 d4! exd4 (or 6... .ie7 7 ctJxf7! ~xf7 8 The alternative is 7 'iVe2?! h6!? (also possi-
'iWf3+ ~e6 9 ttJc3 ctJb4 10 'iWe4 c6 11 a3 ble is 7...'iWd5 8 f4 h6 9 'iWxes+ 'iWxes+ 10
ctJa6 12 'iWxes+ ~f7 13 ctJxds cxds 14 fxes hxgs 11 exf6 gxf6 12 ctJc3 Jid6 and
Jixds+ ~f8 15 0-0 with a huge advantage for Black is slightly better) 8 'iVxes+ Jie 7 9 ctJf3
White) 7 0-0 .ie6 (if 7 ... Jie7 8 ttJxf7 ~xf7 9 0-0 10 0-0 .ig4 11 'iVf4 .id6 with a better
'iWf3+ ~e6 10 ctJc3! dxc3 11 l:te1+ ctJes 12 position for Black whose attack is very dan-
.if4 .if6 13 .ixes .ixe5 14 J:txe5+ ~xes 15 gerous. This was shown in the game Berger-
l:te1+ ~d416 .ixds l:te8 17 'iWd3+ ~cs 18 Zweiberg, corr. 1963-64, which continued 12
b4+ ~xb4 19 'iWd4+ and White won in Mor- 'iVxc4 .ixf3 13 gxf3 l:tb8 14 a3
phy-NN, New Orleans 1858) 8 l:tel 'iWd7 9
ttJxf7 ~xf7 10 'iVf3+ ~g8 11 ::'xe6 l:td8 12
JigS iUxe6 13 .i.xd8 'iWe1+ 14.i.fl iUe6 15
.ih4 and White is much better according to
Euwe.
5 ... b5!?

14... l:tbs!!? (perhaps this is not the best

plan, but it is definitely the most impressive)
15 'iVxbs ctJds! 16 'iVc4 (after 16 'iWxds
.ixh2+ 17 ~xh2 'iVxds Black has a strong
attack with at least perpetual check. White is
of course some material up, but it is sitting
This very interesting idea is copyrighted by idly on the queenside) 16 ... ttJf4 17 ~hl l:te8
the American master Olav Ulvestad, who 18 J:tg1? (it is more difficult to defend than to
wrote an article about this move in 'Chess attack; here White underestimates the
Review' in 1941/1. To this Yakov Estrin strength of Black's attack or else misses some

8
 Introduction and 4 li'Jg5 d5 5 exd5 b5

detail) 18.. .'Yi'h4 19 'iUf1 (if 19 'iUd4 g6! 20 d3 12 ... ~xc6 13 b3 f5 14li'Jc3?!
~h3! and Black wins) 19... ttJd3 20 f4 ttJxf2+ The uncomfortable 14 ttJg5 e4 15 ~h3
21 Wg2 SLc5 and White resigned. 18 d3 was was necessary, when Black can try 15 ... hxg5!?
the best move and if now 18 ...1i'h4? 19 ttJc3! (or 15 ... .tb4 16 ttJxe6 ~xe6 when the game
'iih3 20 .txf4 ~xf3+ 21 Wgl .txf4 22 'iid5 is unclear) 16 ~xh8 ~c5 with strong com-
and White wins. Instead after 18 ... ~g5 19 pensation .
.txf4 .txf4 20 .l:!.gl ~h4 21 l:Ig3 J:!.e1+ 22 14... e4 15 ~h5?!
Wg2 .txg3 23 ~xh4 SLxh4 Black has com- White is surfmg around with the queen as
pensation for the pawns. It is actually likely if it was the Internet. Better was 15 ~e2 i.d6
that he is better here as White has no way to with an attack.
mobilise his pieces. 15 ....i.c5 1S bxc4 gS! 17 ~e2?
7 ... hS 8 li'Jge4 17 ~h4 was better, but the position is
White has no easy choices here. Also pos- very uncomfortable. The white pieces are not
sible is 8 ttJf3 .i.d6 9 'iUe2 0-0 10 0-0 and playing at all.
according to ECO Black has compensation 17 ... .i.xc4!
for the material. This is defInitely the case. Now Black wins.
After something like 1O ...J:te8 White has a 18 d3
diffIcult game in front of him, e.g. 11 'iixc4? If 18 ~xc4 .i.xf2+.
e4 12 ttJel .txh2+! 13 Wxh2 ttJg4+ 14 Wg3 18 ... exd3 19 cxd3 J:.xd3 20 'YWc2 .i.aS 21
h5! 15 'iYb4 h4+ 16 Wh3 as 17 'iUa3 ttJe3+ 18 li'Jd1 .i.b7 22 li'Je3
Wh2 ttJxf1+ 19 Wgl ttJxd2 and Black wins. 9
0-0 might be an improvement, but still Black
has a very dangerous attack.
8 .. .cLlxe4 9 li'Jxe4 ~d5 10 ~f3 .i.eS 11
0-0 O-O-O!

22 ....l:ixe3! 0-1
If 23 fxe3 then 23 ... ~xe3+ and the queen
hangs.

Game 2
After the feeble 11...~e7 White would Grau Ribas-De Groot
have time to complete his mobilisation in Email 1997
peace.
12 J:.e1?! 1 e4 e5 2 li'Jf3 li'JcS 3 .i.c4 li'JfS 4 li'Jg5
This basically just loses a tempo. Better d5 5 exd5 b5!? S .i.xb5?!
was 12 b3 'iUxc6 13 bxc4 f5 14 ttJg 5 'iixf3 15 This move loses the two bishops which
-1Jxf3 i.xc4 16 d3 e4 17 dxc4 exf3 and Black can be crucial in such a sharp position.
IS slightly better. S.. :~xd5

9
Two Knights Defence

Black of course goes for the initiative.

Nevertheless, possible was 8... ~xf3 9 ctJxf3
e4 10 ctJe5 i.a6 11 b3 ctJd5 12 ..ia3 ctJb4
with some compensation for the pawn.
9 'iWb3 ~c5!
Black does not want to waste his time pro-
tecting pawns. It is more important that his
rook will come quickly to the f-ftle.
10 'iWxf7 + It>d8

7 ~xc6+
White also has:
a) 7 ~e2 has been played a few times by
weak players. Their games cannot really be
seen as solid indicators for the way play
should proceed, so I have tried to fmd my
own path: 7...~xg2 8 ~xe5+ j,e7 9 .ttfl 0-0
10 'iVg3 (10 Ji.xc6 'iixc6 and Black is much
better) 1O ... 'iVd5!? (10 ... ~xg3 11 fxg3 ctJd4 12
i..a4 Ji.d7 and Black also has excellent play) 11 0-0
11 ctJc3 'iVc5 12 'iVxc7 ctJd4 13 ~xc5 i..xc5 White is in trouble. Alternatives were:
14 i..a4 h6 15 ctJge4 ctJxe4 16 ctJxe4 Ji.b6 a) 11 ctJc3 l::tf8! 12 Wixg7 J:!.g8 13 'iVh6
v.rith terrific compensation. i..xf2+ 14 ~fl i.a6+ 15 ttJe2 Ji.d4 16 ~e1
b) 7 Ji.e2 Ji.b 7 8 d3 ctJd4 9 ctJf3! (best) Ji.xe2 17 ~xe2 e3 and White has problems.
9... ctJxe2 10 ~xe2 Ji.d6 11 ctJc3 'iVc6 12 0-0 b) 11 ~3 l::tfS 12 0-0 ttJ g4 13 d4 iLxd4
0-0 13 Ji.g5 ctJd7 with compensation, 14 ~4 (if 14l:td1 ~d6! 15 ttJxe4 ~xh2+ 16
Mestrovic-Smejkal, Ybbs 1968. ~fl i.a6+ 17 c4 c5 and Black is much bet-
c) 7 ctJc3 ~xg2 8 'iVf3 ~xf3 9 ctJxf3 Ji.d7 ter, Essegern-Brauer, carr. 1981) 14... Ji.c5 15
100-0 (10 d3 Ji.d6 with equality, or 10... ctJd4 l::tdH ~e8 16 'iVxe4+ 'iVxe4 17 ttJxe4 ttJxf2
11 iLxd7+ ~xd7 12 ctJxd4 exd4 13 ctJe2 Ji.c5 18 ttJxf2 l::txf2 19 lIe 1+ ~f7 20 Ji.e3 iLxe3
and Black is probably a little better) 1O ... i..d6 21 J::txe3 .llxc2 and Black is better in the end-
(10 ...ctJb4!? 11 ttJxe5 ctJxc2 12 ttJxd7 ttJxd7 game; White must lose b2 or g2 now (22 b3
13 l::tb 1 0-0-0 with an unclear game ahead) 11 iLb7 or 22l::tb3? ItcH 23 <;t>f2 ..if5).
i..xc6 i.xc6 12 ttJxe5 i.xe5 13 l::te 1 0-0-0 14 11."h6?
l::txe5 l:!.he8 15 d4 J::txd4 16 lIxe8+ ctJxe8 17 This move gives away the h-pawn for no
iLe3 with an even endgame. reason. It might seem that the game is not
7" :~xc6 8 'iWf3? about pawns, which is true, but still they can
This move is simply a waste of time. have their function in the mating attack.
There is no chance in hell that Black will Black had a win by force here: 11...l:tfS! 12
exchange the queens, even though it is not ~xg7 (for 12 ~3 ttJg4 see 11 ~3 above)
bad at all. Instead, 8 0-0 can be seen in Game 12 ... l:!.g8 13 'iWh6 ..ib 7 14 <;t>h 1 l::tg6! (besides
3. the tempo, we now see why the pawn should
8".e4! be on h7!) 15 'ii'h4 'iVd5 16 ctJh3 (if 16 b4 e3!

10
 Introduction and 4 tDg5 d5 5 exd5 b5

17 f3 e2 18 l:!.e 1 .uxg5 19 bxc5 'iHxf3! and vulnerable.

Black wins, or 16 f4 e3 17 'iVh3 lDe4 18 1S ... ctJf2+ 19 Wg1 ctJh3+ 20 Wh1 .i.xc6!
lDf7+ ~e8 19 lDc3 lDf2+ 20 .uxf2 exf2) Black is winning. There is no way that the
16 ... e3 17 lDf4 ~xg2+ 18 lDxg2 .txg2+ 19 white kingside can withstand this pressure.
~gl .tb7+ 20 ~g3 exf2+ 21 .l:!.xf2 .txf2+ 22 21 l:xe2
~xf2 lDe4+ and Black wins.
12 ~xg7 J:.gS 13 'iYxh6 .i.b7

21 ....i.xf3!
A decisive though not terribly difficult
White is under a lot of pressure. This combination.
looks almost like a beginner's game, where 22 d4 .i.xe2 23 dxc5 .i.f3! 24 .i.g5 .i.b7
White has been running around with his 25 tDdS ctJxg5 26 ctJxb7 Wxb7 27 tDc3
queen taking every pawn possible. Of course ':'aeS 0-1
these strong analysts are not beginners, but
sometimes you have a bad year in email Game 3
chess, just as you can have a bad day in nor- Kan-Konstantinopolsky
mal chess. MOSCOlV 1945
14 Wh1
14.i:!.dl? is met strongly by 14... ii.xf2+ 15 1 e4 e5 2 tDf3 ctJc6 3 .i.c4 tDf6 4 tDg5
Wf1 e3 16 dxe3+ cj;e 7 17 ~h3 l:!.xg5 and d5 5 exd5 b5 6 .i.xb5?! 'iYxd5 7 i..xc6+
Black wins. 'ilt'xc6 S 0-0
14... e3 15 f3 e2 16 .l:te1 tDe4 17 ctJf7+!?
White is taking some chances here. 17
fxe4?? was not possible due to 17 ...'iVxh6 18
:tJf7+ ~d7 19 lDxh6 l:!.af8 and Black wins.
But after 17 'iVxc6 lDf2+ 18 'it'g1 lDh3+ 19
'ith 1 Black is forced to take a draw because
of lDe6+. This was perhaps best.
17...WcS 1S 'iYxc6??
A terrible blunder - a very seldom guest in
correspondence games. Instead, after 18
:xe2! lDf2+ 19 .l:!.xf2 .txf2 20 'iVxc6 .txc6
21 lDe5 .i:!.e8 22 lDd3 .td4 Black has very
~od practical compensation for the materiaL
\\bite still needs to develop and his king is Much better than 8 'iVf3.

11
Two Knights Defence

8 ....ltb7 9 'ii'f3 (necessary was 16 .:te2 .l:!.e 7 17 d4! l:lxe2 18

Now this is forced. White cannot live with CLlxe2 i.xd4 19 CLlxd4 ~xd4 20 i.e3 'iYxb2
the pressure on the long diagonal. After 9 21 .l:!.c1 l:ld8 22 'iVg4 with some kind of dy-
CLlf3?! O-O-O!? Black has an excellent attack; namic equality) 16 ...i.xe4 17 ~xd7 i.xc2 18
e.g. 10 d3 e4 11 CLle5 'iVd5 12 d4 e3 13 CLlf3 'iVb5+ i.b6 19 CLle3 J::td8 20 a4 i.d3 21 'iVb3
exf2+ 14 .l:i.xf2 i.d6 and so on. Also possible ~a8 and Black was much better in Reiter-
is 9... i.d6!? 10 d3 e4 11 l:tel 0-0-0 12 CLlh4 Repp, corr. 1986.
h6 13 d4 ~d7 14 f4 CLld5 with an attack, 14 d3! exd3
Wolminkin-Krol, corr. 1961. 14... i.xc3 15 bxc3 exd3 16 cxd3 ~xc3 17
9 f3? is a stupid move. After 9... i.c5+ 10 i.e3 'iVxd3 18 CLlxf7 and with the safer king
'it>hl h6 11 CLlh3 0-0-0 12 d3 g5 Black had a White stands better.
crushing attack in Best-Muir, corr. 1968-69. 15 cxd3 1.lhd8 16 .lte3 .ltxc3 17 bxc3
9 ... e4 10 'ii'b3 0-0-0 11 "tWh3+ ~xc3
White should not be greedy. After 11 If 17 ....u.xd3 18 CLlxf7 :e8 19 nab 1 and
CLlxf7? e3 12 f3 e2 13 .l:!.el Jtc5+ 14 ~hl White is better.
l:lde8 15 d4 (or 15 CLlxh8 i.f2) 15 ... i.xd4 16
'iVd3 i.f2 17 l:lxe2 i.a6! 18 'iVf5+ 'it>b7 19
Itxe8 .l:i.xe8 20 i.d2 ~c4 Black wins.
11 .. .'ii,b8

18 llab1?
This gives up the d-pawn needlessly. Bet-
ter was 18 d41, and although it closes the
white bishop'S diagonal, the knight gains a
Estrin was convinced that Black is better strong outpost after CLlg5-f3-e5. White can
here, but perhaps it is not so clear. then proceed with the attack on the b-flle
124Jc3 and is much better. 18...'it>a8 19 CLlf3 i.xf3
12 CLlxf7? is still bad: after 12 ... l:ld5! 13 20 'iYxf3+ CLld5 does not seem to be a reliable
CLlxh8 l:lh5 14 ~c3 i.c5 15 ~hl e3 16 f3 defence for Black.
CLlg4! mate is coming - in a maximum of 7 18 ... ~xd3 19 ~f3 "tWd5 20 'ilVxd5 4Jxd5
moves according to Fritz 8. 21 .ltd4
12 ... 1.ld7 13 J:!.e1 .ltb4? 21 CLle4 CLlxe3 22 CLlc5 J::td1 23 J::txb7+
The problem with this move is that the 'it>c8 24l:tbl J:i.xe1+ 25 .:txe1 CLlc2 and Black
bishop on b4 is not threatening anything. should win with an extra pawn in the ending.
Black should play for the attack! Therefore 21 ... h6 224Je44Jb6 23 .ltc3
stronger was 13 ...i.c5! 14 CLlgxe4 CLlxe4 15 Not 23 i.xg7? i.xe4 24 .l:.xe4 f5 and
l:lxe4 (if 15 CLlxe4 l:le 7 16 d3 i.b6 17 .l:i.e2 Black wins .
.l:;!he8 with compensation) 15 ...'iVf6 16 CLld1? 23 .....txe4

12
 Introduction and 4 ttJg5 d5 5 exd5 b5

Necessary, otherwise White will install a thing you should know: as this game was
strong knight on c5. played about 60 years ago, there were ad-
24 J:!.xe4 J:!.d1 + 25 l:te1 J:!.xb1 26 J:.xb1 journments after 4 or 6 hours of play, and
the games fmished some hours or sometimes
even days later. So perhaps Black was hoping
to make his advantage last all the way to
move 60, after which he could take it home
and prepare a winning plan without the ten-
sion and pressure of the tournament hall.
42 1:I.a3 J::.d7 43 :a4 :f7 44 Wd1 ttJd445
.l:(a6+
Not 45 ~xd4? ~b5 46 l::ta3 cxd4 and
Black should win.
45 ... Wd5 46 :g6?
Better was 46 ~xg5 l::txf2 47 lIxa 7 lIg2,
although Black has excellent Willrung
In this kind of position the chances are chances.
equal, which means 50% odds for Black 46 .. .'it>e4?
winning, and 50% odds for a draw. But of There was no need to give up the g5-pawn
course if you are Anatoly Karpov or Ulf now. Better was 46 ... tDe6! and Black should
Andersson, you will have a 99% chance of win quickly.
winning the game. 47 J:.xg5 J:.b7 48 We1 a5?
26 .. .f6 27 h4 e5 28 Wf1 We7 29 We2
We6 30 g4!
The right plan. White wants to exchange
all the pawns on the kingside. This follows
the old rule that if you have a worse end-
game, you should try to exchange pawns, and
if you are better you should exchange pieces.
30 .. Jid7 31 g5?
But this is impatient. After 31 lIc1 tDd5
32 ibd2 White retains his equal chances as
before - that is a 25% chance if you want to
be pessimistic.
31 ... hxg5 32 hxg5 ttJd5 33 j,d2 fxg5
Now the position is more or less lost, This gives ~'hite an unexpected chance.
though it is still not easy to win for Black. After 48 ... ~d3 49 l::td5 as Black appears to
34 :g1 ttJf4+ 35 ~d1 ttJe6 36 We1 J:!.d5 be better prepared for the rook endgame -
37 J:!.e1 Wd6 38 l:te3 J:!.f5 39 J:!.d3+ We6 his main threat now is l::tb4 - but the follow-
40 j,e3 J:!.d5 ing line shows this is not true: 50 .txd4 cxd4
Better is 40 ... a5! 41 lIa3 c4 and step by 51 lIxa5 lIf7 52 lIg5 lIxf2 53 lIxg7 l:!.fH 54
step Black wins. ~b2 ;t>e2 55 a4 d3 56 l::te7+ ~d1 57 as l::tf6
41 J:!.b3 J:!.d8 58 Wc3 d2 59 I:!.d7 Wc1 60 Wc4 d1~ 61
Of course a move like this cannot be a l::txdH Wxd1 62 ~b5 ~d2 63 a6 and the
mistake, but it proves that Black has no idea endgame is drawn. The correct move was
how he is going to win the position. One 48 ... J:.d7! maintaining a clear advantage.

13
Two Knights Defence

49.txd4! This variation is now considered dubious.

White correctly evaluates the rook ending The usual continuation 6... ttJd4 7 c3 trans-
as a draw. poses to 5... ttJd4 6 c3 b5 7 ..tf1 covered in
49 ... ~xd4 Chapter 2.
49 ... cxd4 50 l:txa5 .uf7 is met strongly by Another alternative is 6... ttJxd5 7 i.xb5
51 .l:!.g5 lhf2 52 l:txg7 ~c3 (if 52 ....uxa2 53 ..tb7 (7 ... ..td7?! is met strongly with 8 'iVh5!
.l:!.g3! and White makes a draw with the Phili- g6 9 ~f3 'iVxg5 10 'iVxd5 'iif6 11 ttJc3 and
dor position; i.e. after 53 ... d3 54 l:i.g8 the Black is a pawn down without compensa-
black king can no longer hide from the tion) 8 d4! exd4 9 0-0 ..te7 (if 9.. :iif6?! 10
checks) 53 lIc7+ rJtd3 54 a4 and White l:.e1+! ttJde7 11 ~f3 and White is clearly
should draw. better; less clear is 10 ~f3 ttJde7 11 'i'xf6?!
50 ~c2 J:!.f7 51 .l:!,g4+ ~d5 52 l:!.g5+ gxf6 12 ttJe4 0-0-0 13 ttJxf6 'ud6 14 ttJe4
~c6 53 ~b3 J:!.f3+ 54 ~c4 l:!.f4+ 55 :g6 when Black has compensation accord-
~b3 a4+ 56 ~a3 ~b5 57 l:!.xg7 li!.f3+ 58 ing to Obukhov) 10 'iVh5 g6 11 'iih6 'iYd6 12
~b2 a3+ 59 ~b1 J:i.xf2 60 J::tb7+ ~c4 61 'ilg7 'iYf6 13 'ilxf6 i.xf6 14l:te1+ (this is not
.li!.b8 J:th2 a good moment for reaping: after 14 ttJe4
Black can win the a-pawn by 61...,Sf1+ ii..e 7 15 l::i.d 1?! 0-0 16 ..txc6 ..txc6 17 l:hd4
and 62 ....l:!.a1, but this endgame is a book .l:!.ad8 Black has great compensation for the
draw. pawn) 14... ~f8 15 ttJe4 and White is slightly
62l:!.b7 J:!.g2 YO-YO better.

Game 4
Howell-Volzhin
Calcutta 1996

e4 e5 2 tUf3 tUc6 3 .tc4 tUf6 4 tUg5

d5 5 exd5 b5 6 .tf1 !

7 tUxf7!
This move is the refutation. The alterna-
tives are less critical:
a) 7 ttJe6?! only looks dangerous: 7.. .fxe6 8
dxc6 i.c5 9 d4 (if 9 i.d3?! 0-0 10 0-0 ~d4
with a great attack) 9... i..xd4 10 i..d3 0-0 11
0-0 ~d5 12 'iie2 a6 13 c3 ..tb6 and Black
This paradoxical move is the strongest was slightly better in Morozevich-Piket,
here. One point is that it protects g2 so that London (rapid) 1995.
6.. :iixd5? can be answered by 7 ttJc3. An- b) 7 i..xb5? ~xd5 8 ttJc3 ~xg2 9 ~f3
other is that the bishop is not attacked, as ~xf3 10 CDxf3 i..d7 and Black is slightly
after 61Le2 ttJd4. better.
6 ... h6?! c) 7 dxc6 hxg5 8 'ile2 (if 8 d4 ttJg4! 9 h3

14
 Introduction and 4 tDg5 d5 5 exd5 b5

exd4 10 SLxb5 'iVd5 11 'iie2+ SLe6 12 .td3 ttJxf6 17 ~f3 i.g4 18 'iVf1 ttJh5 and Black
ttJf6 with good compensation for the pawn) wins - PilkCivi) 11...~e8 12 lIVg3 'iVxg3 13
8... SLd6 9 d4 (or 9 ~xb5 as 10 ~a4 e4 with hxg3 ~f8 14 i.e2 ttJxf2 15 b4 .tb6 16 Wh2
compensation) 9...l:.h4 10 f4 .tg4 11 ~xb5 ttJg4+ 17 Wh 1 ttJf2+ with a draw.
exf4 and Black had the initiative in Alberny-
Schaller, corr. 1992.
d) 7 ttJf3 (natural) 7...'iixd5 8 ttJc3 'iVe6
and now if 9 SLxb5?! i.b7 10 'iVe2 0-0-0 with
very good compensation, e.g. 11 .i.xc6 ~xc6
12 d3 e4 13 dxe4 .ta6 14 'iVe3 .tc5 15 ttJe5
~d6 16 ttJxf7 'iib6 17 'iYf3 l:td7 18 ttJxh8
ttJg4 with a strong attack in Strange-Aagaard,
Aalborg 1994; while after 9 ttJxb5 'iYe7 10 d4
e4 11 ttJe5 ttJxe5 12 dxe5 'iixe5 13 .i.e3 .tc5
14 .td4 SLxd4 15 'iVxd4 ~xd4 16 ttJxd4 0-0
and the game was equal in Bronstein-G.Flear,
Hastings 1994/95.
7 .. .'J;;xf7 9 ... tDe4
After 9... i.xf2+ 10 ~xf2 ttJe4+ 11 'i&tf1
Ii:f8 12 d3 Wg8+ 13 i.f3 i.g4 14 ~e2 and
White wins - Howell.
100-0

7... 'iVxd5 8 ttJxh8 .tg4 9 f3 .th5 10 ttJc3

~e6 11 .txb5 0-0-0 would be the 19th cen-
tury way to play this line for Black. It is
probably just as unsound as most 19th cen-
tury games, but for a blitz game ... why not!? 10 .....txf2+?
8 dxc6 ..tc5 Black is in a difficult situation here. The
If 8.. :iNd5 9 'iVf3 ~c5 10 'YWb3+ .te6 11 alternatives were:
'iNxb5 ttJ g4 12 ~xc5 SLxc5 13 SLa6 .i.xf2+ a) 1O...'iVh4?! 11 lIVel l:.f8 12 d3 ttJxf2?!
14 ~e2 and White is much better according (better 12... ttJd6 but Black's compensation is
to PilkCivi. gone) 13 .l::txf2+ ~g8 14 SLf3 and White
9..te2! wins.
The best move. After 9 i.xb5?! ttJg4 10 b) 1O...l:!.f8?! 11 d4 .txd4 12 ttJd2 ttJxd2
0-0 'iVh4 White is under heavy attack; e.g. 11 13 i.xd2 .txb2 14 :b1 i.d4 15 .l:i.xb5 and
'iNf3+ (not 11 h3? .txf2+ 12 ~hl We8 13 White is much better here.
'iNf3l:.f8 14 'iVa3l:.f6! 15 SLe2 .td4 16 .l:txf6 c) 10...'iVf6!? is the best try and is consid-

75
Two Knights Defence

ered in Game 5. if'xc6

11 J:txf2 + 4:Jxf2 If 18... aS 19 .tb2 and White is much bet-
ter.
19 4:Je4 if'xc2 20 ..ta3 a5 21 ..tc5..te6
After 21..."iWb2 22 1:1el ius (not
22 .. .'iWxa2? 23 g4! J.xg4 24 ii.xg4 ~xb3 25
'iVh3 and wins) 23 .thS g6 24 .td1 l:td8 25
ctJf2! White has a virtually winning position.
22 'ifd2 ~xd2 23 4:Jxd2 a4 24 bxa4
l:!.xa4 25 a3 J:rd7 26 J:tc1

12 'ifn!
Black had probably counted on 12 'iio>xf2?
~d4+ 13 'iio>f1 1:1f8 14 c3 'iio>g8+ 15 .tf3 ~h4
16 'iio>gl (16 d3 'ii'xh2 17 ctJd2 .te6 18 ~e2
might give White an advantage, but Black is
allowed too much counterplay all the same)
16... e4 17 g3 ~f6 18 .tg2 .tg4 19 'iVe 1 J.f3
and Black has good compensation.
12 .. J:tt8 13 ~xf2+ ~g8 14 ~e3 'tifh4 White is winning as the black pawns are
very weak. For example, if we moved the
black pawns from eS to f6 and bS to b 7, the
position would be slightly better for Black!
26 .....tg4 27 ..tf1!
No exchanges, as they would only grant
the black rooks more freedom on the board.
27 .....tf5 28 l:!.c3 c6 29 ..tb4 J:ra6 30 Wf2
..te6 31 We3 ..td5 32 4:Jf3 J:rf7 33 ..te2
..txf3 34 ..txf3 J:rf6 35 l:tc5 1-0
Black has no counterplay. He decided to
call it a day.

Game 5
15 g3 Leisebein-Grott
More accurate was 15 .txbS! .th3 (if Correspondence 1998
IS ...1:1b8 16 .td3 .l:!b4 17 c4 .tfS 18 J.xfS
1:1xfS 19 d3 and wins) 16 gxh3 1:1fS 17 d3 1 e4 e5 2 4:Jf3 4:Jc6 3 ..tc4 4:Jf6 4 4:Jg5
1:1af8 18 ctJd2 'iVhs 19 'ilt'el ~xh3 20 .tc4+ d5 5 exd5 b5 6 ..tf1! h6?! 7 4:Jxf7! Wxf7
'iio>h8 21 J.dS 1:1f2 22 'iixf2 J::txf2 23 Wxf2 8 dxc6 ..tc5 9 .te2! 4:Je4 10 0-0 'tiff6!?
~xh2+ 24 J.g2 "iHh4+ 25 'It>e2 and White If Black has enough compensation after
wins - Howell. this move then the variations with 6... h6 have
15 .. J!ff6 16 d3 ..th3 17 4:Jd2 J:tf7 18 b3 a right to live. But as I see it, the compensa-

16
 Introduction and 4 !jjg5 d5 5 exd5 b5

tion is only of a practical nature, and with If 15 'iVd5+ .i.e6 16 'iWxc5 .l:!.xg2! 17 f4
accurate play White should keep the advan- .l:tag8 and White has no defence against
tage. 18....l:!.xh2+ and 19 .. :~h4 mate.

11.th5+? 15 ... .l:!.xg2!!

This fares badly. 11 ~el ttJg5 12 ttJc3 Black shows no restraint.
.l:!.b8 13 d3 .i.d4 with some practical chances, 16 >t>xg2?!
but nothing more. On 16 'iWxg2 .l::!.g8 17 'iVf3 .l:!.g4! 18 ttJc3
11...g6 12 ~f3 iLe4 19 'iVxe4 .l:i.xe4 20 ttJxe4 'iVf3+ 21 'itgl
After 12 .i.f3 ttJxf2 13 'iWe2 e4! Black has 'iVxe4 wins because of the threat of ... ~f3
a strong initiative; e.g. 14 ~xb5?! 'iVd4 15 followed by ... h4-h3. In such a position the
.te2 ~e8 and Black wins because of the pawn can become an attacking piece. Or if
threat of ... ttJh3+ and smothered mate, or if 17 'iWd5+ .i.e6 18 ~e4 iLh3 (18 ... .l:!.g4 19 f4
16 h3 .l:i.f8 17 'ith2 .i.d6+ 18 g3 .i.xg3+ 19 'iVg7 20 d4 .i.xd4 21 iLe3 iLxe3 22 ttJc3 is
'itxg3 ~d6+ and White is mated. not so clear) 19 d4 i.xd4 20 'iVd5+ .i.e6 21
12 ... gxh5 13 't!Vxe4 J:!.g8 14 ~h1 ~e4 .l:!.g4 22 f4 l:th4 23 iLe3 .i.xe3 24 f5
Or 14 d3 .i.h3 15 g3 .l:i.ad8 with a power- .i.c4 25 'iWxe3 iLxfl and Black is much bet-
ful attack. tef.
14 ....tf5 15 Vi'f3 16 ... J:!.g8+ 17 >t>h1 e4 18 Vi'c3 .tg4 19
't!Vxf6+ ~xf6 0-1
There is no defence against the mate.

17
Two Knights Defence

Summary
Against S... bS!? White's only chance to fight for the advantage is with 6 iLf1!, after which Black
should probably transpose to Chapter 2 with 6... ttJd4. The alternatives 6... ttJxdS and 6... h6
both seem dubious and should be played at your own risk - without blaming the author of this
book for the consequences, unless you are victorious of course!

1 e4 e5 2 tDf3 tDc6 3 i.c4 tDf6 4 tDg5 d5 5 exd5 b5 (D) 6 i.f1

6 dxc6 bxc4 7 ttJc3 - Game 1
6 iLxbS ~xdS 7 iLxc6+ ~xc6 (D)
8 ~f3- Game 2
80-0- Game 3
6 ... h6
6... ttJd47 c3 - S... ttJd4 (Chapter 2)
7 tDxf7 ~xf7 8 dxc6 i.c5 9 i.e2 tDe4 100-0 (D)
lO ... .ixf2+- Game 4
1O ... ~f6 - Game 5

5 ... b5 7... Wlixc6 100-0

18
 CHAPTER TWO I
Fritz Variation:
4 liJg5 d5 5 exd5 liJd4

1 e4 e5 2 ttJf3 ttJc6 3 il..c4 ttJf6 4 ttJg5 tiJxf3+ 10 gxf3 i.xf3 11 ~xeS+ iLe7 12 tiJc3
d5 5 exd5 ttJd4 'ifd7! (or 12... 0-0 with compensation) 13l:i:el
S... tiJd4 is called the Fritz Variation, (not 13 ~g3? 'ikfS! and there is no defence to
named after Aleksander Fritz (1857-1932), ... tiJhS) 13. .. tiJhS 14 1i'xe7+ 'i¥xe7 15 .l:i.xe7+
who was a German Master. He was a player ~xe7 16 tiJxbS tiJf4 and Black is much bet-
who never recorded great successes, but who ter.
could still beat anyone on a good day. His b) 6 d6? iVxd6 7 iLxt7+ We7 8 iLb3 (if 8
sword drew the blood of Steinitz, Paulsen c3 h6 9 cxd4 hxgS 10 i.b3 exd4) 8.. .'~Jxb3 9
and Mason among others. He wrote an arti- axb3 h6 10 tiJf3 e4 11 tiJg1 ~t7 12 tiJc3
cle about S... tiJd4 in a 1904 issue of the 'iVc6 and Black was much better in Bo-
Deutsche Schachzeitung, and three years later he golubow-Rubinstein, Stockholm 1919.
wrote another article about S... tiJd4 in the c) 6 tiJc3?! h6 7 tiJf3 .tg4 8 .te2 (8 d3
Swedish journal Tidskriftfor Schack. .tb4 9 0-0 0-0 10 a4! [10 .l:i.el bS 11 ~b3
The standard position comes after 6 c3 bS .l:i.e8 12 iLd2 as 13 a4 i.xf3 14 gxf3 tiJxb3 15
7 ~f1!, which can also arise by the move cxb3 bxa4 16 bxa4 tiJxdS and Black is much
order S... bS 6 iLf1! tiJd4 7 c3. The main better] 10...'i¥d6 and Black is slightly better)
move now is 7... tiJxdS, but we are getting 8...iLxf3 9 iLxf3 i.b4 100-00-011 .l::i.el .l:i.e8
ahead of ourselves. 12 a3 iLxc3 13 dxc3 tiJxf3+ 14 'iVxf3 'iVxdS
r------------------" and Black is slightly better according to Gli-
Game 6 gorie.
G.Lee-luldachev 6 ... b5
Gent 2002 Other moves are just weak, e.g. 6... tiJfS? 7
' - - - - - - - - - - - - - -... 'i¥e2!? tiJxdS 8 iVxeS+ tiJfe7 9 tiJe4 f6 10
, e4 e5 2 ttJf3 ttJc6 3 il..c4 ttJf6 4 ttJg5 'i¥g3 tiJfS 11 'i¥f3 and White is a pawn up
d5 5 exd5 ttJd4 with a better position.
Besides S... tiJaS, this is the most serious 7 ~f1!
move. Not 7 iLd3?! i.fS! 8 iLxfS tiJxfS 9 'iff3
6 c3 'ifd7 and Black is slightly better according to
Other moves are worse: ECG. White has also tried 7 cxd4 bxc4 and:
a) 6 O-O? bS 7 iLb3 h6 8 tiJf3 iLg4 9 'i¥el a) 8 1i'a4+ iYd7! (8 ... tiJd7 9 tiJf3 exd4 is

19
Two Knights Defence

less strong because of 10 O-O! - a new idea; cording to Estrin.

after 10 tDxd4 i.c5 Black is better - 10 ... .l1.e7 8 cxd4!?
11 tDxd4 0-0 12 tDc6 tDb6 13 tDxd8 tDxa4 Also possible are 8 h4 and 8 tDe4, as can
14 tDc6 .l1.f6 15 tDc3 tDb6 16 a4 i.d7 17 as be seen later in this chapter. With 8 cxd4
tDxd5 18 tDxd5 ..txc6 19 tDxf6+ gxf6 with White takes the money and runs!
equality) 9 ~xc4 (probably better is 9 ~xd7+ 8 liJxf7?! does not really work: 8...'iitxf7 9
i.xd7 10 dxe5 liJxd5 although Black has cxd4 exd4 10 ~f3+ (if 10 i.xb5 ~e7+ 11
compensation for the pawn) 9.. :iVxd5 10 ~e2 ~xe2+ 12 i.xe2liJb4 13 ii.c4+ 'iitg6 14
~xd5 liJxd5 11 liJc3 tDb4 12 0-0 liJc2 13 0-0 i.fS 15 liJa3 i.d3 16 ii.xd3+ liJxd3 17
tDd5 (if 13 .i:!.bl liJxd4) 13..JIb8! 14 Itbl c6 liJc4 J:.e8 with excellent compensation for
and Black is much better. the pawn) 10...liJf6! and now:
b) 8 dxe5 ~xd5 (attention should also be a) 11 ~xa8? .l1.c5 12 ~c6 (if 12 .l1.xb5
paid to 8...liJxd5!? 9 "iVf3 ~xg5 10 "iVxd5 :e8+ 13 'iitf1 i.a6 14 'it'c6 'ti'e7 15 "iVxe8+
.l:!.b8 11 0-0 ii.b7 12 ~5+ 'it>d8 13 f3 a6 14 ~xe8 16 i.xa6 ~a4 17 ..te2 d3 18 ii.h5+
'ti'xc4 "iVxe5 with compensation in Avtono- liJxh5 19 tDc3 ~4 and Black wins)
mov-Estrin, Moscow 1948, while if 9 ~a4+ 12... ~e7+ 13 'iitdl .l1.d7 14 ~7 c6! 15 ~a6
'ti'd7 10 ~xc4 as! 11 tDc3 liJb4 and Black liJ g4 16 ii.c4+ bxc4 17 :f1 ~e4! 18 ~xc4+
has excellent attacking chances according to i.e6 19 ~a6 (or 19 'it'e2 'it'xg2 20 f3 ~xe2+
Estrin) 9 exf6 (after 9 0-0 .l1.b7 10 'iVf3 ~xf3 21 ~xe2 i.c4+ 22 d3 l:!.e8+ 23 'iitd2 tDe3 24
11 tDxf3 liJd7 and Black is at least equal, or if 'ue 1 .l1.b4+ 25 tDc3 .l1.a6) 19...~xg2 20 b4
9 liJf3 liJd7 10 0-0 .l1.b 7 11 tDc3 ~c6 12l::te 1 liJe5 21 i.a3 d3 and Black wins.
0-0-0 with full compensation) 9 ...~xg5 10 b) 11 .l1.xb5 ..te6 and then:
'iVf3 .l:!.b8 11 ~e3+ (or 11 0-0 ~xf6 12 ~xf6 bl) 12 b3 i.d5 (not 12... .l:!.b8 13 ~e2 i.d6
gxf6 with a fine position) l1...'iVxe3+ 12 dxe3 14 0-0 .l1.xh2+? - Estrin was convinced that
gxf6 and here the two bishops and the half this sacrifice wins, but... 15 ~xh2 liJg4+ 16
open g- and b-ftles give Black suffIcient 'it>gl! ~4 17 ~f3+ 'iitg8 18 ~3 and White
compensation for the bad pawn structure. wins) 13 .l1.c4 i.xc4 14 bxc4 'it'e8+ 15 'iitf1
7 ... tt::lxd5 (weaker is 15 ~e2?! d3 16 ~xe8+ l::txe8+ 17
~f1 .l1.b4! and Black is much better)
15... ~e6 with more than adequate compen-
sation, e.g. 16 ~xa8? ~xc4+ 17 ~e1 ~xc1+
and Black wins.
b2) 12 0-0 'ub8 13 i.a4 i.d5 (13 ... ~d5!?
also gives enough counterplay) 14 'it'e2 i.c5
15 d3 ':f8 (or 15 ... c6 16liJd2 :e8 with com-
pensation - Pilkovi) 16 liJd2 ~g8 17 ii.b3
'iith8 18 i.xd5 ~xd5 19 liJf3 :be8 20 'it'dl
.l1.d6 21 ~3 ~5 and Black's initiative was
more than suffIcient compensation for the
pawn in V.Ivanov-Kobalia, Moscow 1996.
8 .. :i!Vxg5 9 i.xb5+
7... h6? is no good; after 8 cxd4 hxg5 9 This is the most logical move, although
dxe5liJxd5 10 i.xb5+ i.d7 11 i.xd7+ ~xd7 others have also been on the scanner:
12 tDc3 tDf4 13 d4 liJxg2+ 14 ~f1 'iVh3 15 a) 9 ~3? exd4 10 .l1.xb5+ i.d7 11
~a4+ ~d8 16 ~c6 liJh4+ 17 ~el liJf3+ 18 i.xd7+ 'iitxd7 12 0-0 .l1.d6 gives Black excel-
~dl l::tc8 19 i.e3 White is much better ac- lent attacking chances.

20
 Fritz Variation: 4 Ci'Jg5 d5 5 exd5 Ci'Jd4

b) 9 'iVe2 ctJb4 (also interesting is 9 ... a6!? Black does not have enough compensation
10 ctJc3ctJf4 11 "iWxe5+ "iWxe5+ 12 dxe5 ~b7 for the pawns.
13 d4 ctJxg2+ 14 ~xg2 ~xg2 15 .l:!.gl ~b7
with even chances, or if 11 ~e4 .l:!.b8 12 d3
~b 7 13 i..xf4 'iixf4 14 'iixe5+ 'li'xe5+ 15
dxe5 ~c5 Black has compensation on the
dark squares) 10 ctJa3 ~d6 11 dxe5 0-0 12 d4
'li'h4 13 i..e3 c5 14 ctJxb5 cxd4 15 i..xd4
~a6 16 g3 ~xb5 17 gxh4 ~xe2 18 ~xe2
.l:!.fe8 19 f4 f6 and the game was unclear in
Norris-Nishimura, Jakarta 1993.
c) 9 ctJc3 exd4 10 i..xb5+ .td7 11 i..xd7+
Wxd7 12 0-0 ctJf4 13 'iia4+ Wc8 14 'li'e8+
Wb7 15 ~e4+ c6 16 d3 ~xg2+ 17 ~xg2
ctJxg2 18 'iitxg2 dxc3 19 bxc3 gives an equal
position. This is a typical ending for this line, 12 Ci'Jc3?!
as we shall see in the main game. Black is also slightly better after 12 'li'g3?!
9 ... 'ii?d8 10 ~f3 ~xg3 13 hxg3 exd4 or 12 d3?! ~g6 13 ~g3
Or 10 0-0 ~b7 11 "iWf3. exd4 14 .l:!.e 1 .Jtd6 15 'iVxg6 hxg6. White
10 ... ~b7 should play 12 dxe5, as can be seen in Game
7.
12 ... exd4
Black could also try either 12... ctJxc3 13
dxc3 ~xf3 14 i.xg5+ £6 15 gxf3 .l:!.xb5 16
i..el exd4 17 nd1 with equality, or 12...ctJe3!?
13 'iVh3 "iWxg2+ 14 ~xg2 ctJxg2 15 dxe5
ctJh4 (better then 15 ... ctJf4?! 16 .l:!.e1 ctJh3+ 17
'iitfl as 17 ....tc5?! fails to 18 d4! .txd4 19
i..e3 i..xe3 20 l:!.ad1+ ~e7 21 ':xe3 and
White is much better) 16 ~e2ctJf3+ 17 i..xf3
i..xf3 18 d4 nb6 with good compensation
for the pawn. Black's big dream here is to
win the h-pawn and race his own to hI.
" 0-0 13 d3
Not 11 'iVxf7? ctJf6! 12 'iVc4 'iVxg2 13 .l:!.fl 13 .l:!.e1? is best met by 13. .. i..d6! when
a6 14 i..a4 'iVe4+ 15 'iVe2 exd4 and Black is Black is just better. The alternative 13 ... ctJe7
better. 14 'li'h3 c6 seems to favour Black, but then
11 ctJc3 exd4 12 O-O!? J::tb8 transposes to White is forced into 15 ctJe4 'li'xb5 16 ~g3
the game, but note that 12... dxc3 is risky after .l:!.c8 17 ctJd6 ctJf5 18 ctJxf7+ 'it>d7 19 ~g4
13 dxc3 'iWf6 14 'iWg4 'iWe6 (not 14... ~d6? 15 with a strong attack despite the bishop defi-
.l:!.dl c6 16 ~a4 'iitc7 17 c4ctJf6 18 ~g5 ~e7 cit.
19 i..f4+ 'itc8 20 "iWas and White wins) 15 13 ... Ci'Je3
i..g5+ i..e7 16 i..xe7+ "iWxe7 17 lHe1 when Here I think I have an improvement with
White has a very strong attack for the piece. 13...ctJe7! which has not been considered
" ... J:tb8 before. After 14 i.xg5 (if 14 'iVxb 7?! .l:!.xb 7
Not 11...e4? 12 'li'xe4 i..d6 13 .t!.e1 and 15 i..xg5 dxc3 16 ~a6 l:!.xb2 17 .l::i:fel f6 18

27
Two Knights Defence

..ie3 tbd5 and Black is better, or 14 'ii'h3 ~e3 '>t>c6

'ii'f5 15 'ii'xf5 tbxf5 16 tbe4 ..ixe4 17 dxe4 This ending is lost for White.
Mxb5 18 exf5 Mxf5 and Black is a pawn up) 26 l:!.b2 l:!.a3 27 l:!.c2 f5 2S h3 ~e7 29
14... ..ixf3 15 ..ic4 dxc3 16 gxf3 cxb2 17 ~c1 l:!.aS 30 c4 b4 31 ~f3 .lif6 32 .lif4
Mab 1 f6 18 ..iel tbc6 19 ..ixb2 jLd6 it ap- .l:tdS 33 We2 g5 34 ~g3 g4 35 hxg4?!
pears that Black is slightly better. Also play- Losing by force. White would have more
able is 13 ...'ii'e5 14 tbe4 tbf6 15 ..ic4 jLd6 16 chance of survival after 35 h4, though the
tbxd6 ~xd6 17 ~g3 with equality. passive position of the white pieces and the
14.lic6 weakness of the h-pawn (after 35 ... h5) should
Now comes a long forced line. be enough to secure victory for Black.
14 .. :~xg2+ 15 'iYxg2 liJxg2 16 .lixg2 35 .. .fxg4 36 J::tc1 J:l.aS 37 .:tc2 ~a3 3S
.lixg2 17 ~xg2 dxc3 1S bxc3 .lif4 h5 39 ~c1 J:raS 40 .lif4 J:US 41 ~e3
h4 42 l:td2 ~c3 43 J:l.c2 h3 44 '>t>f1 l:tdS
45 J::tc1 J:l.xd3 46 '>t>e2 J:l.d7 47 J:tg1 h24S
J:!.h1 .lie5 49 f4 gxf3+ 50 '>t>xf3 .l:ta7 51
We4 .lid6 0-1

Camel
Sermek -Olarasu
Nova Corica 2002

1 e4 e5 2 liJf3 liJc6 3 .lic4 liJf6 4 liJg5

d5 5 exd5 liJd4 6 c3 b5 7 ~f1 liJxd5 S
cxd4 ~xg5 9 i..xb5+ ~dS 10 ~f3 i..b7
11 0-0 .l:tbS 12 dxe5!
As noted earlier, this is a typical ending for
this line and chances are about equaL How-
ever, White needs to show more caution
because of the weakness of his kingside.
1S ....lid6 19 ~e3 a6 20 .!:tab1 ~d7 21
.lia7?
It looks innocent, but actually this is the
decisive mistake! Now White cannot avoid
Mxb5 axb5, after which he has a lost end-
game due to the weakness of his a- and h-
pawns. 21 a4 with an even endgame was
clearly better. Then White could start think-
ing about 22 jLa 7.
21 ... l:!.b5! This is the critical position for 8 cxd4.
Black does not give up the open flle. Black has a wide range of discovered attacks.
22l:!.xb5 12 ... liJb4?!
White has no good options here. If 22 c4 This is not the best. Also weak is
J::tg5+ 23 ~f3 jLxh2 and Black is clear pawn 12 ... tbf4?! 13'iUg3 ~xe5 14 d4 tbh3+ 15
up, or 22 jLe3 llhb8 23 nxb5 axb5 and 'iYxh3 'iYxb5 16 tbc3 'iWb6 17 jLf4 and White
Black is much better. is much better.
22 ... axb5 23 ~d4 1::!.aS 24 J:tb1 c5 25 Black should play 12 ... tbe3! 13 'iUh3

22
 Fritz Variation: 4 tDg5 d5 5 exd5 tDd4

'iVxg2+ 14 'iWxg2 ttJxg2 15 d4 ttJh4 (not

15 ... ii.e7?! 16 ii.e2 ttJh4 17 f4 when Black
has no compensation for the pawn; e.g.
17 ... ttJf5 18 ttJc3 ttJxd4 19 .l:.d 1 c5 20 ii.e3
'ite8 21 ii.xd4 cxd4 22 ttJb5 g5 23 f5 f6 24
ttJxd4 ii.c5 25 l:tac1 ii.b6 26 e6 and White is
much better) 16 iL.g5+ iL.e7 17 ii.xh4 (after
17 ii.xe7+? r:j;xe7 and Black is already better)
17 ... iL.xh4 18 ttJc3. So far Wiech-J edynak,
Warsaw 1994, and now Black should have
continued 18... ii.f3! 19 b3 (19 ii.e2 ii.xe2 20
ttJxe2 l:hb2 21 l::tabl l::!.b6 is fIne for Black)
19 ...l:tb6 20 iL.d3
White gives up this pawn in order to open
the d-ftle towards the black king. Black is in
trouble here. If instead 14 'iVg3? 'iVxg3 15
fxg3 ttJc2 16 .tg5+ j"e7 17 ii.xe 7+ r:j;xe 7 18
ttJc3 ttJxa 1 19 .l:.xa 1 f6 20 exf6+ r:j;xf6 21
l::!.f1+ <J;;e7 22 l::te1+ ~d6 and with the d4-
pawn as a weakness it is White who will fIght
for a draw.
14.•..txd5 15 J::!.d 1 ~xb5 16 tDc3 ~c6 17
~f5!?
White decides to keep up the pressure.
White is also better after the simple 17 ttJxb5
'iVxb5 18 'iVxf7 <J;;c819 "iVf5+r:j;b7 20 a3.
20 ... l::!.g6+!! (not 20 ... r:j;e7?! 21 l::!.ac1 l::!.d8 1 7 ... h6 18 a3 I!.a5 19 l:.b 1 ~e6 20 ~h5
22 ttJb5 .l:!.b7 23 l::!.c4 and White is much 'it>c8 21 axb4 .i.xb4 22 tDxd5 l:txd5 23
better) 21 ii.xg6 hxg6 22 l::!.fel ii.g5 23 e6 f5 J:!:.xd5 ~xd5 24 ~g4+ ~d7 25 e6!
24 e7+ (or 24 ttJe2 ii.d2 25 l::!.edl i.g5 26 Again White sacrifIces a pawn for a strong
ttJ g3 ii.xdl 27 ~xdl cJ;;e7 28 d5 l:td8 29 ttJe2 attack on the enemy king. After 25 'iWxd7+?
i.f6 30 ttJf4 g5 31 ttJg6+ rJid6 and Black cJ;;xd7 26 ii.e3 the position is just equal.
should draw) 24...cJ;;e8 25 .l:!.e6 i.f4 26 .l:!.xg6 25 ... fxe6 26 .i.e3 .i.d6 27 f!.a1
l::!.xh2 27 r:j;f1 r:j;xe 7 28 ttJe2 ii.h6 29 J::tg3 (if
29 ttJg 3? r:j;f7 30 l:ta6 f4 31 ttJe2 j"g4! and
White has problems) 29 ... ii.e4 and Black has
compensation here; for example 30 f4 (if 30
f3 j"d3 31 f4 l::!.h1+ 32 l::!.gl l::!.h2 33 l::!.el
j"xf4 34 l::!.xg7+ r:j;f6 and White will have to
tight for survival) 30 ...l::!.h1+ 31 l::!.gl l::!.h2 32
l::!.el ~f6 and the bishop on e4 is worth more
than the white rook. Tarrasch talked about
two bishops equalling rook and knight. Here
the question is whether or not the rook and
knight equal the two bishops.
13 d4 ~g6 14 d5!

23
Two Knights Defence

Black's position is more or less lost. ues actively. All authors are convinced that
27...e5 12... ii.d6 13 ii.xb5 .ue8+ 14 'it'f1 ii.a6 15
If 27 ... c5 28 b4 and the attack should 'Yi'c6 ~e7 wins for Black; but after the simple
crash through. 16 g3! we see how cruel life is: White wins.
28 ~e4 'it>d8 29 ':'xa7 c3;;e7? Much stronger is 12 ... ii.c5! 13 ii.xb5 (if 13
Necessary was 29 ...J::i.f8 and maybe Black d3 'lie7+ 14 ~d2 i..b4+ 15 tDc3 'lic5 and
can still fight. Black is "wning, l\fichalczak-Hermann,
30 ~g6 J::!.g8 31 .i.xh6 'tWe6 32 .i.g5+ Cuxhaven 1994) 13.. .'Yi'e7+ 14 ~f1 ii.a6 15
'it>d7 33 'iie4 'tWb3 34 g3! ~c6 J::i.e8 when 16 g3 is answered by 16 ... d3!
This is worth remembering: when your 17 'lif3 ~e2+ 18 'iixe2 dxe2+ 19 ~g2 i..xb5
position is completely winning, please do not and Black has a very strong attack for the
allow your opponent to mate you! exchange. It is not clear that White can hold,
34 .. :~d1 + 35 c3;;g2 l:!.f8 36 it.e3 rJ;;e7 37 e.g. 20 J::i.el g5 21 tDc3 i..c6+ 22 ~gl gxh4
~e5 rJ;;d7 38 .i.xd6 'i'xd6 39 'iYg4+ ct2e7 23 .uxe2 .uxe2 24 tDxe2 ii.xf2+!! 25 ~h2
40 'iYxg7 + l:tf7 41 'iVg5+ c3;;d7 42 'iYg4+ .i.f3 26 gxh4 i..xe2 and Black wins.
rJ;;e7 43 'iYe4 1-0 9 ... tUe6
Compared with 8 tDe4, Black obviously
Game 8 cannot consider 9...'iVh4 here.
Chandler-P. Littlewood 10 .i.xb5+ ~d7 11 ~a4
London 1996 11 ii.xd7+? 'lixd7 12 0-0 ii.e7 gives Black
an easy game; for example 13 h5 f5 and
e4 e5 2 tUf3 tUe6 3 ~e4 tUf6 4 tUg5 Black is better.
d5 5 exd5 b5 6 .i.f1 tUd4 7 e3 tUxd5 8 11...tUdf4
h4!? Black needs to play actively. After
11...ii.e7 12 ~xd7+ ~xd7 13 'lixd7+ ~xd7
14 f3 tDdf4 15 g3 tDd3+ 16 ~e2 tDxc 1+ 17
J::i.xc1 J::i.hb8!? there is insufficient compensa-
tion for the pawn.
12 d4

White protects the knight, which is not

such an unnatural idea.
8 ... h6
Black needs to be persistent. After
8... tDf5? 9 i..xb5+ i..d7 10 ii.c4 White is
much better. 12 ... f5?
9 tUe4 Here Black is too optimistic. Attacks are
Dubious is 9 tDxf7 ~xf7 10 cxd4 exd4 11 usually better performed with pieces than
~f3+ tDf6 12 ~xa8 as long as Black contin- pawns. Especially if the pieces are on the

24
 Fritz Variation: 4 0,g5 d5 5 exd5 0,d4

back rank! better.

Better was 12... ctJxg2+ 13 ~f1 ctJgf4 14 22 cxd4 g6 23 0,h4 cxd4 24 0,xg6 ':'g8
~xf4 ctJxf4 15 dxe5 ctJd3 16 ctJbd2!? (or 16 250,e5+!
.itxd7+ ~xd7 17 ~xd7+ ~xd7 18 ~e2 This knight belongs on d3 as a blockader.
ctJxe5 and the position is equal - P:ilki:ivi) After 25 ctJxf8+? .I;lbxfS 26 .l:!.e1 (not 26
16 ...ctJxe5 17 .l:!.e1 ~e7 18 ..ixd7+ ~xd7 19 ctJd2? ctJg3+ 27 ~e1 l:te8+ 28 ~d1 ctJe2 and
~xd7+ ~xd7 with equality. Black is much better) 26 ... d3 gives Black
12... exd4?! is less convincing: 13 ~xf4 some chances.
ctJxf4 14 cxd4 ctJxg2+ 15 ~f1 ctJf4 16 ctJbc3 25 .. .'it>d6 26 0,d3! .i..e7 27 0,d2
iLe7 17 ctJc5 ~xb5+ 18 'iUxb5+ ~fS 19 White has a pawn more and a better posi-
ctJd7+ ~g8 20 'iUf5 provides White with a tion.
dangerous initiative. 27 ... <.t>d5 28 l:!.c7 ~g5 29 0,f3 ':'g7 30
13 0,g3 0,d3+ 14 ~e2 0,xc1 + 15 J:.xc1 J:tc4 ..tf6 31 l:!.c5+ ~e4 32 :!:I.e 1 + ~xd3
exd4 1 6 0,xf5 J:Ib8 33 J:te2 1-0
16 ... d3+!? 17 ~f1 .l:!.b8 18 ..ixd7+ 'iVxd7
19 'iVe4 ~f7 20 ctJd2 and White is much Game 9
better, because 20 ... .l:!.xb210ses to 21 ctJc4!. Narciso Dublan-Kuzmin
17 ~xd7 + 'it'xd7 18 'it'xd7 + ~xd7 19 Balaguer 1997
b3! c5 20 h5
It is always useful to fIx the black pawns e4 e5 2 0,f3 0,c6 3 ..tc4 0,f6 4 0,g5
on the dark squares in such a position. Then d5 5 exd5 0,d4 6 c3 b5 7 ..tf1 0,xd5 8
g 7 and h6 are potential weaknesses later in 0,e4
the game.

8 ctJe4 is the most popular move, but

20 ... 0,f4+? White cannot count on an advantage.
Black is so irritated by the h-pawn that he 8 ... 0,e6!
decides to exchange it immediately, but this This move is suffIcient for equality. The
gives White good time to develop. Better was ultra sharp, but also dubious, 8...'iVh4?! can
20 ... ..ie7 21 l:td1 ..if6 22 ctJa3 ~he8 and be found in Game 10.
Black has some drawing chances. 9 .i..xb5+ ..td7 10 ..txd7+
21 ~f1 0,xh5 Best. 10 'iVa4?! is problematic due to the
Better was perhaps 21...d3, although after weakness of the d3-square. Black now has:
22 l:td1 l:td8 23 ctJa3 J:tg8 24 ctJc4 ~e6 25 a) 10 ... f5!? 11 ctJg3 ctJc5 12 ..ixd7+ 'iVxd7
ctJg3 g6 26 J:te 1+ ~f7 27 l:te3! White is much 13 'iVxd7+ ~xd7 14 d4 exd4 15 ctJxf5 J:te8+

25
Two Knights Defence

16 ~d 1 tLld3 17 l:tfl c5 with enough com- c) l1...f5?! is too early. 12 tLlg3 g6 13 d4

pensation for the pawn, Zaitsev-Mohrlok, exd4 14 cxd4 ~g7 15 tLle2 l::td8 16 tLlbc3 0-0
corr.1959. 17 l::te I! and it is diffIcult for Black to prove
b) 10... tLldf4 11 0-0 (if 11 d4 tLlxg2+ 12 any compensation here.
~fl tLlgf4 13 ~xf4 tLlxf4 14 tLlbd2 .:tb8 15 12 d4 exd4 13 cxd4
~xd7+ 'iVxd7 16 'iVxd7+ 'ii;>xd7 and Black is
slightly better) l1...tLld3! with excellent com-
pensation. The knight will stay on d3 forever.
10 .. :ilfxd7

13 ... 0-0
Black can also try 13 ... tLlb6!? 14 i.e3l::td8
15 tLlbc3 (if 15 l:!:e1 tLlxd4 16 'iVxd4 'iVxd4
17 i.xd4 lhd4 18 tLlg3 l::td7 is roughly
11 0-0 equal) 15 ... tLlxd4 16 i.xd4 'iVxd4 17 'iVe2!?
White also has 11 d4 exd4 12 cxd4 (if 12 (or 17 ~xd4 l::txd4 with equality) 17...0-0 18
0-0 dxc3 13 tLlbxc3 tLlxc3 14 tLlxc3 'iVxdl 15 .:tac1 .l:i.fe8 19 tLlb5 'iVe5 20 tLlec3 'iVxe2 21
l:txdl ~c5 with an equal game) 12 ... tLlb4 13 tLlxe2 i.g5 22 .l:!.c2 c5 23 tLlxa7 .l:!.a8 24 tLlb5
a3 (13 O-O? tLlxd4 14 .l:!.e 1 looks tempting at l::txa2 and the endgame is obviously level,
fIrst, but after 14... 0-0-0! 15 tLla3 tLldc2 Black Pilgaard-Biro, Budapest 2003.
is much better) 13 ...'iVxd4 14 'iVa4+ c6 (Keres 14 l2Jbc3 l:i.fd8!
was convinced that this position is much The best move. After 14... l::tad8?! 15 ~e3
better for Black; but to err is human!) 15 0-0 f5 16 tLlxd5 'iVxd5 17 tLlc3 gives White some
'iVxe4 16 axb4 l:tc8 17 tLlc3 'iVxb4 18 'iVxb4 advantage: 17... 'iVc4 (17...'iWb7? 18 'iWe2 and
~xb4 19l:txa7 ~c5 20 l:ta5 r:J;;e7 with equal- White is much better) and now, rather than
ity. It is still not a bad line to play for Black, 18 'iVb3 'iVxb3 19 axb3 f4 20 ~c1 tLlxd4 21
of course. .l:!.xa7 l:td7 with an equal position, 18 d5! sets
11 ... i..e7 Black has some problems.
This developing move is the only really 1 5 i..e3 l2Jxc3 16 bxc3 f5 17 l2Jc5 i..xc5
logical move here. All the alternatives are 18 dxc5 f4
faulty: Or 18 ...'iVc6 19 'iVh5 f4 20 ~d4 tLlxd4 21
a) 11 ... tLldf4? 12 d4! .l:!.b8 (12 ... exd4? 13 cxd4 l:txd4 22 l:tad 1 l::tc4 23 .l:!.d5 l:te8 24
.txf4 tLlxf4 14 'iV f3 tLld5 15 tLlg5 and White l:tfdll:tce4 with dynamic equality.
wins) 13 l:te 1 .te 7 14 ~ f3 and White is 19 i..d4
much better. 19 'iWxd7 l:txd7 20 .tc1 is also possible.
b) l1...c5? 12 d4 cxd4 13 cxd4 exd4 14 but White can hardly hope that the extra
tLlg5 tLldc7 15 ~f3 with initiative for White pawn will generate an advantage. After
according to Estrin. 20 ... ~f7 (20 ... .:td3!?) 21 l:tb1 l:tad8 22 l::tb7

26
 Fritz Variation: 4 0,g5 d5 5 exd5 0,d4

.l:!.d 1 23 .ta3 l:. 1d2 24 l:!.xa 7 l:!.xa2 25 .l:!.a5 LDf4 12 0-0 ~xd4 13 .tf3 e4 14 .ig4+ ~b8
l:.dd2 26 .iel .l:!.xa5 27 .txd2 ~xc5 the posi- 15 LDf5 ~g5 16 LDxd4 h5 17 d3 hxg4 18
tion is equa1. .ixf4 "iVxf4 19 g3 'Yi'h6 20 h4 g5 21 dxe4
19 ... 0,xd4 20 cxd4 ~xd4 21 ~b3+ gxh4 22 LDf5 and White won in Wernst-
21 ~f3 'iWxc5 22 ~xf4 .l:!.d4 23 .l:i.ael ~d6 Huizmann, Limhamn 1978.
24 ~f3 .l:i.b8 is equa1. 10f3e4!?
21 .. .'~;>h8 22 J:lac1 as Black needs to go all the way. After
This looks a bit suspicious. Better was the 10 ... LDfS 11 .ixb5+ ~d8 12 0-0 .ic5+ 13 d4
natural 22 ... ~ab8. exd4 14 LDe4! White was clearly better in
23 J:rfd1 ~b4 24 ':'xd8+ J:!.xd8 25 'iYxb4 R.Webb-Lees, England 1977.
axb4 26 ~f1 g5 27 J:!.c4 l:td5 28 h4 h6 11 cxd4 ~d6 12 ~xb5+ ~d8
29 hxg5
Also after 29 ~e2 Wg7 30 Wf3 ~f6 31
':xb4l:!.xc5 Black will keep the equilibrium.
29 ... hxg5 30 c6 .l:.d6 31 .l:!.xb4 .l:!.xc6 32
nbS J:lc1 + 33 ~e2 J:!.c2+ 34 ~f3 g4+
35 ~xf4 J:!.xf2+ 36 'iiig3 ':'xa2 37 .l:!.c5
%-%

Game 10
Pilgaard-N .Pedersen
Danish Championship, Greve 2002

1 e4 e5 2 0,f3 0,c6 3 .i.c4 0,f6 4 0,g5

d5 5 exd5 0,d4 6 c3 b5 7 .i.f1 0,xd5 8 13 't'ib3!
0,e4 iVh4?! 13 0-0 is ECO's recommendation. They
claim that the position is unclear. In myopin-
ion the position is virtually winning for
White, but you need to show that you know
the position better than your opponent, and
that you can calculate very well. In this re-
spect it can be compared to the Dragon: very
dangerous, but also dubious. And while
dragons might be dangerous, no one cries
when they die.
After 13... exf3 we have:
a) 14 .l:!.xf3 .l:!.b8 15 a4 a6! 16 ..tfl (16
.txa6?! .l:!.e8 17 LDc3 ..txf3 18 ~xf3 ~xd4+
19 ~f2? .l:!.e1+ 20 .1i.fl ii.c5 and Black is
This is the famous Berliner variation and better, Nordenbxk-Pedersen, Denmark
the most common move here, but it appears 1994; 17 ... LDf6!? is also possible) 16...l:!.e8
at the moment that it does not provide the (16 ... .l:!.b4!?) 17 LDc3 LDf6!? 18 d3 .ixf3 19
comfort of equality for Black. ~xf3 ~xd4+ 20 ~hl LDg4 21 LDce4! with a
9 0,g3 .i.g4 total mess.
9... .tb7? is sharp and good - but only in b) 14 'iVb3! is still the better move, when
blitz games. After 10 cxd4 0-0-0 11 .ie2! Black must choose between:

27
Two Knights Defence

b1) 14.. .'~Jb4 15 l:!.xf3 c6 (or 1S ....t!.b8 16 17 ... gxh1'ilV+

CDa3 c6 17 :e3) 16 Mxf7 (16 l::te3 is also 17...~xg3 18 hxg3 ~xdS was later dis-
strong- Westlund) 16 ... cxbS 17CDc3l:te818 cussed as possible improvement on the
CDxbS ~e6 19 ~c3! (19 CDxd6?! .lixb3 20 game. But it is hard to believe that Black
CDb 7+ with a draw by perpetual check in should have enough compensation for the
Brower-Hodges, corr. 1992-94; sometimes piece here if White develops soundly; e.g. 19
beauty is a horrible attraction) 19... i.xf7 20 .l:i.g1 :e8 20 CDc3 i.f3+ 21 'iitc2 nb8 22 d3
CDxd6 ne1+ 21 ~f2 l'ote7 22 d3 as 23 ~gl followed by ~f4 or CDe4.
~g4 24 a3 ~c7 2SCDxf7+ l:hf7 26 ~e3CDdS 18 .txh1 'iYxg3 19 hxg3
27 ~ c6 and White wins. White is much better. Black has no real
b2) 14... fxg2 15 lIxf7! (not 15 .l:tf2? l:tb8! compensation for the material deficit.
16 ~xdS l:txbS! 17 ~xbS l:te8 with an 19 ... J:.b8 20 d3 hS 21 ~e2 f6 22 dS
enormous attack - Pliester) 1S ... i.e6 16 i.g4 23 i.e3 h4 24 gxh4 J:!.xh4 2S ttJd2
l:td7+! i.xd7 17 'iVxdS :tb8 18 i.xd7 i.xg3 ~d7 26 i.e4?!
19 i.h3+ i.d6 20 ~f5 and White was much Clearer was 26 .lixa7 l:!bh8 27 .lig2! l:!h2
better in Schiiler-Leisebein, caIT. 1998. 28 l:tg1 1:.a8 29 .lics fS 30 CDf1 Mh4 31 CDe3
13 ....txg3+ and White wins.
Black has no choice. After 13 ... CDb4? 14 26 ... fS 27 .tg2 J:.h2?!
fxg4 i.xg3+ 15 ~d1 i.d6 16 h3 c6 17 i.e2 Black has more practical chances after
White is a piece up for nothing, or 13...i.e6? 27 ...l:te8 28 i.xa7 .l:i.h2 29 ~f1 gS 30 a4 'iitd6
14 fxe4CDb4 15 dS .lixg3+ 16 "iixg3 'iYxe4+ 31 ~g 1, though White should still win.
17 ~f1 i.xdS 18 d3 and White wins. 28 J:.g1 J:tg8 29 ~e3 gS 30 ttJe4 i.e2 31
14 ~d1 i.e6 1S .te6! exf3 ttJeS+ ~e8 32 i.xa7 f4 33 i.e4! g4 34
Black can also try 1S ... CDe7 16 dS! CDxc6 i.d4 .l:!.hS 3S ttJg6
17 dxe6 CDeS 18 'i'dS+ ~e7 19 WUxe4! (not Now Black loses material and the game.
19 b3?! exf3 20 i.a3+ Wf6! 21 ~b2 "iVhS!
and the position has started to become un-
clear) 19 ... fxe6 20 CDc3 l:i.hd8 21 WUxh4+
i.xh4 22 b3 CDd3 23 .lia3+ 'it>f7 24 ~c2 and
White is much better according to palkovi.
16 .txdS fxg2 17 'iYxg3

3S .. Jhg6 36 i.xg6 .l:!.gS 37 i.e4 g3 38

i.f6 .l:!.hS 39 .td4 J:lgS 40 ~d2 f3 41
We3 f2 42 ~xe2 fxg 1N + 43 i.xg 1 ~d7
44 ~f1 ~d6 4S .te3 :g4 46 ~g2 ~e5
47 ~f3 J!tg7 48 i.f4+ ~d4 49 i..xg3 eS
SO dxe6 1-0

28
 Fritz Variation: 4 0,g5 d5 5 exd5 0,d4

Summary
The Fritz Variation is still alive and kicking as a serious alternative to theory's darling 5.. .'~Ja5.
But only if, after 6 c3 b5 7 i.f1 tbxd5 S tbe4, Black follows Game 9 (S ... tbe6), and not Game
10 (s .. :iih4) where someone needs to introduce a serious new idea to be back in business.

1 e4 e5 2 0,f3 0,c6 3 ~c4 0,f6 4 0,g5 d5 5 exd5 0,d4 6 c3 b5 7 ~f1 (D) 0,xd5
S cxd4 ~xg5 9 i.xb5+ ~dS 10 'iWf3 i.b7 11 0-0 J::tbS (D)
12 tbc3 - Game 6
12 dxe5 - Game 7
S h4-Game8
8 0,e4 (D)
S... tbe6 - Game 9
S.. .'ti'h4 - Game 10

7 ~f1 11 ... 'JJ.b8 80,e4

29
 CHAPTER THREE I
4 ctJgS dS S exdS ctJaS:
Introduction

1 e4 e5 2 lLlf3 l1Jc6 3 i..c4 lLlf6 4 lLlg5

d5 5 exd5 lLla5
In this chapter we shall look at the minor
lines connected to the absolute main line of
the Two Knights. 5... 4Ja5 is the main move
here, and one to which we shall be dedicating
three chapters. Over time it has become clear
that 6 i..b5+ is the only serious move here.
After that Black has 6... i.d7!? (Games 12 &
13). The main move is 6 ... c6 7 dxc6 bxc6
and then 8 .lte2 is the subject of Chapter 4,
but 8 ~f3?! (Games 14 & 15) has also been
played a lot. 8... h6! (Game 15) is the strong-
est reply, guaranteeing Black a great game. 8~e2
After this White manages to keep an extra
Game 11 pawn, but not equality. Black has no prob-
Rudnick-Pichler lems in the position. White has also tried:
Correspondence 1985 a) S dxe4? (an impressive move invented
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _.. by David Bronstein - but he played it only
1 e4 e5 2 lLlf3 lLlc6 3 i..c4 lLlf6 4 lLlg5 once) S...lLJxc4 9 ~d4 and now D.Bronstein-
d5 5 exd5lLla5 6 d3?! Rojahn, Moscow Olympiad 1956, continued
Also dubious is the rare 6 b3?! h6 7 lLJf3 9... lLJb6?1 10 c4 c5? (Black is slightly better
e4 SlLJe5 a6! 9 a44Jxc4 10 bxc4 c6! 11 4Jc3 after 10....lte7 11 e5 4Jfxd5 12 cxd5 'iVxd5 13
.ltd6 12 d4 exd3 13 4Jxd3 cxd5 14 4Jxd5 'iVxd5 4Jxd5) 11 'iHd3 when White proved
4Jxd5 15 cxd5 'li'a5+ 16 ~d2 'iHxd5 and that the pawn centre offered good compen-
Black is better. Perhaps all other moves than sation for the piece. Stronger is 9...4Jd6! 10
6 .ltb5+ are mistakes! 4Jc3 lLJfxe4! (1O ... c6 also gave Black the bet-
6 ... h6 7 lLlf3 e4! ter game in L.Bronstein-Rai, Mar del Plata
This aggressive move is the trouble with 6 1969) 11 4Jxe4 'fie7 12 0-0 4Jxe4 13 ~el £5
d3. 14 4Jd2 ~c5 and Black has a clear advantage

30
 4 t'iJg5 d5 5 exd5 t'iJa5: Introduction

(Euwe), e.g. 15 'iVe5+ 'It>d8 16lbxe4 fxe4 17 ..ITLg3lbh5! 16lbxc5 (or 16 gxf3 exf3 17 'It>hl
l::txe4 (or 17 'iVxe4 'iVe7! with exchange of lbxg3+ 18 fxg3 f2 19 .t!.xf2 l::tae8 and Black
queens) 17... .ii.d7 18 ~e3'iVd6 19 ~d4 .ii.f5 wins) 16 ... lbf4 17lbxe4
and Black wins.
b) 8 lbd4 c6 9 lbc3 a6!? (simpler is
9... ~e7 10 i.e3 .ii.g4 11 'iVd2 exd3 12 .ii.xd3
lbxd5 13 lbxd5 ~xd5 and the position is
equal) 10 a3! cxd5 11 i.a2 i.g4 12 lbde2
lbc6 13 h3 .ii.h5 (if 13... .ii.xe2 14 'iVxe2 lbd4
15 ~dl and with the two bishops White is
slightly better) 14 g4 exd3 15 cxd3 d4! 16
gxh5 dxc3 17 lbxc3 .ii.c5 18 l:!.g I! and White
has the initiative, though Black is also alive
after 18 ... 'lt>f8!. This position is hard to judge.
S ... t'iJxc4 9 dxc4 il..c5!

17...~h3!! 0-1 Field-Tenner, USA 1923.

d) 10 lbfd2?! 0-0 11 lbb3 ~g4 12 'iVfl
.ii.b4+! 13 c3 (weakening d3, but if 13 lbc3
c6! with terrific compensation) 13 ... .ii.e7 14
h3 ..ITLh5 15 i.e3 lbd7 16 g4 i.g6 17 lbld2
lbe5 18 0-0-0 b5 19 cxb5 lbd3+ 20 'It>b 1
'iVxd5 21 c4 ~e6 22 lbd4 'iVe5 with a huge
advantage for Black according to Keres.
10 ... 0-0 11 t'iJh2 c6
An interesting alternative was 11...e3!? 12
i.xe3 i.xe3 13 fxe3 lbe4 14 lbfl! (if 14 0-0
lb g3 15 ~d3 lbxfl 16 lbxfl 'ikg5 17 'It>hl
Black develops and prevents lbf3-d4. In- ..ITLfS and Black is slightly better according to
ferior are both 9... .ii.g4?! 10 h3 .ii.h5 11 g4 Keres) 14".'iVh4+ 15 g3'iVf6 16 c3 .ii.f5 with
.ii.g6 12 lbc3 .ii.b4 13 i.f4 0-0 14 lbe5 and compensation for the pawns in Korchnoi-
9... .ii.e7?! 10 lbd4 c6 11 lbc3 0-0 120-0 cxd5 Sliwa, Bucharest 1954.
13 cxd5 ..ITLg4 14 1Wb5 and White is much 1 2 dxc6 e3 13 il..xe3 il..xe3 14 fxe3 t'iJe4
better according to Keres.
10 h3
White has no alternative that includes any-
thing remotely resembling a survival kit
a) 10 .i.f4? 0-0 11 lbfd2 .ii.g4 12 ~fl c6!
and Black's lead in development is decisive.
b) 10 c3? b5! 11 b4.ii.e7 12lbfd2 .i.g4 13
f3 exf3 14 gxf3 i.h5 15 cxb5 0-0 and Black
was much better in Grob-Keres, Dresden
1936.
c) 10 O-O?! 0-0 I1lbfd2 .ii.g4 12 ~el ~d7
13 lbb3? (but if 13 lbc3 Uae8 with huge
compensation) 13 .....Itf3! 14 i.f4 ~g4 15

31
Two Knights Defence

1S 0-0 Rash. One thing you should never do in a

If 15 .l:i.g1? bxc6 16ctJf3 'iVf6 17 c3 .l:i.b8 worse endgame is to create weaknesses in
and the black attack is worth more than two your own pawn structure. If you cannot gen-
pawns. Maybe the best try was 15 ctJfl! erate realistic counterplay, it is better to wait
~h4+ 16 g3 'iYf6 when Black has compensa- and see what the opponent has to offer. An
tion for the material according to ECO. Let important point is that it can sometimes be
us try to elaborate a bit on this: 17 c3 ~xc6! more difficult for the opponent to win the
(stronger than 17 ... bxc6 18 ctJbd2 itf5 19 position, than for you to draw it. So why not
ctJxe4 ~xe4 20 l:th2 l:tad8 with compensa- let him do the work? Here White should
tion) 18 ~f3 ~6 19 b4ctJg5 20 ~g2 ~d7 have played 24 ctJc5 l1e 7 25 ~f2 with a
and White's development is pathetic, or if 17 worse but playable position.
ctJbd2! ~xb2 18 ~bl 'iYxa2! 19 cxb7 .l:i.b8 24 ... fxg4 2S hxg4 J::tf8 26 ltJd2 ~df7 27
and Black is better. l:!.f1 .i:.xf1 28 ltJxf1 .txc2 29 ltJc3 .td3+
1S ... ltJg3 16 'ilfd2 0-1
If 16 ~f3ctJxfl 17ctJxfl 'ikb6 18 b3 bxc6 There is no sense in playing on in a posi-
and Black is better. tion like this in correspondence chess.
16 ... ltJxf1 17 ltJxf1 bxc6
Even stronger was 17 ... ~6! 18 cxb7 (or Gamet2
18 b3 bxc6) 18... ~xb 7 19 b3 .l:tad8 20 ~e2 Short-Hector
~g6 21 ctJa3 f5 and White is under heavy Lanzarote 2003
attack.
18 'ilYxd8 J:!.xd8 1 e4 eS 2 ltJf3 ltJc6 3 .tc4 ltJf6 4 ltJg5
If the black pieces are played by a Master dS S exdSltJaS 6 .tbS+ .td7!?
or Grandmaster, then I think this position is
more or less won. White has more material
(2ctJ+28 for l:i.+~), but he also has a weak-
ness on e3 and no good squares for the
knights. White can fight for a draw, but it is
very hard work, and probably unrewarding
too.
19 ltJc3 .tfS 20 l:!.c1 J:!.d7 21 'it>f2 J:tad8
22 ~e2 .ltg6 23 ltJa4 fS

This is an underestimated move and an

excellent weapon again 'Informanfs children'.
i.e. chess players who have learned lots of
variations by heart.
7 ~e2 .lte7
For 7 ... ~d6 see Game 13.
8 ltJc3
Others:
a) 8 b4 i.xb4 transposes to 7...~d6 8 b4
24 g4? (see Game 13).

32
 4 lDg5 d5 5 exd5 lDa5: Introduction

b) 8 d4? looks impressive, but it is incor- 14 tLle4 Itac8 gives Black more than enough
rect: 8... exd4 9 b4 O-O! 10 bxa5 i..b4+ 11 play for the pawn) l1...i.xc6 12 d3 lIe8!?
~dl .l:te8 12 'iUc4 i..xb5 13 'iUxb5 tLlxd5 (12 ... tLld5 also looks promising, e.g. 13 tLlxd5
with a decisive attack, e.g. 14 ~d3 ctJe3+ 15 ~xd5 14 ctJf3 .1i.d6 15 i.d2 .l:.ae8 and the
i..xe3 dxe3 16 ~xd8 e2+, or 14 h4ctJc3+ 15 black initiative is worth a pawn) 13 ctJge4 (13
tLlxc3 dxc3+ 16 ~d3l::te7!, or 14 a3 i..xa5 15 i.e3 b5! - a typical move in this kind of posi-
~xa5 tLlc3+ 16 ~d2 'iVf6 17 tLlxc3 dxc3+ 18 tion - 14 tLlge4 tLld7 15 tLlg3 g6 with excel-
~d3l:tad8+ 19 ~c4 'iUc6+ 20 ~b3 .l:!.d5 and lent play for the pawn) 13. .. tLld7 14 tLlg3 g6
Black wins. 15 ~h1 f5 16 f3 .l:!.c8 17 i.d2 b5 with very
c) 8 O-O?! ctJxd5 9 .1i.xd7+ 'iVxd7 10 d3 good play for the pawn, Gikas-Skembris,
(not 10 ~xe5?? f6) 10 ... ctJc6 and Black is Athens 2003.
slightly better. 10 f3
8 ... 0-090-0 As we shall see this is really risky. Proba-
White has also tried 9 .1i.xd7 (or 9 ctJge4 bly better is 10 'it'xe5!? i.d6 11 ~e3 (11
ctJxe4 10 tLlxe4 .1i.f5 11 'iV f3 i..g6 12 0-0 ~d4? fails to l1...c5 12 'it'd3 a6 13 h3 .1i.c8
~h8!? with excellent play for Black) 9.. :~Vxd7 14 i.a4 b5 15 tLlxb5 axb5 16 i.xbS c4 17
10 0-0 .l:.fe8! (10 ... ctJxd5?! is weaker because li'd4 i.b7 18 d3 i.xdS 19 i.e3 ~c7 20
of 11 'iVxe5 c6 12 d3 .l:i.fe8 13 i..d2 i..d6 14 .l:tadl i.h2+ 21 ~hl i.e5 Short-Xie Jun,
~d4 and White is better) 11 d3 (after 11 a3 Jinan 2002, and three pawns are not enough
tLlxd5 12 ~xe5 ctJxc3 13 ~xc3 i..xg5 14 for the piece here) 11....1i.f5! (after l1...a6 12
~xa5 .1i.f6 Black has excellent compensation i..e2.l:te8 13 ~d3 i.xe2 14ctJxe2 .1i.xh2+ 15
for the pawn) l1....1i.b4 12 ctJge4 tLlxd5 13 ~xh2ctJg4+ 16 ~gl 'iixgS 17ctJc3 'iUf4 18
tLlxd5 ~xd5 14 'iVg4 ~e6 and a draw was ~g3 'iVxg3 19 fxg3 tLlc4 20 b3 the endgame
agreed in A.Sokolov-Kunte, Bled 2002. is slightly better for White, Herbrechtsmeier-
Nunn, Germany 1984) 12 f4 i.xc2 13 d4 a6
14 'iVe2 i.f5 15 .Jta4 bS 16 i..c2 l::te8 and
Black is doing very well!
10 ... ~h5

9 ... ~g4?!
This is not the best way to get equal play.
Black has also tried:
a) 9....l:!.e8 10 tLlge4 c6 11 dxc6 ctJxc6 12
!bxf6+ .1i.xf6 13 i.xc6 .1i.xc6 14 d3 :re8 with 11 ~xe5
compensation in Felgaer-Skembris, Lido 11 i.d3?! gives White problems finishing
degli Estensi 2003. his development. Sergeev-Berezjuk, Tatran-
b) 9... c6! 10 dxc6 tLlxc6 11 .1i.xc6 (11 ska Lomnica 1998, continued 11..J::te8 12
~f3?! tLld4 12 tLlxd4 exd4 13 i..xd7 'iYxd7 ~h1 c6 13 dxc6 tLlxc6 14 g4!? (very commit-

33
Two Knights Defence

tal ~ the weakening of the king's position 20 ~f3 .l:.ad8 21 d3 lLlf5 22 ~d2 lLld4 23
ensures that Black has permanent counter- ~d1 c6 24 dxc6 'iVxc6 with some practical
play) 14... .ig6 15 ~xg6 hxg6 16 d3 lLld4 17 chances.
~g2 Sc8 18 1:tf2 'ilVb6 19 lLlce4 lLlxe4 20 17 ~f2 f5 18 g4! fxe4 19 gxh5 i.h4
lLlxe4 lLle6 21 g5 z:.ed8 22 h4 ~c5 23 .l:i.f1 White is better after 19 ... exd3 20 hxg6
..te3 24 .ixe3 'iVxe3 25 'iVf2 ~f4 with full ~c5 21 ~g2 dxc2 22 gxh7+ ~xh7 23 d3
compensation. ~d4 24 'ii'xc2 .txc3 25 d4+ ~h8 26 ~xc3
11 ... ..tg6 'iVxd5 27 .if4 as Black does not have enough
Also tempting is 11...~d6!? 12 'iVe3 a6 13 compensation for the pawn .
.i.e2 Se8 14 lLlge4 ~e7 15 ~f4 ~g6! (not 20 ~g2 exf3?!
15 ... lLlxd5?? 16 ~f5 and Black loses a piece) Simplifying the position does not work for
16 lLlxf6+ ..txf6 17 .id3 ~xd3 18 cxd3 c5! Black. He had no choice but to play 20 ... exd3
with very good compensation. 21 hxg6 dxc2 22 gxh7+ ~xh7 23 d4 ~h8 24
12 4:lge4 a6 13 ..td3 l:te8 14 ~h 1 b5 ~xc2 ~f6 25 lLle2 lLld6 26 lLlf4 lLlf5 and
Or 14...lLlh5!? 15 'iVd4 (if 15 g3?! f5 16 d6 although Black probably does not have
cxd6 17 ~d5+ ~h8 18 lLlf2 lLlf6 19 ~d4 d5 enough compensation, especially after 27
with good attacking chances) 15 ... lLlc6! 16 ~g2!, he does have many chances of cheat-
~c4 lLle5 17 ~3 lLlxd3 18 cxd3 lLlf4 with ing White ~ either through some kind of
compensation for the pawns. elaborate trap, or simply because the position
15 a3 IS messy.
21 iYxf3 i.xd3 22 ~xd3 iYg5 23 b3 4:ld6
24 i.b2 iYxh5 25 'iWh3 ~h6 26 1:.g1!

15 ... 4:lb7?
Too slow. Black should have played for
the initiative exploiting the exposed white Now Black cannot really avoid exchangin/!
queen with 15 ... lLlh5! 16 'iVd4 (if 16 g3 .if6 queens.
17 lLlxf6+ lLlxf6 18 ~g5 h6 19 1Wh4 ~xd3 26 ... i.g5 27 ~xh6 i.xh6 28 d3
20 cxd3 lLlxd5 and Black is better) 16 ... c5 17 White is a clear pawn up and should win.
'iVe3 c4 18 ~e2 lLlf6 19 d3 lLlxd5 20 lLlxd5 28 ... l:te7 29 J:!af1 ~ae8 304:ld1 g6
~xd5 and Black's pressure is worth more If 30 ....l:i.e2 31 .l:.g2 l:Iel 32 ~gl and
than a pawn. slowly, step by step, Black will lose this posi-
16 ~g3 4:lh5 tion.
Possible was 16 ...lLld6!? 17 lLlxf6+ ~xf6 31 c4 i.g7 32 ..txg7 ~xg7 33 c5 4:lf7 34
18 ..txg6 hxg6 19 f4 (after 19 d3 lLlf5 White b4 1:.e2 35 J:tg2 4:lh6 36 1:.xe2 l:txe2 37
has some problems with the queen) 19...'iVd7 4:lc3 ~d2 38 d6! cxd6 39 4:le4 1:.c2

34
 4 0,g5 d5 5 exd5 0,a5: Introduction

If 39 ....l:txd3 40 c6 ttJg8 41 Mel! ~f8 42 c7 Better than 9 0-0 ttJxdS 10 .txd7 ttJxc3 11
0,e 7 43 ttJcS dxcS 44 Mxe 7 and White wins. dxc3 'iVxd7 with equality, or 10... ttJf4!? 11
40 cxd6 0,f7 41 d7 0,d8 42 0,cs as 43 'iWg4 hS 12 'iWf3 'iixgS 13 d3 with unclear
.l:!.e1 axb4 44 0,e6+ 1-0 play - objectively White might be better, but
this has little practical importance.
Game 13 9".'i!fxd7 10 a3!
Morozevich-I.Sokolov
Sarqjevo 1999

, e4 eS 2 0,f3 0,c6 3 i..c4 0,f6 4 0,gS

dS S exdS 0,aS 6 i..bS+ i..d7 7 ~e2 i..d6

The best move. \x'hite prevents .....td6-b4

and does so with tempo by threatening b2-
b4.
After 10 0-0 Black has two ways to
achieve counterplay:
This is a real pawn sacrifice. Black closes a) 10... b6 11 d3 t!.ae8 12 ttJge4 iLe7 13
the d-flle and neglects taking back the pawn ttJxf6+ Sa..xf6 14 ttJe4 Sa..d8 15 c4 fS 16 ttJc3
on dS for a few moves. In my opinion this ttJb 7 and Black has some compensation for
move is not good enough to achieve full the pawn.
compensation, but the game is complicated b) 10... c6! 11 dxc6 ttJxc6 12 d3 ttJd4 13
and interesting. ~dl .l:!.ac8 14 Sa..e3 (14 a3!? is a possible im-
80,c3 provement, targeted against Black's next
If 8 b4 i..xb4 (this position can also arise move) 14... Sa..a3! 15 ..txd4 (15 iLc1 was
after 7... i..e7 8 b4 i..xb4) 9 'iVxeS+ 'iVe7! probably better, but Black has good play for
(Estrin recommends 9... ~f8!? 10 i..xd7 the pawn) lS ... exd4 16 ttJge4 ttJxe4 17 ttJxe4
'~xd7 11 0-0 .l:!.e8 12 'iVg3 iYxdS with a un- Sa..xb2 18 t!.bl iLa3 and Black is slightly bet-
clear game) 10 .txd7+ ttJxd7 11 'iWxe7+ ter, Sulskis-Beliavsky, Koszalin 1998.
iLxe7 12 ttJe4 ttJb6 13 ttJbc3 0-0-0 and the 10".b6 11 d3
position is roughly equal. If 11 0-0 ttJb7 12 b4 as with counterplay
8".0-0 according to Pilkovi.
Maybe Black should try to improve with 11...c6?!
8... c6!? 9 dxc6 (or 9 ttJge4 ttJxe4 10 ttJxe4 Better was 11...ttJb7 12 ttJ ge4? (if 12 0-0
.te7 11 dxc6 ttJxc6 12 c3 0-0 13 0-0 a6 14 a6 13 ttJf3 .l:lae8 or 12 b4 as! with counter-
.ta4 bS 15 i..b3 ttJaS 16 iLc2 fS) 9... ttJxc6 10 play) 12... ttJxe4 13 dxe4 fS 14 O-O?! (castling
iLxc6 .txc6 11 0-0 0-0 12 d3 iLc7 13 .l:.el short the white king will only be safe in his
iYd7 with compensation. grave) 14.. .f4 15 f3 Sa..cS+ 16 Whll:.f6! with a
9 i..xd7 deadly attack in Rabello-Costa, Brazil 1997.

35
Two Knights Defence

White should prefer 14 i.e3 or else 14 exfS But not 17 'iVh3? f6! and the knight on h7
followed by 1S i.e3 and 16 0-0-0. is trapped! Arnold-Iruzubieta, Oropesa del
Mar 1996, continued 18 0-0 CZJd8 19 f4 exf4
20 ':xf4 'iYe6 21 'iYhs \if7 22 ~fS 'iYe6 23
~S ~f7 24 'iYfS and the game was drawn.
Most likely Sokolov did not know this game,
as the improvement is rather easy for a
strong grandmaster to see, though computer
programs do not understand such things as
trapped pieces.
17 .. .f6 18 ltJh3 it'xc2 19 0-0
White has an extra pawn and a safe king.
19 ... ltJd8 20 .i.e3 it'c6 21 d4!
Eliminating the last weakness in the white
camp. From here on it is just technique for a
12 b4ltJb7 13 dxc6 it'xc6 14ltJce4ltJd7 world class player like Morozevich.
In my opinion it was better to play 21 ... ltJf7 22 l:1ad1 ltJf8 23 "iYg4 iVe6 24
14...cLixe4 1S ~xe4 ~xe4+ 16 CZJxe4 i.e7 17 "iYxe6 ltJxe6 25 d5 ltJf8 26 d6 .i.d8 27 g4
~d2 fS 18 CZJg3 g6 with compensation. J:tc4 28 f3 g6 29 Wh 1 !
15 it'f3! .i.e7?! But not 29 CZJhf2?? fS! 30 gxfS gxfS and
After 1S ... fS 16 CZJxd6 \ixf3 17 CZJxf3 White loses a piece!
CZJxd6 18 ~d2 e4 19 CZJd4 ~fe8 20 i.b2 29 .. J:tac8 30 ltJhf2
Black could sing along to the hit song in this
variation: 'Where is my compensation?'. Oust
imagine some lousy beats and a skinny young
blonde singer and you are there!) However,
this was still a better solution.
16ltJxh7!
The queen hangs on c6, so the knight is
taboo.
16 .. J:!.fc8
Of course not 16...~xh7?? 17 CZJgS+.

30 ... ltJd7
If 30 ... fS 31 gxfS gxfS 32 .l:!.g1+ ~h8 33
CZJgS and White wins.
31 J:tg1 Wf8 32 ':g3! J:!.c2 33 h4! .i:te2 34
.i.d2 f5
Or 34...l:!.c4 3S 'it'g2 fS 36 gxfS gxfS 37
~f1 .l:!.xd2 38 CZJxd2 l':txh4 39 'it'e2 and White
wins.
35 gxf5 .i.xh4 36 .i:th3 .i.xf2
Black cannot escape. If 36 ...gxfS 37 .l:!.xh4
17 ltJhg5! fxe4 38 CZJxe4 l':tc4 39 l:i.g4 and White wins.

36
 4 fijg5 d5 5 exd5 fija5: Introduction

37 fxgS! 0-0, or 9... J.c5 10 0-0 0-0 11 b4! i..xb4 12

The black knight 1S strangely out of ctJc3) 10 'ikf3 J.b7 11 'iVe2 i..e7 12 d3 ctJc6
squares. 13 c3 0-0 140-0 ctJd5 15 ctJh3! Me8 16 ctJd2
37 ... .!:l:.xd2 f5 17 ctJb3 and White is much better -
If 37 ... ctJd8 38 .l:th7! and White has a mat- Estrin.
ing attack. b) 8.. .'~c7?! is a bit slow, e.g. 9 i..d3 i..e7
38 fijxd2 fijxdS 39 fije4 fijxe4 40 I:th8+ (or 9... i..d6 10 ctJc3 J.g4 11 ctJb5 i..xf3 12
~g7 41 J:Ixc8 1-0 ctJxc7+ i..xc7 13 gxf3 ctJd5 and White is bet-
r-----------------" ter after 14 a3!? or 14 h4!?) 100-00-0 11 ~f5
Game 14 ~b7 12 d3 c5 13 'iVh3 g6 14 ~e4 and White
Spiegel-Mari Arul stands slightly better, Kamishev-Sopkov,
Calcutta 1997 USSR 1949.
c) 8... ~e7!? (another interesting move,
1 e4 eS 2 fijf3 fijcs 3 ..Itc4 fijfS 4 fijgS about as strong as 8...l:!b8!?) 9 i..xc6+ (if 9
dS S exdS fijaS S ..ItbS+ cS 7 dxcS bxcS ~d3 0-0 10 ctJc3 h6 11 ctJge4 ctJd5 12 ctJg3
8 ~f3?! g6! and Black has compensation - Van der
Wiel) 9... ctJxc6 10 'iWxc6+ i..d7 11 'iVc4 0-0
12 ctJc3 l::i.c8 13 'iVe2 h6 14 ctJf3 e4 15 ctJe5
~e6 and Black has full compensation since
White has problems developing; e.g. 16 b3?
l::i.xc3 17 dxc3 'iVa5 18 .i.d2 'iVxe5 and Black
is much better, or 16 O-O?! 'iWd4! 17 'iVb5
.i.c5 18 ctJc6 'iVd6 19 ctJa5 i..b6 with a killer
attack (20 ...ctJg4 is a great threat, and 20 h3
i..xh3 does not improve things!), or if 16 h3
'iVd4 17 ctJg4 ctJxg4 18 hxg4 l::tfd8 with
strong compensation for the pawns. Proba-
bly 'advantage Black' is a more accurate
evaluation of the position.
This move was quite popular in the mid- 9.i.d3
dle of the 19th century. It was reintroduced White has no reasonable alternative:
in tournament practice in the 1920's by a) 9 J.e2? .i.e7 10 ctJc3 0-0 11 d3 ctJd5! 12
grandmaster Efim Bogolubow and after that ctJge4 f5 and Black is simply better.
was a frequent guest in tournaments until the b) 9 .i.a4? 1:tb4 10 i..b3 ctJxb3 11 axb3 h6
1980's. Now the reputation of this move is 12 ctJh3l:t.e4+ 13 ~f1 ~g4 and Black wins.
bleak. Black receives more active and dan- c) 9 ~xc6+? ctJxc6 10 'iWxc6+ ctJd7 11 d3
gerous play for the pawn(s) than in the main (if 11 ctJf3 .i:b6 12 'iWe4 ~b7 13 'iVe2 l:rg6
lines with 8 ~e2. with an attack, or 11 d4 .i.e7 12 h4 h6 13
8 .. Jl:b8!? ctJe4 0-0 14 ctJbc3 l::i.b6 and Black has full
Not as strong as 8... h6 (see Game IS) but compensation for the pawns) 11...i..e7 12
interesting nevertheless. The alternatives are: ctJf3 0-0 13 ctJc3 .l:tb6 14 'iVa4 .i.b7 15 'iWxa7
a) 8... cxbS? only looks interesting. After 9 ctJcS 16 'ika5 fS and Black has a strong attack
'iWxa8 Black does not have enough compen- according to palk6vi.
sation for the material; e.g. 9... 'iWd7 (if 9 ... hS
9 .. .'fic7 10 ctJc3 i..cs 11 'iVf3!, or 9...ctJdS 10 9... .i.e7 10 0-0 0-0 11 ctJc3 h6 12 ctJh3
-1Jc3 tDc7 11 'iWxa7 ctJc6 12 'fib6 ctJd4 13 .i.g4 13 'iWg3 'iVd7 14 i..e2 i..xe2 IS ctJxe2

37
Two Knights Defence

i.d6 16 d3 e4 also gave Black good compen-

sation for the pawn in Sakharov-Voronov,
USSR 1971.
10 CUe4CUd5!

14 d3?
White's main problem is the knight on f4.
It was a good idea to get rid of it with 14
lDe2! c5 (14 .. :iVf6? 15 'it'c3! and wins is a
A standard move in this line. Black avoids clever pointl) 15 ~c3lDg6 (wild is 15 ...lDd5!?
exchanges as the white pieces are exposed in 16 ~xe5+ 'itn 17 ~xf5+ ltJf6 18 lDbc3 g6
the centre, while the black pieces are mostly 19 ~h3 i..d6 with a completely unclear posi-
threatening. tion, but one probably easier to playas Black)
11 b3 16 lDg3 f4 17 i..d3 lDh4 18 lDe4 lDc6! and
White needs to get his bishop out some- Black has wonderful compensation for the
time. After 11 lDg3?! g6 12 0-0 i..g7 13 lDc3 pawn. But still... it is a fight.
0-014 ..te2 .l:!.b4 15lDxd5 cxd5 16 'i*'a3lDc6 14 ... e5 15 'i¥g3 iif6!
Black is better, Estrin-Ragozin, USRR 1955. Black is much better here.
11 lDbc3?! lDf4 12 i..f1 f5 13 lDg3 g6 also 16 i..xf4 exf4 17 'i¥xf4 ..td6 18 'i¥a4+
leaves Black in the driving seat. cue6 19 i..e2 0-0
11 ... CUf4! And now he is winning.
The right method of annoying White. The 20 'i¥e4+ Wh8 21 0-0 cud4 22 cud2
slower l1...g6?! 12 'it'g3lDf4 13 i..b2 i.g7 14 Losing a piece, but there is no salvation. If
i..a3 lDb7 15 ..ta6 c5 16 f3 leaves White 22 i..d 1 'it'h4 23 g3 '*'h3 24 f3 i..xg3 25
slightly better according to Van der Wiel. hxg3 ,*,xg3+ 26 'ith1 .l:!.f6 and Black wins.
12 i..f1?
In times of emergency, all troops must re-
turn home! Of course this is a bad idea in a
sharp chess middlegame. Necessary was the
unpleasant 12 i..b2 lDxd3+ 13 ~xd3 ~xd3
14 cxd3 f6 with excellent compensation for
Black.
12 .. .15 13 cuee3
After 13lDg3 g5! 14lDe2 e4 15 ~e3 ~c7
16 lDbc3 i..g7 17 lDxf4 gxf4 18 ~c5 i..f8 19
'iVd4 .l:!.g8 Black is clearly better.
13 ... i..b7

38
 4 0,g5 d5 5 exd5 0,a5: Introduction

22 ... 'ifg5 23 0,de4 fxe4 24 dxe4 l:.be8 terrible and after 13 .. .f5 14 ctJg3 .l:!.b8 Black is
25 iVd3 0,xe2+ 26 0,xe2 ~xe4 27 'tlVh3 much better) 13 ... f5 14 ctJec3 (not 14 ctJg3??
ii.f5 28 f4 'ilVg6 29 'iff3 ~g4 30 'tlVd5 ~f6 and White loses the queen) 14....l:!.e8 15
~xe2 31 f5 'ilVf6 32 J:!.f2 'YiVxa1 + 0-1 ~d4 (the only move) 15 ... i.f6 16 ~a4 (an-
other only move) 16 ....l:.b8! and White's posi-
Game 15 tion is unco-ordinated and his scattered
Van der Wiel-Spassky forces will always be passive. 12 'iUxe5 f5 13
Reggio Emilia 1986/87 ctJec3 i.d6 14 1Vd4 ctJf4 gives Black similarly
strong play. So it seems that Palkovi is cor-
1 e4 e5 2 0,f3 0,c6 3 ~c4 0,f6 4 0,g5 rect and that after 10 i.e2?! White is balanc-
d5 5 exd5 0,a5 6 ~b5+ c6 7 dxc6 bxc6 ing on the edge.
8 iif3?! h6! 10 ... cxb5 11 0,xd5 ~b7
Also good is 11....i.e6 12 ctJe3 .l:!.c8 13 0-0
~d7 14 ctJg3 h5 15 c3 ctJc6! 16 Mdl h4 and
Black was much better in Gikas-Balashov,
Lugano 1988.
12 0,e3 'ilVd7 13 0-0 0,c6 14 d3 0-0-0

This move was fIrst introduced in a game

by two Masters in the Soviet Union in 1955.
After this game grandmaster Mark Tairnanov
was sure that Black is doing well in this line.
Grandmaster (in correspondence chess) Ya-
kov Estrin was convinced that White is bet- Black has excellent compensation for the
ter. Who was right? Well, see the game! pawn.
90,e4 15 c3 g6
After 9 b4? ~g4! 10 .ii.xc6+ ctJxc6 11 Preparing .. .£7 -f5.
~xc6+ ~d7 12 ~c3 hxg5 13 ~xe5+ ~e7 14 16 a4 b4
'iVxe7+ i.xe7 Black is better. The three There is no reason for Black to allow
pawns are not enough for the piece here. White to open the a-flie for his rook.
9 ... 0,d5 10 0,bc3 17 0,f6 iVe6 18 0,fd5?
10 ~e2 i.e7 11 ~g3 is also possible. The After this White is in a nasty pin and
authors of EeO believe that this position is without counterplay. Better was 18 .l:!.dl!
much better for White, whereas Pilkovi be- planning to exchange queens with 19 'iUg4.
lieves that Black has a strong initiative after Nevertheless Black has 18 ... i.g7! 19 ctJe4
11...0-0 12 d3 ~h4. However, they only give ctJa5 20 ~e2 ctJb3 21 .l:!.bl f5 22 ctJd2 ctJxd2
words and assumptions, not moves. So in- 23 .ii.xd2 f4 maintaining his attack.
stead of believing them, we will check the 18 .. .f5 19 c4
position: 13 ~xe5 (necessary; 13 ~f3? looks A sad but necessary move. Now the black

39
Two Knights Defence

knight enters the glorious d4-square. 27 1:.e8+ l:1.d8 28 l:1.e7?

19 ... CDd4 White was given a last chance and should
have taken it with 28 'iVxg6! ~xg6 29 ~xf8
i..xd5 30 ~xd8+ 'it'xd8 31 cxd5 ctJb3 32 ctJg3
and Black would have to play very carefully
to win this ending.
28 .. .J::txd5!

20 'iWh3?!
Slightly preferable was 20 ~d1 f4 21 ctJc2
ctJb3 and Black is much better.
20 ... g5 21 .!:Ie1 .l:!.g8 22 "ti'h5 g4 23 CDf1
If 23 as Itg5 24 'iVh4 f4 25 ctJf1 ctJf5 and
Black wins. 29 l:1.c7+ ~d8 30 'ii"h4+
23 ... CDc2 24 .i.f4 CDxa1 25 .l:!.xe5 'ii"g6 26 Or 30 ~xg6 ttxg6 31 cxd5 i..xd5 and
l:.e7! Black wins.
A nice move, but insufficient of course. 30 .. .'it'e8 31 cxd51:.g7
26 ... .!:Id7? White does not have any real compensa-
Often tricks like this are useful to gain tion for the piece.
time on the clock, but not here. Black should 32 CDe3 CDb3 33 h3 CDd4 34 ~f1 l:!.xc7
play 26 ...l:i.xd5! immediately and the game is 35 .i.xc7 .i.e7 36 'ii"g3 f4 37 'ii"xg4
over. "Yi'xd3+ 38 ~g1 fxe30-1

40
 4 ttJg5 d5 5 exd5 ttJa5: Introduction

Summary
6... .td7!? is a good alternative to 6... c6 and, most importantly, there is plenty of room for in-
dependent ideas and analysis. In the line 6... c6 7 dxc6 bxc6 then 8 'irYf3?! is a just weak move.
Nevertheless, Black must know how to meet it, and the best way is with 8... h6!.

1 e4 e5 2 ttJf3 ttJc6 3 .lic4 ttJf6 4 ttJg5 d5 5 exd5 ttJa5 (D) 6 .lib5+

6 d3 - Game 11
6 ... c6
6... .td7 7 it'e2 (D)
7... i..e7 - Game 12
7... .i.d6 - Game 13
7 dxc6 bxc6 8 "iff3 (D)
8... l:!.b8 - Game 14
8... h6 - Game 15

5 ... ttJa5 7 "ife2 8 "iff3

41
I CHAPTER FOUR I
4 ctJgS dS S exdS ctJaS:
Main Line

1 e4 eS 2 lUf3 lUe6 3 .lte4 lUf6 4 ctJgS drawn. Usually this line does not give rise to
dS S exdS lUaS 6 .ltbS+ e6 7 dxe6 bxe6 so many draws, and it can therefore be useful
8 .lte2 h6 when a win is required and a draw is equiva-
This has been the main line of the 4 l2lg5 lent to half-point loss (which should really be
Two Knights since the great Russian Mikhail the case in all games!).
Chigorin demonstrated Black's resources at
the end of the 19th century. To this day it is Game 16
still played occasionally by strong grandmas- Malakhatko-Timoshenko
ters, most recently by Morozevich and Kiev 2003
Sutovsky. The line does not seem to offer
White an advantage - for the pawn Black has 1 e4 eS 2 lUf3 lUe6 3 .lte4 lUf6 4 lUgS
space and a lead in development - but the dS S exdS lUaS 6 .ltbS+ e6 7 dxe6 bxe6
position is complicated enough for both 8.lte2
players to play for a full point.
9 l2lh3 is a very old idea by Wilhelm
Steinitz, though it did not bring him a lot of
success in his games against Chigorin. In the
1960's Robert Fischer brilliantly reintroduced
9 l2lh3 to the top tournaments, and the same
happened in the 1990's when Nigel Short
had success with the move. Recentlv Ukrain-
ian players have contributed enormously to
the development of the variation. At the
beginning of 2003 there was a very strong
theme tournament in Kiev, in which all the
games started from the position after 9 l2lh3.
There it was convincingly proved that Black's The main alternative, 8 'iWf3, was exam-
chances are at least equal: Wrote's results + 12 ined in Games 14 & 15 in the previous chap-
=20 -16 say it all. What is most surprising, ter. White has also tried two inferior bishop
though, is that over 40% of the games were retreats:

42
 4 0,g5 d5 5 exd5 0,a5: Main Line

a) 8 Ji.f1? h6 9 tDh3 Ji.c5 10 d3 ~6 11 and now:

~e2 i.g4 12 f3 Ji.xh3 13 gxh3 0-0-0 and a) 11 ... 'iVc7 12 0-0 Ji.xh3?! (better is
Black is much better, Steinitz-Chigorin, Ha- 12.. .l::tb8 with compensation; there is no rea-
vana 1892. son to capture on h3 just yet as the knight
b) 8 Ji.d3? tDd5 9 tDe4 f5 10 tDg3 tDf4 11 has nowhere to go) 13 gxh3 .l:!.ad8 14 ~ e 1
i.f1 Ji.c5 12 c3 i.b6 13 d4 tDg6 14 Ji.d3 0-0 l:tfe8 15 Ji.f3 tDb7 16 i.d2 'iVc8 17 Ji.g2 i.f8
15 b4 tDb7 and Black is much better, Cas- 18 'iVe2 tDd6 19 tDe4 and White was much
taldi-Keres, Stockholm 1937. better in Taborov-Kruppa, Kiev 2003.
8 ... h6 b) 11..J::b8 (the most natural move) 12
There are some minor alternatives here: 0-0 ~c7 (weaker is 12 ... .l:!.b4?! 13 'it'hl Ji.xh3
a) 8 ... Ji.e7!? is very rare and nothing spe- 14 gxh3 l:i.h4 15 l:i.gl .l:'txh3 16 ~f1 and
cial. Nevertheless, it is not as bad as many White is better) 13 f4 Ji.xh3 14 gxh3 exf4 15
theoreticians and might be a useful weapon Ji.xf4 Ji.d6 16 .i.xh6 .i.e5 (16...~xh2+? 17
against players who know all theoretical lines '.toh1 gxh6 18 llxf6 Ji.e5 19 'iig1+ 'it'h7 20
but nothing about playing chess. After 9 d3 tDe4 would give White a crushing attack) 17
0-0 10 tDc3 tDd5 11 tDge4 f5 12 tDg3 tDb7 .i.el .ixh2+ 18 'it'hl .l:!.fe8 with compensa-
Black has the standard play for the pawn. tion for the pawn according to Timoshenko.
b) 8... Ji.c5?! is risky, as after 9 d3 0-0 10 11 ... c5 12 0,f3
tDc3 h6 11 tDge4 White has managed to re- If this is where White wants the knight
treat in a much more comfortable manner. then 9 tDf3 might come into consideration!
Fischer-Gould, Houston 1964, continued Of course what White wanted was to avoid
l1...tDxe4 12 tDxe4 Ji.e7 13 0-0 tDb7 14 'it'hl ... e5-e4, but giving two tempi to do so is too
Ji.f5 15 Ji.f3 and White is much better. much.
90,h3!? 12 .. .'i¥c7 13 0,bd2 J:.d8 140-0 c4!
Black's prospects in the position are based
solely on his lead in development. Therefore
he has no second thoughts about giving up a
second pawn to accelerate the assault on the
white position.
15 'i!Ve1 iLb7 16 0,xc4 0,xc4 17 dxc4

9 ... iLe7
According to the Ukrainian GM Georgy
Timoshenko, 9 ... Ji.e7 is underestimated by
theory. The other black moves 9...g5, 9 ... ~d6
and 9 ... ~c5 are considered in Games 17, 18
and 19 respectively.
10 d3 0-0 11 0,917 White has managed to win another pawn,
In an open position a move like this but not to develop his queenside. It shortly
should always bad. Preferable was 11 tDc3 becomes clear that his priorities have not

43
Two Knights Defence

been the best. 32 ~f2 .l:i.d3 33 'ii'e5 'ii'g4! 34 ~el f6 35

17 ... e4 18 lUd2?! ~e7 ~g5 and Black wins all the same. Nev-
The most natural square, but the knight ertheless Black had plenty of chances to mess
was needed on the kingside. Better was 18 up the attack here, and should have been
ttJh4, although after 18 ... .ltc8! 19 g3 .lth3 20 given the opportunity to do so.
ttJg2 'iVb 7! Black is aiming forcefully at both
b2 and g2, e.g. 21 b3 e3 22 f3 .tc5 23 a3 as
\mUte is under great pressure.
18 ... .td6

23 .. :iYxf3! 0-1
After 24 gxf3 .i.xf3+ 25 ~gl ttJg4 mate
on h2 cannot be prevented.

19lUb3 Game 17
If 19 'it>h1 .i.xh2 20 g3 e3+ 21 ~xh2 exd2 Timoshenko-Vysochin
22 .txd2 l:!.d4! 23 f3 .l:!.e8 and Black wins Kiev 2003
because of 24 .ltf4 l:!.xf4 25 gxf4 'iVxf4+ 26
'it>gl 'iVg5+ 27 ~h2 ttJh5 28 l:!.f2 ttJf4 and 1 e4 e5 2 lUf3 lUe6 3 i.c4 lUf6 4 lUg5
\mUte is tangoed. d5 5 exd5 lUa5 6 i.b5+ e6 7 dxc6 bxc6
19 ... i.xh2+ 20 >t>h1 .te5 21 c3 8 i.e2 h6 9 lUh3 g5!?
21 'iVa5 'iVe7 22 'iVc5 ~d6 23 'iVe3 .lta6
24 ttJd2 l:!.ac8 also grants Black fantastic
compensation.
21 ... e3 22 i.xe3 ~e6 23 i.f3??
\mUte completely forgets about his king's
frailty. 23 f3 ttJh5 24 ttJa5 was necessary,
when Black has to fInd 24...'iVg6! 25 ttJxb7
ttJg3+ 26 ~g 1 and then:
a) 26 ...l:!.e8? 27 ~dl! (if 27 ttJc5 'iVh5 28
.i.d3 ttJe2+ 29 ~f2 .ltg3+ 30 ~xe2 ..txe 1
and Black wins) 27 ... ttJxfl 28 'iVxfl ~g3 29
i.c2 'iVh2+ 30 ~f2 .tg3+ (30 ...'iWg3+ 31 ~gl
only gives a draw) 31 'it>e2 i.f4 32 ..te4
i.xe3 33 ~xe3 f5 and Black has a strong A very promising move in the sense that it
attack for the piece. should be good for tournament play. The
b) 26...~5! 27 f4 ttJxe2+ 28 ~f2 ..txf4 reason is that Black has a simple plan: pure
29 ..txf4 ttJxf4 30 ~e3 ttJxg2 31 ~xg21lVg6+ murder one!

44
 4 t'i'Jg5 d5 5 exd5 t'i'Ja5: Main Line

10 d3 13 ... t'i'Jd5
Also possible is 10 c3 lVd5 11 f3 ~xh3 12 Black need not insist on keeping the
gxh3 ltJb7 13 'iVa4ltJc5 14 'iVc4 ltJe6 15 b4 queens on the board. After 13 ... e4!? 14 dxe4
ltJf4 16 'iVxd5 cxd5 17 .ib5+ WdS with un- 'iVxd1+ 15 .ixd1 ltJxe4 16 0-0 ltJc4 and
clear play. Black has a good initiative for the pawn; e.g.
10 ... ..tg7 17 ~b3 Ji.a6 lS ltJd4 ltJa5 19 .l:i.d1 .l:i.bdS
Black has tried or considered several other with good play.
moves: 140-0 g4 15 t'i'Je1
a) 10... .l:i.gS!? 11 ltJg1 (if 11 ltJc3 .l:i.bS 12 Or 15 ltJfd2 f5 with compensation.
ltJg 1 c5 13 ltJf3ltJc6 140-0 .ie6 15 b3 g4 16 15 ... f5 16 g3?!
ltJd2 ltJd4 17 ltJde4 ltJd7 with good com- This seems a little irrational. One should
pensation for the pawn) 11...g4 12 ltJd2 ~e6 try to avoid moving pawns in front of one's
13 ltJf1 h5 14 ltJe3 .l:tbS 15 c3 c5 16 "ib'a4+ own king, as it will be easier for the attacker
~d7 (more natural than 16 ... ltJd7?! 17 h3 f5 to fInd a way to open the position. While this
lS hxg4 hxg4 19 f3 gxf3 20 .ixf3 and White of course does not count in all positions, for
is much better, Kruppa-Kosikov, Kiev 2003) this one it certainly does. Nevertheless, after
17 "ib'c2 .ic6! with excellent play. 16 b4 ltJb 7 17 a3 .ie6 Black has good
b) 1O ...g4!? 11 ltJg1 Ji.c5 12 ltJc3 .l:i.bS and compensation anyway.
Black has good play. 16 ... h5 17 t'i'Jg2
c) 10...l:tbS 11 ltJg 1 c5?! (this is too slow 17 c4?! would leave the d4-square weak
and gives White more time to consolidate; it for ever, and after 17... ltJc7 lS ltJc3 ltJe6
also occupies the c5-square from where both Black's attack is probably decisive.
the knight and bishop can be very active. 17 ... c5?
Better was 11...g4!? to keep the white knight Black has compensation for the pawn be-
on gl) 12 ltJd2ltJc6 13ltJc4 g4 14 c3 .l:i.gS 15 cause of his lead in development. But the
h3 h5 16 hxg4 hxg4 17 g3 and White was slow manoeuvring of his knight to c6 costs
better in Timoshenko-Sergeev, Kiev 2003. two moves, and allows White to put a knight
11 t'i'Jg1 on c4 in the meantime.
After 11 ltJd2 0-0 12 ltJg 1 l:tbS 13 ltJb3 Instead Black should act \vith great virility
ltJxb3 14 axb3 a6 Black has fIne play. He is and play 17... f4!
close to being fully developed, whereas
White is not even in the neighbourhood.
11 ... 0-0 12 c3 nbS 13 t'i'Jf3

lSltJd2! (after lS f3? Black has lS .. J:hb2!

19 Ji.xb2 "iVb6+ 20 .l:!.f2 lVxb2 21 ltJd2 ltJxc3
22 1Vc1 ltJxe2+ 23 .l:i.xe2 'iVd4+ 24 ~h1

45
Two Knights Defence

'iVxd3 and wins as the white position simply This is a typical position in which oppo-
collapses) 18... f3 (another possibility is site-coloured bishops ensure a decisive attack
18 .. :~Wb6!? 19 ttJh4 .i.a6 20 ttJe4 ttJb7 21 b4 rather than a draw.
f3 22 ttJxf3 gxf3 23 .i.xf3) 19 ttJxf3 gxf3 20 31 ... J:.xd1 32 J:.xd1 J:!.d8 33 J:txd8+ ~xd8
.i.xf3 and the game is unclear. Black has won 34 ~d2!
a piece and retains a greater activity, but Or they give a winning endgame because
White has three, possibly four pawns for the the c5-pawn is weak and Black cannot gener-
piece and his king seems safe enough for ate any counterplay against the a2-pawn.
now. The position is a mess. 34 ... ~f6 35 'ifd6 ~f7 36 .te5 ~g5 37
18 lLla3 lLlc6 19 lLlc4 .te6 i.f4 ~f6 38 'ifd2 ~e8 39 <t;>f2 'ife7 40
Now Black has lost momentum. After ~e3 ~d7 41 'ifd6 'ife7 42 ~d2 'iff6 43
19 ... f4?! 20 f3 fxg3 21 hxg3 'iid7 22 ttJge3 ~b8+ ~f7 44 "ilVxa7+ ~g6 45 'ifxcs
White is much better as the black attack will 'ifh8 46 ~f2 ~a8 47 a4 1-0
never really get there.
20 'ifc2 lLlb6 Game 18
If 20 ... f4 21 f3 'iVc7 22 ttJd2! and, with the Vysochin-Shishkin
knight soon ftrmly planted on e4, White is Kiev 2003
better.
21 .i.e3? 1 e4 e5 2 lLlf3 lLlc6 3 .tc4 lLlf6 4 lLlgS
Virtually encouraging Black to push the f- d5 5 exd5 lLla5 6 .i.bS+ c6 7 dxc6 bxc6
pawn. After 21 f3!? ttJxc4 22 dxc4 'ile7 23 8 .i.e2 h6 9 lLlh3!? .td6
.i.e3 White is better. Recently this has been the main line. As
21 ...'ife7? we shall see, it does not give Black as easy
Black again plays too slowly. Necessary playas after 9...i.e7 or 9...g5, both of which
was 21...ttJxc4! 22 dxc4 f4 23 gxf4 .i.f5 24 promise more counterplay in my opinion. My
~c1 ttJd4! and Black obtains a dangerous conclusion on 9... Jl..d6 is that it gives White
initiative as the knight on d4 is untouchable some advantage.
(if 25 cxd4 exd4 26 i.d2 d3 and White is
crushed).
22 f4 gxf3 23 .l:!.xf3 e4 24 l':tf4 lLlxc4 25
dxc4 lLle5 26 b3 lLlg4 27 .txg4 hxg4 28
J:!.d1 .l:l.bd8 29 J:tff1 .te5 30 lLlf4 i.xf4 31
.txf4

10 lLlc3
The most natural move, but not necessar-
ily the best. White has also tried:
a) 10 d4?! O-O! (better than Fischer's rec-
ommendation 10 ... e4 when after 11 ttJf4
~c7 12 g3 0-0 13 0-0 White is doing quite

46
 4 0,g5 d5 5 exd5 0,a5: Main Line

well) 11 dxe5 .ixe5 12 ~xd8 l:txd8 13 ttJf4 1992.

.if5 14 ttJd3 SLd4 15 ttJd2 .l:.e8 16 ttJf3 SLb6 12... 0,xc3?
17 ttJfe5 c5 with excellent attacking chances This exchange does not improve the black
for Black in the 2000 Internet blitz game, position in any way. Better was 12 .. :iWc7!?
Short-Golod. This would be rather insignifi- with the standard plan: ....t:i.b8, ... c6-c5-c4 etc.
cant if it were not that Short is the greatest 13 bxc3
expert on the 9 ttJh3 line, and his games, Now due to Black's last move, White can
even blitz games, are always interesting. finish his development and take control over
b) 10 d3 (the most flexible move, but of- the centre. The price for this is very low: a
ten it will simply transpose) 10... 0-0 11 0-0 mere pawn.
(11 ttJc3 transposes to the game) 11....t:i.b8 12 13 .. :tWh4 14 ~h1! .txh3?!
'It>hl (after 12 c3 c5 13 ttJd2 'iVc7 14 ttJe4 In fact this pawn is not worth the bishop.
SLe 7 15 f4 ttJxe4 16 dxe4 nd8 17 ~c2 ibxh3 White also gets the open g-ftle. Black still has
18 gxh3 exf4 19 .ixf4 .id6 20 .ixd6 nxd6 some compensation after 14...g5 15 ttJg 1
Black was able to claim compensation in 'iVa4 although White is better here.
Timoshenko-Sergeev, Kiev 2003) 12... c5 13 15 gxh3 'ii'xh3
ttJg l 'iVc7 14 ttJd2 .l:td8 15 'iVel ttJc6 16 c3
.ifS 17 ttJc4 .ifS 18 f41 e4 19 dxe4 ttJxe4 20
ttJf3 and White is slightly better according to
Timoshenko.
10 ... 0-0 11 d3 0,d5
11...J::tb8 12 ibf3 'iVc7 13 ttJg l c5 14 ttJge2
c415 ttJg3nd8 16 0-0 ~a6 17 ~e2 SLb4 18
.l:td 1 ttJc6 led to a draw in Malakhatko-
Sergeev, Kiev 2003
120-0!

White has returned the pawn, but now has

the bishop pair and the open g-ftle which
give him the better game. The white king
might seem fragile at first glance, but Black
having no light-squared bishop, it is all an
illusion.
16 Iitg1 f5
16 ... e4 17 .l:tg2 .l:tfd8 was probably better.
Now it is easy for White to improve his posi-
tion.
White should not fear ....ixh3 as the 17 J:tg3!
weakness of the light squares is balanced by White takes over the initiative. 17 'iVfl
Black giving up his light-squared bishop. 'iVxfl 18 ~xfl ~h7 19 .ig2 J::tab8 20 c4 is
Also possible was 12 ttJe4!? ~c7 13 c4 ttJe7 only slightly better for White.
14 0-0 f5 15 ttJc3 g5 16 ~hl ttJg6 17 b4 17 ... ~h4 18 ~g1 ~h8!? 19 ~g2!
ttJb7 18 f3 ttJd6 19 SLe3 and White was White wants to win without granting the
slightly better in Kamsky-Yusupov, Tilburg opponent any counterplay. After 19 .l:txg7

47
Two Knights Defence

.l:!.g8 20 1Ig6 f4 21 ~g4! 1Vxg4 22 :xg4.l:!.xg4 possible? Well, fashion also rules chess. Many
23 i.xg4 i.c5 24 f3 l:tb8 Black would have people analyse the positions after 15 moves
some drawing chances. trying to improve on previous players'
19 ... f4 20 1:I.g4 "iie7 21 i..d2 )lab8 22 choices, when all their problems could be
l:!.g1 J:tf7 23 i..f3 "i\Vf8 solved by rewinding a few moves.
Personally I prefer to die in battle, so my
choice is this position would have been
23 ... .l:!.b2. Nevertheless, after 24 ~3 ~f8 25
~h5! .:tJb7 26 .txc6 the battle would be
quite short.
24 i..e4

10 d3!
A tricky move order. Others:
a) 100-0 g5! is similar to the 9...g5 line and
then the game could continue 11 ~hl (11
c3!? i.b6 12 b4 ttJb7 13 d4 exd4 14 i.f3
ttJd5 15 Ite1+ ~f8 leads to an unclear posi-
White is orchestrating a light square sym- tion) 11 ...g4 12 ttJgl ttJe4 13 .txg4 'iVd4
phony - a requiem to mourn the death of the (13. .. Ct:Jxf2+ 14 l:txf2 .txf2 15 i.xc8 is
black king. Fischer's analysis, and White has full com-
24 ... i..e7 25 .l:.g6 i..f6 26 c4 lLlb7 27 pensation for the material; after 15... .txgl 16
i..xc6 .ta6! White plays against the knight on as)
Black's position is deteriorating rapidly. 14 .txc8 ':xc8 15 Ct:Jh3 Ct:Jg5 16 c3 ~h4 P
27 ... lLlc5 28 i..d5 J:.c7 29 i..c3 "VJiIe7 30 b4 ttJxh3 18 gxh3 'iVxh3 19 bxc5 l:tg8 20
"iih3 ~h7 31 "VJiIf5 ~h8 32 "VJiIh5 "VJiIf8 33 J:!.gl J:!.xg1+ 21 'iVxgl 'tWf3+ with perpetual
i..xe5 1-0 check in Neumarker-Schefter, corr. 1984.
b) 10 ttJc3 'iVd4 11 d3 0-0 120-0 'iYh4 13
Game 19 ~h 1 .txh3 14 gxh3 .txf2 and according to
A. Petrosian-Mikhalchishin Gligoric the position is unclear. Taking on h3
Dortmund 1998 is solely justifIed by winning the f2 instead of
the h3-pawn. Still White might have the bet-
e4 e5 2 lLlf3 lLlc6 3 i..c4 lLlf6 4 lLlg5 ter chances here.
d5 5 exd5 lLla5 6 i.b5+ c6 7 dxc6 bxc6 c) 10 c3 i.xh3 11 gxh3 ttJe4 12 O-O! (not
8 i..e2 h6 9 lLlh3 i.c5 12 'iVa4? i.xf2+ 13 ~d1 ~d5 and Black is
In the most recent edition of ECO this is much better) 12... i.b6 13 b4 ttJb7 14 i.f3
the second main line (after 9... .td6). But in 'iYd315 .txe4 'iYxe416 ~g4 ~xg4+ 17 hxg4
my opinion Black will fInd an easier game in h5 18 g5 0-0-0 gives Black excellent compen-
the sidelines 9...g5 and 9... .te7. So why are sation in a complicated queenless middle-
these moves not the main lines? How is this game.

48
 4 0,g5 d5 5 exd5 0,a5: Main Line

10 ... 0,d5 Perhaps preferable was 13 ttJxd5 cxd5 14

An interesting alternative was 10...g5!? c3 ttJc6 15 d4 exd4 16 ttJf4 and White is
with the idea of 11 ttJc3 g4! 12 ttJg1l:tbS and slightly better according to Mikhalchishin.
Black has good counterplay. In fact this is the 13 ... 0,b7 141t>h1 g5!
same position that arises after 9...g5 10 d3 g4 Black prevents £1-f4 and retains his space
11 ttJg1 i.c5 12 ttJc3l:l.bS. advantage.
11 0,c3 150,a4
If 11 0-00-0 (not 11..:iih4?! 12 ~e1 ttJb7 Black is much better after 15 ttJxd5?! cxd5
13 i.f3 i.xh3 14 gxh3 'it'f6 15 ~e2 ttJd6 16 16 c3 i.b6 when the black pawns looks very
l:te1 0-0 17 ttJc3 ttJxc3 1S bxc3 or impressive.
11...i..xh3?! 12 gxh3 ~4 13 i.f3 'iixh3 14 15 ... ~d6 16 f3 ~e6 17 0,f2 "VJIie7 18 c4
i.g2 ~e6 15 !le1 White is better) 12 c4 ttJc7 0,f6 19 i.e3 c5 20 0,c3 0,d8!
13 ~hl i.xh3 14 gxh3 ttJe6 15 i.e3 Heading for d4.
(Veinger-Rytov, USSR 1972) and now not 21 g4!
15 ... ttJd4 (as played by Rytov) as 16 ttJc3
gives White the slightly better game, but
15 ...l:l.bS! 16 ~d2 i.d4 with excellent play.
11 ... 0-0
Black can also play l1...ttJb7 120-00-013
~h1 g5!? 14 i.f3 f5 with good play, or
11...l:tbS 12 0-0 g5!? with an unclear game.
120-0
This position is very similar to that after
9... i.d6, the only difference being that here
the bishop is on c5.
12 ... f5
A very natural move indeed. If instead
12 ... ttJxc3?! 13 bxc3 ~h4 14 i.f3! (here 14 Just in time! Now '>X'hite gets control over
~h1?! does not work since because the f2- e4.
pawn is weak, i.e. 14... i.xh3 15 gxh3 i..x£1) 21 ... 0,c6 22 gxf5 ~xf5 23 0,fe4 0,d4 24
14... i.xh3 15 gxh3 ~xh3 16 i.g2 and White 0,xf6+
stands better. 24 h4? looks tempting, but after 26 ... i.xe4
25 fxe4 'iVe6 26 ~g2l:tabS 27 b3 gxh4 White
has problems.
24 .. :~!Vxf6 25 0,e4 "YIig6 26 "VJIid2 ~e7 27
1:[g1 It>h8 % -%
The position is about even: the black
knight on d4 is as valuable as the white one
on e4. Nevertheless both players should be
ashamed for not playing on.

Came 20
Ciocaltea-Nezhmetdinov
Bucbarest 1954

13 ~d2 The following game does not have great

49
Two Knights Defence

theoretical value. It is, however, very instruc- Black is much better) 13... 0-0 14lbc3 ~bS 15
tive. W'hen I teach my pupils and present 'iVc2 .l:!.eS gives Black excellent play for the
them with a position, they often ask me pawn.
where is compensation for the pawns? I 11 l2Jg4?
show this game and the questions are an- An instructive mistake.
swered. 11 ... .ltxg4!
1 e4 e5 2 l2Jf3 l2Jc6 3 .ltc4 l2Jf6 4 l2Jg5 Time is more important than the relative
d5 5 exd5 l2Ja5 6 .ltb5+ c6 7 dxc6 bxc6 values of bishop or knight in this position.
8 .lte2 h6 9 l2Jf3 e4 10 l2Je5 12 .ltxg4 .ltc5 13 .lte2
White also has problems after 13 0-0 h5
14 ..te2 lb g4 15 g3 lbxh21 (weaker is
15... lbxf2? 16 Mxf2 h4 17 d4 exd3 1S 'Yi'xd3
'Yi'b6 19 'i¥f51 .txf2+ 20 'Yi'xf2 hxg3 21 'Yi'xb6
gxh2+ 22 'it>h1 axb6 23 .tf4 with unclear
play) 16 'it>xh2 h4 17 ~g2 i.xf2 18 Mxf2
hxg3 19 'Yi'g1 gxf2 20 ~xf2 'Yi'f4+ 21 'it>e1
Mh2 22 ~d1 ~4 23 .tfl lbc4 and Black
has an winning attack.
13 .. Jid8 14 c3

Now Black has three good moves:

10...'Yi'c7 (the current game), 1O .....tc5 (Game
21) and 1O ... ..td6 (Games 22-24) - and one
not so good: 10 ...1Vd4?1 when after 11 f41
..tc5 12 Mfl White is better in all lines, e.g.
12... .td6 (if 12.. :~tNdS 13 c31 lbd5 14 'Yi'a4
'Yi'h4+ 15 'it>d1 0-0 16 'Yi'xe4 MdS 17 d4, or
12... .tb6 13 c3 'iVd6 14 b4lbb7 15 lba3 0-0
16lbac4 'iVc7 17 a4, or 12...g5 13 c3 'iVd6 14
d4) 13 c3 'iVb6 14 'Yi'a4 0-0 15 b4 lbb7 16
'Yi'xc6 'Yi'dS 17 lba3 as 1S b5 MeS 19 lbac4 14... l2Jb7!
.i.c5 20 .ta3 1:.e6 21 ..txc5 l:txc6 22 lbxc6 This example is worth remembering.
'Yi'eS 23 .td4 and White was clearly better in Black improves the position of his worst
Kuperman-Van Oosterom, corr. 19S5. placed piece.
10 .. :~c7!? 150-0 h5!
With normal play this move should trans- Targeting the kingside dark squares.
pose into 10.....tc5 or 10... i.d6 lines. Here 11 16 d4
d4 exd3 12 lbxd3 .td6 is Games 23 & 24, This does not look good, but it is hard to
while 11 f4 exf3 12 lbxf3 .td6 is covered in find a good alternative; e.g. if 16 b4 .tb6 17
Game 22. 'Yi'e1 ~f8! followed by lb g4 and the black
After 11 f4 Black can also play 11 ... .tc5 attack is probably decisive.
when 12 c3 i..d6 (or 12... lbb7) is Game 21. 16 ... exd3 17 .ltxd3l2Jg4 18 'iVe2+ ~f8!
Note that here 12 d4?1 exd3 13 cxd3 (if 13 There is no need to worsen the black
lbxd3? .tb6 14 b3 0-0 15 i.b2 lbd5 and bishop's position. After 1S... .te7? 19 g3 the

50
 4 0,g5 d5 5 exd5 0,a5: Main Line

position would be less clear. This move is underestimated by theory.

19 g3 "Yi'd7 20 ~e4 h4 21 ~f4 11 c3
The best reply, preparing d2-d4 or b2-b4.
If 11 f4 ~6 12l:tfl .tgl!, or 11 O-O?! 'iVd6!
(not 11...'iVd4? 12lLlg4 .txg4 13 .txg4 e3 14
~f3! exf2+ 15 'It>hl and White is better) 12
lLlg4 3Lxg4 13 jLxg4 h5 14 ~e2 lLlg4 and the
black attack is mortally dangerous.
11 ... ~d6
Black has also tried 11...'iVc7 (11...0-0!? is
possible too) and now:
a) 12 d4?! exd3 13 lLlxd3 3Ld6 seems to
give Black excellent compensation. Com-
pared with 10 ... 3Ld6 11 d4 exd3 12 lLlxd3
"iVc7 (Games 23 & 24), the additional c2-c3
21 ... 0,xh2! does not improve White's position. For ex-
Simple, but nice. ample: 14lLld2 jLf5 15 b4lLlb7 16lLlc4 !:i.d8
22 :ete1 (also interesting was 16 ... iLxh2!? since if 17
If 22 'It>xh2 hxg3+ 23 'it'g1 "iVh3 and mate g3?! Ji.xg3 18 fxg3 "iVxg3+ 19 Wd2 0-0-0 with
is coming. a crushing attack, while after 17 .te3 ~e6 18
22 ... 0,g4 23 .1&.f3 0,xf2 24 ~e3 hxg3 25 lLld2 the game is unclear) 17 ~e3 O-O! 18
~xc5+ 0,xc5 26 ~xc6 0,h3+ 27 ~f1 .txa7?! (18 h3, preparing to castle short,
"Yi'f5+ 0-1 looked much better) 18 .. .'~Jd5 19 Ji.d4 .l::tfe8
28 "iVf3 lLlf4 29 ~xg3 lLlfd3+ 30 ~f3 20 lLlxd6 lLlxd6 21 0-0 (at fIrst sight it might
.l:!.h 1+ 31 'it'g2 lLlxe 1+ is terminaL seem that White is winning, but this is an
r-----------------. illusion - actually he is under great pressure)
Game 21 21...lLlb5! 22l::tc l
Groszpeter-Hazai
Hungary 1998

1 e4 e5 2 lLlf3 lLlc6 3 ~c4 lLlf6 4 0,g5

d5 5 exd5 0,a5 6 ~b5+ c6 7 dxc6 bxc6
8 ~e2 h6 9lLlf3 e4 10 lLle5 ~c5!?

22 ... lLlbxc3! 23 .l:!.xc3? (the sad alternative

was 23 3Lxc3 lLlt4 24 .l::tel i.xd3 25 .tf3
.te2 26 "iVc2 3Lxf3 27 gxf3 c5! and Black is
better) 23 ... lLlxc3 24 ~xc3 .l:!.xe2 25 'iVxe2
.txd3 26 'iVg4 f5 27 'iVh4 i.xfl 28 'it'xfl
'iVd7 29 f3 ~d3+ 30 Wf2 l::ta8 and White

51
Two Knights Defence

resigned in Sutovsky-Postny, Tel Aviv 2001. 1996, saw 17 ... c4?! which is impressive, but
b) 12 f4 ttJb7 (12 ... i.d6 transposes to the not correct. Nevertheless, modern chess is
game) and then: much more than just mathematics and pat-
bl) 13 d4?! exd3 14 ~xd3 0-0 15 ttJd2 tern recognition. It is also a psychological
ttJd6 16 ..if3 ..if5 17 'iVe2 ttJd5 and Black is fight. To play the attack some material down
much better. is easier in practical terms than defending,
b2) 13 'iVa4 ..ltd7 14 ttJa3 0-0 15 b4 ..ib6 and so such risk taking can be justified. The
16 ttJac4 ttJd6 17 ttJe3 (or 17 ttJxb6 axb6 18 game continued 18 ttJdxc4 ttJxc4 19 ttJxc4
'iWb3 ..ie6 19 ~1 b5 with compensation) 'iVh4 20 g3 ..ic5+ 21 ..ie3 ttJxg3 22 i.xc5
17...a5 18 'iVc2 ttJd5 and Black had full com- ttJxe2+ 23 'iVxe2 ]:tfe8 24 ~f2 ~5 25 ..ie3
pensation for the pawn in Estrin-Dannberg, ':'ad8 26l:!.xd8 ':'xd8 27 ttJd2 ~d5
corr.I965.
b3) 13 b4 ..ib6 (after 13. .. ..id6 14 d4 exd3
15 'ii'xd3 0-0 160-0 as 17 i..f3 ..ixe5 18 fxe5
'iVxe5 19 i.xc6 axb4 20 ~5 'iVxb5 21 i.xb5
the position was equal in Mednis-Spassky,
Antwerp 1955) 14 a4 as 15 b5 0-0 16 bxc6
ttJc5 17 ttJa3 ttJd5 18 g3 ..ih3 and Black had
a strong initiative in Jovcic-Bohak, corr.
1972. The game continued 19 ..ifl f6 20
ttJb5 'iVc8 21 ttJd7 ttJd3+ 22 ..ixd3 exd3 23
ttJxb6 ~xc6! and White was in difficulties,
since if 24 ttJxa8? ':'e8+ 25 '>tf2 ttJxc3! leads
to mate.
12 f4 28 ttJf3?? (A horrible move; instead after
28 '>tfl! Black's attack does not compensate
for the sacrificed material, e.g. 28 .. J::td6 29 c4
'iVh1+ 30 'iVg1 ~c6 31 f5 lH6 32 'it'e2 and
White should win) 28 ... .:.d6 (White was
probably hoping Black would be satisfied
with regaining some material) 29 ':'fl .l:!.g6+
30 'it'h1 ~5! (White has no real defence
against the primitive ...'iVh3 and ....l:!.g3) 31
..ixa7? (but if 31 f5 'iVxf5 32 ..id4 'iVh3 33
..ie5 Itg5 34 .tf4 l:!.f5 35 'iVg2 ~xg2+ 36
'it'xg2 ':'xf4 and the endgame is clearly better
for Black) 31...~h3 32 ~e2 ':'g3 0-1.
13 0-0 0-0 14 d4 exd3 1 5 .ltxd3
12 .. .'iiHc7 If 15 'iVxd3 ':'d8 16 'iVc2 ttJd5 17 b4 ttJb7
Here 12... 0-0!? looks better, e.g. 13 0-0 c5 (Skrobek-Sydor, Lodz 1980) 18 iH3 ..ie6
(13 .. :~Vc7 returns to the game) 14 d4 exd3 and Black has enough compensation for the
(14... cxd4 15 cxd4 ':'e8 16 ttJc3 ~6 is an pawn.
alternative) 15 'iVxd3 ..ib7 16 :d1 ttJe4 (or 15 ... J:!.d8 16 'i¥e2 l:te8 17 b4 li:Jb7 18
16 ... ..ic7!?) 17 ttJd2 and now 17 ... ttJxd2 18 li:Ja3.ltg4
..ixd2 'iVb6 19 ..ie3 l.'tad8 with compensa- Or 18... a5 19 ttJac4 axb4 20 ttJxd6 'iVxd6
tion. Instead Chandler-Hebden, England 21 cxb4 'iVxb4 22 ..ib2 and White is slightly

52
 4 CDg5 d5 5 exd5 CDa5: Main Line

better because of the powerful dark-squared lIad8) 14...lZ'ld5 15 b4 lZ'lb7 16 ~b2 (16
bishop. lZ'lxd5 cxd5 17 d4 f6 18 c4 fxe5 19 dxe5 dxc4
19 ~f2 .Ili.xe5 20 fxe5 'i1Kxe5 21 .Ili.b2 20 exd6 lZ'lxd6 21 'iVd5+ ~h8 22 Sl.b2 is met
%-% by 22 ... c3! 23 ~xc3 ~xc3 24 ~xd6 ~e3+ 25
The game was agreed drawn, though .l:!.f2 .l:.ac8 with fIne compensation) 16 ....l:.ae8
White is slightly better after 21..:i¥c7 22 17 g3 as 18 lZ'lc4 axb4 19 lZ'lxd5 cxd5 20
.l:tae1. lZ'lxd6 'ifb6+ 21 ~g2 lZ'lxd6 22 axb4 lZ'lc4
.....- - - - - - - - - - - - - - - - , . with suffIcient compensation in Timman-
Game 22 Gligoric, Bad Lauterberg 1977.
Vukcevich-Romanishin 12 CDxf3 0-0
Hastings 1976/77 Black should develop fIrst, then attack.
The wild 12... lZ'lg4? 13 0-0 ~c7 14 h3 .th2+
e4 e5 2 CDf3 CDc6 3 .Ili.c4 CDf6 4 CDg5 15 ~h 1 h5 does not really threaten anything,
d5 5 exd5 CDa5 6 .Ili.b5+ c6 7 dxc6 bxc6 and after 16 d4 ~g3 17 ~d3 Sl.e6 18 'iV e2
8 .Ili.e2 h6 9 CDf3 e4 10 CDe5 .Ili.d6 0-0-0 19 c4 White is much better. Black can
The most popular move. also play 12 .. :YWc7 fIrst, transposing below
11 f4 after 13 0-0 0-0.

After this Black has no problems at all, al- 13 d4

though 30 years ago Estrin was convinced 13 0-0 will transpose to the next note if
that this was the best solution. The usual 11 White follows with 14 d4. Otherwise:
d4 is covered in Games 23 & 24. a) 13. ..'iVc7 14 b3?! (14 d4) 14....l:.e8 15
Weak is 11 lZ'lg4? ~xg4! (better than Sl.b2? lZ'lg4 16 h3 Sl.c5+ 17 d4 lZ'le3 and
11...lZ'lxg4 12 .txg4 'ifNh4 13 .txc8 l:1xc8 14 Black was much better in Djordjevic-Truta,
h3 0-0 15 'iVg4 'iVe7 16lZ'lc3 f5 though Black corr. 1980.
still has compensation) 12 ..txg4 ~c7 13 b) 13. .. c5!? 14 b3 (14 d4) 14... ..tb7 15 ..tb2
~h3 0-0 14 g3 lZ'ld5 15 0-0 l:i.ae8 and Black l:!.e8 16 lZ'la3 ..tc 7 17 lZ'lc4 lZ'lc6 gives Black
is much better. good compensation for the pawn.
11 ... exf3 13 ....:.e8
The best response. After 11...'i¥c7 12 0-0 Also interesting are:
White has some chances of gaining an advan- a) 13. ..'i¥c7!? 14 0-0 cS 15 lZ'lc3 a6 16 d5
tage, though Black will still have compensa- ':e8! (better than 16 ... ..tb 7 which blocks the
tion; e.g. 12... 0-0 13lZ'lc3 Sl.f5 14 a3 (or 14 d4 escape route of the knight on as and closes
exd3 15 Sl.xd3 Sl.xd3 16 'iVxd3 ~fe8 17 ~e3 the half open b-file; even so after 17 'itthl

53
Two Knights Defence

J:!.feS 18 ttJh4 J..e5 Black is fIne) 17 'it>h1 This looks natural, but the white queen
l:!.bS 1S a3 ttJg4 19 h3 ttJe3 20 J..xe3 l:l.xe3 really has plenty of squares. Better was
21 l:!.b1 'iIIe7 and Black had great compensa- 1S ... ttJb4! 19 ttJe1 ttJe4 and then if 20 'fih5?
tion in Estrin-LevenfIsh, USSR 1949. g6! (not 20 ... ttJf6? 21 'iIIfS ttJbd5 22 J..f3 and
b) 13 ... c5!? 14 0-0 (after 14 dxc5?! J..xc5 White keeps the extra pawn with a good
15 'iIIxdS l:!.xd8 White has some problems position) 21 'iVxh6 ttJxc3 22 bxc3 l:!.xe2 23
with his king) 14... cxd4 15 'It>h1 J..c5 16 c3 J..g5 ii.f8 24 'iYh4 'fid5 25 ii.f6 J..g7 26
dxc3 17 ttJxc3 with equality. Black's activity ii.xg7 Wxg7 27 cxb4 l:he1! and Black wins,
will give him a draw, but probably nothing while if 20 'iIIg4! ~e6 21 a3 4Jxc3 22 bxc3
more. ttJd5 with excellent play for the pawn.
140-0 c5 19 ..td2 liJg6
Not 19... ttJfg4?! 20 'iVxd8 l:taxd8 21 l::tad1
ttJg6 22 J..b5 and White is better.
20 iYd4liJe4?
Black is on the wrong track here. Both al-
ternatives looked better:
a) 20 ... ii.c7!? 21 'iIIxdS l:taxdS 22 .l:i.ad1
ttJg4 with some compensation.
b) 20 ... ..txh2!? 21 'iYxd8! (if 21 J..xh6?!
iYxd4 22 ttJxd4 i.e5 23 i.e3 ttJe4 with a
dangerous initiative for Black, or 21 'it>xh2?!
..txf3 22 'iYxd8 l:i.axd8 23 i.d3 J..e4 and
Black is slightly better) 21...1:!.axdS 22 ..txh6
J..d6 23 ii.g5 CLJh5 24 J..xd8 CLJg3+ 25 'it>g 1
15 'it>h1 ii.c5+ 26 Wh2 CLJxf1+ 27 ':'xfl l:!.xdS with
White cannot open the game. If 15 dxc5? compensation for the pawn.
J..xc5+ 16 Wh1 'iVxd1 17 i.xd1 ii.a6 1Sl:!.e1 21 ..tc4 ..tf4 22 ..txf4liJxf4
ttJe4 and Black wins.
15 .....tb7 16 liJc3
Or 16 ii.b5 .l:i.e7 17 ttJc3 ttJc6 1S dxc5
i.xc5 19 i.f4 'iVb6 with compensation.
16 ... cxd4 17 'tlfxd4 liJc6 18 'ifh4

Black has some compensation which, after

a hard defence, should be enough for draw,
but nothing more.
23 £lad 1 'iVxd4 24 J:t.xd4 J:tad8 25 £lxd8
J:txd8 26 liJxe4 ..txe4 27 'it>g 1 g5 Y:z - Y:z
18 ... liJe5 Black should have to fIght for a draw here,

54
 4 tbg5 d5 5 exd5 tba5: Main Line

but White assisted him by simply offering it centre; e.g. 12 c3 (if 12 ..ibS+ ~e7! 13 4Jc3
to him. Naturally Black accepted. Offering 1Wb6 14 dxcS i.xcs 1S 'iYe2 lId8 gives Black
the draw was White's worst move in the fine compensation, but not 12... ~f8? 13 dxcS
game! Instead, after 28 c3! Black has no good and White wins) 12... 0~0 13 O~O ~c7 14 f4
moves, e.g. 28 ...4Jxg2 29 .ltxf7+! <j;;;g7 30 .l:.b8 with compensation for the pawn. Still,
<j;;;xg2 <j;;;xf7 31 <j;;;f2 or 28 ... .ltxf3 29 ':xf3 the main line seems to be more natural.
'ud1+ 30.ltfl or 28 ...<j;;;g7 294Jd4 and White 12 lLlxd3 "iIIc7
is clearly better.

Game 23
L. Belov-Nezhmetdinov
Omsk 1961

1 e4 e5 2 lLlf3 tbc6 3 ~c4 lLlf6 4 lLlg5

d5 5 exd5 lLla5 6 ~b5+ c6 7 dxc6 bxc6
8 ~e2 h6 9 lLlf3 e4 10 lLle5 ~d6 11 d4

13lLld2
Here White has tried a wide range of
moves:
a) 13 b3 is covered in Game 24.
b) 13 ..id2?! O~O 14 .txaS 'iYxaS+ IS ~d2
'iVc7 16 4Jc3 ttb8 17 h3 cS with strong com~
pensation in Niemi~Tuomala, Finland 1996.
c) 13 f4? is just a weak move, and Black
stands better after 13 ... 0~O 14 O~O ..ifS 1S
In recent years this has been the most 4Jc3 'uad8 16 ~el lIfe8.
popular move, and it also looks the soundest, d) 13 i.e3?! cS 14 4Jc3 4Jc4 1S lic1 O~O
so probably 11 d4 will continue to be the 16 ..if4 ..ixf4 17 4Jxf4 ttJxb2 also leaves
main line for some time to come. Black with a better game.
11...exd3 e) 13 b4?! is not prudent. After 13 ... ttJc4
The best solution. The knight on eS is dis~ 14 ttJd2 ttJeS! (better than 14... ttJxd2 IS
turbing Black's game too much. Also, Black .txd2 O~O 16 h3 .tf5 17 O~O a draw was
has a lead in development and should there~ agreed in Mikhalchishin~Geller, Dortmund
fore open the position when he has the 1991) 1S a3 4Jxd3+ 16 i.xd3 O~O 17 .tb2
chance. 'ue8+ Black has a strong initiative.
11...lic7?! is too slow. After 12 .ltd24Jb7 £) 13 h3!? is interesting; White prepares to
13 O~O O~O 14 4Ja3 ..ie6 IS ~ c1 ~fd8 16 castle as soon as possible. Nevertheless, after
4Jac4 ..ixc4 17 ..ixc4 .ltxeS 18 dxeS ~xeS 13 ... 0~O 14 O~O ..ifS 1S ttJd2 'uad8 16 'ue1
19 .tf4 White was much better in Bogol~ ttJd5 17 .ltfl cS Black had the usual compen~
jubow~Zimrnermann, Zurich 1928. sation in Kholmov~Geller, Elista 1995.
l1...cS!? (Nenashev's recommendation) 13 ... ~a6 14 lLlf3 0-0 , 5 0-0 J:tad8 16 b3
could be another way to attack the white l:tfe8

55
Two Knights Defence

troduce the knight on as to the game. White

cannot survive against this attack.
22 c4 .ii.xe1 23 <t>xe1 lLlxc4!
Black is unstoppable.
24 bxc4 iLxc4 25 <t>f2
If 25 'It>f1 J:Ixe2 26 'It>xe2 ':xd3 27 iVxd3
~xd3+ 28 ~xd3 ~f6 29 ~bl ~f5+ and
30 ... ~xbl, or 25 ~dl ~a6 26 ctJfel ~xd3
27 ctJxd3 (or 27 ~xd3) 27 ... c4 and Black
WillS.
25 ... iLxd3 26 .ii.xd3 c4+ 27 <t>g3 l:txd3
28 "i¥b2
Against the two rooks White is helpless.
Black is now fully developed and will im- 28 .. :YWg6+
mediately start to attack the white position. Or 28 ... ~xG+!? 29 gxG (29 'It>xG ':e3+
White is not ready for this, but if he devel- wins the queen) 29...~g1+ 30 ~g2 ~e1+ 31
ops, he might ride out the storm. This does ~f2 iVe5+ picks up the rook on al.
not happen in the game! 29 <t>f2 'iVe4 30 iLd2 J:!.xf3+! 31 gxf3
17 l:te1? ~h4+ 0-1
This weakens the dark squares around the After 32 ~f1 ~xh3+ 33 'It>f2 iVh2+ 34
white king, something Black immediately 'It>f1 ~e2 wins.
exploits. Better was 17 ~b2 ctJe4 18 c4.
17 ... lLlg4 18 h3 Game 24
If 18 g3 ~c5 19 ~f1 ctJxf2! 20 ':xf2 "i¥b6 Morozevich-Onischuk
and White loses material. Moscow 1996
18 ... lLlxf2! 19 <t>xf2
Forced, since if 19 ctJxf2 ~h2+ and the 1 e4 e5 2 lLlf3 lLlc6 3 iLc4 lLlf6 4 lLlg5
queen is lost. d5 5 exd5 lLla5 6 .ii.b5+ c6 7 dxc6 bxc6
19 ... ~b6+ 20 <t>f1 iLg3 21 ~d2 8 iLe2 h6 9 lLlf3 e4 10 lLle5 .ii.d6 11 d4
Otherwise something would take on d3. exd3 12 lLlxd3 'iflic7 13 b3

21 ... c5!! The main line. This move has two ideas. It
The brilliant point of the combination! allows the fianchetto development of the
This superb move has but one idea: to rein- bishop, and takes control of the c4-square.

56
 4 0:.g5 d5 5 exd5 0:.a5: Main Line

13 ... 0-0 bl) 16 O-O?! (Tal thought this should give

This is the most natural reply. However, White better play, but...) 16 ... CLlxe2+ 17
13 ... c5!? has also been investigated: CLlxe2! (or 17 'iVxe2 .ixh2+ 18 ~hl .id6 and
a) 14 CLla3 l:tb8! 15 .if3? (15 c4 is unclear) the position is unclear) 17 ... .ixh2+ 18 ~h 1
15 ... 0-0 16 c4 .ifS 17 CLlb5 l:txb5! 18 cxb5 c4 .id6 19 CLld4 .l:tb8 20 c4 i.e7! 21 'iVh5 i.f6
19 bxc4 l:td8 and Black's was crushing in and Black is very close to equality.
Fritz-Malinin, corr. 1989. b2) 16 CLlxf4 ~xf4 17 h3 (after 17 g3 .l:i.d8
b) 14 c4 0-0 (White is better after 18 .i.d3 l:!.e8+ 19 CLle2 .i.g5 20 h4 .i.e 7 21
14... CLlc6!? 15 CLlc3 a6 16 .ie3, or 14... ..tb7?! 'iVd2 c5 and Black has sufficient compensa-
15 CLlc3 .ixg2 16 CLlb5 'iYe7 17 l:tgl and tion according to Tal) 17 ... .l:tb8! (if 17 ...l:td8
White has the initiative according to Blatny) 18 ..td3 c5 19 0-0 c4 20 CLlb5 WNc6 21 bxc4
15 ..tb2 CLle4 16 CLlc3 CLlxc3 17 .ixc3l:te8 18 CLlxc4 22 .i.e4 and \x'hite is better according
h3 .ifS 19 0-0 Mad8 with excellent compen- to Palkovi; 20 ... "iVb 7! is a possible improve-
sation for the pawn. ment with an unclear position, but still,
c) 14 .ib2 c4 15 bxc4 CLlxc4 16 .ixf6 gxf6 17...l:tb8 looks stronger) 18 WIId4 c5 and
17 O-O!? is an idea of Sutovsky, and then Black has great compensation; e.g. 19 'iVe4
17 ... .ixh2+ 18 ~h 1 .ie6 19 CLlc3 l:td8 (or J:tb4 20 'iVo .i.b7 21 CLld5 .ixd5 22 Wlixd5
19 ...CLla3!? 20 CLle4 f5 21 CLlf6+ ~e7 with an l:te8 and White is in trouble.
unclear position) 20 'iVel i.e5?! (better is 150:.c3
20 ... CLla3 since if 21 g3? CLlxc2 22 'iVd2 CLlxal If 15 CLld2 f5!? with unclear play.
23 ~xh2 h5 gives Black a strong attack) 21 15 ... i.f5?!
f4 ..txc3 22 'iYxc3 .l:!.g8 23 .io f5 24 l:tfel This move seems natural, but now White
l:tc8 25 i.d5 CLlb6 26 WIIxc7 Mxc7 27 ..txe6 can seriously consider exchanging on e4,
fxe6 28 l:txe6+ ~d8 29 l:tdl and White won opening up the game for his bishops. Better
in Sutovsky-Acs, Tel Aviv 200l. was 15 ... f5 16 h3 (if 16 O?! CLlc5 17 WIId2l:td8
14 ..ib2 0:.e4 or 16 f4 .i.a6 17 0-0 .l.:!.ad8 with compensa-
Black has the following alternatives: tion) 16 ... ..ta6 (or 16 ... .i.b 7!? 17 0-0 l:tad8 18
a) 14...1:i.e8 15 h3 CLle4 16 0-0 'iVe7 17 CLlc3 WIIel :Ife8 with good compensation) 17 0-0
CLlxc3 18 i.xc3 was tried in Kasparov- .l:i.ad8 18 'iVe1 c5 was Morozevich-Nenashev,
Timman, Moscow 1994, and now after the Alusta 1994, and now after 19 ~h 1 .ib 7 20
strongest 18 ... CLlxb3 19 cxb3 Wlixe2 20 l:tel .i.o the game is unclear according to Moro-
'iYxdl 21 .l:!.axdl White is slightly better. zevich.
b) 14... CLld5!? 15 CLlc3 CLlf4 and now: 16 h3 .l:!.ad8 17 0-0 .l:!.fe8

57
Two Knights Defence

1S il..f3?! enter the attack (bar the rook on a1).

Better was 1S ctJxe4! l::txe4 19 ..tD J::te6 26 .. :YWc7
(or 19 ...lIees 20 l::tel and White is much If 26 ... ctJxe4 27 'iVxg6+ ~f8 2S dxe4 'iVf7
better here too) 20 ..tg4 ..txg4 21 Vi'xg4 and 29 'iVxh6+ ~eS 30 'iVhS+ Vi'fS 31 ~h5+ ~f7
White was clearly better in Howell-Pein, 32 ~xa5 and wins.
Wrexham 1995. 27 il..e5 'iVcs 2S <1:Jf6+ >t>g7
1S ... <1:Jg5 19 il..g4 il..g6 20 >t>h1 'ifb7?
A very slow move, indicating ... ctJa5-c4,
but not really doing much else. Stronger was
20 ... j,b4! and Black still has compensation
for the pawn.
21 f4
White is clearly better.
21 ... il..bS 22 il..h5 il..xd3 23 cxd3 J:!.e3
After 23 ...ctJe6 24 Vi'g4 ctJd4 25 ctJe4 wins.
Now Black hopes for 24 fxg5? 'iVc7!, but...
24 'tIVg4! g6

29 'iVxcS
Simplest, even for a grandmaster. When
an excellent GM like MorOl~evich sees a win-
ning endgame he will often play it immedi-
ately. However, stronger was 29 .tb2! ctJxh3
30 ctJd5+ .te5 (30 .. .<.ii'fS 31 ctJxe3) 31 fxe51~
Vi'xg4 32 e6+ with a nice mate after 32 ... 'iith-
33l::tf7+ <.ii'gS 34 ctJe7.
29 ... J:.xcS 30 il..xbS J:txbS 31 fxg5 hxg5
32 <1:Jg4 nxd3 33 fl.ae1 J:tb7
Black is lost because of his weak pawns.
Again if 24... ctJe6 25 ctJe4 ctJd4 26 j,xd4 weak knight and weak king.
l::txd4 27 ctJg3! and White wins; the threat of 34 fl.eS :ad5
ctJf5 is too strong. Black cannot save the game with 34 ... l::td2
25 il..xg6! due to 35 l::tffS .l:i.xa2 36 :tgS+ <.ii'h 7 (or
White shows no restraint, but simply 36 ... 'iitf7 37 ctJe5+) 37 ctJf6+ 'iith6 38 l::te6!
hacks his way through to the black king. and White wins.
25 .. .fxg6 26 <1:Je4! 35 J:!.ffS c5 36 :acS J:.e7 37 :gS+ ~f7
Now all the remaining white pieces will 3Sl:!.cfS+ 1-0

58
 4 0,g5 d5 5 exd5 0,a5: Main Line

Summary
Against Steinitz' 9lbh3 the latest edition of ECO recommends 9... ..Itd6 and 9... ..ItcS as the best
moves, but it appears that Black will fInd his best chances with 9...gS! as in Game 17. After this
move White does not appear to have any chance of obtaining an advantage.
In the main line 9 lbf3 e4 10 lbeS Black can hope for good counterplay with all three stan-
dard moves: 10.. :Yi'c7, 10 ... ..ItcS and 10.....Itd6. The usual sequence these days is 10 ... i.d6 11 d4
exd3 12 lbxd3 VJlic7 13 b3, which offers a complicated and dynamic middlegame in which
where Black has compensation for his pawn, but probably no more than that.

1 e4 e5 2 0,f3 0,c6 3 Jic4 0,f6 4 0,95 d5 5 exd5 0,a5 6 Jib5+ c6 7 dxc6 bxc6 8
Jie2 h6 9 0,f3
9lbh3 (D)
9... i.e7 - Game 16
9 ...gS - Game 17
9... ..Itd6 - Game 18
9...i.cs - Game 19
9 ... e4 10 0,e5 (D) Jid6
lo ...1Wc7 - Game 20
1O ... i.cS - Game 21
11 d4
11 f4 - Game 22
11...exd3 12 0,xd3 Wic7 (D)
13lbd2 - Game 23
13 b3 - Game 24

90,h3 100,e5 12"."Iic7

59
 CHAPTER FIVE I
Traxler Gambit:
4 l2Jg5 Jtc51?

1 e4 e5 2 tbf3 tbc6 3 i..c4 tbf6 4 tbg5

i..c5!?
This enterprising sacrifice, offering a rook,
was first played by the Czech player
K.Traxler in the 1890's. Over a century later
not much has changed. If you want to play
4... i..cS as Black you must be good at tactics
and have a good memory. White has three
serious moves at his disposal here: 5 d4, 5
ctJxf7 and 5 i..xt7+l. Only the last move gives
White a real chance for an advantage.

Game 25
Grott-Leisebein 5... exd4? loses to 6 ctJxt7 'fiIe7 7 ctJxh8,
Correspondence 1998 but Black can also try 5... ctJxd4 6 ctJxt7 (if 6
' - - - - - - - - - - - - - -.. ~xt7+ ~e7 7 i..c4 .l:.f8 8 ctJc3 h6 9 ctJf3 d6
e4 e5 2 tbf3 tbc6 3 i..c4 tbf6 4 tbg5 with good play for Black in Gofstein-
i..c5!? 5 d4 Nakonechny, USSR 1961) 6.. :fie7 7 ctJxh8
After this move White does not have an d5 8 c3!? if (8 ~e2 dxe4 9 i.e3 i..fS 10 c3
advantage. The pseudo-Evans Gambit 5 b4? 0-0-0 11 cxd4 exd4 12 ~g5 ~b4+, Mednis-
has also been seen, but Black has a strong Santasiere, USA 1955, and according to ECO
counter-stroke in the form of 5... d5! 6 exd5?! the position is unclear) 8... dxc4 9 cxd4 ~xd4
(but if 6 bxc5 dxc4 7 ctJc3 ~g4 8 ctJf3 ctJd4 10 ctJd2 (if 10 ctJc3 i.g4 11 f3 0-0-0 12 ~a4
and Black is much better) 6... ctJxb4 7 d6 (not i.e6 13 ctJg6 hxg6 14 ~g5 'YWc5) 1O ... i..g4 11
7 O-O? i..g4 and wins) 7.. :fixd6 8 i..xt7+ (if 8 'YWa4+ ~d7 12 ~xc4 0-0-0 13 ctJf3 i..g4 14
ctJxt7 ~d4 9 0-0 i..g4! 10 'iVel ctJxc2 and i.e3 ctJxe4 15 ctJt7 ~e6 16 'YWa4 ~xt7 17
wins) 8...<;.t>f8 9 i..b3 (if 9 ctJc3 ~c6) 9... ~d4 ~xd4 exd4 18 0-0 'fiIc5 19 lladl ~e8 with
10 'iVe2 'fiIxal 11 'iVc4 i..xf2+! and Black good compensation for the exchange.
wms. 6 i..xd5?!
5 ... d5! If 6 exd5 ctJxd4 7 c3 ctJf5 8 0-0 ctJd6 9

60
 Traxler Gambit: 4 4:Jg5 i.c5

.lib3 0-0 and Black is at least equaL wins.

White's best continuation may be 6 dxc5 c) 9 f3 is equally horrible. See for yourself:
dxc4 7 ~xd8+ lbxd8 with a roughly equal 9...lbxd5 10 fxg4 lbb4 11 lba3 'iVh4+ 12 g3
ending. Then if 8 lbc3 h6 (probably best; if ~h3 13 c3 ~g2 0-1 Lichtanen-
8... lbe6?! 9 lbb5 lbxg5 10 .lixg5 and White Ostroverchov, corr. 1968-69, since after 14
was slightly better in Starostits-Trojacek, cxd4 ~xh1+ 15 cj;;d2 'ii'xe4 White will not
Trinec 1998, while after 8... 0-0 9 lbf3 lbc6 10 survive the attack.
.lig5 lbe8 11 0-0 f6 game is just equal) 9 d) 9 ~d3 lbxd5 10 c3 lbb4 11 cxb4
lbb5 O-O! (9 ... .l:tb8 10 lbf3 lbc6 11 lbxc7+ .lixb4+ 12 lbc3 (if 12 .lid2 O-O-O! with a
~d8 12 lbb5 lbxe4 13 .lie3 .lig4 140-0-0+ crushing attack) 12... 'ii'd7 13 cj;;f1 ii.xc3 14 f3
would leave White slightly better) 10 lbxc7 (not 14 bxc3?? .i.e2+) 14... .lia5 15 fxg40-0-0
.l:!.b8 11 lbf3 lbc6 with an even game, which with excellent compensation for the materiaL
is far from the same thing as a drawn. White is really in danger here.
6 ... 4:Jxd4 7 ... We7 8 i.c4
Not 6... lbxd5? 7 dxc5 lbf6 8 ~xd8+ After 8 c3 h6 9 cxd4 'ii'xd4 10 'iHxd4
lbxd8 9 lbc3 h6 10 lbf3 lbc6 11 i.e3 .lig4 i.xd4 11 .lib3 hxg5 Black was better in
120-0-0 and White is much better. Gobza-Rohlichek, corr. 1956.
7 i.xf7+ 8 ... b5 9 i.d3 h6
Risky is 7 lbxf7?! 'iVe7 8 lbxh8 .lig4 and Also interesting is 9...1:.£8 10 .lie3 h6 11
now: lbf3 i.g4 12 lbbd2 ~d6 'With compensation;
for example if 13 h3lbxf3+ 14 gxf3 .lie6! 15
'iHe2 a6 and Black has excellent play and no
need for the f-pawn anyway.

a) 9 ~d2lbxd5 10 exd5 (if 10 c3 lbf6 11

cxd4 .lib4 12 lbc3 lbxe4 13 'iVc2 lbxc3 and
White is caught in one hell of a storm)
lO ... ..tfS 11 0-0 lbxc2 12 g4 'iVh4 13 ~g5 10 c3?
'ilVxg4+ 14 'iVxg4 .lixg4 15 ..te3 ..txe3 16 This weakens the central light squares, al-
fxe3 lbxa 1 17 lba3 cj;;e 7 18 .l:!.f7+ cj;;d6 19 lows Black to open the h-ftle, and leaves the
.l:!.xg7 ..te2 20 lbf7+ ~xd5 21 Mxh7 and centre as a highway for Black. Much better
Black is better because of his more active was the simple 10 lbf3 ..tg4 11 lbbd2 'iHd6
pIeces. 120-0.l:!.hf8 13lbxd4 ii.xd4 14 .lie2 i.d7 15
b) 9 ..tf7+? cj;;£8 10 f3 is the computer's lbf3 .lic5 16 'ii'xd6+ cxd6! though Black has
favourite, but after 1O ... lbxe4 11 fxg4 ~h4+ good compensation because of the c-ftle and
12 g3lbxg3 13 .lig5 ~xg5 14 hxg3 'iWe3+ 15 a lead in development.
~f1 lbf5! it fmally understands that Black 10 ... hxg5 11 cxd4 'i!Vxd4

61
Two Knights Defence

Now White has a serious problem: how clear that the position is very complicated
will he complete his development? and White should be very careful. It is per-
12 'Vic2 haps playable in correspondence chess or
If 12 ~e2 .i:!.d8 13 i.c2 .ib4+ 14 ctJd2 after long and hard preparation, say for ex-
i.b 7 15 f3 g4 and Black is much better. ample 100 blitz games in addition to inde-
12 ... lLlg4 pendent analysis.
Black must do this right. After 12 ... i.b4+?
13 ctJc3 ~d8 14 i.xb5 'iVxe4+ 15 ~xe4
ctJxe4 16 0-0 i.xc3 17 i.c6 ':b8 18 bxc3
ctJf6 19 i.xg5 White is much better.
13 iLxg5+
After 13 ctJc3 'iYxf2+ 14 'iYxf2 ctJxf2 15
i..xg5+ ~e6 16 i.xb5 ctJxhl 17 i..c4+ ~d7
18 0-0-0+ i.d4 19 J:.xh 1 i.b 7 White does
not have enough for the exchange.
13 ... Wf7 14 iLxb5
Not 14 O-O? ':xh2 15 i.e3 ~d8! with the
idea of ... ~h8 or ... Mh1+ with mate coming,
or if 14 ctJc3 ctJxf2 15 ~fl 'it>g6 16 .i:!.xf2
'iVxf2+ 17 'iVxf2 i.xf2+ 18 'it>xf2 Wxg5 and 5 ... i..xf2+!
Black is much better. The point. Nothing else makes sense.
14... ~g6! 6 <;t>xf2
Avoiding the potential exchange of The alternative 6 Wfl is considered in
queens on c4. Game 27.
15 h4 6 ... lLlxe4+ 7 ~g1
Or 15 i.cl c6! 16 f3 ctJxh2 and Black Enormously complicated is 7 'it>e3!?, after
wins, e.g. 17 i.fl (if 17 i.c4 ctJxf3+ 18 gxf3 which Black has two possibilities:
l:!.xh1+ mates) 17 ...ctJxfl! 18 lhfl (18 J:.xh8
~g1) 18... i.a6.
15 .. J:tb8 16lLlc3 lLlxf2 17 l:!.f1 l:txb5! 18
lLlxb5
White could have strung things out a bit
with 18 'iVxf2 'iVxf2+ 19 Mxf2 i.xf2+ 20
Wxf2 l:!.xb2+.
18 ... lLld3+ 19 ~d2 iLb4+ 0-1
White resigned in view of 20 'it>e2 'iVxe4+
21 i.e3 Md8 22l:tf2 i.c5.

Game 26
Weir-Smits
Email 1994 a) 7... ~e7 and then:
a1) 8 c3 d5 (not 8...'iVc5+? 9 d4 exd4+ 10
1 e4 e5 2 lLlf3 lLlc6 3 iLc4 lLlf6 4 lLlg5 cxd4 ~e7 11 Me 1! l:!.f8 12 'it>f3 d5 13 i.xd5
iLc5 5 lLlxf7 i. g4+ 14 Wxg4 ctJf2+ 15 'it>g3 ctJxd 1 16
There was a time when 5 ctJxf7 was i.xc6+ bxc6 17 l::txe7+ ~xe7 18 ctJe5 and
thought to be the principal move. Now it is White wins) 9 i..xd5 ~c5+ 10 d4 (10 ~xe4

62
 Traxler Gambit: 4 tDg5 i.c5

iHS+) 10.. :~xdS 11 CLlxeS CLlf6 12 CLlf3! (12 CLle2+ and Black wins.
CLlxc6? ~e4+ 13 'it>f2 CLlg4+ 14 'it>g3 O-O! 15 8 ... tDxg3
1:.f1 Mxf1 16 'iVxf1 ~xc6 with an attack -
Palkovi) 12... 0-0 13 'ilVb3 i.e6 14 'iVxdS
CLlg4+ 15 We4 CLlf2+ 16 ~e3 CLlg4+ and the
game would end in an attractive perpetual
check.
a2) 8 CLlxh8! (critical) 8... ~gS+ (if 8... dS 9
'iUhS+! is strong) 9 Wxe4 dS+ 10 ~xdS i.f5+
11 WD ~g4+ 12 'it>f2 ~xdl 13 i.xc6+ bxc6
14 Mxd 1 'iVhS 15 1:.f1! 'iVxh2 16 d3 0-0-0 17
CLld2 followed by CLlD and ~e3, White has
good chances of consolidating.
b) 7 .. :~·h4 is very interesting and can be
analysed very deeply. First of all White only
has one move: 8 g3 CLlxg3 9 hxg3 'iVd4+ 10 9 tDxh8
'it'f3 and then: The only move since otherwise the black
b 1) 1O ... 0-0!? (this is very risky, but after all rook will live:
chess is only a game) 11 1:.h4 e4+ 12 Mxe4 a) 9 hxg3? ~xg3+ 10 'it'f1l:tfS 11 'iYhs dS!
(perhaps better is 12 'it>g2 dS transposing to gives Black an overwhelming attack, e.g. 12
lO ... dS) 12...CLleS+ 131heS ~xc4 14 ~g2 (if i.xdS CLlb4 13 ~c4 bS! 14 ~b3 (or 14
14 MfS ':'xf7 15 J!!.xf7 'iVxf7+ 16 'it>g2 b6 17 i.xbS+ c6 15 i.c4 CLldS 16 ~xdS cxdS and
d4 ~b7+ 18 dS iLxdS+ 19 Wgl iLD 20 ~f1 Black wins) 14...CLlxc2 15 d4 ~b7! 16 'iVxeS+
~hS and Black wins) 14...'iVxf7 15 ~e2 b6 ~xeS 17 dxeS CLlxal 18 .ti.xh7 CLlxb3 and
16 'it>gl iLb7 17CLlc3 "iUg6 18 ~d3 'iVg4 and Black won in the game Schatunov-Garin,
Black is no worse here despite having a piece corr.1973.
less. b) 9 d4? is strongly met with 9 ...CLle4! 10
b2) 10... dS 11 .l:.h4 e4+ 12 ~g2 0-0 when ~e3 exd4 11 CLlxh8 dxe3 and the pawn has
White has several tries: similar properties to those of a minor piece.
b21) 13 'iYhS Mxf7 14 'uf4 (14 ~xdS It is dangerously active and wickedly close to
'iYf2+ 15 'it>hl "iUf1+ is an immediate draw) the white king. Fedjanov-Tokarev, corr.
14... iLe6 15 iLb3 CLleS or IS ...g6 with an 1977-78, continued 12 iLf7+ (after 12 'iYD
unclear game. CLleS! 13 ~f7+ 'it'f8! Black has the following
b2) 13 CLlc3!? dxc4 (13..:iVxc4 14 Mf4!) 14 forced line at his disposal: 14 'iVxe3 'iYg4+ 15
'iYhs and now with 14... CLle7! (Bennedik) Wf1 'iYd1+ 16 'it>g2 'iVxc2+ 17 Wgl 'iYd1+ 18
Black has good counterplay, e.g. 15 CLlxe4 'it>g2 ~g4+ 19 ~f1 CLlxf7 20 CLlxf7 ~xf7 and
iLfS 16CLlfgS h6 17 CLlc3 iLg4 18CLlh3 'iVf6+ wins) 12... 'it>d8 13 "iUD CLld4 14 'ii'xe3 CLlxc2
19 'ii'xg4 'iYf1+ with a draw. 15 'ii'D 'iVe1+ 16 ~f1 'iVe3+ 17 Wg2 dS 18
b3) 13 i.b3 Mxf7 and now if 14 'iYgl 'iVeS 'iVD ~h3+! 0-1.
ISCLlc3 gS! 16 Mh6 ~fS or 14 'iYe2 ~e6 with 9 ... d5?
an attack, but the continuation 14 J!!.f4 Mxf4 This looks natural, but actually it loses be-
15 gxf4 iLe6 16 CLlc3 might give White an cause White has the extra options of 10 'iVD
advantage. and 10 ~e1. Also after 9... CLle4? 10 'iVf3!
7 ... ~h4 8 g3 White wins. Instead Black should play
Not 8 ~f1?? .l:.fS 9 d3 CLld6 10 CLlxd6+ 9 ... CLld4! when again we have a wide range of
cxd6 11 'iYe2 CLld4 12 ~d2 "iUg4 13 iLdS possibilities:

63
Two Knights Defence

Black need not take the perpetual after

14.. .'iVg4+, but can play for more with
14... 0-0-0! since if 15 'iVxd4 lbxd4 16 lba3?
.!:tffi and White is mated.
10 .. :~Wd4+
There are no sensible alternatives.
10 ... lbf5 11 .Jixd5 lbcd4 12 'iYe4 'iYg5+ (if
12...lbxc2 13 .Jit7+ Wd8 14 'ii'xh4+ lbxh4 15
d4lbfS 16 .tg5+ ~d7 17 dxe5 lbxa 1 18 Wf2
and White is winning) 13 'ii'g2 'ii'h4 14 h3
lbxc2 is unclear according to the Russian
master Lepeshkin, but after 15 lbc3! lbxal
16 lbb5 Itb8 17 lbxc7+ ~ffi 18 b3 it seems
a) 10 d3? lbge2+ 11 ~f1 d6 and Black that White has a winning game.
WillS. 11 "ilHe34:Jxh1
b) 10 iLt7+? 'i;e7 11 hxg3 'iUxg3+ 12 'i;f1
d6! (12 ...'iYf4+ 13 ~g2 ~g5+ is only a draw)
13 iLh5 (if 13 lbc3 iLg4! 14 "iWe 1 iLh3+)
13. .. iLe6! 14 d3 .!:tffi+ and Black wins.
c) 10 ~g2? lbxhl 11 Wxhl 'iVe4+ 12 ~gl
lbf3+ 13 ~f2 'iYh4+ 14 ~e2 (14 ~e3 d5)
14... lbd4+ 15 ~f1 d6 and Black wins.
d) 10 c3?! d5!? (if this doesn't work Black
also has 10 ...lbe4 11 'iVh5+ 'iYxh5 12 iLt7+
'iYxt7 13 lbxt7lbf3+ 14 ~g2lbh4+ 15 'iith3
lb g6 16 lbxe5 lbxe5 17 d4 d5+ 18 'iitg2 lbd3
with good compensation for the exchange)
11 cxd4 iLg4 12 'iVa4+ c6 13 iLxd5lbe2+ 14
~f1 'iVf6+ 15 ~g2lbf4+ 16 ~g3 lbxd5 and 12 i.b5?
Black is obviously OK here, e.g. 17 'iVb3 White lets the win slip away with this
'iVf4+ 18 ~g2 'iWe4+ 19 ~g3 'iYxhl 20 ~xg4 move.
'iVg2+ 21 'iWg3 ~e4+ 22 ~h3 lbf4+ 23 ~h4 Correct was 12 ~xd4! lbxd4 13 .tb3 and
'iVe2! and wins. Any problems should occur then:
earlier. a) 13. ...Jih3 14 lba3! (De Zeeuw)
e) 10 hxg3! (best) 10 ...'iVxg3+ 11 ~f1 14...lbxb3 15 axb3 'iitd7 16 d3 l::txh8 17
'iVf4+ 12 ~g2 'iVg5+ with a draw by perpet- 'it>xhl l:.ffi 18 .te3 as 19 ~gl b6 20 .l:.el c5
ual check. 21 iLf2! l':te8 22 .Jig3 and Black has no real
10 'iVf3! compensation for the piece.
The downside to 9... d5 as I said. Instead: b) 13. .. .te6 if 14 'i£;>xhl 'iitd7 (if 14... 0-0-0
a) 10 'iYel! 'iVd4+ 11 'iVe3 transposes to 15 d3 l:.xh8 16 iLe3 and White wins) 15 d3~
the game. (not 15lbg6? hxg6 16 d3 as 17 c3lbxb3 18
b) 10 hxg3 'iVxg3+ 11 ~f1 iLh3+ 12lhh3 axb3 iLf5 19 d4 exd4 20 cxd4 l:.e8 21 lbc3
'iYxh3+ 13 ~gl 'iVg3+ 'with a draw in Muta- ~e1+ 22 ~g2 'it>c6 and Black has a lot of
fov-Sapundzhiev, corr. 1967. counterplay as there is no easy way for White
c) 10 .Jixd5 .th3 11 'it'D 'iWd4+ 12 'iVe3 to get his queenside pieces into play)
'iVxd5 13 ~xg3 'iVd4+ 14 'iYe3 and now 15 ....:txh8 16 iLe3 lbxb3 17 axb3 .:tffi 18

64
 Traxler Gambit: 4 0,g5 ii..c5

~g2 d4 19 1i.f2 1i.d5+ 20 ~g1 a6 21 ttJd2 ply dances away.

and White has an endgame he should win in 20 'ifxg7+ 'it>d6 21 t2Jc4+ 'it>c5 22 .iLxc6
a very high percentage of cases. ':g8 23 'ij'xg8 'ifh1 + 24 'it>f2 'ifxh2+ 25
12 ... ~g4+ 13 'it>xh1 d4 'it>f1 .iLxg8 26 0,xe5 bxc6 27 0,f3 ~h3+
13 ...1i.f5 is also possible and then: 28 'it>f2 .iLd5 29 0,e1 'ifh2+ 30 'it>f1 'it>d6
a) 14 d3 ~fS (if 14... 0-0-0? 15 1i.xc6 bxc6 0-1
16 ttJc3l::tfS 17 ~g3 ~h5 181i.d2 and Black
does not have enough for the material in- Game 27
vestment) 15 ttJd2 'iVd1+ 16 ~g2 'iWxc2 17 I.Belov-Pankratov
'iVf3 ttJe7 18 ~5 with an unclear game, Correspondence 1995
according to De Zeeuw.
b) 14 ttJc3 d4 15 ~xe5+ WfS 16 ~xc7 1 e4 e5 2 0,f3 0,c6 3 .iLc4 0,f6 4 0,g5
~g8 17 ~g3! 'iVxg3 18 hxg3 dxc3 19 1i.xc6 .iLc5 5 0,xf7 .iLxf2+ 6 'it>f1
bxc6 20 d3 l:te8 21 bxc3 ~xh8 22 l::tb1 and
the position looks like a dead draw.
14 ~e2
White cannot prove an advantage any-
more. If 14 ~xe5+ WfS! 15 ~xc7 (15 SLxc6
bxc6 16 'iWc5+? is a bad idea: 16 ... ~g8 17
~xc6 1i.b7! 18 'iixb7 lIe8 and Black wins)
15 ...1i.e6 16 1i.xc6 bxc6 17 'iixc6 ~d1+ 18
~g2 'iie2+ 19 ~g1 'iie1+ 20 ~g2 'iWe2+
with a draw.
14... 'iff415'it>g1 .iLe616'ifh5+ 'it>d717
t2Jg6 hxg6 18 'ifxg6 'iff3

If you rely on statistics YOU should not

play this line. In my database (2,500,000
games) Black has a score of 77% from this
position! The idea of 6 'litfl is of course that
Black cannot now attack with repeated
checks. But on the other hand he retains the
bishop in the enemy camp as a true avant-
garde soldier.
6 .. :ilfe7 7 t2Jxh8 d5!
Naturally Black is not preparing for the
endgame.
8 exd5
190,a3?? Here White has also tried:
White is not completely up to date on the a) 8 1i.xd5? 1i. g4 9 ~xf21i.xd1 10 1i.xc6+
situation. After 19 'iVg2 ~d1+ 20 'iWfl 'iVg4+ bxc6 11 l:txd1 'iWc5+ and Black wins.
21 'iWg2 the game would have ended in a fair b) 8 d3?! dxc4 (8 ...1i.g4 9 'iWd2 1i.h4 may
draw. be even stronger) 9 ~xf21i.g4 10 'iVd2 0-0-0
19 ... .iLd5! 11 ttJc3 cxd3 12 cxd3 :'xh8 Black is at least
Now Black is winning. The white king is slighdy better here.
caught in the open, while the black king sirn- c) 8 1i.e2?! just has to be bad - I only

65
Two Knights Defence

found one game where 'W'hite wins after this. ttJxc2+ 16 Wdl ttJxal 17 Ji.b2 ttJxb3 IS axb3
Also the games played with this move are .l:i.xhS 19 i.xe5 1-0 Schiiler-Leisebein, corr.
mostly irrelevant to the actual evaluation of 1995) 13 i.a3 ~g5 14 'iVf7 ttJd6 15 ..Itxd6
the position, so here I will follow my own .l:!.xd6 16 'it>gl ttJe2+ 17 ~f2 ttJd4 IS ~gl
track: S... J..b6 9 d4 (if 9 d3 ~c5 10 'ikel Ji.g4 ttJe2+ with a draw - Pilk6vi. Again this posi-
11 ttJc3 0-0-0 or 9 Ji.h5+ 'it>fS 10 ~e 1 ttJxe4 tion can be analysed for ages without a more
11 "ii'f3+ 'it>gS 12 d3 ttJf6 13 ttJg 6 "ii'c5 14 defInite conclusion being reached. It's your
ttJc3 ttJd4 15 'iVdl hxg6 16 .i.xg6 .i.g4 with choice if this is worth your time.
advantage for Black) 9... ttJxd4 10 .i.h5+ 'it>fS
11 .i.g 5 ~c5 12 ttJc3 ttJxh5 13 'iWxh5 ..te6
14 "ii'xh7 ttJf3 15 ..th4 dxe4 16 gxf3 .i.h3+
17 'it>e1 ~e3+ 1S ttJe2 exf3 19 ..Ite7+ 'it>eS 20
'iVg 6+ 'it>d7 21 .l:i.d1+ .td4 and Black wins.

9 d6
White has also tried:
a) 9 c3 Ji.g4 10 'iVa4+ ttJd7 11 'it>xf2 (if 11
cxd4 ~f6 12 dxe5 ~f4) 11...~4+ 12 g3
~f6+ 13 ~el (or 13 Wgl ttJe2+ and wins)
8 .. .tt:Jd4 13. .. ~f5 14 cxd4 ~e4+ 15 'it>f2 'iVf3+ 16
S... .i.g4 seems less dangerous, and after 9 ~e 1 'ikxh 1+ 17 .i.n 0-0-0 IS d3 :fS 19 .i.e3
..te2 ..Itxe2+ then: ~xf1+ 20 'it>d2 'iVe2+ 21 ~c3 ttJb6 22 'iVa5
a) 10 'it>xe2 ttJd4+ 11 Wxf2 ttJe4+ 12 'it>e3 ~xe3 23 ~b3 .td1+ 0-1 Maasen-Stadler,
'iVg5+ 13 Wxe4 'iVxg2+ 14 'it>d3 'iVh3+ 15 corr.1954.
'it>e4 (not 15 Wc4? b5+ 16 Wb4 a5+ 17 'it>c5 b) 9 Ji.e2 ..th4 and now:
'ikh4 and Black wins) 15 .. :~g2+ with perpet- bl) 10 g3 J..h3+ 11 ~e1 ttJe4 12 ..tb5+'
ual check according to Gligoric. Let's look a (an improvement on Wead-E.Larsson, corr.
bit further: 16 'it>xe5!? ttJf3+ 17 We4 0-0-0 IS 1967, which concluded 12 d3 ttJxg3 13 .i.e3?
'it>d3 ttJe5+ 19 'it>c3 'iVxd5 20 ~n 'iVc6+ 21 ttJe4+ 14 ..tf2 J..xf2 mate) 12...'it>fS 13 d3
~b3 'iVb6+ 22 'it>a4 ttJd3 23 'iVf5+ 'it>bS 24 ttJxg3 14 hxg3 i.xg3+ 15 'it>d2 .i.f4+ 16 'it>c3
~xd3 (not 24 'iVb5? ttJc5+ and Black wins) ~c5+ 17 Ji.c4 ttJb5+ IS 'it>b3 ttJd4+ 19 'it>c3
24 ... .l:i.xd3 25 cxd3 'iVc6+ 26 Wb3 ~xh1 27 with perpetual check - Palk6vi. Black can
~c2 ~xh2 with a complicated position. This also try 11...0-0-0!?, e.g. 12 d3 .l:!.xhS 13 gxh4
analysis is of course far from conclusive. Let ttJxd5 14l:!.gl! 'iVc5 15 ..tg4+ 'it>bS 16 .i.xh3
us just say that the position is very unclear. ttJxc2+ 17 'it>e2 ttJxa 1 IS .l:i.xg7 l:tfS 19 .I:t.g2
b) 10 'ikxe2! sets Black the greatest chal- ttJc2 and the game is very unclear.
lenge. 1O ... ttJd4 11 'ikxf2 0-0-0 12 b3 (12 d6!? b2) 10 c3 ttJxe2 11 'iYxe2 .tg4 12 'iVb5+
might be an improvement) 12... ttJe4 (worse ttJd7 13 g3 ~f6+ (also interesting is 13. .. 0-0-0
is 12... ttJxd5? 13 Ji.a3 c5 14 ~ell1fS 15 'iVg3 14 'it>g2 lIxhS 15 h3 .i.f5 16 d3 .i.g5 17 l:tn

66
 Traxler Gambit: 4 0,g5 i.c5

a6 18 'iWc4 ..txc1 19 .l:!.xc1 ttJb6 20 'iWb3 'iid6 17 ~c4 .i.e6 18 'iia6!? (risky but after 18
with excellent compensation for the ex- 'iie2 Black has a draw) 18... J::!.d8 and Black
change) 14 ~gl ..th3 15 'iie2 0-0-0 16 d4 has compensation.
lIf8 17 ttJd2 e4 gave Black a winning attack a2) l1...ttJd7 12 ~xf2 'iVf6+ 13 ~e1 and
ill Bar-Holzhauser, corr. 2001. now not 13. .. 0-0-0 14 J::!.f1 'iWh4+ 15 J::!.f2
c) 9 h3 ..th4! (this seems better than ttJb6 16 g3 'iVh3 17 ttJf7! ttJxa4 18 cxd4
9... ..tg3 as the g3-square is designed for a J::!.xd4 19 d3 'iWh5 20 ..te3 .l:i.d7 21 ttJg5 J::!.d8
black knight) 10 c3 Black now has: 22 ttJc3 ttJxc3 23 bxc3 and after the storm
White is much better, but immediately
13 ... ~h4+! 14 g3 ttJf3+ 15 ~f2 'iif6 16 d4
(or 16 ~e3 ~g5+) 16 ... ttJxd4+ 17 ..tf4 0-0-0
with a dangerous attack.
b) 10 d3 ..tg4 11 ttJf7 ~6 121iVd2 ..te2+
13 ~xf2 ttJg4+ 14 ~e1 it'f6 15 'iWxe2 ttJxe2
(Schiller-Uhlig, email 1996) and now 16
~xe2! when it seems that the white king may
be able to evade the checks, e.g. 16 ...'iif2+ 17
~d 1 'iWxg2 18 z:te 1 ttJf2+ 19 ~d2 ttJe4+ 20
~e3 ~f2+ 21 ~xe4 ~xe1+ 22 ..te3 'iWh1+
23 ~xe5 1iVxh2+ 24 ~d4 with four pieces for
the queen, and if now 24.. :iUxc2 25 ttJc3 c5+
c1) 10... ttJf5 11 d4..td7 12 g4 (12 d6 may 26 ~xc5 J::!.c8+ 27 ~d4 J::!.xc4+ 28 ~xc4
be better here) 12... ttJd6 13 ..te2 ttJxd5 with ~xf7 29 J::!.f1 + ~e6 30 J::i.f2 unexpectedly
an unclear game in Estrin-Jezek, corr. 1964. traps the queen.
13. .. 0-0-0!? is also possible. 10 .. :~·c5 11 d3
c2) 1O ... ttJe4!? 11 cxd4 exd4 12 ..tb5+ (12 11 ttJxe5? is refuted by 11...'iWxe5 12 c3
d6? does not work: 12... cxd6! 13 ttJf7 ..td7 Ji. g4 13 cxd4 "iVf5 14 Ji.e2 ~xd4+ 15 .i.f3
14 ttJxd6+ ttJxd6 15 Ji.d3 Ji.f5 16 Ji.xf5 0-0-0 16 ttJa3 iLb6! (pa!kovi's move) 17 ttJc4
ttJxf5 17 ~gl d3! and Black is much better, l:td3 18 ttJxb6+ axb6 19 ~f2 ttJe4+ 20 ~e 1
while if 14 d3? ttJ g3+ 15 ~gl ttJe2+ and (if 20 ~f1? iLxf3 21 1iVxf3 lIxf3+ 22 gxf3
wins) 12...Ji.d7 13 Ji.xd7+ ~xd7 14 'iVg4+ 'iixf3+ 23 ~e 1 ttJf2 24 d4 ttJd3+ 25 ~d2
~d6 15 'iWe6+ (not 15 b3? J::!.f8+ 16 ~gl ttJf4 and White is mated) 20 ...J::i.xf3 21 gxf3
.i.f2+ 17 ~h2 ~e5+ 18 g3 .i.xg3+ 19 ~gl ttJc5 with a clear advantage to Black.
ttJc5 and wins) 15 .. :iVxe6 16 dxe6 ttJg3+ 17
~gl ttJe2+ 18 ~f1 ttJg 3+ with a draw in
Gorkov-Sapundzhiev, corr. 1966.
9 .. :~xd6
Black should be dynamic. After 9... cxd6?!
10 ..te2 ttJxe2 11 ~~e2 ..tb6 12 d4 ..tg4 13
~5+ ttJd7 14 'iVc4 White was much better
in Da Fonseka-Celio, Brazil 1999.
10 CiJf7
White has also tried:
a) 10 c3 .i.g4 11 ~a4+ and then:
al) l1...b5!? 12 .i.xb5+ ttJxb5 13 'iixb5+
c6 14 ~c4 Ji.e6 15 ~e2 Ji.b6 16 ttJa3 .i.g4

67
Two Knights Defence

11 ... e4? This move is for those who want to play;

This does not achieve terribly much. The for those who want to draw there was 15
idea is to swing the queen to the kingside, but "fed1 ..tg4 16 'ifa4+ with repetition. How-
this is not as big a threat as the players ever, it is not really clear that White can keep
seemed to think. Better was 11...Ji.g4 12 b4 the balance after 15 ttJd6+ so maybe he
"ike7 13 "ikd2 Ji.e2+ 14 ~xf2 ttJg4+ 15 ~e1 should have reconsidered. The attempt to
(not 15 liitgl? 'it'h4 16 g3 "iWh3 0-1 Kulev- play with three pieces for the queen by 15
Klimov, corr. 1989) 15 ..."iWf6 16 'it'xe2 ttJxe2 ~xd7+ fails to 15 ... ~xd7 16 cxd4 (or 16
17 ~xe2 'ilif2+ 18liitdl "iWxg219 ~el ~f3+ iLxd4 ~f5+ 17 ~gl c5) 16 .. :iff5+ 17 liitgl
20 ~d2 ~f2+ 21 .:te2 "ii'f4+ with a draw ttJg4 18 ttJe5+ ttJxe5 19 dxe5 :U8 and wins.
according to P:ilk6vi. l1...iLh4!? is also pos- 15 ... cxd6 16 .ltf7+ ~xf7 17 'ilUxd4
sible.
12 c3?
The correct move was 12 b4! 'iVf5 (if
12...'iVb6 13 dxe4 ..tg414 "iWd2 ttJxe4 15 "ikf4
..tf5 16 c3 ttJc2 17 'ijxf5 'ijc6 18 ..tb5 1-0
Soldini-Moro Lorente, email 2000) 13 dxe4
ttJxe4 14 g4 "iWf6 15 ~g2 and White is much
better according to P:ilk6vi.
12....lth4!
The only move. After 12... ttJe6? 13 'ilie2
e3 14 d4 'iffS 15 ttJe5 ttJg5 16 ~xe3 ..txe3+
17 ~e1 iLf4 18 l:.f1 ttJge4 19 g3 White was
on his way to winning in the earlier game
Blank-Pankratov, corr. 1993. 17 ... 'iif5+
13 .lte3 .ltg4 14 'iia4+ Also playable was 17 .. :iVxd4!? 18 iLxd4
The only move. If 14 "ii'd2? 'iVf5+ 15 iLf4 exd3 19 ttJd2 ..tg5 20 ttJf3 ..tf4 21 .l:.d1 ttJe4
..te2+ 16 ~gl ttJg4! and Black wins, or 14 22 g3 iLh6 23 Iiitg2 d2 and Black is far from
'ifc1 "fef5+ 15 iLf4 ttJe2 16 dxe4 "iWc5 17 being worse.
'ilie3 'ilixc4 18 ttJd2 'ifxf7 and Black is much 18 ~g1 'iig6 19 .ltf2 .ltg5
better. Black trusts his long term compensation.
14 ....ltd7 which is a good plan. After 19 ... ..th3?! 20
iLg3 iLxg3 21 hxg3 'ifxg3 22 'iff2 "ikxf2+ 23
~xf2 iLf5 24 dxe4 ttJxe4+ 25 1iitf3 .l:te8 Black
still has some compensation for the ex-
change, but White is basically a bit better.
20 ttJa3
If 20 dxe4 iLc6 21 ttJd2 iLxd2 22 ~xd2
ttJxe4 and Black has the initiative.
20 ... .ltc6 21 ttJc4?
An understandable mistake. White wants
to defer the pressure against g2 and does not
care much for the pawn on d3. But actually
this pawn becomes powerful, so White
should have played 21 dxe4 ttJxe4 22 'iVc4+
15 ttJd6+ d5 23 'iVd3 and he can probably hold.

68
 Traxler Gambit: 4 Cjjg5 iLc5

21 ... exd3 22 .ig3 ~g8 23 .l:!.d1 b5 .idS (Games 2S-32) and 6 .ib3 (Games 33-
36).
6.id5
The most popular move. White has also
tried:
a) 6 d4?! ttJxd4 7 c3 ttJc6 S i.b3 .ufS 9
i.e3 i.xe3 10 fxe3 d6 and Black is slightly
better.
b) 6 b4? ttJxb4! 7 d4 J..xd4 S c3 J..cS 9
J..b3 l:IfS! and Black is better, e.g. 10 i.a3
(not 10 cxb4? J..d4) 1O ... ttJa6 11 ttJxh7 ttJxh7
12 'i¥hs d6 13 'it'xh7 J..xf2+ 14 ~dl 'it>d7 15
~xg7+ 'it'e7 16 'iVxe7+ 'i;xe7 and White has
problems with the king and completing his
24 Cjjxd6 development.
White has no way out. 24 ttJaS .i:teS! or 24 6 .. .J::U8
ttJe3 ~eS 25 l:i.xd3 lte4 26 't!Vxa7 l:Ia4 27 Black can also play 6... d6 (see Games 31 &
'ikbS+ ttJeS 2S ~cS 'it'xd3 29 'iVe6+ 'it>hS 32) or 6... ~eS which will usually transpose.
wins for Black. After 6... ttJb4? White has 7 d4! exd4 S 0-0
24 ... Cjjg4 25 .l:!.xd3 iLd8 26 c4 .ib6 27 ttJbxdS 9 exdS :eS 10 'it'd3 h6 11 'iVg6! hxgS
c5 iLxc5 28 ~xc5 ~xd3 29 ~c3 ~d1 + 12 ~xg7+ 'it>d6 13 3t.xgS I:.fS 14 c4 and
30 ~e1 ~d4+ 31 ~f1 J:l:f8+ 32 ~e2 White won in Estrin-Vajs, carr. 1971.
iLxg20-1

Came 28
Losev-Isaev
Kherson 1990

1 e4 e5 2 Cjjf3 Cjjc6 3 iLc4 Cjjf6 4 Cjjg5

iLc5 5 .ixf7 +
The most serious try for an advantage.
5 ... ~e7

7 iLxc6!?
Black does not have serious problems af-
ter this, at least not theoretically. In the game
things are less clear. Nevertheless, if White
wants to capture on c6 he should wait a
move and play 7 0-0 d6 S J..xc6 when Black
cannot accelerate his development by recap-
turing with the d-pawn. 7 0-0 is considered in
Games 29 & 30.
Less dangerous is 7 ttJf3 d6 (worse is
Here White has two promising moves, 6 7... ttJd4?! S ttJxd4 J..xd4 9 0-0 c6 10 c3 i.b6

69
Two Knights Defence

11 .ltb3liJxe4 12 'iVh5 and White is better) S e.g. 9 0-0 (or 9 d3 "it'g6 10 liJf3 liJh5 with
c3 .ltg4 9 .ltxc6 bxc6 10 d4 exd4 11 cxd4 compensation) 9... ~g6 10 d3 .ltg4 11 "it'el
.ltxf3 12 gxf3 ..Itb6 13 .te3 'iVd7 14liJd2 (or h6 12 b4 .td6 13 h3 .th5 14 c4 <Jtd7 and
14 liJc3 ~h3 with compensation) 14...1hes Black is better.
15 'iVa4 c5! 16 ~xd7+ liJxd7 17 dxc5 liJxc5 9 f3 ltJf2?
with a roughly equal position, which was This is a suicide mission. It was better to
agreed drawn in Reithel-Walther, corr. 1979. retreat again with 9... tLlf6 10 d3 h6 11 liJh3
7 ... dxc6! J..xh3 (not l1...g5 12 tLlf2 i.e6 13 ..Ite3
.txe3 14 'iVxe3 'iVd6 15 tLld2 and Black has
no compensation) 12 gxh3 "it'd7 13 "iVg2
~dS, though White is better nonetheless.
10 l:!f1 h6 11 d3! ~d4

Black plays for quick development. If

7... bxc6 S d3 d6 9 liJf3 .tg4 10 .te3 J..xe3
11 fxe3 'iibs 12 b3 'iib6 13 ~d2 and White
is much better.
8 'YWe2 12.l:!.xf2!
If S d3 liJg4! (not S...'iVeS?! 90-0 J.. g4 10 A very sound decision. Black's attack is
liJf3 i.d6 11 h3 i.e6 12 tLlg5 and White is now history.
much better) 9 tLlh3 tLlxf2! (9 ... h6!? is an- 12.. .'YWxf2+ 13 'YWxf2 i.xf2+ 14 'It>xf2
other idea and it seems to me that Black has hxg5 15 i.xg5+ 'itd6
enough compensation because of the weak White has two pawns for the exchange
knight on h3, e.g. 10 0-0 .te6 11 tLld2 "iVd6 and a far superior pawn structure, and to-
12 liJb3 J..b6 13 ..Itd2 g5 and Black looks gether these give him a huge advantage in the
[me) 10 i.g5+ and now, rather than 10...<JteS ending.
11 ~5+ g6 12 ~xh7 ~d4 13 ~xg6+ l:tf7 16 ltJd2 ~e6 17 ~h4!
14 l:tn! (De Zeeuw) when White is much Targeting the e5-pawn which cannot be
better, Black can improve with 10 ... <Jtd6! 11 defended.
.ltxdS (if 11 ~h5 J..g4 12 ~xh7 "iVd7) 17 ... c5
l1...liJxdl 12 J..xc7+ Wxc7 13 <Jtxdl l:tf6 Black loses the e-pawn in all variations:
followed by ... .l:!.g6 and Black regains the e.g. 17 ... -UaeS IS J..g3 c5 19 c3 g5 20 d4 cxd4
pawn with a big advantage; the knights are 21 cxd4 c6 22 i.xe5+ and White wins.
no match for the bishops. This is from a 18 i.g3 l:l:ad8 19 ~e3 l:'Ue8 20 c3 i..f7
letter to New in Chess) 'earbook by Tobi Usher, 21 d4 cxd4+ 22 cxd4 c5 23 d5
who is of course completely right. Better than 23 i.xe5+ <Jte6 24 tLlb3 cxd4+
8 ... ltJg4?! 25 i.xd4 when White has more technical
This time S... "it'eS!? was an improvement, problems.

70
 Traxler Gambit: 4 4:Jg5 i.c5

dxc5 17 bxc5 .txc5 18 llYa5 J.d4 19 tLld2

l::td3 20 .l:!.b1 'iVg6 also gives strong initiative,
but 14 'iUdl 'ti'g6 15 ~hl was a possible
improvement) 14...'Dh5 15 'iith2? (15 'iUdl
was still better) 15 ... l:.xf3! 16 gxf3 tLlf4 17
tLle2 'itb5 18 tLlxf4 exf4 19 l::th 1 J.xh3 20 d4
.l:!.f8! 0-1 Bruinenberg-Nische, corr. 1964;
there is no defence against ...l:.f6-g6 and
mates.
a2) 9 tLlf3 'Dd4! 'With a further branch:

23 ... i.xd5
There is no other defence against 'Dd2-c4.
Black has no real chance of saving the end-
game now, though there is always hope.
24 exd5 'it'xd5 25 4:Je4 'it'e6 26 .l:l.e1 b6
27 b4 l:!.d5 28 ~e2 lld4 29 bxe5 b5 30
i.f2 .l:!.d5 31 i.e1 ~a8 32 i.a5 .l:l.d4 33
.l:!.d1 .l:!.xd1 34 ~xd1 ~d5 35 ~e2 .l:l.h8 36
h3 l:ih6 37 i.d2 l:tg6 38 g4 ~e4 39 h4
J:!.a6 40 ~b2 ':'e6 41 h5 ~d5 42 ~b3
~e8 43 ~b4 ~e6 44 i.e3 lIf8 45 4:Jg5 a21) 10 c3? tLlxf3+ 11 gxf3 ..th3 12 l::tel
.l:!.f4+ 46 'it'b3 'it'd5 47 e6 a5 48 c7 a4+ 'iUe8 13 'iVd2 tLlxd5 14 exd5 l::txf3 15 'iUg5+
49 ~a3 .l:!.e4 50 i.a5 Wd6 51 4:Je4+ ~d7 ~d7 16 d4 'iVf7 17 J.e3 h6 0-1 Matajev-
52 g5 .l:!.e2 53 h6 gxh6 54 gxh6 .!':te6 55 Dobrotin, Moscow 1996.
h7 J:th6 56 4:Jd6! 1-0 a22) 10 J.xb7? 'u'b8 11 J.a6 'iUe8 12
A neat fmish. tLlbd2 'iUg6! 13 'lithl tLld7 with a huge advan-
tage for Black according to De Zeeuw; e.g.
Game 29 14 c3? tLlxf3 15 tLlxf3 'iUh5 16 d4 l::txf3 17
Shabalov-I.lvanov dxc5 tth3! and wins, or 14 tLlh4 .Jl.xd1 15
US Championship, Parsippaf!Y 1996 'Dxg6+ hxg6 16 l::txd 1 tLlxc2 and Black is
clearly better.
1 e4 e5 2 4:Jf3 4:Je6 3 i.e4 4:Jf6 4 4:Jg5 a23) 10 tLlbd2! (the only move) 10... tLlxf3+
i.e5 5 i.xf7+ ~e7 6 i.d5 .l:l.f8 7 0-0 d6 (10 ...'iUe8 11 h3! ~h5 12 .tc4 'Dxf3+ 13
8 e3 'Dxf3 ..txf3 14 llYxf3 'iUxf3 15 gxf3 tLlh5 is
White prepares d2-d4 to curtail the black equal) 11 tLlxf3 tLlxd5 12 exd5 ..txf3 13 gxf3
bishop on c5, while also giving the queen an ~d7 14 J.e3 l::tf5 15 J.xc5 dxc5 16 ~e2
escape route on the queenside. Others: 'iUf6 17 ttael ~d6 and Black is slightly better
a) 8 d3?! is strongly met by 8... J.g4! and according to De Zeeuw.
White is already in trouble: The plan of ... .i.g4 followed by ... tLld4
al) 9 'iUd2 h6 10 h3 .td7 11 tLlf3 'iUe8 12 gives Black a strong attack.
.i.xc6 bxc6 13 b4 .i. b6 14 tLlc3 ? (after this b) 8 h3, preventing ... ..tg4, is considered in
White cannot guard the kingside anymore; the next game.
instead 14 c4?! tLlxe4! 15 dxe4 l::txf3 16 c5 c) 8 J.xc6 bxc6 9 tLlf3 is a more promis-

71
Two Knights Defence

ing continuation, eliminating the dangerous 12... ~b6 13 cxd4 h6.14 es! or 12... h6 13
knight. After 9... ~g4 (if 9...'i'e8 10 d4!) 10 c3 cxd4 J:tb4 14 'irVc3 J:txd4 15 ~e3) 13 es! cxb2
'i'e8 (if 10 ... ttJxe4 11 d4 ~b6 12 J:tel or 14 exf6+ ':xf6 15 l::te1+ .lie6 16 J:txe6 ~d7
10... ds 11 d4 exd4 12 es liJe4 13 cxd4 and 17 'iVg4! and wins. .
~'hite is better) 11 d4 exd4 (11...~b6 12 e) 9...~b6! 10 Ji.xc6 (if 10 h3 h6 11 d4
dxes dxes 13 liJbd2) 12 cxd4 Si.xf3 13 gxf3 hxgs 12 hxg4 'irVe8!? or 11 liJf7 'iUd7 12 hxg4
~b6 14 ~h 1 followed by ~e3 and liJd2 'iVxg4) 10 ... bxc6 11 h3 h6 12 hxg4 hxgs 13
~te has reasonable prospects of consoli- d4 'iUd7 14 'iUdl and White is better accord-
dating. Black might consider 9... ~f7!? 10 d4 ing to De Zeeuw, but 13 .. .'ii'e8!? may be an
(or 10 c3 ~g8 11 d4 exd4 12 cxd4 ~b6) improvement, e.g. 14 'i'dl 'i'g6 15 f3 J:tf7!
1O ... exd4 11 liJxd4 ~g8 12 'iWd3 'iWe8 with followed by ...':'h8 with an attack as the rook
counterplay similar to the Spanish 3 .libs fs on the f-ftle deters the white king from run-
4 d3 variation. ning off via f2.
8 ... .1£.g4 9 ~e1 9 ... h6!
~te has also tried 9 ~3!? (if 9 'iUa4?! The white knight has no healthy retreat.
'iUe8 10 .lixc6 'iYxc6 11 'iUxc6 bxc6 and 10 d4 .1£.b6!
Black is slightly better - De Zeeuw) and
now:

Black keeps the tension since it is ~te

who has a problem to solve. If 10 ... exd4? 11
a) 9... liJas? 10 'iUc2 c6 11 b4 cxds 12 bxcs Si.xc6 bxc6 12 es liJds 13 h3 Ji.f5 14 b4 Ji.b6
dxcs 13 d4! liJc4 14 liJa3 cxd4 15 cxd4 ~c8 15 c4 and ~te is much better.
16 liJxc4 J:txc4 17 'iYb3 and ~te wins. 11 h3 hxg5 12 hxg4 ~d7 13 ctJa3
b) 9... h6?! 10 'iYxb7! llbS 11 'iUxc6 hxgs (if Not 13 Ji.xgs? 'i'xg4 14 Ji.xf6+ l:txf6! and
11....lid7 12 liJe6!) 12 ~a6 .lid7 (or 12....l::i.h8 the black attack is very dangerous; or if 13 a4
13 d4 exd4 14 b4 .lib6 15 Si.xgs) 13 'iUe2 as 14 'i'e2 exd4 15 Si.xg5 d3! 16 'iUdl (or 16
Si.g4 14 'iUel liJxds 15 exds 'it>f7 16 b4 Si.b6 'iUxd3 'i'xg4 17 ~xc6 bxc6 18 Ji.e3 'i'h4)
17 a4 and White is close to winning. 16... .a:h8 and Black takes over the initiative.
c) 9... 'iUc8 10 Si.xc6 bxc6 11 d4 exd4 12 13 .. :ii'xg4 14 f3 ~h5 15 ctJc4 g4 16
'iUc4 dxc3 13 liJxc3 .lid7 14 liJf3 and ~te ctJxb6 axb6 17 iVg3 exd4 18 .1£.xc6 bxc6
is better - De Zeeuw. 19 cxd4 gxf3 20 gxf3 g5 21 e5
d) 9...J:tb8 (trying to 'WID a tempo after The position is very unclear; e.g. 21...liJh7
.lixc6 bxc6) 10 d4 (10 d3 'iUd7 is unclear) (the only move) 22 'it>g2 J:tae8 23 .lid2 'ikg6
10... exd4 11 Si.xc6 bxc6 12 'iVc4 and White 24 J:tael ~d8 25 :h1 with a mess. Unfortu-
has a strong threat in e4-es, e.g. 12... dxc3 (if nately the players decided to agree a draw.

72
 Traxler Gambit: 4 tUg5 i.c5

~xg5 15 <Ji>h2 'ifh5! 16 gxf4 ~xh3+ 17 <Ji>gl

.i.g4 18 f3 exd4! and White has no defence)
11 ... exf4 12 .txf4 h6 13 ctJf3 ctJxd5 14 exdS
l::txf4 15 dxc6 bxc6 and Black was slighdy
better in Wegelin-Pohl, corr. 1993.
9 .. :YWg6!
Black should not take action before he is
ready and after this move he is.
Nevertheless, also interesting is 9... h6!? 10
ctJf3 ~h5 11 .txc6 bxc6 and now:

Came 30
Bex-Donev
Biel1995

My annotations to this game are based on

those by Maarten de Zeeuw in Neui in Chess
Yearbook #65.
1 e4 e5 2 tUf3 tUe6 3 i.e4 tUf6 4 tUg5
iLe5 5 iLxf7+ ~e7 6 iLd5 .l:!.fS 7 0-0 d6 a) 12 ctJc3 g5! gives Black a strong attack;
S h3?! whereas after 12... .tg4 13 ctJa4 .i.xf3 14
'ti'xf3 'ti'xf3 15 gxf3 i.b6 16 f4! Wrute parries
threat of ... g7-g5 with a better game, e.g.
16... exf4 17 i.xf4 ctJxe4 18 ctJxb6 axb6 19
i.xh6 gxh6 20 dxe4 and White is slighdy
better.
b) 12 i.e3 i.xh3!? 13 gxh3 ctJxe4 14 dxe4
(14 ctJxe5 'ti'xeS 15 d4 'YWf5 16 dxc5 'ti'xh3 17
cxd6+ 'it'd7 18 dxc7+ cJJxc7 19 ctJd2 ctJxd2
20 .i.xd2 l::i.f3 21 i.aS+ cJJb7 22 'ti'd4 l::i.g3+
23 fxg3 WVxg3+ also leads to a draw)
14... l::i.xf3 15 ctJd2 ':'afS 16 ctJxf3 ':'xf3 17
.txcS dxc5 (17...~xh3 18 J:.el is less clear)
18 l::i.e 1 ~g6+! 19 <Ji>n .uxh3 intending
This looks like a beginner's mistake. White 20 ...Mh1+ 21 <;te2 ~xe4+ with perpetual
prevents ... .i.g4 but creates a weakness in his check.
own camp and loses valuable time while do- 10 ~h1
mgso. 10 .txc6 was probably an improvement.
S .. :~!VeS 9 d3 The bishop is not doing anything for the
After 9 c3 'iVg6 10 d4 .i.b6 11 f4 (if 11 defence, while after a quick ...ctJc6-d4 the
.i.e3 ctJh5! 12 'iVd2 ctJf4 and Black has a knight can suddenly turn into one hell of an
strong initiative, e.g. 13 .i.xf4 ':'xf4 14 g3 attacker.

73
Two Knights Defence

10 ... hS 11 lLlf3 Black is much better - White is not a piece

up, he is essentially a rook down!
14... ii.xh3 15 gxh3 litxf3 0-1
White is mated in 8 moves.

Game 31
Kruger-Moormann
Email 1998

1 e4 e5 2 lLlf3 lLlcs 3 ii.c4 lLlf6 4 lLlg5

ii.c5 5 ii.xf7+ rJ;;e7 6 ii.d5 d6

11 ... lLlg4?
I1.Jbxds! was simple and also the best.
12 exds ibb4 13 ibc3 ~hs 14 ibel (not 14
a3? i..g4! 15 axb4 i..xf3! 16 ~xf3 ':!'xf3 17
bxcs 2:.f6 and Black will win) 14.. .'~xdl 15
ibxdl ibxds 16 c4 ibf4 17 ibc3 i..e6 and
Black is slighdy better.
12 ii.xcS?
Now there is no time for this. Better was
12 iVe2! (not 12 hxg4 .l1.xg4 13 ibbd2 ~hs+
14 ~gl ibd4 with a decisive attack) 12.. :ikhs Black does not fear the thrust 7 ibf7 at all,
13 ii.xc6 bxc6 14 ibbd2 gs 15 c3 ibf6 16 d4 and it is also only an illusion; after 7 ibf7?!
~b6 17 dxes dxes 18 ibc4 ~a6 19 b3 ~e6 ~fS 8 ibxh8? i..xf2+! 9 ~fl i..g4 and Black
with an unclear position. WillS.
12 ... lLlxf2+! 13 J:!.xf2 ii.xf2 7 c3
The most popular move, but Martin de
Zeeuw is convinced that 7 d3 is stronger; this
is investigated in Game 32.
7 .. :YWeS S d4?!
Better is 8 d3 2:.fS 9 ibf3 (or 9 0-0 i..g4 10
.ixc6 bxc6 11 ibf3 iVg6 12 ibbd2 ibhs with
compensation for the pawn) 9...'Yig6 10 i..gs
..tg4 11 ..txc6 bxc6 12 h4 h6 13 .ixf6+ ':!'xf6
14 b4 i..b6 15 ibbd2 i..xf3 16 gxf3 ~g2 17
.l:tfl ~h2 18 iVa4 with an unclear position;
but not 11 ibbd2? "iVxgs! 12 ibxgs ~xdl 13
.l:txdl (or 13 ~xdl ibg4) 13. ..ib g4! and
White had problems in Gikas-Wedberg,
14 ii.a4?! Lugano 1989.
Not 14 'Yifl ~g3! 15 ibc3 .l:txf3! and S... exd4 9 ii.xcs
Black wins, while after 14 iVe2 .l1.b6 15 ~ds Worse is 9 cxd4?! ibxd4 10 ibc3 c6!
i..xh3 16 ibbd2 .ig4 17 "iVel c6 18 .ic4 .l:tf4 (10 ... ~hs 11 'Yid3 .l:tfS 12 b4 ..tb6 13 iba4

74
 Traxler Gambit: 4 l'fjg5 .ic5

tbc6 14 tbxb6 axb6 was unclear in Karpov- 13 l'fjd5 h6 14 l'fjf3 ii.g4 15 ii.e3 l'fjxd5
Beliavsky, USSR 1983) 11 ..tn (if 11 i.c4 16 exd5 ~d7 17 ii.xb6 axb6 18 ~b3
'iVhs 12 0-0 i.g4! 13 ~d2 h6 and Black is ~f7?
much better) 11...~d8 12 i.b3 (12 i.e3 ~6 A grave positional error. Necessary was
13 0-0 ..tg4! followed by ... h7-h6 and Black is 18 ... i.xf3! 19 'iVxf3+ 'Yin 20 ~c3 ~g8 21
much better) 12....l:tfS 13 0-0 i.g4 14 "iVel %:tac1 lic8 22 "iVd3 hS when White is better,
tbxb3 lS axb3 .l::!.e8 Black is slightly better but Black has good drawing chances.
here. He has active pieces, a better pawn 19l'fjd4!
structure and the two bishops. But then, he
also has the king in the centre for a few more
moves.
9 .. :~xc6 10 cxd4 .ixd4

The problem in this position is that White

now obtains a strong post for the knight on
e6.
19 .. .l::the8 20 h3 ii.h5 21 l'fje6 ~g8 22 f4
1O... i.g4? does not work: 11 f3 i.xd4 12 iH7 23 f5 "Yi'a4 24 "Yi'd3 1-0
fxg4 tbxg4 13 tba3! i.f2+ 14 'ot>fl and the Black resigned as this was a correspon-
attack is gone. dence game and in such games an advantage
11 0-0.ib6? of these proportions is decisive; e.g.
Unfortunately this does very little for the 24 ... i.xe6 2S dxe6! 'iVh4 (if 2S ... z:tf8? 26 f6
black position. Both alternatives were better: wins) 26 f6 ':xe6 27 n+ 'it>fS 28 'iVh7 with a
a) l1..."iVcS 12 tbd2 ..tg4 (if 12..."iVxgS 13 terrible attack.
tbf3 "iVg6 14 "iVxd4 i.g4 lS tbgs and White
is OK) 13 tb gf3 i.eS and Black looks better. Game 32
b) 11 ... ..teS!? 12 "iVb3 J:i.f8 13 tbc3 ..txc3 Anand-Beliavsky
14 bxc3 h6 lS tbf3 tbxe4 16 %:tel 'ot>d8 and unares 1991
White does not have enough compensation
for the pawn. 1 e4 e5 2 l'fjf3 l'fjc6 3 ii.c4 l'fjf6 4 l'fjg5
12l'fjc3 ~f8 .ic5 5 ii.xf7+ ~e7 6 ii.d5 d6
This looks like a waste of time, but it does The game actually began 6...~e8!? 7 d3
make some sense: Black avoids the check on d6.
dS and then develops. After 12... .tg4 13 7 d3 ~e8
~3 'uafS 14 tbdS+ 'it>d7 lS i.e3 i.xe3 7... "iVfS?! is less good; Black does not have
(G.Lee-Rumens, British Championship 1981) compensation after 8 i.e3 i.xe3 9 fxe3 tb g4
then 16 fxe3! and White is clearly better ac- 10 ~d2 "iVf6 11 tbf3 'iVh6 12 'ot>e2 %:tf8 13
cording to PatkCivi. tbc3 and 14 'uafl.

75
Two Knights Defence

But Black can still play 7...:fS and then:

bl) 14...'iite8 15 ttJd2 .tg4 (15 ...g5? 16

a) 8 i.e3 i.xe3 (8 ... i.g4?! 9 '*'d2 i.xe3 10 ~5+ Wd7 17 0-0-0 ttJf2 18 .l:.fl White
fxe3 'iVd7 11 h3 h6 12 ttJf3 .ixf3 13 gxf3 wins) 16 ttJb3 as! (not 16 ... c5?! 17 c4 ~f7 18
ttJxd5 14 exd5 ttJb8 15 'iVg2 and White is ~xg4 ~f2+ 19 'iitdl ~f1+ 20 Wc2 'i¥xd3+
much better, Koetsier-Kaupat, corr. 1998) 9 21 'iitxd3 ttJf2+ 22 We2 ttJxg4 23 h3 and
fxe3 ttJa5! (the idea is to exchange the White is much better in the endgame) 17 h3
queen's knight for the white bishop - the (if 17 ~xg4 ~f2+ 18 Wdl 'iVxe3 19 'ifUh5+
other knight will be useful on the kingside; if 'iite7 20 ~e2 '1Wg1+ 21 'iitd2 ~xh2 22 ~g4
instead 9... ttJg4 10 ttJxh7 and White is much ~g3 and Black is no worse) 17 ... i.e6 18 We2
better) 10 0-0 (prophylaxis does not work l:i.b8 19 J::txhl a4 20 J::tfl 'iVg8 21 ttJc1 .l:hb2
here: 10 a3? c6 11 i.a2 'iVb6 12 ~c1 ttJg4 13 22 'iitdl i.xa2 and the position is very un-
b4 ttJxe3 14 ~d2 i.g4 15 bxa5? ttJxg2+ 16 clear.
'iixg2 ,*,e3+ and mates, or 15 ttJc3 '1Wd4 16 b2) 14... ~f6 15 'ii'xf6+ gxf6 16 Wfl f5 17
ttJf7 ttJxg2+ 17 ~xg2 "iVxc3+ 18 Wf2 i.e6 Wg1 fxe4 18 dxe4 ttJg3 19 hxg3 i.g4 20
and Black wins) 1O ... c6 11 i.b3 ttJxb3 12 ttJd2 i.h5 21 .t:rfl i.g6 White is better here,
axb3 ttJg4 13 l:i.xfS 'iNxfS 14 ~e2 ~f6 15 but Black should make a draw.
ttJf3 ~h6 16 h3 ttJf6 (not 16 ...'1Wxe3+? 17
~xe3 ttJxe3 18 ttJa3! and the knight is
trapped) 17 ttJbd2 g5 with compensation for
the pawn.
b) 8 i.xc6! bxc6 9 i.e3 is the most dan-
gerous. Black must be very vigilant in order
to get enough initiative for the pawn:
9... .txe3 (9 ... .tg4?! 10 llVd2 i.xe3 11 fxe3
~e8 12 ttJf3 and White was much better in
Van Vugt-Eastwood, corr. 1998) 10 fxe3
ttJg4 11 ttJxh7! (a very hard move with which
White takes back the initiative; if instead 11
~c1 'iite8 12 ttJf3 .l:.xf3! 13 gxf3 ~g4+ or 11
ttJf3 ttJxe3 12 ~d2 ttJg4 13 ~g5+ Wf7 14 h3 8..1txc6!
ttJf6 15 ttJc3 h6 16 ~g3 .l:!.b8 and the game is If 8 i.e3?! i.xe3 9 fxe3 ~g6 10 ttJf3
unclear) l1...ttJf2 12 'iVh5 ttJxhl 13 ttJxf8 ttJxd5 11 exd5 ttJb4 and Black is OK, while 8
'1Wxf8 14 ~4+ and then: c3 returns to 8 d3 in the notes to Game 31.

76
 Traxler Gambit: 4 ti'Jg5 iLc5

8 ... bxc6 9 iLe3 "tWg6 If 16... ttJxe3 17 ttJg5 or 16 ...'iVxe3 17

If 9...i.b6 10 0-0 .i.g4 11 ttJf3 ~5 12 'iVxe3 ttJxe3 18 ttJa3 and the knight is
ttJbd2 J:!.a£8 13 'iVe2 h6 14 i..xb6 cxb6 15 d4! trapped.
and %ite is better. 17 ti'Jbd2 .b6 18 0-0-0
10 ti'Jf3 iLxe3 11 fxe3

18 ... <t>e7 19 h4
11 .. :~Wxg2?! 19 l:.dgl!? looked better.
After this Black loses all his counterplay. 19 ... l:Iag8 20 .l:!.g5 ~f8 21 ti'Jh2?!
Up to now g2 was weak, now it is exchanged. Again 21 :dgl h6 22 :5g2100ked better.
Better was 11...l::tb8 12 b3 'iVh6 (worse is 21 ... 'ilVf7 22 b3 h6 23 J:!.f5?!
12... ttJg4 13 'iVe2 'iVh6 14 'it'd2 d5 15 ttJc3 d4 This brings the black bishop back into
16 ttJdl M£8 17 h3 i.a6 18:£1 ttJf6 19 WeI play. Better was 23 l:tgg1.
c5 20 ttJxe5! and %ite is much better) 13 23 ... ~c8 24 .l:!.f3 ~h5 25 J:!.df1 .l:!.g3?!
'iVe2 g5 140-0 g4 15 ttJfd2 'iVg5 16 c4 c5 17 After 25 ... i..h3 26 .l:!.1 f2 i..g4 27 ttJxg4
ttJc3 c6 and it seems that Black has compen- ttJxg4 28 %Ih3 White is still better, but only
sation for the pawn, e.g. 18 'iVf2 :£8 19 ttJe2 so much.
ttJd7 (an important move preventing ttJg3- 26 d4 .l:!.xf3 27 ti'Jhxf3 ti'Jg4 28 "tWd3 l:If8
f5) 20 'iVg3 ttJf6! and Black is not necessarily 29 'ilVc3 ~d7 30 ~b2 exd4 31 exd4 ~d8
worse here. 32 e5 J:!.f4 33 .l:!.e 1 ~f5
12 .i:tg1 "lWh3 13 l:txg7 + 'It>d8
If 13. .. We8 14 :g3 ~6 15 'iVe2 and
%ite is much better.
141:.g3
%ite is a clear pawn up. There now fol-
lows a game in which %ite is winning for
39 moves, but then throws it all away. At this
time Anand was still a very young man who
had a tendency to play too fast. Black tries to
hang on, but even though %ite makes many
mistakes, it is only after the fInal blunder that
the advantage switches to Black.
14... ~h6 15 "lWe2 ti'Jg4?
Black achieves nothing by this. 34 <t>c1
16 h3 ti'Jf6 34 'iVa5! wins without effort.

77
Two Knights Defence

34 ... <.t>eS 35 a4 d5 36 ~b2 il.eS 37 J::i.e2 White's extra f-pawn is fIxed, while after 8
'tifS 3S a5?! lLlc3 h6 9 lLlf3 i..g4 Black has counterplay.
This move is not bad, just unnecessary. Refutation attempts with lLlg5-f7 turn out
3S ... ~b7 39 a6+?! ~xa6 40 e6 ~b7 41 fIne for Black: 7 lLlf7 i..xf2+ 8 ~xf2 (if 8
J:te1 'ifd6 42 b4 tUf2 43 tUb3 tUe4 44 ~f1? d5) 8...lLla5! 9 Mfl (not 9 lLlxh8?
tUe5+ ?? lLlxe4+ and mates) 9... lLlxb3 10 axb3 ~xf7
A horrible mistake, losing a piece. 44 ~e3 lLlxe4 8 0-0 lLlxf2 9 ~5 d5 10 ..ixd5 ii.e6
~f8 45 lLlfd2 would still have maintained or 7 0-0 d6 8 lLlm lLld4! 9 lLlxh8 i..g4 10
White's advantage. "iNel ~c8! 11 'Ot>hl ~f3! 12 Mgl 'ifh3! 13
44 ... 'ifxe5! 0-1 gxf3 lLlxf3 and wins according to Estrin.
7 d3!
Game 33
Elison-K. Werner
Email 1999

1 e4 e5 2 tUf3 tUe6 3 il.e4 tUf6 4 tUg5

il.e5 5 il.xf7 + ~e 7 6 il.b3

The best solution and the most dangerous

plan against the Traxler. White simply ex-
changes dark-squared bishops and keeps the
centre solid.
7 ... d6
Others:
Here the bishop is less exposed than on a) 7...~e8?! 8 ..ie3 d6 9 lLlf3 i..g4 10
d5, but White no longer has the useful ex- lLlbd2lLld4 11 i.xd4 ..ixd4 12 c3 i.b6 13 h3
change ii.xc6. ~h5 14 Mfl .txf3 15 lLlxf3 lLld7 16 'ife2
6 ... .l:!.fS l:If4 17lLld4 ~xe2+ 18 lLlxe2 and White was
Others: just a pawn up in Kariakin-Zubov, Krama-
a) 6... d6 7 d3 lLla5?! (7....l:tf8 returns to the torsk 2002.
game) 8 i..e3 (not 8 lLlm ..ixf2+ 9 ~fl ~f8 b) 7... h6! (the best move) 8lLlf3 d6 9 i..e3
10 ~xf2 lLlxb3 11 axb3 'iVxf7 12 ..ig5 h6 13 (not 9 h3?! ~e8 10 lLlc3 lLld4! 11 lLlxd4
..ixf6+ "iNxf6+ and Black is better) 8... ii.xe3 9 ii.xd4 12 0-0 1Wg6 13 ~h2 .td7 with full
fxe3 lLlxb3 10 axb3 and White simply has a compensation for the pawn; this is basically
pawn more. the kind of position Black hopes for in the
b) 6...'iVf8!? is interesting as it is the only Traxler) 9... .tg4 10 i..xc5 dxc5 11 lLlbd2
move that prevents the immediate d2-d3, i.e. lLld4 12 h3 ii.h5 13 0-0 (not 13 g4lLlxg4 14
7 d3? i..xf2+ or if 7 0-0 d6 8 d3 i..g4 9 lLlf3 lLlxd4?? ~xd4 0-1 Paoli-Steiner, Reggio
lLld4 10 lLlbd2 lLlh5 11 c3 lLlxf3+ 12 lLlxf3 Emilia 1951) 13. .. lLlxf3+! (if 13. .. ~d6 14 g4
~xf3 13 "iNxf3 ~xf3 14 gxf3 Mhf8 and lLlxf3+ 15 ~xf3 lLld7 16 ~g3 .tg6 17 lLlc4

78
 Traxler Gambit: 4 ljjg5 i.c5

and White is much better) 14'bxf3 ,*,cS! 15 De Zeeuw says this position is much bet-
~h2 g5 16 ,*,e2 g4 (White was planning 17 ter for White. The problem for Black is that
i¥e3 so there is no time for moves like there are no real white weaknesses. Black has
16 ... a5 and ... l:ta6). 17 hxg4 'bxg4+ 1S ~g1 the f-file, but there are no targets there any-
WHeS and Black has compensation here. more.
8 i..e3 9 .. :~e8
S 0-0 is examined in Games 34 & 35. If 9 ... 'bg4 10 'bxh7 and White comes out
a pawn up.
10 ljjc3 ~g6 11 ljjf3 iVxg2?!
White would have more problems con-
verting his advantage after 11....i.g4 12 0-0
.l:!:aeS.
12 :g1 'ifh3 13 ':'xg7+ ~d8 14 J:tg3
"YIVh5 15 ~e2 i..g4 16 0-0-0
Black has problems completing his devel-
opment and he is a pawn down. In short,
White has a winning position.
16 ... ljja5 17 i..a4 i..d7 18 i..xd7 >£?xd7
19 d4 exd4 20 exd4 ljjc6 21 e5 1-0

8 ... i.xe3?!
This is often a very bad idea. All Black's
compensation is on the f-file, whereas now
he has to share it. Instead:
a) S... .i.g4?! 9 ~d2 h6 10 .i.xc5 dxc5 11
h3 .tcS 12 'bf3 '*'d6 13 'ba3 'bd4 14 'bc4
'bxf3+ 15 gxf3 ,*,e6 16 ~e3 and White is
clearly better.
b) S... 'iVe8 9 .i.xc5 dxc5 10 0-0 'iVg6 11
'bf3 Ji.g4 12 'bbd2 l:tadS, intending ... 'bh5-
f4, and Black has compensation according to
Schneider.
9 fxe3 Black had simply had enough.

Game 34
Paoli-Wagman
Comspondence 1965/66

My annotations here are based on those

by Martin de Zeeuw in New in Chess Yearbook
#66.
1 e4 e5 2 ljjf3 ljjc6 3 i..c4 ljjf6 4 ljjg5
i.c5 5 i.xf7 + ~e 7 6 i.b3 J::tf8 7 d3 d6 8
0-0
S ~e3 as played in Game 33 looks much
stronger, and it is not just appearances! After

79
Two Knights Defence

White castles short Black's attack has a fIxed iVe2 ~d7 16 i..dl with the idea of tiJc4,
target and every white mistake runs the risk iVd2 and tiJe3.
of being the last one.
S... .tg4 9 ltJf3 ltJd4!
The alternative way to attack witl1 9.. :~e8
is considered in Game 35.

11...'i¥h5
Also possible was l1...tiJxe4!? 12 .ixd4!
(the only move; if 12 dxe4? tiJxf3 and Black
wins) and now 12 ... iVh5! is the same as in the
10.i.e3!? game. Worse is 12...iLxd4?! 13 dxe4 i..xf3 14
10 tiJbd2 tiJh5 (or 1O .. :iVe8 11 h3 iVh5 12 gxf3 .ixb2 15 tiJd2 i..xal 16 iVxal 'iVh5 17
c3 tiJxf3+ 13 tiJxf3 .ixf3 14 iVxf3 iVxf3 15 'ifc3 when the position is unclear.
gxf3 tiJh5 and the position is about equal) 11 12 ltJbd2
c3 tiJxb3! (not l1...tiJe6? 12 h3 .ixf3 13
tiJxf3 tiJhf4 14 i..xe6 tiJxe6 15 tiJg5 with
huge advantage for White, while if 11 ....ixf3
12 tiJxf3 tiJxf3+ 13 gxf3 tiJf4 14 d4 .ib6 15
~hl) 12 axb3 tiJf4 13 tiJc4 (De Zeeuw be-
lieves this is much better for White; I dis-
agree) 13. .. iVd7 14 .ie3 (if 14 tiJe3?! tiJh3+
15 ~hl .ixe3 16 Exe3 tiJg5 regains the
pawn) 14... .ib6 and Black keeps up the pres-
sure.
10 .. :i¥eS?!
a) 10 ... tiJxf3+ 11 gxf3 .ih3 12 .:tel iVe8
13 ~hl iVh5 (or 13 ... tiJh5 14 .:tgl ~d8 15
c3) 14 .:tgl g6 15 tiJd2 'ifh4 16 'iVe2 .ib6 17 12 ... ltJxe4!
c3 .:tae8 18 d4 and White is better. You should remember this typical Traxler
b) lO ... .ixf3! (a new move) 11 gxf3 iVc8 tactic as it might come in handy. The number
12 c3 tiJxb3 13 axb3 'iVh3 with excellent of black pieces targeting the white king now
compensation for the pawn, e.g. 14 tiJd2 a6 becomes too great.
15 ~hl .:tf7 16 .l:i.gl i..xe3 17 Exe3 .1:.afS 18 It is also useful to compare these lines in
d4 ~d8 19 .l:!.g3 iVh6. the Traxler (after 6 i..b3 and 6 .td5) with the
11 'ith1? Janisch Variation of the Spanish (1 e4 e5 2
White should have played 11 .ixd4! .ixd4 tiJf3 tiJc6 3 .ib5 f5!? - specifIcally the line
12 c3 .tb6 13 tiJbd2 ~g6 14 <t;hl 'iVh5 15 with 4 d3 Exe4 5 dxe4 tiJf6 6 0-0 iLc5) as

80
 Traxler Gambit: 4 CiJg5 i..c5

they have a lot in common. 'iYd7 25 .ixb7 l:.e8 Black has a clear advan-
13 i..xd4 tage.
After 13 dxe4 l:txf3! 14 .i.g5+ 'iVxg5 15 15 ....txd4 1S c3
ctJxf3 ~5 16 ctJxd4 .i.xd1 17 ctJf5+ ~f8 18 If 16 gxf3 'i'xf3+ 17 1:.g2 1:.f4 18 c3 ~b6
:axd1 g6 Black has a technical won position. 19 ~dl 'iVc6 and Black still has a strong
13 ... CiJxd2 14 ~xd2 attack still. For one thing, how are the white
pieces going to get back into the game?
1S ... .tcS 17 cxd4 '=:f4 18 i..d1

14....i.xf3?
Black did not play this game well. Instead
with 14....l:!.xf3! 15 .i.d5 (if 15 .i.e3 .1:.h3! 18 ... ~f5?
wins) 15 ... SLxd4 16 .i.xf3 .i.xf3 17 c3 .i.b6 Black misplays his attack terribly. Instead
18 gxf3 'iVxf3+ 19 ~gl :f8 Black wins as 18 ...'i'h4! 19 dxe5 'i'xf2 20 exd6+ ~xd6 21
...:f6-g6 cannot be prevented; 20 'iVdl ~xf2 l:!.xf2 22 .i.f3 1:!.xf3 23 gxf3 ~xf3+ 24
~xf2+ 21 .1:.xf2 'iVxf2+ 22 ~hl 'iVxb2 IS .1:.g2 l:.e8 and Black wins.
clearly hopeless. 19 dxe5 ':'xf2?!
Now the game is drawn by force. It was
better to keep up the pressure and play
19 ... dxe5 20 f3 l:!.d4 21 l:!.el 'iitd7 22 .ib3
.l:te8 when Black keeps a slight edge.
20 exdS+ cxdS 21 'YlW'e3+ >i>f8 22 Jtg4
~f4 YO-YO
After 23 'i'xf4+ l:.xf4 24 l:!.gfl the endgame
is equal.

Game 35
Braunsdorf-Augustat
ComspOltdeltCe 1993

15.l:!.g1? 1 e4 e5 2 CiJf3 CiJcs 3 .ltc4 ttJfS 4 ttJg5

It was better to suffer with 15 .i.xc5 :f6 Jtc5 5 i..xf7+ ~e7 S .tb3 ':'f8 7 0-0 dS
16 .l:.fe1 .l:!.g6 17 J:i.xe5+ ~xe5 18 gxf3 ~xc5 7... h6!? 8 ctJf3 d6 is again interesting, and
19 ~f4, although after 19 ...l:!.h8! 20 d4 'i'g5 now:
(now Black would like the queens off) 21 a) 9 h3? ctJxe4 10 c3 ctJxf2 11 .l:hf2 .i.xf2+
'i'e4+ 'iitd8 22 f4 ~5 23 a4 'i'c6 24 SLd5 12 'iitxf2 .ixh3! (a novel idea) 13 gxh3 e4 14

81
Two Knights Defence

'iVel Mxf3+ 15 ~g2 d5 and Black has a clear If 11 CDbd2 g5! 12 h3 .ixh3 13 gxh3
advantage. 'iVxh3 14 d4 exd4 15 CDxg5 Mg8 16 .ixg8
b) 9 d4?! CDxd4 10 CDxd4 itxd4 11 c3 I:i.xg8 17 CDdf3 dxc3 18 'iVb3 CDe5 19 CDxe5
..Itb6 and Black is slightly better. 'iVg3+ 20 ~h1 'ifh4+ with a draw - De
c) 9 d3 .ig4 10 .ie3 CDh5 and Black has Zeeuw.
the initiative as in mam' similar positions in 11...ttJd7!
the Traxler.
8 d3
8CDc3 'iVe8 transposes to Game 36.
8 ... ~g4 9 ttJf3 'tWe8?!
A standard plan which gives Black some
practical compensation. Nevertheless 9... CDd4
was better as in Game 34.

A risky move, and according to De Zeeuw

it is a mistake, though in my opinion Black
has sufficient resources. Also possible is
11....ixf3! 12 'iVxf3 (not 12 gxf3? .ixe3 13
fxe3 'iVg5+ 14 ~h1 'tlVxe3 with a huge advan-
tage for Black) 12 ...CDg4 13 'iVg3 .ixe3! 14
fxe3l:.xf1+!? (or 14... 'iVxh2+ 15 'iVxh21hf1+
10 c3?! 16 ~xf1 CDxh2+ ,-,,1.th equality - De Zeeuw)
Both alternatives In this position were 15 ~xf1 MfS+ 16 ~g1 CDxh2 (not 16 ...'iVh6?
stronger: 17 CDa3 and White consolidates) 17 CDd2
a) 10 h3 .ixf3 (if 10... .ih5?! 11 c3 h6 12 CDg4 18 ..Itd1 'iVg5 19 .ixg4 h5 20.l:.f1 hxg4
..Ite3 ..Itxe3 13 fxe3 g5 14 g4 itg6 15 CDbd2 with an equal endgame.
with a clear advantage to White) 11 'iVxf3 12 ttJbd2 J:!.xf3!?
'iVg6 12 'iVf5! 'iVxf5 13 exf5 CDd4 14 .ie3 If 12... .ixe3 13 fxe3 .l:rf6 14 .idS 'li'h6 15
CDd7 15 itxd4 itxd4 16 c3 .ib6 17 g4 and 'iVe2 .l:!.afS and it is hard to flnd compensa-
White is slightly better. tion here.
b) 10 .ie3! 'iVh5 11 CDbd2 CDd4 (if 13 ttJxf3 :f8 14 d4!
11...itxe3!? 12 fxe3 'iVh6 13 'iVe2 ith5 14 c3 A standard reaction. If 14 'iVd2? Mxf3 15
g5 15 d4! - against a flank attack we should .ig5+ ~e8 16 gxf3 .ixf3 and Black wins, e.g.
always pay attention to the counter-blows in 17 .ie6 CDfS 18 .ic8 CDd7! and White is
the centre) 12 .ixd4 .ixd4 13 c3 .ib6 14 mated.
'iV e2! (not 14 h3? ..Itxh3 15 gxh3 'iVxh3 16 14... exd4 15 cxd4 ii.b6
CDh2 'iVxd317 .ic4 ~g3+ 18 ~h1 i..xf219 Demonstrating remarkable restraint. After
'iVe2 .ib6 ,-,,1.th good play for Black in Baer- 1S ... ':'xf3?! 16 dxcS! Mxe3 17 'iVdS l::!.xb3 18
Leisebein, corr. 1997) 14...'iVg6 15 ..Itd1 and cxd6+ cxd6 19 'iVxb3 CDd4 20 'iVdS CDe2+ 21
White is much better. ~h1 'iVxdS 22 exdS CDf4 the endgame is very
10 ... 'tWh5 11 ..Ite3 unclear.

82
 Traxler Gambit: 4 0,g5 iLc5

to 6... l:i.f8 positions and ending in the same

place.
One independent line is 7 d3 ttJd4!? (not
7...'Yi'g6? 8 .ltf7 'i¥h6 9 ttJf3 and wins, while
7...l::tf8 would reach the notes to Game 33) 8
i.e3 'iVg6 (8 ... l::tf8 9 0-0 d6 10 ttJf3 .ltg4 is
Game 34) 9 i.f7 'iVh6 (threatening
1O ... ttJxc2+ followed by ... i.xe3 and ... ~xg5)
but after 10 .td2! White has a clear advan-
tage, e.g. 10 ... ttJg4 11 ttJe6! ttJxf2 12 i.xh6
ttJxd 1 13 i.xg7.
7 0-0 Zif8 8 0,c3

16 ~h1?
16 .tc4! was the best move and then
16 ....ltxf3! (not 16 ...J:txf3? 17 ~e2 l:tf4 18 f3
and White consolidates) 17 gxf3 l:tf6 18
'it>hl! (if 18 l::tel ttJxd4 19 i.xd4 i.xd4 20
~xd4 'iVh3 21 'iYxf6+ 'it>xf6 22 l::te3 ttJe5 23
i.e2 ttJg6 24 f4 'ikh4 25 f5 ttJf4 and the
black initiative is very dangerous, or 18 f4
'iVh3 19 f5 ttJxd4 20 f3 ':h6 21 l::tf2 ttJxf5 22
~xh6 .ltxf2+ 23 'it>xf2 ~xh2+ 24 'it>e1 'iVxh6
25 exf5 'iVh4+ 26 'it>f1 ~xc4+ and Black is
much better) 18... ttJxd4 19 l::tg1 ttJxf3 20
l::txg7+ 'it>d8 21 J:.g8+ ~e7 22 l:i.g7+ with a The principal line.
draw. Also possible was 16 ~d3!? i.xf3 17 8 ... d6
i.d 1 i.xd 1 18 .:tfxd 1 'it>d8 19 ~ac1 when This position can also arise via 6 i.b3 l:i.f8
the position is unclear. 70-0 d6 8 ttJc3 'iie8.
16 ... ':'xf3 17 gxf3 9 0,d5+ ctJd8
Sometimes a quick death is preferable to After 9... ttJxd5?! 10 exd5 ttJd4 11 c3
prolonged suffering. And there is no escape ttJxb3 12 axb3 h6 13 d4 .ib6 14 ttJe6! i.xe6
since if 17 ~ d2 ttJxd4 18 i.d 1 (or 18 i.xd4 15 dxe6 ~xe6 16 dxe5 dxe5 17 .te3 White is
l::th3 19 'i¥f4 g5 20 'i¥f7+ 'i¥xf7 21 i.xf7 much better according to palkovi.
~xd4 22 l:i.ac1 'it>xf7 23 f3 l::th4 24 fxg4+ 10 c3 h6
'it>e6 25 lIxc7 lIxg4) 18... ttJe5 19 l:Igl ttJe2 1O .. :tWg6? 11 d4! exd4 12 ttJf4 ~e8 13
20 .ltxe2l:!.xe3 21 fxe3 i.xe2 and Black wins. ttJge6+ and White wins.
17 ... iLxf3+ 18 'i'xf31lVxf3+ 0-1 11 d4
On 11 ttJxf6?! l:i.xf6 12 d4 i.b6 Black has
Game 36 compensation for the material according to
Howell-AI. David Howell.
Groningen 1995 11 ... exd4 12 0,xf6
This is stronger than 12 e5 ttJxd5
e4 0,c6 2 0,f3 e5 3 .tc4 0,f6 4 0,g5 (12 ... .ltg4?! 13 ttJxf6 gxf6 14 ttJf7+ ~xf7 15
il..c5 5 il..xf7+ ctJe7 6 il..b3 ~e8!? ~xg4 ttJxe5 16 ~e4 ~g7 17 cxd4 .ltxd4 18
Usually this is just a different move order 'iYxb 7 left White much better in Winkel-

83
Two Knights Defence

mann-Koch, corr. 1971) 13 ..ixdS dxeS 14 Also interesting was 16 ... ~f8!? when after
ctJe4 ..ib6 IS cxd4 ctJxd4 16 ..ie3 c6 17 ..ic4 17 Wh 1 ctJg4 18 ~ e2 dS 19 ctJf3 c6 20 .JtxcS
WVg6 was given as unclear by Howell. In my ~xcS 21 I:t.ae1 .Jtd7 (not 21...l:!.e6?! 22 WVd2
opinion \X1hite does not have enough com- l:!.xel 23 ~xe1 ctJf6 24 'iieS and \X1hite is
pensation here; e.g. 18 ctJg3 ..ie6 19 ..ixe6 better) 22 h3 l:!.e6 23 'iVc2 nxel 24 ctJxel
'iVxe6 20 .l::tel 'iVg6 21 ctJe2 cS 22.l::tc1 .l::tf7 23 ctJf6 2S "iWg6 'iVe7 26 ctJd3 ctJe4 Black is OK.
ctJg3 .l::td7 24 ctJe4 Mc8 and Black is better. 17 J:!.e1 CDxe3?
12 ... ':xf6 13 e5 litf5 This exchange is meaningless as \X1hite
If 13. ...i:tf8 14 exd6 hxgS IS .txgS+ ':f6 quickly develops his remaining forces. Better
16 ..ixf6+ gxf6 17 dxc7+ lfJxc7 18 l':.c1 and was the active 17 ... "iWeS! when after 18 ctJf3
\X1hite is better - Howell. ctJxe3 191:!.xe3 "iWf4 20 l:!.el ':f8 21 ~e2 'i¥f6
14 CDf3 22 .l::tadl .Jtg4 23 J::td3 c6 24 WVeS 'iVxeS 2S
l:!.xeS the position is more or less equaL
18 J:!.xe3 'ilff8 19 ~e2!

14 ... CDxe5
Black could also consider:
a) 14... dxeS IS .Jtc2 ':hS!? with an The tripling of the heavy forces on the e-
interesting mess. If instead IS ...l:!.f6 16 b4 fIle assures \X1hite of a solid advantage .
.tb6 17 bS and \X1hite is slightly better 19 ... c6 20 J:!.e1 ~d7 21 ~e6 ~xd4 22
according to Howell. cxd4 ~xe6 23 J:!.xe6 ':xe6 24 'i!Vxe6 l:!.c8
b) 14... dxc3 IS exd6 i.xd6 16 bxc3 WVhS 25 J:te3 rJ;;c7?
17 .lli.a3 .l::tf6 and a draw was agreed in Leise- Better was 2S ... dS, although after 26 l:.a3
bein-Schiiler, corr. 1998. \X1hite has a huge advantage.
15 CDxd4 %:f6 16 ~e3 CDg4 26l:H3 1-0

84
 Traxler Gambit: 4 4Jg5 .i.c5

Summary
Traxler's 4... i.cS can perhaps only be refuted in the solid variation 5 .i.xt7+ cj;;e7 6 .i.b3 Mffi 7
d3!. In my opinion Black has sufficient compensation after 7... h6!, but practical testing is obvi-
ously required. The Traxler is a good line for amateurs and club players and Black scores well
after both 5 'Llxt7 and 5 d4. There are of course more pressing problems with 4 ... i.cS 5
i.xt7+ than after 4 ... dS, but this is the high risk life. That a player such as Beliavsky has played
4 ... i.cS a few times does not necessarily guarantee that it is completely sound, but it means at
least that he thinks it gives him adequate chances as a surprise weapon, even against Karpov.

1 e4 e5 2 4Jf3 4Jc6 3 .i.c4 4Jf6 4 4Jg5 .i.c5 5 .i..xf7 +

5 d4- Game 25
5 'Llxt7 i.xf2+ (D)
6 'iitxf2 - Game 26
6 cj;;n - Game 27
5 ... ~e7 (D) 6 .i.b3
6 i.dS
6... d6
7 c3 - Game 31; 7 d3 - Game 32
6 ... MfS
7 .i.xc6 - Game 28
70-0 d6
S c3 - Game 29; S h3 - Game 30
6 .. JU8 7 d3
70-0 d6
S d3 - 7 d3; S'Llc3 it'eS - Game 36
7 ... d6 8 0-0
S ~e3 - Game 33
8 ... .i..g4 9 4Jf3 (D)
9...'Lld4 - Game 34; 9.. :~eS - Game 35

5 ... .i.xf2+ 5 ... ~e7 94Jf3

85
 CHAPTER SIX
I
4 d4 exd4:
Introduction

1 e4 e5 2 tLJf3 tLJc6 3 ii.c4 tLJf6 4 d4 common move order is 2 liJf3 liJc6 3 d4

exd4 exd4 4 J..c4 liJf6, though there Black can
In this chapter we deal with the position also consider 4 ... i.c5.
after 4 d4 exd4 - in particular, the variations 5 tLJg5?!
following 5 e5 where Black avoids the imme~
diate counter 5... d5 (which is covered in
Chapter 7). As you will see below there is no
reason to disregard either 5...ctJe4 (Games 38
& 39) or 5 .. .'~Jg4 (Games 40~42); both are
perfectly playable moves without existing
refutations. At lower levels 5...liJg4 might
prove especially effective as many players
might be tempted to play 6 J..xt7+, but as
shall be revealed in the notes to Game 40 this
is not sound at all. Apart from 5 e5 and 5 O~O
(the subject of Chapter 8) White has another,
weaker possibility,S liJ g5?!, as seen in Game
37. On the package from the variation manu~
facturer it says: 'Please note that this should
Game 37 only be used in games with a fast time con~
Bucan-Geller trol. In classical games, the use of this move
Bad If/iirishrifen 1992 may lead to serious injuries, and should be
ventured only at the customers own risk.'
1 e4 e5 2 ii.c4 tLJf6 3 d4 exd4 4 tLJf3 Why? Well, White is breaking one of the
tLJc6 main rules of the opening by playing liJf3-g5
After a common alternative move order so soon. He is attacking before fInishing
we have arrived at the starting position for development, when the advised behaviour is
this chapter. It should be said that 4 ... liJxe4 is the other way round. And while there was an
perfectly possible, but that belongs to the 2 argument for discounting that rule in the case
i.c4 system and not this book. Another of 4 liJ g5 due to Black's vulnerability at t7,

86
 4 d4 exd4: Introduction

here there is little difficulty in defending that with a very strong initiative) 16 ... 'iVd5! and
square. Black is better, perhaps a lot better, as White
S... dS! has big problems with his development.
b) 6 i..b3 (more dangerous) 6... h6 7 f4
hxg5 S fxe5 l2Jxe4 and then:

The voice of Wilhelm Steinitz speaks

clearly from heaven above: 'Against an attack
on the flanks, you must counter-attack in b1) 9 'iVxd4 l2Jc5 is harmless, 9...l2Jc5 10
centre!' Of course it is not always so simple, l2Jc3 d6 11 i.e3 l2Jxb3 12 axb3 dxe5 13
but here it is. 'iVxe5+ 'iVe7 14 'iVxg5 'iVxg5 15 i..xg5 .i.d7
The alternative, 5...l2Je5, is not as good, and Black was slightly better in Pfleger-
especially because by playing this move Black Spas sky, Hastings 1965/66.
also violates general principles, and instead of b2) 9 i.d5!? is an old (13 years old any-
developing the pieces he makes a second way) idea of by Jurij Zezulkin, a present day
knight move as well. White then has: grandmaster. It leads to a very unclear game,
a) 6 'iVxd4l2Jxc4 7 'iVxc4 d5 S exd5 'iVxd5 and is another good reason not to play
9 "*,e2+ i..e7 10 0-0 and here Black should 5...l2Je5. The best response is 9.. .f5! (that the
play very carefully in order to keep the bal- line is dangerous is shown by 9... i.b4+? 10
ance: c3 dxc3 11 bxc3 l2Jxc3 12 .txf7+ WfS 13
a1) 1O ... JLg4? 11 f3 h6? 12 l2Jc3 'iVc6 13 ~3 with a strong attack in Zezulkin-
fxg4 hxg5 14 i.xg5 and White is clearly bet- Kalesnik, l\1insk 1990, while if 9...'iVe7 10
ter, while if 11...i.f5 12 l2Jc3 'iVc6 13 i.f4 i..xe4 'iVxe5 11 lIVe2 Black's three pawns are
and White has a dangerous initiative - usually not enough for the knight in my opinion) 10
you have to sacrifice a pawn or two to get to exf6 l2Jxf6 11 'iVd3 'iVe7+ 12 Wd1 WdS 13
this kind of position; here White has sacri- i.xg5 'iWe5 14 'iWg6 c6 with an unclear posi-
ficed nothing. tion in Watson-Adams, English Champion-
a2) 10... .i.d7! is correct, when White must ship 1991.
be content with an equal game after 11 l2Jc3 6 exdS "tWe7+!
"*,c5 12 i..e3 "*,f5. Instead 11 .l:!.e1?! allows This is the downside to White's attacking
11...0-0! 12 'iVxe7 (if 12 l2Jxh7?! l2Jxh7 13 plan. There is no sensible way to defend
'iVxe7 i.c6 14 f3 .l:!.feS 15 'iVb4 l:t.xeH 16 against this check.
'iVxe1 .l:!.eS 17 .i.e3 "*,e5 1S Wf2 l2J g5! and 7 'iio>f1
the question is not if, but when) 12...lIaeS 13 This is the only move. If 7 Wd2? 'iVb4+
'iVb4 'iVc5 14 i..d2 ~xeH 15 i.xe1 "*,xg5 16 wins, or 7 'iVe2 'iVxe2+ S ~xe2 l2Jb4 9 i..b5+
i.d2 (if 16l2Jd2 i..c6 17 f3l:!dS or 17 ... B.eS!? i.d7 10 i.xd7+ l2Jxd7 and Black wins a

87
Two Knights Defence

pawn with no hint of serious compensation S...CtJxc4 9 'iYxc4 h6 above.

for White. 9 ... tLJxc4 10 tLJxfS + ~xfS 11 'iVxc4 ii.dS
7 .. .cue 5 S 'iVxd4

12 tLJc3?
S ... hS White allows Black to develop effortlessly.
This is simply a matter of move order. Much better was 12 'iVe2+ ~e7 13 CtJc3 0-0
Black can also take the bishop immediately where Black merely has very good compen-
with S...tDxc4 9 'iYxc4 and then: sation for the pawn.
a) 9.. :iVc5 10 'iVxc5 (if 10 'iVe2+ i.e7 11 12 ... 0-0 13 tLJe4?
c4 CtJxd5 12 CtJe4 'iYc6 13 i.g5 CtJf6 14 White hopes to ease his defence through
CtJxf6+ gxf6 and Black is at least slightly bet- exchanges, but trading the knight on d6 will
ter) 1O ... i.xc5 11 CtJc3 ~f5 12 i.f4 0-0 with leave Black with a deadly attack enhanced by
sufficient compensation for the pawn. the opposite-coloured bishops, as White will
b) 9... h6!? 10 CtJc3 (if 10 CtJf3 'iVc5 11 have nothing to resist him on the light
'iVxc5 i.xc5 12 c4 i.f5 and Black has more squares. Preferable was 13 ~e3 i.f5 14 .l:!.c1
then enough compensation for the pawn) .l:!.feS 15 Wg1 and although Black is much
1O ... hxg5 11 i.xg5 'iVc5 12 .l:!.eH Wd8 13 better, White still has some chances to sur-
'iVf4 (if 13 'iVe2 Ji.d7 14CtJe4 Ji.b5! 15 ~xf6+ vive.
gxf6 16 CtJxc5 ~xe2+ 17 ':xe2 Ji.xc5 and
Black wins) 13 ... i.e7 14 h4 and now
14... Ji.d7?! 15 h5 'itcs 16 ~h4 ~d6 17 'iVf3
CtJeS 1S h6 gxh6 19 i.xh6 f5 was played in
Carleton-Franzen, corr. 1991-93. Franzen
believes that this position is slightly better for
Black, and as he spent two \'ears playing this
game, possibly he is right. Nevertheless Black
can play more strongly bv bringing the as-
rook into the game after 14... a5! 15 'iVe5 .l:!.a6
with a clear, possibly even decisive advan-
tage. Many chess players forget that the rook
can also develop forwards and not just to the
side. 13 .. ."Yi'gS 14 tLJxdS cxdS 15 ii.f4 ii.f5
9 tLJe4 With the king's rook boxed in on the h1-
9 CtJc3 CtJxc4 10 'iVxc4 transposes to square, White is virtually playing a rook

88
 4 d4 exd4: Introduction

down.
16 ~b4 i.xc2 17 i.xd6 as 18 'YWcS
Or IS 'tWf4l:tfdS 19 il.c7l:i.xds and White
can only wait for the end.
18 ... l:l.fc8 19 ~a3 i.b1!

6~e2
The main moye 6 i.ds is considered in
Game 39. White can also play 6 0-0 with two
possibilities:
a) 6oo.i.e7!? 7 lbxd4?! (7 .l:Iel ds S exd6
Now the second white rook is set out of lbxd6 9 i.ds lbfs transposes to 6oo.ds)
play too. The game is virtually over. One 7oo.lbxes Slbfs il.f6 9 'iVdslbxc4 10 'tWxe4+
cannot defend playing two rooks down. lbes 11 f4 ds 12 'iVe1 i.xfs 13 fxes i.h4 14
20 i.f4 .l:tc4 21 ~g3 'YWd3+ g3 il.h3 15 gxh4 il.xfl 16 Wxf1 'iVd7 and the
There is no reason to make it difficult. Hungarian 1M Jozsef Pilkoyi, who found
This is more than good enough to fmish off this line, reckons that both players have
the game immediately. chances. In my opinion Black is clearly better
22 'YWxd3 i.xd3+ 23 ~e1 Iixf4 24 ~d2 as White has ongoing problems with the
~d4 0-1 safety of his king.
b) 6oo.ds 7 exd6 (7 i.bs transposes to the
Game 38 soo.ds main line in Chapter 7) 7oo.lbxd6 S
V.Gurevich-Jonkman il.ds lbfs 9 .l:Ie 1+ il.e 7 10 i..xc6+ bxcG 11 g4
Germany 2002 lbh6 and now:

1 e4 eS 2 ttJf3 ttJc6 3 d4 exd4 4 ii.c4

ttJf6 S eS
This has been the modern way of han-
dling the position since the classical lines
with 5 0-0 were exhaustively investigated.
Black now has three options: s... lb g4!?
(Games 40-42), s ... ds (the main line covered
in Chapter 7) and the text.
S... ttJe4!?
This move, like many others, has been a
victim of fashion - unfairly, as it gives Black
a good game. Of strong grandmasters only
Romanishin and l\1ikhalchishin have played it bl) 12 'iVxd4 il.xg4 13 i..xh6 'iVxd4! (infe-
with any regularity. rior is 13oo.i..xf3?! 14 'tWxg7 Wd7 15 Wfl!

89
Two Knights Defence

when 15 ...'iVg 8? loses to 16 l:!.xe7+! rtlxe7 17 13 ... ~xdS

'iVe5+ 'it>d7 18 'iVf5+, while after the forced Black has no worries about accepting an
15 ... i..dS 16 tDc3 .ic4+ 17 tDe2 .id6 18 isolated pawn, as he is quite active here.
'iVd4 .ie6 19 tDg3 White is simply better) 14
tDxd4 gxh6 15 tDxc6 .ie6 16 tDxe7 rtlxe7 17
tDc3 l:!hg8+ and Black has good attacking
chances to compensate for his ruined pawn
structure.
b2) 12 i..xh6 gxh6 13 ~xd4 'iVxd4 14
tDxd4 ~f8 15 tDxc6 .if6 16 tDc3 ..ixc3 17
bxc3 .ixg4 and Black is slightly better.
b3) 12 .igS ..ie6 13 i..xe7 'iVxe7 14 'iVxd4
tDxg4 15 tDc3 tDh6 16 tDe5 0-0 17 tDxc6
'iVgS+ was Sveshnikov-Smikovski, Togliatti
2003, and for a position like this for Black
one pawn is a very small price.
S... ttJc5 7 0-0 ~e7 14 ~e3 e5 15 .l:!.d2 'iVe8 1S ttJe4 'iVgS 17
Simple and good. Also interesting IS f3 ~f5 18 l:!.ad1 ~xe4 19 ~c4+ 'iVf7 20
7... tDe6 8 l:!.dl (if 8 i..xe6 dxe6 9 l:!dl .ics ~xf7+ ~xf7 21 fxe4 %-%
10 c3 i..b6 11 .ie3 'ii'e7! 12 .ixd4 tDxd4 13
tDxd4 i..d7 with complete equality) 8... dS 9 Game 39
.ibS ..ics 10 c3 .i.d7 11 i.xc6 i..xc6 12 Kozakov-Jonkman
cxd4 ..ib6 13 tDc3 0-0 and both players have Lvov 2001
their chances according to J\1ikhalchishin.
8 l:id1 ttJeS 9 ~xeS fxeS!? 1 e4 e5 2 ttJf3 ttJcS 3 d4 exd4 4 .i.c4
ttJfS 5 e5 ttJe4!? S ~d5 ttJc5
Too risky is 6... i..b4+?! 7 c3! dxc3 8 O-O!
and White has terrific compensation.
7 ~xcS?!
After this Black obtains the two bishops
and at least two moves for free. The alterna-
tives were:
a) 7 tDxd4 tDxd4 8 'iVxd4 tDe6 9 'iVc3 d6
100-0 dxeS 11 'iVxeS ..id6 12 iVhS 0-0 with
equality.
b) 7 0-0 i.e7 8 'iVe2 (other moves also
lead to equality, e.g. 8l:!el 0-09 tDxd4 tDxd4
10 'iVxd4 d6 11 tDc3 tDe6 12 'iVe4 c6 13 exd6
Very logical, but this was still a novelty. ..ixd6, or 8 tDxd4 tDxd4 9 'iVxd4 0-0 10 tDc3
Instead after 9... dxe6 10 i.e3 0-0 11 tDxd4 c6 11 ..in d6, or 8 tDbd2 tDe6 9 tDb3 0-010
White has some advantage, e.g. 11...'iVdS 12 l:!el tDb4 and Black is at least equal) 8... 0-0 9
f4 tDxd4 13 l:!xd4 'iVc6 14 tDc3 and Black l:!dl 'iVe8 10 tDa3 (not 10 i.f4?! b6! 11 ..ixc6
still has difficulty in developing. d3! 12 cxd3 dxc6 13 d4 tDe6 14 ..ig3 .ib7 15
10 ttJxd4 ttJxd4 11 l:txd4 0-0 12 ttJc3 d5 tDc3 l:!d8 and in Khmelnitsky-Romanishin,
13 exdS Sibenik 1990, Black had the advantage with
Otherwise Black takes over the centre. his two bishops and control of the light

90
 4 d4 exd4: Introduction

squares) 10... tbb4 11 ilLc4 d3 12 cxd3 dS 13 side majority.

exd6 ..IlLxd6 14 'iWxe8 l:txe8 15 tbbS ~g4 was 7 .. .dxc6 8 VWxd4 i..f5
equal in Khmelnitsky-Malaniuk, Sibenik
1990.
c) 7 c3!? is the dangerous move, but it is
unlikely to take any S... tbe4 players by sur-
prise. After 7... dxc3 8 tbxc3 Black has:

Black already has the more comfortable

development.
9 VWc3
9 iLgS is solidlv met by 9...'iWc8 10 tba3
tbe6 11 ~c3 .ics (l1...iLxa3?! 12 'Yi'xa3!
c1) 8... ~e7 9 ..IlLe3 0-0 10 'Yi'e2 tbb4 11 tbxgS 13 tbxgS ..IlLxc2 wins a pawn, but
0-0-0 c6 12 ..IlLxcs ..IlLxcs 13 ~b3 dS 14 a3 White has reasonable counterplay here as
tba6 15 h4 bS 16 tbgs b4 (it looks as if Black has diffIculues getting the rest of his
White will cross the fInish line in second pieces into the game) 12 iLh4 0-0 and it
place) 17 ..IlLc2 g6 18 axb4 tbxb4 19 .ibl was seems that Black has a slight advantage due
Hector-Nunn, Vejle 1994. In my opinion to his better placed pieces and two bishops.
Black's attack is the stronger here, although 9 ... ti:Je6 10 i..e3 iYd5!
playing such a position is akin to strolling I prefer Black here, whose control over
around in a minefIeld - for both sides of the light squares is very important.
course. Now Nunn continued 19 ....l:i.b8, but I 11 ti:Jbd2 i..e7
prefer 19 ...'i¥b6 with good play.
c2) 8 ...tbb4!? (a new and untested idea, de-
signed for those who hate defending - Black
wants to use the weakness of the d3-square
to his advantage) 9 ..IlLc4 (after 9 0-0 tbxdS 10
~xdS i..e7 11 ..IlLe3 tbe6 and Black's position
looks bullet-proof, e.g. 12 tbbS a6 13 tba7 c6
14 ~b3 bS and there is no reason why White
should be better; more likely he is just a
pawn down) 9... dS 10 tbxdS tbxdS 11 ..IlLxdS
i..e6 12 .ixe6 ~xd1+ 13 ~xdl tbxe6 14
'It>e2 iLe7 with a comfortable endgame for
Black. The knight on e6 has strong control
over the centre and White has no easy way of 12 a3?!
removing it. If allowed Black will play ... c7-cS White is preparing to castle queenside, but
and ... ~d7-c6 in hope of utilising his queen- this is a misunderstanding of the position.

91
Two Knights Defence

Better was 12 0-0 0-0 13 ctJb3, with reason- as White now has no control at all over the
able chances to equalise. light squares in his position.
12".0-0 13 0-0-0 c5! 22 ttJd2 .l:.d8
Black takes control over d4, preventing What Black is trying to achieve here is not
the manoeuvre ctJd2-b3-d4. dear. Perhaps it is simply that Jonkman is
13 ...'ilVa2? surely looks attractive, but after very fond of the endgame and has a tendency
14 'ilVb3 'iVa1+? ~ (14 .. .'iixb3 15 ctJxb3 is to seek it for no better reason than that it is
roughly equal), 15 4Jbl in reality all Black has possible. Of course he has a great under-
done is risk losing the queen, though there standing of the endgame to assist him, once
are no guarantees of course. he makes it there. Personally I prefer 22 ... bS!?
14 ttJb3 iVc6 15 J::.d2 J:tfd8! (my hand will often make attacking moves
In a position where you have the advan- like this without consulting me about central
tage and will win 'if nothing happens', it is control); after 23 ctJb3 b4 24 axb4 c4 the
usually important to prevent counterplay. target is set and Black will probably win by
Here Black exchanges a pair of rooks and direct attack.
thereby decreases the significance of the 23 ttJf1 liIxd1 + 24 ~xd1 .tg4+ 25 ~c1
open d-file. Had he not done so, he would 't!IYd5
have to consider exchanging all the rooks Black centralises and exchanges into a
markedly lessening his attacking chances, or dearly superior endgame.
else allow White to penetrate at some point. 26 b3?
This creates a new weakness at a3, which
becomes immediately apparent after Black's
next move.
26."lt'ld4!
Now there is no defence. Both ... 4Je2+
and .. .'iVxeS are threatened.
27 .txd4 cxd4 28 'Wixc7 .txa3+ 29 ~b1
iVh1 !

16 J:thd1 liIxd2 17 liIxd2 a5

Here comes the attack.
18 ttJxa5
This does not look sound, but White is in
a pretty bad fix. If 18 :d 1 a4 19 ctJbd2 b5
and Black's attack is easy to play and re-
markably strong.
18".'t!IYa6 19 ttJb3 't!IYf1+ 20 .l:.d1 't!IYxg2
21 ttJe1 'Yi'c6 Rarely has the difference between bishops
The cS-pawn is preventing all White's and knights been so dearly exhibited as in
counterplay and is therefore far more impor- this position.
tant then the pawn on h2. The exchange of 30 iVc4 .th3 31 e6 fxe6 32 b4 't!IYxf1 33
a-pawn for g-pawn, however, benefits Black 't!IYc8+ >itf7 34 't!IYd7+ ~g6 0-1

92
 4 d4 exd4: Introduction

r-----------------" iLxc7 d6 10 iLbS iLd7 11 iLxc6 ~xe2+ 12

Game 40 ~xe2 iLxc6 and Black is no worse; his next
Borge-Hector moves are ... ~d7 and ... l:te8+ which can be
Copenhagen 1994 hard to meet for White) 9.. :~xe2+ (also in-
teresting is 9... dS 10 iLbS 'iVxe2+ 11 ~xe2
1 e4 e5 2 CDf3 CDc6 3 i..c4 CDf6 4 d4 a6! 12 iLxc6+ bxc6 13 iLxc7 iLg4 14 iLeS cS
exd4 5 e5 CDg4!? 15 .l:!.hel and the game is unclear) 10 ~xe2
..ics (the safest option; after 1O ... d6 11 .l:!.hel
..if5 12 ~f1+ ~d7 13 CDb3 iLxc2 14 CDbxd4
CDxd4 15 CDxd4 iLg6 16 iLe6+ ~d8 17 iLh3
White has compensation for the material
according to Sveshnikov) 11 .l:!.hel dS 12
~f1+ ~d8 13 iLd3 CDb4 14 CDb3 ii.b6 15
CDbxd4 CDxd3 16 cxd3 iLd7 with an equal
position. Sveshnikov-Ehlvest, Helsinki 1992,
continued 17 CDgS ~c8 18 ~acl l:te8 19
CDge6 c6 20 h3 CDhS 21 iLh2 iLxe6 22 CDxe6
'it>d7 23 CDcS+ iLxcs 24 l:i.xe8 the players
agreed a draw.
9 CDbd2
The second ugly sister to the beautiful 9 0-0 'iYxe2 10 iLxe2 CDgeS 11 CDbd2 iLd6
S... dS. As with S... CDe4, this move is not 12 iLg3 CDxf3+ 13 CDxf3 iLxg3 14 fxg3 d6
openly attractive, but it has a good character was equal in Tzermiadianos-Socko, Istanbul
indeed. 2003.
6 "iife2 9 ... d6
White wants to prevent ... d7-d6 and pro-
tect his e-pawn at the same time. The alterna-
tive, 6 0-0, is considered in Game 42.
A famous mistake is 6 iLxf7+? ~xf7 7
CDgS+ ~g8! 8 ~xg4 (8 'iYf3? iLb4+! 9 c3
CDgxeS 10 "iVdS+ ~f8 is even worse: if 11 0-0
~e7 12 cxb4 CDxb4 or 11 cxb4 CDxb4! and
Black wins) 8... M 9 CDf3 d6 10 'iVg3 (10
'iVe4? dxeS 11 CDxeS ~e8! wins by force after
12 f4 iLd6 13 0-0 CDxeS 14 .l:!.el .if5 15
'i'dS+ 'i'f7 16 'i'xb7 .l:!.f8 17 fxeS iLe4! and
everything is as clear as can be) 10... "iVe8! 11
0-0 dxeS 12 I:te 1 iLfS and Black is much
better. 10 CDb3
6 .. :~lVe7 7 i..f4 f6 White needs to do something to get an ac-
7... d6 is examined in Game 41. tive game. After 10 "iVxe7+?! ~xe7! (better
8 exf6 gxf6!? than 10 ... ..ixe7?! 11 .ibS! when White has
The idea behind this move is simple - some initiative, although Black is probably
Black wants to build a blockade on the eS- still OK) 11 0-0 CDgeS 12 .l:!.fe1 'it>d8 and it is
square. not so easy for White to prove compensation
Also acceptable is 8...CDxf6 9 CDbd2 (if 9 for the pawn.

93
Two Knights Defence

10 .. .tLlee5 should focus on stopping the black pawns.

Also fIne is 1O .. ."iVxe2+ 11 .ltxe2 d3! 12 bite his lip and put the knight back on b3,
cxd3 (12 ~xd3 lLlb4 and Black will gain the although Black is still better after 16lLlb3 b6
two bishops, which may eventually be a deci- 17 l:tfel i.b 7 18 lLlh4 f5 19 'iVg3 'iVf6 20
sive factor) 12 ... i.h6 13 .ixh6 lLlxh6 14 h3 lIadl .l:!.ae8.
i.d7 15 g4 0-0-0 with full equality in Zelcic- 16 .. JH7 17lLle2 i.e6 18 Vi'e3 d5
j\1ikhalchishin, Nova Gorica 2002.
11 0-0 lLlxe4 12 ~xe4 lLle5
The blockade on e5 has certainly proved
itself usefuL
13 Vi'xd4 i.g7
13 ... lLlxf3+!? is interesting, if rather incon-
sistent. After 14 gxf31Ig8+ 15 Whl "iVf7 (not
15 ... ..ih3? 16 Mfel i.g2+ 17 Wgl .ih3+ 18
..ig3 and White wins) 16 l:i.gl lIxg1+ 17
l:!xg1 it will be diffIcult for the black king
ever to fmd safety. So Black's choice in the
game is understandable. Still, you have to
take some risks sometimes.
14 Vi'e3?! Black stands much better now. White has
On c3 the queen faces unemployment. 14 nothing with which to resist the pawns.
"iVe3 looks better and then: 14... 0-0 15 lIfe 1 19lLlh4 d4 20 'i/Hg3 Wh8 21 i.e1?
'iVf7 16 lLlbd4 .ltd7 17 lLlh4!? (Myrvold- White wants to bring the knight to f4, but
Karpatchev, Gausdal 1993, saw 17 a4?! this is not really realistic. 21 .s.fel would be
which is hard to understand - moves like better. Of course this is not a honeymoon -
these just creates weaknesses; after 17 ... a6 18 Black probably plays 21....s.g8 with an attack
lLld2 Mfe8 19 'tib3 lLlg6 20 .ie3 f5! Black - but White is still alive.
was already slightly better) 17.. .f5 18 "iVg3 21 ... .l:!.g8
~h8 19 lLlhf3 lLlxf3+ 20 lLlxf3 and White is 21....ic4!? also looks strong. After 22 .s.el
better here, since if 20 ... .ixb2 21 lLlg5 "iVg 8 .s.e8 White cannot hold on to his material
22 l:lab 1 l:!.ae8 23 J:i.xe8 l':!.xe8 24 h3 and after and his best chance (in view of the threat-
25 .s.xb 7 Black will surely regret his actions. ened 22 ... lLlc6) is to play 23 .id2, but after
An important point here of course is that 23 ... .ixe2 24 :rxe2 lLlf3+ 25 lLlxf3 ~xe2 26
24 .. :iVxa2? fails to 25 l:lxb2 ~xb2 26 lLlf7 Mel 'iVb5 there is no real hope. Black has just
mate. won the exchange for nothing.
14... 0-0 15lLlbd4? 22 lLlf4
This looks natural but is a very bad move! The fIrst time I saw this game I half ex-
The knight was better on b3 than it will be pected White to play Whl and lLlg1.
on e2. 15 lIfe 1 was more logical, as the rook 22 ... i.h6
had yet to join the game. After 15 ... b6! 16 From here on Black misplays his position
Me3 ..tb7 17 :rael 'iVf7 the two bishops and somewhat. Simplest was 22 ... .if8! 23 ~a3
a more clear plan ensure some advantage for .ic4 24 .s.dl f5 25 lLlf3 lLlxf3+ 26 ~xf3
Black. "iVe4! 27 'iVxe4 fxe4 and the black position is
15 ... e5 16 'ifb3+?! simply overwhelming.
This is an excellent illustration of what can 23 Vi'a3
happen when you play without a plan. White The only vacant square.

94
 4 d4 exd4: Introduction

23 ... .i.g5 24 4Jxe6 ~xe6 25 4Jf3 ture. The advantage of 7.. .f6 was that the d4-
White cannot play 25 j.,xgS fxgS 26 f4 (if pawn could still be protected with ... c7-cS
26 ctJf3 ':xf3! 27 gxf3 'iVh3 is the end) eventually.
26 ... gxh4 27 fxeS .l:!.xg2+! 28 ~xg2 'iVg6+ 29 8 exd6
~h1 ~e4+ 30 ~gl l:tg7+ and it is all over
Casanova.
25 ... .i.xc1 26 4Jxe5 l:i.fg7 27 g3 .i.f4?
This works out in the game, but actually
White can defend his position. Stronger was
27 ... j.,xb2 28 ~xb2 fxeS with a clear extra
pawn and a continued attack by ... e5-e4-e3.
28 4Jf3 ~h3
Not 28 ... ~e2 29 iVd3 and White is on the
way to a preferable endgame!
29 4Je1??
29 ~h1! with unclear play was the only
move here. Black has no way in on the light
squares and will not get any further with the 8 ... cxd6!
attack now. Black has slightly better chances, The only move promising reasonable
as White is still under some pressure, but it is chances for equality. If 8...'iVxe2+ 9 j.,xe2
nothing serious. ~xd6 10 i..xd6 cxd6 11 ctJa3! itfS 12 ctJbS
and ctJfxd4 gives White the slightly better
game. However, Black can try 9... ctJb4?!?
(Adorjan's idea, which creates an amazing
mess on the board and, being relatively un-
known, is a good weapon for quick games)
10 CDxd4 (not 10 dxc7? CDxc2+ 11 ~d2
ctJxa1 12 i.b5+ i.d7 13 .l:teH i.e7 14 itd6
ctJe3! 15 fxe3 CDc2 16 ~xc2 d3+ 17 j.,xd3
itxd6 and Black wins according to Pilk6vi)
10... cS!? (wild, and probably unsound, but
after 1O ... itxd6 Black is slightly worse) 11
ctJbS! CDxc2+ 12 ~d2 CDxa1 13 f3 'it>d7 (but
not 13 ... CDf2? 14l:tfl ~d7 15 CDc7 i.xd6 16
29 ... .i.e3! 0-1 i..xd6 ~xd6 17 CDxa8 i.fS 18 CDa3 and
There is no defence against 30 ... .:xg3+ White wins - Pilk6vi) 14 fxg4 a6 15 .l:.e 1.
and mates. Supposedly the game is unclear here, but is
this really the case? After lS ...g5 16 i.g3 hS
Game 41 17 gxhS f5 18 i.eS .l:!.h7 19 CDc7 it does not
Palkovi-Wells look as if Black will survive, while 15 ... ~c6 is
Zalakaros 1998 bad because of 16 CDc7 ltb8 17 i.f3+ ~d7
18 Ite7+! (improving on Pilk6vi's 18 ctJdS?!
1 e4 e5 2 4Jf3 4Jc6 3 .i.c4 4Jf6 4 d4 i.xd6 19 ctJb6+ ~c7 20 i.xd6+ ~xd6 21
exd4 5 e5 4Jg4!? 6 'i!fe2 'Wie7 7 .i.f4 d6 CDc3 with only a slight advantage) 18... i.xe7
This is more natural than 7... f6, but also 19 dxe7 ~xe7 20 CDdS+ 'it>e6 21 j.,xb8.l:.d8
gives Black a weakened central pawn struc- 22 CDbc3 bS 23 ~cl and White is probably

95
Two Knights Defence

winning here. Improvements on this analysis ttJbxd4 ttJxf3+ 14 ttJxf3 i.e7 15 llac1 i.e4
are called for the assessment of unclear is to 16l:tfdl .l:!.d8 17 ttJd2! and the white pressure
be justified. 1S mcreasing.
9lLlbd2 13 cxd3 i.xd3 14l:Ue1?!
After 9 0-0 ttJge5! and Black is at least White is drifting a bit. Stronger was 14
equal; in particular if 10 .i.b5 i.g4 11 ttJbd2 i.xe5! i.xe2 (or 14... ttJxe5 15l:Uel i.xe2 16
0-0-0 the real question is if White can hold l:he2 f6 17 lLlfd4 'it'd7 18 l::td 1 g6 19 lLlb5
equality or not. Alternatively 9 lLla3 a6 Oess with initiative) 15 .i.xg7 i.xfl 16 .i.xh8 i.c4
clear is 9... lLlge5 10 0-0-0 lLlxc4 11 ~xc4 as and the position is roughly equal.
Black has problems getting his king into 14 ... i.xe2 15 libe2 f6 16 lLlfd4
safety) 10 ~dl ~xe2+ 11 'it'xe2 i.e6 with an
even game.
9 ... i.f5 10 O-O?!
A very ambitious move; White sacrifices a
pawn to unbalance the position. Instead after
10 lLlb3 (but not 10 i.b5?! 'iVxe2+ 11 ~xe2
0-0-0 and Black is better) 1O ... d3 11 cxd3 (or
11 i.xd3 j,xd3 12 cxd3 ttJb4) 11...'iVxe2+ 12
~xe2 ttJ ge5 the position is equal.
10 ... iYxe2 11 i.xe2

16 ... lLlxd4?!
This does not really make a lot of sense, as
it invites the white knight on b3 back into the
game for no reason. After 16... 0-0-0!? it is
hard to prove that White has sufficient com-
pensation. Palk6vi writes in his annotations
that White will have enough play, and per-
haps that is so, but only enough for a draw'
For example, 17 l:tc2 (or 17 .l:!.dl g5 18 .i.g3
h5 19 lLle6 l:te8) 17...'it>b8 18 ..ixe5 ttJxe5 19
11...lLlge5?! ttJe6 l:tc8 20 .l:!.ac1 .i.e 7 21 ttJxg7 lLld3 22
Black is being unnecessarily careful. l:txc8+ l:!.xc8 23 l:txc8+ ~xc8 24 lLlf5 ~d7
Stronger was 11....i.xc2 12 i.b5 d3! (not and White has some problems, because after
12 ... i.f5?! 13 l:tac1 j,d7 14 l:tfe1+ ~d8 15 the exchange of b-pawns the white knight
ttJg 5 with an attack - Palk6vi) 13 Mfe1+ (or will probably fmd itself in trouble (e.g. 25
13 lLld4 lLlge5 14 gacl ~d8 15 ttJxc2 dxc2 ttJa5 .i.d8 26 ttJxb7? .i.c7 or 25 ttJxe7 ~xe~
16 .i.xe5 lLlxe5 17 .uxc2 a6 18 j"e2 .uc8 and 26 lLla5 lLlxb2 27 lLlxb7? lLlc4); also Black
White has insufficient compensation) has a much more active king.
13 ... ~d7 14 .i.c4 lLlce5 15 lLlxe5+ dxe5 16 17 lLlxd4 Wd7 18 J:!.d 1 l:tc8 19 i.g3
.i.xe5 ttJxe5 17 l:i.xe5 lld8! and Black is Now White has full compensation for the
clearly better. pawn due to his lead in development and
12lLlb3 d3 pressure on d6.
12... .i.xc2?! is risky now because of 13 19 ... g6

96
 4 d4 exd4: Introduction

If 19 ... a6 20 f4 ttJf7 21 ~f1 and White exd4 5 e5lDg4 60-0!?

keeps the pressure.
20 lDb5l:1:c5 21 lDc3
Temptation is resisted. If 21 ttJxa7? l:!.a5
22 f4 ttJg4! (controlling the f2-square; not
22 ... ttJf7?! 23 ~f2) and now after 23 h3 .l:!.xa7
24 hxg4 .l:!.xa2 25 f5 .J:!.g8 26 .l:!.ed2 gxfS 27
~xd6 .1i.xd6 28 .l:!.xd6+ ~c8 Black has win-
ning chances in the endgame.
21 ... ~e7 22 f4 lDc4!
Black plays really well here. If instead
22 ... ttJc6?! 23 ttJd5 .id8 24 ..if2 .l:!.b5 25
.l:!.ed2 White would have a strong initiative
for his pawn.
23 ~f2lDxb2! A very mean move (in the gunfighter
sense). White has already gambited one pawn
and definitely hopes to get an advantage in
the near future through a direct assault.
6 ... d6!
6... ttJgxe5?? 7 ctJxe5 ctJxe5 8 .l:!.el costs a
piece.
7 exd6
White has no real worthwhile alternative.
If 7 e6? fxe6 8 ctJg5 ctJge5 9 .1i.xe6 h6 10
..ixc8 ~xc8 11 ~h5+ g6 12 ~h3 'iHd7 13
ttJe6 ctJd8 14 ttJxd4 'i'xh3 15 gxh3 .1i.g7 and
Black was much better in Borisov-
Kuznetsov, USSR 1961. Equally hopeless are
Black sacrifices the exchange for two 7 l:!.el? ctJgxe5 8 ctJxd4 ..ie7 and 7 ..ig5?
pawns and strong positional compensation. ..ie7 8 ..ixe7 ~xe7 9 exd6 "iVxd6 and Black
24 l:tde1 l:txc3 25 ~d4 ~d8 26 i..xc3 has an extra pawn.
ctJd3 27 l:!.f1 i..b6+ 28 ~h1 .!:tc8 29 i..b2 7 ... i.xd6
f5 30 g3 CDxb2 7.. :~xd6 has one drawback: White can
After this the game is drawn. My impres- play 8 ttJa3! a6 9 h3 ctJf6 10 ctJg5 ctJd8 11
sion is that after 30 .. J::tc4! followed by ... ctJc5- l:!.e1+ .1i.e7 12 "iVe2 ctJe6 and then 13 f4! with
e4 Black has winning chances. the advantage in Sax-Grochakov, Groningen
31 l:I:xb2l:!.c4 32 :d1 Y2-Y, 1971/72.
Black might be microscopically better 8 J:.e1 + ~f8 9lDa3 'i'f6
here, but the result will always be a draw, so Not 9... ttJceS? 10 h3! ttJxf3+ 11 ~xf3 ttJf6
{he players agreed to it immediately. 12 ttJb5 and Black is in difficulties.
10 i..g5!?
Game 42 This gives Black a chance to go wrong.
Plachetka-Smejkal After 10 ~e2 ..id7 11 ..ig5 ~g6 12 ctJb5
Ostrava 1994 ttJxh2 13 ctJxd6 ctJxf3+ 14 ~xf3 cxd6 15
..if4 h5! 16 ..ixd6+ 'it'g8 Black is slightly bet-
, e4 e5 2 CDf3 CDc6 3 ~c4 CDf6 4 d4 ter in Maciejewski-Sodor, Poland 1976.

97
Two Knights Defence

14... lDxf3+
After this White has the advantage, but if
14.. J::tb8 15 'iYxd4! etJxf3+ 16 gxf3 ~h2+ 17
'.t>f1 'iYxc7 18 fxg4 h5 19 Me5! was danger-
ous, e.g. 19 ... .ixg4 (not 19 ... hxg4? 20 .ixf7!
'.t>xf7 21 'iVf4+ ~g8 22 Me8+ and wins) 20
i..xf7! .l:i.h6 21 i..b3 and White is better,
though Black still has fighting chances.
15 ~xf3 'iWh2+ 16 Wf1 lDe3+ 17 fxe3
'Wixc7 18 i.d5 h5 19 exd4 i.g4 20 'Wia3+
Wg8 21 ~g1?!
This allows Black counterplay. Simpler
was 21 c3l:!.d8 22 SLb3 ~h2 23 ~e7 .l:i.f8 24
10 ... i.xh2+?! 'iYe5 and White is much better.
Stronger was 1O .. .'iVg6! 11 etJb5 etJxh2 12 21 ... ~d7 22 c4 i.e6 23 i.xe6 'iWxd4+
etJxd6 etJxf3+ 13 'iVxf3 cxd6 14 ~f4 i..d7 15 24 iVe3 ~xe3+ 25 J:.xe3 fxe6 26 .!:l:xe6
~xd6+ '.t>g8 with a similar position to that ':'c8 27 b3 ~f7 28 J:te5 Wf6?!
after 10 'iVe2, albeit a tempo down (... h7-h5) After 28 ... Ike8 the position is equal.
for Black. 29J:.d5
11 lDxh2 ~xg5 12 lDf3 'iVf4
PilkCivi considers 12... ~f4 to be a mistake
and that 12...'iVc5?! gives Black a clear advan-
tage. I disagree! After 13 ~d2! etJge5? 14
etJxe5 etJxe5 15 'iV e2 f6 16 'iVh5 g6 17 'iib6+
'.t>e8 18 .l:i.ad1 White has a strong attack, e.g.
18 ... ~g4 19 'iVh4! ~e7 20 .l:i.xe5 fxe5 21
'iVxg4 with a clear advantage. 13 ... ~f5 14
'iVf4 etJh6 15 'iVxc7 is also good for White,
while if 13 ... f6 14 Madl ~f5 15 'iVf4! White
has a strong initiative (not 15 etJxd4?! ~xd4
16 'iVxd4 etJxd4 17 .l:.xd4 when White has no
more than compensation for the material).
13lDb5lDce514lDxc7 29 .. Jlhd8?
Was this a sacrifice to gain counterplay? If
so, it was a great illusion. Black should have
used the other rook, i.e. 29 ....l:i.cd8 30 .l:i.fl+
'.t>g6 and the position is still about equal.
30 .!:l:xh5 l:.d2 31 l:rh3 J:.e8 32J:tf3+
32 l:!.fl+ was stronger. It seems likely that
both players were very short of time here.
~g6 33 J:.g3+ ~h6 34 ':'f3 g5 35 l:!.f6+
'it>h5 36 I:rf5? J:.g8? 1-0
36 ....l:i.ee2 would have put Black back in
the game. Presumably he lost on time while
making his move.

98
 4 d4 exd4: Introduction

Summary
After 4 d4 exd4 5 ctJg5?! is only dangerous for White. Apart from 5 0-0 (which can be found in
Chapter 8), the normal way to complicate the position is with 5 e5. Nevertheless, Black should
be able to find equality in all lines. Here I have tried to draw your attention to the attractions of
5...ctJe4 and 5... ctJg4, which both promise good play, but are less well known than 5... d5.

1 e4 e5 2 CDf3 CDc6 3 ~c4 CDf6 4 d4 exd4 (D) 5 e5

5 ctJg5 - Game 37
5 ... CDg4
5... ctJe4 (D)
6 ~e2 - Game 38
6 i.d5 - Game 39
6 WVe2
60-0 - Game 42
6 .. :~e7 7 ..tf4 (D)
7... f6 - Game 40
7... d6- Game 41

4 ... exd4 5 ... CDe4 7..tf4

99
 CHAPTER SEVEN I
4 d4 exd4 5 e5 d5

1 e4 eS 2 liJf3 liJc6 3 .ltc4 liJf6 4 d4 6 ...ctJd7?! is weak: 7 0-0 1i.e7 8 1i.xc6 bxc6 9
exd4 S eS dS ctJxd4 ctJb8 10 ctJc3 (or 10 f4!? with good
5... d5 is the most natural and also the attacking chances) 10... c5 11 ctJdb5 c6 12
most popular reply to 5 e5, and then only 6 ctJd6+ 1i.xd6 13 exd6 0-0 (if 13. ..'iVxd6?! 14
1i.b5 ctJe4 7 ctJxd4 makes sense of the posi- ~e1+ ii.e6 15 ctJe4 'iVe7 16 1i.g5 with a
tion. The game often revolves around a fight strong initiative) 14 .Jif4 and White had the
for control of c5; Black will try to gain coun- advantage in Barczay-Smejkal, Raach 1969.
terplay on the kingside while White starts 7liJxd4
building his nest there.

Game 43
Rogers-Wong Chee Chung
Singapore 1998
1 e4 eS 2 liJf3 liJc6 3 d4 exd4 4 .ltc4
liJf6 S eS dS 6 i.bS liJe4

7 ... i.cS!?
An interesting move order. Now 8 i.e3
1i.d7 9 1i.xc6 bxc6 10 0-0 (see Game 44)
reaches the same position as after 7... .lfLd7 8
1i.xc6 bxc6 9 0-0 1i.c5 10 .lfLe3, but Black has
avoided lines with 10 f3 ctJg5. Since Game 44
is perfectly OK for Black, a critical question
is whether White can achieve an advantage
There is not really a choice about this. by other means; in particular 10 ctJd2 (Games

100
 4 d4 exd4 5 e5 d5

45 & 46) or else by accepting the offered walk blindfold through a minefield and dis-
sacrifice on c6 (see below). arm twenty bombs on the way. Theoretically
8 0-0 it can be done, but in practice you need to
8 l2Jxc6!? is obviously a critical response, have prepared your defence at home before-
but White comes under a dangerous attack. hand.
The position may be defensible but White
should have both prior knowledge and iron
nerves. After 8... i..xf2+ 9 'it>fl ~h4 we have
the following possibilities:

cl) 15 g4 i..g6 (or 15 ...i..xg4!? 16 llfl

'iVc5 17 i..xc6+ bxc6 18 ~el! i..xf3 19 ~xf3
'iVc4+ 20 ~e3 ~xc2 and the black attack is
strong) 16 l2Jd4 cxb5 17 l2Jf5 i..xf5 18 gxf5
a) 10 'iVxd5? i.b6 11 l2Jd4+ c6 12 g3 'ifxf5 19 ktfl 'ilVxe5 20 'iVe2 0-0 21 c3 b4!
'iYh3+ 13 'it>el 'iVg2 14 i..xc6+ bxc6 15 and Black has very good compensation for
'iYxc6+ ~f8 16 ktfl i..xd4 17 'iYxa8 i..f2+ 18 the piece, if nothing more perhaps.
~e2l2Jxg3+ 19 hxg3 'iixa8 and Black wins. c2) 15 l2Jd4 i..g6 16 J::tfl l2Jd2+ (this se-
b) 10 l2Jc3 l2Jxc3! (the hardest - and I am cures a draw by perpetual; possibly better is
a hard man!) 11 bxc3 bxc6 12 i..xc6+ ~f8 13 16 ... 'ilVxg2! 17 <Jo>e3 cxb5 which looks very
'ilVe2 i..c5 (Medvegdy-.Mihalincic, Hungary dangerous for White, though it is not clear at
1993) and the game, though unclear, is a all) 17 'it>c3 'ilVe3+ 18 i..d3 l2Je4+ 19 ~b3
pleasure to play for Black. Also interesting, if l2Jc5+ 20 'it>c3 l2Ja4+ 21 <Jo>b3 l2Jc5+ with a
slightly risky is 1O ... 0-0!? 11 l2Jxe4 dxe4 12 g3 draw. Note that 21...'ifxd4?! lets the king
i..xg3 13 l2Je7+ ~h8! (not 13 ... 'iVxe7?! 14 escape by 22 a3! l2Jc5+ 23 Wa2 i..xd3 24
hxg3 'iVxe5 15 'ifh5 'iixh5 16 lIxh5 ii.e6 17 cxd3 l2Jxd3 25 l2Jd2 and White is at least
ii.f4 as the three pawns is not equal to the slightly better.
bishop here) 14 l2Jxc8 s'axc8 15 i..g5 'ilVxg5 One final intriguing possibility for White
16 hxg3 'ilVxe5 17 'ilVh5 'iVxh5 18 ~xh5 f5 is 9 ~e2!?, which it seems no one has ever
and Black is probably OK. tried - probably because allowing 9...i..g4+
c) 10 l2Jd4+ c6 11 l2Jf3 l2Jg3+ 12 ~xf2 looks crazy. But after 10 ~fl! i..xdl? 11
l2Je4+ 13 'it>e3 'ilVf2+ 14 ~d3 i..f5. Theoreti- l2Jxd8+ 'it>xd8 12l2Jc3! l2Jxc3 13 i..g5+ White
cians have treated this position with terrible comes out a piece up, while if 9...'ifh4?! (as
laziness. They write one after another that after 9 ~fl) 10 'ilVxd5! a6 11 l2Jd8+ axb5 12
Black has a decisive advantage. Is this true? 'ilVxf7+ 'it>xd8 13 l::td1+ l2Jd6 14 exd6 i..g4+
In my opinion, Black has sufficient compen- 15 ~fl and White is clearly better. Black
sation for the heavy material investments, but should perhaps opt for 9...i.g4+ 10 <Jo>fl and
nothing more. However, White will have to then 10.. :iVd7! 11 e6! (forced) l1...i..xe6 12

101
Two Knights Defence

ctJd4 c6 when the weak position of the white Chiburdanidze-Ma.Tseitlin, Moscow 1989)
king, together with the two central pawns l1...fxe5 12 J::txfS+ "iVxfS 13 c3! with some
and lead in development guarantees Black chances of saving the position (whereas after
long-term counterplay. 13 i..e3? exd4 14 i..xd4 i..g4 Black just
8 ... 0-0! wins).
b) 10 ctJc3?! ctJxc3 11 bxc3 f6 12 i..f4 fxe5
13 i..xe5 'iVd7! 14 "iVd2 J::tf7 15 f4 i..a6 and
Black stood better in Novikov-Sulskis,
Koszalin 1997.
c) 10 i..e3 "iVe8! 11 ctJd2 (if 11 f3?! ctJd6!
or 11 c3 f6! 12 exf6 J::txf6 and the black at-
tack looks murderous) 11 ... ctJxd2 12 'i¥xd2
i..b6! and Black is already slightly better, e.g.
13 c3 (13 f4 c5 14 ctJb3 d4 15 ..Itf2 i..b7 and
'X'hite is weak on the light squares) 13... c5 14
ctJb3 c4 15 ctJd4 "iVxe5 and Black was just a
pawn up in An. Gonzalez-Rossi, De la Roja
Cup 2003.
This interesting pawn sacrifice is the point 10 ... ~d7
of Black's move order, and guarantees an After 10 ...'iYh4 11 i..e3 i..a6 12 g3! (12
even game thanks the tempo saved omitting J::te1? ctJxf2 13 "iVd2 ctJg4 clearly favours
... i..d7. Instead 8... i..d7 9 ..Itxc6 bxc6 would Black) 12... "iVh3 13 i..xc5! i..xfl 14 "iVxfl
transpose to the 7... i..d7 lines in Game 47. "iVxf1+ 15 ~xfl ctJxc5 16 ctJe7+ ~h8 17
9~xc6 ctJxd5 and White has the better ending ac-
Black is quite safe after 9 ctJxc6 bxc6 10 cording to Rogers.
i..xc6 i..a6! but maybe 'X'hite is not! For 11 lDd4 '¥Ye7 12 ~f4 f6 13 ~e3!
instance, if 11 i..xa8? ..Itxfl 12 ~xfl (not 12 This surprising retreat is virtually the only
i..e3? i..xe3 13 fxe3 i..xg2 and wins) move. If 13 e6 (or 13 ctJb3? i..xf2+! 14 J::txf2
12... "iVh4 and Black is much better. So 'X'hite ctJxf2) 13. ..i..xe6 14 ctJxe6 "iVxe6 15 i..xc7
must play 11 "iVxd5 i..xfl 12 "iVxe4 i..b5! 13 ctJxf2! 16 .l::!.xf2 i..xf2+ 17 ~xf2 .l:!.ac8 18 i..a5
ctJc3 i..xc6 14 "iVxc6 ii.d4 15 i..f4 and then "iVf5+ 19 ~g1 .l::!.xc2 and Black stands better
15 ....l::!.b8! (a very natural new move from according to Piikovi.
Piikovi; after 15 ... i..xc3 16 'ii'xc3 'X'hite was
slightly better in few games) 16 .l::!.b1 "iVh4
(my addition to Pilkovi's analysis; 16..."iVe8
17 "iVxc7 'Yi'e6 18 Ii.d1 lUc8 19 "iVd6 i..xc3
20 bxc3 .l::!.b2 21 a3 Ihc2 is equal according
to Pilkovi) 17 ..Itg3 "iVg5 and 'X'hite has
problems even maintaining equality.
9 ... bxc6 10 lDxc6
It seems risky to accept the pawn, but
though 'X'hite has some alternatives here, he
has none that maintain equality.
a) 10 f3?! is answered by 10 .. .£6! 11 fxe4 (if 11
exf6?! "iVxf6 12 i..e3 i..a6 13 .l::!.e1 .l::!.ae8 14 c3
i..d6 and Black's attack was unstoppable in 13 .. .fxe5?!

102
 4 d4 exd4 5 e5 d5

This leads by force to a bad endgame. The 23 ... ..Itc8 24l:tab1 ..Itt5 25 t3!
alternatives were: There is no need for White to defend the
a) 13. .. i..xd4 14 'iixd4 ~xeS 1S Md1 useless c2-pawn.
~xd4 16 ilxd4 ..te6 and I do not see any 25 ... ..Itxc2 26 l:Ixb8 l:.xb8 27 .l:.e1
reason why White should be better here.
b) 13 ... ..tb7!? 'with compensation' is rec-
ommended in ECO. For example: 14 exf6
Mxf6! (not 14.. :iVxf6? 1SctJd2 ..td6 16ctJxe4
dxe4 17 'YWg4 when Black must fare without
compensation) 1S .l:.e1 (1S ctJd2? ctJxf2 16
~xf2 J:txf2 17 ~xf2 ..txd4 and Black is much
better) 1S ... Maf8 16 ctJc3 and the game is
unclear; but not 16 ctJd2? Mxf2 17 ~xf2
ctJxf2 18 .l:!.xe7 ctJxd 1 19 .uxd 1 iLxe 7 and
Black has the two bishops and a clear advan-
tage.
14lLlc6 ~d6 15 ~xc5! ~xc5 16 "tiVxd5+
"tiVxd5 17 lLle 7 + 'it>h8 18 lLlxd5 27 ... .l:.b1?
Usually rook and bishop work well to-
gether so, without this exchange, Black has
real drawing chances. Better was 27 ... ~g8!
(centralising the king) 28 .l:.xeS .l:!.b2 29 a4
~f7 30 .l:!.e2 .l:!.b1+ 31 ~f2 ~b3 32 ~g3 iLdS
and later, after hard work, Black might get a
draw. Then again he might not ... Life is so
brutal. Perhaps Black thought there would be
more drawing chances in the minor piece
ending, but if so, he was mistaken.
28 ':'xb1 ..Itxb1 29 a3 g6?
The last try at saving the game was
29 ... ~g8 30 ctJd7 e4 31 f4 ~f7! 32 ctJeS+
\'Vhite has a clear advantage; Black is a ~e6 33 ctJxc6 ~dS, but White can still de-
pawn down and those which remain are very cide the game himself by 34 ctJxa7 (not 34
weak. ctJeS? gS 3S g3 e3! with good counterplay)
18 ... ..Itb7! 34 ... e3 3S ~f1 i..d3+ 36 ~e1 ~c4 (if
The best chance. 36 ... ~e4 37 ctJc6! ~xf4 38 a4 and White
19lLlbc3 wins) 37 a4 ~xc3 38 as 'it'b4 39 ctJc6+ ~cS
Not 19 ctJxc 7? ~ac8 20 ctJbS Mxc2 and 40 ctJd8 ~c4 41 fS! and White should win
Black is better! after ctJe6.
19 ... c6 20 lLlc7! 30 ~t2 cJ;;g7 31 lLld7 e4 32 t4 ~t7 33
A weaker continuation is 20 ctJxe4 cxdS lLle5+ cJ;;e6 34 lLlxc6 ~d5 35 lLlxa7 ~c4
21 ctJcS ~c6 when Black has good drawing 36 'it>e3
chances. Black has no counterplay now.
20 ... lLlxc3 21 bxc3 .l:!.ab8 22 lLle6 kite8 36 ... 'ii;>xc3 37 lLlb5+ ~b3 38 g4 'it>c4 39
23lLlc5 lLlc7 h6 40 h4 ~c2 41 lLle6 cJ;;d5 42 lLlt8
This is the right place for the knight. g5 43 txg5 hxg5 44 hxg5 cJ;;e5 45 lLlg6+

103
Two Knights Defence

'it>d5 46 ct:le7+ 'it>e5 47 ct:lc6+ 'it>d6 48 pawn.

ct:ld4 i.d1 49 g6 rJ;;e7 50 ct:lf5+ 'it>f8 51 13 ~d2
rJ;;xe4 'it'g8 52 'it'e5 1-0 a) 13 f4?! is too optimistic; after 13 ...lLle6
14 c3 ..tb6 15lLlfS (or 1slLld2 f6! and Black
Game 44 was slightly better in Boyle-Wicknes, Scot-
Wendland-Grober land 1992) 15 ... ~d8 16 "iUg4 f6! 17 lLlh6+
Correspondence 1997 ~h8 18 fS fxeS (also interesting is the tactical
18 ... ~e7!? 19 .ixb6 axb6 20 fxe6 ~xe6 21
1 e4 e5 2 ct:lf3 ct:lf6 3 d4 exd4 4 i.c4 ~xe6 .ixe6 22 lLla3 gxh6 and Black is
ct:lc6 5 0-0 i.c5 6 e5 d5 7 i.b5 ct:le4 8 slightly better) 19 fxe6 WUf6 20 lLlfS .ixe6 21
ct:lxd4 i.d7 9 i.xc6 bxc6 10 i.e3 ~xg7+ "iUxg7 22lLlxg7 ~xg7 23 .ixb6 axb6
After a quite different move order we 24 .l::i.xeS ~f6 and Black has the better end-
reach the position which arises after 5 e5 d5 game.
6 .¥i.b5 lLle4 7 lLlxd4 .1i.c5 8 .ii.e3 .¥i.d7 9 b) 13 lLld2 .¥i.b6 14 a4lLle6 is slightly bet-
.1i.xc6 bxc6 10 0-0. ter for Black according to Gligoric, but after
10 ... 'iVe7! 15 lLl2b3 the position is in my opinion more
or less equal.
13 .. .f61

The best solution and one of the points of

the 7... .¥i.c5 move order. The idea is simple:
to be able to answer 11 f3 with 11 ... lLld6! as Black is ready to break down the white
queen pins the white e-pawn against the centre. 13. .. lLle6!? 14 lLlc3 .l::i.ab8 15 b3 with
bishop. Instead if 1O ... .ltb6 11 f3! lLlg5 12 an equal game is also OK; if instead
~d2 h6 13lLlb3 as 14 a4 ~e7 1slLlc3 and 13. .. .ixd4 14 .ixd4 .1i.f5!? (if 14... ctJe6 15
White had some advantage in Marzoll- .if2 f5 16 c4 White had a slight advantage in
Weiger, Germany 1997. Kupreichik-I.Zaitsev, USSR 1969) 15 lLla3
11 lIe1 lLle6 16 .¥i.f2 J:tab8 1711ab1 d4 and the posi-
After 11 f3 lLld6! 12 .ltf2 lLlfS 13 c3 0-0 tion looks rather unclear.
14 !tel .¥i.b6 15 ~c2 ~gS the position is 14 ct:lc3
equal. 14 ~h1?! is met by 14... h6! (threatening
11 ... 0-0 12 f3 ct:lg5 .. .f6xeS) 15 .ixg5 hxg5 and Black has a nice
By defending the bishop White has pre- game while after 14 c3 .l::i.ae8! it is difficult for
vented the knight's retreat to d6. The draw- White to develop his queenside.
back is that the white rook really belongs on 14... i.b6
f1 to support the further advance of the f- A prophylactic move; Black places the

104
 4 d4 exd4 5 e5 d5

bishop on a safe square in advance of any attacking the enemy king? How many of the
later tactical skinnishes. 14".h6? is now a enemy's pieces protect the king? Clearly the
waste of time after 15 ctJa4! ..ib6 16 ctJxb6 attacking forces are in the ascendancy.
axb6 17 i.xg5 hxg5 18 e6 and White has a 18 i..xb6 cxb6 19 ltJg3 ':'af8 20 ':'f 1
clear advantage, while if 14".fxe5? 15 i.xg5 i..xh3 21 ':'xf3 .l:!.xf3
..ixd4+ 16 'ii'xd4 ~xg5 17 l:!.xe5 and White
stands much better according to Pilk6vi.
15 lLlce2??
A grave blunder. Any of the following was
an improvement:
a) 15 l:!.adl ctJe6 16 exf6 ~xf6 17 ctJxe6
i.xe6 18 ctJa4 ~g6 with an unclear position
in Sokolsky-Shapovalov, corr. 1962/63.
b) 15 ctJa4!? (recommended by Pilk6vi)
15".ctJe6 16 i.f2 l:!.ae8 17 ctJxb6 axb6 18
..tg3 f5 and the position is more or less
equal.
c) 15 ..txg5 makes less sense, since after
15".fxg5 16 Whl ~4 17 ctJb3 i.f5 18 l:!.adl Black has been very successful. The white
l:!.ae8 Black's pieces are more active. knight on b3 only exists 'on paper'; it is not
15 ... lLlh3+!! taking part in the actual game.
22 £!.e1 ~f6 23 'iVe2
\X-rute might consider selling the rights of
this game to Hollywood as a catastrophe
fllm. After 23 ctJcl e4 24 c3 h5 there is no
defence against the move 25".h4 winning the
house.
23 ... e4

A surprising check, but this is a desperado

position in which, according to Lasker, every
piece will try to sell itself as dearly as possi-
ble.
16 gxh3
White has no choice.
16 ... fxe5 17 lLlb3
This, too, is the only move. 24ltJd2?
17 ... .:.xf3 This allows a deadly finish. 24 c3 was the
Some sacrifices does not need to be calcu- best try, but White will not hold.
lated and this is one of these cases. Instead 24 ... "iVd4+ 25 <;t>h1 ~f2 26 "iVe3 "iVxe3
you can think: How many of my pieces are 27 lhe3 ~xd2 0-1

105
Two Knights Defence

.ixc5 27 .Jl.xc5 h5 2S h3 a4 29 'iYd3 a3 30

Game 45 .Jl.d4 hxg4 31 hxg4 iVe6 32 'iYdl ~a6 1/2_1/2
Sveshnikov-Zaja Rybak-Stand, corr. 2000.
Bled 2001 13 .. :iYe7 14 lithe1 a5 15 a3 a4 16 ..tg5
'i'e6 17liJd4 ~g6 18 f4!?
1 e4 e5 2 tUf3 tUc6 3 ..tc4 tUf6 4 d4
exd4 5 e5 d5 6 ..tb5 tUe4 7 tUxd4 ..tc5 8
..te3 ..td7 9 ..txc6 bxc6 10 tUd2!

Very risky, and possibly not very sound.

18 .....ta5
If Black gets tempted by material a sur-
At the present moment this seems to be prise awaits him: lS ... .Jl.g4!? 19 h3! (of course
the most dangerous line. White planned to be aggressive; not 19 lbf3?
10 ... tUxd2 'iYh5 and Black is just better) 19 ... .ixd1 20 f5
This theoretical move is not enough for 'iYh5 21 lbxc6 h6! (necessary, since after
equalising. For other moves see Game 46. 21....Jl.xc2 22lbe7+ ~hS 23 f6 White's attack
11 ~xd2 0-0 is very strong) 22 ~e7 .l:!.feS 23 f6 (not 23
Or 11...'iYe7 12lbb3 .Jl.b6 13 'iYc3 0-0 14 'iYf4? .l:!.a6! as 24 .l:!.xdl then loses to
0-0 .l:!.feS 15 f4 f6 (if 15 ... 'iYe6 16 .l:!.ae1 'iYg6 24 ...'iYe2) 23 ... .ixc2 24 'iYxc2 'iYh4 25 l:.d1
17 .l:!.f3 .Jl.f5 IS I::t.g3 iVh6 19 I::t.g5 with a 'iYf4+ 26 ~b1 'iYe4 27 g4 and this position is
strong attack) 16 .Jl.c5 fxe5 17 fxe5 'iYg5 IS difficult to assess. Probably White has
.l:!.ael and White was slightly better in Kar- enough compensation for the exchange as it
patchev-Frolov, Tomsk 19S5. is hard to see how the black rooks can get
12 liJb3 ..tb6 13 O-O-O?! into the game at all.
A brave decision, but not the best. Even- 19 c3 c5 20 tUe2?
tually the white king will be exposed on the A grave error. White cannot afford to be
open b-ftle. Instead White should play for a passive with a black offensive coming on the
long term advantage with 13 O-O! and be safe queenside. Better was 20 f5! 'iYb6! (20 ....Jl.xf5
and wealthy. After 13 ... f6 14 exf6 (not 14 f4? 21 lbxf5 'iiVxf5 22 g4! gives White the initia-
fxe5 15 fxe5 'iYg5! and Black takes over the tive; e.g. 22 ...'iYxg4? 23 .l:!.gl 'iYh5 24 .ihG
initiative) 14...'iYxf6 15 lbc5 .Jl.f5 16 c3 .l:!.aeS \vins, while if 22 .. :iVe6 23 'iYxd5 or 22 ...'iYg6
17 .id4 'iYg6 IS f3 White is slightly better. 23 .Jl.e7 .l:!.feS 24 'iYxd5 and White is better)
Nevertheless, Black can hold the position, as 21 lbe2 (21 e6? is tactically flawed: 21...cxd4
illustrated after lS ....Jl.cS 19 b4 as 20 ~h 1 22 exd7 dxc3 23 bxc3 l::tabS and mate is
lle7 21 llael I::t.feS 22 .l:!.xe7 .l:!.xe7 23 .l:!.el dose) 21...~xf5 22 'iiVxd5 .l:i.abS 23 .l:!.d2 and
'iYeS 24 .l:i.xe7 'iYxe7 25 g4 'iYn 26 ~g2 the game is undear.

106
 4 d4 exd4 5 e5 d5

20 ... d4 21 ~e7 .l:!.e4.

There is no time for 21 h3 Mfb8 22 g4 30 ... iYc4+?
~e6 23 ~c2 i.b5 and the white position This check loses a tempo for no reason.
cracks. Instead 30 ...'iWa2! 31 'it'c2 ~c4+ 32 ~b1 c5
21..JUb8 22 ~xc5 dxc3 23 lLlxc3 ~g4 and wins.
24 h3 ~xd1 25l:l.xd1 ~xc3 26 'i!Vxc3 31 'iVc2 'i!Va2
Also possible was 31...1Wxf4+ 32 ~b1 ~f5
33 'it'el ~e6 and Black is much better, but
still it can be hard to win a position like this.
32 Wixc7?
The position is of course very difficult,
but this should lose outright. Better was 32
~e2 and 'X'hite can still fight.
32 ... .l:.d1 + 33 ~c2 ~b1 + 34 ~b3

Black has a multitude of advantages in this

position: \Xfhite has an exposed king, his
bishop has no scope whatsoever, and his
pawns are both weak and obstructing his
counterplay. Whereas Black has no weak-
nesses whatsoever, and furthermore is the
exchange up. Technically the game is over.
26 ... l:td8 27 l:te1 lad3 28 'i!Vc2 l:tad8 29
~b4 'i!Ve6 34 ... .l:t 1 d3+??
29 ... ~g3!? was also strong. A grave blunder. Black has done really
well and now throws it all away. Instead
34 ... ~d3+ wins quickly, e.g. 35 ~c3 ~d5+
36 'it'c2 ~xg2+ 37 'it'b3 .l:.1d3 or 35 'it'a4
Ma8+ 36 .i.a5 J::i.xe 1.
35 .rt..c3 l:.'txc3+
Black has no choice. If 35 ...iVxel 36
~xd8+ Mxd8 37 i.xe1 turns the tables.
36 'iVxc3 l:tb8+ 37 ~a4 .l:.a8+ 38 ~b5
.l:.b8+ ?!
A quick check in time trouble presumably.
Better was 38...~f5! 39 .l:!.c1 ~e6 40 ~c6
~b3+ 41 'it'c5 Ua5+ 42 'it'd4 .s.a4+ 43 'it'c5
Ma5+ 44 'it'd4 with perpetual check.
30 'i!Vxa4?! 39 ~a5 'iVa2?
Slightly better was 30 'it'b 1, though Black The queen is not really performing any
wins by 30 ... ~3 31 ~xb3 axb3 32 ltel Mg3 service to the black community from here
33 f5 Me3! 34 g4 Mxe5 35 Mxc7 l:te3 36 h4 (after \Xfhite prevents ... ~d5+). The correct

107
Two Knights Defence

move was again 39 ... 'it'f5' (not 39 ... na8+? 40 the chess games are played with faster and
'>t>b6 nb8+ 41 '>t>c7 and wins) 40 'it'c6 'it'xf4 faster time controls, it becomes increasingly
and Black has enough counterplay to draw, difficult to defend positions like this.
maybe even some chances for an advantage
if White is not careful.
40 l:td 1 h5 41 J:.d6 1:txb2 42 e6 J:.c2?
A final mistake. Black could still have of-
fered some resistance with 42 ..J::tb8!? 43 e7
na8+ 44 Wb6 'it'b1+ 45 'iWb4 ~g1+ 46 nd4
'>t>h 7 though White wins eventually.
43 l:td8+ ~h7 44 "ti'd3+ 9645 exf7 1-0

Game 46
Kristensen-Hebden
Kopavogur 1994

1 e4 e5 2 tLlf3 tLlc6 3 d4 exd4 4 i.c4 11 tLlxe4

tLlf6 5 e5 d5 6 i.b5 tLle4 7 tLlxd4 iLc5 8 White has two serious alternatives:
iLe3 i.d7 a) 11 0-0 ii.b6 12 c3 0-0 13 f3 ttJxd2 14
There is no sense in exchanging the ir'xd2 c5 15 ttJe2 and the position is more or
bishop for only one tempo. After 8... .ixd4?! less equal.
9 ir'xd4 (9 ii.xd4 0-0 10 .ixc6 bxc6 11 ttJd2 b) 11 ttJ4f3 and then:
is also good) 9... 0-0 10 .ixc6 bxc6 11 ttJc3 bl) 11...~e7 12 .ixc5 ttJxc5 13 0-0 nb8!
ttJg5 12 .ixg5 ~xg5 13 f4 i¥h4+ 14 g3 'it'h5 (White has problems with defending the
15 0-0 and White was better in Sveshnikov- pawn on b2 without putting his pieces in an
Balashov, Elista 1997. unnatural position) 14 ttJb3 ttJe6 15 'YWd3 0-0
9 iLxc6 bxc6 10 tLld2 'ii'h4!? and despite appearances to the contrary
This looks the strongest. Other moves are: Black's position is good; for example, if 16
a) 1O ... ttJxd2 was examined in Game 45. c4?! ttJf4! and Black takes over the initiative.
b) 10... ttJg5 11 c3! (weak is 11 ttJxc6? b2) 11.....txe3?!? is exciting. Objectively
.ixc6 12 .ixc5 d4! with a very strong initia- Black does not have enough compensation
tive for the pawn) 11....ib6 12 f4 ttJe6 13 0-0 for the queen but it is an interesting bluff and
g6 14 '>t>hl 0-0 15 'it'el (threatening 16 f5) calculation includes severe psychological
15 ... ttJg7 16 b4 and \'V'hite was better in shock. Realising the material superiority is
Tzermianos-Pavlovic, Agios 1995. not at all easy, and in the game several inac-
c) 1O...'it'e7?! is strongly met by 11 lbxe4 curate white moves will give Black the initia-
dxe4 12 e6! fxe6 (not 12... .ixe6? 13 ttJxe6 tive. It is possible that this sacrifice also
.ixe3 14 ttJxg7+ '>t>f8 15 ttJf5 'iVb4+ 16 c3 benefits from the human tendency towards
and wins) 13 ttJxc6! (better than 13 ir'h5+ g6 giving gifts: Black gives a queen, what will
14 ~e5 a-a-a!? when White is better but the White give? Kotronias-Barbero, Budapest
position is very messy) 13 ... .ib4+ 14 ttJxb4 1988, continued 12 ttJxh4 .ixd2+ 13 '>t>f1
~xb4+ 15 ~d2 ~xd2+ (15 ...'it'xb2?! 16 0-0 ii.a5 14 ~e2? (even in a dream two bishops
gives White a terrible attack; it is hard to say are not a match for the queen in this posi-
that he has compensation for the pawn here tion, so White is returning the generosity -
- the pawn does not matter), 16 '>t>xd2 with a after this move Black wins the exchange and
good endgame for White. In these days when the position becomes more unclear; instead

108
 4 d4 exd4 5 e5 d5

14 ctJf3 was correct and I cannot see how here. The problem is in transferring the rook
Black can develop an initiative after this sim- to the kingside as White's minor pieces are in
ple move, e.g. 14... .lib6 15 ctJd4 and White is the way. Better is 15 ctJb3 and position is
close to winning) 14...g5 15 ctJf3 .lib6 16 c3 unclear.
ctJxf2 17 ctJd4 ctJxh 1 18 'iitg 1 0-0-0 19 b4 1S ... 0-0 16 .i.e 1 ?!
.l:!.de8 20 Wxh 1 and White was better after all. Consistent, since without this the rook
11 .. :iVxe4 12 0-0 .i.b6 faces unemployment. But the bishop is
A strong prophylactic move; Black moves needed to fight for the c5-square. One of the
his bishop out of danger and the way of his hardest things in chess, and in life, is to admit
c-pawn. After 12... 0-0?! 13 .l:te1! ~g6 14 ctJe6 one's own stupidity and correct one's mis-
i..xe6 15 .ixc5 White has a big advantage takes. Here it would have been better to for-
and his bishop is much more active. get about the rook manoeuvre and play 16
13J:.e1 'iVg6 ctJb3!, and after 16 ....lifS 17 .lixb6 cxb6 (not
17 ... .ixc2? 18 'iYd2 cxb6 19 ctJd4 .ie4 20
l::!.g3 and Black has problems) 18 ctJd4 the
position is roughly equal.
16 ... .i.g4 17 ~d2 eS 18 tLlbS 'iVe6 19
':g3 .i.fS

14 a4
14 ctJb3 0-0 15 .ic5 .l:!.fe8 16 I:!.e3 as 17 a4
.lif5 is unclear. Instead 16 a4!? is interesting
from a practical point of view. Wrute pre-
pares two pawn sacrifices, in exchange for
which he seize the initiative and condemn 20 h4?
the opponent to a passive defence: 16... .if5 Steinitz turned in his grave when White
(16 ... a5!? keeps the tension) 17 l:tc1 l:tab8 18 played this! The right to attack comes with
lIe3 .lixc5 (a brave decision, but Black has having the better position, and here Black is
already decided to do this with 17. ..l:tab8). 19 better! The pawn on h4 is just another weak-
ctJxc5 ':xb2 20 ~d4 l::!.xc2 21 ~xc2 .ixc2 22 ness and does little good for White. Better
h3 and White had a strong initiative in was 20 b3 when White is worse but far from
Doghri-Matsuo, Yerevan 1996. Aaron lost.
Nimzowitsch (second only to Wilhelm 20 ... 'it>h8
Steinitz in the history of chess theoreticians) A useful move. Black does not want to
would be proud seeing this position. His risk being at the wrong end of the stick on
thoughts about the effectiveness of the the g-file.
blockade in chess is still very much relevant! 21 'YIHd1?!
14 ... as 1 S .l:!.a3? Making way for the bishop, but it is the
Usually moves like these are good, but not wrong diagonal. 21 b3 and .lib2 was better.

109
Two Knights Defence

21 " .J:!.ae8 22 .Itf4 and ~c3) 15 tbf3 (or 15 tbe2 d4 16 .if2 0-0
Even now 22 b3 looks better. 17 c4 - Sveshnikov) 15 ... d4 16 .if2 .ltc6 17
22 ... h6 23 e3 .lth7 24 'ilVd2 J:le7 25 lite1 iLh4 'iHd7 with an unclear position.
f6! b) 9... iLe7?! is even more passive. After 10
It is time to open the position. f3 tbg5 11 f4 tbe4 12 f5 c5 13 tbe2 iLb5 14
26 e4 d4 27 exf6 l:txf6 28 litb3 'iWg4 29 tba3 iLc6 15 c4 d4 16 tbf4 iLg5 17 tbd3
.i.g3 l:te2 30 'ii'd1 l:tfe6 31 ':f3 J:tb8 18 'iVe2 h6 19 iLf4 llb6 20 ]:tae1 iLa8
Or 31 tbxc7 .Jtc2 and wins. 21 'iVg4 White had a clear advantage in
31 ... i.e2! 0-1 Sveshnikov-Fercec, Nova Gorica 1996.
c) 9...g6 is risky. After 10 f3 tbc5 11 f4
Game 47 tbe6!? (provocative; but if 11...iLg7 12 b4!
Sveshnikov-Zaitsev and White has a stable advantage) 12 f5!?
Podolsk 1992 tbxd413 ~xd4 iLxf5 14 e6 f6 15 tbc3 White
had fIne compensation in Shipov-Sointsev,
1 e4 e5 2 tbf3 tbc6 3 .i.c4 tbf6 4 d4 Moscow 1964.
exd4 5 e5 d5 6 i.b5 tbe4 7 tbxd4 i.d7 d) 9... c5?! 10 tbb3 c6 (or 10 ... iLc6? 11 f3
This move is slightly passive and reduces tbg5 12 tba5 and White is better) 11 c4 d4
Black's possibilities. Since it is quite possible (or 11...dxc4 12 tb3d2) 12 f4 and White has a
to play the active 7... .Jtc5 (as we have seen clear advantage.
Games 43-46), Black should probably do so. 10 .i.e3 i.e 7
8 i.xc6
The only serious move. After 8 tbxc6?!
bxc6 9 .Jtd3 i.c5! Black has the advantage as
10 .Jtxe4 is answered by 1O ... ~4! and Black
obtains the bishop pair.
8 ... bxc69 0-0

11 tbd2?!
More accurate is 11 tbb3! 0-0 12 tbld2
and White is slightly better - PaIkovi.
11 ... tbxd2 12 'iVxd2 c5 13 tbf3 'iVe4 14
l:!.fe1
In a position like this it is always good to
9 ... 'iWh4!? try to fInd some possibility for forcing the
The only independent move to justify opponent's king to stay in the centre. Here.
playing 7... .Jtd7. Other moves are: however, it does not work. 14 iLg5?! .Jtxg5
a) 9... iLc5 10 iLe3 transposes to Game 44, 15 tbxg5 ~d4! and Black has a fme position.
but White can also play 10 f3 tbg5 11 f4 tbe4 with potential for an advantage.
12 .lte3 .ib6 13 tbd2 tbxd2 14 ~xd2 c5 14... i.h3!
(otherwise White takes control of c5 by tbb3 It is always useful to make a mess of the

110
 4 d4 exd4 5 e5 d5

enemy camp. 26 .. :iib6?! 27 l:!.dl d4 (27...~b8?? 28 .l::!.d3

15 ~f4 'iHg6 16 ~g5? c4 29 'iHe7 c6 30 l:!.a3 and White wins) 28
A mistake which passes by unpunished. 'iVa8+ 'iHb8 29 iVxb8+ ~xb8 30 b3 and
Necessary was 16 Jig3 Jie6 17 Jih4 .ixh4 White has winning chances.
18 ttJxh4 ~5 19 ttJf3 0-0 where the posi- 27 'iHxb4 cxb4
tion is about equaL Now the position is different. White can
activate his rook immediately.
28 a3 bxa3
Or 28 ....l:.e8 29 axb4 .l::!.xe5 30 ~f1 Wd7
and the position is more or less equaL
29 l:lxa3 ~d7 30 J:!.g3 f6 31 h4 fxe5 32
hxg5 .l:!.b8 33 b3 J:!.b4!
It is important to prevent White from cre-
ating a passed pawn with h4.

16 ... 0-0-0?
Stronger was 16 ... h6! 17 gxh3 (forced; 17
'iixd5? 'ud8 18 ~3 Jie6 and Black wins)
17 ... Jixg5 18 'iWd3 (very risky would be 18
'iixd5!? .id2+ 19 'It>h 1 .l:.d8 20 'iixc5 Jixe 1
21 .l::!.xel ~6 22 'iia3 with the idea of
22.. :~Vxf2 23 e6 with play, as Black replies
22 ....l::!.d7! whereafter he is better) 18...'iVxd3
19 cxd3 'ub8 and Black is slightly better. 34 h3
17 gxh3 h6 18 ~a5! Look at this. Two strong grandmaster
This is the point! After having castled have a drawn position and they continue to
queenside Black has problems with his king. fight! Why? Because they know that mistakes
18 ... ~xg5 19 ct:Jxg5 hxg5 20 Wixa7 'iHc6 are human.
21 .l:!.e3 .l:th4 22 J:!.a3 J:!.b4 23 lIa6 .l:!.b6 24 34 ... ~e6 35 J:rc3 ~d6 36 l:lf3 ~e6 37
J:!.xb6 'iHxb6 25 ~a8+ ~g2 c5 38 g6 c4 39 .i:l:f7 cxb3 40 cxb3
The transition to the endgame is a very ':xb341 h4
important moment. After 25 'iVxb6? cxb6 41 .l::!.xg7 1:tb8' 42 h4 'It>f6 43 .l::!.d7 'It>xg6
Black has a much better rook ending, despite and the draw is near.
the pawn deficit, because of the white pawn 41.. .J:!.b8 42 ~f3 1:.h8 43 ~g4 d4 44
weaknesses. J:!.xg7 ~f6 45 1:.a7 ~xg6 46 h5+ ~h6 47
25 ... ~b8 26 ~a3 Wib4 I!.a6+ ~h7 48 J:!.a7 + Y:z -Y:z

111
Two Knights Defence

Summary
4 d4 exd4 S eS dS is a good but perhaps too well-travelled road. After 6 i..bS ctJe4 7 ctJxd4
i..cs accepting the sacrifice on c6 - either immediately or following 8 0-0 0-0 - seems to give
Black sufficient counterplay. The only way for White to fight for an advantage is in the line 8
~e3 i..d7 9 i..c6 bxc6 10 ctJd2!, when Black should be very careful. Now 10... ~h4!? is the best
chance for equality, as 10 ... ctJxd2 11 'iVxd2 allows White a small edge.

1 e4 e5 2 tiJf3 tiJc6 3 i.c4 tiJf6 4 d4 exd4 5 e5 d5 6 i.b5 tiJe4 7 tiJxd4 i.c5 (D)
7... ~d7 8 ~xc6 bxc6 9 0-0 (D) ~h4 - Game 47
8 i.e3
80-0 - Game 43
8 ... i.d7 9 i.xc6 bxc6 10 tiJd2 (D)
10 ... iVe7- Game 44
10 ... ctJxd2 - Game 45
10 ... 'iVh4 - Game 46

7... i.c5 9 0-0 10 tiJd2

112
 CHAPTER fiGHT I
4 d4 exd4 5 0-0

1 e4 e5 2 tLlf3 tLlc6 3 .1£.c4 liJf6 4 d4 in the rest of the chapter.

exd4 5 0-0
If you are a grandmaster 5 0-0 is close to
being a small mistake! Why? Well, after cas-
tling White has no real chances of fighting
for an advantage. In fact, there is a real
chance White will emerge with a slighdy
worse position. The game takes on a very
forcing nature and Black stands well here.
The safest and most popular response is
5... tUxe4! which is the subject of Games 48-
54. Grandmasters favour this move enor-
mously (75% of Grandmasters, who ex-
pressed a preference, said ...).
The alternative, 5... Ji,c5 (Games 55-57) is 7 ~d5!
known as the Max Lange Attack. It was very Also good is 7 tUxd4 d5 (not 7... i.c5? 8
popular about 100-150 years ago. However, l:txe4! fxe4 9 "ifh5+ and wins) 8 i.b5 and if
after it became well known that 5... tUxe4 8... i.d6 (hoping for 9 f3 'iUh4 or 9 tUxc6
gives Black an equal game, the Max Lange i.xh2+ 10 1itf1 'iWh4 with counterplay, or if 9
Attack was seen seldomly in tournament h3 O-O! 10 tUxc6 bxc6 11 Ji,xc6 ~h2+ 12
play. This is a shame as Black has some nice Iitxh2 ~d6+) 9 g3! prevents all threats and
ideas in this line. White wins material.
7 ... ..tb4
Game 48 7... tUe7 is met with 8 'iUxd4 tUxd5 9 ~xd5
Ellner-Andruss ~e7 10 tUgS! (a very energetic move; not 10
Correspondence 1977 'iUxf5? d5 and Black is OK) 10...i.xg5 11
- - - - - - - - - - - - - - -..... l:txe4+! Ji,e7 (if 11...fxe4 12 Ji,xg5 wins) 12
1 e4 e5 2 tLlf3 liJc6 3 ..tc4 tLlf6 4 d4 l:te1 c6 13 ~xf5 d5 14 'iVf3 and White has a
exd4 5 0-0 tLlxe4 6 l:!e 1 f5? clear advantage as the black king is too ex-
The only move is 6... d5 which is examined posed. Wolfe-Lheureux, email 2001, contin-

113
Two Knights Defence

ued 14.. :iVd6 15 ~g5 ..Iie6 16 ~xe7 'ii'xe7

17 CDd2 O-O-O? (but 17 ... lIfS!? 18 'ii'h5+ g6 19 Game 49
'ii'e5 is also good for White, e.g. 19 ... ~f6 20 Struik-Mikhalchishin
CDb3 0-0-0 21 CDc5 b6 22 CDxe6 ~e8 23 Zwolle 2003
'ii'c7+! 'ii'xc7 24 CDxc7 J::!.xe1+ 25 ~xel Wxc7
26 J::!.e7+ Wd6 27 J::!.xh7 .l:le6 28 Wf1 and 1 e4 eS 2 .l¥..c4 lLlf6 3 d4 exd4 4 lLlf3
White has serious winning chances) 18 iVe3 lLlc6 S 0-0 lLlxe4! 6 lire1 dS 7 .l¥..xdS
~he8 19 'iNxa7 ~f6 20 CDb3 i.d7 21 c4 dxc4 This is the only serious move. 7 CDxd4? is
22 CDc5 'iYxb2 23 .l:leb 1 1-0. bad for many reasons, one of them being
8.liLgS! 7... CDxd4 (another is 7... i.e7!? 8 ..Iib5 i.d7 9
A powerful move that gives Black a lot of i.xc6 bxc6 10 f3 CDd6 11 ..Iig5 f6 and White
problems. does not have compensation for the pawn) 8
8".lLle7? 'iYxd4 i.e6 9 i.xd5 'iVxd5 10 ~xe4 'iYxe4 11
After this the game is lost. 8...i.xel! is the ~xe4 0-0-0 and Black clearly is better.
best chance, but Black is still in trouble after 7 CDc3?! is seen occasionally, and is 'better
9 i.xd8 ..Iixf2+ 10 Wf1 'Ot>xd8 11 i.xe4 (sim- than its reputation' according to the great
plest; instead 11 CDbd2 CDf6 12 i.xc6 CDg4! is Russian theoretician Yakov Estrin. In my
not completely clear) 11...fxe4 12 Wxf2 exf3 opinion the bad reputation is justified, as
13 iVxf3 d6 14 ~t7 ~e8 15 iVxg7 ..Iid7 16 after this move White can no longer keep
~xh 7 and White should win. equality, but has to fight to stay alive:

9.:axe4! a) 7... dxc4 8 ~xe4+ i.e7 9 CDxd4 f5 and

This sacrifice ensures that the few white then:
pieces already developed get free play against al) 10 ..Iih6? (though not very important,
the black king. the following line is nice) l1...fxe4 11 ..Iixg7
9".fxe4 10 lLleS ~f8 .l:i.fS! (in ECO they only give I1...Wt7 when
R.Adams-Roe, email 1999, finished differ- the position is unclear) 12 'iVh5+ lIt7 13 J::!.dl
encly: 10...l:!.fS 11 'iVh5+ g6 12 iVxh7 CDxd5 (if 13 CDxc6 bxc6 14l:!.dl ..Iid7 15 'iNxh7 i.fS
13 ~xg6+ ~t7 14 'iVxt7 mate. 16 iVxe4+ 'iNe7 and at the end of the day, a
11 .liLxe4 rook is a rook; White should lose) 13... ..Iid7
And, unsurprisingly, there is no defence 14 CDxc6 bxc6 15 CDxe4 ~b8 16 iVxh7 ..tfS
against iVf3+. 17 'iNg6 i.xg7 18 CDg5 'i¥e719 CDxt7 WfS 20
11...d6 12 'i!ff3+ .liLfS 13 .liLxfslLlxfS 14 CDg5 llb5 and Black wins.
'i!fxfS+ 1-0 a2) 10 ~f4! 0-0 (also possible is 10 ... ..tg5

114
 4 d4 exd4 5 0-0

11 'iVe2+ 'WIe7 12 'WIxe7+ 'Dxe7 13 'Ddb5 10 ... .td6

~xf4 14 i.xf4 ~f7 15 'Dxc7 l::tbS 16 'D7b5 The main alternative, 10....i.b4, is consid-
l::taS 17 'Dc7 with a draw by repetition) 11 ered in Game 50. Also seen is 10 ... h6 11 .i.f6
'Dxc6 'iVxd1+ 12 'Dxdl bxc6 13 l::txc4 c5 and then:
where both players have their chances, even
though Black might have an easier game,
practically speaking.
b) 7... dxc3! S ~xd5 ~e6 (if S... f5?! 9 'Dg5!
with a strong attack according to Tarta-
kower) 9 ~xe4 'iVxd1 10 l::txdl cxb2 11
~xb2 f6 and Black is slightly better. White's
lead in development is only temporary (Black
will complete the mobilisation of his forces
in two moves), and while White can regain
the pawn, structurally he has some problems;
e.g. 12 'Dd4 (if 12l::te1 ~f7 13 .i.c3 .i.c5 14
l::tabl i.b6 15 a4 as with a clear advantage to
Black) 12...'Dxd4 13 ~xd4 0-0-0 (not a) 11...'i¥d5? 12 c3 d3 13 'Dd4 'Dxd4 14
13 ... c6?! 14 ab 1 and White has the initia- cxd4 and Black IS under terrible pressure.
tive) 14 .i.xa7 .i.d6 and Black is to be pre- After moves like 14.. :i~b5 15 'Dc3 'iic4 16
ferred. d5 gxf6 17 dxe6 fxe6 1S 'iih5+ ~dS 19 'i¥f7
7 .. :~xd5 8 lLlc3 White has a winning attack.
b) 11...'WIa5?! 12 'Dxd4 'Dxd4 (if 12...gxf6
13 'Dxf6+ ~e 7 14 b4! 'Dxb4 15 'Dxe6 and
White wins; one line is 15 ... ~xf6 16 'iid4+
~g6 17 'WIxhS fxe6 1S 'iVgS+ J.g7 19 ':'xe6+
~f5 20 'i¥f7+ ~g5 21 'iVg6+ ~h4 22 l::te4
mate) 13 'i¥xd4 c5! (the only move; again if
13...gxf6 14 'Dxf6+ cJi;e 7 15 'Dd5+ and White
wins) 14 'i¥c3 'i¥xc3 15 i.xc3 and White is
somewhat better as Black cannot easily de-
velop his kingside.
c) 11...'iig6! (the only move) 12 'Dh4 'i¥h7
13 'WIh5 (after 13 f4.i.e7 14 i.xe7 'Dxe7 15
'i¥xd4 0-0 Black does not appear to be any
8 .. :~h5! worse) 13 ...J.b4 (or 13 ... ~d7!? 14 l:tad1 l:teS
This is a generally underestimated move. 15 i.xd4 cJi;cs 16 'iib5 a6 17 'WIa4 i.e7 1S
The queen looks very natural on the kingside 'Df3 with an unclear position in Helvenstein-
in this position. The reason for this move's Ye Rongguang, Netherlands 1996; the black
lack of support is that S.. :iVa5 (Games 51-54) queen is oddly placed on h7, but it will soon
equalises easily and holds no traps or tactical return to the centre, and White has no obvi-
pitfalls and potholes. ous ways to attack the black king immedi-
9 lLlxe4 .lie6 10 .lig5 ately) 14 c3 dxc3 15 bxc3 ~a3 16 'Df5 'WIxf5!
If 10 'Deg5 0-0-0 11 'Dxe6 fxe6 12 l::txe6 17 'Dd6+ i.xd6 IS 'i¥xf5 gxf6 19 'i¥xf6l::th7
.i.d6 13 .i.d2 l::theS and Black is clearly bet- 20 l::tadl l::tdS with a very unclear situation.
ter; ... g5-g4 is a tertible threat. Usually three minor pieces should favour

115
Two Knights Defence

Black, but here his co-ordination is not too White wins.

good. Maybe White should just play 21 ~f5 11 ... cxd6
Itg7 22 ~f6 ~g8 23 ~xe6+ fxe6 24 ~xe6+
4Je7 25 ~xh6 when he has all the passed
pawns to work with.

12 ~f4 "iVc5!
ECO gives only 12...iYd5 13 c3 ':c8 (or
13. .. 'it>d7 14 iYa4 b5 15 ~a6 .l:.hb8 16 4Jxd4
11 lDxd6+ 4Jxd4 17 cxd4 g5 18 SLd2 h5 with unclear
11 SLf6? is answered by the strong play in Wirschell-Hector, Berlin 1993) 14
11...SLxh2+!' It is interesting that this move 4Jxd4 4Jxd4 15 'ii'xd4 iYxd4 16 cxd4 'it>d7
was introduced as a novelty in Chess Infonnant with equality. The text is better since the
#68 (from 1997). In the real world this move black queen now will not be hanging after
has been well known since the beginning of ... d4xc3.
the 90's! Van Wely-Van de Oudeweetering, 13 c3 dxc3
Rotterdam 1990, continued 12 4Jxh2 (not 12
'it>f1? SLc4+ and it is all over) 12... ~xdl 13
.l:!.axd 1 gxf6 14 4Jxf6+ 'it'fS 15 4Jf3 .l:!.d8 16
a3 h5 17 .l:!.d2 .l:th6 and Black was simply a
pawn up.
White can also try 11 c4!? 0-0 (the most
natural response) 12 c5 ~e5 13 4Jxe5 ~xdl
14 Itaxdl 4Jxe5 15 .l:!.xd4 f6 16 ~f4 SLxa2 17
4Jc3 SLt7 and the position was roughly equal
in Sveshnikov-Bezgodov, St. Petersburg
1994. 11....i.b4 is probably more or less even
too. But not 11...dxc3? 12 4Jxd6+ cxd6 13
~xd6 and White has a decisive attack. To
prevent 14 .l:!.adl (and 15 ~d7+) Black has to 14.l:!.c1?
try 13. .. c2 (if 13 ... cxb2 14 .l:!.abl! and the This is a grave error. White hopes that the
threat is simply .l:!.xb2 and l:!.xb7, which Black pin will allow him to get some initiative, but
can do nothing about; e.g. 14... h6 15 .l:!.xb2 in reality he only loses the b2-pawn - as well
hxg5 16 J::txb 7 ktc8 and then 17 ktxe6+ fxe6 as the initiative. The correct move was 14
18 iYd7+ leads to mate) 14 ktxe6+ fxe6 15 SLxd6 iYa5 15 ~c2! (White needs to play
Mel! and Black has no defence; if 15 ... c1~ energetically; if 15 bxc3 0-0-0 Black is at least
16 .l:!.xc1 'it>t7 17 'ii'd7+ 'it>g8 18 ~xb7 and slightly better, as White has big problems

116
 4 d4 exd4 5 0-0

with the bishop on d6; e.g. 16 c4 .i.xc4 17 tLlf6 5 0-0 tLlxe4 6 'u'e1 d5 7 .i.xd5 'iYxd5
l:tc1 ~ d5! 18 ttJe5 .:!.xd6 19 ttJxc4 'iVxd 1 20 8 tLlc3 ~h5 9 tLlxe4 .i.e6 10 .i.g5 .i.b4!?
llexd1 l:txdH 21 l:txd1 .l:!.d8 and Black wins!)
15 ... 0-0-0 16 .i.e5 ttJxe5 17 Sxe5 .i.d5 18
.l:!.ae1 ~b8 19 bxc3 and the position is un-
clear. Black has good long term prospects,
but White has some initiative which may be
sufficient to grant him even chances.
14 .. :iff5! 15 .i.xd6?
This allows Black to create a nasty pin. In-
stead, after 15 .tg3 cxb2 16 l:tb 1 0-0 17
.u.xb2 IHe8 18l:txb7 i.d5 Black has an extra
pawn, but White has some fighting chances.
15 ... cxb2 16 J::1.c2 0-0-0 17 l:lxb2 .i.d5!
18 tLlh4?
Preferable was 18 i..g3 .i.xf3 19 'iVxf3 Black wants to see blood!
'ii'xf3 20 gxf3. Of course this is lost, particu- 11 c3
larly against such a strong player as l'vfikhal- This is too ambitious. Calm play is to be
chishin, but White is still alive for the time preferred here; i.e. 11 ttJxd4 'ii'xdl 12 lIexdl
being. ttJxd4 13 l:txd4 .i.e 7 and then:
18 .. .'ilVf6! a) 14 l:te 1 l:td8 15 l:txd8+ (if 15 l:ta4 a6 16
.i.xe7 ~xe7 17 ttJc5 .l:!.d2! and Black has
sufficient counterplay) 15 ... ~xd8 16 l:td H
~e8 17 i..e3 f5 18 ttJc5 .i.xc5 19 .i.xc5 \vith
equality - or if you are Anatoly Karpov, with
a slightly better position for White.
b) 14 i..xe 7 cJ;;xe 7 15 ttJc5 l:tad8 16 ttJxe6
fxe6 was Van der Tuuk-Piket, Netherlands
1993; the position is equal, but not a draw -
as Piket proved by winning this game.
11.. .dxc3 12 bxc3 .i.a5 13 h4
White has also tried 13 'iWc1 0-0 14 ttJg3
'ii'g6 15 ttJh4 ~d3 16 ttJe4 cJ;;h8 17 'ii'f4 and
White's compensation is only of a practical
Now White has no way to avoid losing nature; objectively the position is good for
material - a truly spectacular position where Black. For example, 17 ... .txc3! 18l:tadl 'iVc4
most of the white pieces are hanging loosely 19 ttJxc3 'ii'xc3 20 ~xc7 Sab8 (stronger than
around the board. 20 ... 1:!.ae8 as in Kamsky-Kupreichik, Palma
19 'i¥xd5 'YIVxb2 20 tLlf5 l:!.he8 21 J:i.d1 de Mallorca 1989, and although he was still
J:!.e6 22 'i¥c5 l:.e5 0-1 better for a long time Black eventually man-
aged to lose this game) 21 .l:te3 'iVc2 22 I!.de1
Came 50 'iVxa2 23 'ii'g3 ~d5 24 I!.d3 'iYc4 and after
Hoogervorst-Simmelink overcoming some technical problems Black
Correspondence 1988 should win.
13 ... ~g4 14 ~b1
1 e4 e5 2 tLlf3 tLlc6 3 d4 exd4 4 .i.c4 White does not have an easy life here:

117
Two Knights Defence

a) 14 tZJg3 ..l1..b6 15 .i:tb 1 h6! (suggested by tZJxh4 and Black is much better.
Palk6vi) 16 l:te4 ~xg3 17 l:!.xb6 ~ d6 18 l:td4 22 ... i..b6
axb6 19 .l:txd6 cxd6 20 i..e3 0-0 Black is 22 ...':f6 23 .l:txd5 ~b 1+ 24 ~g2 .l:td6
much better. looked interesting; the main idea is 25 l:!.xa5
b) 14 'iYa4 0-015 tZJh2 ~f5 16 .l:tab1 ..l1..b6 tZJd4 26 'iYxa7 tZJxe2 but then 27 'iUa8+ 'iVd8
17 .l:tb5 ~g6 18 h5 "iVxh5 19 i..e7 'iYxb5!? 28 'iYxe4 gives White an excellent game!
(or 19...~6 20 ..l1..xf8 ':xfS and Black is bet- 23 J:!.xd5 'ii'e6 24'tIHa3 ttJe7
ter - Palk6vi) 20 ~xb5 tDxe7 21 tZJg5 l:.ad8 Not 24... tZJe5 25 l':i.xe5 'iUxe5 26 c5 and
22 ~e2 .l:td6 and Black is much better. After White is back in the game.
a probable exchange on e6 Black gets more 25 J:.dd2 ttJg6?!
than his share of the action on the f-flie. 25 ... c5! seems much stronger, blocking
14 ... 0-0! White's only possible counterplay.
Some poor sou1s have tried 14... ..l1..b6?? 26 c5 lLlxh4 27 cxb6 lLlf3+ 28 ~g2
and after 15 tZJh2! Black loses at least a piece; ttJxh2
e.g. 15 .. :~f5 16 tZJd6+! cxd6 17 ,*,xfS. Better looks 28 ... ~e5 29 tZJf1 tZJxd2 30
15 'ilVxb7 i..d5 .l:txd2 axb6 31 tZJg 3 .l::.bd8 32 'iVb3+ ~h8 33
'Vib4 'iUe6 34 ,*,xe4 'iUxe4+ 35 tZJxe4 l:!.xd2
36 i..xd2 .l:ta8 and Black has more chances in
this endgame.
29 bxc7?
After 29 ~xh2 'iVxg4 30 'iVe7 White is so
active that Black has nothing better than
30 ...'iVh5+ 31 ~g2 'iVf3+ 32 ~g1 'iVg4+ 33
~f1 ~3+ with a draw.
29 ... 'ii'xg4+ 30 ~xh2 l:!.b1 31 f4

Material equilibrium has been restored,

but white pieces are not ready for an even
fight.
16 ttJh2 ~f5 17 g4 'i!id7 18 J:Iad 1 'u'ab8
19 'ilVa6 f6
Also interesting is 19... f5!? 20 tZJc5 '*'f7 21
~d3 i..e4! (the point) 22 tZJxe4 fxe4 23 ~c2
h6 24 i..e3 'iVc4 and Black stands better,
since the white king's position is very fragile.
20 i..e3 f5 21 ne2! 31 ... J:!.b5?
The only move. If 21 tZJc5 'iUf7 22 ~d3 In this fantastically complicated position
i..xa2 23 tZJd7 .l:tbd8 24 i..g5 ~b3 25 i..xd8 Black makes a mistake. Instead, if 31...g5 32
(if 25 .l:td2 i..c4! wins) 25 ... .l:txd8 26 .l:td2 i..e6 ,*,xf8+!! (not 32 c8~? .l:txc8 33 fxg5 .l:tfS and
27 .l:txe6 ~xe6 and Black is much better. there is no hiding for the white king)
21 .. .fxe4 22 c4 32 ... ~xfS 33 fxg5 ~f1 34 .l:tf2+ l:!.xf2+ 35
If 22 .l:ted2 ~e6 23 ~xd5 i..b6 24 'iVe2 .l:txf2+ ~e7 36 .l:tc2 and Black must take the
tZJe5! 25 ~xb6 .l:txb6 26 ttd8 tZJf3+ 27 ~h1 draw by 36 ... ~h4+ 37 ~g2 'iVg4+ 38 ~h2

118
 4 d4 exd4 5 0-0

with a draw. 9 t21xe4

The best option is 31 ... exf3! 32 .l::!.d4 'WIh5+ There are no serious alternatives:
33 ~g3 'WIe5+ 34 l:i.f4 l::tg1+ 35 ~h2 (not 35 a) 9 .l::!.xe4+?! i..e6 10 It'lxd4 0-0-0 leaves
~f2? l:!.g2+ 36 ~f1 fxe2+ 37 ~xg2 elN+ 38 White struggling to equalise after 11 i..e3
~g3 'i¥g5+ 39 ~f2 .l::!.xf4+ 40 i..xf4 .l::!.xf4+ 41 It'lxd4 12 .l::!.xd4 i..b4 (12 ...i..a3!? also looks
~xel 'WIxc7 and wins) 35 ... .l::!.xf4 36 c8'W1+ good) 13 It'le4 .l::!.xd4 14 'iVxd4 .l::!.d8 Black is
l:i.f8+ 37 ~xgl .l::!.xc8 38 'iVb3+ ~h8 39 l::tb2 better, because of the tactical trick 15 'iVxg7
h6 and with the white king so exposed, Black 'iVxa2!.
can fight for a win at no risk. b) 9 It'lxd4? It'lxd4 10 'i¥xd4 f5 11 i..h6!?
32 iLc5? (not 11 It'lxe4?? 'i¥xel mate or 11 f3?? i..c5,
Here White misses the chance to make a while if 11 i..g 5 ~f7! 12 It'lxe4 fxe4 wins, or
fantastic draw: 32 'iVxf8+! ~xf8 33 f5!! block- 11 i..d2 'iVc5 12 'iYa4+ 'it'f7! 13lt'lxe4 fxe4 14
ing the black queen's defence of c8 (since if .l::!.xe4 'i¥c6 and Black kept the piece in
33 ...'i¥xf5?? 34 l:!.f2 wins), so Black must Volkov-Mazurenko, USSR 1955) l1...i..d7!
again take the perpetual check by 33 ... 'i¥h4+ (not now 11...~f7? 12 It'lxe4 fxe4 13 .l::!.xe4
34 ~g2 'iVg4+ etc. 'iVb6 14 l::tf4+ ~g6 15 'i¥e5 1-0 Cappello-
32 .. :iVxf4+ 33 'iiVg3 'YiVxg3+ 34 'it>xg3 Lucidi, corr. 1953; after 15 ... ~xh6 16 .l::!.h4+
J:txc5 ~g6 17 ~5+ ~f6 18 .l::!.f4+ 'it'e 7 19 .l::!.e 1+
Black is simply two pawns up - and in a i..e6 20 'iVf7+ 'it'd8 21 ':xe6 i..d6 22 'WIxg7
correspondence game this is equivalent to wins) 12 It'lxe4 O-O-O! 13 i..g5 (if 13 i..d2
being a queen up in a normal game. 'WIa4!) 13 .. .fxe4 14 i..xd8 ii.c5 15 "ilixg7 .l::!.xd8
35 1:1.d7 ':c3+ 36 '.th4 e3 37 ~g3 h6 38 and Black was clearly better in Blomquist-
'.tg2 '.th7 39 J:[d8 1:1.g8 40 lbg8 '.txg8 41 Muir, corr. 1991. This is an important lesson:
J:txe3 J:txc 7 42 J:ta3 ~f7 43 l:t.a6 g6 0-1 rather than going for material gains, Black
defends through developing his pieces and
Game 51 this way ends up with two bishops and better
Ojurhuus-Blees co-ordination.
Gausdal1993

1 e4 e5 2 t21f3 t21c6 3 iLc4 t21f6 4 d4

exd4 5 0-0 t21xe4 6 f!.e1 d5 7 .1I..xd5
'iiVxd5 8 t21c3 'YIVa5

9 ... .1I..e6
9... i..e7?! is an old mistake, answered by
10 i..g5! and then:
a) 1O ... i..e6 11 i..xe7 It'lxe7 12 It'leg5! (12
"ilixd4 0-0 13 ~c5 It'lc6 14 It'le5 'WIxc5 15
This is the most popular route to equality. It'lxc5 It'lxe5 16 .l:i.xe5 is only equal) 12... 0-0

119
Two Knights Defence

13 CDxe6 fxe6 14 'iVxd4 and White is better 1943) 13. .. .i.e7! (better than 13 ... dxc3 14
because of the weak pawn on e6. bxc3 as the opening of the flies is in White's
b) 1O ... f6?! is risky because of 11 CDxf6+! favour) 14 ~g3 O-O-O! when White is under
gxf6 12 ~xf6 .i:!.f8 (12 ... J:tg8 13 .t!.xe7+ CDxe7 pressure, as illustrated by 15 CDxd4 'iVb6 16
14 'iVe2 'iVb4 15 Mel and the attack against .l:!.d2 CDxd4 17 cxd4 f5 18 d5 fxe4 19 dxe6 e3!
the black king cannot be parried, as there are 20 .l:!.xd8+ .l:!.xd8 21 'iVf3 .l:!.d3! and White is
no pieces who can come to his rescue) 13 truly struggling to keep his position together.
.i.xe7 CDxe7 14 ~xd4 and White has fantas- 10 ... 0-0-0 11 tLlxe6 fxe6 12 ':'xe6
tic compensation for the piece.

This is a critical position for the system.

Now Black has: Now Black has four moves: 12... h6 (as in the
a) 14... ~f5 15 .l:!.e5 'iVb6 16 .l:!.ael .l:!.f7 17 game), 12...'iVf5 (usually transposing to
'iVxb6! (not 17 Mxe 7+? Mxe 7 18 'ifh8+ ~f7 12... h6 - see below), 12 ... ~e7!? (see the next
19 .l:!.xe 7+ ~xe 7 20 'iVxa8 'iVxb2 and the po- note), and the probable best move 12....i.d6
sition is very unclear) 17 ... axb6 18 CDg5 .l:!.g7 (Game 52).
19 .l:!.xf5 and White wins. 12 ... h6
b) 14 .. :iVb6 15 'iVe5 ~f6 16 ~xc7 .l:!.g8 (if This move prevents ..Itc1-g5 and prepares
16 ... ~g4 17 CDe5! 'iVxf2+ 18 ~hl and White an attack on the white king with ...g7-g5. But
wins) 17 lIe3 .i.h3 18 g3 ~f8 19 .l:!.ael .l:!.g7 it is somewhat slow. White now has real
20 ~xb7 with a clear advantage to White, chances for getting an advantage.
Gayson-Howell, British Championship 1989. A better method is 12... .i.e7!? which is
c) 14....l:!.xf3! (Black needs to get rid of the another living inheritance from Akiba Rubin-
attacking white pieces in order to survive) 15 stein. Keres once said that if a recommenda-
gxf3 ~g5+ 16 r,i>hl ~f7 17 'iVc4+! CDd5 18 tion was given by Rubinstein or Alekhine,
~e4 'iVf5 19 .l:!.adl 'iVxe4 20 fxe4 CDf4 21 you can always play it. Black wants to con-
.tld8 and White is clearly better. tinue ... .tf6 and defend the d4-pawn without
10 tLleg5 creating a weakness (. .. h7-h6 and ...g7-g5) on
10 .i.g 5? is met strongly by 10 ... h6 11 the kingside. It is a very safe plan, and
.i.h4 .i.b4! (now the black king has the f8- though 12 ... .te7 is not very popular at the
square, which is very useful in many varia- moment, you can use it to receive a nice and
tions) 12 Me2 g5 13 c3 (if 13 CDf6+? ~e7! 14 original game; e.g. 13 CDe5 (if 13 'ife2 .tf6 14
a3 .i.d6 15 b4 'iff5 wins, while after 13 a3 .tf4 'iVf5 15 .i.g3 d3 16 cxd3 'iVxd3 with
.i.e7 14 b4 'iVd5 15 ~g3 0-0-0 Black was a complete equality in Kabanov-Sofronie,
clear pawn up in Medina-Keres, Madrid Techrighiol 1998) 13 ...CDxe5 14 l:!.xe7 .l:!.d7

120
 4 d4 exd4 5 0-0

(or 14.. .l:lde8!? 15 ..td2 lIVb6 16 lIxg7 "iWf6 opmion Black is better after 16 h3, e.g.
17 lIg3 ZlhfS with compensation for the 16 ...lIhg8 17 Zlel h5 18 "iWd3 g4 19 li'lxd4
pawn) 15 lIxd7li'lxd7 16 ..tf41Ie8 and Black li'lxd4 20 lIxd4 'iVf6 with full compensation
was at least equal in Blauert-Van der Sterren, for the pawn, and his position is easy to play.
Groningen 1989. As in the main game, with h2-h3 White cre-
13 'iVe2 i.d6 ates a weakness in his own camp - without
The position after 13 .. :iWf5 often arises af- Black putting any pressure on him to do so!
ter 12.. :~'f5 13 "iWe2 h6. The idea is simply b2) 16 lIel .l:IhfS 17 a3 ~b8 18 b4 and
that White cannot play 14 ..td2 because of White is slightly better; e.g. 18 ...g4 19 li'lh4
14... "iWxc2. Instead after 14 lIe4 g5 15 ..td2 'iVf6 20 l:i.xg4 ~g8 21 l:!.xg8 ..txh2+ 22 ~hl!
iLg7 (15 ...lIh7!? looks like an interesting new .l:!.xg8 23 'iVe6 and the weakness of h6 will
idea; Black can play either ... lIe7 or ...1In on probably decide the game in White's favour.
the next move, both of which seem fully 14 ... 'iVf5 15 a3
satisfactory) 16 .l:le 1 ..tf6 17 h3 (this looks Black has the initiative after 15 lIe4 g5! 16
risky, but it works!) 17 ... h5 18 h4! g4 19 CLlg5 ii.d21Ihg8.
and White is slightly better. 15 ... ~d7!?
Black wants an endgame. 15 ... g5 was also
interesting, and if White plays 16 b4 Black
can follow 16...'iitd7 17 l:!.e4 lIde8 18 lIxe8
lIxe8 19 'iVdl .l:!.e4 transposing to the game.
16 l:.e4 '!:'de8 17 :'xe8 .u.xe8 18 'iVd 1
Forced. After 18 ~d3 'iVxd3 19 cxd3
li'la5! (the weakness of b3 is here exploited to
the maximum) 20 b4 (if 20 iLd2 li'lb3 21
lIdl c5 and Black has a big advantage; he will
play ... ~c6-d5 and then ... b7-b5 and ... c5-c4
with strong pressure on the queenside)
20 ... li'lb3 21 l:i.bl li'lxcl 22 lIxc1 1:Ie2! 23
li'lxd4 .l:ld2 24 li'lb5 lIxd3 and the endgame
14 h3?! is very uncomfortable for White.
White simply loses time with this move - 18 ... l:l:e4 19 b4 g5
and when ...g5-g4 comes, it comes more Possible was 19 ... a6!? with unclear play.
strongly. The best move was 14 ..td2! when 20 i.b2 g4 21 hxg4 ':'xg4 22 'iYd3
Black has:
a) 14.. :iVh5!? 15 'iVe4! looks better for
White. At least he should not fall for 15 lie I?
d3! (a typical tactic in this line) 16 'iVe4 (16
~xd3?? ..txh2+ and 16 cxd3?? li'ld4 both win
for Black) 16... dxc2 and Black is slightly bet-
ter.
b) 14...'iVf5 15 lIe4 (if 15 'iVe4 "iWxe4 16
lIxe4 .l:lhe8 17 lIae 1 lIxe4 18 lIxe4 ..te 7
followed by ... iLf6 and Black has a well ten-
able position) 15 ...g5 and then:
bl) 16 h3?! is given by Palkovi with the as-
sessment that White is slightly better. In my

121
Two Knights Defence

22 ... iYf6? tLlf6 5 0-0 tLlxe4 6 .l:!:.e1 d5 7 i..xd5 ~xd5

Black overestimates the power of his own 8 tLlc3 ~a5 9 tLlxe4 .te6 10 tLleg5 0-0-0
attack and underestimates the weakness of 11 tLlxe6 fxe6 12 tlxe6 i..d6
the d-pawn. Instead, after 22 ...'fixd3 23 cxd3
b5 24l:!.e1 h5! Black is completely OK in the
endgame.
23 ne1 ~g7 24 g3 ~d8

The most popular, and probably best

move.
13 ~e2
The alternative is 13 .tg5 Mde8 14 'iVe2
25 J:te4! ~d7 15 Mxe8 (15 Me4 is also equal)
By exchanging rooks White eliminates all 15 ... Mxe8 16 'iVd3 h6 17 .td2 'iVh5 18 g3
Black's counterplay. 'iVg4 19 l:tel .l:tf8 20 ~g2 'iVf5 with equality
25 ... tLle5 in Berend-Potapov, Pardubice 1997. Note
After 25 ....iVg6 26 llxg4 'iVxg4 27 CLlxd4 that 15 Mel? is a mistake because of
CLle5 28 'fif5 White simply has a pawn more. 15 ...'iVxe1+! 16 CLlxe1 .ilxe6 17 'fig4 Mhe8
25 ...Mg6 26 CLlxd4 (or 26 b5) does not im- and Black is clearly better.
prove anything either. An important -litcie sideline is 14 'iVe1!?,
26 tLlxe5 .txe5 27 i..xd4 .txd4 28 .l::t.xg4 which leads to a draw after 14...'iVxe1+ 15
'iYxg4 29 c3 ~e7 30 ~xd4 ~xd4 31 .l:taxel Mxe6 16 Mxe6 ~d7 17 Me4 Me8 18
cxd4 .l:i.xe8 ~xe8 19 ~f1 ~f7 20 .td2 h6 21 ~e2
A pawn ending with a pawn less is almost ~e6 22 ~d3 ~d5 23 CLlxd4 CLlxd4 24 c4+
always lost. 'iite6! (a very important move; for some rea-
31 ... >ite6 32 f4 ~f5 33 ~g2 ~e4 34 son John Emms missed this in his book PIC1)'
~h3 a5 the Open Games as Black, and was very con-
Or 34 ... ~xd4 35 ~g4 ~e4 36 ~h5 ~f5 cerned about this line for Black, since the
37 ~xh6 c6 38 ~h5 b6 39 g4+ ~f6 40 ~h6 alternative 24...'iite5? 25 f4+ 'iitf5 26 ~xd4
and wins. .txf4 27 .txf4 ~xf4 28 b4 is close to being
35 ~g4 axb4 36 axb4 ~xd4 37 ~h5 b5 lost) 25 ~xd4 .te5+ 26 ~e4 .txb2 and
38 f5 1-0 Black was no worse in Hacat-Hughey, Ed-
monton 2000.
Game 52 13 ... ~h5 14 'iYe4
Kamsky-Ye Rongguang The alternatives are worse:
Manila 1990 a) 14 .td2? d3! (a tactic revisited from the
notes to Game 51) 15 'iV e3 (once more 15
1 e4 e5 2 tLlf3 tLlc6 3 d4 exd4 4 .tc4 cxd3?? CLld4 or 15 'iVxd3?? .txh2+ wins)

122
 4 d4 exd4 5 0-0

15 ... dxc2 and Black is better. 'It>f2 with an equal endgame.

b) 14 iLg5? d3! (again) when 15 1I¥e3 can 20 ... b6 21 h3 a5 22 a4 "tlVd5 23 b3 .l:!:.e4
be met by several good moves; one is 24 "tlVf3 cJ;?b8 25 J:1f2 "tlVe6 26 cJ;?f1?
15 .. .':t:Jd4! 16 IIxd6 (forced) 16 ....l:i.xd6 17 This loses a pawn to a simple tactic. 26
4Jxd4 ~e8 18 'iVf4l:td5 19 ctJf3 d2 20 ttdl g4!? was a good move for active counterplay.
'iVxg5! 21 'iVxg5 .l:i.xg5 22 .l:i.xd2 .l:i.g4 and 26 ... ii.xf4!
Black has excellent winning chances. Since if 27 iLxf4?? l::te 1 mate.
c) 14 h3?! avoids the ... d4-d3 trick, but it 27 "tlVd3 g5 28 J:tf3 i.xd2 29 ~xd2 h6 30
loses time as well as weakening the kingside. 'iYf2 "tlVe5 31 J:tf8+ cJ;?a7 32 ~f3 J:te1 + 33
After 14....l:i.de8 15 iLd2 ctJe5 16 !:!'xe8+ .l:i.xe8 cJ;?f2 'iYe3+ 34 cJ;?g3 "tlVe5+
17 ctJxd4 'iVxe2 18 ctJxe2 ctJc4 19 iLe3 ctJxb2 Also possible was 34... ~xf3+ 35 .l:i.xf3
20 'It>f1 ctJc4 and Black is now very slighdy 'It>b 7! (you cannot win an ending without the
better. king) 36 l::tf6 ~e2 37 .l::.xh6 l::txc2 and Black
14... .l:!:.de8 15 ii.d2! lLJe5 16 J::!.xe8+ J:txe8 wms.
17 lLJxe5 ii.xe5 35 cJ;?g4
If instead 35 'It>f2 then 35 ...1I¥e3+ 36 'It>g3
'iUxf3+ etc.
35 .. .l::te4+ 36 cJ;?h5 g4+ 37 'iYf5 gxh3 38
gxh3 ~e7
The rook ending should also be winning,
but it is stronger for Black to continue the
attack.
39 ~f6 J::ie5+ 40 cJ;?h4 J::ie4+
A draw is not the agenda. Black repeats
the position simply in order to reach the time
control, after which he can work out the win
at leisure.
41 c,t>h5 J:te5+ 42 cJ;?h4 "tlVc5! 43 'iYf3
After this Black has a very small advan- .l:!.e4+ 44 c,t>g3 'ii'e5+ 45 cJ;?g2 'i'g5+ 46
tage. Instead 17 ....l:i.xe5 18 1wd3 1I¥e2 19 cJ;?f2 "tlVd2+ 47 cJ;?g3 "tlVe1 + 48 ~f2 J:l:e3+
'iVxe2 l:txe2 20 l::!.d1 'It>d7 21 'It>f1 is com- 49 >t>h4 "iVh1 50 'iYf1 'iYh2 51 :gS .l:!.e4+
pletely equal. 52 cJ;?h5 ~e5+ 53 cJ;?xh6 .l:!.h4+ 54 cJ;?g6
18 f4 ~g3+ 55 'it>f7 .l:I.h7+ 0-1
18 iLf4!? iLxf4 19 'iYxf4 'iVe5 20 'iUxe5
.l:i.xe5 21 'It>f1 is equal; Black has nothing Game 53
after 21....l:i.c5 22 .l:i.c1 d3?! 23 c4. But Bachler-Colias
18... ..ltf6! 19 'iVd3 'iVe2 puts White under USA 1991
some pressure.
18 ... ii.d6 19 ~d3 g6 20 lirf1?! 1 e4 e5 2 lLJf3 lLJc6 3 .tc4 lLlf6 4 d4
This is not really a mistake. White's mis- exd4 5 0-0 lLlxe4 6 .l:!.e1 d5 7 .txd5
take is his whole strategy here. He plays wait- ~xd5 SlLJc3 'iYa5 9lLJxe4 ..te6 10 ..td2!?
ing moves without any plan at all. This is one This is the modern attempt to squeeze
of the worst things you can do in chess. something out of the position. It is clearly
Even a bad plan is often preferable to playing more dangerous for Black than 10 ctJegS.
without any plan at all. Better was 20 b4 1I¥d5 Therefore it is important to prepare against
21 c4 'iVe4 22 'iVh3+ ~f5 23 1I¥xf5+ gxf5 24 it.

123
Two Knights Defence

~xe7 13 c4! ~xc4 (or 13. .. dxc3 14 ~c2 with

compensation) 14l:tc1 ~d5 15 l:tc5 'YWd7 16
'YWc1 with good attacking chances for the
pawn.
b3) 11...i.b4!? 12 c3 3i.a5 13 b4 3i.b6 14
a4 a6 15 i.f6! (better than 15 b5?! axb5 16
axb5 ctJa5 as in Fette-Steczkowski, Copenha-
gen 1985) 15 ... 0-016 as i.a7 17ctJxd4l:tfe8
18 ctJxc6 ~xc6 19 "iWh5! ~5! with an un-
clear position in Weber-Grzelak, corr. 1992.
But not 19 ... h6? 20 i.xg7! ~xg7 21 'iWe5+
~h7 22 ctJf6+ ~g6 23 h4 and White has a
strong attack.
10 .. :tlVa4!? c) 10... 3i.b4 (another solid equaliser) 11
This is a risky, but playable move. The ctJxd4 (if 11 c3!? dxc3 12 bxc3 3i.e7 13 c4
main line with 1O .. :iWf5! is seen in Game 54, ~a6 14 i.g5! l:td8 15 ~1 0-0 16 3i.xe7
while 10 ...'YWh5 11 i.g5 transposes to 8... ~h5 ctJxe 7 17 ctJc5 'iUxc4 18 ctJxe6 fxe6 19 'iUxb 7
9 ctJxe4 i.e6 10 i.g5 in Games 49 & 50. ctJd5 20 ~xa7 was level in Bielczyk-Panczyk,
Black has also tried: Polish Championship 1982) 11...ctJxd4 12 c3
a) 10 .. :tib6?! (misplacing the queen) 11 i.e7! (safest, though both 12... 0-0-0 and
i.g5 h6 12 ~h4 ~e7 (not 12...g5?! 13 ctJf6+ 12 ... 0-0 are also playable) 13 cxd4 ~d5 and
~d8 14 i.g3 with a dangerous attack in then:
A.Geller-Neishtadt, Leningrad 1956) 13
i.xe7 ~xe7 14 b4! and White has a strong
initiative. It will take a long time before Black
will get his king into safety; e.g. 14... .l::tad8 (or
14.. :iWxb4 15 J:i.b1 "iVc4 16 l:!.xb7 and White
has more than enough compensation for the
pawn in Tringov-Lilienthal, Sofia 1962) 15
ctJc5 l:the8 (or 15 ... ctJxb4 16 ctJxe6 fxe6 17
ctJe5 with excellent compensation) 16 ctJxe6
fxe6 17 'iWd3 ctJxb4 18 'iig6 ~d6 19 l:tabl
'YWc5 20 ctJxd4 with a very strong attack in
Djonev-Radulov, Bulgaria 1991.
b) 10.. :~d5 (this also seems quite safe) 11
i.g5 and now: c1 ) 14 l:tc1?! c6 15 ~g5 i.xg5 16 l:1c5
bl) 11...3i.d6 12 ~f6 0-0 13 ctJxd4 ctJxd4 ~xa2! 17 l:.xg5 (17 ctJxg5 0-0-0 18 ctJxe6
14 'iWxd4 "iWxd4 15 ~xd4 l:!.fd8 16 ctJxd6 fxe6 is no improvement) 17 ... 0-0-0 18 ~d2
l:txd6 17 ~e3 with a likely draw, unless true l:td5 19 'i¥b4 l:!.xg5 20 ctJxg5 'YWc4 and Black
fighters take over. was slighdy better in Sorensen-Palciauskas,
b2) 11...i.e7?! (to playa move like this corr. 1978-83.
you need to have the psyche of Victor c2) 14 ~f4 c6 15 ctJc3 'iVd7 (or 15 ... ~f5
Korchnoi; objectively it is a normal move, 16 d5 'iUxf4 17 dxe6 0-0) 16 ~a4 b5 17 'YWc2
but in real life it means that you will spend 0-0 18 l:!.adl .l:!.fe8 with a level position in
the next 20 moves under heavy attack; there- Vesovic-Kretschmar, corr. 1980.
fore it is only for connoisseurs!) 12 i.xe7 c3) 14 ~b4 i.xb4 15 'YWa4+ 'iVc6 16 'iWxb4

124
 4 d4 exd4 5 0-0

0-0-0 17 J::!.ael (17 ctJc3 'iVb6 18 'ilVxb6 axb6 ctJf6+ 'itf8 15 .i.g3 .i.e7 16 ctJe4 l:td8 17
19 .:tadl c6 is also level, but worse seems 17 ctJe5 iVh5 18 ctJd3 J::!.d7 19 ~f1 h5 Black was
ctJc5?! ~d5 18 'uael fUg6 19 'g3 'iVb6 when much better in Estrin-I.Zaitsev, USSR 1983)
Black has the better chances due to the weak 13. ...i.xe 1 14 ctJxe6 'ilVxe4 (not 14... fxe6?? 15
light squares and the prospect of ... h5-h4) ctJc5! and wins) 15 ctJxc7+ 'iitf8 16 ctJxa8
17...'iVb6 18 ~c3 1:.xd4 19 ctJc5 lthd8 20 .i.b4 17 .i.g3 and White is much better.
ctJxe6 fxe6 21 Mxe6 .l:i.d1+ 22 .:tel and a draw b) 12...iVh4?! 13 a3 'iVxb2 14 .l:!.bl ~xa3
was agreed in Radulov-Smejkal, Raach 1969. 15 ctJxd4 and White has terrific compensa-
11 .i.g5! tion.
11 b3 'ilVa3 12.i.el 'ilVa5 13 .i.d2 has also c) 12 ...g5 13 ctJf6+ cj;e7 14 ctJd5+ <Jitd8 15
been played. Now 13...'ilVa3 is the best, but ctJc3 'ilVc4 16 ..tg3 .i.g7 17 ctJe5 ctJxe5 18
also a sad solution - a draw. It is always a pity .i.xe5 ..txe5 19 J:txe5 with compensation for
that if you want play for a win, you can end the pawn. This line can of course be dis-
up being forced to take very risky decisions. cussed. Maybe 'X'hite should invest more
Here if Black wants to fight for a win he energy in the attack and have fewer material
must try the risky 13. ..'iVf5!? 14 .i.g5 .i.b4 15 constraints.
ctJxd4! ctJxd416 ~xd4 .i.xel17'ilVxg7 ~xe4
(after 17... .i.xf2+?! 18 ctJxf2 1:.f8 19 ltel f6
20 ..tf4 J:.c8 21 g4 'ilVd5 22 ~xh7 White has
more than enough compensation for the
exchange) 18 iVxh8+ <Jitd7 19 'ilVxa8 .i.xf2+!
20 <Jitxf2 ~d4+ 21 .i.e3 'ilVf6+ 22 <Jite2 ~xal
23 ~xa7 ~c3 and Black had good chances
for a draw in Skachkov-Yandemirov, USA
1991, which in the end he managed to
achieve. Nevertheless, in a later game be-
tween the same players Black took the draw
by repetition after 13 ...'ilVa3!.

12 'Dxd4!
A typical tactic for this variation, which
was practically forced here. After 12 c3 iVxdl
13 1:.axd 1 dxc3 14 bxc3 .i.a3 15 ~f4 0-0
Black is slightly better.
12....i.xe1 13 tLlxe6 fxe6
Black is forced to do this, either here or
after 13. .. ..txf2+ 14 'ith1 fxe6 when 15 iVh5+
g6 16 'ilVg4 transposes to the game. If
13. ..'ilVxe4? 14 ctJxc7+ cj;f8 15 ctJxa8 ..txf2+
(or 15 ... .i.a5 16 'ilV d7) 16 'iitxf2 iV f5+ 17 'itg 1
iVxg5 18 iVd7 'ilVe7 19 iVc8+ and White is
11 ... .i.b4? clearly better.
This leads more or less to a lost position. 14 'VWh5+ g6 15 iVg4 .i.xf2+ 16 ~h1!
Necessary was 11 ... h6 12 .i.h4 and now: White is a rook down, but the situation
a) 12... .i.b4 13 ctJxd4! (a new move but for Black is not easy at all.
similar to the game; after 13 J:.e2? g5 14 16 ... tLld4 17 iVf4!

125
Two Knights Defence

Targeting the black king. 17 CLlxf2? gives White will be very happy if he makes a draw.
Black what he needs most of all ... time. After 22 ~d2! ctJd6
17 ... Mf8 18 .J1.e3 (if 18 CLld3? ~xa2! wins) If 22 ... ~d4 23 'iVxd4 CLlxd4 24 Mxd4
18... 0-0-0 19 c3 h5 20 ~4 "i¥c2 21 cxd4 White should win the endgame.
'iVe2 22 ~3 l'1xf2 23 .J1.xf2 ~xf2 24 "i¥xe6+ 23 ~c3 'it>d7
r;i;>b8 Black has a wonderful major piece end- Or 23 ...Mf8 24 CLlxd6+ r;i;>b8 25 CLle4 and
game. The white king is in danger from ... h4- WillS.
h3 and the excellently placed black queen. 24 ctJc5+ ~c6 25 ctJa4+ 1-0
17 .. :~a5 18 "iVf6 ~d7
Game 54
Weber-Grzelak
Com:spondeJlce 1992

1 e4 e5 2 ctJf3 ctJc6 3 i..c4 ctJf6 4 d4

exd4 5 0-0 ctJxe4 6 :i.e1 d5 7 i..xd5
'iVxd5 8 ctJc3 'iVa5 9 ctJxe4 i..e6 10 i..d2
'¥ff5

19 '¥fxf2?
Here 19 CLlxf2! was correct - then if
19 ... r;i;>c6? 20 'iVxd4 'ii'xg5 21 CLld3 and the
black king cannot escape (e.g. 21...b6 22
CLlb4+ r;i;>b7 23 'iVe4+ and wins), or 19 ...CLlf5
20 b4! ~xb4 21 Md1+ ~c6 22 CLld3 'iVa3 23
'iVxe6+ CLld6 24 CLle5+ r;i;>b5 25 Mxd6! ~xd6
(or 25 ... cxd6 26 'iVc4+ ~b6 27 CLld7+ ~a5 28
i..d2+ mates) 26 a4+ ~a5 27 tt:\c4+ and wins, This gives Black safe play.
while after 19 ... 'iVd5 20 c4! 'ii'd6 21 CLle4 11 i..g5 h6
'i!¥b6 22 Mdl White regains the material with Black has two alternatives. One decent
a clear advantage. and one indecent:
19 ... ctJf5 20 .l:r.d1 + ~c8 21 g4 a) 11...i..d6 12 tt:\xd4 tt:\xd4 13 'iVxd4 0-0
White still has some compensation which 14 tt:\xd6 'iVxg5 15 tt:\e4 'iVd5 and a draw was
perhaps is enough for a draw, but should agreed in Sveshnikov-Geller, Sochi 1983.
never offer anything more. b) 11....i.c5? 12 CLlh4! wins material after
21 ... 'iVb6?? 12...'iVd5 13 c4! 'iVe5 (13. .. 'iVxc4 14 Mel
A tactical blunder. Instead after 21.:.CLld6! 'iVxa2 15 CLlxc5 was Baird-Halprin, Vienna
22 'iVd4 Me8 Black seems to ride out the 1898) 14 f4 d3+ 15 r;i;>hl 'iVd4 16 CLlf3 'iVxc4
storm; e.g. 23 c4 e5 (not 23 ... b6? 24 tt:\xd6+ 17 Mel, or 12...'iVe5 13 f4 ~d5 14 f5 d3+ 15
cxd6 25 ~xd6 ~b7 26 'ii'd7+ ~a6 27 i..e7 CLlxc5 'iVxc5+ 16 .J1.e3 'iVc4 17 fxe6 because
':xe7 28 'iVxe7 and White is much better of 17 ...'iVxh418 exf7+ ~xf7 19 'iVf3+ ~f6 20
because of the weak black king) 24 'iVd3 (or ~d5+ ~g6 21 .l:i.f1 and, unfortunately, Black
24 CLlxd6+ cxd6 25 'iVxd6 'iVc7) 24 .. :iVa6! and is toasted.

126
 4 d4 exd4 5 0-0

12 ~h4 17 ....l::.f8!? is untested and dangerous, but if

12 'iVd3? is well met by 12.. :iVa5! 13 S/..d2? you want to win in chess you need to run
if 13 i.h4 g5 14 S/..g3 0-0-0 and Black is some risks. Then again, after 18 'iVd4 'iVxc2
clearly better) 13 .. :iiVa6!! and Black is a pawn 19 ctJe4 .l:tg8 20 ctJf6+ ~f8 White has a draw
up for no compensation. The doubled pawns all the same with 21 ctJh7+! 'iVxh7 22 'iVc5+
after 14 'iVxa6 bxa6 would not matter as they 'it>g7 23 'iVe5+ ~g6 24 .l:tf1! 'iVg 7 25 S/..f6
cannot be attacked; more important is that 'iVf8 26 'iVe4+ <Jo>h5 27 'iVh4+ ~g6 28 'iVe4+
Black has the two bishops. etc. White cannot win but he does have a
12 .....te5 more attractive draw with 27 k!.f5+!? i.xf5 28
Not 12...g5?? 13 ctJxd4 ctJxd4 14 'iVxd4 'iVxf5+ .l:!.g5 29 g4+ <Jo>h4 30 Jtxg5+ hxg5 31
and White wins. 'it>g2! threatening 32 h3 or 32 'iVf3 and mates,
13 b4! thus forcing Black to play the continuation
The only way of breathing life into the po- 31...'iVa3 (31..:iVc8) when perpetual follows
sition. by 32 'iVh7+ ~xg4 33 'iVe4+ ~h5 34 'iVh7+
13 .....txb4 etc.
For those players who do not dance, I can 18 J:td1+ WeS 19 ~e3+ "VIVe5 20 ~f3+
recommend 13. .. .ltb6 14 a4 as 15 b5 ctJb4 16 ..td5 21 "VIVfS+ ..teS
ctJxd4 S/..xd4 17 'iVxd4 0-0 when the game is 21...~b5 22 ~2+ cJtc6 also draws, but
completely equal. not 22 ... <Jo>a6?? 23 J:!.d4! i.c4 24 ctJe4 with a
14lUxd4lUxd4 1 5 ~xd4 ..txe1 1S ~xg7 decisive attack.
22 'iVf3+ ~d5
Not now 22 ... 'it>b5?? since after 23 .lte7!
'iVxe7 24 I:i.b1+ <Jo>a4 25 'iVf4+ <Jo>a5 26 'iVd2+
'it>a4 27 'iVd4+ ~a5 28 'iVc3+ <Jo>a4 29 ctJd3
Black soon will find himself mated.
23 'iVfS+ Yo -Yo

Game 55
Leygue-Flear
St. Affnque 2001

1 e4 e5 2 lUf3 lUeS 3 ..te4 lUfS 4 d4

exd4 5 0-0 ..te5 S e5
1S .....txf2+! 6 c3!? transposes to the Italian Game,
Black should shake things up. Terrible is where 3 ii.c4 Jtc5 4 c3 ctJf6 5 d4 exd4 6
16 ... ~d7? 17 .l:!.xel b6 (or 17 ... b5 18 'iVd4+ O-O!? is the most usual move order. But we
~c6 19 .ltf6 .l:!.hc8 20 i.e5! with a very can have a short look all the same: 6 ... ctJxe4!
strong attack in Vytopil-Kelstrup, corr. 1966) (the only serious move; 6 ... dxc3 gives White
18 ~c3! J:!.ad8 19 f3 'it>c8 20 .ltg3 and White good compensation in more than one way,
was much better in Maliakin-Timoshenko, and 6 ... 0-0 7 cxd4 gives Black a lot of prob-
Katowice 1992. Note that the opposite- lems) 7 cxd4 d5! (the point; without this
coloured bishops favour the player who has move the gambit would be rather dangerous)
sacrificed the exchange. Since Black cannot 8 dxc5 dxc4 9 'iVxd8+ (or 9 'iVe2 'iVd3! 10
force any further exchanges, the white pieces .l:!.el f5 11 ctJbd2 0-0 12 ctJxe4 fxe4 13 'iVxe4
can roam the board without being matched. S/..f5 14 'iVf4 .l:!.ac8 and Black is no worse)
17 lUxf2 Wd7 9 ... ~xd8 (worse is 9 ... ctJxd8?! 10 .l:!.el f5 11

127
Two Knights Defence

ctJc3 0-0 12 ctJxe4 fxe4 13 i:txe4 i.e6 14

ctJd4 SLd5 15 Me5 c6 16 i.d2 and White is
better) and this ending should be equal.
White will get some fast moves, but he is a
pawn down and with the queens off the
board, there is no real attack. After 10 Md1+
i.d7 we have:

7 .1Lf4!
White strengthens the centre and prepares
to irritate the misplaced black knight. Besides
this obvious and sound move, he has a wide
range of alternatives:
a) 7 Mel? d3 and White is in trouble.
b) 7 i.g5?! i..e7 8 i.f4 is no clear im-
a) 11 ctJg5 ctJxg5 12 i.xg5+ f6 13 i.e3 (13 provement, Black can play either 8".f6 9 exf6
i.f4!? is a possible improvement according ctJxf6 10 ctJxd4 ctJxd4 11 'iVxd4 d5 with
to some annotators, but 13".ctJb4 14 ctJa3 equality, or try the mad-looking 8".g5! ctJxg5
ctJd3 seems more than fine for Black, though (9 i.g3 h5 10 ctJxd4 ctJcxe5 seems to favour
the position is not completely clear) 13".~e7 Black as White will have to trade off his
14 ctJa3 SLe6 15 !tael? (White could have dark-squared bishop next) 9".d5! 10 exd6
kept the balance with 15 ctJb5! Mhc8 16 SLf4 i.xg5 11 Me1+ ~f8 12 dxc7 'iVf6 and White
ctJe5 and Black is hardly any better) 15".ctJe5 does not have enough for the piece. One
16 l:tc3 Mhd8 17 Mxd8 ~xd8 18 h3 Md 1+ 19 important thing to note is that after 13 i.xg5
~h2 Mal! 20 f4 ctJc6 21 ctJxc4 Mxa2 22 i.el Black should attack with 13".~xf2+ 14 ~hl
ctJd4 23 .l:e3 ~f7 24 ctJd2 !tal 25 Mel i.d5 Mg8! as White cannot save the bishop;
26 g4 ctJe2! 27 Mxe2 Mxel 28 b4 as 0-1 15,,:~h4 follows if the bishop moves away,
Jablonsky-Konikowsky, Poland-Germany while on 15 MEl (or 15 ~d2) comes simply
1991. 15".Mxg5! winning.
b) 11 i.e3 ~e7 (also possible is 11...~c8 c) 7 c3 d5 is safe for Black (even 7".dxc3
12 Mel i.e6 13 ctJa3 c3 14 bxc3 b6 with could be considered here) 8 i.b5 (8 i..b3
equality) 12 ctJa3 i.e6 13 ctJb5 .l:thc8 14 ctJg5 dxc3 9 ctJxc3 0-0 10 SLf4 gives an unclear
ctJxg5 15 iLxg5+! f6 16 i..f4 ctJe5 (the posi- game in which Black is at least not worse)
tion is level) 17 i.xe5 (not 17 Mel? ctJd3 and 8".dxc3 9 ctJxc3 (too optimistic is 9 ~a4?!
Black won in Petronis-Kreuzer, corr. 1988- 0-0 10 i.xc6 bxc6 11 ~xc6 cxb2 12 SLxb2
90) 17".fxe5 18 l:i.el ~f6 19 Me3 i.d7 20 Mb8 13 ~xc5 Mxb2 and Black is slightly
ctJc3 i.c6 with equal play in Estrin- better) 9...0-0! and 10 ~xd5 ~xd5 11 ctJxd5
Krzyszton, corr. 1972-75. ctJgxe5 12 ctJxe5 ctJxe5 13 ctJxc7 Mb8 14 i.f4
6 ... lLlg4?! i.d6 is equal. The threat of ".ctJf3+ gives
The logical and normal 6".d5 is consid- Black time to avoid any bad side effects of
ered in Games 56 & 57. being pinned.

128
 4 d4 exd4 5 0-0

d) 7 ~xf7+!? is a very interesting move Black has problems developing. After

which, unfortunately (or thank God), is not 13 ... ~fS 14 liJbS .l:i.c8 15 -idS Black would
well known. After 7...Wxf7 8 liJgS+ 'it>g8 (if be struggling to keep his bits together; at the
8... 'it>e8 9 'iVxg4 liJxeS 10 'iVg3 with a dan- minimum he will lose the pawn back, but
gerous initiative) 9 'iVxg4 then: probably more.
a h3 ltJhS 9 .i.xhS gxhS 10 c3 d5

dl) 9... liJxeS 10 'it'g3 and White has com-

pensation for the pawn. The game could 11 .i.d3?!
continue 1O ... d6 11 Mel 'iVd7 (not l1...liJm The bishop targets h 7 from here, but this
12 b4 ~b6 13 'iWb3! dS 14 'iVxdS! and Black gives up a lot of the pressure on the black
is in trouble, or l1..:iiff6? 12 liJe4 'it'e6 13 f4 centre. More standard is 11 iLb3! -if5 (if
and White is close to winning) 12 ~f4 h6 l1...dxc3? 12 liJxc3 d4 13 liJdS and White
(after 12 ... liJg6 13 'iYb3+ dS 14 iLxc7 White has a clear advantage) 12 cxd4 iLb6 13 liJc3
is much better) 13 liJe4 and White has more ~e4 14liJxe4 (after 141:tel? -ixf3 15 'iVxf3
than enough compensation for the pawn and liJxd4 16 'iVg4+ ~h8 17liJxdS .l:i.g8 18 'iVe4
is at least slighdy better. liJxb3 19 axb3 .l:i.g6 and the position was
d2) 9... h6! 10 'iVf3 (if 10 liJf3 d6 11 'iVhS equal in Louma-Dobias, Prague 1943)
dxeS 12 liJxeS liJxeS 13 'it'xeS iLd6 14 'iVe4 14... dxe4 15 dS' exf3 16 dxc6 fxg2 17 'it'g4+
't:!ff6 and Black keeps the extra pawn and a 'iWgS 18 Wxg2 bxc6 19 f4 and White has real
good position, though White is not markedly winning chances in this endgame, especially if
worse after 15 c3!) 1O ...'t:!fe7 (lO ... hxgS 11 he manages to penetrate to the seventh rank
't:!fdS+ ~h7 12 'it'xcs is somewhat better for and attack f7. Should the f7-pawn fall White
White, as the black king has nowhere to hide) would be able to push his two pawns straight
11 'iVdS+ Wfs 12liJf3 (12liJe4 ~b6 13 iLf4 to the fInishing line. Probably Black's best
't:!ff7! gives Black a better endgame with the now is 19 ... ~xg4+ 20 hxg4 -id4 21 ~ac1 cS
two bishops) 12...gS and the game is unclear. 22 Mh1 ~g7 23 ~f3 when White has a
7 ... 0-0 strong attack based on Mc2-h2xh6 assisted
The passive 7... d6 does not really work by his f-pawn and king.
out: 8 exd6 ~xd6 9 lie 1+ 'it>fs 10 -ixd6+ 11...dxc3
'iVxd6 11 c3 'iWcs 12 'it'e2 -id7 13 cxd4 'iWd6 A very risky move. More calm was l1...f6
14 'iVd2 hS 15 liJc3 and White was better in when White has nothing better than 12 cxd4
Reid-Will, Scodand 1994. l1...dxc3 12 liJxc3 iLxd4 13 liJxd4 liJxd4 14 -ixh7+ ~xh7 15
'iVxdl 13 Maxd1 is not a credible alternative 't:!fxd4 fxeS (lS ... i.fS is also playable) 16
either. White has a strong initiative, while 'iVxeS l:;Ig8!? (or the very solid 16 ... c6 when

129
Two Knights Defence

the position is even) 17 '.t>hl l:!.g5 18 'iVe2 so he should playas risky as possible, simply
l:tg6 19 f4 j"fS 20 liJd2 with unclear play. to change the progression of events. One try
12 ttJxc3 i.e6 13 ttJe2 i.e 7 14 ttJf4 is 21...liJb4 22 'iVxc7 liJxa2 23l:txd4l:tfc8 24
14 a3!? with the idea of ii.c2 and 'iVd3 was 'iVa5 liJc1 and while White is much better
an interesting plan. after 25l:td7 at least Black has some play.
14... i.g5 15 ~a4?! 22 i.xc6!
Now White exchanges pieces favourably
and ruins the black pawn structure at the
same time. The position is winning already.
22 ... bxc6 23 ttJxd4 i.xh4 24 g3 J:tg8 25
~xe6
White would love to play an endgame.
25 ... ~h5
If 25 ...'i1'xe6 26 liJxe6 l:tbe8 27 liJd4l:txe5
28 liJf3 l:te4 29 liJxh4 .l:i.xh4 30 l:td7 and
White will win this rook ending without too
many problems. After the c7-pawn falls.
Black will have serious problems with the
seventh rank - he can never exchange all the
White is not playing very actively here and rooks, as the pawn ending will be lost.
seems to be waiting for Black to make a mis- 26 'iVf5 J:rg5 27 ~f3 .l':!.xe5 28 ~xh5
take (which duly appears). Better was 15 J:rxh5 29 ttJxc6
liJh5!? f5! (15 ... ~h8!? is also possible, though
White is better after 16 liJxg5 hxg5 17 f4!
with attack against the black king) 16 'iVb3
'iVe7 17 l:tae 1 '.t>h8 18 liJxg5 hxg5 19 f4 with
good compensation for the pawn. One pos-
sible continuation is 19 ...'iVb4!? 20 fxg5
'iVxb3 21 axb3 with even chances.
15 ... d4?
Now the light squares in the black camp
will become terribly weak. Black should not
have allowed the exchange of the bishop, and
especially not by simultaneously opening
lines for the white bishop. Instead 15 ... ..ltd7!
and Black is probably better. Although material is equal White has a
16 ttJxe6 fxe6 17 i.e4 Itb8 18 J:rad 1 ~e8 technically winning position. All the black
19 h4!? pawns are weak, shattered and isolated and
If you are the active player in the position, his pieces are completely unco-ordinated.
it usually stronger to keep the pressure on Besides that the black bishop has no future
rather than to exchange pieces. Nevertheless, potential, while the white knight is as happy
after 19 ..ltxc6 'iVxc6 20 ~xd4 White is also as can be.
better. 29 ... J:tg8 30 r;t>g2 i.f6 31 ttJb4 J:tb5 32
19 ... i.e7 20 'iVc4 iif7 21 b3 r;t>h8?! ttJd5 i.e5 33 l:!.fe1 i.d6 34 ttJe7 J:tf8 35
Black is just waiting for death to come to ttJd5 J:rc5 36 ttJe3 J:rc3 37 J:re2 J:rf7 38
his door. He has not a lot to lose anymore, ttJc4 J:td7 39 .l':!.d5 J:rd8 40 .l':!.a5 J:tf8 41

130
 4 d4 exd4 5 0-0

lLlxd6 cxd6 42 J:!.xa 7 l:tcf3 43 J:td7 1-0 7 ... dxc4

In my opinion this posmon is at least
Game 56 equal for Black. The two pawns on d4 and c4
Steinitz-Meitner give Black a very strong centre.
Vienna 1860 8 l:.e1 +
Inferior is 8 fxg7?! .:i.g8 9 ii.g5 ~e7 10
1 e4 e5 2 lLlf3 lLlc6 3 d4 exd4 4 iLc4 ~xe7 cttxe7 11 ctJbd2 ~xg7, e.g. 12 ctJxc4
iLc5 5 0-0 lLlf6 6 e5 d5! ~e6 13 l:tel Wffi and Black was somewhat
better in Foltys-Stulik, Czechoslovakia 1940.

This is far more logical than 6.. .'~jg4. In

my humble view, if you want to play the 8 ... .te6
black side of the Max Lange Attack, this is 8...cttf8!? is risky after 9 ~g5 gxf6 10
the key move to justify doing so. And be- ii.h6+ 'it>g8 11 ctJc3! i.f8 (not 11...~g4?1 12
sides, after this you should be able to find ctJe4 ~e7 13 h3 i.h5 14 c3 d3 15l::te3 with a
some interesting blank spots on the map and strong initiative - Pilk6vi) and now:
fill them up with your own analysis. Remem- a) 12 ctJxd4 i.xh6 (not 12... ctJxd4? 13
ber that this line was popular 100 years ago. 'iWxd4! 'iexd4 14 .l:.e8 'iVd6 15 ctJd5! and
This means that not only have the lines have wins) 13 ctJxc6 'iit'xdl 14 ctJe7+ ~g7 15
never been properly computer checked, but l:taxd 1 ~e6 is equal.
also that as chess has changed so radically b) 12 ii.xf8 'it>xffi 13 ctJe4 ~f5 14 'iVd2 h5
such a dynamic position as this can be 15 'iff4 and 'Xmte has a strong attack
viewed from a new perspective. For example, according to Pilk6vi. Better is 13... f5, though
it has only been really understood in the last White has lasting counterplay on the weak
50 years how much compensation it is possi- dark squares.
ble to have for an exchange sacrifice. And 9 ltJg5 ~d5
since Kasparov there has been a completely 9... ii.ffi? is bad due to 10 'ifg4! gxf6 11
different understanding of the initiative. All ctJxf7! cttxf7 12 'ifxe6+ r:JiJg7 13 'ifg4+ cttf7
of this can assist you tremendously in analys- 14 'ieh5+ ctt g8 15 ctJd2 with a fantastic attack
ing and/ or playing this line. (e.g. 15 ... .tb4 16 l:te4!) and has invested no
7 exf6 material for it! Black was hoping for 10
This is the move that is characteristic for ~xe6+? fxe6 11 ctJxe6 'ifxf6! 12 ctJxc7+ ~d7
the Max Lange Attack. 7 ~b5 ctJe4 8 ctJxd4 13 ctJxa8 iLd6 (better than 13 .. :~e6 14 ~f4
transposes to Chapter 7, though White is ~d6 15 ~xd6 ~xd6 and now 16 ctJa3!?
now committed to the 8 0-0 line (Game 43). looks interesting) 14 ~g4+ 'ife6 15 'ii'xg7+

131
Two Knights Defence

i..e7 with a messy position, e.g. 16 ~d2 .l:.g8 15 ~e2!?

17 'iVxh7 'iVg4 18 g3 CDe5 19 i..el ':g720 It is not obvious that this is a bad move at
'iVh6 ':g6 21 iVh8 CDf3+ 22 'it>h1 CDh4! and all. Here Steinitz suggested 15 ~h6 as win-
Black won in Aldrete Lobo-Oim, corr. 1998. ning, but that is not clear, as Black would
10 lbc3 'iVf5 11 g4?! respond 15 ... i..d6! and the position is a mess.
This is too optimistic. 11 CDce4 is better, But stronger is 15 CDdxc7+! ~f7 16 CDg5+!
as considered in Game 57. (16 ~f3+ ~g8 17 ~f5 CDd8! 18 CDxa8 ~xe6
19 ~xc5 ~xg4+ 20 ~£1 I/Z_I/Z Korsano-
Dobrey, Sharjah 1985) 16 ... ~g8 (not
16 ... ~g6? 17 ~f3! and wins) and now:
a) 17 CDxa8 ~d6?! (but if 17 ... ~e7 18 ~f3
t2Je5 19 iVe4 ~xg4+ 20 ~xg4 t2Jxg4 21 CDc7
CDxh2 22 ~xh2 i..d6+ 23 ~g2 i..xc7 24 CDf3
and the endgame is winning for White, while
after 17 ... h6 18 CDe4 i..f8 19 'iVe2 d3 20 cxd3
cxd3 21 'iWd1 and Black has no compensa-
tion) 18 'iVe2 CDe5 19 i..f4 d3 20 'iWe4 'iVxg4+
21 ~hl and White won in Shue-Wood, Can-
berra 1996.
b) 17 ~e2! may be even stronger; e.g.
11 ... v.lVxf6? 17 ... d3 18 ~e4 i..xf2+ 19 ~g2!! (19 ~xf2
Black falls for the trap. Instead 11 ... 'iVg6! ':f8+ 20 'it>e3! b5 21 cxd3 CDe5 22 iVd5+
was correct and after 12 CDce4 (if 12 CDd5? 'iWxd5 23 CDxd5 CDxg4+ 24 ~d4':£1 25 dxc4
0-0-0 13 CDf4 ~xf6, or 12 CDxe6 fxe6 13 ':d 1+ is less clear) 19 ....l::!.c8 20 'iVxc4+ ~f8 21
':xe6+ ~d7 14 f4 ~xc2! 15 'iVxc2 d3+ 16 ..tf4! and Black cannot defend himself satis-
~g2 dxc2 and Black is better) 12... i..b6 13 f4 factorily. One line goes 21...':xc7 22 ~xf2!
0-0-0 14 f5 i..xf5 15 gxf5 'iVxf5 Black had a and Black has no good moves, as after
very strong attack in Blackburne-Samisch, 22 ... ':c8 23 i..d6+ ~e8 24 .l:i.e1+ ~d8 25
Bled 1931. CDf7+ he is history.
12 lbd5 v.lVd8 13 .!:!.xe6+! fxe6 14 lbxe6
What a mess! If 14.. :~d6 15 i..f4 and
White wins.
14 ... v.lVd7

15 ... ~e7?
The reason why Steinitz was unhappy
with his play was presumably 15 ... ~f7!?, but
White still wins after 16 t2Jg5+ ~g8 17 t2Jxc7!

132
 4 d4 exd4 5 0-0

d3 18 'iVe4 .i.xf2+ 19 <t>g2 as illustrated in exd4 5 0-0 ~e5 6 e5 d5! 7 exf6 dxe4 8
the previous note. But of course Black has J:re 1 + i..e6 9 tUg5 ~d5 10 tUc3 ~f5 11
more opportunities to create a mess here. tUce4
16 tUdxe 7 + ~f7 17 ~xe4
Black cannot hold the position now.
17 ... tUe5
17 ...<t>g6 18 'fgd3+ <t>f7 19 ~3! 'it>g6 20
lL\f4+ c;t>f6 (20 ...<t>h6 21 ~3+ is the idea
behind the queen shuffling) 21 g5+ <t>f5 22
~3+ and Black is mated in five moves, if
White is not satisfied with winning the
queen.
18 ~b3 ~d6
18...<t>f6 is met by 19 .tg5+ <t>g6 20 .txe 7
~xe7 21 lL\f4+ c;t>g5 22 h4+ c;t>xg4 23 ~h3+
<t>xf4 24lL\d5+ and wins.
19 f4! 11 ... 0-0-0
Black has two other possibilities:
a) 11...~b6?! should be bad, being a slow
move in a wild tactical position. It might be
effective as a surprise weapon since the cor-
rect method of attack is not obvious, but
objectively White is better.
a1) 12 lL\xf7? O-O! and Black has a clear
advantage.
a2) 12 lL\g3 'ifg6 (not 12 .. :iVxf6? 13 lL\h5
and the black position collapses) 13 lL\xe6
fxe6 14 .l:!.xe6+ 'it>d7 15 lL\h5 ~he8 16 lL\f4
(or 16 .l:!.xe8 .u.xe8 17 lL\xg7 .l:!.f8 and Black
does appear to be worse) 16 .. :~Vf7 17 ~f3
Steinitz has no reason to be unhappy with and here ECO gives White as having a clear
his choices in this game. He plays the attack advantage, but it not at all clear that is the
excellently. case. Strongest is probably 17 ... .l:!.ad8! (if
19 ... tUxg4 20 tUg5+ ~g6 21 ~d3+ 'iith5 17 ... l:i.xe6 18 'ifd5+ <t>c8 19 lL\xe6 d3 20 ~e3
22 ~h3+ Iitg6 23 ~xg4 ~b6 ~xf6 21 cxd3 "iVxb2 22 I:td1 ~e2 23 dxc4
Now it is just mate. Black's best option and White wins) 18 ~d2 gxf6 19 .l:!.ae1
was 23 ... ~xg5 24 'fgxg5+ <t>f7 25 'fgf5+ <t>e7 .l:!.xe6? (a weak move, exposing the king; in-
26 lL\xa8 and White wins. stead 19 ... lL\e5! 20 'iVd5+ <t>c8 21 .l:!.xe8 'iVxe8
24 tUge6+ ~f6 25 ~g5+ Iitf7 26 ~xg7 would leave Black with the advantage) 20
mate lL\xe6 .l:!.e8 21 lL\g5 .l:!.xe1+ 22 .txe1 'WIe7 23
r - - - - - - - - - - - - - - -.. ~f5+ <t>d8 24 .td2 ~e2 25 ~xf6+ lL\e7 26
Game 57 h4 (Black has overpressed and is now worse.
Delaney-Hebden White wins the game in nice style) 26 ... d3 27
Kilkenny 1999 lL\f7+ <t>c8 28 'lWh8+ <t>d7 29 'fgd8+ <t>e6 30
lL\g5+ ~f5 31 'fgf8+ 1-0 Chigorin-Charousek,
1 e4 e5 2 tUf3 tUe6 3 ~e4 tUf6 4 d4 2nd match game, Budapest 1896.

133
Two Knights Defence

a3) 12 fxg7 .1:.g8 13 g4! (the critical varia- safety, his bishop is out of the game and he
tion; 13 lbxf7 ~xf7 14 lbgs+ ~xg7 IS will lose the c4-pawn too.
lbxe6+ ~h8 is very good for Black as the b) 11....ifS!? should also be a small mis-
white king is exposed here) 13 .. .'iVg6 14 take in my opinion. Perhaps I am wrong,
lbxe6 fxe6 IS .igS 'uxg7 16 'iVf3 and Black because this is an idea from Akiba Rubinstein
has large problems with his defence. - and as Paul Keres taught us, Rubinstein's
ideas are immortal! But let us get serious! The
bishop retreats from a good post in the cen-
tre back to the starting position - and this in
an open game! White now continues 12
lbxf7! ~xf7 13 lbgs+ ~g8 14 g4 'iYg6 (not
14... 'ilVxg4+?? IS 'iYxg4.i.xg4 16 f7 mate) and
now:
b 1) IS fxg7?! .idS! 16 gxh8'iV + '.toxh8 17
f4 (or 17 lbh3 SLd6 and Black has counter-
play according to Rubinstein) 17 ....icS 18 f5
d3+ 19 ~f1 .1:.fS 20 i..f4 and ECO claims a
clear advantage for White, but 20 ... h6! gives
Black a strong attack, e.g. 21 fxg6 l:.xf4+ 22
This has been known since an article by lbf3 1:i.xf3+ 23 ~g2 .1:.£2+! 24 ~h3 i..d6 25
G.Abels in Deutsche Schachzeitung Nr.ll, 1900! .1:.e8+ ~g7 26 'iYgl .l:.f3+ 27 ~h4 lbe7! 28
Black now has the following tries: 'iVd4+ ~xg6 when White would require a
a31) 16 .. J:W 17 lbf6+ .1:.xf6 18 'iVxf6 good deal of luck to survive.
'iVxf6 19 i..xf6 ~f7 20 gS and White is b2) IS .l::!.xe6 gxf6 16 'ilVf3 ~g7 17lbe4 (17
clearly better according to Gligoric. .if4 was played in Surowiak-Jaworsky, corr.
a32) 16...~d7 17 lbf6+ ~c8 18 .1:.xe6 1994, when Black has many moves, the most
'iVxgS 19 h4! 'iVg6 20 hS 'iYgs 21 .1:.ael lbd8 solid being 17 ....l:.e8 18 .1:.xe8 'iYxe8 19 lbe4
22 .1:.6eS and the white attack should win the 'ilVg6 and Black will quickly play ... .1:.g8 and
game, e.g. 22 ... 'iVh4 23 'ue8 c6 24 'iYfS+ ~b8 ... ~h8 with a good game) 17 ...lbeS
(or 24... ~c7 2S .l:i.le7+ .l::!.xe7 26 .1:.xe7+ ~b8 (17 ... SLe7!? is as a possible improvement, e.g.
27lbd7+ ~c7 28 'iYeS+ mates) 2S h6 .1:.gS 26 18 SLgS .l::!.afS 19 l:l.e 1 .1:.f7 with chances for
'iVd7 and wins. both sides in a messy position) 18 'ilVxf6+
a33) 16 ... eS 17lbf6+ rJJe7 (or 17...~f7 18 'ilVxf6 19 lbxf6 .id6 with an unclear game
h4! h6 19 lbe4+ ~e6 20 hS 'iVf7 21 i..f6 according to Keres .
.l:i.gg8 22 'ilVf5+ '.t>dS 23 b3! and White won in 12 g4
Chigorin-Teichmann, London 1899) 18 h4 12 fxg7?! achieves nothing. 12 ... .1:.hg8 13
(after 18 i..h4 .l::!.f7! 19lbxh7+ ~d7 20 'iYdS+ g4 (or 13lbxcs 'ilVxcs 14 .l::!.xe6 fxe6 ISlbxe6
rJJc8 21 lbgs .l::!.f8 the position could go ei- 'ilVdS 16 lbxd8 .1:.xg7 and Black is better)
ther way; White is probably forced to enter 13. ..'iYxg4+! 14 'iYxg4 SLxg4 IS lbxcs .1:.xg7.
an unclear endgame bv 22 'ilVe6+ 'ilVxe6 23 for example 16 ~f1 d3 (or 16 ... h6!?) 17 f3
lbxe6 .1:.g8 24 h3) 18... h6 (if now 18....1:.f7 19 SLfS 18 cxd3 cxd3 19 i..d2 h6 20 lbge4
.1:.xeS!+ lbxeS 20 .1:.el '.tofS 21 .1:.xeS wins i..h3+ 21 ~£2 .1:.g2+ 22 ~e3 lbd4 and Black
according to Chigorin; e.g. 21...~g7 22 .1:.fS wins.
~h8 23 hS 'ilVg7 24 lbd7! ~g8 2S 'ilVdS) 19 12 .. :tWe5
lbg8+ ~e6 20 lbxh6 and White is clearly The only move. 12.. .'iVd5? 13 fxg7 .l:.hg8
better. While Black can get his king into 14lbf6 'iYd6 IS lbge4! trapped the queen in

134
 4 d4 exd4 5 0-0

Chigorin-Albin, Berlin 1897, while if much better.

12.. :iixg4?! 13 ~xg4 i.xg4 14 tiJxf7 gxf6 15 13 ... fxe6 14 .i.g5
tiJxd8 tiJe5 16 i.f4! tiJf3+ 17 ~g2 tiJxe 1+ 18 White does not have a bright future after
l:txel i.b4 19 c3 with a clear advantage to 14 fxg 7?! l:thg8 15 ..ih6 d3! (Black needs to
White. get his pieces working) 16 c3 and then:
a) 16 ... d2 17 .i:!.e2 .i:!.d3 is highly unclear;
e.g. 18 tiJxc5 (if 18 ~f1 ~d5 19l:tdl ~e7 20
~g2 tiJe5 21 'uexd2 tiJf7 22 g5 l:i.xd2 23
l:txd2 ~f5, with the idea of ... ~g6 followed
by ... tiJxh6, might be slightly better for Black)
18... ~xc5 19 .l:!.xd2 tiJe5 20 l:txd3 cxd3 21
~g2 ~d5+ 22 'it>g3 tiJf7 23 ~d2 with un-
clear play in Radulov-V.Sokolov, Yugoslavia
1961.
b) 16 ... .te7 seems stronger and if 17 f4
~d5 18 ~d2 (as in Friedmann-Marthinsen,
corr. 1984) then 18 ... i.h4 19 l:i.e3 tiJe7!, in-
tending ... ~c6 and ... tiJd5, and Black is bet-
13 ctJxe6 ter.
Here White could also think of:
a) 13 f4? d3+ 14 ~f1 (14 ~g2 ~d5 15
fxg 7 llhg8 with a strong attack in Rossolimo-
Medina, Malaga 1968; here ECO suggests 16
cxd3 cxd3 17 i.d2 as unclear, but after
17 ...l:txg7 Black should win without any
problems; something like 18 i.c3 i.d4 gets
White nowhere) 14... ~d4!? (14... ~d5) 15
i.e3 ~xe3 16 l:txe3 ..ixe3 17 fxg 7 l:thg8 and
Black has a strong attacking position; e.g. 18
tiJxe6 fxe6 19 cxd3 l:txg7 20 ~ e 1 i.xf4 and
Black stands better.
b) 13 tiJf3? ~d5 14 fxg7 l:thg8!? (or
14... i.xg4! 15 gxh8~ l:txh8 16 tiJf6 ~xf3 17 14 ... .i.b6!
tiJxg4 ~xd 1 18 l:txd 1 l:tg8 19 h3 h5 with a I prefer this move and not only because it
clear advantage to Black 111 Zuev- is cool! Black has also tried:
Hamatgaleev, Russia 1999) 15 tiJf6 ~d6 16 a) 14...l:td7?! is weak because of 15 fxg 7
..ih6 ..idS 17 tiJxg8 1:txg8 and White is in l:tg8 16 ~f6 'iVd5 17 tiJxc5 'iVxc5 18 l:txe6
trouble; e.g. 18 g5 'iVf4, or 18 ~e2 ~xh6!! 19 and \X/hite was better in Faas-Pukshansky,
~e8+ l:txe8 20 l:txe8+ ~d7 21 g8~ ~h3! Leningrad 1975.
and Black wins, or if 18 i.el l:i.xg7 19 tiJg5 b) 14... h6!? is a nice idea. Saether-Vajs,
tiJe5 20 ~f4 l:txg5! 21 i.xg5 i.c6! followed corr. 1978, saw 15 fxg 7 hxg5 16 gxh8~
by ... 'iVd5 and mate comes soon. l:txh8 17 tiJg3, when Black should probably
c) 13 fxg7 l:thg8 14 tiJxe6 transposes to have continued 17 ... 'it'd5! with ideas like 18
the next note, and 14 f4 d3+ 15 ~f1 ~d4 is ~e2 d3 19 ~xe6+ (if 19 cxd3 tiJd4 and it is
the same as 13 f4 above, while if 14 tiJxc5 all over bossa nova) 19 ... ~xe6 20 l:txe6 tiJd4
~xc5 15 tiJe4 ~e5 16 ~h6 d3 and Black is 21 l:tf6 dxc2 22 ~g2 ~b4 23 l:tel .td2 24

135
Two Knights Defence

ctJe2 i.xc1 25 ctJxc1 .l::teS and wins. But sition is playable for White, but my advice is
White can improve on this by inserting 15 ... play it as Black!
ctJxc5! ~xc5 and then 16 fxg7 hxg5 17 17".'~jb4
gxhS~ MxhS IS .l:!.xe6 with a clear advantage An alternative was 17 ... Sl.a5!? IS J:.fl ~f41
in Cafferty-Sombor, Bognor Regis 1965. with unclear play.
c) 14...g6 15 f7 i.e7 16 f4 ~g7 17 iLxe7 18 g5 ~h4 19 tLlg3 ~f4 20 ':'e4 ~d6 21
ctJxe 7 IS ctJg5 d3 19 ctJxe6 'ilVxf7 20 ctJxdS .ii.xh8
~xdS and Black has good compensation for Or 21 .ie5 ~c6 22 'iVd2 ctJd5 23 .ixhS
the exchange in Shkurovich Khazin-Krantz, MxhS 24 'ii;>gl MfS 25 Mfl with some chances
corr 19S1. for a save.
d) 14... .I:!.dgS 15 f4? d3+ 16 'ii;>hl 'ilVd5 17 21".l:txh8 22 a3?!
cxd3 cxd3 IS fxg 7 Mxg7 19 i.f6 was played Preferable was 22 ~g4 ctJxc2 23 Mxe6
in Blauert-Caldouras, Germany 19S9, and ~d5+ 24 'ii;>gl 'Ot>bS 25 t!.dl and though
now after 19 ...Mf8! 20 iLxg7 Mxf4 21 ~3 Black is slightly better White can still fight.
.l:!.xe4 22 ~xd5 exd5 Black is much better 22".tLld5 23 ~g4 ':'f8 24 nf1 tLlf4+ 25
according to Pilk6vi. Instead 15 ctJxc5!? ~h1 d3 26 cxd3 cxd3 27 tLlh5 tLlxh5 28
gives an unclear game. One line possible line ~xh5 'iVd5?!
is 15 ... ~xg5 16 f7 ~xc5 17 fxgS~+ MxgSlS It turns out that the queen is misplaced on
Mxe6 .l::tdS and Black has sufficient counter- d5. Instead 2S ... d2! would give Black a clear
play, but hardly anything more. advantage; after something like 29 .l::tdl 'iVc6
15 fxg7 'iVxg7 16 ~f6 30 'iVg4 .l::txf2 31 ~xe6+ ~xe6 32 .l:txe6 'ii;>d7
Not 16 i.xdS? MxdS 17 ctJg3 d3 and 33 l:!f6 ~e2 and Black should win the end-
White can take his exchange \vith him to the game without any troubles. The d-pawn is
grave. fabulous.
16".'iVh6 29 'iVg4??
Pilk6vi suggests 16...~g6!? and if 17 Necessary was 29 f3 and then, as Black
iLxhS?! .l:i.xhS IS ctJg3 .l:!.f8 with a strong has nothing-after 29 ... ~xe4 30 fxe4 Mxf1+ 31
initiative. 16... ~f7!? is also possible. 'ii;>g2 Mg1+ 32 'Ot>h3 'ii;>dS 33 ~f3, it might be
17 ~g2!? best to give the king some air before begin-
ning the fight with 29 ... a6!? 30 .i::tdl 'ii;>bS 31
~g4 ~3 with good play for Black, though
White is still very much present in the game.
29".1:1f5
Now Black is winning, though 29 .. J:hf2~
was even stronger.
30 f3 d2 31 .l:!.d1 l:txg5 32 'YIHf4 c6??
Here the game has obviously entered into
the time trouble phase. 32 ...Mg1+! 33 MxgJ
iLxgl wins very easily.
33 ~f8+ .ii.d8 34 ~f7 .!:1.g6 351:1f4??
After 35 'ilVf4 it is at all not easy for Black
to make progress.
Smart play. White understands that with- 35".'YIHd3!
out the bishop (i.e. after 17 iLxhS MxhS) his Now it is all over again.
dark squares are very weak. Perhaps this po- 36 I!.b4 b6 37 ':'g4 ':'f6 0-1

136
 4 d4 exd4 5 0-0

Summary
S 0-0 is a poor opening choice for White. It's tempting for me to say that it is even inaccurate.
Why? Well, after a few easy moves (S .. .'~Jxe4 6 ~el dS 7 ~xdS 'iVxdS SlLlc3 'iVhS) Black gets
an excellent position without any problems. So the love this line receives from club players all
around the world is completely unjustified. S...'iVaS is by the way also fine for Black; it is less
ambitious, but more solid and more popular.
The Max Lange Attack is an interesting alternative to S... lLlxe4. The difference is that while
S... lLlxe4 gives an easy game with simple and clear positions, S... ~cS creates a messy struggle,
albeit one in which Black is no worse equipped to participate. The theoreticians - and the vast
majority of grandmasters - prefer S... lLlxe4 because it leads to simple equality, which they quite
rightly regard as a success for Black. But really it is just a matter of taste.

1 e4 e5 2 tLlf3 tLlc6 3 i..c4 tLlf6 4 d4 exd4 5 0-0 (D) tLlxe4

S... ~cS 6 eS .
6...lLlg4 - Game 55
6... dS
7 ~bSlLle4 SlLlxd4 - Game 43 (Chapter 7)
7 exf6 dxc4 S ~e1+ .i.e6 9lLlgs 'iVdS 10 lLlc3 'iVfS (D)
11 g4 - Game 56
11 lLlce4 - Game 57
6 J:!.e1 d5
6 ... fS - Game 48
7 i..xd5 WUxd5 8 tLlc3 ~a5
S...'iVhS 9lLlxe4 ..te6 10 ..tgS
10 ... Ji.d6 - Game 49; 1O .....tb4 - Game 50
9 tLlxe4 i..e6 (D)
10 i..d2
10 lLlegS 0-0-0 11 lLlxe6 fxe6 12 ~xe6
12... h6 - Game 51; 12... ..td6 - Game 52
10 ... WUh5
1O ... 'iVa4 - Game 53; 10... 'iVfS - Game 54
11 i..g5 - 8 .. .'i/Vh5

5 0-0 10... WUf5 9 ... i.e6

137
 CHAPTER NINE I
4 d3

1 e4 e5 2 4Jf3 4Jc6 3 i..c4 4Jf6 4 d3 use his lead in development to launch a dev-
The move 4 d3 is less sharp than the other astating attack. But here it is acceptable be-
lines analysed in this book, but is no less cause White has already played the slower
important from a practical point of view, as 4 d2-d3, so that if (or rather when) he advances
d3 is quite popular as well. Just see how Ma- d3-d4, Black will have had his little move for
cieja wins against Ivanchuk in Game 65 be- free.
low. So although 4 d3 does seem a bit less
aggressive than the alternatives, it should not
be rejected as completely harmless.
From our starting position for this chap-
ter, Black has four options: 4 ... h6 (Game 58),
the very rare 4... d5 (Game 59), and 4... ~e7
(Games 60-66) which is the main line for this
variation. The obvious fourth choice,
4... ~c5, transposes to the Italian Game (3
~c4 Jtc5) and is therefore not covered in
this book.

Game 58
Kogan-Svidler 50-0 d6
Oakham 1992 Black reinforces his centre before playing
...g 7-g6. This move order also gives him
1 e4 e5 2 4Jf3 4Jc6 3 .tc4 4Jf6 4 d3 some additional possibilities.
h6!? 6 l:te1
The idea behind this move is simple. If 6 c3 g6 7 d4 'iVe7 8 .:tel ~g7 9 tZJbd2
Black prevents tZJf3-gS and intends to de- 0-0 10 h3 tZJh 7 11 tZJf1 tZJg5 12 tZJ 1h2 tZJxf3
velop with ... g7-g6 and ... Ji. g7. Generally in 13 tZJxf3 ~h7 14 a4 f5 and in Ghinda-
the Open Games Black cannot afford to Beliavsky, Lvov 1981, both players had their
waste time on such prophylactic measures. chances. Black can also try 6... g5!? which is
White would open the centre with d2-d4 and not so stupid as it is looks. Gelfand-Bareev.

138
 4 d3

Munich 1993, continued 7 tZJbd2 i.g7 S 'iVd3 tZJd7 11 i.e3 i.g7 and the position is
i.b3 O-O?! 9 tZJc4 i.e6 10 ~el and White is unclear.
slightly better. (l\1ost importantly, it is very c) 7 i.d5 c6! 8 i.xt7+ ~xt7 9 b4 g5! (a
hard for Black to neutralise this advantage, as new idea; after 9... c5 10 bxa5 'iVxa5 11 c3
it is based on a clear weakness in his own i.e7 12 d4 White is slightly better) 10 i.b2
camp). Alekhine played similar ideas to the (if 10 ~d2?! g4 11 tZJh4 tZJxe4! 12 dxe4
one played here by Black (in the Philidor), 'YWxh4 and Black is better) 1O ...g4 11 tZJfd2
and he used to delay castling for as long as i.e6 and only White will find problems here.
possible, in order to remain flexible. There- 7 d4 "fJiie7
fore I recommend S... tZJe7! with the idea of This is a standard move in this system .
... tZJg6-f4 and a kingside attack. Maybe Black Black reinforces the e5-square. Note that if
will castle queenside later. There is at least no White delays d3-d4 for too long Black will
reason to omit the possibility. So I find that have time for ... i.g 7 and ... 0-0, and then be
this move is really worth trying. The position able to play the preferable ... ~eS, making
in unclear. tZJc3-d5 is less disturbing and ... tZJc6-e7 a
possibility.
7... ~g4!? has also been tried, but it cannot
really be recommended. After 8 i.b5 tZJd7 9
i.xc6 bxc6 10 tZJbd2 i.g7 (if 10... exd4 11 h3
i.xD 12 tZJxD c5 White exploits his lead in
development by opening the position: 13 e5!
dxe5 14 tZJxe5 tZJxe5 15 .l:i.xe5+ i.e7 16 'iV D
Wf8 17 ~f4 with more than a pawn's worth
of play) 11 h3 i.xD 12 tZJxf3 exd4 13 tZJxd4
tZJe5 14 f4 c5 was Makarichev-Nenashev,
Russia 1993; and now 15 tZJD gives White a
slight edge according to Makarichev.
8 ttJc3 .tg7 9 ttJd5 "ilVd8 10 dxe5
6 ... g6
Here 6... tZJa5!? is a good road to heaven
(for which read 'equality'). It is important to
remember that if you can exchange White's
light-squared bishop without compromising
your own position it is nearly always good to
do so. For that reason, as soon as Black plays
... d7-d6, White will create a safe home for the
bishop, either by c2-c3 (allowing the bishop
to retreat to c2, via b5 and a4), or by advanc-
ing his a-pawn (as for example in Games 64-
66).
Here White has the following options:
a) 7 i.b5+ c6 S ~a4 b5 9 i.b3 c5 10 i.d2 10 ... dxe5
i.e7 11 i.xa5!? ~xa5 12 a4 b4 13 tZJbd2 0-0 In my opinion 1O ... tZJxe5! is better and
14 tZJc4 'iVdS 15 tZJe3 tZJh 7 and the position leads to equality. After 11 tZJxe5 dxe5 White
is equal according to Arkhipov. has two theoretical continuations to choose
b) 7 tZJc3 g6 S a4 tZJxc4 9 dxc4 i.e6 10 from:

139
Two Knights Defence

a) 12 'iVf3 tLlxd5 13 ~xd5 0-0 14 a4 c6 15 12... tLlxd5?! 13 exd5 tLlxb4 14 ~xe5 ~xe5
~c4 'iVc7 16 b3 J:.e8 17 ~a3 ~e6 18 ~xe6 15 tLlxe5 is good for White. His pieces are
J:.xe6 and Black slowly solved his opening better placed and he has a strong presence in
problems in Sermek-Malaniuk, Pula 1999. the centre.
After 19 ~adl he has 19 ... b5!? with counter- 13 h3 ~xf3 14 'iVxf3 cuxd5 15 exd5
play. There is no compensation for the pawn
b) 12 ~d2 a5! (it is important to take con- after 15 ~xd5 tLlxb4, as 16 ~xb7?! ~b8 17
trol of the b4-square; less accurate is .l:.edl 'iVf6 18 ~xf6 ~xf6 19 c3 tLlc2!? 20
12 ... tLlxd5 13 ~xd5 0-0 14 ~b4 J:.e8 15 .l:.e3 J:.ac1 tLle3 21 fxe3 .l:.xb7 gives Black a better
c6 16 ~b3 and White stands slightly better endgame.
according to Svidler) 13 ~c3 (13 'iVe2!? with 15 ... cud4
an even game is of course also possible, as Not 15 ... tLlxb4?! 16 'i¥b3! and Black loses
are other relaxed moves) 13 ... c6 14 tLle3 at least a pawn.
'iVxdl 15 tLlxdl b5 and Black is OK. .. at 16 Wid3
least! 16 ~xd4 exd4 leads to equality.
16 ... '¥Vd6

11 b4!
White begins a queenside initiative. Now 17 f47
the disadvantage of keeping the knights on This advance is completely out of touch
the board becomes apparent. The knight on with the position. Artur Kogan is an inven-
f3 is attacking e5, the knight on c6 is defend- tive and highly original grandmaster, but
ing, but the white knight is not about to be when this game was played he was still a
displaced. junior (albeit a very strong junior) and often
Instead, 11 ~e3 is too routine: after took unrealistic risks. Better was here 17 a3!
11...lDxe4! 12 ~xh6 J:.xh6 13 .l:.xe4 ~f5 14 .l:.ad8 18 iLa2 and White has a slight advan-
J:.el "ilYd7 15 tLle3 0-0-0 Black is at least no tage.
worse. White needs to escape into the end- 17 ... '¥Vxb4 18 .l:!.ab1 7
game as soon as possible, when he might be White continued with his plan of active
able to maintain equilibrium. play, probably not fully aware of the strength
11 ... 0-0 12 ~b2 of Black's response. Preferable was 18 iLa3
Not yet 12 b5 tLla5 13 tLlxf6+ 'iVxf6 14 'ib6 19 'i&thl e4 20 J:.xe4 .l:.fe8 and Black is
~d3 a6 15 bxa6 bxa6 16 'ifd2 tLlc6 17 'iVc3 slightly better. It is somewhat similar to the
and the position is about equal. game, but it is a superior version for White
12 ....ltg4 without the exchange of the bishops, as

140
 4 d3

White still has some counterplay on the dark If26 'iVg 3 ~d4 and White loses material.
squares. Or else White could put on the 26 ... a5 27 ~d4
breaks with 18 i.xd4 exd4 19 f5, though Or 27 ~d4 b5 and Black wins the bishop.
even here the opposite-coloured bishops 27 ... 'ilVf1 28 .l:l.b3 b5!
cannot guarantee a draw. Black is still better,
not because of the extra pawn, but because
the white king is vulnerable on the dark
squares. This becomes transparent with the
weak continuation 19 ~ab 1 'iVd6 20 ~xb 7
'iVxf4 and mating patterns with ... j,e5 will
soon appear.
18 ... e4!
Black returns the pawn correctly. Instead
of being tied down by pins and overloaded
pieces, he exchanges the dark-squared bish-
ops and gets a position where his knight is
fabulous on f5/d6, while White's bishop on
c4 is terribly limited by his own pawns. The key move and the bishop is simply
trapped. What a glorious end to Black's strat-
egy of strong knight against weak bishop.
29 ':'f3 ~e2 30 SLb3 a4 31 SLxa4 bxa4
32 ~xa4 lLlf5 33 ~a3 ~xc2 34 'ilVc3
~xa2 35 'ilVxc7 'ilYxd5 0-1

Game 59
Tagansky-Glazkov
Moscow 1975

1 e4 e5 2 lLlf3 lLlc6 3 i.c4 lLlf6 4 d3

d5!?

19 J:l:.xe4 lLlf5 20 SLb3 'iVc5+ 21 ~h2

j,xb2 22 .laxb2 .l:!.ae8 23 SLa4?
It is often difficult to play bad positions
simply because whatever move you investi-
gate, the inevitable outcome will be that you
are worse. Here White commits the common
mistake of not choosing the lesser evil, even
though that can sometimes be very hard to
determine. Better here was 23 c4 ~xe4 24
'iVxe4 lLJd6 25 'iVd3 ~e8 and Black has good
chances of winning the ending thanks to his
more active pieces and superior structure.
23 ... ':'xe4 24 'iVxe4 lLld6 25 'iYd3 'iVf2! Virtually all authors believe this move to
White cannot save the position anymore. be a mistake. The truth is as often otherwise.
26 ':'b4 It is clear to me that if Black can play such a

141
Two Knights Defence

passive move as 4 ... h6 ~rithout being pun- least the position is very complicated. White
ished, he should also be able to play 4... d5. has a whole range of alternatives that should
'Safe sex' is what my coach, master Wojciech be considered:
Ehrenfeucht, called 4 d3. White usually plays a) 7 lLlxeS lLlxeS 8 .l:!.el ~e6 9 .l::txeS
this move in search for a quiet game. 4 ... d5 ~xf2+! 10 ~h1 (not 10 ~xf2? ~f6+) 1O ... c6
declines this suggestion is therefore interest- 11 'iVf3 ..id4 12 .l::te4 ~f6 and the position is
ing for that reason alone. Besides which, about equaL
4 ... dS is not very well known and with new b) 7 'iVe2 0-0 8 lLlxeS lLld4 9 'iVhs ~e6 is
analysis it also very dangerous. I have played unclear. Black has a lot of play for the pawn.
it myself in internet games with short time But he should not fall into temptation and
controls and to the present date I have won play 9...lLlxc2?? because of 10 lLlxf7! ~xf7 11
every game! ~xdS 'iVd7 12 ..ixf7+ 'iVxf7 13 'iVxcs and
5 exd5 lLlxd5 6 0-0 White should win.
White can also choose to attack the e- c) 7 ..ibS! is in my opinion the most an-
pawn at once, but this seems to be very risky. noying move for Black here. The dual threat
After 6 ~e2 i.e7! (not 6... i.g4?! 7 h3 ~hS? of giving a Black doubled pawns or simply
8 g4 i.g6 9 lLlxeS lLld4 10 ~bS+! c6 11 taking the e-pawn cannot be easily dismissed.
lLlxc6+! lLlxe2 12lLlxd8+ 'ltoxd8 13 ~xe2 and After 7... ..ig4 (probably the only move) 8
wins) 7 lLlxeS (otherwise White's last move lLlbd2 0-0 9 ..ixc6 bxc6 10 h3 i..xf3 11 'iVxf3
did not make much sense) 7...lLld4! 8 ~dl f5 12 lLlb3 and White was better in
(the only serious move; not 8 ~e4?? lLlf6 9 Deszczynski-Pinski, Warsaw 1997. In this
..ixf7+ ~f8 10 ~h4lLlg4 11 lLlg6+ ~xf7 and line I want to improve with 11...lLlb4 12 ~d1
White is dead meat, while 8 ~S? is simply a fS 13 lLlb3 ~d6 and though White still looks
waste of time: Black plays 8... ~e6! and the better, Black's position is playable. This
white queen will have to go to dl all the needs practical testing. But for white players
same, as after 9 ~b3 lLlf6 10 ~dl lLlxb3 11 7 ~bS is certainly still my recommendation.
axb3 'iVdS Black has a strong development 7 .. .0-0 8 lLlxe5 'ifh4
and the two bishops for his pawn) 8...i.f6 9
lLlf3 ~g4 and Black has full compensation.
6 ... i.c5

9 l:tf1?
This is already a decisive mistake and
clearly illustrates the dangers of this line. If
This is the critical position for 4 ... dS!? instead 9 ..ixdS i..xf2+ 10 ~h1 ..ixe1
7 l:te1 (10 ... lLlxeS 11 .l::txe5 ..ig4 is also strong) 11
After this move Black looks OK, or at lLlf3 'iVhS 12 ~xc6 ..ig3! 13 ..ie4 i..xh2 and

142
 4 d3

Black is clearly better. 9 ... ttJxe5

The only move was 9 'iVf3 when Black Black is already winning!
now continues 9.. .':t:Jf6 10 tLlxc6 (10 g3? is 10 i.xd5 i.g4?
bad because of 1O ... tLlxe5 11 i:rxe5 ~d4 12 This smooth developing move seems
l:.e1 i..g4 13 'iUf4 l:.ae8 14 l:.xe8 l:.xe8 15 natural, but is actually a blunder. Black can
'iUxd4l:.e1+ 16 ~g2 i..xd4 17 tLlc3 i.d7! and attack with only three pieces as White has no
Black is much better) 10... tLlg4! (not 10... bxc6 representation on the kingside light squares.
11 i.f4 and Black is without compensation; Thus 10 ... tLlg4! and the game is finished, e.g.
time is very important in such a razor sharp 11 i.f4 (or 11 h3 tLlxf2 12 l:!.xf2 ~xf2+ 13
line) 11 d4 ~xh2+ 12 ~f1 i.d6 13 tLle7+ ~h1 i.g4! and kingdom come is near)
i..xe7 14l:.xe7 ~h1+ 15 ~e2 tLlf6 and now: l1...i..xf2+ (l1...tLlxf2 12 '*'f3 i.g4 also wins
a) 16 ~e3? i.g4+ 17 ~d2 b5! 18 i.b3 instantly) 12 ~h1 g5 and Black wins a piece
(not 18 i.xb5? '*'d1+ 19 ~c3 tLld5+ and because of 13 i.xc7 tLlxh2! 14 i.xh2 i.g3
wins) 18 ... c5 19 c3 .l:i.ae8 was played in and mate is imminent.
Konicek-Rybak, corr. 1999, and Black's at- 11 ~d2 J::!.ad8 12 ct::lc3?
tack is probably decisive. The only move Once again a natural developing move is a
now is 20 i.xf7+ l:.xf7 21 l:.xe8+ tLlxe8 22 terrible mistake. This game should probably
~xe8+ l:.f8 23 'iVe1 'iVxg2 24 ~c2l:.xf2+ 25 not be shown to beginners. Instead 12 i.xb7
i.d2 i.f5+ 26 ~b3 .l:!.e2 and the white king is intending 'YWg5 was necessary. Black of
in trouble; after something like 27 'iVh4 course has terrific compensation for the
'*'d5+ 28 ~a3 as the end seems very near. pawns, for example after the natural move
b) 16 ~d2 (much better) 16... i.g417l:.e1 12...l:!.fe8, but White is still alive.
(if 17 'iVf4?! 'iVd1+ 18 ~c3 b5 19 i.b3 as
with a terrible attack in Nolan-Harding, corr.
1989) 17 ... 'iVh2! (if 17 ...'iVxe1+?! 18 ~xe1
i..xf3 19 gxf3 and White has two bishops for
the rook) 18 'iVxb7 -Uab8 19 'iVc6 l:tb6 20
~c5 'iVxg2

12 ... J:!.xd5! 13 ct::lxd5 ct::lf3+ 14 gxf3 i.d6!

15 h3
White has no defence anymore.
After 15 tLlf6+!? (or 15 f4 i.f3 and mates)
15 ...gxf6 16 ~h6 i.h5 17 h3 Black is clearly
better after something like 17 ... ~h8, but
My computer is close to fainting when it strongest is simply 17 ...l:!.e8! where the impo-
sees this position. White is apparently com- tence of the white pieces becomes apparent
pletely winning. However, I am sure that my to all. Black will play 18... ~xh3 19 i.f4 i.xf4
idol Mikhail Tal would bet a bottle of vodka 20 'iVxf4 .l:.e5 and win the queen.
on Black here! 15 ... .th2+ 0-1

143
 Two Knights Defence

15 ... .ixf3 also won, but why not finish in and the game is unclear) 11 .ib3 (11 ~b5
style? SLd6 12lbe4 ~e8 13 SLg5 f6 14 ~h4 a6! 15
1""""---------------__. SLxc6 bxc6 is also unclear) 11...~h8 12 lbe4
Game 60 f5 13 lb g3 SLg6 14 lbxe5 lbxe5 15 ~xe5 f4
Psakhis-Geller and Black had good compensation for the
S ochi 1984 material in Radulov-Spassky, Slavija-Solingen
1984. White will find it difficult to finish his
1 e4 e5 2 l2Jf3 l2Jc6 3 .llc4 l2Jf6 4 d3 i.e7 development, and those of his pieces already
developed have problems finding good
squares.

This quiet developing move is the most

natural and also the most frequently played
(ahead even of 4....ic5). 8 ... ~d6
5 0-0 8... f6! is probably better and if 9 h3 'iVd6 1
Also possible of course is 5 lbc3, but (a new idea) 10 lbbd2 a6 and Black is OK. 9
moves like this are dangerous only for White. l:tel is met strongly with 9....ig4! 10 lbbd2
The game after this move might very well be a6 11 .ixc6 bxc6 and the weakness of
rather boring, e.g. 5... d6 6 h3 0-0 7 0-0 (too Black's pawn structure is compensated by
optimistic is 7 g4?! lba5 8 .Jte3 c6! with swift counterplay against d3 and the slightly greater
counterplay in the centre) 7... lba5 8 ~b3 space. Kutschenko-Wedberg, Copenhagen
lbxb3 9 axb3 c6 10 lbe2 lbe8 11 g4 g6 12 1991, continued 12 h3 .1i.h5 13 lbn c5 14
lbg3 lb g7 and the position is about equal. lbg3 .if7 15 ~e2 l:te8 16 lbfS .ifS and
5 ... 0-0 6 c3 d5!? Black had an excellent position.
This advance is still somewhat risky, but Also possible is 8... .Jtg4 but White can
also gives Black lively play. The safer 6... d6 is then play 9 h3 .1i.h5 (9 ... i.xf3 10 'iWxf3 White
perfectly possible, and will often transpose to is slightly better) 10 g4 i.g6 11 .1i.xc6 bxc6
Game 63 below after 7 ~b3. 12lbxe5 when 'the question of Black's com-
7 exd5 l2Jxd5 8 i.b5 pensation is problematic', to phrase it in the
The best chance to fight for an advantage. words of a politician or his spin doctor. (I am
Instead 8l:i.e1 is met with 8....ig4 9 h3 ~h5 a political scientist myself.) The translation
10 lbbd2 (too risky is 10 g4 SLg6 11 lbxe5 would be something like this: Black is a pawn
lbxe5 12 l:txe5 lbb6 13 SLb3 i.f6 14 l:te1 down, has problems with his ruined pawn
~xd3 15 ~xd3 ~xd3 when Black has supe- structure, and no real counterplay.
rior development and structure) 10 ... lbb6 (or 9.l:!.e1
1O ... lbf4 11 lbn lba5 12 .ixf4 exf4 13 SLb5 White should not go after the e-pawn with

144
 4 d3

9 3i.xc6 bxc6 10 l:.e 1, as Black would react 22 .l:i.e7 \X!hite has the initiative in the end-
with great pace and be fIne: 10... 3i.g4 11 h3 game.
i.h5 12 g4 (if 12 CDbd2 f5 with unclear play) 20 ~f3 f5?
12... i.g6 13 CDxe5 ~h4 14 ~g2 i.xe5 15 This weakening of the g5-square is now
l:txe5 f5 and Black has full compensation for tactically flawed. The idea, of course, is to use
the pawn in the form of a terrifIc lead in the hook of white g4-pawn in order to create
development and a very weak white king. open lines to the white king, but in real life it
9 ... .tg4 10 h3 .th5 11 CDbd2 'ilih8?! does not work out like that, as \X!hite is able
This attempt at a pawn sacrifIce does not to complete his development with gains of
work out, as \X!hite can also choose simply to tempi. Preferable was 20 .. .lHe8 21 i.d2 ~b6
develop his forces and thereby get a slight 22 .l::!.adl i.c2 23 .l:!.c1 i.g6 and Black per-
pull. Better were either l1...f6 12 d4 exd4 13 haps stands slightly worse, but nothing more.
CDe4 CDe5 14 g4 CDxf3+ 15 'iUxf3 ..to and
Black keeps the balance, or l1...CDb6!? trans-
posing to a sub-line (11 3i.b5) to 8 .l::!.e1 in the
notes above.
12 g4 .tg6 13 CDe4 f6 14 d4

21 .tg5!
Black cannot take on g4 because the
queen is en prise after 21...fxg4 22 3i.xd8!.
21 .. Jl:d7
Black is clearly worse. Also after 21...CDf6
\X!hite now stands slightly better. His pres- 22 gxf5 ..th5 (the tactical 22 .. J:td5 23 i.xf6
sure in the centre is a little uncomfortable for .l:!.xf5 fails to 24 'i¥xf5! 3i.xf5 25 .l::!.e7 'iUf4 26
Black to meet, and now he even snatches the iLxg 7+ ~g8 27 ..ixf8 ~xf8 28 .l::!.e3 and
bishop pair. \X!hite should win the ending) 23 'iUg 3 'iUd7
14 ... exd4 15 CDxd6! 24 .l::!.e6 and \X!hite stands much better.
\X!hite goes for the bishops. After 15 22 :tadl CDf6?!
iLxc6?! bxc6 16 CDxd4 'iUd7 Black would This allows 'W'hite to simplify to a position
have good attacking chances. that requires only technical accuracy. Better
15 .. :~xd6 16 CDxd4 CDxd4 17 ~xd4 c5 was 22 ...CDb6 and Black is still alive. After 23
18 'ilt'dl .l:!.ad8 19 .tfl! gxf5 .l::!.dO 24 f6 gxf6 25 3i.h4! \X!hite is much
The bishop is transferred to g2 where it better, but still there is hope of some compli-
will not only protect the white king, but also cations. Note that 25 ..tf4? walks into the
create strong pressure on the long diagonal, trap 25 ... i.h5! and Black is OK.
making it diffIcult for Black to operate freely. 23 .txf6 J:txf6 24 .l:!.xd7 'ilt'xd7 25 J:!.dl
19 .. :~Vc7 YWe8
After 19 ... f5 20 3i.g5 CDf6 21 'iUxd6 ':xd6 25 ... nd6 is met simply by 26 .l::!.xd6 'iUxd6

145
Two Knights Defence

27 ~xb 7 and White has a '-'inning endgame. This move order is designed to meet
26 'iYxb7 fxg4 27 i..bS! 6... d5. By delaying c2-c3 White does not have
a weakness on d3 to bother about after
6... d5, while on 6... d6 he will play 7 c3 after
all to save the bishop from the black knight.
Black can now choose to be active or pas-
Sive.
6 ... dS
Against careful play Black plays aggres-
sively! Of course Black can very well play the
calm 6... d6 (see Game 63), but it does not
guarantee equality.
7 exdS
The only try for an advantage. If 7 liJbd2
dxe4 8 dxe4 .i.c5 and the position is equal.
27 .. :YWg8 7...ct'lxdS 8 J:l:e1
Black has few options now. If 27 .. :~f8 28 The prophylactic 8 h3 is considered in
'iVc7! .l:!.e6 29 hxg4 and White should win. Game 62.
28 hxg4 h6 29 'iYxa7 ri:f4 30 i..d7 VoiVf8 8 ... .tg49 h3
31 VoiVc7 'iff7?! This move is more or less obligatory and
This loses by force. But 31....l:!.xf2 32 Si.f5! gives Black an interesting choice between
.l:!.e2 33 l:!.d8 .te8 34 ~d6 was little im- two perfectly sound options.
provement.
32 'iYc8+ 1-0
After 32 ...<;t>h7 33 .te6 wins, while 32 .. .'~f8
is a lost ending.

Game 61
R. Perez-Gild. Garcia
Santa Clara 1996

1 e4 eS 2 ct'lf3 ct'lc6 3 .lii.c4 ct'lf6 4 d3 i.e 7

S 0-0 0-0 6 i.b3

9 ....thS!?
This move is for players who like compli-
cations, particularly if they are also fans of
the Marshall Attack. 9....txf3 is normally
considered the main move here, and it is also
perfectly fine. After 10 'iVxf3 liJd4! (the key
idea) 11 'i¥e4 (not 11 'iVxd5? ~xd5 12 ..Itxd5
liJxc2 13 Si.d2 liJxa 1 14.l:!.c1 c6 15 .tf3 .l:!.ad8
16 liJc3 .l:!.xd3 and Black had a clear advan-
tage in Dizdar-Mikhalchishin, Zenica 1989)
l1...liJxb3 12 axb3 liJb4 13 liJa3 liJc6 and

146
 4 d3

there is no reason in the world why Black to exchange a bishop which has already lost
should be worse here. its scope (after the pawn advanced to f4).
10 g4 Better was 17 lDc3 but still the position is
Of course. extremely dangerous for White: 17 ... lDxc3
10 ... ..tg6 11 liJxe5 liJxe5 12 J:.xe5 c6 13 (N ogueiras gives 17 ... tLJe 7 18 h5 ~f5 19 tLJe4
iif3 tLJd5 and Black has the initiative) 18 bxc3
13 ~xd5 is best met by 13 ... ~d6! 14 .l:i.el h6!? (Black can also play more calmly with,
cxd5 with perfect compensation for the for example, 18... 'iVd7) and now White
pawn. should be very careful. If 19 gxh6? 'iVxh4 20
13 ... ..td6 14 'u'e2 f5! hxg7+ ~xg7 and in my opinion only a truly
forgiving God would be able to save White.
The black rooks will go to h8 and f5-g5.
17 ... iid71SliJc4?
18 tLJf1 is better, but White is still under a
lot of pressure. One idea is simply 18... .tf5!?
(intending ... ~g4) 19 tLJh2 .l:i.ae8 and Black is
for preference, though the game is not de-
cided.

Black takes over the initiative with this

move, which is actually an avalanche of pawn
sacrifices. But these are too dangerous to
accept. White needs to get his pieces into
play.
15 g5
a) 15 ~xd5+ is too dangerous. White im-
mediately loses all control over the light
squares: 15 ... cxd5 16 'iVxd5+ ~h8 17 tLJc3 1S ... ..th5!!
(not 17 .l:i.e6?? ~h2+) 17... .l:i.c8 18 'iVg2 ..te8 A fantastic decoy sacrifice. White has no
19 tLJd5 'iVh4 and the position is unclear defence now.
according to Nogueiras. Black seems to have 19 'iVxh5 iih3 20 J:.e4
enough compensation for the pawns. Everything loses here. If 20 lDe5 f3 21
a) 15 tLJc3! fxg4 16 'iVxg4 is very compli- 'iVg4fxe2 and Black has won material, or 20
cated. White is a pawn up but he is lacking in g6 h6 21 tLJe5 f3 etc., or 20 tLJxd6 f3 and
pawn cover for his king. Nevertheless, if he White must part with the queen.
is to have the advantage in any line after 20 ... g60-1
9... ~h5!? it is probably here. Only practical
tests or a month of isolation with a Pentium Game 62
5 PC could give a good hint about the true Dubiel-Macieja
evaluation of this position. Biala Podlaska 1994
15 ... f4 16 h4 WhS 17liJd2?!
White begins a horrible knight manoeuvre 1 e4 e5 2 liJf3 liJc6 3 ..Itc4 lZJf6 4 d3 ..Ite7

147
Two Knights Defence

5 0-0 0-0 6 ..ib3 d5 7 exd5 ltJxd5 8 h3 tbf4 13 .lixf4 .l:!.f6 14 'iV e2 ':xf4 when Black
White is so scared of 8... ~g4 that he de- had two bishops, an active rook and there-
cides to prevent it. But it is hard to imagine fore clearly the brighter future in Berezjuk-
that Black should be worse after such a cau- Ulak, Frydek i\1istek 1996.
tious move. 10 ... ltJb6 11 ..ib3
11 c3 is again met with swift action:
11...tbxa4 12 'iVxa4 tbe2+ 13 ~hl .if6 14 d4
~xh3! 15 gxh3 'iVdS+ 16 ~h2 ~xeS+ 17
dxeS tbxc1 18 .l:!.xc1 'iVxeS+ 19 ~h I! (not 19
f4?? ~e2+ 20 ~hl ~f3+ 21 ~h2 .l:tae8 and
Black wins) 19...~dS+ 20 ~h2 with a direct
draw. Black can also try 17 ... ~xeS+!? as in
Shirov-Mozetic, Tilburg 1993, when after 18
f4 (best) 18 ... 'iVhS 19 .ltd2 'iVg6 20 .liel
.l:!.ad8 21 'iVbs Md3 Black has sufficient com-
pensation according to Mozetic. He also
notes 14 'iVe4 tbxc1 15 .l:!.xc1 SLxeS 16 ~xeS
~xd3 17 ~xc7 .ixh3 18 'ilVg3 ~xg3 19 fxg3
8 ... a5! without assessment. Does he think the posi-
In my opinion this is the best move. The tion is equal or does he just want that readers
alternatives are: of Chess Informant to think so? The truth is
a) 8 ... fS? 9 tbc3 ~e6 10 tbxeS! tbxeS 11 that Black stands much better! He has rook
Mel 'ilVd6 (or 11....i.f6 12 d4) 12 tbbS ~d7 and bishop vs. rook and knight, and a supe-
13 .l:!.xeS and White has a clear advantage. rior structure.
b) 8... .i.f6 (with the idea of ... tbaS, but this 11...a4!
does not really work out) 9 .l:!.e 1 .1i.e6 10 These are necessary tactics. After
tbbd2 tbf4 11 tbe4 (stronger than 11 .1i.a4 11...tbxb3 12 axb3 tbdS 13 .l:!.el SLd6 14 tba3
tb g6 12 .1i.xc6 bxc6 13 tbe4 .1i.e 7 and the 'iIVf6 15 tbac4 bS 16 tbxd6 cxd6 17 tbf3 ~b 7
position offers both players good chances 18 tbgs Black does not have enough com-
according to Dolmatov) 11....ixb3 12 axb3 pensation for the pawn.
and White has some advantage. 12 ltJxf7
9..ia4?!
As we shall see there is a purely tactical
reason why this move does not work. And
having played 6 SLb3 and 8 h3 it would be
strange if White was able to refute a move
like 8... aS.
Better here was 9 a3 a4 10 .ia2 ~h8 11
.l:!.el f6 12 d4 exd4 13 tbxd4 tbdb4 14 axb4
'ilVxd4 15 c3 'ilVxdl 16 .axdl SLfS 17 tba3
tbeS 18 tbbS cS 19 bxcS SLxcS 20 tbd4 .l:!.fd8
and the game was equal in Kramnik-
Kasparov, New York 1995.
9 ... ltJd4! 10 ltJxe5
After 10 c3 tbxf3+ 11 'iYxf3 .l:!.a6! Black 12 ... axb3!!
rapidly develops the initiative; e.g. 12 SLb3 The big idea behind it all. Black gets three

148
 4 d3

minor pieces for the queen. Black had difficulties in converting a better
13 tUxd8 bxc2 14 ~e1 cxb1'iIY 15 '!:'xb1 position to a \Ninning position. White should
i..xd8 sit tight with 28 ~e 1.
In the middlegame three minor pieces are 28 ... tUxe3 29 fxe3 tUd5
nearly always stronger than the queen. The Now Black is much better - actually it is
two extra white pawns are not so important practically a winning position.
right now. More significant is it that Black 30 !:te2 .!:.e7 31 J:l:de1 h6!
has nice play and, as we shall see, White lacks Remember to play such moves as this .
the time to get his pieces to good squares. ... h 7-h6 is both useful and safe with several
16 i..e3 i..f6 17 ~b4 .l:td8 18 ~fe1 tUd5 minor functions, but most importantly it
Not 18 ....l:Ixa2?? 19 i..xd4 i..xd4 20 ~3+ throws the ball back to White, who must
and White wins. then ponder over what to do with his posi-
19 'iVc4 tUc6 20 d4?! tion.
This restricts the bishop and creates a 32 e4!?
weakness. Much better is 20 i..f4 lbaS (not This move seems necessary because of
20 ... 'it>h8?? 21 ~xdS) 21 ~e4 'ua6 with an tricks with ...lbxe3, but still it is unpleasant to
unclear position. play.
20 ... i.f5 21 li!.bd1 'iioh8 22 a3 tUb6 23 32 ... tUf4 33 e5 tUxe2+ 34 ~xe2 Ji.g5 35
~c3 J:id7! d5 J:ted7 36 e6 l:txd5 37 ~xc7 b5 38 e7
Now we understand why 20 d4 was a .l:!.e8 39 ~c6 i..f7 40 ~f2?
weak move. Black has consolidated his posi- A mistake in time trouble. But the posi-
tion. So now White has problems, not least tion was lost anyway.
that now he can only wait. Active play will be 40 ... i..e30-1
punished.
24 li!.d2 ~ad8 25 ned1 i..g6 26 b4 tUd5 Game 63
27 ~c5 tUce7 Gelashvili-Gokhale
Dubai 2002

1 e4 e5 2 tUf3 tUc6 3 i..c4 tUf6 4 d3 i.e7

5 0-0 0-0 6 i..b3 d6 7 c3

28 g4?
The skill of suffering patiently is worth at
least a 100 Elo points. I will quote the first
World Champion Wilhelm Steinitz: 'If you
have the advantage you should attack, if your This is one of the main positions in the
position is worse you have to play defensive Two Knights. It is a kind of odd Ruy Lopez
moves.' Without the weakness at g4 to attack without the pawn moves ... a7-a6 and ... b7-bS.

149
Two Knights Defence

The position should objectively be more or bxc6 11 tLJxd4 c5 12 tLJxe6 fxe6 13 'iVb3 lIf6
less equal, but in positions like these the best 14 f4 with a slight edge for White in Yudasin-
player will win in almost 100% of cases. Klovans, Kostroma 1985.
Robert Fischer said that in the Sicilian d) 7... tLJd7 8 .i.e3 tLJc5 9 JLxc5 dxc5 10
Dragon a 2000 player can hope to beat a .i.d5 .i.d6 11 JLxc6 bxc6 12 tLJbd2 i.a6 13
GM. Here it is impossible. In positions 'iYc2 and White was better in Kramnik-
where there are no direct tactics or obvious Meister, Kuibishev 1990.
attacking moves weaker players will fail to e) 7...~h8 8 tLJbd2 tLJg8 9 d4 i.f6 10 dxe5
place their pieces well and will quickly get tLJxe5 11 tLJxe5 dxe5 12 tLJc4 JLe6 13 tLJe3
into trouble. This game is an illustration of 'iYxdl 14 .:!.xdl J:!.ad8 15 Mxd8 .:!.xd8 16
how simple moves can 'WID simply by being .i.xe6 fxe6 17 ~f1 and White is slighdy bet-
more natural. ter in this ending according to Magomedov.
7 ... h6 8 lLlbd2 lLlh7?!
Usually this kind of move looks like a be-
ginner's move. If 7... h6 is played to prepare
....:!.e8 then it makes perfect sense, but
Black's idea in the game is not very good.
Still almost anything can be played here.
Others:
a) 7 ... iLg4 (in my opinion this is not as
good as Black's other options; ... .i.g4 should
only be played when the white pawn is on d4
and the centre is fluid) 8 tLJbd2 tLJd7 9 h3
iLh5 10 .i.c2 tLJc5 11 g4 (this is one of the
possibilities White gets against an early
... iLg4; another is l:i.el, tLJf1-g3, though that
is a bit slow here) l1...iLg6 12 d4 exd4 13 This idea seems to both too slow and
cxd4 tLJd7 and after 14 d5 the game was very positionally unjustified. If .. .17- f5 the squares
unclear in the game Hjartarson-Piket, Manila around the black king will be weak, while
1992. My recommendation is keep control ... tLJh7-g5 loses a lot of time merely to ex-
over the centre with 14 a3 when White is change pieces of even value. 8....l:.e8 is the
litde bit better. better move here, vacating ffi is for the
b) 7... tLJa5 8 .i.c2 c5 9 tLJbd2 tLJc6 10 l:i.el bishop. After 9l:i.el .tffi 10 h3 JLe6 11 iLa4
Me8 11 tLJf1 h6 12 h3 JLffi 13 tLJ g3 JLe6 (af- iLd7 12 tLJf1 tLJe7 13 i.c2 tLJg6 14 d4 c5 15
ter 13. .. d5 14 exd5 tLJxd5 15 i.b3 i.e6 16 tLJ g3 cxd4 16 cxd4 l:i.c8 the position was
i.a4! White keeps some pressure) 14 d4 cxd4 about equal in Gelfand-Onischuk, Gronin-
15 cxd4 exd4 16 tLJxd4 tLJxd4 17 'ilVxd4 tLJd7 gen 1996.
18 iLf4 with a slight advantage to White in 9 lLlc4 ~f6?!
Kobalija-Atalik, Istanbul 2003. If Black wants to place the bishop on f6, it
Also interesting is 9 b4!? cxb4 10 cxb4 was better to play 7... tLJd7. Now the knight
tLJc6 11 b5 tLJa5 12 d4 (or 12 h3 iLd7 13 a4 on h7 is unemployed. Instead if 9... tLJg5 10
VJlic7 14 i.b2 and the game is unclear) tLJxg5 i.xg5 11 f4! exf4 12 JLxf4 iLe6 13
12... exd4 13 tLJxd4 iLd7 14 tLJd2 l:i.c8 15 tLJe3 i.xf4 14 .l:!.xf4 and White is slighdy
JLb2 tLJg4 16 a4 i.f6 with active play for better.
Black in Kramnik-J .Polgar, Moscow 1996. 10 lLle3lLle7 11 h4!
c) 7... i.e6 8 .i.a4 tLJd7 9 d4 exd4 10 JLxc6 This is a strong prophylactic move, pre-

150
 4 d3

venting .. .'~jh7-gS. Now all the black pieces are misplaced

11 .. .~eS 1 2 g3 cS White opens the game to exploit it with di-
rect tactics. The game suddenly becomes
very concrete, but no matter what Black
does, the lines do not work in his favour.
19 ... exd4
If 19 ..."iYc7 20 ~b3 .ie6 21 dS cxdS (or
21...~d7 22 dxc6 .ixc6 23 .ie3 and White is
much better) 22 exdS .i.d7 23 iLc2! with an
attack.
20 cxd4 ~a5 21 ~e3 h5?
This looks bad, but Black has no easy
choices here. After 21 ... dS 22 eS i.xfS 23
~xfS White retains strong pressure. Probably
best was 21 ... .l:tad8 waiting for a better future
13 liJh2! and allowing White the chance to mess up
White regroups his pieces based on the the attack.
weakened light squares. The obvious idea is 22 l2lh2 l2lfS
~f3, ctJh2-g4 and ctJe3-fS with an attack! The only way to defend the hS-pawn.
13 ... J:.e8?!
Black is just waiting here. Better 1S
13 ....ixb3 14 axb3 dS with the idea ... ~c7,
...lIfd8 and the knight on h7 can rejoin the
struggle via the f8-square.
14 ~c2
White decides to keep this bishop for
later.
14 ... .th3 15 .l:!.e1 liJgS 16 Wif3 .td7 17
l2lf5

23l2lhS+!
Now the tactics start to arise - as they al-
ways do when the pieces are well placed.
23 ... gxh6 24 'YIixf6 .l:!.eS 25 ~f3 ~g7 2S
a3!
There is no reason to rush things. Black's
game is a positional ruin which he fmd very
hard to improve.
26 ... ~b5 27 .:tab1 ~c4 28 ~d1! 't1e7 29
i.d3 ~eS 30 ~xh5
White stands much better. Black has no Now besides having a clearly better posi-
counterplay and his minor pieces have no tion, White has an extra pawn too.
way of getting active. 30 ... <;t>h7 31 f4 ~b3 32 .l:Ibd1 f5 33 exf5
17 ... .te7 18l2lg4 .tf8 19 d4! l2lxf4 34 gxf4 l:txe3 35 l:txe3 ~xd4 1-0

757
Two Knights Defence

Resignation was a sensible decision. After

36 f6+ mate comes in no more than four
moves.

Game 64
Kovchan-Malaniuk
Swidnica 1999

1 e4 e5 2 lbf3 lbc6 3 ~c4 lbf6 4 d3 iL.e7

5 0-0 0-0 6 lbbd2 d6 7 a4

In positions like this Black has only two

plans and both involve pawn breaks: to play
... d6-d5 or ... f7-f5. In most cases .. .f7-f5 is
better. First of all the pawn because the e5-
pawn retains its defence, and secondly, Black
can make good use of the f-flle for attacking
purposes.
9 ... lbgS
Interesting is also 9...lbh5!? This active
move is possible because 10 lbxe5?? fails to
The white play here differs somewhat 10 ... lbxe5 11 '*IVxhS .ig4 and the queen is
from the Ruy Lopez, as White goes for a lost. Therefore White should choose be-
space advantage on the queenside straight tween 10 .l::i.el lbf4 with the initiative, and 10
away, while also preserving his bishop against d4 exd4 11 cxd4 lbf4 12 lbb3 fS! 13 eS dxe5
exchange by .. .'~ba5. (13 ... d5!? is completely unclear) 14 dxeS .ie6
7 .. .'>¥ihS 15 i.xe6 lbxe6 and the ending is more or
Intending ... lbgs and .. .0-f5 with coun- less equal.
terplay. Also possible is 7...~e6 Sl'.Iel '*IVd79 10 d4 exd4
c3 lUeS 10 .ib3 i.f8 11 as i.xb3 12 '*IVxb3 This is a good sound move, but Black
a6 13lbn d5 14 .ig5 dxe4 15 dxe4 '*IVe6 and could also play very energetically with
position is equal, Grosar-Gostisa, Slovenia 10 ... fS!? when White has two main moves:
1994. Inadvisable is S... ~xc4?' forcing White a) 11 dxeS dxeS (remember this! - when
to complete his development by 9 lbxc4 and the white knights are on f3 and d2 and
then \X/hite is slightly better according Mi- thereby taking each others squares, and
chael Adams; put simply, Black exchanged White at the same time has less space, Black
his good bishop and his position is now solid should not exchange pieces) 12 ..lidS!? i.f6
but rather passive. 13 i.xc6 (removing the knight before Black
S a5 a6 supports it with ... lbg e7) 13. .. bxc6 14 ~e2 f4
The alternative S... .l:!.bS!?, intending ... b7- 15 b3 g5! 16 lbc4 (if 16 h3 g4 17 hxg4 i.xg4
b5, is an interesting new idea, which can be with unclear play) 16 ...'iYe8 17 ..Ita3 J:[,f7 18
compared with 6 l:!:e 1 d6 7 a4 ~h8 8 as l:.b8 l:!:fd 1 g4 19 lbe 1 l::.g7 and here Black has the
in Game 66. initiative while White has a better pawn
9 c3 structure. In positions like this I prefer play-

152
 4 d3

ing the black pieces, but this is a matter of 12 WVb3 fxe4 13 CLJxe4 is probably mostly
taste; White has his own chances. Except strongly met with 13 ... .l:!.xf3! (also possible is
that in blitz games - which we all seem 13. .. CLJf6 14 CLJxf6 i.xf6 15 d5 CLJe5 16 ~e2
mainly to play these days - attacking is much ~e7 and question is whether White can keep
more effective than defence. Of course this is the balance or not) 14 gxf3 (if 14 'iVxf3 d5 15
not true if you are Petrosian arisen from he CLJg5 ~xg5 16 ~xd5 'iVxd5 17 i.xd5 ~f6
dead for one last round in the ring. and Black is better) 14... CLJxd4 15 ~d3 CLJc6
Black has also some compensation after with excellent play for the exchange. Most
11...fxe4 12 exd6 ~xd6 13 CLJxe4 ~g6 14 attacking players would prefer Black here.
CLJeg5?! ~xg5 15 CLJxg5 h6 16 CLJf3 i.g4 17 Fritz 8 thinks the position is quickly 0.00 -
i.d3 'iVh5 and the pressure will get stronger, sometimes I wonder how it is that it often
or if 14 'iVd3 ~f5 15 Mel Mad8 with coun- gets to 0.00 in messy positions - but then
terplay. But this is of course quite compli- after some time White declines to -0.03,
cated and risky; additional investigation of which of course is basicall," the same.
the position can be recommended for those 12 ... dS 13 i..d3 f4! 14 CLlb3 i..g4 1S h3
with enough time on their hands. This does not look good, but is necessary.
b) 11 d5 CLJb8 12 exf5 ~xf5 13 l:tel CLJf6 Black has a very simple plan: ... 'iVd7 and
14 ~3 'iVc8 (also strong is 14... CLJbd7!? 15 ... CLJd8-e6 with a superior position.
'iVxb 7 CLJc5 16 ~4 ~d3 17 ~a2 ~g6 and 1S ... i..hS 16 g4
Black has good compensation for the pawn; White has no choice. After others move
maybe 15 CLJg5 is an improvement in this he runs the very likely risk of losing without a
line) 15 CLJd4!? exd4 16 l:txe7 d3! and Black's fight.
chances looks good, although the position is 16 ... fxg3 17 fxg3 ~e8 18 i..d2 CLld8!
very unclear; e.g. 17 CLJf3CLJbd7 18 ~g5CLJc5
19 iVdlCLJfe4 with a complex struggle.

Black executes his simple plan. A blockad-

ing knight has been famous ever since
11 cxd4?! Nimzowitsch wrote Blockade almost a century
After this the pawn will be weak on d4, ago, and it now forms part of the basics for
and Black \vill be able to put up a strong any serious chess player.
blockade of the centre while attacking on the 19 'it'g2 i..g6 20 "YI:Yc2 il..xd3 21 ik'xd3
kingside. Better was 11 CLJxd4CLJxd4 12 cxd4 CLle6 22 h4?
f5 13 e5 dxe5 14 dxe5 'iVd4 15 ~el i.d7! 16 Moves like these are always bad. Remem-
'iVc2 ~c6, though Black still has good play. ber the Steinitz quote earlier in this chapter!
11 .. .1S 12 eS Here \x'hite wants to take control of g5 and

153
Two Knights Defence

play 4Jf3-g5 with some offensive ideas. But 4Jb4 ':xfl 34 .i:Ixfl ':c4 35 4Jxd5 'iVxd4 36
this is unrealistic, while the weakening of the ~xd4 ':xd4 37 4Jxb6 4Jxe5 38 ':f4! White is
light squares is very real. Instead, after the able to make a draw, if only just; e.g.
solid 22 .l:lf2! White can think about the fu- 38 ... .i:Id2+ 39 l:!:f2 ':xf2+ 40 ~xf2 4Jd3+ 41
ture. ~e3! 4Jxb2 42 ~d2 'it'h7 43 ~c2 ~g6 44
22 ... 'ilVh5 ~xb2 ~f5 45 4Jd5 ~g4 46 4Je3+ ~xg3 47
The light squares around the white king 4Jf5+ and draws.
are going to create a (k)nightmare. 31 ... e5! 32 ~xf8+
23 .\tg5 J:.ae8 Or 32 dxc5 d4 33 ~d2 ':xfl 34 l::txfl
If 23 ... ..ixg5 24 4Jxg5 4Jxg5 25 hxg5 :fxe5 and Black wins.
~xg5 26 4Jc5 and White has some compen- 32 ... J:.xf8 33 J:l:f1 ':'xf1 34 ~xf1 exd4 35
sation for the pawn, though Black is still 'iVe2 'iVf5+ 36 'iitg2 bxa5 37 e6 d3 38 e7
better after 26 ... ~g4. 'Dxe7 39 'Wixe7 d2 40 ~e8+ 'iith7 41
24 .iLxe7 'Dxe7 25 'Dg5 'iVe2 ~e2 0-1
If 25 ':ac1 4Jf5 26 4Jg5 ~g6! and the un-
protected queen gives Wrute problems; after Game 65
27 ~gl 4Jxg5 28 hxg5 'iYxg5 and Black is Macieja-Ivanchuk
close to winning. Hyderabad 2002
25 ... 'Dxg5 26 hxg5 'ilVxg5 27 'De5 'iVg4
28 l:lh1 1 e4 e5 2 'Df3 'De6 3 .te4 'Df6 4 d3 .te 7
Not 28 4Jxb7 4Jg6 29 4Jc5 ':f4! and 5 0-0 0-0 6 J:.e1 d6 7 a4
Black wins. This plan seems more flexible than 6
28 ... 'Dg6 29 l:laf1 h6? 4Jbd2. The knight can also move to the king-
Black is wasting time on luxury moves. side via a3-c2-e3.
His chances were here right now and should 7 ... 'Dd4!?
have been milked. The right path was
29 ... :ff4! 30 :fxh7+ ~g8! 31 ':h2 Mxd4 and
Black wins.
30 'iVe3! b6

A double edged move. More natural IS

7... ~h8 which is considered in Game 66.

8 'Dxd4 exd4 9 'Dd2
Also possible is 9 c3 dxc3 10 4Jxc3 c6 11
31 'Dxa6? 'iVb3?! (better is 11 d4 with equality
A tactical error which is easy to under- 11...4Jg4! 12 d4 ~h4 13 g3 ~f6 14 :fe2 ~g6
stand, as White's defensive task is extensive. 15 ~h1 ..if6 and Black's initiative was very
Instead, after 31 axb6 cxb6 32 4Jxa6 :fc8 33 dangerous in Nevednichy-Tseshkovsky, Igalo

154
 4 d3

1994. After 16 f3 .Jtxd4 17 fxg4 .Jtxg4 18 or later) 13 CDxe5 dxe5 14 ~h5 .Jtf6 15 f4
.l::tg2 .Jtf3 Black kept the advantage. ~c7 16 f5 and White is much better.
9 .. .tlJg4?! 12 .. .'~xh4 13 i.f4 llJe5 14 i.g3 'ile7 15
The knight is not very well placed here. b3
Better was 9... .Jtg4!? 10 f3 .lte6 11 .ltxe6 fxe6 15 ..I1.d5 is answered by 15 ... i.g4!? 16 f3
12 e5 (if 12 f4 d5 13 e5 CDe8 14 ~g4 ~d7 15 .Jte6 blocking the dl-h5 diagonal with equal
CDf3 c5 16 b4 cxb4 17 CDxd4 ..I1.c5 18 .Jte3 play.
CDc7 and the game is more or less equal) 15 ... llJxc4 16 bxc4 i.d7
12...CDd5 13 CDc4 and Macieja believes that Better 16 ... f5! 17 exf5 i.xf5 18 .l:!.el ~d7
White has a small advantage here. However, and the position is equaL
with a strong knight on d5 and the half-open 17 a5 ~c6?!
f-file as well as the weak dark squares in the Black is wasting too much time. Better
white camp, Black has good reasons to hope again was 17 .. .f5 18 exf5 ..I1.xf5 19 .l:!.e 1
for good counterplay: 13 ... dxe5 14 l:!.xe5 (or though Black must play very carefully in or-
14 CDxe5 .Jtg5) 14... ~d7 15 ~e2 .l:!.ae8 16 der to keep the balance: 19 ... .lte6! (the most
l:!.xe6 i.g5! 17 .l::te5 .Jtxc1 18 .l:!.xc1 ~xa4 and important thing is to control the f3-square; if
Black is OK. 19 ...'iVd7 20 a6 b6 21 ~f3 and White is
10 llJf3 i..h4?! slightly better) 20 ~d2 .l:!.ae8 with equality.
Better is 10 ... c5 when White has only a 18 ~g4 J:tae8 19 'ifif4 'ifie5 20 'ilg4 ~e6
slight advantage. 21 h3?!
11 litf1! Better was 21 f3 and \Xlhite retains a small
Black has was hoping to provoke 11 g3?! edge according to Macieja.
when after 11...i.f6 12 as c6 13 i.b3 .l:!.e8
the light squares around the white king are
quite weak, so Black is OK.
11 ... c5

21 ... f6
After 21...~xg4?! 22 hxg4 .l:!.d8 23 g5 .l:!.d7
24 .l:!.fel .l:!.e8 25 f4 White has full control
over the position. Perhaps a draw is a realistic
12llJxh4? hope, but Black should be prepared for 50
Why White wants to bring the black moves of hard fight in order to survive.
queen to h4 is not clear. Instead 12 h3 and Black could still have tried 21...f5!? and after
there is no comfort for Black in 12...CDe5 22 exf5 .l:!.xfS 23 .Jtxd6 h5 24 ~g3 'iih6! 25
(even worse is 12... .txf2+? 13 .l:!.xf2 CDxf2 14 .l:!.ael! (if 25 'ti'h2 .l::te2 with a strong attack)
~xf2 ..I1.e6 15 i.d5 ..I1.xd5 16 exd5 'iVf6 17 25 ... .l:!.xel 26 .l:!.xel .l:!.g5 27 'ti'f4 .l:!.xg2+ 28
~d2 and White will win this ending sooner ~f1 the game is completely unclear.

155
Two Knights Defence

22 iYf4 iYe5 23 ~d2 'ilYe6 24 Vif4 iYe5 it is Black who has the winning chances!
25 ~d2 't\Ve6 26 il..h2?! 33 ... l:th7 34 'iitf2 'ilYf7
Better was 26 J:.ael and White has a slight 34... b5 is of course risky, but quite neces-
advantage. sary. White is much better after 35 cxb5
26 .. JU7?! .ixb5 36 f4, but Black has some real chances
Again 26 ... f5! leads to equality. Ivanchuk to survive and that is what matters.
seems to have made a conscious decision not 35 J:th1 'ilYg7 36 l:t.xh7 ~xh7 37 ~c1 'iitf7
to consider this move. 38 .l:.h1 ~g7 39 g5! 'iite8
27 J:tae1 ~d7 If 39 ... fxg5 40 'iYxg5 'iYf6 41 Mh7+ ~g8
If 27...l:tfe7 28 g4! and White maintains 42 'iUh6 (with the deadly threat of ~h4)
strong pressure on the black position. 42 ...g5 43l:th8+! and White wins.
28 g4 40.:th6
Now the attack is killing.
40 ... ~d7 41 ~h1 .l:!.e7 42 gxf6 'ilYxf6 43
i.h4 ~f4 44 il..xe7 'iitxe7 45 ~h4+ 1-0

Game 66
Macieja-Pinski
Polanica Zdroj 1999

1 e4 e5 2 l2Jf3 l2Jc6 3 i.c4 l2Jf6 4 d3 il..e7

50-00-06 l:te1 d6 7 a4 \t>h8

28 ... h5?!
Old rules says that if your opponent is at-
tacking on the flank you should play in the
centre; if that is impossible, you should pre-
pare counterplay on the other side of board.
Thus 28 ... b5! 29 axb6 axb6 30 l:tbl b5 31
cxb5 ~xb5 32 ~xd6 ~xd3 33 cxd3 liVxd6
34 l:tfel and now White has switched to the
queenside, then 34... h5! with counterplay on
the other wing.
29 ~f4 hxg4 With similar ideas as after 6 CDbd2 d6 7 a4
29 ... .ia4! is best met with 30 .:tel! and ~h8 in Game 64.
White keeps the advantage. But not 30 8 a5
'iYxd6? liVxd6 31 .ixd6 ~xc2 32 ~xc5 Or 8 c3 CDg8 when White has tried:
~xd3 and Black wins material, or 30 l:i.e2? a) 9 d4 ~g4 10 dxe5 (if 10 d5 CDb8 11 h3
hxg4 31 hxg4 f5! ""ith terrific counterplay. ~c8! and Black has good counterplay com-
30 hxg4 J:re6 ing with .. .f7-f5; this is better than l1...~xf3
30 ... ~a4!? was still possible. 12 'iUxf3 ~g5 13 as a6 14 CDd2 CDe7 15 'iVh5
31 i.g3 ~e8 32 ~d2 g6 33 f3 when White's greater space gives him the
Not 33 f4? f5! 34 exf5 gxf5 35 Mxe6 'iYxe6 advantage) 10... CDxe5 (worse is 1O ... dxe5?! 11
36 g5 liVe3+ 37 liVxe3 dxe3 38 l:tel l:te7 and CDbd2 ~d6 12 as a6 13 llVb3 liVd7 14 h3

156
 4 d3

3l.h5 15 ctJh4 ctJge7 16 ctJf1 and White was White, as usual, was offered the central
better in Armas-Adams, France 1991; but pawn. But the costs are high, i.e. 13 g4 3l.g6
1O ... i.xf3 11 ~xf3 ctJxe5 12 ~e2 ctJxc4 13 14 ctJxe5 ctJxe5 15 l:!.xe5 c6 and Black has
~xc4 f5 14 ctJd2 fxe4 15 ~xe4 d5 16 ~d3 good compensation.
ctJf6 was also OK in Tiitta-Sepp, Finland 13 .. .16 14 lLlg3
1997) 11 i.e2 3l.xf3 12 3l.xf3 3l.g5 13 i.e2 In 1999 this was a new move to theory,
i.xc1 14 ~xc1 ~h4 15 ctJd2 (or 15 ~e3 f5 but not to me. I has analysed the position
16 exf5 ctJh6 17 ctJd2 ctJxf5 with good coun- after S... l:!.bS a few weeks earlier.
terplay for Black) 15 ... l:!.aeS with equality in 14.....tf7 15 lLlh4
Armas-Hebden, France 1991. 15 ctJf5 3l.c5! is flne for Black. If 16 b4?
b) 9 as a6 (weaker is 9... f5?! 10 a6 fxe4 11 ctJxc3 17 ~3 3l.xc4 IS dxc4 i.xb4 19 3l.b2
dxe4 bxa6 12 i.xa6 and the weak pawn on ~d3! and Black wins.
a7 gives White the advantage; but 9...l:!.bS! is 15 ... J:!.e8!
stronger in my opinion - it is always better to
prepare ... b7-b5 this way, as then if 10 a6
Black has counterplay with 10 ... b5!) 10 d4 (10
ctJbd2 f5 11 ~3 was interesting) 1O ... 3l.g4
11 d5 ctJbS (as in Macieja-Johansson, Ber-
muda 2002) 12 3l.d3! ctJd7 13 b4 and White
is slighdy better according to Macieja.
8 ... ~b8!

This simple move was shown to me be-

fore the game by my coach Wojciech Ehren-
feucht (we both played in the Polish Cham-
pionship that year). The idea is simple: ...3l.fS
and the white attack is over.
16 lLlhf5 ..tf8 17 ~b3 ~d7 18 lLle3 J:!.ed8
19 lLlxd5 il..xd5 20 iLe3 b5 21 axb6 axb6
22 ~c2 %-%
Of course the 'I' is only my own opinion, I had planned to play 22 ... ctJa5! (but not
but I sincerely believe that the pawn is better 22 ... i.xc4?! 23 dxc4 ~d3 24 ~a4 ctJa5 25 c5
kept on a7. and White is slighdy better) 23 i.xd5 ~xd5
9 c3 ..tg4 10 lLlbd2 d5 11 exd5 lLlxd5 12 24l:!.adl c5 where Black has more space, but
h3 ..th5 13 lLle4 the position is nothing more than equal.

157
Two Knights Defence

Summary
4 d3 is not a dangerous move. Black should equalise in all lines - with the exception perhaps
of 4 ... dS!?, which is interesting and fun to play, but inadequate for equality. The usual move is
4 ... j"e7 gives Black a safe position, while 4 ... h6!? leads to a more complicated game with good
chances for a long and hard fight. And if you are happy in the quiet lines of the Italian Game.
there is also 4 ... ~cS.

1 e4 e5 2 lUf3 lUe6 3 ~e4 lUf6 4 d3 (D) ~e7

4 ... j"cS - Italian Game
4 ... h6 - Game 58
4... dS - Game 59
5 0-0 0-0 6 J::te1
6 c3 dS - Game 60
6 j"b3
6 ... dS 7 exdS tDxdS (D)
S .l::i:e1 - Game 61; S h3 - Game 62
6 ... d6 7 c3 - Game 63
6 tDbd2 d6 7 a4 '.t>hS - Game 64
6 ... d6 7 a4 (D)
7...tDd4 - Game 65
7 ...'.t>hS - Game 66

4 d3 7... lUxd5 7 a4

158
 INDEX OF COMPLETE GAMES I

Anand-Beliavsky, Linares 1991 ............................................................................................... 75

Bachler-Colias, USA 1991 .................................................................................................... 123
Bahram-Hector, Stockholm 1998............................................................................................... 7
Belov.I-Pankratov, Comspondence 1995 ................................................................................. 65
Belov.L-Nezhmetdinov, Omsk 1961 .................................................................................... 55
Bex-Donev, BieI1995 ................................................................................................................ 73
Borge-Hector, Copenhagen 1994 .............................................................................................. 93
Braunsdorf-Augustat, Comspondence 1993 ........................................................................... 81
Bucan-Geller, Bad Wiirishoftn 1992 ......................................................................................... 86
Chandler-Littlewood.P, London 1996 ................................................................................... 24
Ciocaltea-Nezhmetdinov, Bucharest 1954 ............................................................................ 49
Delaney-Hebden, Kilkenny 1999 .......................................................................................... 133
Djurhuus-Blees, Gausdal1993 .............................................................................................. 119
Dubiel-Macieja, Biala Podlaska 1994 ................................................................................... 147
Elison-Werner.K, Email 1999 ................................................................................................. 78
Ellner-Andruss, Comspondence 1977 ..................................................................................... 113
Gelashvili-Gokhale, Dubai 2002 .......................................................................................... 149
Grau Ribas-De Groot, Email 1997 .......................................................................................... 9
Groszpeter-Hazai, Hungary 1998 ........................................................................................... 51
Grott-Leisebein, Comspondence 1998 ...................................................................................... 60
Gurevich.V-Jonkman, Germany 2002 .................................................................................... 89
Hoogervorst-Simmelink, Comspondence 1988 ................................................................... 117
Howell-David.AI, Groningen 1995........................................................................................... 83
Howell-Volzhin, Calcutta 1996................................................................................................ 14
Kamsky-Ye Rongguang, Manila 1990 ............................................................................... 122
Kan-Konstantinopolsky, Mo.'cow 1945.................................................................................. 11

159
 Two Knights Defence

Kogan-Svidler, Oakham 1992 ................................................................................................ 138

Kovchan-Malaniuk, Swidnica 1999....................................................................................... 152
Kozakov-Jonkman, I~vov 2001 ................................................................................................ 90
Kristensen-Hebden, Kopavogur 1994 ................................................................................... 108
Kriiger-Moormann, Email 1998 ............................................................................................. 74
Leisebein-Grott, Com.rpondence 1998 ...................................................................................... 16
Lee.G-Iuldachev, Gent 2002 ................................................................................................... 19
Leygue-Flear, St. Affnque 2001 ............................................................................................. 127
Losev-Isaev, Kherson 1990 ........................................................................................................ 69
Macieja-Ivanchuk, Hyderabad 2002 ..................................................................................... 154
Macieja-Pinski, Polanica Zdrqj 1999...................................................................................... 156
Malakhatko-Timoshenko, Kiev 2003 ................................................................................... 42
Morozevich-Onischuk, Aloscow 1996.................................................................................... S6
Morozevich-Sokolov.I, Sarajevo 1999 .................................................................................... 35
Narciso Dublan-Kuzmin, Balaguer 1997.............................................................................. 25
P:ilkovi-Wells, Zalakaros 1998 ................................................................................................. 95
Paoli-Wagman, Comspondence 1965/66................................................................................. 79
Perez.R-Garcia.Gild, Santa Clara 1996 ............................................................................... 146
Petrosian.A-Mikhalchishin, DOrlmund 1998 ....................................................................... 48
Pilgaard-Pedersen.N, Danish Championship, Greve 2002..................................................... 27
Plachetka-Smejkal, Ostrava 1994 ........................................................................................... 97
Psakhis-Geller, Sochi 1984 ..................................................................................................... 144
Rogers-Wong Chee Chung, Singapore 1998 ...................................................................... 100
Rudnick-Pichler, Comspondence 1985 .................................................................................... 30
Sermek-Olarasu, Nova Gonca 2002 ........................................................................................ 22
Shabalov-Ivanov.I, US Championship, Parsippany 1996........................................................ 71
Short-Hector, L.anzarote 2003 .................................................................................................. 32
Spiegel-Mari Arul, Calcutta 1997 ............................................................................................ 37
Steinitz-Meitner, Vienna 1860 .............................................................................................. 131
Struik-Mikhalchishin, Zu)olle 2003 ...................................................................................... 114
Sveshnikov-Zaitsev, Podolsk 1992 ........................................................................................ 110
Sveshnikov-Zaja, Bled 2001 ................................................................................................... 106
Tagansky-Glazkov, Moscow 1975 ......................................................................................... 141
Timoshenko-Vysochin, Kiev 2003 ........................................................................................ 44
Van der Wiel-Spassky, Reggio Emilia 1986/ 87..................................................................... 39
Vukcevich-Romanishin, Hastings 1976/77 ......................................................................... 53
Vysochin-Shishkin, Kiez! 2003 ................................................................................................ 46
Weber-Grzelak, Comspondence 1992 ..................................................................................... 126
Weir-Smits, Email 1994 ............................................................................................................ 62
Wendland-Grober, Comspondence 1997 ............................................................................... 104

160
 the
two
knights
defence
The Two Knights Defence is one of the trickiest tactical
openings around. If White initiates complications with
either 4 Ng5 or 4 d4, play becomes extremely sharp
and gambits and counter gambits abound. Anyone
who enters the murky waters of the Two Knights
Defence must be well prepared for the mind-boggling
complications that ensue. In this book, openings
theoretician Jan Pinski guides the reader through
both the well-trodden paths of the main lines plus the
less fashionable side variations of this most complex
opening. Using illustrative games, Pinski studies the
key ideas and tactics for both Black and White.

II Up-to-date coverage of a controversial opening

II Written by an openings expert
II Ideal for club and tournament players

International Master.Jan Pinski is a talented player

from Poland who is highly regarded as an openings
theoretician. This is his third book for Everyman
following the Classical Dutch and the Sicilian
Kalashnikov (co-authored with Jacob Aagaard)

ISBN 1-85744-283-0

www.everymanchess.com
Published in the UK by Gloucester Publishers pic
Distributed in the US by the Globe Pequot Press
£14.99 $19.95