The Instructor

Castling on Opposite Sides

One of my articles, in the collection published under the English title Attack and
Defence by Dvoretsky and Yusupov, concerns the outstanding collection of
games by the 12th World Correspondence Champion, Grigory Sanakoev.

However, I did not deal with the engrossing encounter that I now bring to your
attention in that article. I had previously prepared the game for training replay,
relying chiefly on Sanakoev’s comments. Several times, my students have played
out the position arising after White’s 18th move, either with Black (playing
against me) or against one another (using the mutual attack against opposite-side
The castled positions as a training theme). Naturally, we discovered analytical
additions and improvements in the process of examining the game. I would like
to acquaint my readers with the most current version of my notes to this game.
Instructor
Mark Dvoretsky Estrin - Sanakoev
World Correspondence Championship 1968-70

1. e2-e4 d7-d6 2. d2-d4 g7-g6 3. Nb1-c3 Bf8-g7 4. f2-f4 Ng8-f6 5. Ng1-f3 0-0 6.
e4-e5 Nf6-d7 7. h2-h4 c7-c5 8. e5-e6

8. h5 is considered the main line. According to theory, it offers White no

advantage.

8...f7xe6 9. h4-h5 g6xh5!

10. d4xc5

Here I present a long quote from Sanakoev,

which reflects his approach to explaining
opening theory in his annotations:

After the game Estrin maintained that this

move was a major inaccuracy, bringing the
Black knight into action and freeing the light-
squared bishop, which plays an important
role in Black’s defense. To support this view
he gave the variation 10. Rxh5 Nf6 11. Rh4,
and claimed that either of Black's replies would lead to positions where
White's chances were clearly preferable despite his pawn minus:
a) 11...Nc6 12. Be3 cd 13. Nxd4 e5 14. Nxc6 bc 15. fe de 16. Bc4+ Kh8

file:///C|/cafe/Dvoretsky/dvoretsky.htm (1 of 6) [6/5/2005 1:29:03 PM]

The Instructor

17. Qe2.
b) 11...cd 12. Nxd4 e5 13. Bc4+ d5 (13...Kh8 14. fe de 15. Nf3 Qxd1+ 16.
Kxd1) 14. Nxd5 Nxd5 15. Qh5 h6 16. fe. Later, G. Fridstein discovered
that Black could seize the initiative with 14...e6! 15. Nxf6+ Qxf6 16. Nf3
e4 17. Ng5 h6 18. Qh5 Qd4!. Ten years after this game, in Estrin - Nunn,
Lublin 1978, Black demonstrated an improvement two moves earlier, with
12...Nc6!, which gave him the advantage after 13. Be3 Qb6! 14. Qd3 Nb4
15. Qd2 e5 16. fe de.

Quite frankly, I can see little point in trying to establish the ultimate
theoretical truth on the basis of a game I finished many years ago. A
concrete opening problem arose during the game and had to be solved
promptly (even allowing for the special conditions of postal play, with the
relatively free time-limit for thinking about your moves). If, after the
game, many qualified players in various countries exerted themselves and
unearthed refinements, improvements, - even refutations! - in this or that
variation, what does it prove?...

To revert to Estrin’s comment on his tenth move, it must be said that the
general grounds on which he prefers 10. Rxh5 are, of course, valid. Yet
just as many arguments can be urged in support of 10. dc. For instance
the black knight is diverted to the queenside; the pressure is lifted on d4;
Black loses the chance to bring his knight to the defense of his weakened
kingside wiht tempo; he can no longer develop the other knight with
tempo after exchanging on d4, etc. The choice between two continuations
of roughly equal worth is a matter of taste and may also depend on your
mood.

I must admit that what he says here is close to my own way of thinking. Real
over-the-board struggles have always interested me far more than theoretical
discussions, prepared at home with the aid of reference books and computers. In
addition, the current article is devoted to the problems of the middlegame; thus, I
won’t even waste any time presenting the variations referred to above: the reader,
if he wishes, may do so himself.

10...Nd7xc5 11. Rh1xh5 Nb8-d7

11...e5? would be bad: 12. Bc4+! (12. fe Bg4) 12...e6 13. fe d5 14. Bd3
(Sanakoev).

12. Bf1-d3 Nd7-f6 13. Rh5-h4 Bc8-d7

13...e5!? 14. fe Nxd3+ (or 14...de 15. Bc4+ e6) 15. cd de 16. Nxe5 Bf5 would
have led to quieter play.

14. Bc1-e3 Nc5xd3+ 15. Qd1xd3 Qd8-e8

It looks more natural to bring the bishop to g6, not the queen. Of course, this
would weaken e6. Sanakoev considers the variation 15...Be8!? 16. Ng5 Bg6 17.
Qe2 Nd5 18. Bd2 as “leading to difficulties for Black.” However, after 18...Qc8!

file:///C|/cafe/Dvoretsky/dvoretsky.htm (2 of 6) [6/5/2005 1:29:03 PM]

The Instructor

19. Qxe6+ Qxe6 20. Nxe6 Bxc3, or 19. Nxe6 Bxc3! 20. Nxf8 (20. bc Rf6 21.
Ng5 Nxf4!) 20...Bxb2 21. Nxg6 Bxa1 22. Nxe7+ Nxe7 23. Qxe7 Qxc2, the
position clearly favors Black.

16. 0-0-0 Qe8-g6 17. Qd3-e2 Rf8-c8

This rook could have come in handy on the f-file; therefore, it would have made
sense to play 17...Rac8!? (with no need to worry about 18. Bxa7, because of
18...Nh5) Another promising-looking idea is Viorel Bologan’s suggestion:
17...b5! 18. Nxb5 Nd5 (or 18...Bxb5!? 19. Qxb5 Nd5).

18. Be3-d4

This is where we begin the training games;

for the notes which follow, I shall be relying
on the games we played and the subsequent
analyses.

In positions with opposite-side castling, both

sides generally storm the enemy king’s
fortress with pawn attacks. The guiding
principle in such positions is: “Whoever
comes first, wins!”

18...b7-b5! 19. g2-g4!

Of course it would be senseless to throw in the moves 19. a3? a5 - moving the
pawn up to a3 would make it easier for Black to open queenside lines.

In one training game, an interesting pawn sacrifice was essayed: 19. f5?! Qxf5
20. g4 - the point being to gain a tempo for the pawn’s march to g5. However,
this idea was refuted by 20...Qf4+! 21. Kb1 (21. Be3!? Qc4-/+) 21...e5 22. g5
Bg4! (22...Ng4 23. Nd5 Qf7 24. Bc3 is much less convincing) 23. gf Bxf3 24.
Rxf4 Bxe2 25. Rg1 (25. Nxe2 exf4 26. fg e5-+) 25...exf6 26. Rxf6 Bh5 27. Be3
Bg6-+.

And 19. Rdh1!? b4 20. Nd1 Rc7 (but not 20...Rxc2+?! 21. Qxc2 Rc8 22. Nc3 bc
23. bc, with advantage to White) 21. Ne3 Rac8 leads to an unclear position.

19...b5-b4 20. g4-g5!

file:///C|/cafe/Dvoretsky/dvoretsky.htm (3 of 6) [6/5/2005 1:29:03 PM]

The Instructor

20...Nf6-h5!

“The opening of the g-file would spell death

for Black. After 20...bc? 21. gf cb+ 22. Kb1
ef 23. Rg1 Qf7 24. Rhg4, White wins”
(Sanakoev). This analysis is inexact: as
indicated by Bologan, Black meet 22. Kb1?
with the tremendous counterblow 22...Rxc2!!
23. Qxc2 (23. f7+!? Kf8) 23...Qxc2+ 24.
Kxc2 Ba4+ 25. Kxb2 Bxd1, and only Black is
thinking about winning. White can keep the
advantage by choosing 22. Bxb2! Ba4
(22...Bxf6 is bad because of 23. Rg1, and 22...ef, because of 23. Rg1 and 24.
Rhg4) 23. f5!?, with 24. Rxa4 to follow.

Black would have a difficult position after 20...Nd5?! 21. Nxd5 ed 22. Bxg7
Qxg7, and now not 23. Nd4?! (hoping for 23...e5 24. Rdh1! ed 25. Rxh7 Qxh7
26. Rxh7 Bf5 27. Qh5! Bxh7 28. g6i) 23...Rc4!, but 23. Qd3! threatening 24.
Qxd5+ or 24. Rdh1.

21. f4-f5!

On 21. Ne4? Sanakoev examines 21...Rxc2+!?, but 21...Bc6! is simpler.

21...Qg6xf5

21...ef? 22. Nd5 is unfavorable, or 21...Qf7? 22. g6 hg 23. fg, when White has a
dangerous attack.

22. Rh4xh5 b4xc3

The intermediate 22...Be8? is unfortunately met by 23. g6!

23. Bd4xg7

23...Bd7-e8!!

But here, the intermediate move proves

necessary: it’s the only way for Black to
neutralize White’s onslaught.

23...Kxg7? is bad: 24. Nd4! cb+ 25. Kb1 Qg6

26. Rdh1, and White’s attack must succeed,
despite the loss of two or three pawns.

Interesting complications would arise after

23...Bb5?! 24. Qh2 Qxf3

file:///C|/cafe/Dvoretsky/dvoretsky.htm (4 of 6) [6/5/2005 1:29:03 PM]

The Instructor

The normal 25. Rxh7?! leads, after 25...e5!!,

to an unclear position: 26. Bxe5 (26. Rh8+!?
Kf7 27. Bxe5) 26...Qe3+ 27. Kb1 Qxe5 28.
Qh5! (28. Qh6? Bd3!!-+) 28...Qf5! (28...Be8?
loses to 29. Qh6 Rab8 30. b3 Rxb3+ 31. ab
Rb8 32. g6! - just not 32. Rh1? Rxb3+ 33. cb
Qf5+, drawing) 29. Rh8+ (29. Rh1? Bd3!-+)
29...Kg7 30. Qh6+ Kf7 31. Rh7+ (or 31. g6+)
31...Ke8 32. Re1. This was, in fact, the course
of a training game in which Vadim
Zvjagintsev played the black pieces against
me.

White should have taken the other pawn first: 25. Bxc3! Rxc3; only now does 26.
Rxh7! Bd3! (both 26...Rxc2+ 27. Qxc2 Rc8 28. Qxc8+ Kxh7 29. Qxe6 and
26...Qe3+ 27. Kb1 Qe5 28. Qh6 lose for Black) 27. Rh8+ Kg7 28. Qh6+ Kf7 29.
bc lead to an advantage for White.

24. g5-g6!

Much weaker are both 24. Rh2? Kxg7 25. Nd4 cb+ 26. Kb1 Qe5 27. Nxe6+ Kg8
and 24. Nd4? Bxh5 25. Qxh5 Qf7 26. Qxf7+ Kxf7 27. Bh6 cb+ 28. Kxb2 Rab8+
29. Kc1 e5 - in either case, Black is on top (variations by Sanakoev).

24...c3xb2+!

On 24...Qf4+? 25. Kb1 Bxg6, the simplest means to White’s end is 26. Rh4!
followed by 27. Bxc3.

After 24...Qxh5?! 25. Qxe6+ Kxg7 26. Qxe7+, Sanakoev’s opinion is that White
should mate soon.

Black’s position indeed looks very

dangerous; however, I see no win for White
here. Some sample variations: 26...Kg8 27.
gh+ Kh8 (27...Qxh7? 28. Rg1+ Bg6 29. Qf6)
28. Qf6+ Kxh7 29. Rg1 (29. Ng5+ Kg8 30.
Rg1 cb+ 31. Kb1 Qg6) 29...Bg6 30. Ng5+
Kh6! (but not 30...Kg8? 31. Ne6! Qh6+ 32.
Kb1 Kh7 33. Ng5+ Kg8 34. Nf7+-) 31. Ne6
(31. Nf7+ Kh7 32. Ng5+ Kh6!) 31...cb+ 32.
Kb1 Rg8 33. Nf4 Raf8! 33. Rxg6+ Qxg6 34.
Qh4+ Kg7 35. Nxg6 Kxg6, with equality.

25. Bg7xb2

25. Kb1 is worse: 25...Qxc2+! (as long as Black doesn’t fall for 25...Rxc2? 26.

file:///C|/cafe/Dvoretsky/dvoretsky.htm (5 of 6) [6/5/2005 1:29:03 PM]

The Instructor

Qxe6+! Qxe6 27. gh+) 26. Qxc2 Bxg6! 27. Qxg6 hg, when Black has winning
chances.

25...Rc8xc2+! 26. Qe2xc2 Qf5xc2+ 27. Kc1xc2 Be8xg6+ 28. Kc2-d2 Bg6xh5

Black has emerged from the complications with four pawns for the knight. On
the other hand, White still has a way to keep a rough balance. This very sharp
battle soon comes to a peaceable conclusion.

29. Rd1-g1+ Bh5-g6 30. Nf3-h4 Ra8-b8 31. Bb2-c3 Rb8-b1

On 31...Kf7, there follows 32. Nxg6 hg 33. Rf1+ Kg8 34. Rg1.

32. Rg1-g4!

Sanakoev notes that 32. Rg2?! was less exact: 32...Kf7 33. Nxg6 hg 34. Rf2+
Ke8 35. Rg2 Rh1! 36. a4 Rh6 37. Bg7?! Rh4. Now Black has to take Ra4 into
account.

32...e6-e5 33. Nh4xg6 h7xg6 34. Rg4xg6+ Kg8-f7 35. Rg6-g4 Drawn.

For those who would like further practice with opposite-side castling positions, I
recommend my lectures on this theme in the above-cited book, Attack and
Defence.

[ChessCafe Home Page] [Book Review] [Columnists]

[Endgame Study] [ Skittles Room] [Archives]
[Links] [Online Bookstore] [About ChessCafe] [Contact Us]

Copyright 2005 CyberCafes, LLC.All Rights Reserved.

"The Chess Cafe®" is a registered trademark of Russell Enterprises, Inc.

file:///C|/cafe/Dvoretsky/dvoretsky.htm (6 of 6) [6/5/2005 1:29:03 PM]