Anatoly Karpov's

Best Games

Anatoly Karpov

Translated by Sarah J. Young

B. T. Batsford Ltd, London

First published 1 996

Anatoly Karpov 1996
ISBN 0 7 1 34 7843 8
British Library Cataloguing-in-Publication Data.
A catalogue record for this book is
available from the British Library.

All rights reserved. No part of this book may be

reproduced, by any means, without prior pennission
of the publisher.

Typeset by Petra Nunn

and printed in Great Britain by
Redwood Books, Trowbridge, Wilts
for the publishers,
B. T. Batsford Ltd,
4 Fitzhardinge Street,
London W l H OAH

A BATSFORD CHESS BOOK

Editorial Panel: Mark Dvoretsky, Jon Speelman
General Adviser: Raymond Keene OBE
Specialist Adviser: Dr John Nunn
Commissionin Editor: Graham Bur ess

Contents
Symbols
Introduction
1 Karpov - Gik, Moscow University Ch 1969
2 Karpov - Hort, Moscow Alekhine mem 1971
-\Vsmejkal - Karpov, Leningrad IZ 1973
arpov - Polugaevsky, Moscow Ct (6) 1974
5 Karpov - Spassky, Leningrad Ct (9) 1974
6 Karpov - Korchnoi, Moscow Ct (2) 1974
7 Karpov - Vaganian, Skopje 1976
8 Tatai - Karpov, Las Palmas 1977
9 Karpov - Korchnoi, Baguio City Wch (14) 1978
1 0 Karpov - Korchnoi, Baguio City Wch (32) 1978
1 1 Timman - Karpov, Montreal 1979
1 2 Korchnoi - Karpov, Merano Wch (9) 1981
1 3 Karpov - Korchnoi, Merano Wch (18) 1981
14 Karpov - Yusupov, USSR Ch (Moscow) 1983
1 5 Karpov - Kasparov, Moscow Wch (9) 1984/5
1 6 Karpov - Kasparov, Moscow Wch (27) 1984/5
1 7 Karpov- Kasparov, Moscow Wch (4) 1985
1 8 Kasparov - Karpov, Moscow Wch (5) 1985
,!.2_ Karpov - Kasparov, Moscow Wch (22) 1985
Q Karpov - Beliavsky, Moscow tt 1986
2 1 Karpov - Kasparov, London/Leningrad Wch (5) 1986
22 Karpov - Kasparov, London/Leningrad Wch (1 7) 1986
2} Karpov - Kasparov, London/Leningrad Wch (19) 1986
241Karpov - Sznapik, Dubai OL (14) 1986
25 Karpov - A.Soko1ov, Linares Ct (10) 1987
26 Kasparov - Karpov, Seville Wch (2) 1987
CZ1';, Karpov - Farago, Wijk aan Zee 1988
2f Karpov - Timman, Brussels World Cup 1988
29 Karpov - Kasparov, Belfort World Cup 1 988
> Karpov - M.Gurevich, USSR Ch (Moscow) 1988

-}K

5
6
9
13
18
25
29
33
37
41
45
49
52
55

60

65
68
73
78
83
86
88
91
94
98
1 02
105
1 10
1 13
1 16
1 20
1 25

3J Karpov- Yusupov, USSR Ch ( Moscow) 1988

@ Karpov - Malaniuk, USSR Ch ( Moscow) 1988

3 3 Karpov - Hjartarson, Tilburg 1988

34 Hjartarson - Karpov, Seattle Ct (3 ) 1989
35 Karpov - P.Nikolic, Skelleftea World Cup 1989
Karpov - Yusupov, London Ct ( 8) 1989
37,, Karpov - Timman, Kuala Lumpur Ct ( 4) 1990
38 Karpov - Anand, Brussels Ct ( 4) 1991
39 Karpov- Short, Linares Ct ( 7) 1992
40 Karpov - Kamsky, Moscow Alekhine mem / 992
(l Kamsky - Karpov, Dortmund 1993
42 Karpov - Kramnik, Linares 1994
- ::41' Karpov - Beliavsky, Linares 1994

1 28
131
1 33
1 36
1 38
142
145
1 50
153
1 57
1 63
167
1 72

Symbols
1-0
0 1
lf2_lf2

++

#
!
?

!!

??

!?
?!
Ch
Cht
Echt
tt
jr
worn
rpd
Web
z

IZ
Ct
OL
Corr
(D)
(n)

White wins
Black wins
Draw
Check
Double check
Checkmate
Good move
Bad move
Outstanding move
Blunder
Interesting move
Dubious move
Charnpionship
Team championship
European team championship
Team tournament
Junior Event
Women's event
Rapidplay
World Championship
Zonal
Interzonal
Candidates
Olympiad
Postal game
Diagram follows
nth match game

Introduction
In 1968 I started at Moscow State University as a seventeen-year-old, and
in 1969 I won my first tournament in Moscow. You could say that the Mos
cow University Championship was my first step in the struggle to reach
the top in chess. Immediately after it I was selected for the World Junior
Championship, won it, and six years later, in 1975, became 'adult' World
Champion. A quarter of a century has passed since 1 969, and quite re
cently, in 1994, I won the ' super-tournament' at Linares - the 'World
Championship Tournament' , and what is more, I can say without false
modesty that I had a unique result - 1 1 points out of 13, and two and a half
points clear of Kasparov.
So, this anniversary of a quarter of a century in top-level chess, and also
my success at Linares, gave me the idea of compiling a collection of my
best games 1969- 1994.
In my already fairly long chess life I have taken part in around two hun
dred tournaments and matches (not counting team competitions, rapid
plays and blitz tournaments) and won well over a hundred of them.
Although it would be difficult to count up precisely, I think that all in all I
have won about a thousand games. Obviously, out of a quantity like that it
is not at all easy to pick out the very best games. I hope that all of them are
models, which the reader can judge, of the clearest examples of my work,
and many of them are also part of the true art of chess.
It stands to reason that this book includes all my most interesting victo
ries from matches for the chess crown, from Candidates battles of various
years, and from prominent international tournaments. Many of the victo
ries I have chosen won prizes for their beauty, best game prizes, and most
important theoretical game prizes in one of the most popular world chess
publications, lnformator. Incidentally, the author of this book has the
leading number of these lnformator prizes, way ahead of other grandmas
ters.
Of course, over many years chess players remember fewer and fewer of
their old games, as they are displaced by fresher and more important du
els. That is why this book in the main concentrates on my victories over
the last ten years.

Introduction 7

I have also written new commentaries for the older games, or at least re
worked the opening part, as theory has advanced so much since then.
In an overwhelming number of games in this collection I am playing
White. This is not surprising, as beautiful victories are most often gained
with the white pieces.
Let us quickly run through the composition of this collection. Which of
my victories have got into it?
The first game, dated 1969, was played when I was still a master, but the
next one is taken from a strong grandmaster tournament. Then my pro
gress towards the Mount Olympus of chess is illustrated - here the reader
will find one game from the Interzonal tournament and all the Candidates'
matches. My match with Fischer did not take place, through no fault of
mine, and in 1975 I became the 1 2th World Champion. Then there are
some games from prominent international tournaments, and then two en
counters from my duels with Korchnoi in Baguio and Merano. You could
say that my victory in our anniversary national championship ends the
first part of the book.
In 1984 a new era began in chess, linked with Karpov-Kasparov con
frontation. Here the reader will find ten of my victories from my first four
duels with Kasparov. I also bring you some games played in between
these 'epoch-making' matches. After our battle in Seville, our dispute was
interrupted for three years. In this part of the book you will find my victo
ries in the World Cup, our 55th national championships, and other con
tests. After 15 years I again had to climb the Candidates ladder, and most
of the steps are illustrated with a victory, as well as games from my next
duel with Kasparov. This is where you could say the second part of my
book ends.
Soon I was joining the regular cycle of battles for the crown, which also
means regular games from new Candidates matches. Unfortunately, my
encounter with Short did not end happily for me, and, alas, my possible
sixth meeting with Kasparov did not take place. Therefore there are sev
eral wins from tournaments in the early 1990s. In 1993 a schism occurred
in the chess world (it is scarcely worth dwelling on that question in this
book), as a result of which I was able to play another match for the chess
crown. Having won my match against Timman, I was again declared
FIDE World Champion (Kasparov and Short had played for the title of
PCA World Champion).

Introduction

I end the book with two victories from the 'super-tournament' at Li

nares in 1994. Although after that I have won other interesting tournament
games, I think that the full stop in this book should be placed precisely af
ter Linares, where I achieved one of the most convincing victories in the
entire history of chess.
In conclusion it remains for me to thank the chess master and writer
Evgeny Gik for his help in preparing the manuscript.

Anatoly Karpov

Karpov

Game 1
Karpov - Gik
Moscow University Ch 1969
Sicilian, Dragon
The Moscow University Champi
onship was my first tournament in
the capital, and I really wanted to
win it. The champion's title was
decided by this very game, and
who would have thought that my
opponent in my first tournament in
the capital would later become co
author for some of my books?
c5
1 e4
d6
2 lLlf3
cxd4
3 d4
lLlf6
4 lLlxd4
g6
5 lLlc3
g7
6 e3
lLlc6
7 f3
0-0
8 c4
'it'aS
9 ifd2
d7 (D)
10 0-0-0

Gik 9

The Dragon Variation, one of

the sharpest and most fascinating
in chess theory. Both sides' aims
are the same - getting to the enemy
king as quickly as possible. Be
cause of this, White attacks on the
kingside, stopping at nothing, while
Black in his turn attacks on the
queenside. Games played in this
variation are almost always played
in an open fighting spirit, and are
often awarded prizes for beauty.
lLle5
1 1 h4
l:r.fc8
12 b3
Nowadays the queen's rook is
most often placed on c8 (putting
the queen on a5 in that case is not
obligatory), but at that time mov
ing the king's rook to c8 was more
popular.
lLlxh5
13 h5
14 h6
At the end of the 1 960s this posi
tion was subjected to thorough
analysis, and moreover there was a
lively discussion about the knight
check 14 . . . lLld3+. In this game I
was preparing to test one of my
own ideas on this theme, but my
opponent managed to avoid my
preparation.
xh6
14 ...
.:xc3 (D)
15 1Wxh6
The standard exchange sacrifice
in the Dragon - on the one hand
Black protects himself from the
knight lunging onto d5, and on the

10 Moscow University Ch 1969

other he shatters the enemy king's

fortress.

w
16 bxc3
'if:xc3
The black queen here is occupy
ing an ideal position to generate
threats to the white king, and it is
difficult to believe that this move
can already be a decisive mistake.
Either 16 . . .lLlf6 or 1 6 ....:r.c8 was
necessary.
17 lLle2! (D)
In the event of 1 7 ..t>b1 a5 Black
obtains fair counterchances. How
ever, the modest knight retreat to
e2 is extremely unpleasant for
Black. The knight deals with the
problem of ousting the queen beau
tifully, and simultaneously joins in
the attack on the kingside.
'iWcS
17
Alas, after 1 7 ... lLld3+ 1 8 .:r.xd3
'iWa1+ 19 ..t>d2 'iVxh1 20 g4 lLlg3 2 1
'iVxh 1 lLlxh 1 2 2 e3 ! and 2 3 .:r.d 1
Black's knight is lost.

lLlf6
lLlhS (D)

18 g4
19 gS

w
20 .:r.xhS!
There is no time to lose. 20 lLlg3,
which I almost played at the board,
had to be rejected because at the
last moment I spotted the effective
20... .i.g4 ! , which would exclude
the white queen from the game.
20
gxhS
21 .:r.h1
'iWe3+
22 bl!

Karpov

In the Dragon even a slight inac

curacy can spoil the whole game.
Thus, for example, 22 b2? would
have given Black at least a draw:
22 . . . d3+ 23 cxd3 (23 bl loses
after 23 . . .'ifxf3 ! ) 23 ... 'ifxe2+ 24
al 'ifxd3 and Black is guaran
teed a perpetual check.
'i'xf3
22
The white knight is untouch
able: 22... 'ifxe2 23 'fi'xh5 e6 24
'ifxh7+ 25 'fi'h8+ e7 26 'iff6+
e8 27 .::t h8#. 22 ... e6 also does not
work: 23 'ifxh5 'ifxf3 (23 . . .g6 24
'ifxh7+ f8 25 g3 and 26 f5)
24 'ifxh7+ f8 25 d4.
23 .::txhS (D)

Gik 1 1

for White here was only revealed

ten years after this game. The key
move for White was ignored by the
world press, as many commentar
ies only examined, after 24 "ii'xh7+
'ittf8, 25 .::th6 and 25 'fi'h6+:
a) 25 .::t h6 e6 26 .::t xg6 fxg6 27
it'xd7 "ii'xe2 28 "ii'xd6+ g7 29
"ii'e7+ 'itth8 30 "ii'f6+ h7 31 "ii'f7+
h8 32 "ii'xg6 "ii'd l + 33 b2
'ii'd4+ 34 c3 "ii'd2+ 35 .i.c2 'it'd7 !
and Black maintains equality.
b) 25 il'h6+ was also suggested
in my notes, with the following
variation: 25 . . .e8 26 il'h8+ f8
27 .::th7 and now:
bl) 27 . . . e6 turns out to be inac
curate, as the bishop should be
placed on e6, not the pawn; after 28
g6 fxg6 29 d4 'ifxe4 30 'ii'g 8
it'xd4 31 'iff7 + 'it>d8 32 'ifxf8+
'it>c7 33 il'xa8 the white king easily
steals away from the 'perpetual'
check: 33 ...'ii'dl+ 34 'it>b2 'ii'd4+
35 c3 'ii'f2+ 36 .i.c2 "ii'b6+ 37 al
'ifgl+ 38 .i.bl .
b2) 27 ....i.e6 ! 28 g6 (28 d4 is
not dangerous for Black either, as
after 28 . . .'ii'xe4 29 xe6 fxe6 30
g6 .::tc 8 31 g7 il'e l + 32 b2 White
is forced to submit to the perpetual
check; or 28 .i.xe6 fxe6 29 g6
'ii'xe2 and then . . . "ii'el+) 28 . . . fxg6
(28 . . . "ii'xe2 29 g7) 29 .i.xe6 "Yi'xe2
30 .:n 'ii'el+ 31 b2 'ifb4+ 32
.i.b3 'i!fxb3+! (32 . . .d7 33 .::t xf8
.::t xf8 34 'fi'xf8 'ifxe4 with three

12 Moscow University Ch 1969

pawns for the piece) 33 axb3 xf7

with very unclear play. Moreover,
maybe Black does not have to give
up his queen.
It is amusing that old and, as we
shall see, not particularly precise
variations for some time caused
experts to change their attitude to
wards 16 ...'ikxc3, and in some
theoretical publications this move
has been considered quite play
able. But all the same, microscopic
analysis allowed me to return, and I
think definitively, to the old assess
ment: taking on c3 loses for Black.
If after 23 ...lLlg6 24 'i!t'xh7+ f8,
instead of 25 'ii'h6+ or 25 .:h6,
White should make the quiet rook
move, 25 l:hl ! ! , then Black would
be completely helpless.
Having thought of this rook ma
noeuvre, it was easy to find sup
porting variations:
c) 25 .:hi ! ! (D) and then:

c l ) 25 . . ..i.e6 26 lLld4 (but not

the continuation 26 .i.xe6 fxe6 27
lbd4 'ii'f7 !) 26 ... 'ii'xe4 27 lLlxe6+
fxe6 28 :Cl+ 'iti>e8 29 'ii'g8+.
c2) 25 ...e5 26 lLlg3 ! 'ifte8 27 .:n
'itxg3 28 'i!t'g8+ lLlf8 29 'ii'xf7+
d8 30 'ilixf8+ .i.e8 31 'ii'xd6+
and Black is forced to lay down his
arms.
c3) 25 . . . e6 26 ltJd4 'iixe4 (al
ternatively, 26 .. .'iff4 27 lLlf5) 27
.:n .i.e8 28 lLlxe6+ 'ifte7 29 lLlc7.
24 g6!
The courageous pawn sacri
fices itself, guaranteeing a breach
in the enemy defences. The hasty
24 ifxh7+ would have left the
black king at large: 24 . . . f8 and
now not 25 ii'h8+ (nor 25 lLld4 ow
ing to 25 . . .'ilidl + 26 'iti'b2 'ii'xd4+)
25 .. .ri;e7 26 'ii'xa8 due to 26...'ii'xh5.
lLlxg6
24
Not 24 ... fxg6 (24 . . . hxg6 is even
worse: 25 'ilih8#) 25 'ii'xh7+ 'ifilf8
26 'ii'h 8+ e7 27 .:h7+ lLlf7 28
'iVxa8.
However, if, as in the game,
Black takes on g6 with the knight,
it seems that everything is in order
for him: White's attack has been
repulsed, and furthermore he is
two pawns down.
25 'i!t'xh7+
'itf8
26 .:rs ! ! (DJ
This unexpected rook move was
like a thunderclap in a clear sky for
Black ! The elegant geometrical
.

Karpov - Hort 13

.I

w
' ... .
'
\Wr

B
idea immediately decides the fate
of the game. Two lines - the a2-g8
diagonal and the f-file - intersect
on one critical square - f7. The
threat is 27 'ifxf7#, and moreover,
the rook is supporting the queen
along the file, while the bishop - in
the event of 26 . . .exf5 - is support
ing her on the diagonal. Because of
this Black is forced to part with his
own queen.
'ilxb3+
26
exfS
27 axb3
28 lbf4! (D)
Another elegant manoeuvre.
The aS-rook is not defended, and
thanks to this White can com
pletely destroys his opponent's de
fences.
28 ...
.:d8
29 'ii'h6+
A last finesse - the g6-pawn will
drop with check.
e8
29
30 lbxg6
fxg6

. . -

.

.

. .
- . .
.<it?.

B
31 'ii'xg6+
32 'ilg5+!
After 32 exf5 .:fs
still resist.
32
33 exfS
34 'ii'g8+
35 'ii'g7+

ri;e7
Black could

ri;e8
.:c8
q;e7
1-0

Game 2
Karpov - Hort
Moscow Alekhine mem 1971
Sicilian, Keres Attack

cS ,
1 e4
d6
2 lbf3
cxd4
3 d4
lbf6
4 lbxd4
5 lbc3
e6
6 g4 (D)
The Keres Attack - my favourite
weapon against the Scheveningen
in the 1970s and 1980s. This vic
tory is perhaps one of the clearest I
gained in this variation, although

14 Moscow Alekhine mem 1971

8 f4
8 .ie3 a6 (two other popular
lines are 8 ...4Jb6 and 8 ....ie7) 9 h4
(theory has also minutely studied 9
a4, 9'tlfd2 and 9 :g1) is an impor
tant alternative ; 8 4Jdb5 4Jb6 9
.if44Je5 10 'ilih5 has been played
in its time; now 10 .. 4Jg6 ! (after
1o... g6 1 1 it'h3 .id7 12 o-o-o it'b8
13 it'g3 Black is defenceless) 11
.ig3 ( 1 1 .ixd6 .ixd6 1 2 :d1, inci
dentally, does not work, as after the
cold-blooded 12 . 0-0 13 4Jxd6
'ii'e 7 ! , the initiative is on Black's
side in spite of his lack of pawns)
l l . . .a6 1 2 4Jd4 .ie7 (12 . . . d5 ! is
also interesting) 13 0-0-0 .ixg5+
14 'it>b1 0-0 is unclear.
a6
8
9 .ie3
.ie7
9 . . . h6? is very risky: 1 0 4Jxe 6 !
fxe6 1 1 'ii'h5+ e7 1 2 .ic4 'ii'e 8
13 1li'h3. However, 9 ...'tlfc7 de
serves attention.
10 :g1 (D)
10 'tlfd2 4Jxd4 11 .ixd4 e5 1 2
.ie3 exf4 1 3 .ixf44Je5 1 4 0-0-0
.ie6 15 h4 'i!i'a5 16 a3 :c8 174Jd5
led to an even game in Tarjan-Hu
lak, Indonesia 1983.
10
4Jxd4
Black does not usually hurry
with this exchange , but here it is
not easy for him to find another
move : IO. .'i'c7 looks too sluggish;
deciding on 10 ...0-0 when one is
directly under attack is not easy;
.

B
almost a quarter of a century has
passed since .
6
4Jc6
It is interesting that I used the
Keres Attack in the very first game
of my long marathon against Kas
parov (Moscow 1984/5). True, we
did not play it again, except a
modified form of it in game 14 of
our second match, when I chose an
unusual move-order: 2 . . .e6 3 d4
cxd4 4 4Jxd4 4Jc6 5 4Jc3 d6 6
g4 !?, not then expecting the knight
to go to f6. After a sharp battle the
matter ended peacefully. In the first
game of our first match Kasparov
chose 6 . . . h6 in the diagram posi
tion, a more popular continuation
in recent years (6. . . a6 is a third pos
sibility). After 7 h4 4Jc6 8 :g 1
h5 ! ? 9 gxh5 4Jxh5 10 .ig54Jf6 1 1
'ilfd2 I gained somewhat the better
chances, but that encounter also
ended in a draw.
4Jd7
7 gS
...

..

Karpov - Hort 15

B
10 . . . lDc5 1 1 lDxc6 bxc6 1 2 xc5
gives White a clear advantage.
1 1 'ilt'xd4
e5
exf4
12 'ilt'd2
13 xf4
lDe5
1 3 . . . 'ilt'b6 14 llg3 'ilt'xb2 1 5 llb1
Wa3 1 6 lDd5 is obviously bad for
Black.
14 e2
e6
Black could not generate any ac
tivity with 1 4 . . . 'ilt'a5 as the end
game which follows is clearly in
White's favour: 1 5 lDd5 'ilt'xd2+ 16
'itxd2 d8 17 llad 1 e6 18 c 1 .
15 lDd5
Immediately ! Otherwise Black's
queen could take up an active posi
tion, e.g. 15 0-0-0 'ilt'a5 ! .
1 5 ...
xd5
16 exd5 (D)
As a rule, you should try to oc
cupy a blockading square with a
piece. Actually, 16 'ilt'xd5 would
also have left me with an advan
tage, as the d6-square demands

B
constant defence by Black. But
then my e4-pawn would also in
some cases need watching, which
might restrict my light-squared
bishop. Now this bishop is free for
its own action, all the more so as its
black counterpart is no longer on
the board.
lbg6
16
Black's position is appreciably
worse, so Hort seeks a tactical so
lution to his problems. Castling on
either side ( 16 0-0 or 16 ...'ilt'c7
followed by 16 ...0-0-0) would condemn Black to passive defence.
17 e3
h6!?
It is difficult to give a move like
this a symbol. It shows a fighting
mood and a desire to carry out the
battle in a tactical key. But it also
represents a precise assessment of
the position with a clear under
standing of its disadvantages. In
short, it is risky for both sides !
18 gxh6
h4+

...

16 Moscow Alekhine mem 197 1

1 9 d1 (D)

B
The white king is not troubled
by the loss of his castling rights. At
the same time, the black king
maintains that possibility right to
the end of the game, but he does
not manage to make use of it.
19 ...
gxh6
f6
20 xh6
20 ...Wf6 would probably not
have worked out for Black in view
of the fact that the h4-bishop has
perceptibly lost its mobility. Hort
was intending idealistically to re
build his forces, but in chess there
are two sides, and I in my turn took
some necessary countermeasures.
21 c3
eS (D)
Black seems to have achieved
his aim - he is threatening to play
22 ...Wh4+, while after 22 g5 "ilfb6
23 e3 Wc7 my achievements will
not be that great. But White finds a
very strong retort.

22 l:.g4!
The h2-pawn is disturbing no
one (for the time being !). Now the
main thing is not allowing the en
emy queen an active position on
the kingside, where my pieces are
rather precariously arranged.
22 ...
'iVf6
22 ...xh2 would have been a
relatively better choice, restoring
the material balance. But I can un
derstand my opponent; he wanted
to develop.
23 h4!
Now not 23 ... li)xh4? because of
24 g7. In general it will be diffi
cult for Black to win back this until
recently so helpless pawn, which is
now gradually finding its strength.
23 . ..0-0-0 was also impossible in
view of 24 g5.
23 ...
'ii'fS
Defending against the threat
ened 24 h5, Black prepares to cas
tle long.

Karpov - Hon 17

24 l:tb4! (D)

B
A beautiful place for the rook !
In the first place 24 0-0-0 25 .ig4
is impossible, and in any case the
b7 -pawn is under threat of capture.
24
.tf6
25 h5
t[je7
Of course, this retreat is less val
iant than 25 . . . 5. but then his
pieces would not be able to gener
ate any real activity (after 26 l:tf4) .
I should note in passing that during
all this time it has not been feasible
for Black to sacrifice the exchange
with ...l:txh6.
'iWe5 (D)
26 l:tf4
27 l:tf3
The rook, which so often lacks
agility, in this position shows re
markable manoeuvrability. It cre
ates one threat after another, and
moreover is acting productively
not only in attack, but also in de
fence. Thus, Black will now gain
...

w
nothing after 27 ......xh5 28 l:txf6
'iVh l + 29 .tfl (the rook defends
both bishops) 29 ... t[jg8 30 e l+
and White wins.
27 ...
t[jxd5
28 l:td3
l:txb6
Nothing better is apparent - if
28 ... t[je7, then 29 .if4 is decisive.
29 l:txd5
"ii'e4 (D)
,

18 Leningrad IZ 1973

A hymn to the rook! The chess

player's consciousness is used to
the working potential of all his
pieces, but, you will agree, with
rooks this mainly applies to the
endgame.
'ii'h 1+
30
Black is forced to go into this by
no means equal exchange.
31 c2
'iVxa1
i.eS
32 'ifxh6
33 'iWgS
Preventing queenside castling
and creating irresistible threats to
the exposed king. In this desperate
position Hort lost on time.
1-0
The game was nominated as the
best at the Alekhine Memorial, and
then lnformator included it in their
best games. Although many years
have passed, I will confess that I
still love this game.

Game{!)
Smejkal Karpov
Leningrad IZ 1973
Sicilian, Paulsen
-

1
2
3
4
5
6
7
8

e4
ltlf3
d4
lbxd4
lbc3
i.e2
0-0
i.e3

c5
e6
cxd4
lbf6
a6
'fic7
lbc6
i.b4 (D)

w
9 lba4
One of the most popular posi
tions in the Paulsen System, which
was established long ago. The b6square is weakened, and White dis
patches his knight to it. As well as
this he makes use of the fact that
taking the e4-pawn is dangerous:
9 . . . lbxe4 1 0 lbxc6 'iVxc6 1 1 ltlb6
.:.bs 1 2 'i!Vd4 i.f8 1 3 i.f3 d5 (or
1 3 ...f5 14 .:.ad 1 ) 14 c4.
0-0
9 .
Black has a wide choice: 9 ...b5 ?
or 9 .. d5?; retreating the bishop to
d6 or e7, or even 9 . . lbe7. But cas
Hing is the most reliable way of fin
ishing his development.
bxc6
10 lbxc6
Theory shows a clear-cut way
for White to gain an advantage af
ter either 10...dxc6 or 10...'iVxc6.
1 1 lbb6
The most logical, although 1 1
c4 and 1 1 f4 are also known.
1 1 ...
.:.bs (DJ
..

Smejkal - Karpov 19

-
.....
- - .

. . .
.

- -- .

.
-
- . . "
,
-U
U
"

.
. --

. . . .
"
"rf.
U
U
" ,

w
12 tbxc8
l:l.fxc8
Twenty-odd years after this game
I decided to test another capture1 2 ... 'ifxc8 (Salov-Karpov, Buenos
Aires, Sicilian theme tournament
1 994) 1 3 e5 ltJd5 14 i.c 1 i.c5 and
now White, instead of the theoretical recommendations 1 5 i.d3 or
1 5 c4, decided on an aggressive
policy - bringing his queen over to
the kingside with the standard Sicilian manoeuvre 'ii'd 1-d3-g3: 1 5
'ii'd 3 0-0 16 'it'g3l:l.e8!? 17l:l.dl aS
1 8 b3 a4 1 9 i.h6 i.f8 20 h4 'f!lc7.
Thus, pressure on the g7-point is
parried, and the ending arising is
highly favourable for Black.
13 i.xa6
l:l.d8! (D)
At the time when this game was
played, the rook move to d8 was
presented as a novelty, and moreover a very important one. It looks
as though the rook is no worse on
e8, but run ahead for a second and
have a look at the situation i n the

w
game after 24 'ifxf7+. As you can
see, if the rook were on e8, Black
would have had to resign immed.iately.
14 i.d3
14 i.g5 has apparently not been
tested in practice, for example,
14 .. .'ii'e5 1 5 i.h4 'ii'xe4 1 6 .i.g3
l:l.a8 17 i.d3 leads to a complex po
sition.
14 ...
i.d6
15 h1
15 f4 has also been seen more
than once. If 1 5 ... e5 16 f5l:l.xb2 17
g4, then 1 7 . ..'ifa5 18 h 1 i.c5 19
i.c 1 l:l.xa2 20 l:l.xa2 "ii'xa2 21 g5
ltJe8 22 'ifh5 d5 ! 23 g6 lLlf6 is
good for Black.
.i.e5
15
l:l.xb2
16 c3
ltJg4! (D)
17 'it'cl
18 f4
Of course, 1 8 'W'xb2 loses after
1 8 ...i.xc3 ! .
ltJxe3
18

'I

- :

20 Leningrad IZ 1973

21
eS
Later it was established that here
the best move is 2 l . .. g5 !? with
roughly the following variation: 22
g3 'ifd6 ! (an important intermedi
ary manoeuvre) 23 i.e2 i.e5 24
'ifxf7+ Wh8 25 l:td l 'it'c7 26 'ife7
(26 .tg4 d5 ! 27 'ifxe6 i.xc3 28
exd5 cxd5 29 i.f3 We5 30 Wxe5+
i.xe5 with a draw) 26 ... i.xc3 27
i.g4 W'c8 ! 28 i.xe6 l:te8 29 Wc5
l:txe6 30 Wxc3+ g8 3 1 'ii'c4 'ii'b 8
32 a4 'ife5 Razuvaev-Matulovic,
Tbilisi 1 973.
22 g3
'ifd6!
i.gS
23 i.e2
24 'ilfxf7+
h8
25 a4 (D)
Incidentally, 25 i.g4 is also a
promising coninuation for White,
for example, 25 ...i.f6 26 g2
l:tf8 27 'ii'xd7 'ii'c5 28 h3 Wc4 29
l:tf2 preserving an advantage for
White; Adorjan-Matulovic, Novi
Sad 1 973.
..

w
19 'li'xbl
1 9 fxe5 is no good in view of
1 9 'ifb6 with a clear advantage.
i.xf4
19 ...
But not 19 ... lLlxfl ? 20 fxe5 l2Je3
2 1 W'e2 'it'b6 22 :tel .
20 W'f2
He has to return the exchange,
because after 20 l:tf3 lLlg4 Black's
chances are better, as 2 1 h3 is im
possible due to 2 1 ...i.c 1 ! !.
20
lLlxfl
21 :xn (DJ
...

..

Smejkal - Karpov 21

Strangely enough, I had this po

sition on the board when I was pre
paring for the game. True, it turns
out to be weak for Black. I had to
apply more than a little force to ex
tricate myself from it, and then I
managed to swindle my opponent.
25
Jle7
l:.f8
26 aS
27 'i'c4
l:.xfi+
28 Jlxfi
'ili'f6!
The only way for Black to or
ganize his position. The bishop be
longs on c5, and so needs support
from the queen on f8.
29 gl
'ili'f8
30 Jle2
Jlc5
31 Jlg4
'iif2+
d6 (D)
32 h3

w
The critical moment.
33 Jld7
33 'ii'e6 g6 (Black cannot move
into an ending with opposite-col
oured bishops, since 33 ... 'iif l+ 34

Wh4 'ii'f6+ 35 'ii'xf6 gxf6 immedi

ately loses a pawn to 36 Jld7 and
the white king will easily steal his
way through the holes in the pawn
barrier) 34 'i!fe7 ! Jle3 ! (the strongest
continuation because co-operative
variations such as 34 ... h6 35 i.e6
\i'fl + 36 h4 lead to victory for
White), and now Black, by defend
ing against the direct threat, creates
counterplay:
a) 35 Jle6 'ii'f l + 36 g4 'i'e2+
(but on no account 36 ... h5+?? be
cause of 37 h4) 37 h3 'ii'f l + 38
Wh4 g5+ 39 g4 (after 39 'itth 5?
Black forces mate: 39...'ii'e2+ 40
Jlg4 'fi'xh2+ 4 1 Jlh3 'ii'xh3#)
39 ...'ii'e2+ 4(1 'itth 3 'iWfl +, etc., is a
perpetual check.
b) White cannot even advance
his passed pawn with 35 a6 since
after 35 . ..h6 threatening ...'iWfl +
and ...Jlg5+, he would be forced to
give perpetual check himself.
c) 35 Jle2! (not 35 ...'ii'xe2?? 36
ir'f8#).
d) 35 'i!fxd6 also leads to per
petual check after 35 ... 'iWfl+ 36
h4 g5+ 37 h5 'iWf7+ 38 'itth6
ii'g7+ 39 h5 'i'f7+.
Smejkal was counting on win
ning the c6-pawn and quietly going
about attaining an advantage. The
move played is perhaps the correct
practical decision.
g6
33
34 Jlxc6
'ittg7

22 Leningrad IZ 1973

'ifb2!
35 .i.bS
A fine move, the idea of which
Smejkal had apparently over
looked. A threat to the h2-pawn has
suddenly arisen, and both the c3pawn and the b5-bishop find them
selves under attack. White has to
part with the c3-pawn, but in the
first place he has a very dangerous
passed pawn. The misfortune for
Black lies in the fact that he can
never transfer play into an oppo
site-coloured bishop endgame,
because White, by threatening an
exchange, can at any moment
'question' the black queen from almost any square.
36 a6
.i.gl
37 We2
Wxc3 (D)

-
-
-
:
- .,.
- . .
.i. .

. .

. tlli

..wrll. "

,@-
u

.
-

Precise play by Black has al

most equalized the position, but the
presence of queens on the board
makes the situation dangerous for
both sides.

Wet
38 .i.c4
39 Wn ?
39 g2 was correct. Now Black
achieves an advantage.
39
Wh6+
Wxh2+
40 g2
41 f3
Wh5+
42 g2 (D)
Here the game was adjourned.
Analysis shows that with precise
play White has chances for a draw,
but to achieve this he will have to
overcome great difficulties.

B
Wh2+
42
.i.d4
43 f3
44 .i.dS
.i.cS
.i.d4
45 .i.c6
46 .i.b7? (D)
White has fallen into the trap I
set for him and made my problems
easier. 46 .i.d5 ! was correct. I do
not at all want to say that after this
move everything is in order for
White, because that is really not

Smejkal - Karpov 23

the case. In the adjourned position

Black can try various ways of
achieving victory, but White has
defensive resources for every one
of them.

.i..
-

. .

A -
-L:l

.
.
mu

. . . .. .

B
The position which has arisen
on the board is a kind of zugzwang
for White. The king cannot go to
the only free square, g4, in view of
the mate. The white queen cannot
move to either g2 or e2, as in that
case he will be forced to advance
the g3-pawn by means of ...'iVhS+,
and definitively compromise the
position of the king. In fact, only
the bishop can move. In his turn,
Black has to find a plan to
strengthen his position. As we saw
earlier, simple checks here achieve
nothing. I chose a plan linked with
advancing my kingside pawns.
Analysis revealed that this con
tinuation is not the only one, but it
is effective enough. It is better for

Black to begin his pawn movement

with his bishop on c5 or d4, while it
is better for White to keep his own
bishop on c6 or dS. As you can see,
a system of corresponding squares
has arisen. Smejkal feared this and
accordingly spent more time ana
lysing a different continuation, so
he did not find this area of corre
spondence, and took his bishop too
far away. I will show the difference
in later notes .
46
gS!
47 g4
hS+
48 rs
The only move. After 48 xg5
'iVxg3+ 49 xhS (49 fS 'ii'g4#)
49 . . ..i.f2! White has no defence
against mate.
48
'iVxg3
49 e6
'iif2 ! (D)

24 Leningrad IZ 1973

his advantage: 50 'ii'xf2 .i.xf2 5 1

xd6 g4 ! and the bishop does not
manage to stop the pawn from
reaching f1 , whilst in the event of
52 .i.c8 f6 Black advances the
king. The difference lies in the fact
that if the bishop were not on b7,
but on d5 instead, White could
have continued 50 "iixf2 .txf2 5 1
f5 ! g4 52 .tc4 ! and the passed
pawn is held back, while the black
king cannot move forward.
'ii'f6+
50
51 dS
g4
"fle7
52 .tc8
53 .trs
h6
"ilc7
54 "ii'n
'ii'cS+
SS 'ii'el
56 e6 (D)

Black has successfully arranged

his forces and easily parries all
White's attempts to create counter
play.
61 'ii'al
'ii'fl
62 'iVbl
g3
63 .th3 (D)

. . - .
- .

.
- .

0. ?&l.&
w;@
A.

.
.0

.i.
.W%:-?1

.0:?;!;

56
57
58
59
60

wn
'ii'el
'ii'dl+
'ii'aS

gS
'ii'a3
.tcs
'ii'e3
.tb6

"

"'

mu
-

B'ii'B B

B
White has occupied the last line
of defence, but this blockade is
only temporary.
h4
63
'iVgl
64 .tgl
Black forces an exchange of
queens, after which the road to vic
tory becomes quite short.
65 'ii'xgl
.i.xgl
66 d6
Black would have achieved an
effective finale after 66 f6 .td4
67 .tfl g2! 68 .txg2 g3 fol
lowed by ...h4.
.i.d4
66
.txa7
67 a7
g4
68 xeS
.

,__ '

Karpov - Polugaevsky 25

h4
dS
h3
eS
xh3
.txh3+
.tcS!
e6
0-1
This encounter won the regular
competition in lnformator for the
most important theoretical game.
69
70
71
72

Game@
Karpov - Polugaevsky
Moscow Ct (6) 1974
Sicilian, Najdorf
cS
1 e4
2 tbf3
d6
cxd4
3 d4
4 tbxd4
tbf6
5 tbc3
a6
eS
6 .te2
.te7
7 lbb3
.te6
8 0-0
'fic7
9 f4
lbbd7
10 a4
0-0
1 1 h1
12 .te3
exf4
tbeS (D)
13 .:txf4
14 aS
Fourteen moves have gone and I
have not yet made a single com
ment. This is not surprising - this
position is one of the standard pat
terns in the Najdorf Variation, and
it has been seen in practice many
times. In the fourth game of the
match I played the weaker 14 tbd4,
and after 14 ....:tad8 1 5 'ii'g 1 .:td7 1 6

-
..

- -*--
. . - .
a

!@

"
"
u
. i.
u
a <att
w
.:td 1 .:te8 17 tbf5 .td8 1 8 tbd4
tbg6 1 9 .:tffl tbe5, Black had a
lovely game. Later on both 14 'ii'd2
and 14 tbd5 were tested several
times. The second of these moves
deserves special attention, for ex
ample in the encounter Hulak-Por
tisch, Indonesia 1 983, where after
14 tbd5 .txd5 1 5 exd5 tbfd7 1 6
.:tb4 .:tfe8 1 7 a5 .tf6 1 8 .t g1 .tg5
19 tbd2 tbf6 20 tbfl g6 White
could have gained an advantage by
attacking on the queenside: 2 1 c3
.:te7 22 ifb3 h5 23 h3 lbh7 (pre
paring ... f5) 24 .:tb6 followed by 25
.:ta4.
tbfd7
14 ...
It was scarcely worth moving
this knight away from the centre.
The back rank is free of pieces, and
an initial manoeuvre by one of
Black's rooks suggests itself ....:tfe8, ....:ae8, or ....:ac8. For ex
ample, after 14 . . . .:tac8 1 5 tbd4
tbfd7 16 'ii'd2 .:tfe8 17 tbf5 .tf8 1 8

26 Moscow Ct (6) 1974

.:tf2 h8 in Matanovic-Polugaev
sky, Moscow 1 977 a tense, com
plex struggle arose.
1s .:r.n
..tf6
16 lbdS!
..txd5
17 xd5! ?
After the quiet 17 exd5 Black
should continue 17 ...lbc4 1 8 ..txc4
'ifxc4, and White's opening initia
tive is exhausted. In order to pre
serve it, an idea came to me during
the game of sacrificing two pawns
straight away.
17
"ikxc2 (D)

w
Polugaevsky spent more than an
hour wondering whether to take or
not. He might have been thinking
that if I was prepared to sacrifice, it
meant that I had analysed every
thing thoroughly at home, but on
the other hand, he could not find the
forced loss at all, so he kept search
ing through it again and again. If
Black declines the sacrifice he

ends up in a difficult positional

bind.
'ji'xb2
18 lbd4
Not going the whole hog (even
if that is what Black wanted to do)
would lead to a position with
material equality and a noticeable
advantage to White. For example,
after 1 8 . . ...c5 1 9 lbf5 ! 'ji'xd5 20
exd5 the black pawn on d6 cannot
be saved.
'ji'cJ
19 .:tabl
1 9 . ....a3 is worse - the black
queen finds herself on the edge of
the board and totally out of the
game.
20 lbfS (D)
The aim of this move is to de
fend the bishop and not allow the
queen to c5, which might happen
in the variation 20 .:tb3 'ji'cS 2 1
lbf5 'ji'xd5 2 2 exd5 lbc5, when
Black can defend himself.

Karpov - Polugaevsky 27

Polugaevsky is manoeuvring
magnificently. If Black procrasti
nates, White can improve the inter
action of his pieces: 2 1 J.d4 'ii'd2
22 l:.b2 or 2 1 l:.fc l Wa3 22 l:.al
'i!i'b2 23 l:.a2 'i!i'b4 24 J.d2 ! trap
ping the queen. With the move in
the game Black avoids this.
21 l:.be1
Imprecise. After 2 1 l:.fe 1 White
would have gained a big advan
tage. Now Black can construct a
defence.
21
t:hc5
22 t:hxd6
lhcd3
23 J. xd3
t:hxd3 (D)

.1.

$!

..... -
.

. !?L.X
...
.

" .

. - .
. .
n
a.:
w

24 l:.ed1
t:hb4
This is the essential difference.
If on my 2 1 st move I had played
l:.fel, the white rooks would now
be on bl and d l. Consequently,
Black would only have two reason
able possibilities: 24. . . lhf2+ 25
J.xf2 'i!i'xf2 26 e5 J.e7 27 l:.fl , and

24 ...t:he5 25 lhxb7, but in both

cases White has the initiative.
25 'ii'xb7 (D)
I was short of first two, and then
one pawn, and perhaps I was there
fore unconsciously trying to liqui
date my 'material disadvantage' .
However, 25 'i!i'h5 would have cre
ated the threat of 26 l:.xf6 gxf6 27
'ili'g4+ c;i;lh8 28 J.h6, which would
not. have been easy to repulse
(25 ...g6 26 'ii'c 5).

.'ii' . .......
.... . .

. .
.

.,q
.0.

m
v o
. .: :
B
25
l:.ab8
26 'ii'a7
Wc6
The decisive mistake. 26 ...t:hc6
is no good either: 27 'i!i'c7t:hb4 28
'ti'xc2 lhxc2 29 J.b6 with advan
tage to White. But 26 ...'ii'e2 would
have given Black definite counter
play, although here too after 27
J.b6 White has an advantage.
27 J.f4
Later I discovered that in the
press centre Furman was at this
.

28 Moscow Ct (6) 1974

very moment proving that an ex

change sacrifice leads to victory
for White: 27 llxf6 ! gxf6 28 .i.h6
(threatening 29 'ii'e 3) 28 . . . 'iic2 (or
28 . .. tbd3 29 .i.xf8) 29 .U.c l 'ii'd3 30
1t'c5 ! . But when I played 27 .i.f4
(threatening 28 e5) my trainer con
tentedly commented, 'and that is
good as well' . In fact, the threat of
e4-e5 contains Black's activity,
while the b4-knight finds itself out
of the game.
27
.:.as
28 1ifl
.U.ad8 (D)

Otherwise e4-e5 with a decisive

attack. However, the attack has not
been stopped.
-

W%

ta'i

, /,ffl i . i
m

"

}A
iQ)i
,Wi
,,
R

,..."-

r ''
; .
8
1W illti

.\Ulr.
"
i{:y,, u
. . :
--"

w
32 .i.h6
The fire is now directed at two
points at once, f7 and g7.
tbc6
32
In the hope of complicating the
game after 33 tbxf7li'c4 ! [editor's
note: but then 34 .i.xg7 .i.xg7 (af
ter 34 . . .'Wxfl + 35 .U.xfl .i.xg7 36
tbh6+ h8 37 .U.xf8+ .U.xf8 38 tbf5
White has a decisive material ad
vantage) 35 1ixg7+ rJilxg7 36 .U.g3#
is mate].
33 tiJfS! (D)
33
'ii'b2
The game would have had a cu
rious finish in the event of 33 ....i.e5
34 .i.xg7 .i.xg3 35 .U.xg3, when
Black cannot prevent mate.
34 .i.cl!
White frees h6 for the knight
with tempo. There is no defence.
.

w
29 'Wg3!
The threat is stronger than its
execution ! It is a paradoxical prin
ciple, but it is valid surprisingly
often. White does not hurry with
the advance e4-e5, and increases
the pressure.
29
'ii'c3
30 .U.f3
'fi'c2
31 .U.dfi
.i.d4 (D)

Karpov - Spassky 29

B
ti'b5
34
Wh8
35 lbh6+
36 lbxf7+
l:.xf7
Or 36 ...g8 37 lbh6+ h8 38
.:.Xf8+ and mate.
..tf6
37 l:.xf7
Wg8
38 ti'f2
39 l:.xf6
gxf6
40 ti'xf6
1-0
Game 5
Karpov Spassky
Leningrad Ct (9) 1974
Sicilian, Scheveningen
-

1
2
3
4
5
6
7
8
9
9

e4
lbf3
d4
lbxd4
lbc3
..te2
0-0
f4
..te3 (D)

c5
e6
cxd4
lDf6
d6
..te7
0-0
lbc6
..td7

B
The so-called Modern Schev
eningen is seen these days only
rarely, and this match with Spassky
did not play the final role in this. In
my duels with Kasparov in 1 984
and 1 985, where the Sicilian De
fence arose regularly, my opponent
preferred the Classical Scheven
ingen with a more flexible move
order - first ... a7-a6 and ...ti'd8-c7,
then ... lbb8-c6 and ...l:.f8-e8, and
only then .....tc8-d7. Premature de
velopment of this bishop is playing
into White's hands. The alterna
tives are 9 ...ti'c7 or 9 ...e5.
10 lbb3
Black was preparing to ex
change knights in the centre and
advance the bishop with .....tc6.
Thus this knight retreat is quite ac
ceptable and more precise than the
other well-known moves 10 h1
and 10 'fi'el .
With the bishop on d7, Black's
counterplay on the queenside falls

30 Leningrad Ct (9) 1974

behind in comparison with White's

threatening pawn storm on the
kingside, all the more so as the cen
tre is fully under his control.
aS?! (D)
10
Although White also has the
initiative after the waiting moves
1 0 ... a6 or 10 ...'ifc7, playing like
this is even less dangerous. The a
pawn's march is superficially ac
tive, but in strategic terms it is
dubious. White gains the excellent
b5-square for the knight, and hin
ders freeing manoeuvres by the
black pieces. This game may have
been played twenty years ago, but
it is still a good model of how
White should play in situations like
this.

w
lbb4
1 1 a4!
It is more precise to play 1 l ...e5
1 2 'ith1 and only then 1 2 ... lbb4, al
though 1 3 .tb5 ! .tc6 14 .txc6
bxc6 15 fxe5 dxe5 16 'ii'e2 gives

White a clear positional plus in any

case.
12 .tf3
.tc6 (D)
Returning the knight to d4 is not
really to Black's liking, but 1 2 ... e5,
with the passive bishop on d7, is
not very attractive either.

w
13 t2Jd4
g6
Or 1 3 . . . 'ii'b 8 14 'ili'e2 l:l.e8 1 5
l:l.ad 1 '1th8 1 6 lbdb5 l:l.a6 1 7 l:l.d2
l:l.d8 18 l:l.fd 1 d5 19 e5 lbd7 20
'iif2 ! and White had a clear advan
tage in Klovans-Vasiukov, Mos
cow 1968.
14 l:l.f2
e5
15 tbxc6
bxc6
dxe5 (D)
16 fxe5
17 'ii'f l!
The battle is flaring up around
the c4-square, which White is
counting on occupying with one of
his pieces. If Black managed to
prevent this, things would not look
at all bleak for him. It makes no

Karpov - Spassky 31

w
sense to move the rook from the f
file ( 17 l:td2), as it is still unclear
which file will be the centre of ac
tivity.
'ii'c8
17
18 h3 (D)
Of course, there was no point al
lowing an exchange of knight for
bishop ( 1 8 ... lLlg4).

with 19 l:tcl (or 19 g4 followed by

g5 and .i.g4) 1 9 . . . l:tfd8 20 .i.e2
l:td4 21 b3 (in order to establish the
bishop on c4).
h5
19 .i.g4
This severely weakens the posi
tion of the black king. He should
have chosen the simple 19 ...'i'c7,
escaping from the pin and connect
ing the rooks along the back rank.
'i'xd7
20 .txd7
21 'i'c4 (D)
So, the queen has arrived at her
appointed destination.

.th4
21
An unpleasant endgame awaits
Black after 2l ...'i'e6 22 1fxe6 fxe6
23 l:taf l .
22 l:td2
'i'e7
23 l:tO!
I was not tempted by the chance
of winning the exchange; after 23
.i.c5 'ii'g 5 24 l:td7 lLlxc2 25 i..xf8
l:txf8 Black has active counterplay.

lbd7
18
In the event of 1 8 . . . 'Mi'e6, White
could have developed the initiative

32 Leningrad Ct (9) 1974

The d-file does not matter; the

d8-square is securely covered.
Therefore I have to generate ac
tivity in other directions.
l:l.fd8 (D)
23

The time has come to invite the

black knight to move away from its
familiar place.
26
llJa6
27 l:l.e2! (D)

w
24 llJbl!
The craftiness of this move lies
in the fact that I have managed to
choose the most suitable moment
to move the knight into a more ac
tive position (this can happen par
ticularly quickly if Black swaps
rooks). It is amusing that conse
quently this unexpected knight
retreat has become an almost char
acteristic illustration of my work.
'iiib7
24
25 'itth 2!
It is a rare occurrence when, in
the middlegame, the mobility of
your king limits the activity of an
enemy bishop.
25
g7
26 c3

White does not intend to ex

change rooks yet; the major pieces
are needed for an attack on the f
file. This also frees a square for the
knight transfer llJd2-f3-g5, and
moreover threatens 28 g3 f6 29
l:l.ef2 l:l.d6 30 g5.
.:r.rs
27
28 llJd2
d8
29 llJf3
f6
In defending the e5-pawn, Black
is simultaneously trying to cover
the f-file. But White's attack is al
ready unstoppable.
30 l:l.d2!
The white rook's 'hesitation'
might seem illogical. First he occu
pied the d-file, then abandoned it,
and now returns in decisive fashion

Karpov - Korchnoi 33

(that is what is important!) onto the

open file.
30
e7 (D)
After 30 ... o!Db8 3 l.!Dg5 ! I would
have won straight away.
..

w
When Spassky made his move,
I lost my head a little. At first I
thought that I was winning in all
variations, but then I looked and
could not find a win. Fortunately,
my confusion only lasted for a mo
ment.
31 W'e6
l:tad8
This loses by force. Dragging
out resistance is only possible by
means of 3 1 ....!Db8, as if setting up
the pieces for a new game.
32 .:.Xd8
xd8
If 32...l:txd8, then 33.!Dxe5 'ilc7
34 'ilt7+ h8 and 35 'ilxe7 'ilxe5+
36 'ilxe5 fxe5 37 l:tf6.
33 lld1 (D)
Material is level on the board.
Black's king has avoided White's

.
-
. d
- d
.\Ull d

dwd

- .
- .
-
.
d
-d
d -"
u .

.".:
.

B
immediate threats, but Spassky's
position is worsening with every
move. The problem is that the
black pieces are disconnected and
cannot come to each other's help.
Now, for example, it is impossible
to defend the seventh rank by
means of 33 . l:tt7 due to the hang
ing bishop on d8.
33
o!Db8
34 cS
llh8
35 l:txd8!
10
After 35 . . .l:txd8, 36 e7 ! is de
cisive.
..

Game 6
Karpov - Korchnoi
Moscow Ct (2) 1974
Sicilian, Dragon

1
2
3
4
s

e4
o!Df3
d4
o!Dxd4
o!Dc3

cS
d6
cxd4
.!Df6
g6

34 Moscow Ct (2) 1974

In my long chess career the

Dragon Variation has been used
against me about 20 times, and I
think only once has my opponent
managed to gain a draw. But al
though many years have passed, to
this day this encounter with Korch
noi remains one of the clearest ex
amples in the whole history of the
variation. Possibly the level at
which his game was played had a
part in that.
.tg7
6 .te3
lh c6
7 f3
0-0
8 'ifd2
.td7 (D)
9 .tc4

the symmetrical . . .h5 (now or on

the next move). The basic develop
ment of contemporary theory in
this variation has also gone in pre
cisely that direction.
1 1 .tb3
lheS
12 0-0-0
The immediate 1 2 h5 is also
possible.
lhc4
12 ...
This was Black's last chance of
averting the dangerous opening of
the h-file by means of 12 . . .h5.
.:txc4
13 .txc4
14 hS
14 g4 also leads to a sharp game.
lhxhS
14
lh f6 (D)
15 g4

w
10 h4
.:cs
In Game 1 examined above, the
then fashionable line .. .'i!i'a5 and
....:tfc8 was played. Transferring
the f8- rook to c8 was later totally
displaced by putting the queen's
rook there. Besides, in recent years
Black has tended to meet h4 with

w
A very topical position in those
days. Here White has many possi
bilities: 1 6 .:tdg1, 16 lhd5, 16 .th6,
or 16 e5. However, I chose the
modest knight retreat to e2, as I had
in mind one new idea which I had

Karpov - Korchnoi 35

prepared specially for this match. I

should note that a detailed investi
gation of all the finesses of even
one of the above-mentioned lines
of the Dragon would take many,
many pages.
16 lbde2! ?
The idea o f this move i s under
standable - to reinforce the c3square, as exchange sacrifices such
as . . Jhc3 often let B lack gain a
powerful attack; indeed the Dragon
Variation is underpinned by this
blow. However, from e2 the knight,
when the occasion arises, can eas
ily be transferred over for a direct
attack against the enemy king.
16 ...
'iia5
1 6 . . . lte8 seems safer for Black,
as 1 7 h6 ( 1 7 eS? is no good be
cause of 17 . . .lbxg4 ! 18 fxg4 xg4
19 'ii'd3 't!l'c8 with advantage for
Black; Bemai-A.Schneider, Hun
gary 1 976) is met by the retreat
17 ...h8. The game Klovans-Beli
avsky, USSR 1 977 continued 1 8 eS
lbxg4 1 9 fxg4 xeS 20 f4 'ifaS
2 1 xeS 't!l'xe5 22 lbd5 ltxg4 with
roughly even chances.
17 h6
xh6
Here as well it was possible to
play 1 7 . . . h8, in the spirit of GM
Simagin, sacrificing the exchange
but preserving Black's long-range
bishop, which proves useful for
Black in both attack and defence .
On the other hand, 17 . . . ltfc8 leads

merely to a transposition of moves

after the continuation 1 8 xg7
rj;xg7 19 'ii'h6+ rj;gS.
18 'iixh6
ltfc8
19 ltdJ! (D)

B
Only this move at that time was
a novelty - by overprotecting c3,
White frees the e2-knight to par
ticipate in the attack. The theoreti
cal paths 19 ltdS or 1 9 g5 bring
White no great advantage, and fur
thermore Korchnoi was certainly
ready for them. But after the mod
est rook transfer he plunged deep
into thought.
19
l14c5
This move apparently leads to a
forced win for White. However, af
ter 1 9 ... e6 20 gS lbh5 2 1 lbg3
'it'eS 22 lbxh5 gxh5 23 'ii'xhS as
well, he has an appreciable ad
vantage. Botvinnik's continuation
19 ...'ili'd8 was the most stubborn.
20 g5
...

36 Moscow Ct (2) 1 974

The knights on c3 and f6 are

both defending their kings, and
therefore it is precisely these
pieces which are exposed to the
greatest danger (if the black knight
moved from f6, it would immedi
ately invite a white invasion on the
square d5).
llxgS (D)
20

- ..
..

.. ..
-
illit

)";-.
;l
_m
- .e.z.J.

A
.?:.'.

-0
!!!f}

.t
. -

- .
"
j

C
w

w
21 lidS!
Not, of course, 21 lt:Jd5 llxd5 and
the main guardian of Black's for
tress, the f6-knight, remains alive.
llxdS
21 ...
lle8
22 lt:JxdS
Now the queen can no longer
manage to return to her camp:
22 . . .'ifd8 23 lt:Jef4 'ii'f8 24 lt:Jxf6+
exf6 25 'fi'xh7#; if 22 ... lbh5, then
White wins by 23 lbxe7+ h8 24
lt:Jxc8.
23 lt:Jef4
..tc6 (D)
It is necessary to control the d5square, as otherwise lt:Jxf6+ and

- .. ..

.
....
lW

w
lt:Jd5, mating. After the alternative
23 . . . ..te6, I had prepared 24 lt:Jxe6
fxe6 25 lt:Jxf6+ exf6 26 ifxh7+
r.fi>f8 27 ifxb7 'fi'g5+ 28 bl lle7
29 'ii'b8+ lle8 30 'ii'xa7 (but not 30
llh8+?? <l;g7 and Black even wins,
due to the threat of 3 1 . . .'ii'g l#)
30 . . .lle7 3 1 'ili'b8+ lle8 32 'ili'xd6+
with a form of 'windmill' .
24 eS!
Cutting off everything on the
fifth rank. I was almost dazzled by
the wealth of apparently effective
possibilities, but only this continu
ation appears to be decisive. The
straightforward 24 lt:Jxf6+ exf6 25
lt:Jh5 'ili'g5+ (the point !) 26 'ifxg5
fxg5 27 lt:Jf6+ g7 28 lt:Jxe8+
..txe8 would not have won.
24 ...
..txd5
After 24 ... dxe5 25 lt:Jxf6+ exf6
26 lt:Jh5 mate is inevitable.
exf6
25 exf6
The main thing is to be strong to
the finish ! It is still not too late to

Karpov - Vaganian 37

lose: if 26 lLlh5 (to meet 26 . . . gxh5

by 27 l:tg1 + and 28 "ilg7#) then the
sobering 26 ...l:te1+.
26 "ii'xh7+
<iitf8
27 "ilh8+
After 27 .. .rt;e7, 28 lLlxd5+ ii'xd5
29 l:te1+ would be decisive.
1-0
In the second half of 1 974 this
game won the best game prize in
lnformator, with 89 marks out of
90 ! In the almost thirty-year his
tory of this contest, never have the
grandmaster jury been so unani
mous!

support, . . . b7-b5, has already been

prepared. But this rook's pawn
move has one essential defect - it
does little to assist piece develop
ment.
6 dxcS
Apparently the simplest solu
tion to the problem. However, 6
.te2 also gives White reasonable
prospects.
.txcS
6
.tb6 (D)
7 lLlb3
The bishop has many possible
retreats - besides this, it could also
go to e7, d6 and a7. It is difficult to
say which one of these is the safest.
..

Game 7
Karpov - Vaganian
Skopje 1976
French, Tarrasch

e6
1 e4
2 d4
dS
cS
3 lLld2
exdS
4 exdS
a6
5 iLlgf3
In those days I almost always
chose the variation with 3 lLld2 in
this opening. And my opponent,
for whom the French Defence was
then and is to this day a fundamen
tal weapon for Black in reply to 1 e4,
had prepared a quite rare continu
ation. The aim of it is to prevent the
bishop appearing on b5. Besides
that, when the occasion arises he is
prepared to play . . . c5-c4, and its

w
8 .td3
8 .tg5 gives White nothing:
8 ... lLle7 9 'ii'd2 lLlbc6 10 .te2 0-0
1 1 0-0 h6 12 .te3 l:te8 1 3 l:tad1
lLlf5 ! 14 .txb6 'ii'xb6 15 l:tfe 1
.te6.
lLle7
8
lLlbc6
9 0-0

38 Skopje 1976

J. g4
10 :et
Both sides are making their own
moves, seemingly with little regard
for the other. The battle is being
fought around the critical d4square. If Black manages to ad
vance his central pawn to d4, he
will have a reasonable game. If, on
the other hand, White can guaran
tee control of this important square
for his pieces, his advantage from
the opening will not be in doubt.
In connection with this, 1 1 J.e3
would now be a mistake because of
1 l . ..d4.
1 1 c3 (D)
Immediately pushing back the
enemy light-squared bishop by 1 1
h3 makes no sense, as Black can
not castle at the moment in view of
the sacrifice on h7.

11
12 h3
13 J.e3

h6
J.hS
0-0

Exchanging dark-squared bish

ops is in principle not to Black's
advantage, although, for example,
in the event of 1 3 . . . J.c7 he is even
further behind in development.
14 J.xb6
'ifxb6
15 'ife2
This does not look very logical
the queen is standing in front of the
rook on an open file. Nevertheless,
Black does not have time to make
use of this fact: after 1 5 . . .l:tfe8 his
rook is insufficiently defended.
15
l:tfd8
16 l:tad1
aS (D)

w
Vaganian is trying to become ac
tive on the queenside by forcing
the white knight from b3. He prob
ably underestimated the ensuing
pawn sacrifice, after which the
game takes on a forcing character.
17 J.b1
It may seem superfluous to re
peat that White's main idea is to

Karpov - Vaganian 39

control d4, but he has to watch it

constantly. Therefore the move
'it'e3, which enters into White's
plans, would in this specific case
be unsuccessful: 1 7 'it'e3 ti'xe3 1 8
l:.xe3 i.xf3 1 9 l:.xf3 a4, and Black
has everything under control.
i.xf3
17
18 'it'xf3
a4
19 lbd4
'it'x b2
Not 19 ...lbxd4 20 l:.xd4 'it'xb2,
as after 2 1 l:.b4 Black loses a piece
(the rook not only defends the
bishop, but also blocks the a3-f8
diagonal).
20 lbxc6
Black's final piece is diverted
away from defending its king.
20
lbxc6 (D)

c3-pawn, and keeping the major

black pieces out of the game, but is
also creating the threat of a bishop
sacrifice on g6, with a mating at
tack).
22
l:.d7
After 22...l:.e8 I was planning to
continue 23 i.xg6 fxg6 24 ti'xg6+
'it>f8 25 ti'xh6+ 'it>g8 26 l:.e6.
23 i.fS! (D)
The point!
...

B
l:.e7
23
After 23 . . .l:.c7 White would
continue with the simple 24 l:.xd5;
23 ... gxf5 loses due to 24 l:.d3 (it
must be precisely this rook, cover
ing the b1 -h7 diagonal, as the other
white rook has to control the e-file)
24. . .f4 (24 . . . lbe7 25 l:.xe7) 25
'ii'xf4 ti'c2 (White was threatening
26 'ilfg4+, but 25 . . .f6 is probably
more stubborn) 26 l:.g3+ 'it>h7 27
'ilff6 l:.g8 28 l:.xg8.
24 l:.xe7
..

w
g6
21 'ii'fS
22 'it'f6
The white pieces are very har
moniously placed (in particular,
the queen is not only defending the

40 Skopje 1976

I could also have gained an ad

vantage by means of 24 .i.xg6 fxg6
25 llxe7 li:Jxe7 26 'iixe7, but I
wanted more.
li:Jxe7
24 ...
li:JfS
25 .i.d3
The only move, as there are no
other possibilities for the knight
(25 . . . li:Jc6 26 .txg6); 25 ....:te8 26
.:tel 'iia 3 27 .i.b5 is no good either.
Defending the knight with the king
is also unsuccessful: 25 ... f8 26
llbl 'iixa2 27 .:txb7, and now both
27 ... lle8 and 27 ...1\Va3 should be
met by 28 .txg6.
26 .i.xfS
gxfS (D)

27 .:tel!
The black rook must not be al
lowed onto his third rank, a s a fter
....:ta6-g6 he could gain reasonable
counterchances linked with the
dangerous passed a-pawn, by giv
ing up some pawns.
'iixa2
27 ...

28 'it'xh6
The immediate 28 lle3 sug
gested itself, but then Black would
have had an interesting way of
gaining a draw due to his cunning
trap: 28 .. .f4 29 'iixf4 'iVbl + 30 h2
a3 ! 3 1 llg3+ 'tig6 32 .:txg6+ fxg6,
and here White has nothing better
than a perpetual check, as his
queen cannot even approach the
black rooks.
28 ...
a3
29 'figS+
The queen is heading for f6 with
tempo.
29
'i1i>f8
30 'iif6
'it>g8
31 'iixfS
'ifd2 (D)

w
32 .:te7!
A final finesse. The black rook
is now diverted from the a-file.
.:tf8
32
'i1i>h7
33 'ifg4+
'li'h6
34 .:teS

Tatai - Karpov 41

35
36
37
38
39
40

l:th5
'ii'f5+
l:txh6
'iff6+
'ii'xf7+
Wxb7

:as
<j;g7
<j;xh6
<j;b7
<j;bS
1-0

sufficient compensation for the

pawn, as in the game Kaiszauri
T.Georgadze, Sukhumi 1977.
8
e6
9 ltJge4
ltJb6! (D)
It is ridiculous to defend the
pawn with the bishop - 9 . . i..f8,
while 9 'ii'e7 loses straight away
in view of 10 ltJxdS exd5 1 1 ltJc3.
...

Game S
Tatai Karpov
Las Palmas 1 977
English Opening

...

1 ltJf3
c5
2 c4
ltJf6
d5
3 ltJc3
4 cxd5
ltJxd5
g6
5 g3
6 ..tg2
..tg7
The so-called Modem Variation
of the English Opening. The stand
ard continuation here is 7 0-0, but.
7 ii'a4+
An entertaining idea. White
strives to make immediate use of
his small advantage in develop
ment and the lack of harmony
among the black pieces.
ltJc6
7
After 7 .....td7 8 Wc4 ltJb4 9 0-0
ltJ8c6 10 'ii'xc5 it is not easy for
Black to prove that he has compen
sation for the pawn.
8 ltJg5
In the event of 8 Wc4 ltJdb4 ! 9
0-0 'ilfaS ! 1 0 ltJe4 ii'a6 ! 1 1 'ilfxc5
b6 1 2 'ile3 0-0 1 3 ltJe1 ..te6 14
ltJc3 l:tac8 B lack has more than
..

w
c4
10 'ii'b5
1 1 ltJa4
A logical continuation of the
plan which was begun on move 7 .
1 1 'iWcS is repulsed by means of
1 l . . . ..tf8, while 1 1 ltJcS 0-0 1 2
..txc6 bxc6 1 3 'ii'xc6 e5 would give
Black a dangerous initiative.
0-0
11
12 ltJxb6
axb6
13 ii'xc4 (D)
The critical position. White has
got his way, having won the c
pawn. Black meanwhile has man
aged to remove his king from the

42 Las Palmas 1 977

B
centre and is ready to get down to
action. Incidentally, the fact that
the a-file has been opened is un
doubtedly to his advantage (the
weakness of the doubled pawns
will only tell in the endgame, but
that is still a long way off). Be
sides, the white queen has made an
early exit from the centre, and now
one of Black's problems is how to
use this fact wisely, and this, in
conjunction with other threats,
should guarantee him a definite ad
vantage.
e5!?
13
1 3 . . . .i.d7 also looked tempting,
in order to move the rook to the c
file as quickly as possible. In this
case White has to be on the look
out for danger with every step.
However, he does have an interest
ing resource:
a) Not 14 'ii'c 2? lbd4 15 'ii'b 1
.i.a4 16 b3 .i.xb3 17 axb3 l:txa1 1 8
'ii'x a1 lbc2+.
...

b) 14 0-0 lbd4 1 5 lbc3 (the only

defence against 1 5 ....i.b5) 15 . . . b5
1 6 'ii'd3 b4 17 e3 lbb3 is no use to
White either.
c) White can try to take the d4square away from the knight by
means of 14 e3, but then 1 4...e5 is
better for Black.
d) 14 lbc3 ! lbd4 15 'ii'd3 ! .i.c6
( 1 5 . . .b5 looks promising, but here
as well there is a satisfactory de
fence 16 e3 ! lbb3 17 l:[bl lbxcl 1 8
:xc 1 b4 19 lbe2 lha2 20 'ii'b3 fol
lowed by 2 1 d4) 1 6 .i.xc6 ( 1 6 0-0 is
bad due to 16 . . .lbb3 ! 17 'ii'xd8
l:[fxd8 1 8 l:[b1 lbxd2 19 .i.xd2
:xd2 with an advantage for Black)
1 6 . . . bxc6 1 7 e3 ! - the point! 17 . . . lbb3 (White has removed the
important knight, without drop
ping his the queen: 1 7 . . . lbf3+ 1 8
e2), and 1 8 'ii'xd8 l:[fxd8 19 l:[b1
lbxc 1 20 :xc 1 .i.xc3 2 1 bxc3
:xa2 22 d4 gives rise to an equal
ending.
It is not surprising that it took
me an hour to decide between
1 3 . . ..i.d7 and the move played in
the game.
14 'ii'c2
The only move. 14 0-0 is bad
due to 14....i.e6, when the queen
has no good square. 14 lbc3 is also
weak in view of 14 ... .i.e6 15 'ii'e4
( 1 5 .i.d5 is answered by 15 ... b5 !)
1 5 ....i.f5 followed by 16...lbd4 .
14 ...
lbd4

Tatai - Karpov 43

15 ii'b1
fS
e4 (D)
16 ltlc3
1 6 . . .b5 allows White the ade
quate response 17 e3, but perhaps
I should have paid more attention
to the preparatory 16 ... -te6, threat
ening 17 .. .-txb3.

21

'it'xd4 (D)

w
17 d3
1 7 e3 gives White a hopele ssly
weak position after either 17 ... ltlc6
followed by 1 8 . . . ltle5 or 1 8 ... ltlb4,
or 17 ... ltlf3+ 1 8 -txf3 exf3.
17
bS
18 -te3
. Again the only move. 18 e3 ltlf3+
1 9 -txf3 exf3 20 ltlxb5 a5+ 2 1
ltlc3 b5 serves no purpose at all.
b4
18
19 ltld 1
.:.es
fxe4
20 dxe4
21 -txd4
But not 21 -txe4? in view of the
reply 2 1 . . . .:.xe4 22 'it'xe4 -tf5 and
... ltlc2+.

22 a3
22 0-0 allows Black a pleasant
choice between the two possibili
tie s 22 ... -tg4 and 22 . . . ii'd2. The
latter is more promising, as White
is obliged to go for the variation 23
-txe4 -th3 24 -tg2 -txg2 25 <iPxg2
.:.xe2 26 'ilfc1 'ii'd5+ 27 'ittg 1 -td4
with good attacking prospects for
Black.
22
-tg4
Black guesses his opponent's
crafty little idea: 22 ... bxa3 23 0-0,
making use of the undefended po
sition of the aS-rook.
23 c2 (D)
Ruling out the threat 23 . . .-tf3,
because of 24 -txf3 exf3 25 b3+
and 26 xf3 .
'it'd3!
23
It is rare that one can place the
queen under attack from a pawn .
24 exd3

44 Las Palmas 1977

c4 l:lxc2+ 29 xd5 followed by

29 ....i.f3+.
27
l::txc2+
28 xb4
l:lcd2 (D)

- - -

- -

B
Allowing the attack to develop
elegantly. 24 lbe3 would have lost
immediately to 24...'ilfxc2 25 lbxc2
.i.xb2. 24 .:le 1 would not have
brought White any relief due to the
simple 24 ...bxa3, but even the su
perior 24 'ii'd2 would have left
B lack a number of promising pos
sibilities, e.g. 24 ... 'ii'xd2+ 25 xd2
.:ladS+ 26 e 1 , and now either
26 . . J::tc 8 threatening 27 ...l:lc2, or
the immediate 26 ... .i.f3 27 l:lgl
(both 27 exf3 exf3+ 28 fl fxg2+
29 xg2 b3, and 27 .i.xf3 exf3 28
e3 lead to a difficult endgame for
White) 27 ....i.xg2 28 l::txg2.
24
exd3+
25 d2
l:le2+
It is possible that White was not
expecting this check when he was
considering his own 24th move.
26 xd3
l:ld8+
27 c4
Black emerges with an extra
piece after 27 .i.d5+ l:lxd5+ 28

B B B B
.
. ...
... .
.

f .
g .
- ;,- -

- --- D.i.r,; ;;;_

a ttJ a :
,,

/,

0
,

' /

r:'

!;;;:

"

w
Again White's knight is the key
piece ! It contains the action of all
the black pieces, and must be at
tacked immediately.
29 f3
Forced, as otherwise the b2pawn will be lost.
.i.f8+
29
30 aS
After a retreat to any other
square the g4-bishop jumps aside
with check and the g2-bishop is
lost.
.i.d7!
30
The white bishop is under at
tack, and if 3 1 .i.fl (or 3 1 lbe3),
then 3 1 ... .i.c5, and there is no way
of saving White from mate by the
rook.
0-1

Karpov - Korchnoi 45

Garne 9
Karpov - Korchnoi
Baguio City Wch (14) 1978
Spanish, Open Variation

1 e4
e5
2 .!iJf3
.!lJc6
3 .i.b5
a6
4 .i.a4
.!iJf6
5 0-0
.!lJxe4
The Open Variation of the Span
ish, one of the most popular open
ings in both my matches with
Korchnoi for the world chess crown.
One could almost write a whole
book about the games we played
with this opening. In this encoun
ter, for the first time in our theoreti
cal dispute I managed to use an
extremely valuable novelty in this
variation.
6 d4
b5
7 .i.b3
d5
8 dxeS
.i.e6
9 c3
.i.cS
10 .!iJbd2
0-0
1 1 .i.c2
.i.f5
12 .!iJb3
.i.g4
1 2 . .i.g6 has the same value.
13 h3
.i.h5
14 g4! ?
.i.g6 (D)
In opening manuals at the time
this position was said to be totally
in Black's favour. In fact, if you are
referring to the rniddlegame, the
open position of the white king is
the basis for this assessment. But
.

w
I thought of an interesting idea
linked with transferring into the
sort of endgame where opposite
coloured bishops do not diminish,
but on the contrary even increase,
White's advantage.
dxe4
15 .i.xe4
exf3
16 .!lJxc5
17 .i.f4! (D)

B
If White exchanges queens im
mediately by 17 Wxd8, then he
loses a tempo and Black has time

46 Baguio City Wch (14) 1978

to organise counterplay against the

e5-pawn. The point of my play lies
in the fact that after the move
played Black is forced to exchange
queens, which gives White two ex
tra tempi.
17
'ii'x d1
17 ...'Wh4 is insufficient in view
of 1 8 'ti'xf3, defending the h3pawn and attacking the c6-knight.
After 17 ..."ile7 White wins a pawn:
1 8 'ii'd5 .!Lla5 1 9 b4 .!Llc4 20 'ii'xf3,
and 20 . . ..!Llxe5 is impossible due to
2 1 .i.xe5 'ti'xeS 22 .!Lld7.
18 .:.axd1
.!Lld8! (D)
Korchnoi finds the best response
to White's unexpected move. Ob
viously, Black cannot tolerate the
beautifully placed white knight on
cS. Exchanging off into a rook end
ing with opposite-coloured bish
ops promises him some chances
for defence. Naturally, I had fore
seen this possibility.

.!Lle6
19 .:.d7
fxe6
20 .!Llxe6
.:.ac8
21 .i.e3
Exchanging one pair of rooks
(2 l ... .:.f7) only weakens the possi
bilities for defence. After 22 .:.fd 1
White has enough squares to main
tain the rook on the seventh rank.
22 .:.rd1 (DJ
22 .i.cS immediately deserved
attention, so that after 22 . . . .:.fe8
(if 22....:.fd8 { or 22 . . ..:.f7 } . then 23
.:.fd 1 ; 22 ....:.f4 leads to unclear
consequences) White can play 23
.:et .

.
1
-;;; :
,
-.
. . .... .

- - .
0%f;
"
QJ
-

E
.
"

;,:
-

: ;r"'
m1
'I'll'
.
.

. t=. .

B
22 ...
.i.e4
Moving the bishop to d5 in an
attempt to disturb the co-ordina
tion of the white rooks is Black's
only realistic defensive possibil
ity. If 22 . . .h5, then 23 gxhS .i.xh5
24 .:.e7, invading with the second
rook.
.:res
23 .i.cS

Karpov - Korchnoi 47

Here is the whole point of

B lack's play; he does not put his
rook on either f4 or f7, offering to
exchange, but takes the e7 -square
under his control, and the white
rook becomes short of space on the
seventh rank.
24 ll7d4
.i.dS
25 b3
aS
Again the only possible attempt
to create counterplay, as otherwise
nothing will oppose White's gen
eral advance on the kingside.
lla8
26 h2
If 26 ... a4, then 27 c4 bxc4 28
bxc4 .i.c6 29 a3 .
lla6
27 g3
27 ... a4 would have been more
logical, and in reply to 28 c4 both
28 ...bxc4 29 bxc4 .i.c6 (now White
does not have time for 30 a3 in
view of 30 ... lla5 when the e5pawn perishes), and 28 ... .i.c6 are
possible, in order to exchange a
few more pawns, although in this
case as well White has a clear posi
tional advantage.
Korchnoi 's move is part of an in
correct plan.
28 h4
llc6 (D)
Continuing the same mistaken
plan. Korchnoi clearly underesti
mated his opponent's idea, which
is described after the next move.
29 llxdS!
White immediately gains a
bishop and pawn for the rook, and

besides, the beleaguered black

f3-pawn is already under attack by
the king, and the queenside pawns
are also becoming objects for at
tack.
29
exdS
30 llxdS
llce6
31 .i.d4
c6
32 lieS (D)
Korchnoi may have underesti
mated the power of this move. If he
had played 32 lld7, Black would
have had the very strong reply
32 . . . c5 at his disposal, exchanging
off one of White's fundamental
trump cards - the e5-pawn. Now
the white rook is occupying a very
powerful position. White wants to
isolate the queenside pawns by
means of a2-a4, and the f3-pawn is
already under attack.
llf8
32
In the game it would have been
hard to decide on the continuation
32 ...lld8 33 xf3 lld5 34 .:.Xd5 (or
...

48 Baguio City Wch (14) 1978

B
34 f4 lhc5 35 .txc5) 34 . . . cxd5.
There have been many arguments
about this position. I suggest that
the mobility of the white king side
pawns should in the final analysis
lead him to victory, although it is
not as simple as it appears at first
glance. In particular, after 35 e3
35 ....:h6 36 h5 g6 the tempting 37
f4 does not work in view of
37 ...gxh5 38 g5 h4 ! .
35 a3 was my intended continu
ation, in order to create a fortress
on the queenside with b3-b4, and
only afterwards advance the kingside pawns, for example 35 ...g6 36
g3 .:e8 37 f4 and if 37 ....:f8,
then 38 e6 with the plan of h5, and
if ... gxh5, then f5, creating two
connected passed pawns.
Nevertheless, trying to open up
the game on the kingside was
Black's only chance of survival,
since the passive defence to which
Korchnoi condemned himself on

his last move gives him no hopes of

saving himself.
33 a4!
In the first place this fixes the
objects for attack on the queenside,
in order to chain the black rooks
down to their defence and divert at
tention from operations on the op
posite side of the board.
33
bxa4
34 bxa4
g6
35 .:xa5
.:ee8
In the event of 35 . . . h5 the move
36 .:a6 ties Black down to defend
ing the c6-pawn and prepares the
advance of the a-pawn.
36 .:a7
.:r7
37 .:a6
A cunning little move, forcing
the black rook to occupy a passive
position.
37
.:c7
37 ...c5 does not work because of
38 .txc5 .:xe5 39 .:a8+ g7 40
.td4.
38 .tc5 (D)
After the bishop has reached d6,
the position should be considered
won.
.:cc8
38
.:as
39 .td6
40 .:xc6
.:Xa4
h5
41 xf3
42 gxhS
gxh5
.:a2
43 c4
44 .:b6
f7
.:a4
45 c5
..

Karpov - Korchnoi 49

e6
46 c6
d7
47 c7
l:c8
48 l:b8
49 e3
Hoping for 49 e6+? xe6 50
l:xc8 xd6 and the possibility of a
rook ending with f- and h-pawns.
49
l:xh4
50 e6+!
1-0
Since 50 . . .'it>xe6 (50 ... xd6 5 1
l:xc8 l:c4 5 2 l:d8+ 'it>xc7 5 3 e7)
51 .i.g3 is decisive.

Game 10
Karpov - Korchnoi
Baguio City Wch (32) 1978
Pirc Defence
The battle in Baguio lasted for
more than two months, and things
started rather successfully. After
game 27 the score was 5-2, and it
looked as though the match was
coming to an end. But excessive
confidence weakened me, and aided

by the fact that he had nothing to

lose, my opponent equalled the
score at 5-5. In four games I lost
everything I had gained in the pre
vious 27 . I was shattered. All the
same I managed to cast off the bur
den of painful blunders and made
up my mind that the 32nd game
was to be the decisive one. I played
quietly, confidently, and, having
gained an overwhelming position,
instilled myself with the thought,
'just do not hurry ! ' , as I knew that
the desire to realize my advantage
had already let me down more than
once.
1 e4
d6
2 d4
lbf6
3 lbc3
g6
4 lbf3
.i.g7
5 .i.e2
0-0
c5
6 0-0
In the 1 8th game when the same
opening occurred, play proceeded
along the main variation 6....i.g4 7
.i.e3 ltlc6, and here I introduced
the novelty 8 'ilfd3 ? ! . Wishing to
avoid another surprise, Korchnoi
this time chose another plan. In
principle the position after 6 . . . c5
7 dxc5 dxc5 is thought by theory
to be slightly better for White, but
I thought my challenger might
have prepared an improvement,
so I avoided the early exchange of
pawns.
7 d5
lba6

50 Baguio City Wch (32) 1978

8 i.f4
liJc7
9 a4
b6
10 l:le1
i.b7
1 1 i.c4
The prophylactic 1 1 h3, with the
aim of maintaining the bishop on
the h2-b8 diagonal, would have led
to a double-edged game after
1 1 ...'i'd7 12 i.c4 l:lad8 1 3 W'd3 e5.
liJh5? (D)
11 ...

a) 19 liJxd5 is simple and good:

19 ... i.xd5 ( 1 9 ... 'ii'xd5 20 'i'g3 'ifa2
2 1 i.c4 W'xb2 22 lbxt7) 20 c4 i.a8
(or 20 . . .i.e6 2 1 liJc6 'ifxd3 22
lbxe7+) 21 liJd7.
b) 19 liJd7 liJxc3 20 bxc3 liJxd7
21 i.xe7 and Black suffers a loss of
material.
In any other competitive situ
ation I would certainly have played
like that, but in this game I did not
want to take risks.
liJd7
15
Preventing a break in the centre
and hoping for counterplay with
16 ...b5 17 axb5 liJb6.
16 'ife3
i.a8
b5
17 i.h6
18 i.xg7
xg7
ttJr6
19 i.n
axb5 (D)
20 axb5
A pawn attack without a fi
anchettoed bishop on g7 is like an
infantry attack without a prelimi
nary bombardment.
..

Losing valuable time. 1 1 . . .iid7

was necessary, trying at any cost to
provoke tactical complications: 1 2
e5 dxe5 1 3 lbxe5 'i'f5 14 lbxg6
hxg6 1 5 i.xc7 lbg4. However, af
ter 12 'i'd3 White has the better
chances.
12 i.g5
liJf6
13 'i'd3
a6
l:lb8
14 l:lad1
15 h3
15 e5 ! looks logical. 15 ...dxe5 16
lbxe5 b5 ! 1 7 axb5 axb5 1 8 i.xb5
lbcxd5 and now:

Karpov - Korchnoi 51

21 t:De2
..tb7
Black's misfortune lies in the
poor communication between his
pieces. If he had managed to carry
out the manoeuvre ...e5, his pieces
would have had an easier passage
from one side of the board to the
other. However, 2 1 ...e5 22 dxe6
t:Dxe6 23 lDg3 would have weak
ened his pawn structure, and Black
would have been forced to move,
all but crawling, from one flank to
the other along the 8th rank.
:as
22 lDgJ
%:ta4
23 c3
24 ..tdJ
'iia8 (D)

White's knife blow has a firm

tactical basis: 25 ...lDfxd5 26 lDh5+
(or 26 lDf5+) 26 ... gxh5 (26 ...h8
27 'fih6 %:tg8 28 lDg5) 27 'ilfg5+
h8 28 'ilff5.
t:DcxdS
26 'ilfxeS
%:ta7
27 ..txbS

28 lDb4 (D)

B
Now not only do threats hang
over the black king, but the prosaic
c3-c4 is also unpleasant. Thus, af
ter 28 ... i.c6 there is the possibility
of 29 i.xc6 'ilfxc6 30 c4 lDb4 3 1
%:td6 exd6 32 lDh5+ gxh5 3 3 'ii'g5+
h8 34 'ii'xf6+ g8 35 lDf5,
whilst 28 . . .'ii'b 8 29 c4 'ii'xe5 30
%:txe5 leads to a hopeless ending.
[Editor's note: Black then has
the surprising resource 30 ...l:.a5,
when it is not even clear that White
is better, since 3 1 cxd5? l:.xb5
leaves White with some very weak
pawns, while freeing the b5-bishop
is not possible by other means. Af
ter 28 . . .'ii'b 8, White might do bet
ter to try 29 'iig 5.]
28
..tc8
29 i.e2!
You must agree that there is
something attractive about this
bishop move.

52 Montreal l979

29
30 c4
31 'ifxcS
32 n

e6
lbb4
'ifbs
l:r.c8

White has an extra pawn and a

menacing initiative.
l:r.g8
35
36 lbf3
'iff8
37 'ilfe3
g7?!
37 ...l:r.b7 would have prolonged
the battle by preventing movement
by the pawns.
d7
38 lbg5
ila8
39 b4
lba5
40 b5
41 b6
In this position Korchnoi sealed
the move 4 1 . . .l:r.b7, but the follow
ing day resigned the game and the
match.
1-0

Game 1 1
Timman - Karpov
Montreal l979
English, Four Knights
A game from the 'tournament of
stars' which is noteworthy because
the Dutch grandmaster fell into an
opening trap I had stored up for
Korchnoi, but it was not played un
til after the Baguio match.
lbf6
1 c4
2 lbc3
e5
lbc6
3 lbf3
4 e3
4 g3 is played more often, but
because in the Baguio match I had
managed to achieve a good game
in a variety of lines of the Four
Knights English, Timman chose a
rarer move.
4
e7
Here the more popular continu
ation is 4 .b4. At the time this
modest bishop move to e7 had dis
appeared from the scene. However,
this very game changed its assess
ment and it again attracted the at
tention of theoreticians.
5 d4
As will become clear, this natu
ral move gives Black a wonderful
game. Evidently, 5 e2 0-0 6 0-0
is a better option for White, and if
6 . . d5, then 7 cxd5 lbxd5 8 d3 leads
to a reversed Scheveningen Sicil
ian with an extra tempo for White.
..

..

Timman - Karpov 53

exd4
5 ...
6 xd4
Or 6 exd4 d5 ! 7 cxd5 xd5 8
.i.b5 0-0 with a good game for
Black.
0-0
6
bxc6
7 xc6
8 .i.e2
dS
9 0-0
.i.d6
'fie7
10 b3
1 1 .i.b2 (DJ

B
This position had been seen in
practice before our game, and
moreover it was thought to be in
White's favour. Thus, 1 1 . . ..:r.d8 12
cxd5 'fie5 ( 1 2 . . . cxd5 1 3 b5 .i.a6
14 d4 ! .i.xe2 1 5 "ii'xe2 'ii'e5 16
g3 l:te8 1 7 l:tac 1 gives White a po
sitional advantage) 1 3 g3 .i.h3 14
.:r.e1 .i.b4 1 5 'ifc2 .i.f5 1 6 'ilc l
cxd5 17 .i.f3 'fie7 1 8 a3 .i.a5 19 b4
.i.b6 20 xd5 ! is obviously to
White's advantage; Keene-Jans
son, Haifa OL 1976.

1 1 ...
dxc4!
A move which sharply changes
the assessment of the position. This
idea was also prepared for the Ba
guio match.
Black solves his fundamental
problem by completely widening
the action of his pieces, which are
directed towards an attack on the
kingside.
12 bxc4
If White had taken with the
bishop all sorts of attacking ideas
would have appeared for B lack,
linked with the weakened control
over g4, for example 1 2...'fie5 1 3
g 3 .i.h3, 1 2. . .g4 1 3 g 3 xh2 or
1 2 . . . .i.xh2+ !?. Now the queen's
rook joins the battle with tempo.
12 ...
l:tb8!
13 'i!fcl (DJ
Not falling into the trap 13 .:r.b1 ?
.:r.xb2 1 4 .:r.xb2 'ii'e5 ! .

13 ..

g4

54 Montreal l 979

Black's attack is developing

easily and naturally. Not a shadow
remains of White's opening advan
tage, and on the contrary he is
watching with alarm as events de
velop on the kingside.
14 g3
After 14 xg4 xg4 1 5 l:r.e 1
Black increases the pressure with
1 5... l:r.b4 ! .
1 4 ...
l:r.e8
Of course, I could have played
14 . . . lDxh2 1 5 xh2 'ili'h4+, imme
diately forcing a draw. But now the
threat of the knight sacrifice on h2
is not so harmless, since the rook
would be included in the game
along the sixth rank with decisive
effect: 15 . . .lDxh2 1 6 xh2 'ifh4+
1 7 g2 'ifh3+ 1 8 g1 xg3 1 9
fxg3 'ifxg3+ 2 0 h 1 l:r.e6 and the
rook reaches h6.
15 lDd1 (D)
After 1 5 f3 there follows
15 ...'ili'f6 16 xg4 ( 1 6 g2 is bad:
1 6 ... 'ili'h6 1 7 h3 lDe5 ! ) 1 6 . . . xg4
17 f3 h3 1 8 l:r.f2 'ii'g6 with a pow
erful initiative for Black.
15
lDxh2!
The sacrifice which helps Black
to increase his advantage. Timman,
of course, saw this blow, but placed
his hopes on the strength of his
replying intervening move. How
ever, when I began the combina
tion, I had carefully taken into
account all the tactical nuances.

ltt

Wi
- -
A

%,

u
- Bi. B

a NI
: -
, "'Z.J : %=
/.
%;
"

/
/

"

16 cS
Not, of course, 1 6 xh2 'ii'h4+
17 g2 'ii' h 3+ 1 8 g1 xg3 1 9
fxg3 'ili'xg3+ 20 h1 l:r.e4 ! (or
20 . . . l:le6 2 1 f6) 2 1 l:lf4 h3 and
everything is over for White.
16
lDxfl!
17 cxd6 (D)

Korchnoi - Karpov 55

for White, he has absolutely no

choice.
18 fxg3
'ii'xd6
19
'ii'h6
20 ..td4
'ii'h2+
21 'it>e1
'ii'xg3+
22 'it>d2
'ii'g2
23 ltlb2
..ta6
24 ltld3
..txd3
25 'it>xd3
l:.bd8
26 ..tn
'ii'e4+
cS! (D)
27 'it>c3

w
Clearing the final approach to
the white king.
28 ..txcS
'ii'c6
29 'it>b3
l:.b8+
30 'it>a3
:es
31 ..tb4
'ii'b6
0-1
The merit of the novelty used in
this duel was recognized - it won
the lnformator prize for the most
important theoretical game. Re
markably, in the very same issue of

lnformator it also won the prize for

the best game ! In the almost thirty
year history of these prizes, this
double has only happened twice.
Game 1 2
Korchnoi - Karpov
Merano Wch (9) 1981
QGD, Orthodox

1
2
3
4
5
6
7

c4
ltlc3
d4
ltlf3
..tgS
..th4
:et

e6
dS
..te7
ltlf6
h6
0-0
dxc4! (D)

w
Amazingly, this simple move
had never been seen in grandmas
ter practice before. Chess really
is inexhaustible - even in such a fa
miliar opening as the Queen's
Gambit it is still possible to think
up something new on move seven.

56 Merano Wch (9) 1981

8 e3
After 8 e4 llk6 ! Black has quite
a playable game.
8
cS
9 i.xc4
cxd4
10 exd4
The idea of an early exchange
on c4 was used by Kasparov in
game 23 of our first match. I cap
tured on d4 with the knight, and it
quickly ended peacefully after 1 0
lL!xd4 i.d7 1 1 0-0 lL!c6 1 2 lL!b3
%lc8 1 3 i.e2 lL!d5 ! with a mass of
exchanges.
lL!c6
10
1 1 0-0 (D)

bishop on h4 directed me towards

an unusual idea, so I abandoned the
standard idea of playing ...lLlc6-b4d5.
11
lL!hS!
Exchanging bishops is useful
for Black, but the straightforward
1 1 . . .lL!d5 12 i.g3 would allow
White to avoid it, and in the event
of 1 1 ...lL!e4 1 2 i.xe7 lL!xc3 1 3
bxc3 White i s clearly better.
12 i.xe7
After 12 i.g3 lL!xg3 1 3 hxg3
i.f6 White has immediate difficul
ties defending the d-pawn.
lL!xe7 (D)
12

B
A typical position with an iso
lated d-pawn has arisen. But there
is one important nuance which is
favourable for Black. In situations
like this, White's dark-squared
bishop does not usually hurry to
get into the game, but waits for the
best moment. The position of the

w
An important link in Black's
plan; one knight temporarily finds
itself at the edge of the board,
while the second has managed to
take control of d5 . Here White
could rid himself of the weak
pawn: 1 3 d5 exd5 1 4 lL!xd5 lL!xd5
1 5 i.xd5 lL!f4 16 i.e4 'ii'xd 1 1 7

Korchnoi - Karpov 57

:cxd1 i.e6 with equality. Korch

noi did not want to give up his ad
vantage as White so easily, and as a
result he fell into a difficult position.
13 i.b3
lbf6
14 lbe5
i.d7
15 'ilfe2
:cs
16 lbe4
White cannot find a clear-cut
plan, and fails to take account of
the fact that every exchange of mi
nor pieces reduces the dynamic po
tential of the d4-pawn.
1 6 :fe l looks expedient, inci
dentally preventing 16 . . .i.c6 be
cause of 1 7 lbxf7. In that case I
would have preferred to reply with
1 6...:c7 or 1 6...i.e8.
lbxe4
16
17 'ii'xe4
i.c6! (D)

an isolated pawn. In that case his

knight would be an able and reli
able defender of his pawn, while
also attacking the enemy d4-pawn.
At the same time the function of
the white bishop is limited.
18 lbxc6
:xc6!
Very precisely seizing the c-file,
in case White wants to move his rook
to the kingside. After 1 8 ...lbxc6 1 9
d5 exd5 20 i.xd5 the initiative
would have passed into Korchnoi's
hands.
'Mid6
19 :c3
20 g3
It was not worth hurrying to
open an escape hatch. Now White's
prospects on the kingside have dis
appeared.
:ds
20
21 :d1 (D)

58 Merano Wch (9) 1981

the possibility of attacking on the

kingside, it is important for Black
to preserve his major pieces from
exchange so that he can put pres
sure on the isolated pawn. At
tempting to build up on the d-file
does not yet work ( 2 1 . . .'ifd7 22
.i.a4), and he did not want to make
the move . . . a7-a6, as it is not yet
clear what formation the queenside
pawns will need to take on. The
text move chains the bishop to cov
ering the b2-pawn and leaves the
possibility for an attack by ....:tb4.
'iid7
22 'ifel
Consistently carrying out the
fundamental plan. 22 . . ..:tb4 is par
ried with 23 l:lc4.
23 .:tcd3 (D)
Attempts to gain counterplay on
the c-file are refuted by tactics:
23 .Uc5 .Ud6 24 .Udcl lLlc6 25 .i.a4
lLlxd4 ! 26 .i.xd7 lLlf3+ 27 fl
lLlxe1 28 .i.xe6 lLld3.

23

24 'iVe4

.:td6

In this situation relieving one

self of the weak pawn does not
work: 24 d5 lLlxd5, and Black un
ties himself by means of ...'iic 6,
... .:t6d7 and . . .lLlf6. Therefore he
would have to exchange minor
pieces and defend a difficult major
piece ending with a pawn less.
'ifc6!
24
25 'iff4
Here the break 25 'ifxc6 lLlxc6
26 d5 just loses the pawn after
26 ...lLlb4.
25 ...
lLld5
Forcing the queen to occupy an
unfortunate position. Now 26 'ife4
is impossible because of 26...lLlb4
27 'ifxc6 lLlxc6 28 d5 lLlb4.
26 'iVd2
'ifb6 (D)

w
The unpleasant threat of . . .lLlb4
has arisen. White stubbornly does
not wish to play a2-a3, as he is

Korchnoi - Karpov 59

afraid of weakening his bishop's

position. Korchnoi's patience was
by now exhausted, and further
more he had very little time left,
so he exchanged off minor pieces,
depriving himself of any tactical
chances whatsoever and condemn
ing himself to passive defence.
.:.xd5
27 .i.xd5
28 .:.b3
Beginning an unsuccessful op
eration, as a result of which the
white rook turns out to be 'offside' .
If Korchnoi had planned to play f2f4, he should have done it straight
away, without moving the rook
from the d-file.
28
29 '6'c3
30 f4 (D)

B
The only way of preventing
... e6-e5, but at the price of weaken
ing his own king.
30
b6!
..

31 .:.b4
b5
32 a4 (D)
32 a5 was threatened, while af
ter 32 .:.b3 Black should continue
32 ....:.cs and 33 ....:.c4.
...

B
bxa4
32 ...
Not for the sake of winning a
pawn, but to distract the white
pieces from the kingside.
aS
33 '6'a3
34 .:.xa4
Or 34 .:.c4 'ii'b5 35 'iWxa4 'iWxb2.
'6'b5!
34 ...
Chaining the queen to the de
fence of the unfortunately placed
white rook, and incidentally threat
ening to invade on e2. I thought
that White now had to choose 35
b3, in order to free the queen. This
would probably have led to a win
ning rook ending after 35 ....:.bs, al
though I will not hide the fact that I
wanted to finish the struggle in the
middlegame. I was helped in this

60 Merano Wch (18) 1981

aim by Korchnoi's organic dislike

of parting with material.
35 :d2
eS!
Here Black had several tempting
continuations (35 . . .:cs; 35 ... g5),
but, in making the text move, it was
necessary to have a sort of feel for
the geometry of the position. At
first I could not manage this, until I
found the 37th move, which gave
me personal aesthetic satisfaction
and was said not to have been ex
pected by the majority of those pre
sent in the press centre.
36 fxe5
:xe5
37 ii'a1
'W'e8! ! (D)

w
Making full use of the queen's
ability to strike on both the files
and the diagonals.
38 dxeS
:xd2
39 l:.xa5
Wc6
40 :aS+
<ith7
41 'W'b1+
g6
4 t ...:c2 is also good.

42 'ifn
Not, of course, 42... Wxa8? be
cause of 43 'ifxf7+, with perpetual
check.
43 h1
Wd5+
0-1
The idea of the knight thrust to
hS did not go unnoticed. This duel
won the next lnformator competi
tion for most important theoretical
game.
Game 13
Karpov - Korchnoi
Merano Wch (18) 1981
Spanish, Open Variation
My opponent's choice of opening
somewhat surprised me, as the pre
vious two games in the Ruy Lopez
had turned out to be rather difficult
for him, especially game 14. Obvi
ously at the end of the match other
openings seemed even more dan
gerous to Korchnoi.
e5
1 e4
tl:lc6
2 tl:lf3
a6
3 .tb5
4 .ta4
tl:lf6
tl:lxe4
5 0-0
b5
6 d4
d5
7 .tb3
i..e6
8 dxe5
tl:lc5
9 tl:lbd2
d4
10 c3
tl:lxe6
11 .txe6
12 cxd4
tl:lcxd4 (D)

Karpov - Korchnoi 61

w
The Open Spanish underwent
fundamental tests in Baguio and
Merano. In the final stage of the
second of these matches a telling
blow was inflicted on B lack. Now
after 1 3 tbxd4 1Wxd4 14 1Wf3 l:d8
1 5 a4 White achieves only an even
position, as in the old game Capa
blanca-Lasker, St Petersburg 1914.
In game 14 of this match I had al
ready used a novelty here, taking
Korchnoi unawares: 13 tbe4 ! . Hav
ing thought for a record amount
of time, 78 minutes, he answered
1 3 ....i.e7 14 .i.e3 tbxf3+ ( 1 4...tbf5
was the more accurate move my
opponent played in game 1 6, when
he came out of the opening with no
losses), and after 1 5 it'xf3 0-0 1 6
l:fd 1 e8 17 tbf6+ ! .i.xf6 1 8 exf6
it'c8 1 9 fxg7 l:d8 20 h4 ! he found
himself in a critical position. I then
gained my fifth victory. For this de
cisive encounter I had prepared yet
another dangerous surprise.

13 a4!
The fact that Korchnoi spent 54
minutes thinking testifies to the
strength of this novelty. The his
tory of game 14 was being re
peated.
13 ...
.i.e7 (D)
The rook move 1 3 . . . l:b8 loses a
tempo, and after 14 axb5 axb5 1 5
tbe4 .i.e7 1 6 .i.e3 White has a seri
ous initiative; however, 1 3 . . . .i.c5
14 tbe4 .i.b6 1 5 tbxd4 .i.xd4 1 6
tbg5 0-0 1 7 axb5 tbxg5 1 8 .i.xg5
.i.xf2+ 1 9 l:xf2 'ii'xg5 20 bxa6
xe5 deserves attention, as in the
game Tischbierek-Chekhov, Pots
dam 1 985, when after 2 1 h 1?,
21 ... b5 ! would have led to an
even game, although it is true that
after 2 1 'iWfl ! ? White would have
preserved slightly better chances.

w
14 tbxd4 (D)
The manoeuvre 14 tbe4 has no
independent significance, as in this

62 Merano Wch (18) 1981

case play transposes to game 14 of

the match (with 14 a4 instead of 14
e3). Neither 1 4 h3 nor 14 lle l
gives White many prospects. After
14 axb5 axb5 (a complex game
would have arisen after 14 . . .llxb5)
15 llxa8 'iba8 1 6 lbxd4 lJxd4 1 7
'ifg4 lle6 1 8 f4 White has an ap
preciable advantage.

B
lJxd4
14
It must be said that after the
Merano match various defensive
resources after both 13 lle4 and 1 3
a4 were found for Black. But find
ing an antidote demands more than
a little time. In particular, here cap
turing with the queen, 14 . . . 'ir'xd4,
was later tried several times. Here
is one revealing example: 1 5 axb5
'ifxe5 16 bxa6 0-0 1 7 'ir'a4 llfb8 1 8
a7 llb7 1 9 llf3 'ir'b5 20 'ikxb5
llxb5 2 1 e3, which gave White
slightly the better chances in Hjar
tarson-Smejkal, Bundesliga 1990.
...

15 lle4 (D)
Every move has many lines, not
all of which have been exhausted;
here for instance it is also possible
to exchange on b5.

Evidently, this is the primary

cause of B lack's difficulties; the
black knight should have remained
in its place. 1 5 ... 0-0 is more pre
cise, for example 1 6 axb5 llxb5
17 e3 'ir'c8 18 'ir'c2 1i'e6 19 f4
llad8 20 l:ta4 l:td7 2 1 l:tfa1 1i'd5
22 h3 f6 23 exf6 xf6 24 llxf6+
:Xf6 25 llxa6 llxa6 26 llxa6 lJd4
yielded Black some compensation
for the pawn in Adarns-Yusupov,
Hastings 1989/90.
0-0
16 e3
17 f4
Threatening to play 1 8 f5, and if
1 7 ... g6, then 1 8 g4. In a difficult
position Black finds the best de
fence.

Karpov - Korchnoi 63

17
18 lUxd1
19 l:[d7 (D)

'iVxd1
l:[tbS

B
..tfS
19 ...
It was necessary to continue
1 9... ..td8, although after 20 a5 lDf8
2 1 l:[d3 Black's position is not the
most pleasant.
lbd8
20 fS
21 aS!
Emphasizing the passive nature
of the black pieces; 2 1 .z:r.xc7 bxa4
22 ..td4 l:[b4 23 .z:r.d 1 .z:r.ab8 is not so
clear.
lbc6 (D)
21 ...
fxe6
22 e6!
lbeS
23 f6!
Defending against the unpleas
ant threat of 24 f7+ and 25 tbg5.
Simplifying by 23 ....z:r.d8 24 .z:r.xc7
.z:r.ac8 would have led to a hopeless
endgame for Black after 25 f7+
h8 26 .z:r.xc8 .z:r.xc8 27 .z:r.c t .
.z:r.cs
24 .z:r.xc7

w
24 . . . lbc4 25 ..tc5 ..txc5+ 26
lbxc5 gxf6 27 lbd7 .z:r.cs 28 lbxf6+
is no good.
.z:r.xc7
25 .:.act
26 .z:r.xc7
l:[dS (D)

w
27 h3!
There is no hurry, and this quiet
move deprives Black of the chance
of 27 ....z:r.d1 + 28 lbg4+.
h6
27 ...
He must not allow the knight to
invade g5.

64 M1rano Wch (18) 1981

28 :a7!
The most clear-cut way of real
izing the advantage. 28 et:'lcS gxf6
29 et:'lxe6 :d 1 + 30 'iltf2 i.b4 would
have activated the black pieces.
28 ...
et:'lc4
The rook is tied to the back rank:
28 ...:d1 + 29 'iPf2 :b1 30 i.d4 lLlc6
3 1 fl+ 'iPh7 32 :as et:'lxd4 33 :xf8
g6 34 et:'ld6, and the f-pawn will
become a queen.
29 i.b6
:bs
In the event of29 ...et:'lxb6 30 fl+
h7 3 1 axb6 Black cannot cope
with two passed pawns.
i.xcS+
30 i.c5
gxf6 (D)
31 et:'lxc5

a .

W&

"' .

u
-
. . .
" -
W&
u

W&
W&
-

w
At this point Black even has an
extra pawn. However, White's pos
session of the seventh rank, the dif
ference in possibilities of the
knights, and the abundance of weak
nesses in the black camp leave him
no hopes of saving himself.

32 b4
:ds
33 :xa6
f7
34 :a7+
Avoiding a trap: after 34 :xe6
et:'lxa5 35 bxa5 :cs the rook ending
may turn out not to be winning.
34 ...
g6
35 :d7
Since exchanging rooks leads to
the loss of a piece - 35 ...:xd7 36
et:'lxd7 et:'ld6 37 lbb6 - Black is
forced to give up the open file and
resign himself to the loss of his last
queenside pawn.
35 ...
:es
:as
36 a6
37 :b7
'iPfS
38 :xbS
'iPeS
39 :b7
'iPdS
40 :r7!
Precisely here; after 40 :e7 eS
41 :n f5 ! Black is still holding on.
fS
40 ...
41 :r6 (DJ

1. .

a
.

W&
u

W&

. . .
. . -
=
B

Karpov - Yusupov 65

In this position Black sealed the

move 4 l ...e5.
However, the next day Korchnoi
resigned without resuming. The
match, and together with it the era
of our world championship rivalry,
ended.
1 0

13 lDfd4
.txd4
14 cxd4
aS
15 .te3
a4 (D)
Instead 1 5 ... lDb4 1 6 .i.b1 a4 1 7
lDd2 a 3 1 8 'it'c 1 ! gives White a big
advantage.

Game 14
Karpov - Yusupov
USSR Ch (Moscow) 1983
Spanish, Open Variation
In this game from the 50th anniver
sary of our national championships
we see again the Open Variation,
so popular in the World Champi
onship matches at Baguio and
Merano. Not long before this
championship, at Linares, Yusupov
had already chosen this against me,
and after a difficult defence the
game ended in a draw. This time
that situation was not repeated.
1 e4
eS
lDc6
2 lDf3
a6
3 .tbS
4 .ta4
lDf6
5 0-0
lDxe4
6 d4
bS
7 .tb3
dS
8 dxe5
.te6
.tcS
9 c3
10 lDbd2
0-0
11 .tc2
.trs
12 lDb3
.tg6

w
16 lDd2
At Merano (match game 6) I re
treated the knight to c 1 , and as a
result sustained a defeat. Now I de
cided to return to a manoeuvre I
had used in an old game against
Savon (Moscow 197 1 ).
16
a3
axb2
17 lDxe4
.txe4
18 l:tb1
When I was commentating on
my encounter with Savon, I esti
mated the position after 1 8 ...dxe4
1 9 l:txb2 7 20 l:txb5 l:txa2 2 1
'it'b1 1i'a8 22 l:tc l to be favourable
for White. The text was introduced
into practice by Yusupov.
19 l:txb2
'ii'd7

66 USSR Ch (Moscow) 1983

20 .id3
In Ivanov-Yusupov, USSR Ch
(First League) 1979, 20 .ixe4 dxe4
2 1 lhb5 lllxd4 22 .l:.c5 .l:.fd8 led to
an equal position. Three years after
this game, Hiibner, in two battles
with Korchnoi, used the new text
move.
20 ...
.ixd3
An exchange of bishops hap
pens all the same. 20 ...b4 does not
work because of 2 1 .ib5 .l:.fb8 22
.l:.xb4 .l:.xa2 23 .ixc6.
21 'i'xd3 (D)
The pawn structure is more
pleasant for White.

end the game. However, returning

material is not compulsory.
22 lUb1
22 .l:.xb5 .l:.xb5 23 'i'xb5 lllxe5
24 'i'b7 'i'c6 25 'i'xc6 lllxc6 26
.l:.c 1 .l:.a6 only leads to an equal po
sition.
b4
22
23 h3
Following 23 a3 bxa3 24 .l:.xb8+
.l:.xb8 25 .l:.xb8+ lllxb8 26 'i'xa3
'i'c6, Hiibner-Korchnoi, Chicago
1982 ended in a draw after 27 'i'e7
'i'd7 28 'i'a3 although according to
Hiibner, by continuing 27 g4 h6 28
f4 llld7 29 f5 lllb6 30 .if2 White
would have retained an advantage.
Interestingly, this move was my
first independent decision. Strictly
speaking, this has nothing to do
with the pawn move, but with
White avoiding the simplification
on the queenside, which occurred
in the above-mentioned games.
23
h6 (D)
Not, of course, 23 ....l:.a3 because
of 24 'i'xa3 ! . However, a year later
in a game against Popovic (Sara
jevo 1 984), my opponent played
more accurately: 23 ....l:.b6, saving
tempi. White had no great desire to
spend time trying to realize a mi
croscopic advantage, and a couple
of moves later, after 24 'i'c2 .l:.ab8
25 .l:.c l .l:.8b7 26 'i'c5, he had to ac
cept a draw.
24 :et
.l:.b6

B
.l:.fb8
21
In the second Hiibner-Korchnoi
encounter (Lucerne OL 1 982) an
even position arose after 2 1 ...b4 22
.id2 .l:.fb8 23 .l:.fb 1 'i'g4 24 .ie3
.l:.b6 25 h3 'i'c8, although as a re
sult of 26 .l:.c2? b3! 27 .l:.xb3 lllb4
Black won the exchange, and in the
.

Karpov - Yusupov 67

- . .

with a position which has possi

bilities for both sides.

w
In the event of 24... lDa5 there is
the unpleasant 25 'lfbl , whereupon
25 . . . lDc4 loses: 26 l:txb4 l:txb4 27
'iVxb4 l:txa2?? 28 'iVb8+ h7 29
'iih l+.
llab8
25 'ilbl
After 25 ...l:ta7 too, B lack will
suffer, as in the game Popovic
Timman, Sarajevo 1984.
lDd8
26 l:tc5
27 l:tcc2
lDc6
If 27 ... lDe6, then 28 f4, with a
pawn storm; therefore Black re
turns the knight to its place.
28 'ii'c t
l:t8b7
29 l:tc5
li:J.e7
30 h2
lDf5?! (D)
White has a clear initiative on
the queenside, besides which he
can also send forward the kingside
pawns, giving his opponent new
problems. Black decides to give
back the c-pawn to stir up some
counterplay. All the same, 30...llb5
or 30 . . .c6 would have been better,

w
31 llbc2
32 l:txc7
33 l:txc7
34 g4!
34 ...'iVe2 was threatened, and if
35 llc2, then 35 ...%lxg2+ ! 36 xg2
lDxe3+. The defence 34 'iVc2 is
insufficient in view of the reply
34 ... 'ii'f l ! . However, the move by
the g-pawn saves White from all
his problems.
lDh4
34
After 34 ... lDxe3 35 'iVxe3 the
black attack has run dry, and a cru
cial pawn has gone missing.
h7
35 l:tc8+
36 'ii'd l
'ii'a6
37 l:tc2
White's aim is to carry out the
manoeuvre .i.f4-g3, securing his
king position.
f5 (D)
37

--

- . .

.
.

-
u
.
- u

--

.,.m
u
R,

68 Moscow Wch (9) 1984/5

40 xh3 "ili'e6+ 4 1 h2 'ilff5

threatening 42 ..."ili'e4 is dangerous.
Now the rook comes to the aid of
the king.
40
'ii'e6
41 "ili'hS!
Not allowing 4 l ...h5.
41
'ii'e7+
42 'itxh3
'ii'f7
43 l::thl !
Preventing 43 . . .l::tg 3+ 44 h4
lth3+.
43
'ii'd7+
1-0
44 fS
It is interesting that Yusupov and
I were jointly handed the prize for
the most beautiful game in the
championship. The founder of the
prize for beauty understood that
every work of art in chess has two
authors ! Then, a little later, this
victory won the next competition
for the best game in lnformator.
..

w
In the event of 37 ...1i'a3 38 ...e2
b3 39 axb3 "ili'xb3 the position can
be strengthened by means of 40
l::tc 7.
38 g3!
The very same knight which was
meant to bring havoc into White's
camp is cold-bloodedly eaten by
the king.
fxg4
38
gxh3 (D)
39 xh4

Garne 1 5
Karpov - Kasparov
Moscow Wch (9) 198415
QGD, Tarrasch

w
40 f4

In his Candidates matches against

Beliavsky, Korchnoi and Smyslov,
my next opponent for the world
chess crown had used the Tarrasch
Defence successfully. However in
our first duel I managed to find the
key to Black's position.
dS
1 d4

Karpov - Kasparov 69

2
3
4
5
6
7
8
9
10
11
12

c4
lbf3
cxd5
g3
.tgl
0-0
lbc3
.tg5
lbxd4
.teJ
'ili'b3 (D)

e6
cS
exd5
lbf6
.te7
0-0
ltJc6
cxd4
h6
.:r.es

B
Kasparov's opponents in the
Candidates matches had played a
variety of moves here - 12 1t'a4, 1 2
'j/fc2 and 12 a3, and Black achieved
a good game every time. I instead
used a rarer continuation, which I
had specially prepared for the encounter.
lbaS
12
.tg4
13 'ili'cl
14 lbfS
GM Lajos Portisch thought of
this move. It is better to dispatch

the knight to f5 straight away than

to play 14 h3 .th5 1 5 .:tad 1 .:tc8 1 6
g4 .tg6 1 7 lbf5 first, a s given i n
earlier opening manuals .
14 ...
.:r.cs
In our first match the Tarrasch
Defence was played twice - in this
game and earlier in game 7. It was
then that Black first chose to place
his rook on c8, which was at that
point considered a novelty.
15 .td4! (D)
A strong move. In game 7 I had
exchanged on e7 - 1 5 lbxe7+ .:txe7
1 6 .:tadl 'ii'e 8, and Black quickly
equalized. He was defeated not be
cause of his opening formation, but
due to an inaccurate move he made
in the endgame.

.tcs
15
.:txcS
16 .txcS
17 lbe3!
Compensation for the isolated
d-pawn usually takes the form of

70 Moscow Wch (9) 198415

active piece play, but in this case

the black pieces are quite passive.
Thus, the opening dispute has con
cluded in White's favour.
17
.ie6
The fork 17 ...d4 is not danger
ous because of 1 8 :ad 1 .
Wc8
18 :ad1
White was already threatening
1 9 lbexd5 ! xd5 20 e4.
19 Wa4
1 9 Wb1 :d8 20 :d3 deserved
attention, when 20 ... d4 is bad due
to 21 :Cd l 6 22 .ixc6.
:ds
19
20 :d3
a6
21 :rd1 (DJ
After 2 1 Wd l Black would have
to reply 2 1 . . .11t'c6, as 2 1 . . .lLlc4 22
exd5 lLlxd5 23 xd5 .ixd5 24
.txd5 xb2 25 .ixf7+ xf7 26
:xd8 lLlxd I 27 :xc8 :xc8 28
lbdl would lead to a rook ending
with an extra pawn for White.

Now the white rook and knight

are very unusually positioned, in a
letter T (it is as if the d3-rook is
suspended between epaulettes on
c3 and e3), and his pieces are
moreover directed at the most vul
nerable spot in Black's fortress. All
the same, the d-pawn is withstand
ing the pressure.
21 ...
lLlc4
22 lLlxc4
22 lLlexd5 xd5 23 lLlxd5 .txd5
24 .ixd5 gives nothing in view of
24 . . .:dxd5 ! 25 lbd5 :xd5 26
lbd5 b6 27 'it'd4 lLlxd5 28 'it'xd5
'it'c 1 + 29 g2 'it'xb2 with an level
queen ending.
:xc4 (D)
22 ...

w
23 WaS
The natural move 23 'ifb3 would
probably have been stronger. After
23 . . . d4, not 24 'ii'xb7 'it'xb7 25
.txb7 :b8 26 .txa6 dxc3 27 .txc4
c2 28 :d8+ lbeS ! , but the quiet 24

Karpov - Kasparov 71

'ii' b 6! wins a pawn: 24 ...ltld7 25

.:txd4 ! . Black generally has to be
very careful, so as not to lose his
isolani.
23
.:tcS
24 'ii'b6
.:td7
25 .:td4 (D)
Prophylactic measures such as
25 h3 or 25 a3 would probably
have given Black more trouble.

32 Wg2

.:t7c5 (D)

w
.:txc4
33 .:txc4
wrs
34 .:td4
35 .i.e2
.:txd4
We7 (D)
36 exd4
Insufficiently energetic play by
White has almost allowed Black to
stabilize the position; here 36 .. lbe4
37 ltla2 ltld6 38 ltlb4 a5 39 ltlc6
ltlc4 would have given him more
chances to equalize.
.

B
25 ...
'ii'c7
.:tdxc7
26 'ii'xc7
27 b3
Exchanging queens turns out to
be possible because the d5-pawn is
invulnerable: 27 ltlxd5 ltlxd5 28
.i.xd5 .i.xd5 29 .:txd5 .:txd5 30
.:txd5 .:tc2 3 1 .:td8+ Wh7 32 .:td7
.:txb2 33 .:txf7 .:txe2 drawing.
b5
27 ...
g6
28 a3
Wg7
29 e3
30 Wh2
.:tc4
b5
31 .i.f3

72 Moscow Wch (9) 1984/5

37
38
39
40
41
42

lt::!a2

li)b4
f3
h4
'itf2
lt::Jc2 (D)

.tc8
'itd6
lOgS
lt::!h6
lt::!f5

.i..

u
8
-

BlLl ii.=

B
Here the game was adjourned,
and Black sealed the move 42 ... f6.
After 42 ...lt::!g7 43 g4 f6 44 .td3 g5
45 .tg6 ! hxg4 46 h5 White has a
dangerous passed h-pawn, but
42 ... .td7 would probably have
been more reliable.
f6
42
g5
43 .td3
.txf5
44 .txf5
.tbl
45 li)e3
46 b4
gxh4 (D)
Black's bad bishop (his a6-, b5and d5-pawns are placed on the
same coloured squares) determines
White's stable advantage, but is it
decisive? After 46....tg6, piercing
a hole in the fortress would not

w
have been so easy. However, Black
has taken the h4-pawn. Exchang
ing pawns (if it happened) would
make achieving a draw simpler. In
fact, the break g3-g4 is harmless,
and White only has the f4-square
left through which he can penetrate
the enemy camp; his pieces, the
king and knight, cannot both pass
through it at the same time. How
ever, White has at his disposal a
study-like route.
47 lt::!g2! !
Psychologically i t was impossi
ble for Black to foresee such a
move. It is all too easy to assume
that White will automatically re
capture the pawn on h4. The unex
pected knight manoeuvre is based
on a pawn sacrifice. By giving up
one of his own pawns White clears
a square which is needed for his
king. The material balance is soon
restored, after which the white
pieces invade.

Karpov - Kasparov 73

hxg3+
47 ...
48 xg3
e6
49 lDf4+
s
50 lbxh5
e6
d6
51 lbf4+
52 g4
.tc2
53 hS
.tdl
54 g6
e7 (DJ
After 54... .txf3 55 xf6 Black
loses the d5-pawn after all.

. - - .
.
. - .
-

-
- -

R
.

u u

0 -8
.

. -

B B.t.B B

w
55 lbxd5+
e6
d7
56 lbc7+
57 lbxa6
.i.xf3
58 xf6
d6
dS
59 fS
.i.hl
60 f4
61 e3
c4
.i.c6
62 lbcS
If 62 ... .tg2, then 63 lbd3 b3
64 lbf4 .*.b7 65 d3 xa3 66 c3
and 67 d5 is decisive.
63 lbd3
.*.g2
c3
64 lbeS +
65 lbg6
c4 (DJ

w
.*.b7
66 lbe7
66...b3 67 d5 xa3 68 d6
i.h3 69 lDd5 will not save Black
either; the last chance of prolong
ing the game lay in 66...i.h l 67
ltks d5.
.tg2
67 lbf5
b3
68 lbd6+
69 lbxbS
a4
1-0
70 lbd6
Game 1 6
Karpov - Kasparov
Moscow Wch (27J 1984/5
QGD, Orthodox
d5
1 lbf3
lbf6
2 d4
3 c4
e6
.te7
4 lbc3
h6
S .tg5
Here White has a choice be
tween immediately taking on f6
(the Petrosian Variation) or retreat
ing the bishop to h4, in order to

74 Moscow Wch (27) 198415

carry out the exchange at a more

appropriate moment (usually after
...b7-b6). These two plans are seen
alternately in practice, and this is
how it was in my duels against
Kasparov.
6 .i.xf6
.i.xf6
7 e3
0-0
8 'ircl
cS
9 dxcS
dxc4 (D)
Kasparov and I often borrowed
from each other's opening systems,
which was inevitable, as in those
days our repertoires were in many
ways similar. In particular, this
position was seen in Kasparov
Timman, USSR v Rest of World,
London 1984, which was played
some months earlier. On that occa
sion Kasparov gained a convincing
victory after 9 ...'ira5 1 0 cxd5 exd5
1 1 0-0-0 .i.e6 1 2 lL!xdS llc8 1 3
b1 ! . Now he had to fight against
it himself.

'iraS
10 .i.xc4
.i.xc3
1 1 0-0
In the event of 1 1 ...'ii'xc5 1 2
lL!e4 'fke7 1 3 lLlxf6+ 'ii'xf6 14 llfdl
White has tangible pressure.
12 'irxc3
'irxc3
lL!d7
13 bxc3
14 c6
bxc6
lL!b6
15 llab1
16 .i.el
cS
17 llfcl! (D)

B
Moving the rook to c 1 is the first
in a long series of precise moves by
White. 17 llfdl suggests itself, but
in the long term this can only lead
to exchanges on the d-file. On c l
the rook is fulfilling a prophylactic
function (defending the c-pawn),
and is not taking the d 1 -square
away from the bishop. The situ
ation will only become completely
clear in six moves time, and I will
confess that I sensed rather than
saw how I should place my pieces.

Karpov - Kasparov 75

17 ...
.i.b7
After the game this move pro
voked unanimous criticism from
commentators. In fact, 1 7 ....i.d7,
controlling the b5-square, is more
reliable, for example 1 8 fl (there
is another possibility linked with
moving the bishop to a6 and the
knight to e5) 1 8 .. .l:tfd8 1 9 l:.b3
l:.ac8 20 l:.a3 l:.c7 2 1 c4 .i.a4 22
l:.bl .i.e8 23 l:.a5 and White had a
minimal advantage in Novikov
Sturua, Lvov 1 985.
But in the game itself White's
advantage from the start was also
scarcely noticeable. Therefore,
glancing at this almost symmetri
cal position, it is difficult to under
stand immediately why the bishop
is better placed on d7 than on b7.
1s n
.i.ds
At the cost of a tempo it would
have been possible at this point to
prevent the rook from landing on
b5 : 1 8 ...c6, but after 1 9 lDe5
.i.a4 20 .i.b5 (20 .i.a6 l:.fd8 2 1
e2 threatening 22 l2Jd3 i s also un
pleasant for him) 20 ... xb5+ 2 1
l:.xb5 l:.fc8 22 l2Jd3 c4 2 3 l2Jb2
Black has not escaped difficulties.
19 l:.bS (D)
The white rook is headed for an
ideal position - the a5-square. The
a2-pawn cannot be taken in view of
the reply 20 c4 and the bishop is
trapped - the cl -rook is already in
fluencing the course of events.

B
19
l2Jd7
In the final analysis this loses a
pawn; 1 9...l:.ac8 20 l:.a5 l:.c7 21 c4
.taB is more resilient.
20 l:.aS
l:.tb8
21 c4
.i.c6
22 lbel!
Paradoxically, before going on
the attack, the white pieces first of
all retreat.
22
l:.b4
23 .i.d1 ! (D)
..

76 Moscow Wch (27) 198415

If 23 'Lld3 immediately, then

23 ...l:.a4, and the pawn is safe.
Now, however, the march 'Llel
d3xc5 cannot be prevented. The full
extent of White's resources is now
revealed, and the hidden strength
of the quiet move 1 7 l:.fc I becomes
clear.
l:.b7
23
24 f3
l:.d8
25 'Lld3
g5
26 ..tb3
The final preparations; after the
hasty 26 'Llxc5 'Llxc5 27 l:.xc5 l:.b2
28 l:.xc6 l:.dd2 all White's gains
would have dissolved.
f8
. 26
'Llxc5
27 'Llxc5
28 l:.xc5
White still faces a large and la
borious task, but the preliminary
results of the engagement are not
difficult to sum up: a pawn is a
pawn.
28
l:.d6
e7
29 e2
l:.xd1
30 l:.d1
31 xd1
d6
f5
32 :as
h5 (D)
33 e2
Kasparov is trying with all his
might to gain some counterplay.
With his last move he has weak
ened the g5-pawn, and I immedi
ately make use of this.
fxe4
34 e4
..txe4
35 fxe4

. - -
- R.t.li .t. R R
a
-- - .ii.. B R
.
B W B '7}.,

-- - - -

w
36 l:.xgS
..trs
37 e3
The black h-pawn will in the
long run cause White distinct
trouble, and it would have been
worth fixing it with 37 h4. For ex
ample, the following variation is
possible: 37 . ..l:.f7 38 e3 ..tg4 39
c5+ c6 40 ..ta4+ c7 41 ..te8
l:.h7 42 ..tg6 l:.h6 43 ..te4 with an
easy win.
h4
37
38 d4
e5+
39 c3
..tb1
40 a3
l:.e7
41 l:.g4 (D)
White's final move was sealed.
In the process of home analysis it
became clear that the adjourned
game was exceptionally sharp, and
that Black could gain counterplay.
Therefore, in order to find a clear
cut route to victory, I had to solve
various problems straight away. It
is no accident that this ending has

Karpov - Kasparov 77

B
already found its way into mono
graphs on endgame theory.
41
h3!
In the event of passive defence
with 4 1 ...1lh7, I would have fixed
the h-pawn (but now on h4), after
which Black would have had noth
ing to hope for. If 4 l ...e4, then 42
dl .:.n 43 'iit d4 llf2 44 c5+ 'iite7
45 b3 lld2+ 46 c3 lld3+ 47
'iitb2 e3 48 a4 is decisive.
42 g3
Black was counting on my ac
cepting the pawn sacrifice - 42
gxh3, and then 42 ...1lh7 !, notice
ably widening the rook's sphere of
influence.
lle8
42
Preparing to transfer the rook
to the second rank via the f-file.
42 ...1lf7 loses immediately to 43
c5+.
.:.rs
43 llg7
44 llxa7
.:.n
llxh2 (D)
45 b4

White has a beautiful win after

45 ...1lb2. Here is the main line: 46
c5+ 'iitc6 47 c4 c2 48 lla6+ c7
(48 ... 'iitb7 49 llb6+ c7 50 'iitc 3)
49 xc2 llxc2+ 50 d5 llxh2 5 1
lla7+ b8 (alternatively, 5 1 ...'iitc8
52 llh7 llh1 53 'iit d6 h2 54 llh8+
'iitb7 55 c6+ 'iitb6 56 c7) 52 llh7
llh 1 53 'iite4 h2 54 f3 e4+
(54 .. .1la1 55 llxh2 llxa3+ 56 g4
llc3 57 lle2 llxc5 58 f5, etc.) 55
g2 llc l 56 xh2 llxc5 57 llf7
.:r.c4 58 g4 e3 59 g3 winning.

..

w
'iitc6
46 c5+
d5
47 a4+
e4
48 lld7+
48 . .. e6 49 c6 llb2+ 50 b3+
llxb3+ 5 1 'iit xb3 e4 52 lld8
xc6 53 .:r.hs g2 54 a4 f5 55
llh4 will not save Black either.
49 c6
llb2+
llb8
50 'iita S!
If 50 . . . h2, then 5 1 c7 !, and as
soon as the black queen appears on

78 Moscow Wch (4) 1985

the board, she is lost: S l ...hl'ii 52

.i.c6+.
l:.c8
51 c7
e3
52 b6
53 .i.c6!
h2
54 g4! (D)
Taking the f5-square away from
the bishop. Now it is all over.

S .i.gS
h6
.i.xf6
6 .i.xf6
0-0
7 e3
8 'ii'c2
As the queen move had led me
to victory in the game we have just
seen, there seemed no point in
avoiding this popular position a
year later. However, Kasparov had
prepared an interesting novelty.
8
tba6! (D)
.

B
54

ss l:.dl

56
57
58
59

l:.el+
l:.e4+
l:.xeS
l:.e2

l:.h8
.i.a2
f4
g3
xg4
1-0

Game 17
Karpov - Kasparov
Moscow Wch (4) 1985
QGD, Orthodox
1
2
3
4

d4
c4
tbc3
tbf3

dS
e6
.i.e7
llf6

w
A curious manoeuvre, which allows Black to make use of the
white queen's position on c2 in order to simplify - 9 cxd5 lbb4 ! 10
'ii'b3 tbxd5.
cS
9 l:.dl
'iiaS
10 dxcS
tbxcS
1 1 cxdS
12 'ii'd2
After 1 2 d6 .i.d7 followed by
.. .l:.ac8, or 1 2 dxe6 .i.xe6, Black
has sufficient compensation for the
pawn.

Karpov - Kasparov 79

lidS
12
Not, of course, 1 2...exd5 in view
of 1 3 ltlxd5, while 1 2 ....i.xc3 1 3
1Wxc3 1Wxc3+ 1 4 bxc3 exd5 1 5
l:.xd5 is also favourable for White.
13 ltld4
The only attempt to fight for the
initiative. After 1 3 .i.e2 .i.xc3 1 4
Wxc3 Wxc3+ 1 5 bxc3 l:.xd5 the
time has come to agree a draw.
exd5
13 ...
14 .i.e2
'ii'b6
After 14 ...lDe6 1 5 ltlb3 .i.xc3 1 6
bxc3 as well, White i s slightly better.
15 0-0
ltle4 (D)
Definite problems remain for
Black after 15 ... lDe6 16 ltlf3 d4 1 7
ltle4 dxe3 1 8 ltlxf6+ gxf6 1 9
Wxe3.
.

The queen retreats and occupies

a safe place; when she is needed
she can quickly return to the centre.
llc7
19 ...
20 l:.d2
l:.dc8
After 20....i.xd4 2 1 llxd4 all the
black pieces would be tied up de
fending the d5-pawn. Possibly in
this case the game would have
ended in a draw, but passive play is
not Kasparov's style. It was diffi
cult to imagine that changing the
pawn structure would lead in the
future to serious problems for
Black.
21 ltlxe6!
If 2 1 llfd l , Black should con
tinue 2 1 .. ..i.xd4 and then 22 llxd4
llc2 23 .:.4d2 .i.f5 .
21 ...
fxe6 (D)
2 1 . . .Wxe6 22 llfd l is scarcely
any better.

ltlxc3
.i.e6
llac8

w
22 .i.g4

80 Moscow Wch (4) 1985

If 22 llfd 1 , then 22 . . .'iVb4 is pos

sible, threatening 23 .. ."tlt'xd2 ! with
complete equality. For the time be
ing Black's position looks hopeful,
but while his bishop is firing into
thin air, White's bishop is capable
of creating dangerous threats - a
typical motif in opposite-coloured
bishop endings. In the end White's
light-squared strategy is trium
phant in this game. As Grandmas
ter (and pianist !) Taimanov noted,
for the next 1 7 moves I was only
playing on the 'white keys' .
l:lc4
22
'ifc6
23 h3
24 'ii'd3
After 24 'it'g6, there is the retort
24...'it'e8, and White cannot yet in
vade via the light squares.
h8
24
aS
25 l:lfdl
l:lc3
26 b3
l:lf8 (D)
27 'ife2

White would maintain a mini

mal advantage after 27 ...l:lc l 28
l:lxc l 'ii'xc 1 + 29 h2 'ii'c6 30 g3.
But Kasparov, it seems, underesti
mated White's threats.
28 .i.hS!
The bishop bounces off the sides
of the board like a billiard ball,
changing direction and ending up
on another, even more dangerous,
diagonal.
bS
28
28 ...i.d8 29 .i.g6 .i.c7 30 .i.d3
'it'd6 3 1 g3 'it'e5 32 ..g4 'ii'f6 was
necessary, erecting a fortress. The
text move loses time, and allows
White to improve his position.
29 .i.g6
.i.d8
30 i.d3
b4
31 'ii'g4
'ife8
.i.gS (D)
32 e4!
.

w
.:.Xc2
33 l:lc2
After either 33 ...'it'f7 34 l:le2 or
3 3 ...'it'c6 34 'ife2 l:lc8 35 l:lxc3

Karpov - Kasparov 81

bxc3 36 exd5 exd5 37 .i.c2 White's

initiative has not been dampened
down, but B lack should neverthe
less feel calmer.
'ii'c6
34 .i.xcl
35 'ii'el
'ii'cS
36 .:0 (D)
Defending against the threat of
36 ...l:txf2 37 'ii'xf2 .i.e3. The black
queen's activity will be short-lived,
and clouds are gathering over the
black king.

36

'ii'c3
exdS
37 exdS
'ii'dl
38 .i.bl!
39 'ii'e5
39 'ii'e6 was an alternative, but
then a white piece has to go onto a
dark square at some point! Obvi
ously there is no reason to pursue
the pawn - 39 'ii'xd2 .i.xd2 40 l:tdl
.i.g5 4 1 :Xd5 l:td8 and now the op
posite-coloured bishops work in
B lack's favour.

39
l:td8
39 . ..d4 40 'ife4 Wg8 4 1 .i.d3 is
also bad, but 39 ....i.f6 is a more
dogged defence.
Wg8 (D)
40 'ii'fS
.

w
Here the game was adjourned,
and painstaking analysis showed
that White has huge attacking re
sources.
41 'ii'e6+
h8
If 41 ...f8. then 42 .i.g6 'ili'f4 43
l:tel with the deadly threat 'ife8+.
Wg8
42 'ii'g6
43 'ii'e6+
h8
1Wc3
44 .i.fS
g8
45 'ii'g6
46 .i.e6+
h8
g8
47 .i.fS
Wf8
48 g3
'iff6
49 gl
'ili'f7
so 'ifh7
5 1 f4 was threatened, winning a
piece, and if 50...g6, then 5 1 .i.xg6
'ii'g7 52 f4 .i.f6 53 l:tdl .

12 Moeow Wcla (4) 1985

!1 h4

.td2
Other bishop retreats are no bet
ter: 5 1 . . .i.f6 52 l:.e l 'ili'g8 53 'iWg6
'iWf7 54 'ii'g4, or 5 1 ...i.e7 52 l:tel
with the threat of 53 i.e6 or 53
.tg6.
i.c3
52 l:tdl
53 l:td3
l:td6 (D)
If 53 ... 'iWg8, then 54 'ii'g 6 'ii'f7
55 'ifb6 is very strong. Now that
the white rook has entered the
game, events develop swiftly.

w
54 l:tf3!
One imprecision - 54 l:te3 would be enough to allow Black to
free himself by 54 ...g5 ! .
<j;e7
54
The king is forced to abandon
his refuge. If the f-file is recovered,
the appearance of the rook on e3
will quickly be decisive: 54 ...i.f6
55 l:te3 g5 56 'iWxh6+, and the
rook's action is covered; 54 .. .l:tf6
55 l:te3 g5 (55 ...l:txf5 56 'ii'h 8+

'ii'g 8 57 l:te8+ <j;xe8 58 'ii'xg8+

<j;d7 59 f4) 56 'ii'h 8+ covering the
action of the bishop. Problem mo
tifs in abundance !
55 'ii'h 8!
55 l:te3+ would have been hasty
in view of 55 ...<j;d8 56 'ifh8+ <j;c7
57 'ii'c8+ <j;b6 58 'ii'b8+ <j;c5 !, and
Black can defend himself with the
help of ...l:tc6 and ...'ii'c7.
55
d4 (D)
...

w
If 55 ...i.e5, then 56 i.h3 l:tf6 57
l:te3 ! l:txf2+ 58 <j;gl is the end.
l:tf6
56 'ifc8
57 'ifc5+
<j;eS
58 l:tf4
'iib7+
rj}f7
59 l:te4+
It looks as though Black would
escape without a scratch after
59 ...l:te6, because of 60 i.xe6
'ii'xe4+, but again a problem ma
noeuvre is decisive: 60 'ifc4! l:txe4
6 1 'ii'g 8+ <j;e7 62 'ifxg7+ and 63
1Wxb7.

Kasparov - Karpov 83

60 'ii'c4+
8
lU7
61 h7!
'ii'd 7
62 'ii'e6
63 'ii'e 5!
1-0
I had prepared 63 . . .:e7 (if
63 . . .'iWd8, then 64 'ii'c 5+ :e7 65
:f4+ e8 66 'iWc6+ 'ii'd7 67 ..tg6+;
or 63 ...'ife7 64 'iWb8+) 64 'ii'f4+
:n 65 'iWb8+.

12 a4
'ifd7 (D)
This queen move is relatively
rare; the usual continuation here is
1 2 ... h6.

Game 1 8
Kasparov - Karpov
Moscow Wch (5) 1985
Spanish, Zaitsev
1 e4
e5
2 lLlf3
lLlc6
3 b5
a6
4 a4
lLlf6
5 0-0
e7
b5
6 :et
7 b3
d6
8 c3
0-0
9 h3
b7
10 d4
:e8
This rook move was brought
into practice by my long-time sec
ond Igor Zaitsev, and the variation
is therefore named after him. Black
fortifies the centre, and does not
spend time on the prophylactic
... h7 -h6. True, there is a danger that
White will repeat moves with 1 1
lLlg5 :f8 1 2 lLlf3 :e8 1 3 lLlg5.
Thus, if he needs a victory, Black
should choose something else.
i.f8
11 lLlbd2

w
axb5
13 axb5
i.xa8
14 :Xa8
After 14 ... :xa8 1 5 lLlg5 the
black knight is forced to retreat to
the back rank allowing White to
seize the centre: 1 5 ... lLld8 16 lLldf3
exd4 ( 1 6 . . . h6 1 7 lLlxf7 ! lLlxt7 1 8
dxe5; 1 6 ...c 5 17 dxe5 dxe5 1 8
'it'xd7 lLlxd7 1 9 lLlxf7 ! c4 20 lLlxd8
:xd8 2 1 i.a2 :e8 22 e3 and
White had a clear advantage in
lvanchuk-Portisch, Linares 1990)
17 e5 !.
15 d5
Now, as he has taken on aS with
the bishop, after 1 5 lLlg5 Black can
reply 15 ... :e7, and then ...h7-h6.
15 ...
lLla5
In game 46 of our first match I
retreated with 15 . . .lLld8, and after

B4

Mo1cow Wch (5) 1985

1 6 fl h6 1 7 ltl3h2 ! ltlb7 1 8 i.c2

ltlc5 1 9 b4 ltla6 20 ltlg4 ltlh7 2 1
ltlg3 c 6 2 2 dxc6 i.xc6 2 3 i.b3
ltlc7 24 W'f3 ltle6 25 h4 W'd8 26
l:.dl White achieved an advantage.
16 i.a2
c6
17 b4
ltlb7 (D)

.i.

.I

%%

'ii'

--

8
n

-
u

%% -
/";"\. . 8

"LJ

8.
i.

u

7,

!ii

blocked in by its own b7-knight

and c6-pawn, will quickly free it
self and develop energetic activity,
and precisely thanks to his bishop
Black will gain the upper hand.
18 c4
The standard 1 8 ltlfl is better,
and after 1 8 ... cxd5 ( 1 8 ... c5 19 i.g5
i.e7 20 ltlg3 g6 21 'ii'd2, preparing
ltlh2 and f4) 1 9 exd5 h6 20 ltlg3
l:.c8 21 W'd3 ltld8 ! 22 ltlh2 g6 23
h4 White maintains the initiative.
18
l:.c8! (D)
Before retreating the knight to
d8, it makes sense to place the rook
on a more promising file.

. -
- -

w
After 17 ...ltlc4 1 8 ltlxc4 bxc4
1 9 i.g5 ! cxd5 20 i.xf6 dxe4 2 1
ltlxe5 W'a7 22 ltlg4 'ii'xa2 23 ltlh6+
h8 24 ifh5 gxf6 25 1li'xf7 i.xh6
26 '1Vxe8+ g7 27 'ii'e7+ White
has the advantage, since the enemy
bishop has no prospects. However,
now it seems that Black is all but
suffocating from a lack of vital
space. But this impression is de
ceptive. Nimzowitsch once said
that a piece which has been in cap
tivity for a long time finds incred
ible strength when it has broken
free. In fact, in this game my aS
bishop, hiding in the corner of the
board, and for the time being

19 dxc6
Handing over the centre allows
Black immediately to gain a play
able game. Fascinating events arise
after 19 W'e2 ltld8 ! 20 i.b2 bxc4
2 1 ltlxc4 W'a7 ! (2 l . ..l:.b8 ! also
gives him good prospects) 22 l:.a1
cxd5 23 exd5 i.xd5 24 ltlxd6

Kasparov - Karpov 85

xd6 25 xd5 1fxa1 + 26 xa1

l:tcl + 27 h2 l:txa1 and Black has
enough compensation for his small
material loss.
'1Vxc6
19
20 c5?
Instead of this highly risky ven
ture, 20 b2 straight away would
have been better. As it is, the threat
to f7 (20 ... dxc5 2 1 xf7+) is eas
ily repulsed, and Black seizes the
initiative.
20
lbd8
21 b2
dxcS!
22 bxc5
Black also has good chances
after 22 lbxe5 1fa6 ! 23 ..,al c4 24
c3 1fa3 25 l:te3 .i.xb4 26 lbexc4
bxc4 27 xf6 c3 ! 28 l:tg3 lbe6, or
alternatively 22 .i.xe5 lbd7 23
.i.b2 c4.
22
xc5
23 .i.xeS
lbd7
4! (D)
24 .i.b2
.

More precise than 24... ..,c2 25

1fa1 ...a4 26 lbd4 lbc5 27 l:te3
with a sharp game.
25 lbb3?
After 25 1fb1 White's position
would still have been defensible.
The desire to generate some activ
ity has led him into a difficult position.
lbcS!
25 .
26 .i.a1
A forced pawn sacrifice. After
26 lbxc5 it'xb2 ! 27 l:te2 it'a3 28
lbd3 .i.xe4 29 .i.xf7+ lbxf7 30
l:txe4 b4 Black has a clear advan
tage.
.i.xe4
26 ...
27 lbfd4 (D)
Black would also have won ef
fectively after 27 lbg5 .i.c2! 28
'ii'xc2 'ii'xe 1 + 29 h2 .i.d6+ 30 g3
..

lbe4.

86 Moscow Wch (22) 1985

29
30
31
32
33
34
35
36
37
38
39
40
41

lL:!xcS
'ii'g4
.:.dt
'ii'f4
'ii'c l
.:.et
..tb3
'ii'b2

'ii'xcS
.:.e8
..tg6
'ii'b 4!
..te4
'ii'a5
'ii'a8
b4
..tg6
'ii'xe8
lbe4
lDc5
lDd3

Game 1 9
Karpov - Kasparov
Moscow Wch (22) 1985
QGD, Exchange

In game 2 1 , when Kasparov had

White, after 5 ... c6 6 e3 ..tf5 7 g4
.te6 8 h4 lDd7 9 h5, I introduced a
novelty, 9 ... lDh6, but did not man
age to solve my opening difficulties.
6 e3
In game 20 I stopped the bishop
coming out to f5 with 6 'i!fc2 but
the time I lost in doing this led to
rapid equality: 6 ...0-0 7 e3 c5 ! 8
dxc5 ..txc5 9 lDf3 lDc6 10 .te2
d4! .
6
0-0
..trs
1 lDf3
8 h3
Preparing to carry out the stand
ard manoeuvre in this position,
g2-g4. This rook's pawn move had
until that time not been used with
this move order.
8
c6
9 g4
.tg6
10 lDe5
lDfd7 (D)

For the third time in a row in the

match a sharp form of the Queen's
Gambit Exchange Variation arose.
In the first two games White had
the initiative, but they ended in
draws. This time, however, I man
aged to win.
d5
1 d4
e6
2 c4
3 lDc3
..te7
4 cxd5
exd5
5 ..tr4
lDr6

.:.eJ

.:.xe8
'ii'c l
..td5
lDb3
0-1
The final move was sealed. In
the adjourned position Black' s
passed pawn guarantees him vic
tory and Kasparov considered that
there was no point in resuming.

Karpov - Kasparov 87

After 10 ... lL'lbd7 1 1 h4 Black

feels rather hemmed in.
1 1 tt:'Jxg6
fxg6
Following the banal 1 1 .. .hxg6
Black is deprived of all possible
counterplay.
tt:'Jb6
12 ..tg2
h8
13 0-0
1 3 . . .tt:'Ja6 is not sufficient for
equality because of 14 'ikb3 ! ; nor is
1 3 . . . g5 satisfactory, due to the re
ply 14 ..tg3 i.d6 1 5 ..txd6 1i'xd6
1 6 e4 ! .
1 4 tt:'Je2 (D)
Both 1 4 'ikb3 and 1 4 ..tg3 fol
lowed by e3-e4 would have been
promising for White.

( 1 8 ...gxf4 19 tt:'Jxf4) 19 fxg5 ..txg3

20 tt:'Jxg3 'iixg5 21 e4.
18
'iixd6
19 f4
White's kingside pawn advance
guarantees him a definite initiative.
gxf4
19
.:Z.ae8
20 exf4
21 f5 (D)
2 1 .:Z.f3 tt:'Jd7 22 g5 followed by
h4 and lL'lg3 is probably more accu
rate.

B
14
g5
..td6
15 ..tg3
tt:'Ja6
16 'iid3
'iie7
17 b3
18 ..txd6
The immediate 1 8 f4 would have
led to a sharp struggle: 1 8 ....:Z.ae8

21
22
23
24
25
26
27
28
29
30
31

.:Z.f2
g5
h4
.:Z.d1
'iixe3
h2!
tt:'Jg3
tt:'Jn
l:td3
tt:'Jg3

tt:'Jc7
tt:'Jd7
'ike7
'iie3
tt:'Jb5
.:Z.xe3
tt:'Jb6
tt:'Jc8
.:Z.e7
tt:'Jcd6

BB Moscow tt 19B6

Here I was preparing to move

my king to g4 and force back the
knight with a2-a4, maintaining the
pressure. Obviously, Black should
also bring his king up to the cen
tre with 3 l ...g8, or activate the
rook by means of 3 1 ...llel . Instead
of this Kasparov, in severe time
trouble, made a serious mistake.
31
t:be4? (D)
...

. . . .,. .
- . .
8 -
.

8:
8 ail.=

w
dxe4
32 xe4!
33 lle3
t:bxd4
34 h3
By 34 f6 ! gxf6 35 gxf6 White
achieves a strong position, for ex
ample 35 ...lle6 36 llxe4. But be
fore moving into decisive action, I
decided to activate the king.
34

..

lieS

35 g4
h5+?
Kasparov gives an impulsive
time-trouble check; 35 . ..g8 or
35 . . .llfe8 was necessary, defend
ing obstinately.

36 xh5
Black was possibly counting on
36 f4 llexf5+ 37 t:bxf5 lbxf5
with a probable draw (but not 3 8
xe4 due to 3 8 ...liJd6+ 39 e5
llxf2).
t:bxf5
36
37 :xf5
llfxf5
38 t:bxf5
llxf5
h7
39 :Xe4
40 lle7
b5
In the event of 40...llb5, White
can achieve his goal with 4 1 a4
llxb3 42 g6+.
b4
41 llxa7
1-0
42 g4
The adjourned rook ending is
easily winning for White, as the
mobility of the h-pawn is decisive.

Game 20
Karpov - BeUavsky
Moscow tt 1 9B6
QGD, Orthodox
1
2
3
4
5
6
7
8
9
10
11
12

d4
c4
m
t:bc3
i.g5
i.xf6
e3
:et
i.d3
0-0
i.xc4
h3! (D)

t:bf6
e6
d5
e7
h6
i.xf6
0-0
c6
t:bd7
dxc4
e5

Karpov - Beliavsky 89

Both 12 .i.b3 and 1 2 li)e4 used

to be played, but in both cases con
vincing routes to equality have
been found for Black. This modest,
but poisonous peripheral pawn
move was first used by Kasparov in
the 23rd game of our second
match. The idea behind it is that the
light-squared bishop for the time
being stays where it is, and so after
1 2 ... l:leS he still has the possibility
of 1 3 'iib 3. Besides this, the g4square is taken under control, and
after opening the centre the cS
bishop will have problems devel
oping. Thus Black finds himself in
a peculiar zugzwang.

1 S l:le1 l:ldS 19 'iif4 li)d5 20 li)xd5

cxd5 2 1 li)e5 .i.xe5 22 l:lxe5 .i.e6
23 'it'e3 White gained an advan
tage.

12
13 exd4
14 .i.b3

exd4
li)b6
.i.f5 (D)

In the above-mentioned game

against Kasparov I continued with
14 ... l:leS and brought the bishop
out on the next move. After 1 5 l:le1
.i.f5 16 l:lxeS+ 'fixeS 17 'iid2 'iid7

15 l:le1

aS

In the event of 15 ... l:leS 1 6

l:lxeS+ 'fixeS 17 'iid2 the game fol
lows the variation examined above.

16 a3

'iid7

In game 22 of the return match

with Kasparov I replied 1 6 . . .l:le8,
and after 17 l:lxeS+ ..xe8 1 8 'iid2 I
tried a new idea - 1 S ... li)d7. How
ever after 1 9 1i'f4 .i.g6 20 h4 'ifdS
2 1 li)a4 I ran into difficulties. An
important novelty was used in
Gurevich-Van der Sterren, Balcu
1 9S6: 1S ...'ifd7 19 l:le 1 l:leS ! 20
l:lxeS+ 'fixeS 2 1 ..f4 .i.e6 ! , and
Black equalized.

17 li)eS
18 l:lxe5
19 1i'e2

.i.xe5
l:lfe8
l:lad8

Black has decided that there is

no hurry to push his rook's pawn.

90

Moscow tt 1 986
AxeS
a4 (D)

w
It looks as though Black's delay
in pushing his a-pawn makes no
difference, and White must remove
the bishop to a2, when the variation
22 ll:!xa4 ll:!xa4 23 .i.xa4 'ibd4 24
'it'xf5 'it'xa4 cannot worry him.
However, White finds an attractive
intermediary manoeuvre.
axb3
22 'it'cS

23 :e7

'it'd6

Yet another zwischenzug. Before

he gives back the piece, the white
rook penetrates the seventh rank.
Here 23 . . .'it'xd4 loses because of
24 Ae8+! h7 25 'it'xf5+, while if
23 ...'it'c8, then 24 'it'xb6 with an
obvious advantage for White.
lhe8
24 Ae8+

25
26
27
28

'it'xd6
'it'b4
dS!
ll:!xd5

ll:!c4
b5
cxd5
.i.c2 (D)

When I played 27 d5, I envis

aged 29 ll:!e7+ here, but while I
was checking the variation I turned
my attention to the possibility of
29 .. .llxe7 ! 30 'it'xe7 lbxb2 3 1 'it'e8+
h7 32 'it'xb5 ll:ldl 33 'it'c4 f5 with
a study-like draw: the knight leaps
between b2 and d l , and Black
saves himself. In the event of 34 a4
lDb2 35 'it'c6 ll:!dl the queen has to
return to c4, as after 36 aS .i..e 4!
and then 37 . . . b2 Black actually
takes the upper hand.
:ds
29 'it'c3
30 lDf4
Adl+
31 b2
ll:!d2
After 3 1 ...Abl 32 a4 ll:!xb2 33
axb5 the pawn cannot be stopped.

32 b4
Preventing perpetual check by
the knight.

32
33 f3
34 bxg5
35 ll:!e2 (D)

.i.f5
g5
bxg5

Karpov - Kasparov
Avoiding the final trap - 35 lbb5
t'Dfl+ 36 Wgl t'Dg3+ and White's
knight is lost.

8 cxd5
9 'ilt'd2
10 bxc3 (D)

91

t'Dxc3
'ilt'xal

Game 2 1

Karpov - Kasparov

London/Leningrad Wch (5) 1986

Griinfeld Defence

This variation came into prac

tice after the famous game Petro
sian-Fischer, Buenos Aires Ct (2)
1 97 1 , which continued 10 ...'1i'a5
1 1 .i.c4 t'Dd7 12 t'De2 5 1 3 .i.a2
.i.f5 14 .i.xe5 ! .i.xe5 1 5 .!Dd4
tWxc5 1 6 .!Dxf5 gxf5 17 0-0 with a
dangerous attack for White. Later
various lines were proposed for
Black, the most effective being
1 2 ... .!Dxc5 1 3 0-0 0-0 14 f3 e5 ! 1 5
.i.g3 b 5 1 6 .i.a2 'li'b6! 17 Wh 1 a5
with sufficient counterplay.

10
1
2
3
4
5
6
7

d4
c4

lDc3
.i.f4

e3
dxc5
.:.et

t'Df6
g6
d5
.i.g7
c5
'it'aS
t'De4

...

'ilt'xd2+

Kasparov prefers to exchange

queens. Until now this move had
resided somewhat in oblivion, and
only recently have new ideas been
found for Black.

1 1 Wxd2
12 .i.b5

lbd7

t2 LDndol'lllAnlngrad Wch (5) 1986

After 1 2 c6 bxc6 1 3 dxc6 lZ1b6
(or 1 3 . . .le!f6) White has achieved
nothing.
0-0
12
13 -txd7
If 1 3 c6, then 1 3 ...lZ1c5. Now
Black gains the advantage of the
bishop pair, but as will soon be
come clear, both will be crippled.
-txd7 (D)
13

has completed his development,

maintaining powerful pressure.

-
- i.
-i
-i

"
u

mi

- A -

L.l

. .:
.
w

f5
14 e4
e6! ?
15 e5
A n important moment. Why
did Kasparov reject the previously
tested 1 5 ...l:l.ac8 1 6 c6 bxc6 17 d6
exd6 1 8 exd6 l:l.f6 with an advan
tage to Black (Schmidt-Gross, Nal
eczow 1 984)? The answer is that
1 5 . . . l:l.ac8 would be met by 16 c4!
l:l.xcS 17 -te3, as in Seirawan
Adorjan, New York 1 987.
16 c4
l:l.fc8 (D)
After 1 6 ... gS 1 7 -txgS -txe5 1 8
lbf3 -tg7 1 9 l:l.b 1 and l:l.he 1 White

w
17 c6!
He does not keep the extra
pawn, but in returning it, White
will elicit maximum profit by cre
ating a passed pawn and limiting
the movement of Black's rook and
light-squared bishop.
17
bxc6
18 d6
c5
Now the light-squared bishop
finds some freedom, but the dark
squared one begins to suffocate.
Perhaps the fate of the bishops
should have been dealt with differ
ently - 1 8 .. . g5 1 9 -txg5 -txeS 20
cS l:l.cb8, and Black maintains
hopes of counterplay.
h6
19 h4!
20 le!h3! (D)
After the more natural move 20
lZ1f3, the reply 20 . . .-tc6 ! would
have led to a complex game with

Karpov - Kasparov

chances for both sides. However, I

managed to find a most logical
solution to the position. The white
knight is concentrating on the only
true trajectory towards its optimal
square, d3 ! As soon as it has
reached its goal, the g7-bishop will
be caught in a trap constructed by
the white e5- and d6-pawns and the
f4-bishop. Incidentally, White pre
vents the freeing advance ... g6-g5.

20

93

23
24
25
26
27

lbf2
lbd3
llal

hxg5
.i.xg5 (D)

a2
lla3
g5
hxg5

aS

The passed a-pawn is not dan

gerous, so Black should have im
mediately occupied another file
with 20...llcb8.

21 f3
a4
22 llhel ! (D)
Overprotecting the e5-pawn.
After the hasty 22 t:Df2 Black can
muddy the waters with 22 ... g5 ! 23
hxg5 hxg5 24 .i.h2 f4 25 lbd3 .i.e8
26 .i.g1 .i.g6 27 lbxc5 .i.xe5 28
lbxe6 .i.xd6 29 lbxg5.

22

a3

w
27

rM7

After 27 . . . llb8 the duel might

conclude with a study-like draw:
28 .i.f4 llbb3 29 lbxc5 llb2+ 30
<itc1 llxg2 3 1 .i.d2 .i.h6 ! 32 .i.xh6
llc3+ 33 d 1 lld3+! 34 lbxd3
.i.a4+ 35 c 1 llc2+ and Black

 I.IMtllllrtlrtii'IMI Wch (17) 1986

diUwn perpetual check. But a
ample refutation to this 'study' can

bt found: 28 e2! l:tbb3 29 lLlxc5

fl and White has the
upper hand.

lb2+ 30
28
29
30
31
32

.i.f4
l:tecl
l:tc3
l:tcl
lLlcl

l:tb8
.i.c6
.:taS

l:tba8
1-0

Game 22

Karpov - Kasparov
London/Leningrad Wch (1 7) 1986

The critical position in the

Smyslov Variation. Almost every
thing has been tried here, includ
ing 12 'it'xd6, 1 2 lLlb5, 1 2 d5, 1 2 h3
and 12 0-0. In the key game Bot
vinnik-Fischer, Vama OL 1962, af
ter 12 h3 .i.xf3 1 3 gxf3 .:tfd8 14 d5
lLle5 1 5 lLlb5 'it'f6 1 6 f4 lLled7 17
e5, Fischer gave Botvinnik an un
pleasant surprise with 17 ...'ifxf4 ! .
However, after mistakes b y both
sides this fascinating game eventu
ally ended in a draw. However, it
turns out that White has at his dis
posal one other interesting move,
which has not been tested in prac
tice until now.

Grilnfeld Defence
This miniature is a classic example
of a battle which was decided en
tirely in the process of home prepa
ration. White's novelty cannot be
called unusual - it consists of a
simple one-square advance by the
h-pawn, but careful analysis has
shown that after this move Black's
defence becomes noticeably more
difficult.
1 d4
lLlf6

2
3
4
5
6
7
8
9
10
11

c4
lLlc3

lt:lf3
'iWb3
'iWxc4
e4
.i.e3
.:tdl
.i.el
'iVcS

g6
dS
.i.g7
dxc4
0-0
.i.g4
lLlfd7
lLlc6
lLlb6
'iWd6 (D)

12 eS!
At first glance this looks para
doxical. White not only hopelessly
weakens his e5-pawn, but also al
lows an exchange of queens. How
ever, things are not so simple as
that.

12 ...

'iWxcS

Karpov - Kasparov 95
13 dxcS
c8 (D)
In reply to the more natural
1 3 ... d7 White has 14 h3 ! .txf3
1 5 gxf3 ! and the 'doomed' e5pawn is invulnerable, as if it is cap
tured by either knight, 1 6 f4 ! wins
a piece. Otherwise, White will
strengthen the e5-pawn with its
neighbour, and gain a clear advan
tage.

resurrecting his idea. Meanwhile,

when I again set up the position on
the board, I managed to spot that
the c8-knight's prospects can be
significantly limited.
14

.t:xf3

If 14....te6, then 1 5 g5 looks

unpleasant: 1 5 ...xe5 1 6 xe6
fxe6 1 7 f4.
15 .txf3
.txeS (D)
After the alternative capture
Black also runs into difficulties:
15 ...xe5 16 .txb7 :bs 17 c6
4 1 8 :d7.

14 h3!

In game 1 5 , where the novelty

1 2 e5 was used for the first time, I
played 14 b5, and after 14 ... :b8 !
1 5 xc7 e6 ! Black managed to
equalize completely. In fact, the
threat to surround the knight by
means of 1 6...a6 forces White to
lose a tempo by playing 1 6 b5,
and the black c8-knight quickly
transfers to a comfortable position.
The result of that encounter obvi
ously satisfied my opponent, who
thought that White had no way of

16 .txc6!
17 .td4

bxc6
.tf4 (D)

After an exchange of bishops

with 17 ....txd4 1 8 :xd4 :bs 1 9
b 3 a5 the white rook can penetrate
the seventh rank, but after 20 :d7
(20 :a4 is stronger: 20 ...:as 2 1
e2 :ds 2 2 :d 1 :xd 1 23 xd 1
followed by the transfer of the
knight to c4) 20... a7 21 :xc7

96 London/Leningrad Wch (17) 1986

lllb 22 ltlxb5 .:.xb5 23 .:.xc6 a4!

Black has counterplay.

w
aS?!
18 0-0
It was necessary to move the e
pawn. My preparation showed that
after 18 ...e5 19 .i.e3 (19 lbe2 lbe7!)
19 . . ..i.xe3 ( 1 9 ... lbe7 is also possi
ble) 20 fxe3 White has a definite
advantage. However, soon after the
match this position was seen in
Karpov-Timman, Tilburg 1 986,
and the Dutch grandmaster proved
that Black has chances to hold the
position. The game continued thus:
20 ... lbe7 21 .:.d7 lbf5 22 .:.xc7
:res ! 23 .:.d7 :adS 24 .:.fd 1 .:.xd7
25 .:.xd7 lbxe3 26 .:.c7 .:.bs 27 b3
:dB with a swift draw. True, the
feeling remains that somewhere
White could have played some
thing stronger.
19 :rei
White plans to dominate the
centre. The move g2-g3 will chase

away the bishop, whereupon .i.e5

and .:.d7 will make inroads into
Black's position. In view of the
atrocious position of the knight on
c8, Black has great difficulty op
posing this plan.
19 ...
a4?!
The pawn continues its march,
but this is fiddling while Rome
bums. 1 9 ...e6 would not help in
view of 20 lbe4 when White's
pieces co-operate superbly. It is
hard to see what Black should actu
ally play here - maybe 1 9 .. .f5 to
deny White's pieces access to e4,
but to commit oneself to such an
extreme measure would be diffi
cult at the board.
What is not in doubt, is that the
pawn advance chosen by Kasparov
carries no real threat, and consti
tutes a fatal waste of time.
20 .:.e4!
Originally I had intended 20 a3,
to fix the pawn on a4 so as to gang
up on it. However, it occurred to
me that I could ignore the pawn;
especially since if (following 20
a3) the white knight captures on
a4, his black counterpart could
spring out to b5 via a7, with some
chances of real resistance. Clearly
there would be no point complicat
ing matters in this way, when sim
ply centralising my pieces is so
strong.
.i.h6
20
...

Karpov - Kasparov 97
21 i.e5
Black's planned . . .liJa7 is ren
dered rather ineffectual by this
move.
21
a3
22 b3
Clearly the best move.
!lJa7
22 ...
Although this leads nowhere,
there were already no sensible
moves for Black.
23 l:td7! (D)
23 i.xc7 would regain the pawn
and hit e7, but then 23 ...e6 or the
more precise 23 ... i.g7 would pro
vide counterchances. Thus I pre
ferred instead to keep full control
of the long diagonal.

Black seems to want to justify

his a-pawn's moves by using the
b2-square. However, White can
evict the bishop.
24 l:txc7
The e7-pawn is the next item on
the menu.
i.b2
24 ...
Black would do no better with
24 . . .!iJb5 25 l:txc6 or 24 . . .e6 25
l:ta4. Now White must be precise.
25 !lJa4!
This move guarantees decisive
material gains, and is the clearest
way to win. Instead the careless 25
l:txe7 would allow Black counterplay with 25 ...i.xc3 26 i.xc3 !iJb5
and then . . .l:td8, eyeing the a2pawn. Although White is still bet
ter, Black is in the fight again.
!iJbS
25 ...
After 25 ...i.xe5 26 l:txe5 e6
White wins an exchange at least.
26 l:txc6 (D)

98 London/Leningrad Wch (19) 1986

White simultaneously creates a
passed pawn and knocks the sup
port away from the b5-knight. 26
lbe7 would be less logical, since a
knight firmly planted on b5 would
guarantee counterplay for Black.
l:lfd8
26
Black's moves are forced.
27 l:lb6!
This is the main point of White's
plan. Black's knight has had an un
fortunate career in this game, and it
is now hounded out of the one good
square it has found.
l:ld5
27
28 .i.g3
I avoided a last, though transpar
ent, snare here: 28 lL!xb2? :Xe5 29
l:lxe5 axb2 would even win for
Black, e.g. 30 l:le 1 l:lxa2 followed
by 3 1 ...l:la l .
Now the advance of the c-pawn
decides the game quickly.
lL!c3
28
.i.xc3
29 lL!xc3
30 c6
.i.d4
31 l:lb7
1-0
.

to the system brought into practice

by Ragozin, which Najdorf then
worked on in detail. But here as
well an opening surprise awaited
Black.
1 d4
lL!f6
g6
2 c4
3 lL!c3
d5
4 lL!f3
i.g7
dxc4
5 ..,3
6 Wxc4
0-0
lL!a6
7 e4
In the two previous odd-num
bered games in this match Black
preferred 7 ...i.g4.
8 J.e2
cS
9 d5
e6
10 0-0 (D)
10 i.g5 exd5 1 l lL!xd5 .i.e6 1 2
0-0-0 .i.xd5 1 3 l:lxd5 'it'b6 i s not
dangerous for Black.

Game 23
Karpov - Kasparov
London/Leningrad Wch (19) 1986
Griinfeld Defence
Despite his disappointment in the
Smyslov Variation (game 17), Kas
parov did not abandon the Griin
feld Defence. This time he turned

B
10
11 exdS
12 i.f4 (D)

exd5
.i.5

Karpov - Kasparov

In the Seville match I twice

played 1 2 l:td1 l:te8 1 3 d6 h6 and
then 1 4 h3 (game 1 5), or 1 4 .i.f4
(game 2 1 ). Both encounters ended
in a draw.

99

I. B . I. B*B

B B B i B
BlD. 8 B.t.B
BBfl B
B B BlDB
.t
6
8
8

.:

B
:es
12
1 2 . . ....b6 and then 1 3 h3 (or 1 3
.i.e5) has independent interest.
Here a few practical examples have
accumulated, but it is too early to
make a definitive judgement.
lle4
13 l:tad1
Here 1 3 . . .,..b6 has also been
tried, and then 14 ,..b5 (or 14 llh4)
with chances for both sides.
14 ltJbS! (D)
The leap by the knight to b5 was
prepared at home. Until this point
only 14 .i.e3 and 14 .i.d3 had been
tested, and White had achieved
nothing.
The basic strategic conflict in
the middlegame is centred around
the d-pawn. If White manages to

use its potential energy, he will

have the initiative, while if Black
manages to set up a reliable block
ade, his position will become more
promising. Having led the knight
to b5, I was ready to part with the
b-pawn in order to develop pres
sure in the centre after 14 . . . .i.xb2
15 d6 .i.f6 16 .td3 ! and 17 l:tfe2 .
Kasparov refused the offer.
...f6!
14
15 .td3
lJb4
After 15 ....tg4 16 .te5 l:txe5 17
llxe5 .txd1 1 8 .txe4 irxe5 19
l:txd 1 White has gained the initia
tive. Not wishing to condemn him
self to passive defence, Black risks
everything, but the exchange sacri
fice is not justified. The capture
1 5 ......xb2 is definitely risky, al
though later practice in the vari
ation showed that Black can gain
counterplay by continuing with
either 15 ...l:tad8 or 15 ....td7.
llxd3
16 llc7!

lOO LondonllAralragrad Wch ( 19) 1986

l:.xe8
17 lllxe8
18 'ii'x d3! (D)

18 ...
'ii'xb2
Black's misfortune is that nei
ther 1 8...lllxf2 nor 1 8 ...lllg3 works
because of 1 9 'ii'b 5, attacking the
undefended rook on e8.
19 l:.de1
The straightforward 19 d6 (inci
dentally threatening to play 20 g4)
19 ...lld8 ! 20 We3 h5 2 1 llb 1 'ilfxa2
22 l:.xb7 'it'd5 23 l:.xa7 lbxd6 24
.i.xd6 'ilfxd6 25 lbg5 is unpleasant
for Black, and it is therefore prob
ably worth playing. However, I de
cided to pin the knight, limiting the
enemy pieces' activity. Of course
19 l:.fe1 does not work in view of
1 9 . . .'ilfxf2+ 20 h1 'ilfxel + ! 21
l:.xel lbf2+ 22 g1 l:.xe1+ and
23 ... lbxd3.
'i!fb4? (D)
19
This is the decisive mistake. Af
ter 1 9 . . . lbf6 20 l:.xe8+ lbxe8 2 1
..

'ii'e 3 realizing the material advan

tage is not difficult, but in the event
of 1 9 .. .'ii'xa2 20 Wb5 l:.d8 2 1
'ii'xb7 'ii'xd5 Black holds on: 22
'ilfxd5 .:Xd5 23 g4 lbf6!.

w
20 lbd2!
Playing for the pin - the main
motif of White's idea.
'ii'a4
20
21 'ilfc4
Forcing an exchange of queens;
21 lbxe4 is not so clear: 2 1 . . .l:.xe4
22 l:.xe4 .i.xe4 23 'i!fd2 c4 with
counterplay.
21
'ilfxc4
.i.c3
22 lbxc4
22 ...b5 23 lbd2 lbf6 24 l:.xe8+
lLlxe8 25 l:e1 , etc., loses immedi
ately.
23 liJd2
.i.xd2
.i.d7 (D)
24 .i.xd2
During the game many people
thought that Black had got himself
out of his mess: the d2-bishop is
.

Karpov - Kasparov 101

under attack, and . . . .i.b5 is threat
ened.

- - . -

--*- - -
- - -- 8- -

- -- "" A
A
.
-%m;; -"'B
LJ
LJ

B
%

0+

, y

'""'

!.;.;

% -""'

.:
.
. .
.

%-

, !/;::"

37 e2
38 ..td3
39 c4

w
25 .i.f4!
26 f3!

.i.b5
g5

After 26 . . . .i.xfl 27 'itxfl lLlf6

28 l:lxe8+ lLlxe8 29 .i.e5 ! the d
pawn finally shows what it is capa
ble of: 29 . . . f6 30 d6 ! and he has to
give back the piece.

27 .i.xg5

.i.xn

27 . . . lLlxg5 is no better after 28

.:xe8+ .i.xe8 29 h4 ! , and White
wins the knight.

28 xn
29 .i.e7

lLJd6
lLlc8 (D)

29 . . . lLlc4 30 d6 lLlb6 3 1 .:bl

lLld7 32 .:xb7 is also bad.

30
31
32
33
34
35
36

.i.xc5
.:eS

.:r5
.i.d4
.i.xf6
.:g5+
.i.xg5

.:d8
f6
b6
lLle7
.:xd5
.:xg5
lLlc6

f7
e6
lLle5+ (D)

- - - . .
.

.
.

.
.

-<&it- - - .8.

"
8.
-
-
-8u
- - - .
w
lLlc6+

40 d4
41 c4

Here the game was adjourned,

and Black resigned without resum
ing. After 4 1 . ..lLle5+, both 42 'itb5
and 42 ..td4 lLlc6+ 43 ..te4 lead to
victory, as the white pawns are ir
repressible.

1-0

102 DNIHJI OL (14) 1986

Oame 24
Karpov - Sznapik

Dubai OL (14) 1986

Sicilian Dragon
This encounter took place in the
last round of the Olympiad; our
team could not afford to drop even
half a point, so this game was deci
sive. By winning the match 4-0, we
won the Olympiad, ahead of Eng
land by only half a point.
1 e4
c5
d6
2 lDf3
3 d4
cxd4
lDf6
4 lDxd4
g6
5 lDc3
6 ..te3
..tg7
0-0
7 f3
lDc6
8 Wd2
9 ..tc4
..td7
lDeS
10 0-0-0
l:tc8
1 1 ..tb3
12 h4
Sometimes 12 ..tg5 is played im
mediately, but obviously, including
the moves h2-h4 and . . .h7-h5 is in
White's favour for a future attack
on the kingside.
12
h5
Black previously did not prevent
the h-pawn from moving. Its sac
rifice - 1 2 ... lDc4 1 3 ..txc4 l:txc4 14
h5 ! ? - has been seen countless
times, but a definitive conclusion
has not yet been made. However, it
most often favours White, and
..

therefore in reply to h2-h4, the

symmetrical . . . h7-h5 has now
gained popularity as a means of
hindering the movement of the
white g- and h-pawns.
1 3 ..tgS (D)

B
Attempting to blow up the en
emy fortress immediately with the
help of 1 3 g4 is unsuccessful, and
the bishop move to g5 looks more
solid than ideas such as 13 ..th6, 1 3
l:tdgl and 1 3 bl .
13
l:tc5
1 3 ... lDc4 and 1 3 . ..lDh7 are less
convincing.
14 b1
This waiting move, which has a
definite idea that will become clear
later, was suggested to me by Gel
ler while I was preparing for my
match in Merano. I had to wait five
whole years to use the idea in a
game !
14
b5 (D)

Karpov - Sznapik 103

w
15 g4!
Until that time 1 5 l:lhe1 was
played here, simply centralizing.
hxg4
15
A fundamental reply, which we
looked at in 1 98 1 . Some rounds
earlier in the USSR-Bulgaria match,
Kiril Georgiev chose another prob
lematic continuation against me,
1 5 ... a5, which we had not had time
to prepare for. The correct reaction
was 1 6 gxh5 ! , but I went for the
more modest 1 6 i.xf6 i.xf6 1 7 a3
and the game soon ended in a draw.
16 h5!
llxh5
Obviously, it was worth Black
giving up the exchange immedi
ately - 16 ...l:lxc3 ! ? 17 bxc3 llxh5
1 8 l:lxh5 gxh5 1 9 'ifh2 lbc4 20
'ifxh5. White has a certain initia
tive, but the whole game still lies
ahead.
17 lJd5!
Now the idea behind White's
1 4th move becomes obvious. An

analogous position has arisen in

practice several times, but without
the moves c 1 -bl and ...b7-b5. In
that case the manoeuvre 1 6 lJd5
(now it is made a move earlier) af
ter 16...l:lxd5 ! 17 i.xdS 'irb6 is not
dangerous for Black in view of his
pressure on b2 ( 1 8 ...llxf3 is threat
ened). Including the moves b1
and ... bS seriously changes the
situation - the exchange sacrifice
on d5 is now insufficient, as the b
file is covered by his own pawn,
and after 17 ...l:lxd5 1 8 i.xd5 gxf3
19 lbr5 it is only one step to victory.
17
l:le8 (D)

ri@;

'-
...
ri@;

ri@; '
. -

B
m
l. -"l.J- -

. -
Bi.. B B
"
pm

uri@;
B<t>B : :
w

1 8 l:lxhS!
White sacrifices the exchange
himself, hopelessly weakening
the guard around the enemy king.
18
gxhS
19 'ifh2! (D)
This precise position was the
subject of our old analysis. Then

104 Dubai OL (14) 1986

we established that White's small
material outlay is repaid with inter
est when the black king is attacked.
Because the defence cannot man
age without ... e7-e6, Black should
move his queen from d8, and we
considered it necessary here to re
turn the exchange - 1 9 ... llxd5 20
.i.xd5 (20 exd5 llxf3 21 llxf3
gxf3 22 'ibh5 .i.f5 is not so clear)
20 . . . 1ib6. Now if 2 1 'ifxh5, then
2 l . . . e6 is possible, with a resilient
defence, but Black's fortress can be
assaulted by 21 llf5 ! .i.xf5 22 exf5
ltJxf3 23 'ii'xh5 (the pawn must be
taken at precisely this moment; if
23 .i.xf3 gxf3 24 'ifxh5, then
24...'ifc5 !) 23 ...ltJxg5 24 'ifxg5, and
pressure by White on the g- and h
files threatens the most unpleasant
consequences. 19 ... a5 is terrible
for Black, for example 20 'ifxh5 a4
2 1 llh 1 axb3 22 'ifh7+ f8 23
.i.h6 .i.xh6 24 'iVh8#.

19 ...
llc4
Black prefers to give the rook
back for the bishop rather than the
knight, but it does not ease his
problems.
bxc4
20 .i.xc4
21 'ii'xhS
f6
22 f4!
liJf7 (D)
He will be saved by neither
22 . . . e6 23 llxf6+ .i.xf6 24 fxe5
.i.xg5 25 1i'g6+ nor 22 ...fxg5 23
fxe5 dxe5 24 llh1 (24 llf5 ! is even
stronger for White) 24 ... exd4 25
'i'h7+ f7 (25 . . .<it>f8 26 'ii'g6 ! ,
threatening both 27 l:l f l+ and 27
l:lh7, gives White a decisive attack)
26 llfl + <it>e6 27 'ii'xg7 .i.c6 28
'ifxd4.

23 .i.b4!
There is no sense in rushing, as
Black is totally helpless.
'ifb8
23
Again, if 23 ...e6, then 24 llxf6+
.i.xf6 25 1i'g6+ is decisive; after
..

Karpov - A. Sokolov 105

23 ...lbh6 there is 24 f5 l:tf8 25 lbf4
.te8 26 lbg6 .txg6 27 fxg6 fol
lowed by lbf5.
c3
24 l:th1
'ifb7
25 b3
26 r5
lbeS
27 lbe6
1-0
If 27 . . ..txe6, then 28 'iWxe8+,
and otherwise there is no defence
against 28 .txf6 .txf6 29 'iWh7#.
Game 25
Karpov - A. Sokolov
Linares Ct (10) 1987
Queen's Indian

1 d4
lbf6
e6
2 c4
b6
3 lbf3
.ta6
4 g3
.tb4+
b3
5
6 .td2
.te7
7 lbc3
It is curious that the Queen's In
dian was played in all five games
where I had White in this match
against Sokolov, and this position
arose automatically. It must be said
that the results exceeded expecta
tions - I won three games, and the
other two were drawn. However,
my success was due not so much to
good opening play, but to victories
in the ending.
7
0-0
Other games in the match against
Sokolov continued 7 . . . d5, and after

8 cxd5 lbxd5 9 .tg2 0-0 10 lbxd5

exd5 1 1 0-0 lbd7 a position arose
which was well known from the
first Karpov-Kasparov match. In
situations like this, White's altered
move order (.tg2 and lbc3) does
not matter. But now, after Black
has already castled kingside, the
position of the knight on c3 is not
that harmless for him, since White
can seize the centre by means of
e2-e4.
d5
8 e4
9 cxd5
.txn
10 xn
exd5
l l eS
lbe4
12 'iie2 (D)
In game 8 I played 1 2 l:tc l , pre
venting immediate regrouping by
means of ...'iid7 and ...lbc6-d8-e6.
However, this loss of tempo makes
itself felt, and Black could gain a
satisfactory position by playing
...c5 and ...lbc6.

IM UMNI a (10) 11'1

Thll time I managed to find a
more precise arrangement for my
pieces. The position of the queen

has great attacking potential con

nected with preparing f4-f5.
18 -txd4
'W'f5 (D)

on e2 besides everything else has

the important virtue of preventing
... f7-f5; after 12 ...f5 1 3 exf6, the
e4-knight is forced to abandon the
centre, and White has the e-file and
the e5-square at his disposal.
12
xc3
'W'd7
13 -txc3
14 g2
6
The idea behind this somewhat
fanciful knight manoeuvre is to
bring it to e6 to blockade White's
pawn centre.
15 l:the1
Introducing a natural plan:
White will throw the f-pawn for
ward and simultaneously control
the e-file. The move 1 5 l:tac l was
played here previously, allowing
Black to construct a defensive for
tress: 15 ... d8 1 6 -tb2 e6 17
l:thd l l:tae8.
15
d8
16 g1
c5
1 6 ...e6 is still impossible: 17
f4 f5 18 exf6 l:txf6 1 9 f5 ! wins a
piece. In the event of 16 ... f5 17
exf6 -txf6 1 8 f4 followed by f3e5, White has the advantage.
17 f4
cxd4
Black cannot manage without
opening the c-file, and he decides
to do so immediately. 1 7 . . c6
1 8 f3 g6 is risky, since White

19 l:tad1
-tb4
e6
20 :n
It looks as though Black is on
the point of seizing back the initia
tive, but, as so often happens, he is
a tempo short. If the rook were on
c8, White would not have the pos
sibility of 21 'ii'd3 due to 2 1 ...l:tc2+
22 l:tf2 'ii'xd3 23 l:txd3 xd4 24
l:txd4 :Xf2+ and 25 ...-tc5 .
21 'ii'd 3!
Destroying Black's blockade on
the light squares.
21
'ifxd3
l:tac8
22 l:txd3
23 f3
23 f5? is too hasty: 23 ...xd4 24
l:txd4 l:tc2+, when 25 l:tf2? fails to
25 ...l:txf2+ 26 xf2 -tc5.
23
l:tc2+
:res (D)
24 :n

Karpov - A. Sokolov 107

33 lbf3

.:.cl

(D)

Obviously, now Black is simply

placing his hopes on a counterat
tack against the white queenside
pawns. On other parts of the board
he has lost the fight.

25 f5 !
The phalanx of white pawns finally begins its forward march.

25
26
27
28
29
30

lbxd4
xfl
g4
3
g3

lbxd4
lbfl+
.:.et
8
.:.n+
.:.et

Checking from behind achieves

nothing: 30....:.g1 + 3 1 f4 .:.n + 32
lbf3 , while Black already has to
bear in mind the direct threat of 3 1
lbc6.

3t f4
After 3 1 g5, in the event of
3 I .. ..:.g1+ (or 3 l ...e8 32 lbf3) 32
f4 .:.n+ 33 lbf3 things are not
looking good for Black, but after
exchanging rooks with 3 I . . . .:.c3 32
.:.xc3 .i.xc3 33 lbc6 aS 34 f4
e8 his resistance could have been
prolonged.

3t
32 h4

h6
e8

w
34 a4
35 lbd4
36 h5

.:.bl
i..e7

I spent a lot of time on this

move. Of course, the continuation
36 g5 hxg5+ 37 hxg5 would have
given me chances for victory, but it
is more useful to fix and 'seal up'
Black's kingside and his pawns. I
am not exaggerating when I con
firm that at this point I had already
calculated a beautiful combination
a dozen moves long. But at the
same time I will not hide the fact
that I foresaw the possibility of the
pawn break.

36
37

a6
i..cS

/OB LlnaNs Ct ( 10)

1987

38 l0e2

d4
39 l0f4
d7
40 e6+
e8
Both now and later, taking on e6
is impossible since the g7-pawn
would be lost.
41 e4
aS
Here the game was adjourned.
Analysis showed that the game
could be won by study-like means.
42 .Uf3! (D)

B
The sealed move, which seemed
the strongest. The rook lies in an
ambush, anticipating the black
king wishing to cross the f-file.
42
.Ubl
Preparing for 43 d5, in which
case after 43 ....Ugl 44 lJd3 .Uxg4
45 lbxc5 bxc5 46 xc5 fxe6 47
fxe6 e7 Black would have held
the position. However, White's
knight goes off in quite a different
direction.
.Ugl !
43 lJd5!

43 ...f8 caused more trouble in

analysis. At first I examined the
move 44 lbc7, but I stopped every
thing when I found that after
44 ... .Ue 1 + 45 d3 .Ue5 ! Black can
hold the position. Only the next
morning did I manage to find the
correct route: 44 e7+ ! i..xe7 45
xd4 .Ugl 46 c4 .Uxg4+ 47 b5
i..d8 48 .Uc3 ! .Uh4 49 c6 ! .Uxh5
50 <i'd7 .Uxf5 and here 5 1 lbxb6! is
the end. 5 1 . . . i..xb6 allows a rather
funny mate: 52 .Uc8+ i..d 8 53
.Uxd8#. Therefore Black has to give
back the exchange with 5 1 .. ..Uf3,
and after 52 .Uxf3 i..xb6 the inevi
table break b3-b4 is decisive. A
unique case, in which one king is
the main director and executor of
an attack on the other!
44 <i'd3!
.Uxg4
45 f6! (D)

B
This is White's quintessential
idea. He is threatening 46 llc7+

Karpov - A. Sokolov 109

and 47 e7+, and none of his pawns
can be taken: 45 ...fxe6 46 f7+ d7
(46 . . . f8 47 li:Jc7 and 48 li:Jxe6+;
46 . . .d8 47 f8'ii'+ xf8 48 llxf8+
'iti>d7 49 xb6+) 47 li:Jxb6+ 'iti>c7 48
li:Jc4, threatening li:Je5 and li:Jg6.
45 . . . llg5 46 li:Jc7+ 'iti>d8 47 e7+
xe7 48 fxe7+ 'iti>xe7 would have
been more stubborn, with hopes
for the kingside pawns.

45
46 li:Jxb6

d6
llgS

Again 46 . . .fxe6 will not do: 47

f7+ 'iti>f8 48 li:Jd7+. Black is not
helped by 46 . . . gxf6 either: 47
llxf6 ! fxe6 48 llxe6+ e7 49 li:Jd5
leads to a winning pawn endgame
for White.

47 fxg7

llxg7

The final hope for salvation lay

in 47 . . .fxe6 48 li:Jc4 e7 (if Black
plays 48 ... b4, then 49 xa5 ! is
decisive) 49 li:Jxa5 llxg7 50 li:Jc6+
'iti>d7 5 1 li:Jxd4, although I suggest
that I would have coped with the
diffkulties.

48 li:Jc4
49 exf7+
so llxf7

b4
llxf7
'iti>xf7 (D)

The storm of the combination

has abated, and the game has
moved into its technical phase.
This gives the impression that the
worst is over for Black, but there is
more to come.
51 li:JeS+!
6

52 li:Jc6

el

- - .

. . - .
- . . -
d
ttJ

- - d
.

. . - .
w

53 li:Jxb4 was threatened, while

after 52 ...c5 53 li:Jxa5 g5 54
li:Jc6 xh5 55 b4 the a-pawn will
queen.

53 li:Jxd4

b4

Black would not be saved by

53 . . .f2 54 li:Jc6 g5 55 xa5
xh5 56 e2 ! , followed by 57
'iti>f3 and, if necessary, 58 g2. The
queenside pawns can easily move
forwards. Surprisingly, the knight
turns out to be stronger than the
bishop with play on both flanks.

54
SS
56
57
58

li:Jc6
e2
'iti>d3
'iti>c4
li:JxaS!

el
c3
el
..tgs

Both the most effective and the

most striking line. Black cannot re
fuse to take the piece: 58 ...xh5
59 li:Jc6 'iti>g4 (after 59 ... g5 60 b4
h5 White wins a decisive tempo by
means of 6 1 li:Jd4, threatening 62
li:Jf3+) 60 b4 h5 6 1 aS h4 62 a6 h3

1 10 Stvllll Wch (2)

1987

(or 62 .tf2 63 lbd4) 63 a7 h2 64

aB'ii' h 1'ii 65 lbe5+.
58
.txaS
59 b4
.tdS
'itxhS
60 aS
61 '1tb5
.tgS
62 a6
.te3
63 'itc6
1-0
...

Game 26
Kasparov - Karpov
Seville Wch (2) 1987
English, Four Knights
1 c4
I must confess that I was not in
itially expecting Kasparov's con
version to the English Opening in
this match. However, this surprise
at the start strangely turned into a
gift for me straight away - at least,
Kasparov spent almost an hour and
half on his first ten moves ! It is in
teresting that the novelty I used
was prepared at the beginning of
the 1 980s for the Merano match.
However, instead of 7 lbg5, Korch
noi retreated the knight to e 1 , and
so it remained unused.
1
eS
2 lbc3
lbf6
lbc6
3 lbf3
.tb4
4 g3
0-0
5 .tg2
e4
6 0-0
.txc3
7 lbgS
8 bxc3
:es

9 f3

e3! ?

(D)

w
It was precisely this move, sug
gested in its time by Igor Zaitsev,
which plunged my opponent deep
into thought (until then Black had
always taken automatically on f3).
10 d3!
The correct reply, so White has
not wasted his thinking time. Let
us look at two other possibilities:
a) 10 dxe3 b6 ( 10 ... h6 1 1 lbh3
d5 12 lbf4! dxc4 1 3 e4 is worse) 1 1
e4 and then:
a l ) 1 l .. . .ta6 immediately is
dangerous : 12 f4 h6 13 lbxf7 !
cl;xf7 14 e5 .txc4 (14 ... lbg8? 1 5
.td5+ '1tf8 1 6 .ta3+ lbge7 17 'iid3
d6 18 'iih7 ! winning) 15 exf6 gives
White an advantage.
a2) 1 1 ... h6 12 lt:Jh3 .ta6. The
c4-pawn is doomed, and Black can
be quite satisfied at the results of
the opening.

Kasparov - Karpov
b)
.i.b7 !
as in
grade

10 d4 iLia5 1 1 'ifd3 b6 12 c5
and Black has the initiative,
H.Olafsson-Naumkin, Bel
1988.
10
dS
The only move, because other
wise the e3-pawn will quickly be
surrounded.
1 1 'ii'b 3!
Yet another precise move; after
1 1 cxd5 iLixd5 two lines are possi
ble:
a) 1 2 iLie4 f5 1 3 c4 fxe4 14
cxd5 exf3 15 l:txf3 ilJd4 1 6 l:txe3
.tg4 17 .te4 'iff6 1 8 .tb2 'ifb6 ! 19
.txd4 'ii'xd4 20 'iic l .txe2 21 1itg2
.txd3 with a winning position for
Black.
b) 1 2 'ii' b3 ilJa5 ( 1 2 ...'iixg5 1 3
f4 iLixf4 1 4 l:txf4 leads to a sharp
game) 1 3 'iia3 'iixg5 14 'iixa5 'ii'e5
( l4 . . . b6 15 f4) 1 5 d4 ( 1 5 f4? iLixf4
16 'iixe5 iLixe2+ 17 1ith l l:txe5 1 8
.i.b2 .:r.b5) 1 5 ...'ifd6, and the game
is roughly even: 1 6 .ta3 'ifc6 17 f4
'iixc3 1 8 'ifa4 iLif6.
11
ilJaS
12 'ii'a3
c6
13 cxdS
cxdS (D)
14 f4
ilJc6
15 l:tb1
'ii'c7
After 15 ....i.g4 16 .l:tel (the most
precise; after 16 l:txb7 .txe2 17
lbxf7 'ifa5 ! 1 8 'ifd6 .i.xfl 1 9
iLih6+ lith8 20 ilJf7+ White has to
settle for perpetual check) 1 6... h6
( 1 6 . . .'iid7 1 7 h3) 1 7 iLif3 'ifd7 1 8
..

Ill

i. -

.
.
. .

.
d
-

.
.

- .
"' '/!iJAo .
% oA B
"' '"'
'
/,-

!'?

!1,:::

8 Bl 8 -B

-
.:*
w

d4, by directing the knight to e5,

White is better.
16 .tb2
.tg4
17 c4 (D)
Of course, it is very tempting to
threaten the enemy king, but this
gives Black the initiative. 17 iLif3
'ifd7 1 8 .tal .l:te7 ! ? 1 9 l:tfc l .i.h3
is not dangerous either - Black has
sufficient counterplay. However,
by continuing with the quiet 1 7
l:tfe l White would have retained
good chances.
17
dxc4
gxf6
18 .txf6
19 ilJe4
litg7
1 9....:Xe4 ! ? 20 .i.xe4 f5 ! 2 1 .tf3
ilJd4 22 dxc4 i.xf3 23 exf3 e2 24
l:tfel 'ifxc4 is also interesting.
20 dxc4
Now the advantage fully trans
fers to Black. The continuation 20
iLixf6 ! ? litxf6 21 .:r.b5 l:tad8 22
'ifc3+ lite? 23 'ifxc4 would have
led to uncertain complications.

1 12 S1vlll Wch (2)

1987
l:.ad8
20
More reliable than 20....txe2 2 1
lbxf6! .
2 1 l:.b3 (D)
Apparently the decisive error.
After the continuation 2 1 lbc3
li:)d4 22 li:)d5 there is still a stub
born fight to face. Its direction de
pends on the queen's choice of
whether to go to d6 or c4:
a) 22... 'ii'd6 23 'irxd6 l:.xd6 24
l:.xb7 .txe2 25 l:.e1 .i.xc4 26 lbxe3
l:.de6 27 l:.b4 lbe2+ 28 f2 .i.d3
29 lbd5 lbxf4 30 lbxf4 :Xe1 3 1
lbxd3 l:.8e2+ 32 f3 l:.e3+ 3 3 f2
l:.1e2+ 34 1 :Xg2 35 xg2 lhd3
36 l:.a4, and as a result chances are
equal.
b) 22 ...'ii'xc4 23 li:)xe3 'ii'xe2 24
li:)xg4 'ii'xg4, with a tense situation
on the board.
...

B
Here many commentators have ex
amined the more natural 20 'ii'c 3,
and in this case after the correct re
ply 20 . . .'ird8! (20 .. .'1Ve7 2 1 l:.xb7 !
'ii'xb7 22 lbxf6 f8 23 lbxg4 ! and
White has the upper hand) 2 1 dxc4
'ii'd4 22 :Xb7 l:.xe4 23 'ii'xd4 l:.xd4
24 .txc6 the chances are roughly
even.
Another dangerous manoeuvre
is 20 lbd6, and Black has to play
carefully: 20 . . .l:.e6 2 1 lbxc4 l:.d8
22 f5 l:.ee8 23 l:.b2 lbd4 24 l:.xb7
lbxe2+ 25 h1 lbxg3+! 26 g1
lbe2+ (neither 26 ...'ii'xb7 27 .txb7
lbxfl 28 xfl .txf5 29 lbd6, nor
26 . . . lbxfl 27 l:.xc7 e2 28 'ii'xa7
e1 'iV 29 l:.xf7+ h6 30 l:.xf6+ !
g5 3 1 'ii'g7+ f4 32 'ii'h6# will
do) 27 h l 'ii'g3 with perpetual
check. However, in the last vari
ation Black could take a risk:
27 ...'ii'c 8 ! ? 28 lbd6 l:.xd6 29 'ii'xd6
.txf5 with enough compensation
for the exchange.

Karpov - Farago 113

24 l:td3
25 hd8

xe2
hd8 (D)

w
26 l:te1
l:te8
The game is decided. Black has
prepared a simple mating idea: 27
lDd6 ltlxd6 28 1i'xd6 f3 29 :Xe8
1i'fl#.
27 1i'a5
bS
1i'd3
28 ltld2
29 ltlb3
f3
30 xf3
1i':d'3+
l:txe1+
31 g1
32 1Wxe1
lLle3
0-1
Game 27

Karpov - Farago
Wijk aan Zee 1988
French, Winawer

1
2
3
4

e4
d4

lbc3
eS

e6
dS
b4
ltle7

S a3
xc3+
6 bxc3
cS
7 1i'g4
I had prepared this variation for
my match against Korchnoi at Ba
guio. Ten years later, I finally got a
chance to try it in practice.
7
1i'c7
8 1i'xg7
l:tg8
9 1i'xh7
cxd4
One of the key positions in this
opening.
10 ltle2
A versatile manoeuvre: White's
knight is covering the queenside
and simultaneously supporting the
pawn advance on the kingside.
10
ltlbc6
The pawn on e5 is invulnerable:
10 ...1i'xe5 1 1 4 1i'f6 12 cxd4 with
a big advantage to White.
1 1 f4
White's basic idea is to prepare a
pawn advance on the kingside, and
1 1 f4 runs contrary to this. Inci
dentally, l l cxd4 is no good due to
1 1 .. .ltlxd4 ! .
11
d7
dxc3
12 1i'd3
13 1Wxc3 (D)
White takes the pawn immedi
ately, and with it opens up both the
c-file and the entire queenside, so
he has to be extremely careful. 1 3
lDd4, 1 3 ltlg3, 1 3 e3, and 13 l:tb1
have all been played several times,
but with material equality Black
..

1 14 W(Jl Cltlll bt 1988

can easily count on a successful
outcome. There is no difference in
value between taking with the
queen or the knight on c3, and in
any case I have tried both routes. If
the queen takes on c3, the black c6knight turns out to be pinned, while
the queen herself does not feel too
settled (the threat is the advance
...d5-d4 ); if the knight takes it then
the black knight on c6 is free to
manoeuvre.

B
ti::lfS
13
14 l:tb1!
This move has been given an ex
clamation mark not because it is so
cool or particularly strong, but due
to other considerations: this game
is unique in the number of short
moves made by the white rooks out of the 27 moves remaining in
the game, 1 5 are made by the
rooks!
14
l:tc8

15 ..td2
b6
In his game against Tal in the
same tournament, Farago used the
successful novelty 1 5 ... a6 ! and
equalized. I still think that all
White's resources have not yet
been exhausted here.
16 g3 (D)
The exchange 16 ti::lg3 ti::lxg3 1 7
hxg3 ti::le7 1 8 "iixc7 l:txc7 19 ..td3
l:txg3 20 l:th8+ l:tg8 2 1 l:txg8+
ti::lx g8 22 ..tb4 led to a better end
game for White in Martic-Dras
kovic, corr. 1 987, but I prefer to
maintain the tension on the board.
Incidentally, only now does this
game acquire independent signifi
cance. Of course, I would have
liked to play g2-g4 in one go, but if
16 l:tgl , then 1 6 ...1fd8 ! with the
threat of ...ti)cd4 and ...1fh4+.

16
17 .d3
18 l:tg1

.b7
ti::lce7
l:tc4! (D)

Karpov - Farago

1 15

....... . . .
. --
.

--

--
}, .
"

rff// -
\Wr
u
-

"'

mltj
u
- .

,Wit :

i.. "

w
Black's rook is ideally placed
on c4: it is working both along
White's fourth rank and the c-file,
and furthermore it is covering the
fl -a6 diagonal from the possibility
of transferring the queen to a6 or
the bishop to a4 (or b5).
19 g4
ll'lh4
,.c8
20 l:.g3
ll'lhg6
21 l:.b2
22 l:f3! (D)
Overprotection of the f4-pawn.
The white rook on g3 was beauti
fully fulfilling its defensive func
tion on the third rank and the g-file,
but now it is time to get down to ac
tive operations, for which it is nec
essary to regroup.
22
ll'lc6
Black cannot repeat the position
with 22 ... ll'lh4 because of 23 l:.f2 ! ,
and then 2 3. . .l:.xg4 i s impossible in
view of 24 'ilfh3 ! , when White is
better.
.:hs
23 m

B
24 h3
ii'd8
25 'ii'g3!
Following Black's example of
rearranging his pieces on the c-file,
I also decided to swap round my
queen and rook.
'ilfe7
25
ii'cS
26 .:o
27 'itd1
White is planning to begin ma
noeuvring the knight, therefore it
is best to remove the king from the
possibility of a check from c4 be
forehand, simultaneously defend
ing the pawn on c2. Incidentally,
transferring the king to dl now is
better grounded than it was on
move 10.
l:.a4 (D)
27
In a complex position Black is
losing the thread of the game. It
was necessary to play 27 ... d4 28
ll'lc l l:.c3 29 .i.xc3 dxc3 30 l:.b5
'ilfd4+ 3 1 l:d3 'ilfxf4 32 'ilfxf4 ll'lxf4
33 l:.xc3 ll'ld5, and Black holds on,

1 16 Brussels World Cup 1988

although 3 1 i.d3 is stronger for

White.

w
28 .Uc3!
Giving Black no respite. 28 lbc 1
lba5 29 l:c3 30 l:b4 is weaker.
Now Black should return the rook
to c4, as retreating the queen leads
to an almost forced loss.
'ile7
28 ...
29 .!i)ct
The white pieces have unex
pectedly developed stormy activity
exactly where Black usually domi
nates. White is threatening 30 i.b5,
and if29... lba.S, then 30 .Ub4 is rea
sonable.
d4
29
Somewhat spoiling White's in
tentions, but having torn itself
loose, the central pawn will need
constant defence.
.!i)a5
30 .Ucb3
31 .Ub4!
.!i)b7
i.xa4
32 .Uxa4

33 .Ub4
34 .!i)b3
Matters are coming to a head;
34...i.xb3 is impossible because of
35 i.b5.
34
d3
35 'ili'xd3
'ilt'xd3
36 i.xd3
i.xb3
37 .Uxb3
.!i)cS
Not 37 . ..l:xh3 because of 38
i.b5+.
</;e7
38 .Uc3
39 i.n
.Ud8
1-0
40 .Uc4!
Game 28
Karpov - Timman
Brussels World Cup 1 988
Queen's Gambit Accepted
d5
1 d4
dxc4
2 c4
3 e4
The classical continuation is 3
.!i)f3, 4 e3 and 5 i.xc4, immediately
winning back the pawn. However,
sometimes I like playing in pure
gambit style, acting in accordance
with the name of the opening!
.!i)f6
3
4 e5
.!i)d5
.!i)b6
5 i.xc4
6 i.d3
It is still not known which re
treat is better for the bishop, d3 or
b3. Before we continue I would
like to pay a little attention to the

Karpov

second of these: 6 .i.b3 lDc6 7 .i.e3

.i.f5 and now:
a) 8 lbe2 e6 9 0-0 lDa5 I 0 .i.a4+!
c6 1 1 .i.c2 .i.g6 12 lbbc3 lbac4 1 3
'iVc 1 .i.e7 1 4 .i.xg6 hxg6 1 5 lbe4
l:h4 1 6 lD2g3 'iVd5 1 7 'iVc2 0-0-0
1 8 .i.g5 .i.xg5 1 9 lbxg5 l:d7 20
lDf3 l:tf4 21 l:ad1 g5 22 1fc1 ! gave
White an appreciable advantage in
Karpov-Speelman, Brussels 1 988
and the game ended in my favour.
b) 8 lDc3 e6 9 lbge2 .i.e7 10 a3
0-0 1 1 0-0 lba5 I 2 .i.c2 .i.xc2 1 3
'it'xc2 lDac4 14 l:adi lDxe3 1 5 fxe3
and I quickly managed to seize the
initiative in Karpov-Short, Linares
Ct (9) 1 992.
6 ...
lbc6
7 lDe2 (D)
White can also play 7 .i.e3 :
a) 7 . .. lbb4 looks good: 8 .i.e4
f5 ! 9 exf6 exf6 1 0 lbc3 f5 1 1 .i.f3
lD4d5 1 2 .i.d2 .i.e6 1 3 lDge2 'iVd7
1 4 0-0 0-0-0 1 5 l:e 1 ( 1 5 a4 ! ? de
served attention) 1 5 ...l:tg8 ! , and
Black had solved all his opening
problems in Karpov-Short, Linares
Ct (3) 1 992.
b) 7 ... .i.e6 8 lbc3 'iVd7 9 lDf3
0-0-0 10 h3 lbb4 1 I .i.e2 f5 12 0-0
h6 1 3 a3 lD4d5 14 lDe1 lDxc3 1 5
bxc3 .i.c4 1 6 lbd3 e6 I 7 a4 g5 I 8
1fc2 'iVc6 1 9 :tfc I gave White the
better chances in the game Karpov
Ivanchuk, Reggio Emilia I99112.
7
.i.g4!?
8 .i.e3

Timman 1 17

B
After 8 f3 .i.h5 and 9 ... .i.g6
(White must avoid continuing 9
lDf4 owing to 9 ...'iVxd4 10 lbxh5
'iVxe5+) Black has a comfortable
game (the position resembles the
Alekhine Defence, Four Pawns At
tack).
8 ...
.i.xe2
9 .i.xe2
'iVd7
Black does not use up a tempo
playing ...e7 -e6, preferring instead
to increase the pressure on the d4pawn. In a situation like this one
immediately has to resort to tactics.
0-0-0
10 lbc3
11 a4! ?
With the intention of distracting
White's attention from the d4square.
a6!?
1 1 ...
If I l .. .lbxd4, then 1 2 a5 lbxe2
( 1 2 ... lba8 is no better: I 3 a6 b6 I4
.i.f3, and the black knight is
trapped in the corner of the board)

I 18

Brussels World Cup 1988

1 3 axb6 xc3 14 iic2 cxb6 1 5

'it'xc3+ with a very strong attack
for White: 1 5 ... b8 16 e6 iixe6 17
0-0, etc.
12 aS
dS
13 i.f3
db4 (D)
1 3 ... e6? loses immediately in
view of 14 xd5 exd5 1 5 i.g4.
Exchanging on c3 or e3 favours
White, as he gains a strong pawn
centre. However, 1 3 . . .cb4 looks
stronger.

w
14 e6!?
The prelude to all further play
with an unusual correlation of
forces. After the quiet 14 0-0, tak
ing the central pawn is risky:
14 ...xd4 (Black's best chance is
probably 14 ... e6, with a complex
game) 1 5 i.xd4 'it'xd4, and now
alongside 1 6 Wb3 e6 17 llfd 1
iixe5 1 8 i.xb7+ xb7 1 9 llxd8
i.d6 20 llxd6 iixd6 21 a2 c5 22
xb4 cxb4 23 'it'f3+, which gives

White a minimal advantage, he

could immediately take the bull by
the horns with 1 6 i.xb7+ ! xb7
17 Wf3+, and then both 17 ...c6 1 8
llfd I d3 19 lla4 Wd7 20 e6
2 1 d6+ followed by 22 llxd3,
and 17 ...b8 1 8 llfd1 3 1 9 lla4
'it'd7 20 lla3 ! are unfortunate for
Black.
14
'ili'xe6
After 14 .. .fxe6 1 5 a4 Black
does not become disentangled:
15 ... xd4 16 i.xb7+!.
'ili'eS
15 dS
There is another interesting at
tempt, 1 5 ...Wg6. Then the continu
ation 1 6 i.h5 'ifxg2 17 'it'g4+!
looks tempting, but if instead
Black replies 1 6 ...c2+ ! 17 fl
xe3+ 18 fxe3 Wf6+ 19 g1 e6,
the complications are not at all in
White's favour. Probably here as
well everything would have been
reduced to a queen sacrifice for
rook and knight, but in a slightly
different way: 1 6 0-0 e6 17 dxc6
llxd 1 1 8 cxb7+ b8 19 llfxd l
i.d6. White's initiative compen
sates him for his small material
loss, although the chances can be
considered roughly equal.
16 0-0
e6 (D)
17 dxc6!
The logical continuation of the
plan White began on move 1 1 .
17
llxd1
b8
18 cxb7+
.

Karpov - Timman ll9

24 ... lDe5 loses: 25 l:tdxc7 lDxf3+

26 gxf3 Wxa5 27 lDc5 Wxc7 28
lDxa6+ xb7 29 l:txc7+! xa6 30
l:txg7 and White has a technically
won ending.

w
19 l:tfxd1
i.cS
19 . . . i.d6 is more stubborn, pre
venting the white rook from pene
trating the seventh rank.
20 i.xcS
'iWxcS
fS
21 l:td7
lDc6
22 l:tad 1
White was threatening l:td8+,
and in the event of 22 . . . c6 23 h4 !
(but not 23 l:txg7 lDd3 ! 24 l:txd3
'iWe5 ! ) White has an obvious posi
tional advantage.
23 lDa4
23 lDe2 !? is also good: 23 ...'ir'b5
24 l:tcl lDe5 (the most obstinate
continuation is 24 ... lDa7 ! , though
here as well Black experiences se
rious problems) 25 l:tdxc7 lDxf3+
26 gxf3 Wxa5 (Black cannot con
tinue 26 . . . 'iWxe2 due to 27 l:tc8+
xb7 28 l:t l c7#) 27 l:tc8+ xb7
28 l:txh8 'ir'd2 29 fl is insufficient for Black.
23
24 l:tcl

w
'iWxaS
25 .:Xc6
26 l:txe6
a7
27 g3 (D)
27 h4 ! is even stronger, com
pletely depriving Black of counter
play on the kingside.

120 Btlfort World Cup 1988

27
g5!
28 ltxh7!
ltb8
29 hJ
g4
fxg4
30 hxg4
31 .i.g2
'it'a1+
'it'xb2
32 ..th2
'it'a2
33 lthh6
c5
34 ltef6
34...ltxb7 would have lost: 35
.i.d5 ! 'it'a5 (35 ...'it'xd5 36 ltxa6+
..tb8 37 lth8+) 36 .i.xb7 ..txb7 37
ltxa6 'ifxa6 38 ltxa6 ..txa6 39 f4!.
35 ltf4 (D)

B
'it'd2? !
35
35 . . .c4 was necessary, although
it is true that after 36 .i.d5 White
has taken three black pawns for his
own on b7, and finally an endgame
would arise with a rook, bishop
and pawns on f2 and g3, against the
queen, which is good for White.
36 .i.n !
Lb7
..tb8
37 ltxa6+
..tc7
38 ltf8+

39 .i.g2
Co-ordination between the rooks
and the bishop creates decisive
threats.
39
'it'd7
c4
40 lth8
1-0
41 .i.e4
Game 29
Karpov - Kasparov
Belfort World Cup 1 988

Griinfeld Defence
1 d4
llf6
2 c4
g6
d5
3 lbcJ
4 cxd5
lbxd5
lbxc3
5 e4
.i.g7
6 bxcJ
7 .i.c4
The plan of 7 lbf3 c5 8 ltb1 0-0
9 .i.e2 has enjoyed great popularity
in recent years. Another plan of
development for White involves
supporting his mighty pawn centre
with an early .i.e3, followed by
bringing the light-squared bishop
out to e2.
c5
7
lbc6
8 lbe2
0-0
9 .i.e3
.i.g4
10 0-0
It is well known that in the event
of 10 . .. cxd4 1 1 cxd4 lba5 1 2 .i.d3
.i.e6 1 3 d5 White can sacrifice the
exchange, gaining a dangerous at
tack.

Karpov - Kasparov 121

1 1 f3
lLlaS (D)
Now in reply to 12 i.d3 there is
the possibility of 12 ...cxd4 13 cxd4
i.e6, when sacrificing the exchange
with 14 d5 is not so effective, as in
a lot of variations Black has a
queen check on b6. However, in
this case White is not at all forced
to give up the rook for the bishop.

w
12 i.xf7+
I specially prepared this cap
ture on f7 for the Seville match.
Immediately after it had ended
Kasparov remarked that the vari
ation is hopeless for White. But if
you judge from what happened in
the five games with it in Seville,
you cannot say that it is all that
easy for Black to solve the prob
lems of the opening. Moreover the
proposed post-match duel in gen
eral ended miserably for him. Only
later was it established that in the
event of i.xf7+ and the exchange

of light-squared bishops, Black

gains a favourable game.
l:txf7
12
13 fxg4
l:txfi+
14 Wxfl
The point of White's idea is not
to take the pawn, as his activity
compensates Black for his small
loss, but to block in the enemy g7bishop with the help of the pawn
chain c3, d4, e5, g5 and h4.
'iVd6
14 ...
This queen thrust occurred in
four out of five games in Seville,
and it has also been used in many
other encounters. The other con
tinuations, 14 ...cxd4, 14 ...'ili'd7 and
14 .. .'ii'e8 , are not really justified.
15 eS
'iVdS
l:td8 (D)
16 i.f2

w
The threat of a capture on e5
forces the white queen to vacate
d 1 . In this game I moved her to a4,
while later against Timman I

122

B1(fo11 World Cup 1988

played 17 Wc2, a game which also

proved to be very engaging.

17 . . . Wc4 1 8 'ifb2 .i.h6 19 h4, and
now:
a) 19 ...'iff7 20 g1 (20 g5 ! ltX4
2 1 e6 ! is correct; thanks to this
zwischenzug, White diverts the
queen, and does not allow Black to
double on the f-file: 2 1 . . .'iff5 22
g3 Wxe6 23 Wxb7 llf8 24 lle 1 !
e3+ 25 g1 l:txf2 26 gxh6 cxd4
27 cxd4 llf8 28 'ifb2 4 29 llxe6
1 -0 Vyzhmanavin-Emst, Stock
holm 1 99 1 ) 20 ...l:tf8 2 1 g3 c4
22 We2 Wxf2+ ! 23 'ifxf2 .i.e3 24
'ifxe3 xe3 and a roughly even
endgame arose in Karpov-Kaspar
ov, Amsterdam 1988.
b) 19 ... llf8 20 g5 ! (a universal
move which guarantees White an
advantage in this case) 20...'ifd3 2 1
'ii'b 1 ! 'ife3 2 2 1We 1 .i.g7 2 3 g1
We4 24 g3 ! (White returns the
pawn but seizes all the critical
squares) 24...'irxh4 25 4 llxf2
(a desperate sacrifice, similar to
what happened in the Belfort game
I am commentating on) 26 xf2
cxd4 27 lld 1 d3 28 'ire3 ! c6 29
xd3 'ira4 30 'irf3 'ifa5 3 1 e6 d8
32 4 .i.e5 33 llJd5 'ifc5+ 34 h1
1 -0 Karpov-Timman, Rotterdam
1 989.
17 'ifa4 (D)
Now after . . .b6, the queen will
(as we see) retreat to c2. The ques
tion is, why does White choose to

provoke such a useful pawn move

as . . . b6? The fact is that after 1 7
'ifa4 b6 1 8 'ifc2 in some variations,
for example in reply to 1 8 ...'ifc4,
White switches the queen to e4
and, by attacking the knight, wins
an important tempo. Thus, moving
the queen out to a4 has its advan
tages. But I will not undertake to
decide definitively where the
queen is more comfortable, al
though I have tried both possibili
ties in practice.

- -
-

N
- u

tb p
a

B
17 ...
b6
White gets a better game after
17 ...6 1 8 'irb3 c4 19 'ifxb7.
If Black prefers 17 . . .c4, then
after 1 8 f4 1Wf7 1 9 g3 d2+ 20
g2, White has the advantage in
the event of 20. . .lbe4 (Gutman) 2 1
'ifc2 xf2 22 'irxf2, but Black has
the strong reply 20 ...g5 ! as demon
strated by Gutman. He further sug
gested that White should play 1 8

Karpov - Kasparov 123

g5 ! himself, followed by 18 . . .:fs

1 9 g1 Wf7 20 ..tg3 3 2 1 ..tf4
l'bd5 22 :n ; Black's initiative has
been arrested, and his bishop is un
der lock and key.
After the continuation 17 . . .:fs
1 8 g1 1lf7 19 ..th4 ! l'bc4 20 'ifb3
the bishop again does not manage
to break free: 20 . . . ..th6 2 1 g5 !
.i.xg5 22 .i.xg5 11f2+ 23 h1
Wxe2 24 h3 and White is apprecia
bly superior.
18 Wc2
The queen has nothing left to do
on a4.
18
:rs
Now the black pieces fall into a
clamp. 1 8 . . .:cs is more precise,
with counterplay on the c-file.
19 g1
11c4 (D)
Black's bishop still cannot break
out to freedom: 19 .....th6 20 h4 'iff7
2 1 l'Dg3, alternatively 1 9 . . . l'Dc4 20
h4.

20 Wd.2!
White continues to play to limit
the mobility of the bishop. In the
event of 20 'ife4 l'Dc6 ! ? the threat
is to take on e5, and 2 1 Wxc6 Wxe2
is natural, but not a cause for de
light.
20
We6
20 . . .11f7 is no good: after 2 1
l'Dg3 everything turns out happily
for White - the knight moves to e4,
and the queen to e2. Black is not
saved from difficulties by 20.....th6
2 1 11xh6 Wxe2 either, in view of
22 We3 'iVxg4 23 dxc5 bxc5 24
Wxc5.
21 h3
l'Dc4
22 1lg5! (D)
An important moment. Besides
l'Df4, now ..th4 is also threatened
when the time comes.

h6
22
22 . . ...tf6 does not work due to
23 exf6 exf6 24 l'Df4.

124 Be/fort World Cup 1988

23 'ifcl
'ii'f7
In order to gain counterplay
B lack should have chosen 23 . . .b5
with the idea, for instance, of 24
tbf4 iif7 25 tbd3 (after 25 -'.g3
White would preserve his advan
tage) 25 ... b4 !?. On the other hand
23 . . . 'ii'd5 does not achieve Black's
aim because of 24 'ii'c2! , and after
23 ... h5 the queen returns to g5.
g5
24 g3
This move was condemned by
the commentators, but alas, the
recommended 24 ...iid5 25 tbf4
'ii'e4 26 tbe6 would also have led to
a difficult position for Black. Here
are three key variations:
a) 26 ...l:tc8 27 'ii'b l ! 'i!Ve3+ 28
-'.f2 Wxc3 29 'i!Vxg6 'i!Vxal + 30
'it>h2 with inevitable mate.
b) 26 ... liJe3 27 'ii'd2 cxd4 28
cxd4 l:tc8 29 l:tel l:tc2 30 l:txe3
'ii'c6 3 1 d5 and Black's position is
bad. He could try driving the queen
to e2 by playing 27 ... ltJc4 28 'ii'e 1
lLle3 29 'ii'e2 cxd4, but 30 lLlxf8
-'.xf8 3 1 'ii'f3 ! 'i!fd3 32 cxd4 'ii'xd4
33 l:te1 brightens up White's situ
ation completely: 33 ...liJc2+ 34 -'.t2
ltJxe1 does not work because of the
intervening 35 Wb3+.
c) 26 ...cxd4 27 lLlxf8 liJe3 28
'ii'd2 dxc3 29 'ii'e 2 -'.xf8 30 'ii'f3
with a big advantage.
'it'd5
25 'it'c2
b5
26 -'.f2
.:tf7
27 liJg3

Forced, as if 27 ... b4, then 28

liJf5 and after 28 . . .l:tf7, White has
29 e6 'ii'xe6 30 l:te1 Wd7 3 1 cxb4.
b4
28 l:te1
wrs
29 'ii'g6
29 . .. bxc3 loses immediately: 30
liJf5 Wf8 31 e6 l:txf5 32 gxf5 liJd6
33 dxc5.
30 liJe4 (D)

B
l:txf2
30
This exchange sacrifice hardly
improves the situation, and White
only needs elementary accuracy.
31 Wxf2
bxc3
g8
32 'ii'f5+
33 'ii'c8+
h7
iif7+
34 iixc5
c2
35 g1
-'.f8
36 liJg3
..ti>g8
37 liJf5
38 .:tcl
1-0
Black's pieces on the kingside
could not break free. This was de
clared the best game in the next
.

Karpov - M. Gurevich I 25

edition of Jnformator, and as it was

played in the French town of Bel
fort, someone as a joke called the
variation with 1 7 iVa4 the 'Belfort
Variation' .
Game 30
Karpov M.Gurevich
USSR Ch (Moscow) /988
Queen's Indian
-

1 c4
lbf6
e6
2 lbf3
3 d4
b6
4 g3
J.a6
J.b4+
5 b3
J.e7
6 J.d2
In the game Karpov-Short, Am
sterdam 1 988, after 6 . ..J.xd2+ 7
'ifxd2 0-0 8 J.g2 c6 9 0-0 d5 10
'ifb4 lbe4 1 1 l:tc l lbd7 1 2 'ir'a3
J.b7 1 3 cxd5 exd5 1 4 lbc3 f5 1 5
e 3 the dark-squared strategy led
White to victory; after 1 5 ...'ii'f6 16
l:tc2 a5 17 :td 1 l:tac8 1 8 lbe2 g5 1 9
lbc 1 g4 20 li:Jh4 m y opponent had
got nowhere.
7 J.g2
c6
The gradual preparation of ... d5
has recently supplanted immediate
action in the centre by means of
7 ... d5 or 7 ...J.b7 8 li:Jc3 d5. These
move-orders were seen a few times
in my first match against Kasparov.
8 J.c3
dS
9 li:JeS

In game 1 8 of the above-men

tioned match, at a point when my
opponent was not particularly try
ing for an active game, he played 9
li:Jbd2, and after 9 ... li:Jbd7 10 0-0
0-0 1 1 l:tel c5 1 2 e4 dxe4 1 3 li:Jxe4
J.b7 ! the game was totally equal.
9
li:Jfd7
10 li:Jxd7
li:Jxd7
0-0
1 1 li:Jd2
l:tc8
12 0-0
13 e4 (D)

B
13
bS
After 1 3 . . . dxe4 14 J.xe4 b5 1 5
1Wc2 h6 1 6 l:tfd l bxc4 1 7 bxc4
White had the initiative in Ftanik
Adorjan, Szirak 1986. Obviously,
1 3 ...c5 is also insufficient: 14 exd5
exd5 1 5 dxc5 dxc4 1 6 c6! cxb3 1 7
l:tel b2 1 8 J.xb2 li:Jc5 1 9 J.a3 !
gave White a clear advantage in
Gheorghiu-Csema, West Berlin
1 986.
bxc4
14 :tel

126 USSR Ch (Moscow) 1988

In the sixth game of my first

match against Kasparov, I captured
with the other pawn, 14 ... dxc4, and
after 15 bxc4 lbb6 ( 1 5 ...bxc4 would
have led to a position from the cur
rent game against Gurevich) 1 6
cxb5 cxb5 17 l:l.c l .i.a3 1 8 l:l.c2
lba4 1 9 .i.a1 l:l.xc2 20 xc2 as
2 1 'iWd 1 White gained an apprecia
ble advantage thanks to his danger
ous passed d-pawn.
15 bxc4
dxc4
After 1 5 ....i.xc4 16 lbxc4 dxc4
1 7 a4 'iWb6 1 8 .i.fl ..,a6 1 9 a5 c5 20
d5 l:l.fd8 2 1 e5 White had more
than enough compensation for the
pawn in Dol.matov-Ehlvest, USSR
Ch (Minsk) 1987.
16 4! ?
Until now this encounter had
followed game 2 1 of my return
match against Kasparov. Then I
continued 1 6 'ifc2, and after
1 6... ..,c7 1 7 lbfl e5 ! 1 8 lbe3 exd4
1 9 .i.xd4 .i.c5 chances were even.
However, White's resources in this
variation have not yet been ex
hausted.
.i.b5
16
.i.a3
17 ..,c2
18 lbb1 (D)
White could have maintained a
minimal advantage after 1 8 lbxc4
.i.xc4 1 9 ..,a4 c5 20 'iha3 cxd4 2 1
.i.xd4.
18
.i.d6
19 a4
.i.a6

B
20 lbd2
e5?!
20 ... c5 ! is more precise: 21 d5
lbes with a double-edged game.
..,e7
21 l:l.ad1
22 .i.n
e6
.i.xeS
23 dxeS
23 ... lbxe5 24 .i.xe5 .i.xe5 25
.i.xc4 is no better.
24 f4
i.c7
lbb6 (D)
25 e5
After 25 ... .i.b6+ 26 h1 lbcs
the whole game would still lie
ahead, but now White gains a seri
ous advantage.
26 f5!
...h6
lbd5
27 lbe4!
28 .i.d2
.i.b6+
29 h1
h5
30 .i.e2
...h3
31 lbgS
...h6
32 .i.xc4
Taking on f7 would have led
White to his goal more quickly: 32
lbxf7 ! 'ifh3 33 lbd6, etc.
32 ...
.i.xc4

Karpov - M.Gurevich 127

43 ...
44 l:xe6
45 .:eel

33
34
35
36

w
'ibc4
e4
g4
w

hs
.i.fl
..h4
..d8 (D)

B
37 .:n
.i.b6
38 l:del
l:e8
39 l:e2
c7
40 .:et
7
41 c4
c5
42 g2
lbb4
43 e6!
White makes the decisive break.

fxe6
h8
..d7 (D)

w
46 aS!
.i.xa5
46 ... .i.c7 also loses; 47 .i.xb4
cxb4 48 ..xc7 !.
lbd3
4 7 xc5!
lbxel +
48 xas
.:ed8 (D)
49 l:xel

w
50 g3!
51 ..a3

7
l:d7

128

USSR Ch (Moscow) 1988

'tWbS
52 f4
'iWc6
53 l:teS
l:tb7
54 'ife3
'iWa4
ss l:te6
'ifbS
56 l:te4
'iVc6
57 eS
1Wd6+
58 d4
l:tf8?!
59 h3
An oversight caused by time
trouble in a difficult position.
1-0
60 cS
Game 3 1
Karpov - Yusupov
USSR Ch (Moscow) 1988
QGD, Exchange Variation
1 c4
e6
dS
2 lLlc3
3 d4
e7
lLlf6
4 lLlf3
exdS
5 cxdS
c6
6 gS
7 'ii'c2
Instead 7 e3 (and then 'iVc2 and
d3 in one order or another) is
most frequently played, or first 7
'ifc2 and then 8 e3. In any case, at
this point the e-pawn is not moved
more than one square.
g6
7
A year later at Rotterdam, Yusu
pov chose 7 ... lLla6 against me, but
again did not solve his opening
problems, and the game ended in
my favour.
8 e4

Transferring the game from a

closed opening to an open type of
game. Incidentally, the move e2-e4
in this situation was given new life
by Timman in his match against
Short, Belgrade 1 987.
8
lLlxe4
Short's preference was 8 ... dxe4
9 xf6 xf6 10 1Wxe4+ c;li1f8 1 1
c4 g7 1 2 0-0 l:te8 13 'iff4 e6
1 4 xe6 l:txe6 1 5 l:tfe1 l:txel + 1 6
l:txe1 lLld7 17 lbe4 e7 1 8 h4 'iVb8
19 lLle5 f6 20 lLlg5 ! lLlf8 21 hS
gxhS 22 'iff5 'ii'c8 23 lLld7 ! 'i'xd7
24 'ii'xd7 lLlxd7 25 l:txe7+ g6 26
lLlf3. White had the better end
game, and eventually came out on
top. Taking on e4 with the knight
did not seem to have been tried be
fore, but the novelty did not take
me unawares.
xe7 (D)
9 xe7!
Strangely enough, the alterna
tive 9 ... 'it'xe7 loses straight away:
10 lLlxdS ! cxd5 (if lO ...'i'e6, then 1 1
lLlc7+) 1 1 'iVxc8+ 1i'd8 12 bS+
e7 1 3 'i'xb7+ with a wipe-out.
10 lLlxe4
dxe4
1 1 'ifxe4+
e6
'iWaS+
12 c4
1 2 . ..l:te8 is probably more pre
cise, because then 1 3 lLlgS is im
possible on account of 1 3 ...'ii'a5+
and 14... 1t'xg5.
13 !
Yusupov very likely underesti
mated this retreat by the king. After

Karpov - Yusupov 129

w
1 3 lLld2 lLld7 14 0-0-0 .:l.ae8 Black
would have overcome his opening
problems.
'iWfS
13
lLld7
14 'it'e3
1 4 ...f6 loses: 1 5 d5 ..ixd5 1 6
..id3 'irg4 17 We5#. After 1 4...f8,
White has the good continuation
1 5 ..ixe6 'ilfxe6 1 6 'irh6+ g8 17
g3 lLld7 1 8 g2.
15 .:let
.:l.ae8 (D)

16 dS! !

Thanks to this mighty break in

the centre, White gains an impor
tant square on d4 and freedom on
the dark diagonals for his queen, in
return for the sacrificed pawn. I
spent almost an hour thinking
about it, but the time is worth it for
a move like that!
16 ...
cxdS
17 ..ibS!
The key to my idea. Without this
attacking move White would find
himself at a dead end. If 1 7 lLld4,
then 1 7 . . .'iWe5 ! is strong, and 1 8
1t'a3+ leads to an equal position af
ter 1 8 ...'ird6, while if 1 8 Wxe5
lLlxeS 1 9 ..ib5, then the simple
1 9 . . .lLld7 gives Black an excellent
game. Of course, before throwing
forward the d-pawn, I had to fore
see this fine bishop manoeuvre to
b5.
17 ...
a6
After 17 ...f8 there is the possibility of 1 8 'ilfc3, followed by
lLld4xe6+ and ..ixd7. Black should
have thought about returning the
pawn by means of ... d5-d4, and
transferring his queen to c5.
18 ..Wa3+
d8
Right through the game the
black king does not find a safe ha
ven, and now 1 8 ...f6 loses to 19
..ixd7 ..ixd7 20 'iWc3+.
19 'ii'aS+
But not 1 9 lLld4 because of
1 9... 'ii'f4.

JJO USSif Ch (Moscow)

1988

19
e7
The path to the queenside is
fraught with danger, e.g. 1 9 ... c8
20 l:.c l + b8 2 1 'ii'c7+ as 22
li)d4 'ii'f6 23 .txa6 ! l:.b8 24 'it'a5
'it'd8 25 l:.c7 winning.
20 'ii'b4+
6
If 20 ...d8, then 2 1 li)d4 'iVf6
22 .txa6 bxa6 looks reasonable.
Then after 23 l:.c l ! (23 li)c6+ c7
24 l:.c 1 l:.a8 25 lLla5+ d8 26 'ii'b7
e7 27 tLlc6+ d6 28 'ir'b4+ c7
only leads to a repetition of the
position) 23 ...l:.eg8 24 l:.c6 there is
no defence against the threat of 25
l:.xa6 and 26 l:.a8+:
a) 24...'it'e7 25 'iia5+.
b) 24 ... 'ii'e5 25 l:.xe6! .
2 1 'ii'd4+
A pure geometric pattern. The
white queen is moving like a pen
dulum. 2 1 .txd7 does not work in
view of 21 ...'ir'd3+.
21
e7
'ii' bS
22 .td3!
22 ...'ir'f6 cuts off the path of the
king's retreat after 23 'iVb4+ [edi
tor's note: for example 23 .. . d8
24 'ii'xb7 'ifxf3 !? 25 'ir'a8+ (25
gxf3?? .th3+ 26 g1 l:.xe 1 + 27
.tfl l:.xfl#) 25 ... e7 26 'ifxe8+
l:.xe8 27 gxf3].
23 h4! (D)
It is almost as if White is con
structing a study on the theme of
domination.
23
d8

..

B
24 tLlgS
Threatening 25 g4 ..Wh6 26 l:.xe6
and 27 lLlxf7+.
24
l:.hf8
..Wh6
25 .te2!
26 .tf3
l:.e7
Retreating the king does not
help: 26 ...c7 27 ..Wf4+ and 28
tLlxe6; 26 ... c8 27 l:.c 1 + d8 28
.txd5 .txd5 29 ..Wxd5.
lLlf6
27 'ii'b4
28 'ii'd6+
28 'ii'f4! would have been imme
diately decisive, as it attacks the
knight and simultaneously threat
ens 29 tLlxe6+ and 29 Wfb8+.
l:.d7
28
29 'ii'f4
tLlg8 (D)
If 29 ... 'ir'g7, then 30 l:.xe6!.
c8
3 0 .tg4!
Taking on g4 leads to an elegant
mate: 30 ....txg4 3 1 'ifb8#.
31 .txe6
fxe6
32 l:.cl+
d8
33 tLlxe6+
e7

Karpov - Malaniuk 131

3
4
5
6
7

34
35
36
37
38
39
40
41
42
43

w
'ii'xf8+
.!Dxf8
.Uh3
h5
h6+
llf3+
.Ue1+
.Uf6+
g4
.Ue8
1-0

'ii'xf8
xf8
.!De7
g7
6
e6
d6
c7
.!Dc6
d4

Game 32
Karpov - Malaniuk
USSR Ch (Moscow) 1988
Dutch Defence
Malaniuk is one of the outstanding
experts in this opening, and there
fore I prepared particularly thor
oughly for our game in this event.
As a result I managed to come up
with an interesting novelty.
1 d4
f5
2 g3
.!Df6

i.g2
c4
.!Df3
0-0
.!Dc3

g6
i.g7
d6
0-0
'ife8 (D)

w
One of the most topical posi
tions in the Leningrad Variation of
the Dutch Defence. Black is in
tending ...e5, creating a flexible
pawn centre. e8 is the most suitable
square for the queen, as from there
she is influencing the centre, and
might come in useful on the king
side.
8 b3!
The appearance of the bishop on
a3 could essentially hinder Black's
plans. Other lines are less danger
ous for him.
8 ...
.!Da6
An immediate march by the e
pawn, 8 ... e5, has not yet been fully
investigated, but it is evidently not
without danger for Black.

132 USSR Ch (Moscow) 1988

9 .i.a3!
White's idea consists of actively
positioning his pieces, with his
bishop on a3, his queen on d3, and
rooks on d 1 and e 1 , and then mov
ing the pawn to e4. In this game I
managed to carry out this plan in
ideal fashion. Black is forced to set
aside his standard advance ... e5 in
definitely.
9
c6
10 'iVd3
It is interesting that in this na
tional championship Malaniuk
chose the Dutch Defence seven
times ! Two rounds before our game
Gavrikov had played the more
timid 10 llc 1 against him. After
10 . . .h6 1 1 e3 e6?! 1 2 We2 'i!fd7
1 3 lbd2 lbc7 14 'ji'd3 llab8 15 b2
h8 1 6 d5 cxd5 17 lbxd5 lbcxd5
1 8 cxd5 f7 Black built a fortress
which was not easy to penetrate,
and in the end he even won. How
ever, this time matters did not turn
out so successfully for him.
10 ...
d7?!
This bishop move has very un
pleasant consequences. 10 ...llb8
deserves attention, as a series of
games has testified. The fate of the
whole variation possibly depends
on the assessment of this position.
11 llfe1
l:.d8
1 l . . .d5 is more stubborn, al
though after 1 2 lbe5 White main
tains the initiative.

h8
12 llad1
13 e4
Having finished mobilizing his
forces, White generates activity in
the centre of the board.
13
fxe4
14 lbxe4
fS (D)

..

The extremely unpleasant 2 1

Wa5 was threatened.
21 lbxe6
xe6
d7 (D)
22 l:.de1
He should have brought his
bishop back to c8. There now fol
lows a decisive exchange sacrifice,
which, incidentally, has been ma
turing for a long time.
xe7
23 l:.xe7!

Karpov - Hjartarson 133

dxe5
29 l:l.xe5
30 'iVxeS
rM7
31 d6
j_f5
32 c5
Black is in total zugzwang.
h5
32
33 g4
hxg4
34 hxg4
j_d3
If 34 ...j_xg4, then 35 'iVf6+ 'it>e8
36 'ii'xg6+ and 37 'ifxg4 is decisive.
1-0
35 j_d5+!
An effective final blow ended
the game; taking the bishop is im
possible in view of 35 . . . cxd5 36
'iVxd5+ 'it>e8 37 'it'e6+ mating.
Game 33
Karpov - Hjartarson
Tilburg 1988
QGD, Slav Defence

In this tournament there was a rare

occurrence, at least in my experi
ence - I played three straight
games in the same variation, and
moreover it was not in a classical
variation, but a very sharp one, in
which the balance of forces on the
board is disturbed in the first ten
moves. My results were quite fa
vourable, 211213 . Detailed analysis
of these rather complex duels only
took place much later.
d5
1 d4
c6
2 c4
lt:lf6
3 lt:lf3

/34 Tilburg 1988

dxc4
4 tt'lc3
5 a4
i..f5
6 lDe5
This is the so-called Central
Variation (White is preparing f3
and e4).
6
e6
This move has recently com
pletely replaced two other theoreti
cal continuations, 6 ... lDa6, and
6... lDbd7.
7 f3
Logical; after 7 e3 or 7 i..g5 it is
easier for Black to equalize.
7
i..b4
8 e4
Leading to immense complica
tions, as Black is forced to sacrifice
a piece. As a rule after the cautious
8 tt'lxc4 or 8 i..g 5, positions arise
with even chances.
i..xe4
8 ...
tt'lxe4
9 fxe4
Black has three pawns for the
piece, so there is approximate
material equality on the board.
'ii'xd4
10 i..d2
10 .. .'ii'h4+ does not work: 1 1 g3
lDxg3 12 hxg3 'ifxh1 13 'ii'g4 with
a strong attack for White.
1 1 tt'lxe4
'i!Vxe4+
1 1 ...i..xd2+ is no good: 12 lDxd2!
'ifxe5+ 1 3 i..e2 b5 14 axb5 cxb5 1 5
l:la5 a6 1 6 tt'lxc4 'ii'c7 17 lDd6+
<1iie7 1 8 0-0 gave White an irresist
ible attack in Khalifman-Ehlvest,
USSR Ch (Moscow) 1 988.

12 'ii'e2
i..xd2+
Here in the same tournament
Timman introduced the novelty
1 2 . . .'ifh4+ against me, but the in
tervening check did not achieve its
aim: 1 3 g3 i..xd2+ 1 4 'itxd2 'fle7
1 5 'ii'e 3 ! with a patent advantage.
13 'itxd2
t!fd5+
In the ending after 1 3 ... 'flxe2+
14 i..xe2 the black pawns do not
compensate for the missing piece.
It is another matter if the queen ex
change takes place on d5, because
then Black strengthens his pawn
chain.
lDa6
14 'itc2
15 tt'lxc4 (D)

,
.

''

. , . , . .
. . .. . .
ttJ

ft
u
Rc;t>
-
u

a i. :

B
A peculiar, though standard pat
tern from the Slav. Black has quite
a wide choice here, but recently
castling one side or the other has
been seen most frequently. Theory
is sceptical about 1 5 ...l:r.d8, 1 5 ...b5
and 1 5 ...lDb4+; 1 5 ...'iff5+ deprives

Karpov - Hjartarson 135

Black of the possibility of swap

ping queens on d5, while exchang
ing on other squares leads to an
unpleasant endgame.
15
0-0-0
With castling long, as opposed
to short, the rook immediately
lands on d8.
16 'ii'e5
White is aiming at the g7 -square
and is prepared to swap queens,
just not on d5.
16
f6
Defending the pawn and forcing
the queen to clarify her intentions.
Neither 1 6 ... lbb4+ nor 1 6 . ..:he8
has vindicated itself in practice.
17 'ii'eJ!
c5
Black has also tried 17 ...'ii'f5+,
1 7 . . . lbc5 and 1 7 .. .'b8, but with
out any particular success.
18 b3!
In my game against HUhner at
Tilburg I preferred 1 8 .te2, and af
ter the continuation 1 8 ...lbb4+ 1 9
b3 lbc6 2 0 c3 lbd4 21 .tf3
lbxf3 22 gxf3 'ii'd4+ 23 'ii'xd4
:xd4 24 b4 :ds 25 bxc5 :xc5 a
level endgame arose. This unex
pected raid by the king looks rather
strange, but it saves White an im
portant tempo by immediately
freeing the c-file for the rook.
18
lbb4
lbc6
19 :et
lbd4
20 a3
21 lba5!
.

More accurate than 21 :g1 e5 22

.td3 b5 ! 23 axb5 lbxb5+ 24 a4,
when the game is roughly even.
21
eS (D)
2 l . ..b8 is no good because of
22 1i'g3+.

w
b6
22 'ii'c3!
23 lbbJ
'W'xb3+
23 . ..'W'e6 is more obstinate: 24
.tc4 'ii'e7 25 lbxd4 :xd4 26 :hd1
:hd8, and Black holds on.
lbxb3
24 'W'xb3
25 b3
:d4
26 h4
Black has material equality and
a flexible pawn chain, but I quickly
managed to disturb its harmony.
26
:hd8
c7
27 .tc4
28 h5
:g4 (D)
Instead of pursuing the g-pawn,
Black ought to generate some ac
tivity on the queenside: 28 ... a6! 29
.txa6 (Black was threatening ... b5)

J 36

Seanle Ct (3) 1989

29 . . ..:tb4+ 30 c3 (in the event of

30 a3 White could even lose 30 . . . .:ta8 !) 30 ....:txa4 with chances
for a draw.

32
33
34
35
36
37
38
39
40
41
42
43

.Uhxf6
li6f4
.Uf7+
lixg7
lif4
c2
c3
d2
.Uf7+
e3
xa6

e4
h5
lid4
.Ud7
lixg7
lig3+
lig2+
.Ug3+
lig4
d6
a6
1-0

Game 34
Hjartarson - Karpov
Seattle Ct (3) 1989
Spanish, Zaitsev
29 h6!
By stealing along the edge of the
board, this pawn undermines the
black kingside.
29
lbg2
Attempting to preserve a dy
namic pawn chain by means of
29 . . . g6 is unsuccessful on account
of the weakness on h7: 30 .:thd 1 !
.Ugd4 (30 ....:txd 1 3 1 .Uxd 1 .:th4 32
b5 .Uxh6 33 l:td7+ b8 34 a6
with decisive threats) 3 1 lixd4
lixd4 32 g8 ! lid? 33 c2 and
Black will be put in zugzwang.
l:txg7
30 hxg7
.Ud6
31 lien
32 lih6
Black's fortress has been de
stroyed, and this sums up the result
of the opening.

e5
1 e4
llJc6
2 lLlf3
a6
3 b5
lLlf6
4 a4
e7
5 0-0
b5
6 :et
d6
7 b3
0-0
8 c3
lieS
9 h3
b7
10 d4
f8
1 1 lLlbd2
12 a3
As well as 1 2 a4, this restrained
pawn advance is also seen frequently. White guards b4 from invasion by the black knight, and
then prepares to develop the initiative on the queenside.
h6
12 ...

Hjartarson - Karpov 137

liJb8
13 ..tc2
14 b4
liJbd7
15 ..tb2
aS! ? (D)
A move I had prepared specifi
cally for this match; 1 5 ...g6 is a
more widely used continuation.

...
. jilli
...
- ... d

...

.t.

-
.
-
-

.
.

d
- -

-/
fj d

1":\ . fj
u .
.t.z.J.
-

--'<% - 0A
... /
R

.
m
.

'. .

1'1

w
c6
16 ..td3
17 liJb3
axb4
18 cxb4
1 8 axb4 liJb6 1 9 liJa5 "ilc7
would have led to equality.
exd4
18
19 liJfxd4 (D)
Now Black seizes the initiative.
Things are also fine for him after
1 9 ..txd4 c5 20 ..txf6 liJxf6 2 1
..txb5 l:be4. The correct continu
ation is 1 9 liJbxd4 c5 20 bxc5 dxc5
2 1 liJxb5 ltJxe4 22 liJe5 liJxe5 23
..txe4 with a roughly even game.
c5!
19
A very promising pawn sacri
fice, with which Black takes aim at
the enemy centre.
.

B
20 bxc5
20 lbxb5 immediately is much
more precise: 20 ...cxb4 21 axb4
l:lxal 22 ..txal d5 23 exd5 (23 e5
..txb4 leads to an approximately
equal position) 23 ...l:lxel+ 24 1Wxel
liJxd5 25 'lie4 liJ7f6 26 ..txf6
liJxf6 27 "ilxb7 "ilxd3 with a sharp
game, in which Black has enough
compensation for the pawn.
20
dxc5
21 liJxb5
ltJxe4!
22 'lic2?!
This is a serious error. White had
to play 22 "ilf3, when the following
variations are possible:
a) 22 ...liJd6? fails to 23 l:lxe8
'tWxe8 24 liJxd6 ..txd6 25 "ilxb7
l:lb8 26 '1id5.
b) 22 ...'lib6 23 ..txe4 ..txe4 24
l:lxe4 'lixb5 25 a4 is unclear.
c) 22...l:lb8 should be answered
by 23 ..tc4, with a complicated po
sition. Instead after 23 ..txe4?!
..txe4 24 l:lxe4 l:lxe4 25 'lixe4

138 Slcelleftea World Cup 1989

.:txb5 26 'ili'd5 'ilt'b8 27 'i't'xd7 .:txb3

28 cl .:tb1 29 .:.Xbl 'i't'xb1 30 'ilt'd2
c4 3 1 h2 1t'b3 Black is slightly
better.
d) 22 ... df6 !? 23 xf6 xf6
24 .:txe8 xf3 25 .:txd8 .:txd8 re
sults in an equal position.
22
df6
23 c3 (D)
White must certainly avoid the
continuation 23 .:tad 1 'ilt'b6 24 tDc3
xf2! .

Black threatens the deadly

...tZ::lf3+ or ...tZ::lxh3+.
27 .:te3
d6
28 h4
tZ::le6
29 li)d1?
White would fare no better with
29 .:txe6? fxe6 30 'it'g6 'it'e8 . His
best chance was 29 li)b5 !? f4 30
.:th3 li)e4 when Black would still
have some obstacles to overcome.
After his actual choice I was able to
launch a winning attack.
tZ::lg4
29
30 .:txe6
h2+!
31 h1
1t'xe6
32 f3
ile1 ! !
0-1
White's position collapses: 33
fxg4 'it'xh4 wins on the spot, while
33 'ii'c4 gives Black a choice be
tween 33 ...f4 and 33 ....:ta4.
Game 35
Karpov P. Nikolic
Skelleftea World Cup 1989
Bogo-Indian Defence
-

B
g5!
23
24 bS
More resistance was possible by
24 tZ::ld 2, although after 24 ...'ilt'b6
Black threatens 25 ...'ilt'c6.
llxe1+
24
ikc7
25 .:txe1
26 n
Now, however, 26 tZ::ld2 can be
met by 26 ...1t'f4!. Still, 26 .:te3 was
worth considering.
26
'ii'c6!

tZ::lf6
1 d4
2 c4
e6
3 tZ::lr3
b4+
4 d2
A more popular continuation
than 4 li)bd2.
ile7
4
Black has several alternatives,
including exchanging bishops on
d2.

Karpov

0-0
s g3
i.xd2+
6 .tg2
7 Wxd2
d6
e5
8 ltlc3
l:te8
9 0-0
.tg4
10 e4
.txf3
11 dS
12 .txf3
cS
13 l:tael
a6
ltlbd7
14 b3
15 .tg2
l:.ab8
16 a4 (D)
By freezing play on the queen
side, I am decisively moving the
battle's centre of gravity to the
kingside. Now if 1 6 ... a5, then 1 7
ltlb5 b6 1 8 ltla7 and ltlc6, followed
by f4.

'ird8
16
Preventing further movement by
the white pawn. Otherwise after a5
and ltla4, I would have advanced
the pawn to b4 and put pressure on
the weak b7-pawn.

P. Nikolic 139

17 'ird1
First and foremost in order to
control the h5- and g4-squares.
After the immediate 17 f4 Black
could have chosen the exchange
1 7 . ..exf4 in order to gain counter
play by means of ...ltlg4 (or ..ltlhs)
and ...'ifh4.
17
l:te7
Was
18 l:te3
19 l:tfe1
White has reinforced the e4pawn, and it remains for him to
transfer the bishop to h3 and ex
change it for the d7-knight, which
would give him a significant spa
tial advantage.
19 ...
h8
l:tg8
20 .th3
Nikolic prevents the advance 21
f4, after which the game would
now continue 2 l ...exf4 22 gxf4 g5.
In that case White would not have
enough time to rush the pawn
through to e5, and Black would
thus gain a comfortable post for his
knight on that square.
Wd8
21 h1
ltlhS
22 'ird2
g6
23 ltle2
ltlg7
24 aS
White needs time to regroup his
forces on the queenside, but Black
has apparently prepared for activ
ity on the kingside. However...
25 .txd7
Wxd7 (D)
l:tge8
26 f4!
.

140 Skelleftea World Cup 1989

.
-

.
-

-
--

. . .... . . . .

' 8
d
>

!'!

88

-
. A . Ml

pm o.z.J
U

B B

do

"

. !'!

31 lbxe4
ikf5 (D)
3 l . . .'iWg4 32 c5 .:txe4 33 .:txe4
'ii'f3+ 34 gl 'iWg4+ 35 f2 'ii'h4+
36 <ite2, etc., is worse, but 3 l ...lbh5 !
is interesting, and in the event of
32 c5 c;l;>gS 33 cxd6 .:txe4 34 .:txe4
lLlg3+! 35 hxg3 'iWh3+ the compli
cations end with a perpetual check.

w
27 b4!
cxb4
In the event of 27 ...exf4 28
lbxf4 the essential superiority of
the white knight over his opposite
number becomes clear: by landing
on d3 it supports the attack in the
centre.
f5
28 'ii'xb4
Black is prepared to exchange
the e- and f-pawns, then move the
knight over to f5, achieving a posi
tion of dynamic equality. There
fore I have to play very carefully,
but also quite decisively.
29 lbc3
The immediate 29 c5 ! deserved
attention: 29 ... dxc5 30 'ii'xc5 exf4
3 1 lbxf4 fxe4 32 l:Xe4 l:be4 33
.:txe4 .:tc8 (33 ....:txe4? 34 'ii'f8#)
34 .:te7 'ii'g4 35 'ii'f2 , and White's
chances are obviously better. De
laying the break c4-c5 for three
moves complicates the situation.
29
exf4
fxe4
30 gxf4

w
It looks as though everything is
in order for Black now. 32 lbxd6
does not work: 32 .. .l:be3 33 l:lxe3
ifxf4 threatens mate on fl , which
would cost White his knight if he
plays 34 .:txe8+ lLlxe8. After 32
lLlg3 Black should again continue
32 ....:txe3 33 .:txe3 'ili'xf4 34 .:txe8+
lLlxe8 and White has nothing better
than a draw: 35 'iWxb7 'tie l + 36
c;li>g2 1Wxc4.
But not all White's resources
have been exhausted...
32 c5!
The queen joins in to defend the
knight, and White makes a passed

Karpov - P. Nikolic 141

pawn. Events on the board take on

quite a stormy character.
dxcS
32
33 'it'c4
'ifxf4
:Z.eS? (D)
34 d6
Black is still under the impres
sion that his doubled rooks are
strong. 34 ... :Z.d7 would have been
more forceful, preventing the pawn
from advancing. After 35 'ii'c3 'ii'e5
Black holds on, while 35 'ii'xc5 is
not dangerous for him either:
35 . . .'ii'e5 36 'ii'xe5 :Z.xe5 37 :Z.c l
lbe6 38 :Z.c8+ g7 39 :Z.e8 f7 40
liJg5+ [editor's note: White may
try 40 :Z.xe6 xe6 4 1 liJc5+ xd6
42 :Z.xe5 :Z.c7 43 liJd3] 40 ...:Z.xg5
4 1 :Z.3xe6 :Z.d5 42 :Z.8e7+ :Z.xe7 43
:Z.xe7+ f6 44 :Z.xh7 :Z.xd6 45
:Z.xb7 with a draw.

Allowing White to create a

small study. After 36 ...h6 the posi
tion is still unclear, for example 37
:Z.f3 'ii'g4 38 :Z.f7 ...e6 39 :Z.efl
liJfS ! .
'it'h4
37 :Z.f3
38 :Z.f7
Not, of course, 38 xe5? due to
38 ......xel + 39 g2 ,..e2+.
38
Ilg8
39 'it'cl
39 :Z.xg7 :Z.xg7 40 d81i'+ xd8
41 'ii'xe5 is also good, and after
4 l . ..'i'e7 42 liJxc5 xe5 43 :Z.xe5
White achieves a technically win
ning endgame. But I managed to
find a more elegant line.
39
gS
39 . ..liJf5 does not work because
of 40 :Z.xf5, therefore Black is
forced to advance yet another
pawn from his king's shelter.
:Z.dS
40 liJg3
41 :Z.f6
:Z.d8 (D)
The game ends with immediate
mate after 4 1 . .. h7 42 ...bl+.
42 'iib 1 !
An elegant finale.
g8
42
It looks as if Black has extri
cated himself, and with his next
move he will seize the daring run
away on d7, but, as it perishes, the
pawn has the last word.
1-0
43 :Z.e8+!
Let us see the end of the study:
43 ...:Z.xe8 44 dxe8...+ liJxe8 45

w
35 d7!
:Z.d8
If 35 ...:Z.f8 36 'ii'c3 ! :Z.d5, then
37 liJxc5 is decisive.
hS?!
36 'iVcJ!

142 London Ct (8) 1989

Game 36
Karpov - Yusupov
London Ct (8) 1989
QGD, Lasker
This decisive duel was very strik
ing.
li)f6
1 d4
2 c4
e6
d5
3 lf)fJ
i.e7
4 li)c3
0-0
5 i.g5
h6
6 e3
li)e4
7 i.h4
Not waiting for the white bishop
to take on f6. This move constitutes
the Lasker Defence, where Black
in many respects has no preten
sions, but relies on creating an im
penetrable fortress. In games 4 and
6 Yusupov had already turned to
this defence, and in fact I had not

managed to glean the least advan

tage out of the opening.
8 i.xe7
'iVxe7
9 llcl
In game 4 I had continued 9
it'c2, and in game 6, 9 cxd5, and in
both cases Black quickly equal
ized. This time I was better pre
pared for the game.
9
c6
10 i.d3
li)xc3
dxc4
1 1 llxc3
12 i.xc4
Recapturing on c4 with the rook
also deserves attention.
12
li)d7
eS
13 0-0
Neither 1 3 ... b6 nor l 3 ...lld8 is
very promising for Black. It must
be said that from the Lasker De
fence, the game has transposed to
lines of the less weighty Capa
blanca System, and moreover the
position of the pawn on h6 (instead
of h7) is favourable for White, as it
gives him an additional object for
attack.
14 i.b3 (D)
This waiting move with the
bishop was prepared specially for
this game. Here White would nor
mally play 14 h3, 14 jfc2, 14 'iib l ,
or 1 4 dxe5 li)xe5 1 5 li)xe5 it'xe5
1 6 f4.
exd4
14
lDf6
15 exd4
'iVd6
16 lle1
.

Karpov - Yusupov 143

.t. -
B mu m
-
. .
.
- m

. . .
R
U"'l.J.
d
8

B 8_rJ
.\lllr
..

- .
-
B
lDdS

17 ltleS!
18 l:.g3
White does not hide his inten
tions.
.i.fS
18
This move is a novelty. Theory
gives the continuation 1 8 ....i.e6 19
'iid2 h8 20 l:.e4 and 21 l:.h4,
which is somewhat preferable for
White.
19 'tWhS
.i.h7
The paradoxical move 19 ......e6!?,
suggested by Taimanov, is interest
ing.
20 'tWg4!
gS (D)
After 20 . . . g6 there is the very
strong reply 21 h4. 20 . ..'it'f6 does
not work because to 21 ltld7. After
20 ....i.g6, the following variation
is possible: 2 1 ltlxg6 'tWxg6 22
'iixg6 fxg6 23 l:.xg6 l:.ae8 24 l:.xe8
l:.xe8 25 1 l:.e4 26 l:.d6 l:.xd4 27
l:.d8+ with a big advantage to
White.
(6
21 h4

w
2 1 ... f5 is no good because of 22
'6'f4 ! (or 22 'ii' h5 g4 23 ltlxg4); in
the event of 21 ... ltlf6 White has the
strong 22 'itf3, or 22 '6'f4! ltlh5 23
'ii'f3 ltlxg3 24 ltlxf7 'iixd4 25 l:.e7,
etc.
22 hxgS!
22 Wh5 is insufficient: 22... fxe5
23 hxg5 .i.g6 24 gxh6 h7 and the
black king is safely covered.
hxgS
22
In the event of 22 ... fxg5 the re
ply 23 f4 simply becomes more
valid.
23 f4
After the continuation 23 l:.h3
fxe5 24 l:.xh7, White would have
won beautifully in the event of
24...xh7: 25 .i.c2+ g7 (25 ...h6
26 'iWh3+ g7 27 'iWh7+ f6 28
dxe5+ Wxe5 29 'iWh6+) 26 'fi'xg5+
cltf7 27 dxe5 We6 (27 ...'6'b4 28
e6+ e8 29 .i.g6+; 27 .. .'ile7 28
'iWg6#; 27 ....:tae8 28 'iWh5+) 28
.i.f5 . However, Black can hold on

144 London Ct (8) 1989

thanks to 24 ... 'iff6. Besides mobi

lizing the f-pawn, White would
have gained a dangerous initiative
after either 23 'iih5 or 23 ltlf3.
23
.:tae8 (D)
The critical moment. This rook
manoeuvre forces Black's demise.
He should have sheltered the king
in the corner with 23 ... h8, al
though in this case as well I could
have reached a winning ending
with four rooks with 24 fxg5 fxe5
25 g6 'iixg6 26 'ii'xg6 i.xg6 27
.:txg6 exd4 28 .:te4 .:tf7 29 i.xd5
cxd5 30 .:th4+ .:th7 3 1 l:xd4 l:d8
32 .:tg5 .:thd7 33 a3 a6 34 a4 b5 35
axb5 axb5 36 b4 - Black is in zug
zwang.
..

queen sacrifice - 25 gxf6+! ! i.xg4

26 .:Xg4+, and then:
a) 26 ... h8 27 ltlf7+ .:txf7 (if
27 ...h7, 28 .:tg7#) 28 .:Xe8+ .:tf8
(28 ... h7 29 .tc2+ h6 30 .:th8+
and 3 1 .:txh7#) 29 f7 ! ltlf6 (or
29 . . .h7 30 i.c2+ h6 3 1 .:tg6+
'iixg6 32 i.xg6, etc.) 30 .:txf8+
'ili'xf8 3 1 .:tg8+ ltlxg8 32 fxg8'ii+
'ifxg8 33 i.xg8 xg8 34 f2 and
it's all over.
b) 26 . .. h7 27 i.c2+ h6 (al
ternatively 27 ...h8 28 .:th4+ g8
29 .th7+ h8 30 ltlg6#) 28 .:tg6+
h7 (28 . . . h5 29 i.d 1 + h4 30
l:e4+ mates) 29 f7 and White wins.
i.xg6
25 g6!
26 dxeS!
More accurate than 26 .:txe5
h7 27 'iWh5+ g7 (27 ... i.xh5 28
.:txh5+ 'ii'h6 29 i.c2+ h8 30
l:xh6#) 28 .:txg6+ 'ifxg6 29 l:g5
.:te 1 + 30 h2 .:te6 with an unclear
game.
'ili'e6
26
Things are no better for Black
after the continuation 26 ...'iic5+
27 h2! (but not 27 h1 .:tf6 !), or
26 . .. g7 27 exd6 .:txe 1 + 28 h2
.:th8+ 29 .:th3.
cxdS
27 i.xdS
28 'iWxg6+
'iixg6
29 .:txg6+
h7 (D)
If 29 ...t7, then 30 .:td6 .:td8 3 1
.:r.n + e7 32 .:txf8 .:txf8 3 3 .:txd5
winning.
.:cs
30 .:td6

Karpov - Timman 145

.
.
.
- ...
.

. . :
. ' .


80 .8

.

. .
.
w

After 30....:d8 White has the decisive 3 1 .:d 1 .

l:.c2
31 .:e3
g6
32 .:d7+
33 .:xb7
.:e8
The rook ending is hopeless for
Black, but resigning the game was
tantamount to resigning the match
too, and so Yusupov played on,
hoping for a miracle. If 33 ...1:.ff2,
then 34 .:g3+ h6 35 .:bs ..th7 36
.:b4 is decisive.
d4
34 a3
.:xe5
35 .:d3
.:g5 (D)
36 .:xd4
If 36 ....:ee2, then 37 .:g4+ f5
38 .:gg7.
37 .:d6+
h5
g4
38 .:h7+
39 .:d4+
5
g6
40 .:d5+
41 .:g7+
xg7
f6
42 .:Xg5+
43 .:b5
a6
44 .:b6+
e7

Game 37
Karpov - Timman
Kuala Lumpur Ct (4) 1990
Griinfeld Defence
1
2
3
4
s
6
7
8
9
10

d4
c4
lLlf3
g3
j.gl
cxd5
lLlc3
lLleS
0-0
f4

lLlf6

g6

j.g7
c6
d5
cxd5
0-0
e6
lLlfd7
lLlc6

146 Kuala Lumpur Ct (4) 1990

1 1 .i..e3
f6
ltlb6
12 ltld3
'ile7
13 b3
14 a4
.i..d7
.:tfd8
15 .i..c t
16 e3
.i..e8
'ilf7
17 .i..a 3
18 :et
.i..rs
19 .ixrs
Wxrs (D)
As a result of complex manoeu
vres, an advantage has crystallized
for White on the queenside.

I . _ ... . . .
- .

"

Q
- U

r.... rt

"l.JU
i. "

m
m

- ' \lllr

-
"

"

'ile7
20 g4
.:tac8
21 'ild2
l:tc7
22 ltle2
ltlc8
23 l:tc5
24 f5
g5
White is very slightly better af
ter the continuation 24.. . ltld6 25
lLldf4 gxf5 26 gxf5 ltlxf5 27 ltlxd5
exd5 28 l:txf5.
25 ltlg3
25 fxe6 'i!he6 26 ltlc3 ltl8e7 27
e4 is not so clear.

After the text move the threat of

the knight appearing on f5 is more
dangerous for Black: 25 ... .if7 26
fxe6 'ilxe6 27 lLlfS. To defend
against this, Timman sacrifices a
pawn.
25
e5 (D)
.

- -*-
.

- - . .

.
-
M
r...
.

- r/
z.w"l.JU
.
wa.
-
i.
-
- -:=

w
26 Wet
Obviously, the continuation 26
.i..xd5+ h8 27 l:tfcl exd4 28 e4 !
is stronger, returning the pawn;
the inevitable break e4-e5 would
have guaranteed White a serious
advantage. Killing two birds with
one stone (attempting to control
the c- and f-files simultaneously)
does not work, and Black, by liqui
dating the danger in the centre,
achieves a playable game, which
for a long time will be typified by
manoeuvring. However, thanks to
such tardy action the game quickly
moved into an ending filled with
fascinating events.

Karpov

b6
26
e4
27 l:.c2
liJd6
28 liJf2
l:.dc8
29 'ii'd2
aS
30 l:Ucl
31 n
If it did not deprive the white
queen of mobility, this move would
have been very perilous for Black.
Attempts to rebuild with 3 1 'ii'c3
are repulsed by means of 3 1 . .. d7
and 32 ...l2Je8.
31
li)b4 (D)
3 l . . .li)xd4 !? suggested itself,
but after 32 exd4 e3 33 'ii'd3 exf2+
34 xf2 l:.xc2+ 35 l:.xc2 l2Je4+ 36
f3 a draw is not far away (swap
ping the good c6-knight for the bad
one on f2 is not favourable for
Black). Timman's position in the
match obliged him to maintain the
pressure on the board in the hope
of seizing the initiative from me.
.

33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
so

Timman 147

liJdl
l:.xc6
l:.xc6
l2Jc3
f2
el
dl
el
l2Ja2
'ifxa2
f2
'W'b2
e2
el
d2
'ii'c l
el

l:.c6
l:.xc6
'ifxc6
f8
e7
f8
'ii'c8
g7
l2Jxa2
'ii'c7
f8
e7
d8
c8
b7
'W'e7
d7
l2Je8 (D)

i.- .
.
.
- .

-
-88.
Ri8

u
.8.

d
d
u

d .li.m
d
u

.
.

.
.

.
.

w
51 'W'hl !
I did not want to agree to a draw,
so I decided to bring about some
confusing complications, by leav
ing the queenside to its fate, while
actively occupying the kingside.

148 Kuala Lumpur Ct (4) 1990

Calculating variations to the very

end was obviously out of the ques
tion.
'Wib4
51
'ii'xb3
52 h4!
53 hxg5
fxg5
54 'Wxh7
'ii'xa4
55 'Wie7
'ii'c6
56 'ii'xg5
a4
57 'ii'e7
'ii'd6
lbxd6
58 'Wixd6
i.b5?!
59 i.dl
It was barely worth obstructing
the path of the pawn. After 59 ...b5
the game would most likely have
ended in a draw.
a3
60 lbe2
61 lbcl (D)

- - .
.

'
-*-
.
- "
u
-

.
"
.
- ;Q/

.
.

.
.

.
"
i.

B
Here Timman sealed a move he
had been thinking about for almost
half an hour. Black's activity on the
queenside puzzled many commen
tators, and the following day there
were even headlines in the local

papers such as 'Will Karpov Hold

Out?' . In fact, my preliminary
analysis showed that only White
has chances to win, and his pawns
are more dangerous than his oppo
nent's.
61
<j;c7
A good sealed move, but, of
course, we had also analysed oth
ers. In the end it was clear that
6 1 . .. i.d7, 6 l . ..i.e8 and 6 l .. .'.tc6
would have left Black with less
chances to save himself.
lbc4
62 'iti>g3
62...<iti>d7? loses: 63 <ifi>f4 i.c4 64
g5 a2 65 lbxa2 i.xa2 66 g6 ri;e7
(66 ...lbe8 67 i.a4+) 67 g7 <j;f7 68
f6 i.c4 69 i.g4 <itg8 70 i.e6+ lbf7
7 1 <j;f5 i.b5 72 ..ti>g6 i.e8 73 i.xd5
b5 74 i.a2 b4 75 i.b3 .
63 i.e2
But not 63 <j;f4? lbb2 ! with the
threat of ...lbd3+.
i.e8!
63
This is the strongest defence.
63 ...lbxe3 is insufficient: 64 i.xb5
lbc2 65 g5 lbxd4 66 f6 <j;d6 67 g6
<j;e6 68 g7 ..ti>f7 69 i.e8+ <j;g8 70
..ti>f4 winning.
64 ..ti>f4 (D)
Here is a striking variation after
64 g5, ending in a problem-like
mate: 64 ...lbxe3 65 <itf4 lbc2 66
<j;e5 lbb4 67 <ifi>f6 <iti>d6 68 g6 lbc6
69 i.b5 e3 70 g7 l1Je7 7 1 i.xe8 ! e2
72 lbxe2 a2 73 lbc3 ! lbg8+ 74 <j;f7
lbh6+ 75 <ifi>f8 a1 'iW 76 lbb5#.
..

Karpov

It would not have been too bad

ending like this, but unfortunately,
Black has a defence: 64 ...d6 65
.i.xc4 dxc4 66 ltJa2 .i.a4 67 ltJc3
.i.b3 68 tiJb5+ e7 69 ltJxa3 c3 70
f4 b5 ! 7 1 xe4 b4 72 d3 bxa3
73 xc3 .i.d5 drawing.
tiJb2
64

Timman 149

67 ltJal
b5
b4 (D)
68 .i.dl
Bringing the king slightly closer
to the pawns does not work:
68 .. .<iPd6? 69 .i.b3 tiJc l 70 cJ;h6 b4
71 f6 e6 72 Wg7 liJxb3 73 liJxb3
.i.a4 74 ltJc5+! cJ;d6 75 f7 al 'iW 76
f8'iW+ cJ;c6 77 'il'c8+ cJ;d6 78 'il'd8+
Wc6 79 'i!i'd7+ 'it>b6 80 'ii'b7+ cJ;a5
8 1 'iWa6#.

. .... . .

/'

...

o
.... . . .
.JL.
" -

B
65 g5
After 65 ltJa2, the long variation
65 . . . d6 66 g5 e7 67 h6
f6 68 tiJb4 .i.f7 69 .i.fl ! tiJd I 70
g5+ e7 7 1 g7 ltJxe3 72 f6+
e6 73 .i.h3+ tiJf5+ 74 .i.xf5+
xf5 75 rJifxf7 e3 76 'iii>g7 ! e2 77
ltJc2 a2 78 f7 e l 'ii' 79 liJxel al 'i!l
80 f8._,+ xg5 8 1 'iWf2 promises
White good prospects, but after the
correct 65 ... b5 ! nothing appropri
ate can be found.
tiJd3
65
It is too late for 65 ...ltJc4 66 f6
ltJxe3 67 cJ;e7.
66 tiJb3!?
a2

w
tiJcl
69 .i.b3
After 69 ...'it>d6 70 cJ;f6 Black's
king is kept away and (he white g
pawn is irrepressible.
'iii>d6
70 .i.xd5
.i.b5
71 .i.c4
There is also the interesting pos
sibility of 7 l ...b3 72 .i.xb3 liJxb3
73 liJxb3 'it>d5 74 tiJal 'it>c4 75 cJ;f6
'it>c3 76 cJ;e7 cJ;b2 77 'it>xe8 'it>xal
78 f6 cJ;b2 79 f7 al 'iW 80 f8'iW
'ifa4+ 8 1 'it>f7 'i!i'b3+. This position
arose in our analysis, when we
considered that 82 'it>f6 Wxe3 83

150 Brussels Ct (4) 1991

e5 c2 84 'iff4 d3 85 d5
would have given chances for
victory, but in our post-mortem,
Timman demonstrated the more
striking continuation 82 d5 ! 'ii'xe3
83 'ifb4+ c2 84 g5 'iff3+ 85 g7
e3 86 'ifc4+ b2 87 'ifb5+ c 1 88
d6 e2 89 'ifc5+ b2 90 'ffb6+ c 1
9 1 d7 el'if 92 d8'ff 'ffec3+ 93
'ifdf6, etc.
e7
72 i.g8
73 h6 (D)
In the event of 73 f6+ f8 74
.i.e6 b3 (74 ... .i.a4 leads to the
same position as in the game, but
74 . . . .i.d7 ! will save Black, and if
75 i.xd7? then 75 ...b3 76 f5 b2
77 g5 lbe2! 78 g6 lbg3+ 79 g5
lbh5 ! 80 xh5 bxal'ff 8 1 g7+
82 .i.e8+ xf6 and Black unex
pectedly takes the initiative) 75
.i.xb3 lbxb3 76 lbxb3 i.c4 77 lba1
f7 78 f5 White has the upper
hand.

73
Only now does Timman make
the decisive error. 73 ...b3! 74 .i.xb3
lbxb3 75 lbxb3 .i.c4 76 lbal f6 !
would have led to a draw.
74 .i.e6
i.d7
75 g5
b3
76 g6
1-0
If 76 ...i.xe6, then 77 fxe6 b2 78
g7+ g8 (78 ...e7 79 g811f bxal'ff
80 'it'f7+ d6 8 1 'ifd7#) 79 e7
80 g8'it'+ xe7 8 1 'it'g5+ f7 82
'ifd5+ and after two more checks
White takes the b2-pawn.
000

Game 38
Karpov - Anand
Brussels Ct (4) 1991
QGD, Semi-Slav
1 d4
d5
c6
2 c4
3 lbf3
lbf6
4 lbc3
e6
5 e3
lbbd7
6 'ifc2
After 6 .i.d3 dxc4 7 i.xc4 b5 8
i.d3 the classical variation of the
Meran arises. Moving the queen
forward to c2 is a standard Anti
Meran possibility.
6
.i.d6
7 .i.e2
Besides this move, White has a
choice between 7 i.d3, 7 g4 and 7
b3.
0-0
7
000

Karpov - Anand 151

8 0-0
Black's counterplay is linked
with the advance . . .c5 or . . .e5 (or
both), usually after a preliminary
exchange on c4. All my games as
White in this match against Anand
developed in this fashion.

8
9 .ixc4

dxc4
'fle7 (D)

In the second game of this match

after 9 ... a6 10 l:td1 'fle7 1 1 h3 b5 I
did not place my bishop too suc
cessfully: 12 .id3 c5 1 3 lDe4 c4 14
lDxd6 'flxd6 15 .ie2 .ib7 1 6 .id2
l:tfc8, and Black solved his open
ing problems.

'ifc5+ 16 h2 .id7 ! was being

played, and Black had a favourable
game) 1 3 . . . lDe5 14 .ib3 .id7 15
.ie3 lDg6 and everything was in
order for Black.
My dispute with Anand contin
ued to the 8th, and decisive, game
of the match. On that occasion I
changed my move order and
played 10 a3. Then in the event of
10 ...c5 1 1 dxc5 .ixc5 Black has to
bear in mind 1 2 b4 .id6 1 3 lDb5,
but Anand instead advanced his e
pawn - 1 0...e5.
Thus, in this opening variation
White can play a3 or h3, while Black
can play ...c5, or ...e5, and any of
these moves could essentially in
fluence the future course of the bat
tle. Modem opening theory is built
on precisely such nuances. I am
convinced that assessments here
are still to be redefined more than
once.

1 1 e4!
12 l:tdl !

After 12 ... exd4 13 lDxd4 5 14

.ifl White's pressure in the centre
is quite tangible.

w
10 h3

e5
b5

a6

In game 6 Anand improved his

play by means of 10 ...c5 1 1 dxc5
.ixc5 1 2 e4 .id6 1 3 lDd4 (incred
ibly, at that very moment on the
same stage in the quarter-final
match between Korchnoi-Timman,
1 3 lDb5 lDe5 14 e5 .ixe5 15 f4

13 .tn
14 d5

c5
c4?!

A serious inaccuracy; the mod

est 14 ...l:tb8 was correct, and then
in the event of 15 a4 b4 1 6 lDb1 b3
17 'iie2 l:tb4 Black would have
gained counterplay.

15 a4

l:tb8

152 Brussels Ct (4) 1991

Now 1 5 ...b4 does not work be

cause of 16 lbe2 .!Dc5 17 .!Dd2 aS
1 8 lL!g3 and the advanced black
pawns become objects for attack.
axb5
16 axb5
17 .:l.a5! (D)

_... . -

ma
l.

- l.

-
- .
.

a-
--

- ltj

.
d
- \lllr

,.

!;,'

.: d
.i.

I had this position on the board

during my preparation for the

game. The attack by the black
queenside pawns, which is inci
dentally forced, seems dangerous,
but in fact these foot soldiers will
soon be stopped.
b4
17
18 .!Da4
1 8 lL!b5 d8 ! 1 9 .!Dxd6 'ii'xa5
20 .!Dxc4 'ii'c7 !? is not so clear.
'ii'd8
18 ...
19 .:l.a7
b3!?
After 1 9 . . . .!Db6 20 i.e3 b3 2 1
'ii'b l it i s more difficult for Black to
create counterplay, and he does not
have full compensation in the event
of 1 9 ... c3 20 bxc3 b3 2 1 'ii'e 2 lL!c5

22 .!Dxc5 i.xc5 23 .:tal b2 24

i.xb2 .!Dxe4 25 'ii'xe4 .:l.xb2 26
.:l.d2 either.
20 'ii'e2
The c4-pawn for the time being
is invulnerable: 20 'ii'xc4 .:l.b4 2 1
c6 lL!b8.
20
.!DeS
21 .!Dxc5
i.xc5
22 .:tal
c3!
Postponing this pawn's march
until a later date is unsuccessful,
viz. 22 .. .'ii'd6 23 i.d2 i.a6 24
i.c3, and White's superiority is be
yond doubt.
23 .!Dxe5!
Anand had probably only reck
oned on 23 bxc3, and then 23 ...b2!
24 i.xb2 .:l.xb2 (24....!Dxe4 25
1i'xe4 .:l.xb2 26 'ii'x e5) 25 'ii'xb2
.!Dxe4 26 .:l.d2 .!Dxd2 27 'ii'xd2 e4
28 lbel e3 ! would have led to a po
sition with chances for both sides.
After this bold knight move the en
emy pawn reaches White's second
rank, but fortunately its career
comes to an end there.
23
c2
24 .:l.d3 (D)
I chose this continuation, al
though I also examined 24 .:l.el
.:l.e8 25 .!Dc6 'ii'b6 26 'ii'f3 which
has roughly the same value.
1t'e8
24
Now 24 . .. .:1.e8 loses: 25 .!Dc6
'ii'b6 26 e5 .!Dxd5 27 .:l.xd5 'ii'xc6
28 \i'c4.

Karpov - Short 153

-.t.- -
-

- - - -
- 8 - - 8 -

- -:- - 8
B -'iVD
8.

'
'

!:'-'

!:'-'

.- !i!

a m -i.=
+ if?

'

!i!

l:tb6
25 li:)c6
li:)xe4
26 .i.e3
li:)xc5
27 .i.xcS
'iVd7
28 l:te3
l:txc6
29 'iVc4
29 ...'iVd6 30 l:tc3 li:)d7 3 1 l:txb3
is no better.
'iVd1
30 dxc6
'iVd6
31 l:te1
32 'iVcJ!
White has a clear-cut plan to win
a pawn: the queen makes way for
the bishop on c4, after which the
threat of l:ta5 arises.
32
'iVd5
32 ... .tf5 33 'it'e5! 'iVxe5 34l:txe5,
etc., does not help.
33 c7!
33 l:ta5 would have been too
hasty in view of3 3 ... li)e4! 34 l:txd5
li:)xc3 35 l:tc5 li:)a4 ! 36 l:tb5 li:)xb2
37 l:txb3 li:)dl and Black unexpect
edly wins.
33
.i.b7
li:)e4
34 l:taS
.

35
36
37
38
39
40
41
42
43

lbd5
l:td3
lbb3
.i.c4!
l:ta3
l:txcl
l:tcJ
lbcl
.i.n

li:)xcJ
li:)a2
.i.c8
g6
cHi'
li:)xcl
l:te8
l:te7
10

Game 39
Karpov Short
Linares Ct (7) 1992
QGD, Tartakower

1 d4
d5
e6
2 c4
li:)f6
3 ll)cJ
.i.e7
4 .i.g5
0-0
5 e3
h6
6 ll)f3
b6
7 .i.h4
The Tartakower Variation of the
Queen's Gambit, also known to
Russian players as the Bondarev
sky-Makogonov Variation, is very
widely used in modem practice,
and moreover many grandmasters
willingly play it as both Black and
White. When the h-pawn attacks
the bishop, you always have some
hesitation: should you exchange on
f6, or wait until ... b6 has been
played, and temporarily retreat the
bishop to h4? In my encounters
with Kasparov, this bishop di
lemma arose 24 times (equivalent

154 Linares Ct (7) 1992

to a whole World Championship

match !), and most frequently the
bishop avoided immediate ex
change.
8 .i.e2
.i.b7
.i.xf6
9 .i.xf6
exdS (D)
10 cxdS
At first glance the position
seems fairly simple, but this vari
ation is quite poisonous, and de
mands careful play by Black, as
we shall see. Curiously enough,
this position arose eight times in
my matches against Kasparov. In
spite of the stormy nature of our
battles, these disputes strangely al
ways ended in peace.
B lack is already preparing to
carry out ... c5, and the natural reac
tion is b4, and after an exchange on
c5, l:[bl .

w
1 1 b4
cS
By playing I l ...c6 Black could
have avoided hanging pawns, but

this looks rather passive. After 1 2

0-0 consider the following exam
ples:
a) 12 .. .'ifd6 1 3 'ifb3 lLld7 I4
:rei .:r.ad8 I 5 .:r.ab1 .:.res I6 .i.d3
.i.a8 (Speelman-Short, London Ct
( I ) I 988) I7 h3 and then e4 gives
White a small but stable advantage.
b) I 2 . .. .:r.es 1 3 'iib 3 and then
(D):

B
b i ) 1 3 ...'ifd6 I 4 a4 lLld7 1 5 a5
l:tad8 I 6 axb6 axb6 17 l:ta7 'iib 8
IS .:r.a2 b5 I9 lLle1 .i.e7 20 lLld3
.i.d6 2I g3 lLlb6 22 .i.f3 .i.c8 23
l:tfal .i.f5 24 lLlc5 lLlc4 25 lLle2
gave White a secure plus in P.Nik
olic-Short, Manila IZ 1 990.
b2) Black tried to improve with
1 3 . . . a5 in Karpov-Short, Amster
dam 1 99 1 . The game continued I4
a3 lLld7 1 5 b5 c5 1 6 lLlxd5 .i.xd4
( 1 6 ... .i.xd5 I7 'ifxd5 lLle5 1 8 'iib3
ltlxf3+ I9 .i.xf3 .:r.cs 20 .i.c6 is no
good). After 17 .:r.ad 1 lLle5 ! 1 8

Karpov - Short 155

lt:ixe5 -'.xd5 1 9 lt:ic4 'ifg5 20 g3

'iff5 2 1 1lfe1 'ife4 22 f3 'ifxe3+ 23
'ifxe3 llxe3 24 lt:ixe3 -'.xe3+ 25
fl -'.d4 26 llxd4 cxd4 Black
managed to stand firm, but 17 exd4
llxe2 1 8 1lfe 1 ! ? would have given
Black certain difficulties. What
ever, at our next meeting in the
Candidates match, Short pushed
his c-pawn two squares.
bxcS
12 bxcS
'ii'aS
13 llb1
As a whole series of games
against Kasparov showed, 1 3 .. .-'.c6
is more passive.
cxd4
14 'ifd2
-'.xd4
15 lt:ixd4
16 exd4 (D)
Not, of course, 16 'iVxd4 lt:ic6 1 7
'iVd2 d 4 1 8 exd4 -'.a6 an d Black i s
already better [editor's note: but in
The Queen 's Gambit for the At
tacking Player 1 9 -'.xa6 'ifxa6 20
lt:ie2 llfe8 2 1 llb2 is analysed to
an ending with marginal winning
. chances for White].
-'.a6
16
1 6 . . . -'.c6, as seen in the 40th
game of my first match against
Kasparov, is weaker. I was playing
B lack, and I only managed to hold
the position with difficulty.
17 lt:ibS
'ii'd8
18 0-0
Now the black knight has two
possibilities - it can develop to d7
or c6. After 1 8 . . . lt:id7 a rapid rout

B
ended the game Vaganian-Geller,
New York 1 990: 1 9 1lfc 1 lt:if6 20
f3 ! lle8 2 1 a4 1le7 22 -'.d3 -'.c8?
23 'iVf4 ! a6 24 lt:ic7 g5 25 ...xf6
lle1 + 26 1lxe1 'ifxf6 27 hl 'ilfxd4
28 1lbd1 ! , etc.
18
lt:ic6
19 .:.Cd1
Soon after our Candidates match
the Olympiad took place (Manila
1 992), and Short again chose this
variation against Azmaiparashvili.
The Georgian team captain obvi
ously supposed that after our duel
Short would have prepared some
sort of improvement, so he played
something different: 1 9 a4. I have
also played this pawn move, if
slightly later, but here White de
cided there was no time to lose. Af
ter 1 9 . ..'iff6 20 llfd1 1lfd8 2 1 1lb3
llac8 22 h3 ...g5 23 'ilfxg5 hxg5 24
llg3 ! f6 25 -'.g4 1lb8 26 1lc3 Black
had a difficult endgame.
'iVf6
19

156 Linares Ct (7) 1992

20 n (DJ
A modest novelty, which I had
prepared specially for the match. It
is more precise than 20 f3, as in
the game Yusupov-Beliavsky, Lin
ares 1 988, where after 20 ... .:tab8
2 1 a4 .:tfd8 22 'iic 3 c8 23 .:tbc 1
a6 24 'iixc6 'iixc6 25 .:txc6 axb5
26 axb5 the players agreed a draw.

w
26 .:te3!
lbe6
26 .. ..:tc2? loses directly: 27 .:te8+
'iii>h7 28 'iixc2 1fxc2 29 d3+.
27 .:te5
'ii'g4
28 e2!
'ii'g6
29 g3
.:tbc8 (D)

B
20
.:tabS
21 a4
:res
22 .:tb3!
xb5
lbd8 (D)
23 axb5
24 'ifa2!
Now the white rook on the third
rank prepares to move to one of
three important squares, a3, e3 or
f3.
24
.:tc7
If 24 ....:tb7, then 25 .:te3, threat
ening 26 'ifxd5 and 26 .:te8+. The
weak d5-pawn is causing Black a
lot of trouble.
'ji'fS
25 .:ta3!

. . . .

41\a
- IJ

'ii' IJ

/
u u

.:.

30 h5!
The d5-pawn is going nowhere,
and so there is no point hurrying to
take it: 30 .:txd5? .:tc2; 30 jfxd5
.:td8 3 1 'ji'e4 'ji'xe4 32 .:txe4 .:tcd7.
1i'f6
30

Karpov - Kamsky 157

g6
l:tc3
l:ta3 (D)

31 'ii'xd5
32 .te2
33 h4

38 . .. 'ilt'xd4 39 l:txd4 l:taa2, though

it is true that here the precise 40
f l ! would have maintained an
appreciable advantage for White.
xg8
39 l:tg8+
40 'ii'xf6
l:txe2
lbe4
41 :at
l:txa2
42 l:txa2
l:txf2+
43 'ii'd4
l:te2
44 gl
1-0
45 d6
Game 40
Karpov - Kamsky
Moscow Alekhine mem 1992
Grtinfeld Defence

w
34
35
36
37
38

'ii'e4
g2
d5
'ilt'd4
l:te8 (D)

l:tcc3
g7
lbc5
l:tc2

. :

mu
.
."" ...
-

- - :n

. -

-

-
u
.

B .I Bi.B

. : .
B

l:taa2?
38
Black could have shown more
dogged resistance by continuing

This game, played at the Moscow

super-tournament celebrating the
l OOth anniversary of Alekhine's
birth, went, as they say, wonder
fully well. A bitter struggle was
conducted over the whole board,
first on one flank, then the other,
and in the end the black pieces be
came totally disorientated.
lbf6
1 d4
2 c4
g6
.tg7
3 lbfJ
c6
4 g3
First White (by playing 4 g3 in
stead of 4 lbc3) and then Black
(by 4 ... c6 instead of 4 ...d5) avoids
the sharpest continuations in this
opening, giving the game a posi
tional feel.
d5
5 .tg2

1 58 Moscow Alekhine mem 1992

6 cxdS
cxdS
0-0
7 ltlc3
e6
8 ltleS
9 0-0
9 .i.g5 has been played several
times, but it is not compulsory to
fix the position of the dark-squared
bishop at this early stage.
9
ltlfd7
10 f4 (D)

B
Supporting the knight with the
pawn is more logical than retreat
ing it to f3.
10
ltlc6
1 0 ...f6 is also played quite fre
quently, but practice has shown
that it has more disadvantages than
advantages.
1 1 ..t.e3
ltlb6
Here as well 1 1 . . .f6 is possible
(see game 37).
12 .i.f2
.i.d7
In Karpov-Kasparov, Seville
Wch (3) 1 987, Kasparov preferred
.

1 2 .. . ltle7, so that after 1 3 e4 he

could exchange pawns and seize
the d5-square. Although Black
managed to equalize, it was felt
that his development was defi
nitely awkward.
13 e4
ltle7
With the idea of placing the
bishop on c6 and strengthening the
centre.
14 ltlxd7
'ifxd7
15 eS (D)

B
l:.ac8
15
1 5 ...l:.fc8 was Kasparov's pref
erence in Karpov-Kasparov, Se
ville Weh ( 1 ) 1987. This game went
1 6 l:.cl .i.f8, when White could
have sacrificed a pawn with 17 g4
l:.c4 18 f5 !?, achieving a serious in
itiative. Instead the game contin
ued 17 .i.f3 l:.c7 1 8 b3 l:.ac8 1 9
'ilfd2 ltlc6 20 'ilfb2 a6 2 1 .i.e2 'ilfe7
22 ltlb1 lJb4 23 ltlc3 ltlc6, ending
soon in a draw by repetition.
.

Karpov

16 l:lcl
a6 (D)
Now Black has a ' hanging'
knight on b6, which will soon put
him on the spot. 1 6. ..l:lc7 would
have been safer, and if 17 'ifb3,
then 17 . . .ltlc4 1 8 l'Llb5 l:lc6 1 9
ltlxa7 l:la6 20 l'Llb5 l:lb6 with coun
terplay. Generally, while Black's
position on the queenside looks
quite reliable, the less said about
the kingside the better.

Kamsky 159

The first in a series of fine

moves by the queen, which charac
terized this game. In the variation
20 l'Lle2 l:lxc 1 2 1 l:lxcl l:lxc l + 22
tbxc 1 ltlc6 23 lDd3 White main
tains his spatial advantage, but the
position has lost its internal energy,
and his chances on the kingside are
not all that great.

.
.
-
-

--
-
B BiD B

u
8

u
.
8 !!U
" mu

8
M
.JL
U
R
8 8

.
-:
*!
B

lbc6
20
a3
21 fS!
All the black pieces have trans
ferred to the queen side, but for the
time being this is not dangerous,
because the white pieces on the
other half of the board are not yet
co-ordinated. 2 1 .. .exf5 22 gxf5
'ifxf5 is unsuccessful in view of 23
h3 'ifh5 24 xc8 h6 25 'ifh3
xcl 26 'ifxh5 gxh5 27 xb7
b2 28 xc6. 2 l ...b4 would not
have led Black to his aim either,
due to 22 hl intending 23 el .
l'Llb4
22 l:lcd1

17 b3
Emphasizing the awkward position of the knight on b6.
ltc7
17
1 8 'ifd2
:Cc8
19 g4!
The starting gun is fired for
White's attack. The constrained
nature of the black pieces makes
counterplay significantly more dif
ficult.
f8
19
20 '1Ve3! (D)

160 Moscow Alekhine mem 1992

Now the black pieces are getting

under each other's feet. The knight
is blockading his own bishop, a
fact of which I instantly made use.
However, 22 ...lbxd4 23 W'xd4 is
no better, since the b6-knight is
hanging (23 ....tc5 24 Wf4 .txf2+
25 1fxf2).
23 'ii'h6 (D)
Stronger than 23 lbbl lbc2! 24
W'h6 ..tf8.

'iWe8
23
In the new situation 23 . ..lbc2
does not guarantee equality, as af
ter 24 lbe2 the black rooks have no
entry square for their invasion.
23 .. .'ii'e7 is no better: 24 lbbl .tb2
25 ..th4 'ii'f8 26 'ii'd2 .:.c2 27 1Wel ,
and the black pieces are soundly
tied up on the queenside, despite
their superficial activity. 23 ...lbd3?!
does not work either, in view of 24
lbxd5 ! lbxd5 25 .:.xd3 .tf8 26
'ii'g 5 h6 27 fxe6 ! 'ii'xe6 28 .txd5,

when White simply has an extra

pawn.
24 lbbl!
Now the bishop has to desert the
important a3-f8 diagonal.
24
.tb2
25 'ii'd 2!
The white queen's shuttling is
causing Black more than a little
bother.
lbc2 (D)
25
The startling 25 ... a5 ?! can be re
futed by means of 26 a3 (the trap is
26 ti'xb2 l:r.c2 27 'ii'a3 l:r.xa2)
26 . ..l:r.c2 27 W'e l ii'bS (27 ...lbc6
28 l:r.d2 !) 28 axb4 .:.e2 29 it'xe2
"it'xe2 30 bxa5 (or 30 .:.d2 'ii'xg4 3 1
.:.xb2) 30...lbd7 3 1 .:.d2 .
..

..

........,
- a"i
-
8

J
u

- . fli.B
attJa : a : =

w
Black has been demonstrating
the maximum activity for the last
five moves. However, the threaten- ,
ing-looking tandem of the bishop
on b2 and knight on c2 has to be
defended, and only reduces the

Karpov - Kamsky 161

activity of his own rooks. Having

siopped his opponent's aggression,
White begins to develop his attack.
26 hl !
Freeing g l for the bishop, which
in its turn uncovers the fl -rook.
"ile7
26
lbd7
27 i.gl
28 .Uf3
With every move White's attack
is gaining energy, while Black is
still at a dead end.
28
'ifb4
29 'ifh6!
Black has no time to take the d4pawn due to the threat of 30 .Uh3
lbf8 3 1 f6. 29 'iff4 would have
been slipshod: 29 ...i.xd4! 30 i.xd4
lbxd4 3 1 .:Xd4 .:c 1 + 32 .:.n .:Xfl +
33 i.xfl 'ifel 34 lbd2 lbxe5 with
sufficient counterplay.
29
'iff8
30 'ifgS (D)
In the event of 30 'ifh4 the cap
ture 30 ... lbxd4 does not work due
to 3 1 i.xd4 i.xd4 (or 3 1 .. ..l:lc l 32
i.xb2 ! .l:lxdl + 33 .l:lfl ) 32 .Uxd4
.Uc 1 + 33 .Ufl .Uxfl + 34 i.xfl .l:lc 1
35 lbd2 and White maintains a
material advantage. However, both
30 ... 'ifd8 and 30 ...'ifg7 are quite
reliable for Black. Now after either
of these moves White should con
tinue 3 1 'ifd2, and in comparison
with the position on move 28, the
situation has altered appreciably in
White's favour.

The black queen cannot yet re

turn to f8 in view of the reply fxe6.
31
b6
There is no other way of freeing
the 'hostages' on b2 and c2. Black
is preparing 3 1 . ..a5 in order to se
cure the b4-square for his knight to
retreat.
aS
32 .Udn
33 h4
lbb4
34 a3
Obviously not 34 'ifxb2 owing
to 34 ... .Uc2. After the text the rook
invasion on c2 is not dangerous for
White.
.Ucl
34
35 'iff4
lbc6
36 .th3!
Threatening 37 fxe6 fxe6 38 g5
lbf8 39 'ifxf8+ ! ! 'ifxf8 40 i.xe6+
g7 4 1 .l:lf7+! , winning.
lbd8
36
37 i.e3!
..

1 62 Moscow Alekhine mem 1992

The c l -square is under control,

therefore the bishop scrutinises the
h6-square.
37
b5
38 ft3fl! (D)
Having appreciably strength
ened his position on the kingside,
White suddenly changes tactics.
Paradoxically, because of the awk
ward position of his pieces, espe
cially the b2-bishop, Black, after
exchanging rooks, will perish pre
cisely where he once dominated. In
a few moves the white queen be
comes mistress of the position on
the c-file.
..

li)fS 47 'ife7 ! 'ifd7 48 .i.h6 ! - an

amusing finale!
li)xe5
43 'ili'c2
Despair. 43 . ..'ili'f8 44 'ifc7 'ife8
45 f6 is also bad, and if 43 ... li)b8,
then 44 li)d2 ltJdc6 45 lbf3 is suffi
cient, as White is effectively play
ing with an extra piece. But now,
having accepted the sacrifice, I
managed to find a beautiful forced
win.
44 dxeS
'iWxeS
45 'ili'c8!
'ili'e4+
45 ...'ibe3 allows mate in three.
46 i.g2
'ili'xbl +
47 h2
i.b2
48 'ili'xd8+
g7 (D)

-
mit
l
-

z . - -

u
tffi "'

- u
o m .t

-
--

.l[). .:.
"

B
b4
38
39 axb4
axb4
40 llxc2
llxc2
41 :n
ftxfl
42 'ili'xfl
.i.a3
Alas, after 42 ....i.c3 there is the
immediately decisive 43 f6 'iff8 44
'ili'c2 i.el 45 'ifc7 'ili'e8 46 i.fl

w
i.xf6
49 f6+!
h6
50 i.h6+!
51 'ili'xf6
'ifcl
h5
52 g5+
53 gJ!
Strangely, 53 h3? would have
let victory slip from my grasp:

Kamsky - Karpov 163

53 ...1Wf5+ 54 1Wxf5 gxf5 55 .i.f3+

g6.
53
'ii'c7+
54 h3
1-0
For this victory in the Alekhine
memorial, I was awarded a prize
for 'playing in Alekhine's style' .
Strictly speaking, I do not feel this
was not a very precise decision. In
fact, in the work of the fourth
World Chess Champion there were
many examples of the battle being
transferred from the queenside to
the kingside. But here we were ac
tually dealing with a rarer theme,
as on the contrary, I managed to
transfer the battle from the king
side to the queenside with decisive
effect!
Then this beautiful victory also
soon won the best game prize in In
formator.
..

Game 4 1
Kamsky - Karpov
Dortmund 1993
Caro-Kann Defence
1
2
3
4
5
6
6 .i.c4
tinuation.
6

e4
c6
d5
d4
ltJd2
dxe4
ltJd7
ltJxe4
ltJg5
ltJgf6
.i.d3
is another popular con
...

e6

7 ltJ1f3
.i.d6 (D)
One of the key positions in the
modern treatment of the Caro
Kann. Besides 7 ....i.d6 Black may
choose between 7 ... 'fic7, 7 ... i.e7
and 7 ...h6.

w
8 1We2
More active than 8 0-0, which
Kasparov once played against me
(Amsterdam 1 988); then after the
continuation 8 ...h6 9 lbe4 ltJxe4 10
.i.xe4, by playing 10...ltJf6 !? Black
could have equalized. In fact I
chose another plan: 10 ...0-0 1 1 c3
e5 1 2 .i.c2 l:le8 1 3 lle1 exd4 1 4
l:lxe8+ ..,xe8 1 5 'fixd4 'fie7 16
.tf4 .txf4 17 'fixf4, and White
gained an advantage.
8
h6
9 ltJe4
The standard sacrifice on e6 is
not dangerous for Black in this
situation: 9 ltJxe6 fxe6 10 .i.g6+
e7 1 1 0-0 lDf8 1 2 .td3 .td7 1 3
...

164 Dortmund 1993

lbe5 e8 14 f4 d8 1 5 c4 c7 1 6
d2 .:.cs 1 7 c5 e7 1 8 f5 b8 1 9
.:.ael d8, and he had a clear ad
vantage in the game Amason-0st
enstad, Torshavn 1987.
tDxe4
9
10 1i'xe4
In the event of 1 0 xe4 0-0
Black quickly plays ...c5 or ...e5
with a good game.
10
lDf6
1 1 ... h4 (D)
.

B
1 1 'ii'e2 is played more fre
quently, but in fact on h4 the queen
looks quite threatening. 1 1 . .. 0-0
will not do because of 1 2 xh6 !
gxh6 1 3 'ii'xh6, and Black cannot
defend himself from the threat of
14 lLlg5 . After l l ...'iia5+ 1 2 d2
'ifh5 1 3 'ifxh5 lDxh5 1 4 lDe5 0-0
1 5 0-0-0 he has a prospectless posi
tion.
1 1 . . .lDd5 is recommended by
theory, but after 1 2 'ifxd8+ xd8

1 3 c3 White preserves a small but

stable positional advantage. What
should Black do?
11
e7!?
I thought of this paradoxical ma
noeuvre, keeping the king in the
centre, in 1 988, when I was prepar
ing for the above-mentioned game
against Kasparov in Amsterdam. I
was afraid of the white queen being
transferred to the king side, but ex
changing queens struck me as be
ing rather insipid. I sat for some
time over this position before I
found the correct solution. So this
important novelty had to bide its
time for five whole years !
With this odd king move Black
unexpectedly brings harmony to
his forces. Now the threat of ...g5g4 has been created, winnilfg a
piece, and if White is to count on
keeping his opening advantage, he
will have to act decisively.
12 lDeS
The most ambitious solution.
Bearing in mind that the enemy
king is stuck in the centre, White
decides to sacrifice a pawn. Of
course, he will hardly be happy af
ter 1 2 f4 b4+ 1 3 d2 (else
1 3 ... g5) 1 3 ...xd2+ 14 'itxd2.
12
xeS
'ila5+
13 dxeS
14 cJ (D)
1 4 d2 'ifxe5+ 1 5 e3 is also
possible, and it is very dangerous

Kamsky - Karpov 165

for B lack to take on b2, altho1:1gh

the position of the c-pawn is hardly
better for White than in the game
itself.

20 f4 (in order to close the c l-h6

diagonal from checks) 20...gxf4 2 1
i.d4 seems dangerous, for exam
ple (D):

.. . . . .

- .t.
.

-
- mu

- -

Wl1

i.

- u

"
-

.
:,

w
14
15
16
17
18
19

i.e3
0-0-0
'ii'a4
libel
'ii'a3

,\&

'ii'xe5+
b6
g5
cS
..td7
l:.hd8 (D)

B
a) 21 ...ifd5 loses on account of
22 l:.e5 Wxg2 23 l:.xc5! bxc5 24
ifxc5+ e8 25 ..txf6.
b) 2 l ...'it'c7 is also in White's
favour: 22 i.eS ifc6 23 i.xf6+ (a
rare combination arises after 23
..txf4 ! ? 'ilt'a4?? 24 ..td6+! e8 25
l:.xe6+! fxe6 26 i.g6#) 23 ...xf6
24 i.e4 Wa4.
c) 2 1 . . .'it'g5 ! 22 l:.e5 ifh4 ! res
cues Black, and 23 g3 does not
lead White to his goal in view of
23 . .. fxg3 24 hxg3 ifxg3 25 l:.xcS
bxcS 26 ifxcS+ e8 27 i.xf6
iff4+.
20
'ii'c7
The black queen abandons the
centre of the board, and restrains
all approaches to the king.
21 ..td4

20 g3

166 Dortmund 1993

Now after 2 1 f4 there is the

strong retort 2 1 . .. ttlg4 22 .td4
gxf4 23 h3 ttle3.
.te8!
21 ...
22 b1
l:td5
22 . . . .tc6 ! 23 .te5 'ifd7 24 .tc2
'ifb7 is perhaps even better.
.:.Sd8
23 f4
24 .tc2
24 .te5 1i'c6 ! 25 c4 l:t5d7 26
'ifc3 ttlg4 27 fxg5 hxg5 28 l:tfl
l:td4 ! favours Black.
l:t5d6
24xf6
25 .txf6+
hxg5
26 fxg5+
l:txd6
27 l:txd6
28 c4
The only move, as otherwise the
white queen turns out to be ex
cluded from the game.
e7
28
29 ii'e3
f6
30 h4 (D)
Here in mutual time trouble,
Kamsky offered me a draw, but I
turned it down, as Black's chances
are evidently better: White has
been deprived of the advantage of
the bishop pair, and he is still a
pawn down.
gxh4
30
'ifd7
31 gxh4
e5?!
32 W'h6
In time trouble I decided to de
fend the f6-pawn, and missed a
chance to do something much
more effective: 32 ...l:td2 ! 33 l:tfl
..

have been stronger.

34 W'h7+
d8
34 .. ..tf7 35 .tg6 l:tdl + would
have led to a draw (35 ...'ifxc4 36
b3 ii'e6 37 l:tgl ).
35 h6
l:td2
36 W'f5
'ifxf5
37 .txf5
.td7? (D)

Karpov - Kramnik 1 67

38 i.g6?
A mistake; Kamsky failed to
make use of my blunder in time
trouble. After 38 c 1 ! Black would
even have to fight for a draw:
38 ...l:td4 (the rook cannot abandon
the d-file because 39 l:td1 would
win the bishop) 39 i.xd7 ! l:txc4+
(after either capture on d7 there
follows 40 l:td1 , and the h-pawn
will inevitably promote) 40 bl
l:th4 (40 ... xd7? 4 1 l:th l ) 4 1 l:td l
l:txh6.
l:th2
38
e7
39 h7
i.e6
40 i.d3
fS
41 l:tg1
f6
42 l:tg7+
e4 (D)
43 l:txa7

l:txa2
49 l:txb6
0-1
If you exclude Black's mistake
in time trouble, the game was quite
logical, and its result was fully in
order.
Game 42
Karpov - Kramnik
Linares 1994
QGD, Semi-Slav
1 d4
dS
2 c4
c6
3 f3
f6
4 c3
e6
bd7
5 e3
6 i.d3
Around this time I was avoiding
the Meran System, preferring the
Anti-Meran 6 'ii'c2 (see, for exam
ple, game 38), but this did not
mean that I had excluded it from
my opening repertoire for ever.
6
dxc4
7 i.xc4
bS
8 i.d3
After the move 8 i.e2 it would
be more difficult for White to count
on an advantage, as the game Kar
pov-Kasparov, Moscow Web (29)
1 984/5, showed: 8 ... i.b7 (8 ...b4 9
a4 i.e7 10 0-0 i.b7 1 1 a3 a5 1 2
'ii'c2 0-0 1 3 l:td 1 c5 ! 1 4 xc5 lill.c5
15 dxc5 'ii'c7 1 6 axb4 axb4 17
l:txa8 l:txa8 1 8 i.d2 i.e4 ! 1 9 'ifc4
'ifxc5 20 jfxc5 i.xc5 gave Black

1 68 Linares 1994

an equal position in Polugaevsky

M.Gurevich, Antwerp 1 993) 9 a3
b4 1 0 lLla4 bxa3 1 1 bxa3 !l...e7 1 2
0-0 0-0 1 3 !l...b2 c5 112-112.
8
a6
After 8...!1...b7 9 a3 b4 10 lLle4 a5
1 1 lLlxf6+ lt:Jxf6 1 2 e4 !l...e7 1 3 'ii'e2
c5 14 !l...b 5+ 8 15 dxc5 !l...xc5 1 6
!l...d 3 h 6 17 0-0 g5?! 1 8 !l...e 3 !l...xe3
1 9 'iixe3 rlig7 20 li:Jd2 lDd7 2 1
lt:Jc4 'iie7 2 2 'ifd4+ e 5 2 3 'ifd6
'iWxd6 24 lDxd6 !l...c6 25 !l...c4 I
gained a big advantage in Karpov
Shirov, Linares 1994.
c5
9 e4
c4
10 d5
My theoretical duel with the
young chess star soon continued in
Monaco, albeit in a rapid-play. Af
ter the continuation 1 0 . . . 'ilfc7 1 1
0-0 !l...b7 1 2 dxe6 fxe6 1 3 !l...c 2,
Kramnik could have returned to
the game we are examining with
1 3 . .c4, but of course it ended sadly
for him, so he continued instead
with 1 3 ...!1...e7 !? ( 1 3 ...!1...d6 14 lDg5
gives White a big advantage) 14
lt:Jg5 W'c6 1 5 'ii'f3 h6 16 'ii'h 3. Here
1 6...lLlf8 !? would have led to a
rather complicated game, but in
stead Black preferred a very dubi
ous exchange sacrifice: 1 6 . .. hxg5?
17 ...xh8+ rlif7 1 8 ...h3 g4 19 'ilfh4
lt:Je5 20 f3 gxf3 2 1 gxf3. Black has
no compensation for his material
loss, and later I could have taken
the upper hand by force more than
...

once, but I was caught out by the

time control.
1 1 dxe6
fxe6
After 1 1 .. .cxd3 1 2 exd7+ 'ilfxd7
1 3 0-0 !l...b7 14 .:.et , it is well
known that Black faces an uphill
struggle for a draw.
12 !l...c2
!l...b7
'ilfc7
13 0-0
14 lDg5
lLlcS (D)
Black played 1 4 ...'ii'c6? ! in the
game Gligoric-Ljubojevic, Linares
1 99 1 , but the stormy complications
of 1 5 'irf3 ! !l...c5 1 6 'irh3 rlie7 17
lDf3 b4 1 8 lLle2 lDxe4 19 lDed4
'iib6 20 !l...xe4 !l...xe4 2 1 'ifg4 !l...xd4
22 'irxe4 !l...f6 23 'ii'xc4 were
clearly not in his favour.
Moving the knight to c5 looks
reasonable, but...

w
15 e5!
A clever break in the centre.
There are other possibilities, 1 5
'irf3, 1 5 'ire2, and 1 5 f4, but as I

Karpov - Kramnik 169

have the two bishops, I want to

open up the position ! Generally,
the idea of sacrificing a pawn with
e4-e5 is not new, but I am not
aware of it being seen in precisely
this position before.
'ii'xeS
15
The pawn must be taken, as after
1 5 . . .'ifc6 1 6 f3 Black's battery on
the long diagonal is liquidated, and
things are looking bad for him.
'ii'd6
16 :et
17 'i!kxd6!
This turn of events is completely
unexpected; White is a pawn down,
but he nevertheless voluntarily ex
changes queens. Rather paradoxi
cal !
.txd6 (D)
17 ...
.

w
18 .te3!
Quietly finishing his develop
ment, as the e6-pawn is going no
where.
0-0
18 ...

After 1 8 ... d3, White should

play the simple 19 .txd3 cxd3 20
.:r.ad l , and Black loses both pawns
in the centre. 1 8 . . b4 is no better:
1 9 .txc5 .txc5 20 a4, and either
the knight or rook can take on e6.
Finally, 18 ...g4 is no good be
cause of 19 .txc5 .txc5 20 .:r.xe6+
'ifild7 2 1 .:r.dl+ r/;c7 22 . etc.
19 .:r.ad1
.te7
Black has to lose a tempo - after
19 ...:ads 20 .:r.xd6 .:r.xd6 2 1 .txc5
.:r.d2 22 :c l Black does not have
sufficient compensation for the
material he has given up because of
the threat of 23 .te3.
.txcS
20 .txcS
:res
21 xe6
22 h3! (D)
Not so much to open an escape
hatch as to advance the g-pawn.
The raid 22 xc5 .:r.xc5 23 .:r.e7
.:r.bs allows Black to hold the posi
tion.
.

1 70 Linares 1994

The critical moment. The black

pieces are poorly co-ordinated, and
the question is whether he can im
prove their interaction.
i.f8?!
22
22... b4 loses: 23 lba4 i.a7 (oth
erwise 24 lbb6) 24 lbxg7 'it>xg7 25
l:te7+. If 22 ... i.b4, then 23 a3,
forcing Black to give up his bishop,
as after 23 ...i.a5 there is the cap
ture 24 lbxg7 'it>xg7 25 l:te7+. A
later game, Topalov-Lautier, Dos
Hermanas 1994, featured 22...l:tab8,
and after 23 g4 i.f3 24 l:td2 b4 25
lba4 i.a7 26 g5 lbd5 27 l:te5 l:tb5
Black eventually managed to stand
firm. However, the impression re
mained that he was on the edge of
the abyss for a long time.
h6
23 g4
Or 23 ... .tf3 24 l:td4 ! b4 25 lba4
lbd5 26 l:te5 ! , etc. If 23 ... g6, then
24 lbxf8 l:txf8 25 g5 followed by
26 l:td7 is very unpleasant.
i.f3
24 f4
25 l:td2 (D)
25 l:td4 looks more energetic,
and if 25 ...i.c6, then 26 g5 hxg5 27
fxg5 lbd7 28 l:th4 and the black
king falls into a mating net. How
ever, by continuing 25 ...l:te8 Black
could threaten 26 ... l:r.xe6 27 l:r.xe6
i.c5 or 26...i.c5 immediately.
i.c6
25
hxg5
26 g5
lbd7
27 fxg5
28 lbxf8
.

i.
R .
RltJ

08&

""

.t.
-8

8 "i.
u

.
.

. .

B
28 .l:.f2, with the idea of i.g6f7+ and l:tf4, is tempting, but af
ter 28 ... .tc5 (28 ...lbc5? 29 lhf8+)
29 lbxc5 lbxc5 30 l:te5 White has
only a minimal advantage: 30...lbd3
3 1 i.xd3 cxd3.
xf8
28 ...
In the event of 28 ...l:r.xf8 the
rook manoeuvre to d6 becomes
more valid.
29 l:td6
29 lbd5 also looks reasonable,
aiming to exchange off Black's ac
tive bishop.
b4
29 ...
30 lbe4
.i.e8?! (D)
30 ...i.d7 was necessary: 3 1 h4
(3 1 lbg3 l:te8 !) 3 1 ...l:tc6 and Black
can still breathe.
31 lbg3!
I was placing great hopes on this
move - the appearance of the
knight on f5 or h5 is extremely un
pleasant.
lidS
31 ...

Karpov - Kramnik 1 71

Evidently 37 a4 is more logical,

rounding up the black a-pawn later.
l:td3
37 ...
38 gl
l:txb3
39 h4 (D)

w
Now 3 l . . ..i.d7 does not work:
32 l:te7 .i.xh3 33 ll)hS, etc.
l:txd6
32 ll:)fS
.i.g6
33 ll)xd6
34 .i.xg6
ll)xg6
35 ll)xc4
l:td8 (D)
35 ...l:tf8 36 l:te4 ll)f4 37 h4 h7
38 ll)eS is suicide.

. .

.
- -

. D

.ttJ.
. . . -
B

%z

36 l:te4
b3
If 36 ...l:td3, then 37 g2 and the
h-pawn marches forward.
37 axb3

B
ll)f8?
39
Of course, the defence 39 ... f8
40 h5 ll)e7 4 1 l:te5 e8 would
have been more stubborn, although
after 42 l:ta5 (or 42 l:tc5) White,
with his extra pawn, must emerge
victorious .
40 l:te8!
1-0
It is becoming clear that the
black king cannot break out to free
dom; if 40 . .tf7. then 4 1 ll)d6+ !
and the rook is defended. While
Kramnik was grieving over the fact
that moving his king from g8
would not free his pieces from their
paralysis, his flag fell.
It remains only to say that while
my game against Topalov from the
same tournament won the best

1 72 Linares 1994

game prize in lnformator, this

game in the very same edition won
the prize for the most important
theoretical game.
Game@
Karpov - Beliavsky
Linares 1994
Catalan Opening
We have come to the end of the Li
nares tournament, and with it the
end of this book. Victory over Be
liavsky in the final round would
guarantee me a lead of 2 points
over the rest of the field ! But this
game is not only remarkable for
the context of the competition, as
in it I also managed to use a piece
of opening preparation which had
been awaiting its hour of glory for
about twenty years !
So, let us return to the chess
board for the last time.
ltlf6
1 d4
d5
2 ltlfJ
e6
3 c4
i..e7
4 g3
0-0
5 i..g2
dxc4
6 0-0
a6
- 7 ..,cl
8 a4
So far this is all very well
known. For example, I once had
the position after move 7 against
the very same opponent, although
on that occasion we were playing

with opposite colours! After 8

'iVxc4 b5 9 'ifc2 1&..b7 10 i..d 2 lDc6
1 1 e3 ltlb4 12 i..xb4 1&..xb4 1 3 a3
i..d6 14 ltlbd2 .:c8 15 b4 a5 a com
plex position with chances for both
sides arose in Beliavsky-Karpov,
Brussels 1 988.
i..d7
8
9 Wxc4
i..c6 (D)

w
10 i..g5
i..d5
Matters developed differently in
Kasparov-Karpov, Moscow Wch
(22) 1 984/5 : 10 ... a5 1 1 ltlc3 ltla6
1 2 .:act 'ilfd6 1 3 ltle5 i..x g2 1 4
'iPxg2 c 6 1 5 i..xf6 gxf6 1 6 ltlf3,
and again White has preserved
only a small advantage out of the
opening.
c5
1 1 'ii'd3
12 ltlc3
cxd4
'ii'xd5 (D)
13 ltlxd5
It is necessary to capture on d5
with the knight. In the old game
Htibner-Karpov, Tilburg 1 979, the

Karpov - Beliavsky 1 73

w
exchanges 1 3 ... lLlxd5 14 i.xe7
'flxe7 15 lLlxd4 lLlc6 16 lLlxc6 bxc6
left White only a symbolic advan
tage.
Taking on d5 with the queen is a
new move ...
1 4 h4! !
. . . new for Beliavsky, but not for
the author of these pages ! I had
faced this position on the board in
1 974, as I was preparing for my
match against Korchnoi in Mos
cow - the match which was soon to
give me the title of 1 2th World
Champion ! It goes without saying
that it is very rare for an opening
novelty to be used two decades af
ter it was prepared. Furthermore, I
must admit that this novelty is shat
tering: Black's position collapses
like a house of cards.
Capturing on d4 with the queen
is known to lead to a better end
game for Black, and taking with
the knight is impossible, since the

g5-bishop loses its support. Now,

with the bishop defended by the h
pawn, the recapture in the centre
can be made by the knight. Besides
this, the light-squared bishop ac
quires incredible strength, which is
extremely important. This is a very
rare example of an advance by a
rook's pawn practically deciding
the outcome of the game from the
very start.
lLlbd7
14 ...
14 ... lLlc6 suggests itself, but
White has at his disposal the fol
lowing cunning variation: 15 i.xf6
i.xf6 16 lLlg5 !, and again the pawn
on the edge of the board helps, this
time by defending the knight:
16 . ..iif5 17 i.e4 it'a5 18 i.xc6
i.xg5 19 i.xb7 .:.a7 20 i.e4. The
pawn has been won back, and
White's positional advantage is ex
tremely tangible.
15 lLlxd4
'ii'd6 (D)

1 74 Linares 1994

If 15 ...Wa5, then 1 6 lt:Jb3 ! .

1 6 l:lfd1
lt:JcS
17 ...c4
This is stronger than 17 lt:Jf5? !
lt:Jxd3 1 8 lt:Jxd6 lt:Jxb2 19 l:ld2 .i.xd6
20 l:lxd6 lt:Jc4 or 1 7 'i'c2 ...b6 1 8
a5 'ii'b4, when White's advantage
is not that large.
17
l:lfd8 (D)
...

etc.

w
18 b4!
lt:Jxa4
Beliavsky is confused, and this
is another consequence of White's
1 4th move. He should have played
1 8 . . . lt:Jce4 1 9 .i.xf6 lt:Jxf6 (but not
1 9 . . ..i.xf6? 20 .i.xe4 .i.xd4 2 1 e3
.i.xal 22 l:lxd6 l:lxd6 23 .i.xb7
with a big advantage) 20 .i.xb7
l:lab8 2 1 'i'xa6 'ilt'xb4 22 lt:Jb5 .
White would have taken the upper
hand, but the game would have
continued.
19 ...b3! (D)
Now it's all over: Black loses a
piece.

1-0
20 e3
This victory in the final round
was my ninth at Linares; it was
even more remarkable that I won
six straight games in the first six
rounds ! With four draws (I man
aged to avoid being beaten) I
reached an inconceivable score,
1 11 1 3 . And, incidentally, the sec
ond and third placed players, Kas
parov and Shirov, were 2112 points
behind me.
Without false modesty I can say
that in the whole of chess history,
tournaments where on the one
hand, all the stars of the chess
world are gathered together, and on
the other, the winner has demon
strated such notable superiority
over the remaining contestants, can
be counted on your fingers.

Karpov - Beliavsky 1 75

Kasparov once called the tradi

tional battle at Linares the World
Championship Tournament. It is a
pity that this title was not made

official, because it would have

been a wonderful crowning addi
tion to my collection of champion
ship titles !