1 d4 d5 2 c4 c6 3lDf3lD f6 4lD c3 e6

To my grandfather, the man who started it all.

The Semi-Slav
CHESS PRESS OPENING GUIDES

Other titles in this series include:

1 90 1259 05 6 Caro-Kann Advance Byron Jacobs

1 90 1259 06 4 Closed Sicilian Daniel King
1 901259 03 X Dutch Leningrad Neil McDonald
1 901259 10 2 French Advance Tony Kosten
1 90 1 259 02 1 Scandinavian John Emms
1 90 1 259 0 1 3 Sicilian T aimanov James Plaskett
1 901259 00 5 Slav Matthew Sadler
1 90 1 259 04 8 Spanish Exchange Andrew Kinsman
1 90 1259 09 9 Trompowsky Joe Gallagher

For further details of Chess Press titles, please write to The Chess Press
c/o Cadogan Books pIc, 27-29 Berwick Street, London W 1 V 3RF.
Chess Press Opening Guides

The Semi-Slav

Matthew Sadler

ir
rnm

The Chess Press, Brighton

First published 1998 by The Chess Press, an imprint of First Rank Publishing,
23 Ditchling Rise, Brighton, East Sussex, BN 1 4QL, in association with
Cadogan Books pIc

Copyright © 1998 Matthew Sadler

Distributed by Cadogan Books pIc, 27-29 Berwick Street, London WIV 3RF

All rights reserved. No part of this publication may be reproduced, stored in a

retrieval system or transmitted in any form or by any means, electronic,
electrostatic, magnetic tape, mechanical, photocopying, recording or
otherwise, without prior permission in writing from the publishers.

A CIP catalogue record for this book is available from the British Library

ISBN 1 90 1 259 08 0

Cover design by Ray Shell Design

Production by Book Production Services
Printed and bound in Great Britain by BPC Wheatons, Exeter
 CONTENTS

1 d4 d 5 2 c4 c 6 3 ttJf3 ttJf6
4ctJc3 e 6

Bibliography 8
Introduction 9

5 .1l.g5
1 Botvinnik Variation: Main Line with 15 . . . b4
(5 ... dxc4 6 e4 bS 7 eS h6 8 .1l.h4 gS 9 ttJxgS hxgS 10 .1l.xg5 ttJbd7
1 1 exf6 i.b7 12 g3 cS 13 dS 'iib 6 14 .1l.g2 0-0-0 15 0-0 b4) 12
2 Botvinnik Variation: Black's 13th Move Alternatives
(S ... dxc4 6 e4 b5 7 e5 h6 8 .1l.h4 g5 9 ttJxgS hxgS 10 .i.xgS ttJbd7
1 1 exf6 .1l.b7 12 g3 cS 1 3 d5) 31
3 Botvinnik Variation with 1 1 g3
(5 ... dxc4 6 e4 bS 7 eS h6 8 .i.h4 g5 9 ttJxgS hxg5 10 .1l.xgS ttJbd7
1 1 g3) 41
4 Botvinnik Variation: Early Deviations after 5 .i.gS dxc4 54
5 Moscow Variation with 7 e3 (5 . . . h6 6 i.xf6 �xf6 7 e3) 64
6 Moscow Variation: White's 6th and 7th Move
Alternatives (5 . . . h6) 77

5 e3 and other m oves

7 Meran Variation: Main Line (5 e3 ttJbd7 6 .td3 dxc4 7 .1l.xc4 bS
8 i.d3 .1l.b7 9 0-0 a6 10 e4 c5 1 1 dS c4 12 .tc2 'ii'c7) 81
8 Me ran Variation: Move Orders and Sidelines 97
9 Meran Variation with 8 . . . .1l. b7: White Alternatives 110
10 Meran Variation with 8 . . . a6: Old Main Line 9 e4 c5 10 eS
- 123
11 Meran Variation: Systems with an Early . . . bS-b4 131
12 Meran Variation: Odds and Ends 140
13 6 'ii'c 2 (Karpov System): 7 .i.e2 and 7 �d3 145
14 O dds and Ends 154
Index of Complete Games
158
 BIBLIOGRAPHY

Books
Encyclopaedia o/ Chess Openings vol.D, Sahovski Informator, 1987
The Complete Semi-Slav, Peter Wells (Batsford, 1994)
D44, Alexander Beliavsky and Adrian Mikhalchishin (Sahovski Informator,
1993)

Period i cals
In/ormator
ChessBase MegaBase CD-R OM
New In Chess Yearbook
British Chess Magazine
Chess Monthly
 INTRODUCTION

Every chessplayer is both fascinated

and appalled by the Semi-Slav: fasci­
nated by the sheer romance of this
opening where the double-rook sacri­
fices and king hunts seem to recall the
golden years of the 'Immortal' and
'Evergreen' games; but appalled by the
number of complicated variations and
the volume of analysis surrounding it.
Often an initial burst of enthusiasm to
learn the opening is followed by pro­
found despair at the confusion that
such fantasy and complexity brings! 1 6 . . :ii'a6 1 7 dxe6 ioxg2 1 8 e7
Witness the following example, the
game Topalov-Kramnik, Dortmund
19 96 (analysed in full in Game 7 of
this bo ok) . Topalov and Kramnik
have had many great fights over the
years but this clash has to be my fa­
vourite. I just can't see it often
enough.
1 d4 d5 2 c4 c6 3 tLJf3 tLJf6 4 tLJc3
e6 5 iog5 dxc4 6 e4 b5 7 e5 h6 8
ioh4 g5 9 tLJxg5 hxg5 1 0 ioxg5
lLl bd7 1 1 exf6 i.b7 1 2 g3 c5 1 3 d 5
"it'b 6 1 4 iog2 0-0-0 1 5 0-0 b 4 1 6 1 8 . . . i.xf1 1 9 �d5 i.xe7 20 fxe7
rib 1 i.d3 21 tLJe4 i.xb1
-------
9
Th e S e m i- Sla v
22 ltJd6+ rJ;c7 23 �f4 rJ;b6 24
ltJxc4+ rJ;b5 25 ltJd6+ rJ;b6 26
exd8� +
26 . . . l:txd8 27 ltJc4+ �b5 28 ltJd6+
q.,b6 29 ltJc4+ q.,b5 Y.z - Y.z
Since the theory of the Semi-Slav is
1D
so well developed, the 'real' game may
sometimes only begin after move 18
or 19. A common approach when
preparing such an opening is to learn
these theoretical moves by rote in a
few minutes, and to concentrate only
on what comes after. This is a major
error; before one can play the early
middlegame well, it is necessary to
grasp the logic of the opening. This is
achieved not by detailed and time­
consuming analysis, but by describing
the conflict of ideas in the opening in
words: the positional aims of both
sides, the territory they seek, the
squares they weaken.
Question 1: What purpose does this
approach serve?
A nswer: Positional themes from the
opening - the weakening of the oppo­
nent's dark squares, the creation of a
queenside majority, etc. - shape and
define the early middlegame. Posi­
tional understanding of the opening is
the foundation of good play in the
early middlegame. By highlighting the
essentials of the position, this method
ensures that our middlegame analysis
will focus only on the important fac­
tors.
After 1 d4 ds 2 c4 c6 3 !Df3 !Df6 4
!Dc3, 4 . . . e6 introduces the Semi-Slav.
see follo wing diagram
The move 4 . . . e6 locks the light­
squared bishop on c8 inside the pawn
chain. However, by freeing the dark­
squared bishop, Black threatens to win
a pawn with S . . . dxc4 as the typical 6
a4, preventing . . . b7-bs can be met by
6 . . ..ltb4! 7 e3 (intending .ltxc4 )
when 8 axbs cxbs 9 !Dxbs is impossi-

ble as th e knight is pinned to the king.
White has two basic reactions:
1) He can place the dark-squared
bishop outside the pawn chain with 5
In tro duc t i o n
.ltgS, intending to defend c4 on the
next move with 6 e3 . After the consis­
tent S . . . dxc4, White can set up an
'ideal' centre with 6 e4 (threatening
.ixc4 regaining the pawn) when 6 . . . bS
(protecting c4) 7 eS plunges into the
complications of the Botvinnik system
(Chapters 1-6)
2) He can forestall . . . dSxc4 by
calmly protecting the c4-pawn with 5
e3. However, this has the disadvantage
of blocking White's dark-squared
bishop inside his pawn chain. A full
discussion of the Meran variation and
other possibilities after 5 e3 can be
found in Chapters 7-14.
11
 CHAPTER ONE

Botvin nik Variation :

Main Line with 1 5 b4 . . .

1 d4 d 5 2 c4 cS 3 lbf3 lbfS 4 lbc3 is practically forced: if White is al­

eS 5 �g5 dxc4 S e4 b5 7 e5 hS 8 lowed to recapture the pawn with 7
�h4 g 5 9 lbxg5 hxg5 1 0 i.xg5 �xc4, then Black will have given up
lbbd7 1 1 exfS �b7 1 2 g3 c5 1 3 d5 the centre for nothing. However, now
'fibS 14 �g2 0-0-0 1 5 0-0 b4 White can use his central control and
In this chapter we shall consider the the activity of his bishop on gS to at­
main line of the Botvinnik system. tack the knight on f6 with
However, let us start with an explana­ 7 e5
tion of the opening moves that lead to Now the knight cannot move as
tfie main line position. this would lose the queen to 8 �xd8.
With Black must use tactics to stay alive:
5 i.g5 7 . . . hS
White seeks to 'have it all': before
defending against Black's threat of
. . . dSxc4 with e2-e3, White wants to
develop his dark-squared bishop out­
side the pawn chain in order to obtain
the maximum activity for his pieces.
However, the drawback to 5 �gS is
that it does not defend the c-pawn, so
5 . . . dxc4
is Black's most consistent response.
S e4
then grabs the central space that
Black conceded by taking on c4, after counterattacks against the bishop
which on gS. White usually then retreats his
S . . . b5 bishop along the h4-d8 diagonal with

12
 B o t vin n ik Va ria tio n : M a in L in e w i t h 1 5
8 i.h4
again threatening to wm a piece
with 9 exf6.
Black's reply is forced:
8 g5
. . .
attacks the bishop again so that 9
exf6 can be met by 9 . . . gxh4, regaining
the piece. This move also blocks the
h4-dS diagonal so that a retreat of the
bishop with 9 .ll g3 can be met by
9 . ttJdS, leaving Black a pawn up.
..
Therefore White strikes with
9 ttJxg5 hxg5 1 0 i.xg5
By sacrificing his knight for Black's
g- and h-pawns, White has destroyed
the barrier on the h4-dS diagonal and
now threatens 11 .ll x f6, forking the
queen on ds and the rook on hS.
1 0 ttJbd7
...
defends the knight on f6 and starts
to develop the queenside pieces. This
is important since Black's king will
only find a safe{ish!) home on the
queenside - castling kingside would
require a supreme effort of faith! Now
1 1 exf6
regains the piece.
Although 5 .ll g S began as a pawn
sacrifice, White is now a pawn up!
The extra pawn on f6 is a strange one:
. . . b4
Black can easily recapture it, but this
would expose weaknesses in his posi­
tion. For example 11.. .ttJxf6 gives
White an unpleasant pin on the knight
on f6 by the bishop on gS. Conse­
quently, Black usually prefers
1 1 i.b7
. . .
intending the central break ... c6-cS,
which will attack the two points most
weakened by White's tactical efforts to
maintain the pin on the knight on f6:
Black's light-squared bishop now at­
tacks g2, which would usually be
shielded by a knight on f3, and the
pawn on cS attacks d4 which now
lacks the support of the knight on f3.
White also needs a safe place for his
king: he would like to castle kingside,
but with the removal of the h- and g­
pawns, Black has two half-open files
against the white kingside. Thus
White usually plays
1 2 g3
to protect his kingside. This shields
h2 against a combined attack by a
rook on hS and a queen on c7, and
protects g2 against a combination of a
bishop on b7 and a rook on gS. It
looks cheeky to play g2-g3 when
Black has just played his bishop on b7,
13
Th e S e m i- Sla v
but the logical
1 2 . . . c5
threatening . . . �xh l , meets with
1 3 d5
By maintaining the pawn on dS,
White checks his opponent's d-file
play and finds a cunning use for the f6-
pawn. If Black undermines dS with
13 . . . b4 then 14 dxe6! ! is very strong, as
14 . . . fxe6 10ses to 15 f7+, discovering an
attack on the queen.
1 3 . ... ii'b6
side-steps this, protecting e6 and bS
while allowing the black king to castle
queenside. After
1 4 �g2
both sides put their kings to safety:
1 4 . . . 0-0-0 1 5 0-0
and now
1 5 . . . b4
undermines the dS-pawn by attack­
ing the knight on c3, its most impor­
tant defender
Now we have arrived at the staning
position of this chapter. Black intends
to win the pawn on dS and over­
whelm his opponent with his central
control; whereas White will try to
open lines on the queenside against
Black's king - since Black's queenside
14
pawns are so far advanced, he has
plenty of targets!
The first six games in this chapter
deal with the most common move in
this position, 16 ttJa4. 16 llb l is con­
sidered in Games 7-10.
Gamel
Ivanchuk-Shirov
Wijk aan Zee 1996
1 d4 d5 2 c4 c6 3 lLlc3 lLlf6 4 lLlf3
e6 5 �g5 dxc4 6 e4 b5 7 e5 h6 8
�h4 g5 9 lLlxg5 hxg5 1 0 �xg5
lLlbd7 1 1 exf6 �b7 1 2 g3 c5 1 3 d5
�b6 1 4 �g2 0-0-0 1 5 0-0 b4 1 6
lLla4 'it'b5
Question 1: Why does the queen go
here?
Answer: Black has tried three moves
in this position: 16 . . . 'ili'a6 (see Game
6) , 16 .. :ifd6 (see the notes to Game 6)
and 16 ... 'i'bS. The latter seems the
best all-round positional move. The
queen protects c4 and attacks White's
loose knight on a4, preventing the
white queen from leaving the a4-dl
•
diagonal. The queen also stays in con­
tact with the bishop on b7, which al­
lows him to transfer the queen rapidly
to the a8-h l diagonal if White ex­
changes bishops with dSxe6 and
.ixb7. Finally, on bs the queen is rela­
tively safe from the white pieces!
1 7 a3
Now White will open the a-file
with a3xb4 and Black will respond
with ... cSxb4 to prevent the knight on
a4 from returning to c3. This has the
effect of 'diluting' the central black
pawn mass, which makes it much
harder for Black to achieve his desired
 B o t vin n ik Va ria tio n : M a in L in e w i t h 1 5
idea o f . . . dS-d4, once h e has taken the
pawn on dS with his e-pawn.
Question 2: Why does Black wish to
play . . . dS-d4?
Answer: The advance ... dS-d4 acti­
vates Black's pawn mass, which is his
major positional trump. It also invites
the exchange of light-squared bishops,
which generally weakens White's
kingside more than Black's queenside.
(See Game 3 for a graphic illustration
of this.)
17 dxe6 is discussed in Game S.
1 7 exd5
. . .
The text wins back the sacrificed
pawn as 18 .i.xds is impossible due to
18 ... tLleS ! 19 .ixb7+ 'i'xb7 20 'i'e2
tLlf3+ 21 <iith1l:lxh2+ mate.
17 . . . ttJeS and 17 . . . tLlb8 (which may
transpose to one another) are the sub­
ject of Game 4.
18 axb4 cxb4
Black has regained the pawn that he
sacrificed on the kingside by taking
the white central dS-pawn. This has
resulted in a lopsided pawn structure:
White has a 4-1 kingside majority and
a useful passed h-pawn, and Black has
a 4-1 queenside majority and a passed
d-pawn. Black's plan is simple: ... ds-
. . . b4
d4. White's plan is less obvious and
requires some thought:
1) White wants to make it hard for
Black to achieve his own plan. For
example, he may blockade the d-pawn
by occupying d4 either with his
bishop or queen. (Note that after . . . ds­
d4, the black queen on bS will attack
the bishop on gS.)
2) White can try to take advantage
of Black's weak dark squares: his cen­
tral structure leaves the dark squares
around it rather weak, while the f6-
pawn takes control of e7 and g7 away
from Black's pieces.
3) White can open up the black
queenside with b2-b3.
4) The f6-pawn provides two out­
posts for the white pieces: g7, and
most usually, e7. By placing a rook on
e7, White occupies the seventh rank
from a central position, giving him
targets on both wings. This move is
often played as a sacrifice, offering to
trade the rook for Black's dark­
squared bishop on f8 . After . . . .li.xe7,
f6xe7, White gets a passed pawn on
the seventh rank, just one step away
from queening, and removes the black
piece best suited to defending his weak
dark squares: the dark-squared bishop.
S) The black king looks quite safe
on c8, but if the knight were diverted
from d7, the king could be caught in
an unpleasant crossfire with a bishop
on f4 taking b8 from the king and a
queen on g4 or a bishop on h3 deliver­
ing a nasty check. Note that, for now
at least, the rook on h8 prevents
.li.h3+.
6) Finally, the most important tar­
get for White: the a7-pawn. This pawn
15
Th e S e m i- S la v
is a very useful defensive unit, cover­
ing b6 and therefore helping to stop
White's knight from becoming active.
It allows Black's knight to move from
d7 when it desires and also provides a
haven on a8 for the black king. How­
ever, it is also a natural target for
White's pieces: White can play the
positionally desirable moves 'ii'd4 or
i.e3, preventing . . . dS-d4 and at the
same time attacking a key black defen­
sive unit. Black will nearly always seek
to shield it from attack or defend it, as
moving the a-pawn weakens another
dark square: b6.
1 9 �e3
The major continuation. On gS all
the bishop seemed to do was defend
the f6-pawn, whereas on e3 the bishop
attacks the a-pawn and helps to pre­
vent the . . . dS-d4 push. For 19 :e1 see
Game 3 .
1 9 lLlc5
. . .
Shielding the a7-pawn, unmasking
the support of the rook on d8 for the
pawn on dS, and eyeing the outposts
on b3 and d3 . By exchanging White's
knight, Black frees cS for his bishop to
support . . . ds-d4.
20 -'g4+!
16
A rather awkward check since the
natural 2o ... �b8 21 "i'd4! It:'Jxa4 22
"ii'xa7+ and 20 . . . "ii'd7 21 'i'xd7+ ltJxd7
22 l:tfd1 ltJxf6 23 i.xa7!, intending
i.d4 and It:'Jb6+ (Agzamov) are both
dodgy for Black. However, 20 . . . �c7!
was suggested by Ivanchuk. Black is
threatening simply . . . It:'Jxa4, so White
must react rapidly. Then 21 i.f4+
conceals a cunning trap: 21. . .i.d6
blocks the check, but 22 It:'Jxcs 'ii'xcs
23 i.e3! wins the a-pawn. So 21. . .'it>c6!
is best.
Question 3: Well hang on a minute,
I...
Answer: I kno � it looks suicidal,
but it may well be okay! I have a the­
ory that in positions where the pawns
in front of your king have moved far
forward, your king is safest not on the
back rank, but on the third or fourth
rank, close to the pawn wall that
shields the king! Black does indeed
have a large number of pieces on the
queenside to protect his king in this
position, so White will have to sacri­
fice substantially if he is to breach his
opponent's position. After 22 It:'Jxcs
i.xcs, Black intends . . . 'it>b6 and . . . ds­
d4!
 B o t v in n ik Va ria tio n : M a in L in e w i t h 15
20 . . J:td7
After this move the rook on d7 is
pinned to the king by the queen on
g4. White will look to exploit this ei­
ther by playing .th3 or by opening up
the position to make this tactical detail
count for more. However, Black's
rook is a strong defensive piece, cover­
ing Black's seventh rank.
21 'it'g7! !
I told you that White had an out­
post on g7!
21 . . . .i.xg7 22 fxg7 :g8 23 lLlxc5
'
Black has a queen for knight and
bishop. However, the a7-pawn is
likely to fall, while the rook on d7 can
be attacked further by a bishop on h3,
so White will gain further material for
the queen. Several positional factors
are important:
1) White has a passed h-pawn.
2) White's king is very safe whereas
Black's is not.
3) White can blockade the d-pawn
with .td4, which stops Black from
activating his queenside majority and
leaves him with a passive bishop on
b7.
Question 4: What should Black do?
Answer: Black wants to play . . . dS-d4
. . . b4
to activate his queenside majority and
his bishop on b7, but it is hard to
achieve this. He should exchange a
pair of rooks to reduce the danger to
his king and give his queen more
room to enter the white position.
23 . . . d4?
This devilish idea is actually a tacti­
cal blunder. 23 .. Jhg7 is considered in
the next game.
24 i.xb7+ l:[xb7 25 lLlxb7!
Threatening to fork the king and
queen with 264Jd6+.
25 . . . 'it'b6!
This was Shirov's idea; 2S ... c.t>xb7
would simply have allowed 26 �xd4,
protecting the g7-pawn and attacking
a7. Shirov's move defends the d4-
pawn and parries the fork on d6. Un­
fortunately . . .
2 6 i.xd4!
26 . . . 'it'xd4 27 l:[fd 1 'i6'xb2 28 lLld6+
,.pb8 29 l:[db 1 ! 'i6'xg7
Ivanchuk points out that 29 . . . 'i'd2,
attempting to keep hold of the b4-
pawn, is cleverly met by 304Jxc4 'ii'c 3
31 :a4! b3 32 4JaS! b2 33 :b4+!, pick­
ing up the b-pawn, as 33 . . . 'i'xb4 loses
the queen to 34 4Jc6+, forking the
king and queen.
17
Th e S e m i - Sla v
30 l:txb4+ �e7 3 1 l:ta6 l:tb8 32
l:txa7+ �xd6 33 l:txb8 �g4 34 l:td8+
�e6 3 5 l:ta 1 1 -0
A magnificent game!
Game 2
lalic-J. Wilson
London 1 996
1 d4 d 5 2 e4 e6 3 lU13 lU16 4 lUe3
e6 5 �g5 dxe4 6 e4 b5 7 e5 h6 8
�h4 g 5 9 lUxg5 hxg5 1 0 �xg5
lUbd7 1 1 g3 �b7 1 2 �g2 �b6 1 3
exf6 0-0-0 1 4 0-0 e5 1 5 d 5 b4 1 6
lUa4 �b5 1 7 a3 exd5 1 8 axb4 exb4
1 9 �e3 lUe5 20 �g4+! l:td7 2 1
�g7 ! ! �xg7 2 2 1xg7 l:tg8 2 3 lUxe5
23 . . . l:txg7 24 �d4
In the notes to his game against Shi­
rov, Ivanchuk mentioned 24 l:txa7,
but now Lalic's 24 . . . d4!! seems good.
For example:
a) 25 �xb7 'iNxc5!
b) 25 �xb7+ litxb7! 26 ttJxb7
(threatening to fork king and queen
with ttJd6+) 26 .. :iVb6!! Now after 27
.1t.xd4 'iixd4 28 �fa1 �g6! the white
knight on b7 is very short of squares.
c) 25 ttJxd7 .txg2 26 .txd4!? just
leads to a draw after 26 ... .1t.xfl! (greedy
18
and correct - the best sort o f move!)
27 ttJb6+ Wb8 28 ttJd7+ \t>c8, as
27 ... Wd8 allows a lovely mate in two:
28 .tf6+ We8 29 �a8 mate.
Ivanchuk also mentions 24 .th3 f5!
25 .txf5 d4! (seizing the opportunity
to open the a8-h 1 diagonal) 26 �xd4
26 .. Jtgf7 'retaining definite counter­
chances'. I think it is clear that White
has lost control of this position.
The text is the right idea, stopping
any tricks with . . . d5-d4 and threaten­
ing to win material with ttJxb7 and •
then �xg7.
24 . . .1 5 25 �h3?
The bishop on d4 is White's most
important piece; and removing the
bishop from the h 1-a8 diagonal allows
Black to dislodge it by playing a major
piece to e4. 25 ttJxd7 l::txd7 26 �xa7
was better, with a mess, although it
may be a little easier for White to play
this position than Black.
The game illustrates what I mean by
this. Although Black gets a good ver­
sion of this line, he still has to be accu­
rate or his weak king and vulnerable
pawns will lose him the game. As we
shall see, the pressure very quickly
became too much for Black.
 B o t vin n ik Va ria tio n : M a in L in e w i t h 15
2 5 . . . 1:[gf7 2 6 lba7 l:tc7 2 7 lbe6
l:tce7 28 lbg5 �e8 29 lbxf7 �xf7 30
l:tfa 1 �h5 ! 3 1 �g2 f4 32 l:t7a5 f3
33 �f1 l:th7?
Black mIsses his chance! 33 ... l:le4!
(Lalic) was best, and would have been
very awkward for White to meet.
Now, however, Black is in trouble.
34 h4 �f5 3 5 �e3! �c2 36 �f4 l:tf7
37 �h3+ Wd8 38 �d6 l:tf5 39 l:ta7
�c8 40 l:[g7 �xb2 41 l:taa7 �b1 +
42 Wh2 l:tf8 43 �xf8 �xh3 44 �c5
�e6 45 l:ta8+ �c8 46 �b6+ 1 -0
Mate follows on the next move.
There is plenty of scope for practical
tests here, but, in general, I feel that
White has the easier task in a practical
game, even if I cannot say that he is
'theoretically' better.
Game 3
Stean-Rivas
Marbella 1982
1 d4 d 5 2 c4 c6 3 lbf3 lbf6 4 lbc3
e6 5 �g5 dxc4 6 e4 b5 7 e5 h6 8
i.h4 g5 9 lbxg5 hxg5 1 0 �xg5
lZ:lbd7 1 1 exf6 �b7 1 2 g3 c5 1 3 d5
'i'b6 14 �g2 0-0-0 1 5 0-0 b4 1 6
tLla4 �b5 1 7 a3 exd5 1 8 axb4 cxb4
. . . b4
1 9 l:te1
With this move White takes control
of the e-file and dreams of playing
lIe7.
1 9 . . . �h6
I was suspicious when I first saw
19 . . . �h6: Black allows his opponent
access to e7 and the seventh rank
without even having to sacrifice the
exchange! This is, however, a typical
idea in the Botvinnik variation: by
exchanging off the dark-squared
bishop, Black undermines the defence
of the f6-pawn and virtually assures
himself of winning it. He also removes
one of the best pieces for blockading
on d4 and attacking a7, the queen be­
ing the other.
The alternative 19 . . . d4!? 20 'iVxd4
�xg2 2 1 Wxg2 'iWxgS 22 'ii'xc4+ Wb8
23 l:ted 1! gave White a huge attack in
Van Wely-Piket, Wijk aan Zee 1994.
20 �xh6 l:txh6 21 �d4 l:.xf6
Black deals with the threat to his a­
t>awn tactically: 22 li'xa7 is met by
22 .. Jh6, winning the knight on a4.
22 �h3!
White uses the absence of the black
rook from the h-file to play this an­
noying move, pinning the knight on
19
Th e S e m i - Sla v
d7 to the king, and threatening 'iix f6.
22 .. J:td6 23 liJc5
The next few moves are critical.
Black is threatening to unpin with
. . . 'itt b 8, solving all his problems, so
White must take purposeful action.
23 . . :�i'b6!
Black does not mind losing the ex­
change on two conditions:
1} He can keep his other rook on
the board to help push his own pawns
and stop White's outside passed pawn.
2} He can exchange queens to re­
move the last piece capable of actively
blockading the d-pawn.
24 %:ted 1
24 :te7, to increase pressure on d7,
should be met by 24 . . . lii'c7 intending,
yes you've guessed it, 25 . . . 'itt c 6, un­
pinning and putting pressure on the
c5-knight. Remember that this knight
cannot capture on d7 while the queen
on d4 is unprotected. Now 25 J..xd7
:8xd7 26 lhd7+ l:txd7 26 'if e5+ I:td6
27 'i'e7+ lii'c 6! {27 . . . <it>c8 28 tDxb7 :d7
29 "ile8+ wins} gives Black a good po­
sition, while 25 l1d1 , defending the
queen on d4, is met by 25 . . . J..c 8, when
Black is a little tied up but after 26
:xf7 'itt c 6 27 tDe6 'i'xd4 {27 .. JlJce6 28
'i'xd5+} 28 tDxd8+ 'itt c 5 29 ':xd4 {29
tDe6+ :xe6! 30 lhd4 l::[e 1+ wins}
29 . . . 'ittxd4 30 tDe6+ 'itt d3, he escapes!
24 . . . �c7 25 �xd7 l:t8xd7 26 liJxd7
'ii'x d4 27 l:txd4 �xd7
Such unbalanced material endings
are very typical of the Semi-Slav.
Black will try to create a passed pawn
on the queenside by playing his king
to c5 to chase the rook from d4. This
will allow the black d5-pawn to ad­
vance to d4 with two effects: first, the
20
pawn can go 'all the way' with the
support of the rook on d6j and sec­
ond, the advance activates the bishop
on b7. From f3, it can remove a
blockader on d1, and from e4 it can
prevent the passed h-pawn from ad­
vancing to h7. White cannot hold d4
by doubling rooks on the d-file, as
Black will just create another passed
pawn either on the c-file { . . . c4-c3} or
the a-file { ... a5-a4-a3}j and he cannot
support the rook on d4 with his king
on e3, since Black will just check and
drive it away. White's task is the more
difficult, particularly in a practical
game. Passed pawns win endings and
White has just one while his opponent
has several!
28 h4 �c6! 29 h5 �c5 30 %:th4 l::t h 6!
Necessary to prevent h5-h6.
31 g4 d4 32 g5 l::t h 8 33 l::tx a7
33 :a5+ �b6 34 :f5 litd8! 35 'itt f 1
c3 36 bxc3 b3! is very good for Black.
33 . . . �c6 34 l:ta5+ �b5
After the text move the game
quickly peters out to a draw. Black
could still have played for a win with
34 . . . 'itt b 6!?
35 l:txd4 �xd4 36 l::t x b5 l::tx h5 37
l:txb4 %-%
 B o t v in n ik Va ria tio n : M a in L in e w i t h 15
Black's two other possibilities on
move 17 are linked in a very spectacu­
lar way.
Game 4
Kamsky-Kramnik
New York (Candidates match) 1994
1 d4 d 5 2 c4 c6 3 lLlc3 lLlf6 4 lLlf3
e6 5 i.g5 dxc4 6 e4 b5 7 e5 h6 8
i.h4 g 5 9 lLlxg5 hxg5 1 0 i.xg5
lLlbd7 1 1 exf6 i.b7 1 2 g3 c5 1 3 d5
"'b6 14 i.g2 0-0-0 1 5 0-0 b4 1 6
lLla4 'i6'b5 1 7 a3 lLle5
Very logical. Black unmasks the at­
tack of the rook on the d5-pawn and
intends to win it without blocking his
bishop on b7 with . . . e6xd5. From e5,
the knight aims for the defensive
square c6 as well as d3 . However,
White has a wonderful tactical possi­
bility that makes this move unplay­
able.
1 8 axb4 cxb4 1 9 'i6'd4!
Attacking the knight on e5 and the
pawn on a7.
1 9 . . . lLlc6
This position can also be reached
via a different move order: 17 . . . tDb8
18 axb4 cxb4 19 "iVd4 and now
. . . b4
19 . . . tDc6.
20 dxc6! ! l:txd4 21 cxb7+
White's compensation for the queen
is 'two pieces, a passed pawn on the
seventh and a host of tricks', to quote
Peter Wells.
2 1 . . . �c7 22 i.e3 e5 23 lLlc3 ! ! bxc3
24 bxc3 i.c5 ! ?
This was Kramnik's attempt t o re­
habilitate this line after the crushing
White win in Salov-Illescas, Madrid
1993, which continued 24 . . . .:d6 25
llab 1 ! a6 26 k!xb5 axb5 27 !:tal nd8 28
.ie4 .ih6 29 .ic5 .if8 30 .ia7 with an
overwhelming position for White.
25 cxd4 i.xd4
25 . . .exd4 loses to 26 .if4+ .id6 27
.ixd6+ �xd6 and 28 l:tfb 1 followed
by 29 b81'i + .
26 l:tfb 1 !
The stan of a magnificent series of
moves.
26 . . .... c5 27 l:ta6! 1:b8
27 . . . .ixe3 28 �c6+ 'i'xc6 29 .ixc6 is
clearly better for White according to
Kramnik.
28 i.c 1 ! !
Quite superb, bringing the bishop
round to the sensitive d6-square via a3.
28 . . . c3 29 i.a3! 'ii' c4 30 i.d6+ �d7
21
Th e S e m i - Sla v
3 1 �c6+! !
3 1 . . . �e6
3 1 . . .'i.t>xd6 loses the queen to 32
i.bs+.
32 i.b5! �xf2+ 33 'it'xf2 �d4+ 34
'it'f1 �e4 3 5 Ue 1 'ii'h 1 + 36 'it'f2
�xh2+ 37 'it'f3 l:txb7 38 i.xe5+ l:tb6
39 �c4+ 'it'd7 40 l:txa7+ 'it'c8 41
l:tc7+ 1 -0
A magical performance that de­
stroyed two lines - 17 .. .'�JeS and
17 ... liJb8 - in one game!
Now we turn to the old line: 16
liJa4 'i'bS 17 dxe6.
Game 5
Nikolic-Shirov
Wijk aan Zee (match) 1993
1 d4 d 5 2 c4 c6 3 lLlf3 lLlf6 4 lLlc3
e6 5 �g5 dxc4 6 e4 b5 7 e5 h6 8
�h4 g 5 9 lLlxg5 hxg5 1 0 �xg5
lLlbd7 1 1 g3 �b 7 1 2 �g2 'ii' b 6 1 3
exf6 0-0-0 1 4 0-0 c5 1 5 d 5 b4 1 6
lLla4 'ii' b 5 1 7 dxe6
Instead of opening the queenside
with 17 a3, White takes on e6, deny­
ing his opponent the chance to form a
massive queenside pawn chain and
22
keeping his extra pawn. The draw­
back, however, is the activity that
Black receives and that White forgoes.
First, without the open a-file and
a3xb4 to create holes in Black's queen­
side structure, White lacks a target on
the queenside. Second, dSxe6 opens
the d-file for Black's rook and allows
him to weaken White's kingside light
squares by exchanging the light­
squared bishops.
1 7 . . . �xg2 1 8 'it'xg2 �c6+! 1 9 f3
19 'fif3 loses the queen to the deflec­
tion 19 .. Jhh2+! 20 Wxh2 'iixf3 .
1 9 . . . 'ii'x e6
Threatening .. :i'h3+, winning the
h-pawn.
20 �c2
A dual-purpose move: White pro­
tects the second rank to meet
20 .. :i'h3+ with 2 1 Wgl , and gets the
queen off the d-file so that . . . liJeS no
longer comes with tempo.
20 . . . lLle5 2 1 l:tae1 l:td4! !
Savchenko's superb discovery in­
troduces the optimal set-up for the
black pieces. Black wants to develop
his dark-squared bishop to its most
active square on d6, but first brings
the queen's rook to d4 so that the
 B o t vin n ik Va ria tio n : M a in L in e w i t h 1 5
bishop does not block the rook's path
on the d-file. On d4 the rook defends
c4, leaves the d3-outpost free for the
knight, and prevents White from us­
ing e4 for his queen to trouble Black's
queenside light squares. White cannot
win a piece with 22 f4 due to the un­
pleasant 22 . . . 1I'd5+!
22 h4
Stopping Black from playing
. . . 'i'h3+, but weakening the g3-square.
22 . . . �d6
Eyeing g3 and completing Black's
development.
23 a3
This was Nikolic's improvement
over the stem game Rublevsky­
Savchenko, Helsinki 1992, where 23
�e3? showed another brutal point to
Black's set-up: 23 .. J:tdxh4!! 24 gxh4
:xh4
25 �gl (25 llh l lLlxf3! ! 26 �xf3
'i'g4+ wins) 25 . . . lLld3! 26 'ii' g2 �h2+
27 1I'xh2 :xh2 28 Wxh2 lLlxe 1 29
lixe l 'iIIx f6, threatening .. :iWh4+,
when Black had a clear advantage
(analysis by Savchenko) .
23 a3 aims to weaken Black's con­
trol of d4 by swapping off his useless
a-pawn for Black's c-pawn with a3xb4,
. . . b4
. . . c5xb4. White players may like to
investigate Shirov's suggestion of 23
:e2 'i'ld5 24 1I'f5+ rt;c7 25 life l (25
�f4 :e8 is unclear) 25 . . . lLld3 26 'iixd5
�xd5 27 :e7+ .llxe7 28 lhe7+ 'ottc6 29
:'xf7 with a mess, as 23 a3 does not
seem to work.
23 . . ...d5
Centralising and increasing Black's
influence along the d-file, while eyeing
the f3-pawn.
24 'iVf5+ �c7 25 :e2 �c6 ! !
Not a surprise to us now!! This su­
perb id�a prepares to exchange off
queens with . . . lLld3 and allows the
king to recapture on d5 after 'iVxd5, in
order to support the queenside pawns.
Shirov evaluates the position as
slightly better for Black.
26 b3! c3 27 axb4 cxb4 28 l:ta 1 !
lLld3 29 'iVxd5+ l:txd5 30 l:te4! �b5
3 1 l:tc4! l:te8! 32 l:ta2! lLle 1 + 33 �h3
lLlxf3 34 l:txc3 lLlxg5+ 35 hxg5 l:th8+
36 �g2 bxc3 37 lLlxc3+ �c6 38
lLlxd5 �xd5 39 l:xa7 'it>e6 40 l:ta4?
40 'ottf3 l1h3 41 �g4 J1xg3+ 42 �h5
lth3+ 43 'ottg4 Iixb3 44 l:ta6 would
have allowed White to squeak a draw
according to Shirov. The text allows
Black to win the crucial g5-pawn.
23
Th e S e m i- S la v
40 . . . %:I.g8 4 1 %:I.g4 �e5 0-1
After 42 ... 'itf5, Black will win both
the g5- and f6-pawns.
Game 6
Ivanchuk-Shirov
Novgorod 1 994
1 d4 d 5 2 c4 c6 3 lLlc3 lLlf6 4 lLlf3
e6 5 �g5 dxc4 6 e4 b5 7 e5 h6 8
�h4 g 5 9 lLlxg5 hxg5 1 0 �xg5
lLlbd7 1 1 exf6 �b7 1 2 g3 c5 1 3 d5
�b6 14 �g2 0-0-0 1 5 0-0 b4 1 6
lLla4 "a6
On a6 the queen attacks the knight
on a4, keeps in touch with the bishop
on b7 and hence the a8-h 1 diagonal,
and protects the pawn on c4. In con­
trast to 16 . . :VWb5, this move also keeps
in touch with e6 so that after d5xe6 or
. . . e6xd5 the queen can transfer to e6,
eyeing h3 and the white light squares
on the kingside. The drawback is that
a3xb4, opening the a-file, will be more
dangerous as the queen is in the line of
fire of the rook on a 1 .
16 . . .'iVd6!? 1 7 �f4 'iWa6 aims for a
16 . . :ii'a61ine with the white bishop on
f4. Although f4 is a more attacking
post than g5 - taking c7 and b8 away
24
from the black king - it does leave the
f6-pawn unprotected, enabling Black
to recapture on f6 put extra pressure
on the d5-pawn. Ionov-Bjerring, El
Vendrell 1996, continued 18 dxe6 fxe6
19 .ixb7+ �xb7 20 'iYg4 �h6 21 .id6!
'i'c6 22 llfd1! with an initiative for
White.
1 7 a3 �xd 5 ! ?
This imaginative idea i s attributed
to Alexander Shabalov.
1 8 �xd5 lLle5 1 9 '§e2 %:I.xd5 20 axb4
cxb4 2 1 lLlc3 "c6! ?
There i s another interesting idea
here, which Piket played in a TV
game against Lutz in Germany:
2 1 . . J:ta5!? After 22 l:txa5 'i'xa5 23
ttJe4, Piket played 23 . . . ttJd3 24 b3
'ilfe5!? (Lutz mentions 24 . . . �f5!?, aim-
ing for h3), when 25 I,ic 1? ttJxc1 26
'iVxc4+ was easily countered by
26 .. .'iVc7 27 'iVxc1 'ili'xc 1 28 �xc 1 as
with a clear advantage to Black. This
so impressed Lutz that when he got
the chance a little later against
Korchnoi, he decided to play it as
Black (Horgen 1994) . Unfortunately, •
he was once again on the wrong side
of the board! Korchnoi found the
much stronger 25 :td1 ! , and sacrificed
a piece for a vicious attack after
25 . . . ttJc5 26 'ii'xc4 'ii'xe4 27 'i'b5!, aim­
ing for the e8-square .
22 lLlxd5 "xd5
Threatening . . . ttJf3+.
23 f3 �c5+ 24 'it>g2 lLld3
Unfortunately, as Kharitonov
points out, the lovely 24 . . . ttJg4 25 h4
'ii'xg5, to meet 26 hxg5 with 26 . . . .:.h2+
mate, fails to simply 26 fxg4!
25 h4 �b7!
The active king again! Black's idea is
 B o t vin n ik Va ria tio n : M a in L in e w i t h 15
t o play . . . �b6 and allow the a-pawn to
join in the fun with . . . a7-aS.
26 l:ta5
The right idea but not the best exe­
cution. White's idea is to challenge the
knight on d3 by undermining its sup­
port with b2-b3, attacking the pawn
on c4. After Black has taken on b3 or
played . . . c4-c3, White has two ways to
put pressure on the black position:
1) a:fd l , attacking the knight on d3
which is pinned to the queen on dS.
2) lIaS, attacking the bishop on cS
which is pinned to the queen on dS.
In a later game Kharitonov­
Sabanov, Moscow 1995, White played
the immediate 26 b3! , which seems to
give Black a lot of problems. In the
game, Black chose 26 . . . c3 27 l:tfdl �d8
28 �aS! (threatening 29 �xcS 'iVxcs 30
�xd3, winning a piece) 28 .. st>b6 and
now instead of the violent 29 lIxcS,
Kharitonov claims an initiative with
29 'ifia2!, and this does indeed seem
very strong for White. If Black wishes
to try this line, he therefore must find
an improvement on this game - which
is beyond me for the moment!
26 . . :�d4 27 b3 �b6!
In comparison with 26 b3, Black
. . . b4
gains a tempo to put his bishop on the
safe b6-square.
2S l:ta2 c3 29 l:td 1 l:tdS 30 h 5 ! a5 3 1
g4 lLlf4+ 3 2 �xf4 'ii'x d 1 3 3 �xd 1
l:txd 1 34 h6 �a6?
A blunder. 34 . . . eS! 3S h7 (3S .i.xeS
�d2+ 36 \t>h3 �xa2 37 h7 c2 38 i.f4
i.c7! wins, as 39 i.e3 c1 'ii 40 i.xc1
llh2 is checkmate) 3S .. Jld8 36 i.xeS
.l:th8 would have favoured Black ac­
cording to Shirov. Suddenly, White is
wmnmg.
35 g5 l:tdS 36 �f1 �d4 37 �e2 e5
3S �e3 �b5 39 h7 l:thS 40 �xd4
exd4 41 g6! fxg6 42 f7 �c6 43 �d3
�d7 44 l:te2 a4 45 l:teS axb3 46
l:txhS b2 47 l:tdS+ 1 -0
Now we move on to the other
branch of the main line: 16 �b l .
Gamel
Topalov-Kramnik
Dortmund 1996
1 d4 d5 2 c4 c6 3 lLlf3 lLlf6 4 lLlc3
e6 5 i.g5 dxc4 6 e4 b5 7 e5 h6 S
�h4 g5 9 lLlxg5 hxg5 1 0 �xg5
lLlbd7 1 1 exf6 �b7 1 2 g3 c5 1 3 d 5
�b6 1 4 �g2 0-0-0 1 5 0-0 b4 1 6
l:tb 1
By placing his queen's rook on the
b-file, White meets the threat of
16 . . . bxc3, as 17 bxc3 'ii a6 18 Iixb7!
'iix b7 19 dxe6! is clearly better for
White.
Question 5: Why does White want
to keep his knight on c3?
Answer: By holding his knight on
c3, White maintains his support of the
dS-pawn, blocking Black's play along
the d-file and the a8-h 1 diagonal.
White avoids committing his knight
25
Th e S e m i - Sla v
to the rim, and hopes to transfer it
later to a central post such as e4 or dS.
1 6 . . :�a6!
Black renews the threat of ... b4xc3,
as b2xc3 will now not come with gain
of tempo. He wants to force matters
and not to let his opponent consoli­
date behind the barrier on dS.
After 16 . . . .th6 17 .txh6 �xh6 18
b3! cxb3 19 lba4 'iWbS 20 axb3 exdS 2 1
:c 1 Black's position was very loose in
Piket-Illescas, Dos Hermanas 1985.
1 7 dxe6 �xg2!
17 ... fxe6 1 8 lbe4! (centralising the
knight) is good for White, while
17 .. :�xe6 allows 1 8 .txb7+ �xb7 19
'i'f3+ with dangerous play against the
black king. Kramnik's move exploits
the fact that 1 8 exd7+ loses to
18 .. .Ihd7, attacking the white queen
on d 1 , while after 18 'it>xg2 Black can
play 18 . . :�'xe6, threatening both
19 . . . bxc3 and 19 . . . 'iWh3+.
1 8 e7 !
An amazing intermediate move,
forking the rook on dS and bishop on
fS .
1 8 . . . �xf 1
Black can also play l S . . . i.aS, keep­
ing the bishop on the h 1-aS diagonal
26
and hoping to play the queen to c6
and deliver mate on g2 or h I . Peter
Wells considers that 19 lbdS 'i'b7 20
exfs'i' �hxfs 2 1 lbe7+ cJ;;c7 22 .tf4+
lbe5 23 .txeS+ �b6 24 f3! (prevent­
ing . . . 'iWg2 or . . :i!Vh 1 mate) 24 . . .l:hd1 25
l:tbxd1 is fairly promising for White
and I am in full agreement. Cenainly,
none of the top players have tried this
for Black.
1 9 'ii'd 5 ! ?
An astonishing idea o f Yermolin­
sky's. White, a rook and a piece down,
ignores the material on offer and in­
stead creates another threat: 20 'ii'a S+
lbbS 21 exdS'ii' + �xdS 22 'i'xbS+.
19 Wxfl is considered in Games 9
and 10.
1 9 . . . �xe7!
A typical idea, lessening the impact
of 'i'aS+ by defending the back rank.
20 fxe7
Black has protected the back rank
with gain of tempo: now 2 1 'i'aS+
lbbS 22 exdS'fI + l:hdS 23 .txd8 will
win material, but will create no threats
against the black king. Black can.
therefore use his 'spare move' before
this happens to attack the white rook.
20 . . . �d3!
This also gives Black the threat of
. . . b4xc3, which will no longer open
the b-file, as Black can then take the
rook with . . . .txb 1 !
In fact 20 . . . .td3! was Kramnik's
remarkable improvement on his own
20 . . . �dgS (see the next game) .
21 lLle4
Since 2 1 'i'aS+ lbbS and 2 1 .tf4
(threatening 22 'iWa8+) 2 1 . . :i'b7!
achieve nothing, White must bring
another unit into the attack. From e4,
 B o t vin n ik Va ria tio n : M a in L in e w i t h 1 5
the knight can move t o the dangerous
d6-square.
21 . . . �xb 1 !
Greedy but good! White must
hurry as his oppoenent threatens
. . . .ltxe4, removing the attacking white
knight.
22 ltJd6+ �c7
This is the critical position. Now 23
lbxc4 threatens the unpleasant 24
.ltf4+, but then 23 . . .f6 meets 24 .ltf4+
with 24 . . . lbe5, blocking out the check
and discovering an attack on the white
queen, which is winning for Black. 23
lbxf7 also looks interesting, intensify­
ing the attack on the rook on dS,
while also threatening the check on f4,
but the calm 23 . . .l::t dcS 24 .ltf4+ �b6
wins after 25 'i'xd7 'ii'b 7! or 25 'i'd6+
l:!c6! 26 'iYxd7 'iYc8 . Yermolinsky sug­
gests 23 exdS'ii' + .l::lxd8 24 lbxf7; and
this looks like the most critical test, as
24 . . . �eS 25 .ltf4+ \t>b6 26 'iYd6+ \t>a5
27 'ii'xd7 gives White good play.
23 �f4!? �b6!
Avoiding the discovered check.
24 ltJxc4+ �b5 25 ltJd6+ �b6 26
exd8�! ? l:txd8 27 ltJc4+ �b5 28
ltJd6+ �b6
Neither side can avoid the draw by
. . . b4
repetition!
29 ltJc4+ �b5 % - %
Game 8
Kasparov-Kramnik
New York (rapidplay) 1994
1 d4 d5 2 c4 c6 3 ltJf3 ltJf6 4 ltJc3
e6 5 �g5 dxc4 6 e4 b5 7 e5 h6 8
�h4 g5 9 ltJxg5 hxg5 1 0 �xg5
ltJbd7 1 1 exf6 �b7 1 2 g3 c5 1 3 d5
�b6 14 �g2 0-0-0 1 5 0-0 b4 1 6
l:tb1 �a6 1 7 dxe6 �xg2 1 8 e7 �xf1
1 9 �d5 ! ? �xe7 ! 20 fxe7 l:tdg8 2 1
ltJe4!
We've seen this idea before! White
brings the knight to an attacking posi­
tion from where it aims to give an ex­
tremely unpleasant check on d6.
21 . . . l:tg6
Preventing the knight check on d6,
and intending to challenge the white
queen on d5 with .. :ifc6.
22 l:txf 1 ? !
Natural, but as Kasparov shows, he
missed an opportunity to win bril­
liantly here: 22 'iWaS+ lbbS 23 .l::lx fl ,
when 23 . . :iVc6 24 �dl ! ! 'iVxaS 25
:dS+ �c7 (25 . . . \t>b7 26 lihhS .l::le 6 27
lbxc5+ followed by lbxe6 and e7-eS'iY
27
Th e S e m i - Sla v
wins) 26 Jtf4+ Wb6 27 nxh8 l:te6 28
e8'i' l:txe8 29 �xe8 wins for White -
he will pick up the knight on b8, leav­
ing him with too much material for
the queen.
22 . . Jli'c6 23 ili'xc6+ 11xc6 24 %:td 1
!:te8 25 lbd6+ !:txd6 26 l:bd6 f6 27
�xf6?
As Kasparov points out, 27 Jte3, to
push the kingside pawns, was much
stronger, as with the minor pieces on,
Black's queenside majority has a much
harder task advancing.
27 . . .'3;c 7 28 !:te6 lbxf6 29 !:txf6
l:txe7 30 �f1 !:te4
Threatening the unpleasant . . . c4-c3.
31 11f4 l:txf4 32 gxf4 �d6 33 'iPe2
a5 34 a4 c3! 35 bxc3 b3! V2 - V2
After 36 '.t>d2 c4, White's king can­
not leave the queenside due to Black's
protected passed pawn, while White's
kingside pawns restrict his opponent's
king to the kingside after 37 h4 �e6
38 hS <it;fS 39 <it;c 1 <it;f6!
Game 9
K ramnik-Shirov
Monaco (blindfold) 1996
1 lbf3 d 5 2 d4 lbf6 3 c4 c6 4 lbc3
e6 5 �g5 dxc4 6 e4 b5 7 e5 h6 8
�h4 g 5 9 lbxg 5 hxg5 1 0 �xg5
lbbd7 1 1 g3 �b 7 1 2 �g2 �b6 1 3
exf6 0-0-0 1 4 0-0 c5 1 5 d 5 b4 1 6
%:tb 1 ili'a6 1 7 dxe6 �xg2 1 8 e7 �xf1
1 9 �xf 1 !
The only winning attempt. This
quiet move prevents the light-squared
bishop from causing a nuisance and
intends �ds.
1 9 . . . �c6!
An extremely fine move, activating
28
the black queen along the a8-h 1 di­
agonal while preventing White from
doing the same. 19 ... bxc3 20 bxc3!!
leaves Black helpless, as 20 .. :i'c6 (to
stop 'i'dS) is met by 2 1 l:ib8+! !, when
both 2 1 . . .ltJxb8 and 2 1 . . .<it;xb8 lose to
22 exd8'1W + !
The alternative 19 . . . Jtxe7 is consid­
ered in the next game.
20 exd8i1i'+ �xd8 21 lbd5 !:txh2! 22
�g 1 11h8
Here is where the mastery of such
sharp systems lies. First, it is necessary
to calculate the initial flurry of tactics;
but the real skill lies in playing the
extremely unbalanced positions - that
then arise. Material is now equal: the
f6-pawn is no longer an extra pawn,
and with . . . 1:.xh2, Black has eliminated
White's passed h-pawn. White's king
is rather weak as only the knight on
dS prevents Black from giving mate on
h 1 . However, Black also has his prob­
lems: his loose c4-pawn, inappropri­
ately centralised king and inactive
bishop.
23 �f4!?
This prevents the knight from mov­
ing to the dangerous eS-square.
23 . . . �d6? !
 B o t v in n ik Va ria tio n : M a in L in e w i t h 1 5
This solves the problem o f Black's
bishop, but the exchange on d6 dis­
tracts the black queen from the a8-h 1
diagonal, allowing White to unravel
and coordinate his pieces. In Nikolic­
Shirov, Linares 1997, Black tried to
improve with 23 . . . �e6 24 'ilVf3 'ilVh3 25
lld1 b3?! 26 a4 iYh2+ 27 �f1 iYh 1+ 28
We2 'iYxf3+ 29 Wxf3, but it was clear
that this was not the correct path!
23 . . . ..t>c8 ! is an improvement, intend­
ing to solve the more important prob­
lem of Black's king by transferring it
to b7, where it protects the queen on
c6. By stepping off the d-file, Black
also sets up the threat of . . . liJb6, re­
moving the knight from dS.
24 .i.xd6 �xd6 25 �f3 lLle5 26 �e4
l:te8 27 lLle3!
Highlighting the weakness of c4.
27 . . . lLle6 28 �f3 rl;e7 29 lLlxe4 �d4
30 b3 �e4 31 'ii'xe4 l:txe4 32 '1t>g2
lLle5 33 lLle3 'it>d7 34 l:th 1 rl;e6 35
l:th8 Wxf6 36 lIc8 lLld3 37 l:ta8 l:td4
38 l:txa7 rl;g6 39 f4 lLle 1 40 lLle4 f6
41 l:ta6 l:te4 42 Wf3 1:e2 43 lLle5+
�g7 44 l:ta7+ Wh6 45 lLld3 l:te2 46
lLlxe 1 l:txe 1 47 'it>e4 'it>g6 48 f5+ 'it>h5
49 l:th7+ 'it>g4 50 l:tg7+ rl;h3 51 Wd5
l:te2 52 l:tg6 l:txa2 53 l:txf6 l:tb2 54
l:th6+ 'it>xg3 55 f6 l:txb3 56 l:tg6+
�h4 57 f7 l:tf3 58 'it>e6 b3 59 l:tf6
�-�
We shall now examine another at­
tempt that Black has made in this line.
Game 10
Kamsky-Kramnik
l)os}{ennanas 1 996
1 d4 d5 2 e4 e6 3 lLle3 lLlf6 4 lLlf3
. . . b4
e6 5 .i.g5 dxe4 6 e4 b5 7 e5 h6 8
.i.h4 g5 9 lLlxg5 hxg5 1 0 .i.xg5
lLlbd7 1 1 exf6 .i.b 7 1 2 g3 e5 1 3 d 5
�b6 1 4 .i.g2 0-0-0 1 5 0-0 b 4 1 6
l:tb1 �a6 1 7 dxe6 .i.xg2 1 8 e7 .txf1
1 9 'it>xf 1 .i.xe 7 20 fxe 7 l:tdg8
Kramnik had already tried this sort
of idea against Kasparov after 19 'iVdS
(Game 8) . In annotating the present
game, he described it as dubious.
21 lLle4! �e6 22 lLld6+ 'it>b8 23 .tf4
l:txh2 24 'it>e2! 'it>a8 25 �e2 lLlb6 26
'ii'f 5 lLle8 27 e8'ii' l:txe8+ 28 lLlxe8
'i'xe8+ 29 .i.e3
Kramnik considers White to be
clearly better here. Resolute endgame
defence however, saves a valuable half­
point!
29 . . . lLlb6 30 l:td 1 rl;b7 31 �xe5 �a4
32 lId2 e3 33 bxe3 bxe3 34 l:td4
'ii'x a2+ 35 'it>f3 e2 36 l:td2 l:th8 37
l:txe2 'ii'd 5+ 38 'ii'x d5+ lLlxd5 39
lIb2+ lLlb6 40 �e4 l:te8 41 g4 l:te6
42 g5 l:te6+ 43 Wf5 a5 44 l:tb5 a4
45 l:ta5 lLle4 46 l:txa4 lLlxe3+ 47
fxe3 l:txe3 48 l:ta5 rl;e6 49 'it>f6 l:tf3+
50 'it>e7 l:tf 1 5 1 l:ta6+ rl;d5 52 lIf6
:g 1 53 1:f5+ rl;e4 54 '1t>f6 :g4 55
l:te5+ '1t>d4 � - �
29
Th e S e m i - Sla v

Summary
Against 16 ctJa4, 16 .. .'Yi'bS is my recommendation for Black players, meeting the
main line, 17 a3 exdS 18 axb4 cxb4 19 �e3 ctJcS 20 �g4+, with either 20 .. J l
and Ivanchuk's untried 20 . . . �c7. In fact, this is also my recommendation for
White players as my analysis shows there is plenty of scope for both sides!
16 l:Ib 1 'ilia6 17 dxe6 �xg2 1 8 e7 �xfl 19 'ilidS is best met by 19 . . . .i.xe7 20
fxe7 .td3! , as in Kramnik-Topalov, while 19 'it>xfl leads to a balanced position
after 19 ... �c6 20 exd8'iV + �xd8 2 1 ctJds lhh2 22 �gll:rh8 23 �f4 �c8! How­
ever, White should try this: although he is not better there is still plenty of play!

1 d4 d5 2 c4 c6 3 lLlf3 lLlf6 4 lLlc3 e6 5 �g5 dxc4 6 e4 b5 7 e5 h6 8 �h4

g5 9 lLlxg5 hxg5 1 0 �xg5 lLlbd7 1 1 exf6 �b 7 1 2 g3 c5 1 3 d5 �b6 1 4 �g2
0-0-0 1 5 0-0 b4
1 6 lLla4
16 lib 1 'ifa6 17 dxe6 .i.xg2 18 e7 .txfl (D)
19 'iVdS .ixe7 20 fxe7
20 ... �d3 - Game 7; 20 .. J�dg8 - Game 8
19 'it>xfl
19 .. .'iVc6 - Game 9; 19 . . . .txe7 - Game 10
1 6 . . :it'b5
16 .. .'ii' a6 - Game 6
1 7 a3 (D)
17 dxe6 - Game 5
1 7 . . . exd5
17 . . . ctJeS - Game 4
1 8 axb4 cxb4 1 9 �e3
19 lite 1 - Game 3
1 9 . . . lLlc5 20 'it'g4+ J:td7 2 1 'it'g7 �xg7 22 fxg7 :g8 23 lLlxc5 (D) d4
23 . . . l:lxg7 - Game 2
24 �xb7+ - Game 1

1 7 a3 23 lLlxc5

30
 CHAPTER TWO

Botvi nnik Va riation : Black's

1 3th Move Alternatives

1 d4 d5 2 c4 c6 3 liJf3 liJf6 4 liJc3 e6 5 jLg5 dxc4 6 e4 b5 7 e5 h6 8

e6 5 �g5 dxc4 6 e4 b5 7 e5 h6 8 �h4 g5 9 liJxg5 hxg5 1 0 �xg5
1Lh4 g5 9 liJxg5 hxg5 1 0 �xg5 liJbd7 1 1 exf6 �b7 1 2 g3 c5 1 3 d5
liJbd7 1 1 exf6 �b7 1 2 g3 c5 1 3 d5 liJxf6
In this chapter, we analyse Black's
alternatives to the main line. The
popularity of these lines peaked in
1993 when all the leading Semi-Slav
players played one of these- as their
main weapon. My feeling is that they
renounced these ideas not because of
tactical problems, but because the
positional risks that Black takes be­
come less attractive once White play­
ers developed a good understanding of
the appropriate tactical motifs.
The first three games consider 13 . . .ttJxf6 re-establishes material
13 . . . ttJxf6, restoring material parity, equality by capturing White's extra
Games 14-16 deal with 13 . . . �h6 and pawn on f6, and coordinates substan­
Games 17 and 1 8 with 13 . . . ttJb6 and tial pressure against White's d-pawn.
13 . . . ttJe5 respectively. The move's boldness lies in the fact
that it reactivates the pin on the f6-
Game 11 knight, thereby keeping the black
Van Wely-Dreev queen tied to the knight's protection
Bern Open 1993 and thus preventing Black from mov­
ing the queen to prepare queenside
1 d4 d5 2 c4 c6 3 liJc3 liJf6 4 liJf3 castling. Finally, the b5-pawn is still

31
Th e S e m i- Sla v
hanging which gives Black another
loose point to worry about.
1 4 �g2 jLe7
Breaking the pin on the knight on
f6 along the h4-dS diagonal and
thereby increasing the pressure on d5.
The alternative 14 . . . i.h6!? is consid­
ered in Game 13.
1 S 0-0
This puts the king to safety and
protects the bishop on g2, breaking
the pin on the d5-pawn.
1 S . . . lLJxdS 1 6 �xe 7 ctJxe 7 1 7 lLJxbS
Although material is equal and
Black has a large number of files for
his major pieces, his pawn structure is
very weak and his king is rather
draughty in comparison to White's.
Black has to play actively or his posi­
tional weaknesses will cost him the
game.
1 7 . . :�b6 1 S lLJa3
Attacking c4.
1 S . . . c3
The bizarre 1 s . . .11h4 is considered
in the next game.
1 9 lLJc4 �c7 20 bxc3! lLJxc3 2 1 �d2
�xg2 22 �xg2
Wells also suggests 22 'iWg5+!?
22 lLJe4 23 -.e3 lLJf6 24 �f3
. • •
32
After this move, taking control of
the h 1-aS diagonal, White stands
slightly better due to his better pawn
structure.
24 . . J:tabS 2S l::t a b1
Van Wely suggests 25 .:tfe 1 l:Ih5 26
l:!adl followed by h2-h4. The ending,
though not objectively losing, is of
course not very inspiring for Black _
and Van Wely makes it look much
easier for White to win than Black to
hold!
2S . . . lLJd7 26 h4 J:tb4 27 J:tfc 1 J:thb8
28 J:tb3 J:txb3 29 axb3 �b7 30 -.xb7
J:txb7 31 J:ta 1 lLJb6 32 J:taS ! :c7 33
g4 �f8 34 hS ctJg7 3S lLJd6 J:td7 36
lLJe8+ ctJh6 37 lLJf6 :b7 38 l::t x cS
lLJdS 39 lLJxdS exdS 40 J:txdS l::tx b3
41 :as a6 42 l::t x a6+ ctJgS 43 f3 �h4
44 J:tf6 1 -0
Game 12
Van Wely-Kramnik
Biel lnterzonal 1993
1 d4 dS 2 lLJf3 c6 3 c4 lLJf6 4 lLJc3
e6 S �gS dxc4 6 e4 bS 7 eS h6 8
�h4 gS 9 lLJxgS hxgS 1 0 �xgS
lLJbd7 1 1 exf6 �b7 1 2 g3 cS 1 3 dS
lLJxf6 14 �g2 �e 7 1 S 0-0 lLJxdS 1 6
 B o t vin n ik Va ria tio n : B la c k 's 1 3 th M o v e A l t e rn a t i v e s

�xe7 rJixe7 1 7 lLlxb5 �b6 1 8 lLla3 36 l:td3! l:tg4 37 lLld2 a5 38 l:te3

1::t h4! ? ! ? rJid6 3 9 h3 l:tg8 40 lLle4+ rJie7 41
l:tb3 l:td8 42 l:tb5 l:td4 43 l:txe5 lLlb4
44 a3 1 -0

Game l3
Kasparov-Ivanchuk
Linares 1 994

1 d4 lLlf6 2 e4 e6 3 lLle3 d 5 4 lLlf3

e6 5 .tg5 dxe4 6 e4 b5 7 e5 h6 8
.th4 g5 9 lLlxg5 hxg5 1 0 .txg5
lLlbd7 1 1 exf6 .tb7 1 2 g3 e5 1 3 d5
lLlxf6 14 .tg2 .th6 ! ?
A quite stunning move! Kramnik
makes the maximum use of the open
files created by Black's fractured pawn
structure.
1 9 'i'd2!
A masterful and strong reply. 19
gxh4 11g8! wins for Black.
1 9 . . J:td4?
In a later round of the same tour­
nament, T opalov tried the astounding
19 . . .lbf4 against 011 but was rapidly
defeated after 20 lbxc4 'i'c7 2 1 f3!
lbxg2 22 'i'g5+ wf8 23 gxh4. In re­ With this move Black seeks t o ex­
sponse to 20 . . . 'ii' a6 011 recommended change the dark-squared bishops and,
21 lbe3 as slightly better for White, by breaking the pin on the knight on
but in Ehlvest-Onischuk, Philadelphia f6, to increase the pressure on White's
Open 1994, White preferred 2 1 i.xb7 d5-pawn.
lbh3+ 22 �g2 'ii'x b7+ 23 f3 lid8 24 1 5 .txf6
'i'c3, when 24 .. .ihc4 25 'ii'xc4 'ii'x b2+ This removes the knight on f6 and
gives Black sufficient play for the deflects the black queen to f6, thus
draw. There may still be life in this removing a great deal of pressure from
line yet! the d5-pawn. 15 .ixh6 �xh6 16 0-0
20 'i'g5+ lLlf6 2 1 .txb7 l:tg8 22 'it"e5 i.xd5 was fine for Black in Sher­
lLld7 23 "e2 'i'xb7 24 lLlxe4 l:1h4? ! bakov-Korneev, Elista 1996, but 15
25 f3 'i'e7 26 tOe3 �f8 27 lLlg4 l:th5 i.h4!? is worth a thought.
28 l:tad 1 l:td5 29 'iie4 'iid 6 30 lLle3 1 5 . . :ii'xf6 1 6 0-0
l:td4 31 l:txd4 'iix d4 32 l:td 1 'i'xe4 33 16 lbxb5 is less good according to
fxe4 �e 7 34 lLle4 lLlb8 35 rJif2 tOe6 Kasparov, as 16 . . . 'iie 5+ 17 �e2 �xe2+

33
Th e S e m i - Sla v
(17 . . . �d2+ 18 \tfl) 18 Wxe2 0-0-0 19
ctJxa7+ \tb8 20 ctJc6+ (20 ctJbS seems
worth a try) 20 . . . iLxc6 2 1 dxc6 :d2+
gives Black counterplay. 16 ctJe4!? is
also worthy of attention.
With 16 0-0, White not only puts
his king to safety, but also protects the
bishop on g2, thus unpinning the
pawn on dS and threatening dSxe6.
Both sides have several dynamic fac­
tors in their favour: White has a
passed h-pawn and is threatening
ctJxbS, winning the b-pawn; whereas
Black has the two bishops and two
open files against the white king - the
g- and h-files.
1 6 . . . 0-0-0 1 7 tLJxb5 exd5 1 8 tLJxa7+!
'it>b8 1 9 tLJb5
Although White is a pawn up,
Black's position looks quite promising
- he has the two bishops, a trio of cen­
tral pawns, and two open files against
White's king. However, the removal
of the a7-pawn has not only weakened
Black's king position but also pro­
vided a secure outpost for the white
knight on bS, from where it eyes a7
and c7. This spells danger for Black's
king: if White can reach as with his
queen both 'ii' a7+ and 'iWc7+ will be
dangerous. Furthermore, the bishop
on h6 blocks the h-file and this ob­
structs Black's plan of doubling on the
h-file and attacking h2. He will have
to waste time redirecting this bishop
before he can strike at White's posi­
tion.
1 9 . . . .ltg7
Komljenovic-Lupu, Andorra Open
1994, continued 19 ... .tgS 20 f4 �h6!
2 1 fxgS 'ii'x h2+ 22 wf2 d4!
(threatening . . . 'ii'xg2+) 23 l:tg1 l:th4!!
34
A sensational move which threatens
to expose the white king still further
with . . . l:tf4+! and also has the threat of
. . . �e8 , increasing the strength of .. J�f4
by cutting the king from its flight
squares on the e-file. 24 gxh4 fails to
24 ... 'iVxh4+! 25 wfl (25 We2 Ite8+ 26
'it>d2 'ii'f2+) 2S . . . 'iVf4+ 26 �e 1 �e8+ 27
�e2 llxe2+ 28 Wxe2 �e3+ 29 \tf1 c3!?
30 bxc3 d3! 31 �e 1 �f4+! However,
we should borrow from Kasparov and
instead of 20 f4 play 20 �e 1 , eyeing as
and eS: 20 . . . �h6 2 1 �eS+ \ta8 22 h3!,
intending 'iWc7, looks rather unpleas­
ant for Black.
20 a4
20 'iNe 1 immediately was also possi­
ble, as after 20 . . . 1hh2 2 1 �xh2 l:th8+
22 \t;gl 'ii'h6 23 f3 'i'h2+ 24 �f2 l:[h3
25 �e8+ .i.c8 26 Iig1 Black must still
prove that he has enough for the sacri­
ficed rook.
20 . :ii'h6? !
.
Kasparov suggests a different line up
on the h-file with 20 .. J:th6, intending
2 1 . . J:tdh8.
21 h4 .ltf6 22 "ii'e 1 !
The culmination of White's open­
ing plan. It is impressive how Kas­
parov managed to find such a strong
 B o t v in n ik Va ria tio n : B la c k 's 1 3 th M o v e A l t e rn a t i v e s
method of coordinating his pieces
over the board: giving up the bishop
pair, using his queen's knight to grab a
pawn on a7 at great cost in time and
anticipating that this would give him
such a strong attacking set-up on the
queenside. It shows visionary attack­
ing talent!
22 . . . �xh4 23 �a5
Strong, but 23 �eS+ �aS 24 �c7
would have been even better accord­
ing to Kasparov.
23 . . . �e7 24 �c7+ �a8 25 �a5+
'it>b8 26 �c7+ �a8 27 Ufe 1 �d6 28
'i'b6 �b8 29 a5 %:td7 30 %:te8 !
30 . . :i'h2+ 3 1 �f1 �xg2+ 32 �xg2
d4+ 33 �xb7+ %:txb7 34 %:txh8 %:txb5
35 a6 �a7 36 %:tf8 l:txb2 37 %:txf7+
�a8 38 a7 c3 39 %:tf8 1 -0
A magnificent game.
Game 14
Bareev-Filippov
Russia 1995
1 d4 d5 2 c4 c6 3 lLlf3 lLlf6 4 lLlc3
e6 5 �g5 dxc4 6 e4 b5 7 e5 h6 8
�h4 g5 9 lLlxg5 hxg5 1 0 �xg5
lLlbd7 1 1 exf6 �b7 1 2 g3 c5 1 3 d5
�h6
1 4 �xh6 %:txh6
By forcing the exchange of the dark­
squared bishops, Black removes
White's protection of f6. He intends
to capture on f6 with the knight,
without allowing the pin that arises in
the 13 ...lDxf6 line. Moreover, Black
frees fS for his king, which allows him
to consider . . . e6xdS since a check on
the e-file is no longer devastating.
However, the exchange of the dark­
squared bishops costs Black some con­
trol over the central dark squares; d6
in particular is a tempting target for
White's queen's knight, either via e4
or via the unprotected pawn on bS.
Finally, 14 i.. xh6 l:Ixh6 draws Black's
king's rook to an exposed position and
White can use this to gain tempi for
his development.
1 5 �d2
This move attacks the exposed rook
on h6 and achieves several positional
ideas with tempo. First, it tempts
Black to recapture the sacrificed pawn
on f6 with the queen or with the
rook, thus depriving the knight on d7
of its ideal square. Second, it covers
the e I-aS diagonal and enables lDxbS
without fear of . . . 'iWaS+. Finally, it
35
Th e S e m i - Sla v
prepares rapid queenside castling.
1 S . . . �xf6 1 6 0-0-0 rJi>f8
Now that White has castled his king
to safety on the queenside, ttJe4, at­
tacking the queen and threatening
ttJd6+ forking king and bishop, was a
big threat and ttJxb5, threatening
ttJd6+ or ttJc7+, is also possible. After
16 . . .'it>f8 , neither ttJd6 nor ttJc7 will
come with check and this reduces the
power of ttJe4 and ttJxb5. Moreover,
by removing the king from the e-file,
Black makes it safe to take on d5 and
open the e-file.
16 . . . 0-0-0, for example, would have
led to disaster: 17 ttJe4! 'ii'e 5 1 8 "1!fxh6
'i'xe4 19 dxe6 'i'xh l 20 .i.h3! as in
Quist-Thiesing, Dieren Open 1988.
1 7 f3! ?
This blocks the hl-a8 diagonal,
breaking the pin on the d5-pawn (18
dxe6 is now the threat), and covers e4,
supporting the white queen's knight
when it arrives there. Its drawback is
that it is now hard to find an active
square for the white light-squared
bishop - it is restricted by the b5- and
c4-pawns and f2-f3 blocks the hl-a8
and e2-h5 diagonals, while the h3-c8
diagonal is covered by the rook on h6.
36
17 h4 is discussed in the next game,
while 17 f4 is met by 17 . . . ttJb6, with
tremendous pressure on the white d­
pawn.
1 7 . . . exdS 1 8 lLlxdS 'ir'd6 1 9 �gS ! ?
2 0 ttJc7 i s a threat now: the knight
on c7 attacks the rook on a8 and
19 .. :�i'xc7 is impossible as the rook on
h6 hangs.
1 9 . . . 1:[g6?
White's last move is unpleasant to
face unprepared over the board and
here Filippov goes astray immediately.
A little later, however, with the bene­
fit of home preparation, he found the
splendid 19 . . . �b8!!, dealing with the
threat of ttJc7 by calmly placing the
rook on a square where the knight
cannot attack it. Now any knight
move is met by .. :iff6, attacking the
unprotected white queen, when 'ii'xf6
ttJxf6 saves the attacked black knight
on d7! 19 . . . 1:.b8 also protects the loose
bishop on b7, which is always useful.
It may seem a little strange that Black
can just play a quiet consolidating
move in the middle of a tactical battle,
but I think that this is mainly due to
the drawbacks of 17 f3: although it
performed a useful function, it did not
bring any pieces into action; indeed if
anything, it made the bishop on fl
more passIve.
20 "fS! lLleS 21 lLlf4 l:U6 22 "h3!
22 "1!fh5 would fail to the rather
cruel 22 .. :i'ixdl+! 23 'itxdl .i.xf3+!
22 .. JWxd 1 + 23 �xd 1 iLxf3+ 24 �c 1
iLxh 1 2S -.h4 lLld7 26 iLh3 l:td6 27
iLxd7 l:txd7 28 "*h8+ rJi;e7 29 'ir'eS+
�d8 30 h4 b4 31 hS b3 32 axb3
cxb3 33 h6?
The crucial mistake. 33 'i'c3, pre-
 B o t vin n ik Va ria tio n : B la c k 's 1 3 th M o v e A l t e rn a ti v e s
venting Black's next, would have
given good winning chances (Bareev) .
33 . . . .Jtf3 34 ttJd5 �xd5 35 'it'f6+
<l;; c 7 36 'it'xf3 '1Pc6 37 h7 %:e8 38
'i'xb3 �e 1 + 39 'it'c2 �e2+ 40 'it'b1
:h2 41 �a4+ �b6 42 �b3+ % - %
Game 15
Razuvaev-Filippov
Russian Championship 1995
1 d4 d5 2 c4 c6 3 ttJf3 ttJf6 4 ttJc3
e6 5 .Jtg5 dxc4 6 e4 b5 7 e5 h6 8
.Jth4 g 5 9 ttJxg5 hxg5 1 0 .Jtxg5
lLlbd7 1 1 exf6 �b7 1 2 g3 c5 1 3 d5
�h6 14 .Jtxh6 lbh6 1 5 'it'd2 �xf6
1 6 0-0-0 '1Pf8 1 7 h4! ?
This interesting move aims to make
use of White's kingside pawns to at­
tack Black's exposed and hence vul­
nerable major pieces on f6 and h6.
1 7 . . . ttJb6 1 8 g4
Threatening to fork the queen on f6
and the rook on h6 with g4-gS.
1 8 . . . ttJxd 5 ! ?
18 .. .l:hh4 19 g S 'ii f4 2 0 �xh4 'iixh4
21 dxe6 and 1 8 . . . l:.h8 19 gS 'iVg7 20
lLlxbS are both good for White accord­
ing to Razuvaev.
1 9 g5 �g6
Now Black is ready to meet 20 gxh6
with 20 ... ltJxc3 and 2 1 . . . .txh 1 .
2 0 ttJxb5 ttJb4
Threatening ... ltJxa2 mate.
21 ttJc3 �xh 1 22 gxh6
Razuvaev considers this position to
be slightly better for White and I see
no reason to disagree with his assess­
ment.
22 . . . .Jtf3?
22 . . . .tdS was necessary according to
Razuvaev.
23 %:e 1 .Jtd5 24 a3 a5 25 %:e5 �g 1
Now White elected to repeat with
26 1:e1 ? 'it'g6 % - %
This is quite a surprising draw as in
fact White has a forced win! Razuvaev
realised after the game that the not so
difficult 26 h7! 'ii'x f1+ 27 l1e 1 1tJd3+ 28
'5t>b 1 would have won for White.
Thi3' line with 17 h4 is definitely the
most dangerous for Black to face at
the moment and he needs an im­
provement in order to play this line
with confidence.
Game 16
lonov-Popov
St Petersburg Open 1995
1 d4 d5 2 c4 c6 3 ttJf3 ttJf6 4 ttJc3
e6 5 .Jtg5 dxc4 6 e4 b5 7 e5 h6 8
�h4 g5 9 ttJxg5 hxg5 1 0 .Jtxg5
ttJbd7 1 1 exf6 .Jtb7 1 2 g3 c5 1 3 d5
�h6 14 �xh6 %:xh6 1 5 'it'd2 'it'xf6
1 6 ttJe4!?
This is the most direct attempt to
refute 1s .. J W
attacks the queen on f6 and aims for
the d6-square. Peter Wells was panicu­
larly keen on this move in The Com·
plete Semi-Slav, but it has only recently
37
Th e S e m i - Sla v
had its first practical test.
1 6 . . :�f3 1 7 lDd6+ �e7 1 s lDxb7
18 �gl St.xd5 and 18 . . . 'ii'xd5 19
'iixh6 'iix d6 give Black sufficient
compensation according to Yusupov.
1 S . . J:thS ! ?
This move, dealing both with the
threat to the rook and the annoying
idea of �g5+, was initially suggested
by Artur Yusupov. 1 8 . . :ii'xhl 19 d6+!
'.t>e8 20 'iWxh6 �b7 2 1 'iWh4! is given as
winning for White by a certain
Hamovic in ECO.
1 9 d6+ �eS 20 :g 1 e3! ?
Ionov mentions 2 0 . . . l:te5+ 2 1 i.e2
'iix b7 22 0-0-0 and 20 . . .:xh2 21 'i'f4!
'iWxb7 22 0-0-0 as unclear.
21 bxe3 l::t e S+ 22 i.e2 l::tx e2+ 23
�xe2 1t'xe3+ 24 �f1 �xa 1 + 25 c.t>g2
�d4? ! 26 :td 1 �a4 27 �f3?
27 'ii h 5! would have been good for
White according to Ionov.
27 . . . :teS 2S 'iVh5 !
White gets a second chance!
2S . . . �e4+ 29 c.t>g 1 'ifxb7 30 �hS+
lDfS 31 d7+ �xd7 32 l:lxd7 c.t>xd7 33
�f1 a5 34 h4 b4 35 c.t>e2 a4 36 hS
c.t>e7 37 h6 :tdS 3S 'WWe S :teS 39 �e4
A bad mistake. 39 'ikg5+ '.t>d6 40
'iWf6 would have kept White's advan-
38
tage according to Ionov.
39 . . . lDd7 40 h7 :thS 41 �h4+ f6 42
'WWh S a3 43 c.t>d2 b3 44 axb3 a2 45
'ifh 1 �d6 46 �a 1 l::t x h7 47 'iix a2
l::t h 5 4S 'iVa6+ �e7 49 �b7 l::t d 5+ 50
�e3 :td4 51 �e7 l::t b4 52 �e6 V2 - V2
The next game is the classic Botvin­
nik system game. Theoretically, it is
still the latest word on 13 . . . lUb6 and
aesthetically, it never ceases to amaze!
Game 17
Polugayevsky-Torre
Moscow 1981
1 d4 d5 2 e4 e6 3 lDf3 lDf6 4 lDe3
e6 5 i.g5 dxe4 6 e4 b5 7 e5 h6 S
i.h4 g5 9 lDxg5 hxgS 1 0 i.xg5
lDbd7 1 1 exf6 i.b7 1 2 g3 eS 1 3 d5
lDb6
1 4 dxe6! 'iix d 1 +
Lukacs's 1 4 . . . St.xh l 1 5 e7! 'ii'd7
(15 . . . 'ii'xdl+ 16 �xdl transposes to the
main game) does not seem good after
16 'ikxd7+ (16 exf8'ii + 'ittxf8 17 'iWxd7
lUxd7 1 8 0-0-0 i.c6 19 h4 lUe5 20 i.e3!
is also good for White according to
Wells) 16 . . . 'ittxd7 {16 . . . lUxd7 17 lUxb5
i.xe7 18 fxe7 f6 19 i.e3 'ittxe7 20 h4
i.f3 21 i.xc4 �hc8 22 �c 1 lUe5 and
 B o t vin n ik Va ria tio n : B la c k 's 1 3 th M o v e A l t e rn a t i v e s
now 23 b3, intending i.f1 to attack cS,
would have given White a clear advan­
tage in the game lonov-Sherbakov,
Rostov on Don Open 1993, according
to lonov) 17 l::! d 1+ �c6 18 ,Ud8 is very
good for White according to Wells.
1 S l:txd 1 �xh 1 1 6 e7! a6 1 7 h4! !
.ih6 1 8 f4! !
Polugayevsky comments that 'Ha;­
ing given up a rook, White has no in­
tention of regaining the lost material,
but contents himself with the fact that
the rook on h8 is not destined to come
into play for some time.'
18 ... b4 1 9 lId6 l:tb8 20 lbd 1 �xgS
21 fxgS lbdS 22 .lixc4 lbxe7 23 fxe7
�xe7 24 l:tf6 l:thf8 2S lbe3 .lie4 26
l:txa6 l:tbd8 27 l:tf6 l:td6 28 l::tf4 l:td4
29 hS .lid3 30 lbdS+ '1td6 3 1 l::t xd4
cxd4 32 .ib3 .ic2 33 .ixc2 'it>xdS
34 .ib3+?
Alas, it won't be a perfect game
now! 34 h6 followed by h6-h7 would
have won easily.
34 . . . �eS 3S g4 �f4
Black's only chance was 3S . . . d3!
36 g6 '1te3 37 g7 l:tc8 38 �f 1 d3 39
'iPg2 �f4 40 h6 1 -0
Finally, 13 . . . ltJe5.
Game 18
Beliavsky-lIIescas
Linares 1994
1 d4 dS 2 c4 c6 3 lbf3 lbf6 4 lbc3
e6 S .ligS dxc4 6 e4 bS 7 eS h6 8
.ih4 gS 9 lbxgS hxgS 1 0 .lixgS
lbbd7 1 1 exf6 .ib7 1 2 g3 c5 1 3 d5
lbeS ! ?
13 . . .b 4 loses t o 14 ..txc4 bxc3 15
dxe6. The text aims for d3 but releases
pressure on the f6-pawn.
1 4 .lig2 liJd3+ 1 5 �f1 �d7 1 6 dxe6
Natural but not the best. Vujatovic
suggests 16 'ii' f3 (threatening dSxe6),
when 16 . . . exdS 17 h4!! prevents
queenside castling due to 17 . . . 0-0-0 18
i.h3! Apparently this has received the
Royal seal of approval from Gazza
himself, so it is probably best!
1 6 . . . fxe6 1 7 b3! ? 0-0-0 1 8 bxc4
.lih6! 1 9 .ih4! b4 20 lbdS! exdS 2 1
�xd3 dxc4 2 2 .ixb7+! �xb7 23
'i'fS+ 'it>b8 24 1:[g 1 ! .lid2!?
2 4 . . . IId5 2 5 'i'e6 c 3 2 6 f7 'ifbS+ 27
'it>g2 'i'b7 with a draw (Illescas) .
2S f7 l::t c 8 26 'i'eS+? ! �a8 27 �e7?
.ic3 28 'i'd6 .lixa 1 29 f8'i' l::t h xf8 30
.ixf8 b3 0-1
39
Th e S e m i - Sla v

S u mmary
All these systems are worth a go on an occasional basis, but as main defences
they seem a little too risky to place too much reliance on. 13 . . . tDxf6 14 .ig2
.th6!? looks interesting, while 13 ... .th6 continues to survive. However,
13 . . . tDb6 seems to be pretty much busted and 13 .. .'�Je5 devotees need an answer
to 16 'iff3 .

1 d4 d 5 2 c4 c6 3 lLlf3 lLlf6 4 lLlc3 e6 5 �g5 dxc4 6 e4 b5 7 e5 h6 8 �h4

g5 9 lLlxg5 hxg5 1 0 �xg5 lLlbd7 1 1 exf6 �b7 1 2 g3 c5 1 3 d5

1 3 . . . lLlxf6
13 . . . .th6 14 .ixh6 �xh6 15 'ilVd2 'i'xf6 (D)
1 6 0-0-0 �f8
17 f3 Game 14
-

17 h4 Game 15
-

1 6 tDe4 Game 16
-

1 3 ... tDb6 Game 1 7

-

13 . . . tDe5 Game 1 8
-

1 4 �g2 �e7 1 5 0-0 (D) lLlxd 5

14 . . . .ih6 Game 13
-

1 6 �xe7 �xe7 1 7 lLlxb5 �b6 1 8 lLla3 (D) c3

18 . . . l::th4 Game 12
-

1 9 lLlc4 Game 11
-

15 .. .'flxf6 1 5 0-0 1 8 lLla 3

40
 CHAPTER THREE
Botvin nik Variation
with 1 1 9 3
1 d4 d5 2 c4 c 6 3 lbf3 lbf6 4 lbc3
e6 5 i.g5 dxc4 6 e4 b5 7 e5 h6 8
.ih4 g5 9 lbxg5 hxg5 1 0 �xg5
tLlbd7 1 1 g3
As the alert reader will have spot­
ted, Ivanchuk and Kasparov always
prefer the 'standard' move order of 1 1
exf6. However, Kramnik has consis­
tently chosen to play 1 1 g3 and in this
chapter we shall look at the differences
that this move order makes. The first
point to note, however, is that 1 1 g3
.tb7 12 �g2 'i'b6 13 exf6 0-0-0 14 0-0
c5 15 d5 b4 simply takes us back to
the main line position discussed in the
first chapter.
With 1 1 g3 White delays the cap­
ture of the pinned knight in order to
develop his bishop on the hl-a8 di­
agonal one move earlier than usual.
Question 1: Why is this important?
Answer: After 1 1 exf6 i.b7 12 g3,
Black can play 12 . . . c5 opening the at­
tack of the bishop on b7 on the rook
on h I . However, after 1 1 g3 .tb7 12
.tg2, 12 ... c5 is impossible as the
bishop is en prise on b7. By develop-
ing the bishop early to the long diago­
nal, White makes it harder for Black
to achieve . . . c6-c5.
Question 2: What's the downside?
Answer: By cutting down on some
options, you give Black other possi­
bilities. The main alternative is
I l . . J;rg8 to break the pin by sacrificing
the rook for the bishop on g5.
If Black wants to play the main line
with 13 . . . 'i'b6 against 1 1 exf6, then he
can transpose back into Chapter 1 by
following Timman-Tal. However, 1 1
g3 avoids all the main line alternatives
in Chapter 2, so Black advocates of
these variations will need a separate
line against 1 1 g3 .
Game 19
Timman-Tal
Hilversum (match) 1988
1 d4 d5 2 c4 c6 3 lbf3 lbf6 4 lbc3
e6 5 �g5 dxc4 6 e4 b5 7 e5 h6 8
�h4 g5 9 lbxg5 hxg5 1 0 �xg5
lbbd7 1 1 g3 �b7
Threatening the typical . . . c6-c5
41
Th e S e m i - Sla v
counter-thrust.
1 2 �g2 �b6 1 3 exf6 0-0-0 1 4 0-0
cS!
After due preparation, Black has
achieved the freeing . . . c6-c5 break.
Here 15 d5 b4 would transpose to
Chapter 1 . White's only independent
possibility is seen in this game.
The risky 14 . . . 4Je5 is considered in
Game 20.
1 S dxcS
After this move White keeps his ex­
tra pawn. However, in so doing he
allows his opponent to weaken his
kingside light squares by exchanging
the light-squared bishops, opens the d­
file for Black's rook on dS and draws
the black knight to the powerful d3-
outpost via c5.
1 S . . . ttJxcS ! 1 6 'iVe2 �xg2 1 7 'itxg2
�h6
Although Black has weakened his
opponent's kingside light squares, he is
not set for an all-out kingside attack.
Black's strength is on the queenside -
the pawns on b5 and c4 give Black a
queenside space advantage and a
strong outpost on d3 . His queenside
majority is more mobile and much
more potent than White's kingside
42
maJonty, although this will only
really come into play in an endgame.
Black must first solve the problem of
his undeveloped king's bishop on f8
and win back the sacrificed f6-pawn.
Then he can exchange some pieces -
the queen and a pair of rooks - and
utilise his long-term endgame
strengths. 17 . . . St.h6 seeks to exchange
the dark-squared bishops in order to
weaken White's hold on the f6-pawn.
1 8 h4
Not the best. Tal's suggestion of 18
�xh6 l;Ixh6 19 'W f3 sets interesting
problems: 19 . . . a6! is the best reply, to
follow up with . . . 'iib 7 to challenge for
the as-h 1 diagonal.
1 8 . . . .txgS 1 9 hxgS 'iVc6+! 20 f3
Forced.
20 . . J:thS !
Now that f2-f3 has closed the d1-h5
diagonal to the white queen, Black can
play this sneaky move, attacking the
pawn on g5 and threatening on occa­
sion to double on the h-file.
21 a4
This move, undermining the black
queenside, is White's only hope for
counterplay.
21 . . . b4 22 lLlbS

Threatening a fork on a7. Tal
points out that 22 . . :ifd5! was the
strongest here, meeting 23 �fd 1 with
23 . . . lbd3 .
Tal continues: 24 'iVe3 'iVxg5!
(threatening 25 . . :i!i'xg3+!! 26 'it>xg3
lig8+ 27 'ilJ'g5 llg8+ mate!) 25 'Yi'e4
'i'd5! 26 �xd3 ! cxd3 27 'ii'x b4 d2!
(preventing l::tc 1+) 28 lbxa7+ (28 .l::t a3,
threatening 11c3+, is countered by
28 . . . d1lb!, covering the c3-square!)
28 .. .'�c7 29 lbb5+ Wb8! 30 lbc3+ 'ii'b 7
31 'i'xb7+ 'it>xb7, which is assessed as
unclear by Tal, Krnic and Velickovic
but looks rather nice for Black to me!
It therefore seems that this line poses
few problems for Black.
22 . . .'l;b7 ? ! 23 l:tad 1 lLld3 24 'iVe3
'i'b6 25 'iVe4+ 'it>b8?
A critical mistake after which
Timman's peerless play wraps up the
game for White. 25 .. :iVc6 26 g6 a6 27
g7 axb5 28 axb5 'ifNxe4 29 fxe4 �g5
(Tal) would have led to great compli­
cations.
26 g4 l:th4 27 �g3 ! ! l:tdh8 28 ltxd3!
cxd3 29 'iVe5+ �a8 30 'iVe4+ 'it>b8
31 'iVe5+ 'it>a8 32 g6 l:th3+ 33 ..t>f4
a6 34 ite4+ 'it>b8 35 ite5+ �a8 36
'iVe4+ 'it>b8 37 'iVd4! �xd4+ 38 lLlxd4
B o t vin n ik Va ria tio n w i t h 1 1 9 3
fxg6 3 9 l:td 1 1 -0
White has too many pawns for the
exchange.
This game is all that the main line
Black player needs to know in order
to meet 1 1 g3 with confidence.
Game 20
Yermolinsky-Kaidanov
USA Championship 1993
1 d4 d5 2 c4 e6 3 lLlf3 lLlf6 4 lLlc3
c6 5 i.g5 dxc4 6 e4 b5 7 e5 h6 8
i.h4 g5 9 lLlxg5 hxg5 1 0 i.xg5
lLlbd7 1 1 exf6 i.b 7 1 2 g3 �b6 1 3
i.g2 0-0-0 1 4 0-0 lLle5
This risky move seeks to exploit the
pin on the d-pawn by the rook on d8 .
However, White has a radical solution
to the problem.
1 5 dxe5! l:txd 1 1 6 l:taxd 1
White has a rook, knight and pawn
for the queen. Black's main chances lie
in his 4-2 queens ide majority, but
White should be able to neutalise this
with accurate play.
1 6 . . . b4 1 7 lLle4 ita5
Attacking e5.
1 8 i.f4 c3 1 9 bxc3 i.a6! 20 cxb4
i.xb4 2 1 a3!
43
Th e S e m i - Sla v
Diverting the dark-squared bishop
from b4, thus allowing the rook on fl
to move to e l .
2 1 . . . �xa3
2 1 . . .'ii'x a3 22 l:ial .Jtxfl 23 l:txa3
�xg2 24 �xa7 �b8 25 �a4 (defending
e4) wins for White according to Yer­
molinsky and Shabalov.
22 1:[fe 1 �d3 23 ltJd6+ �xd6 24
exd6 ¢>d7 25 l:te5 �b5 26 l:te1 it'a4
27 l:tee5
Our two annotators assess this posi­
tion as slightly better for White.
27 . . . a6 28 �e5 ? !
28 h 4 allows Black the annoying
28 . . . litxh4 29 gxh4 'iix f4, breaking up
the white kingside, so Kaidanov sug­
gests the interesting 28 h3, patiently
giving the white king an escape hole.
28 . . J:th5 29 f4 �a3 30 l::t 5 e3 it'b4
3 1 h4 l:txe5 32 fxe5 'ifd4+ 33 ..t>h2
"xe5 34 �xe6+ .ltxe6 35 l:txe6
�xf6 36 l:t6e2 ¢>xd6 37 :ta2 ¢>e7 38
:le6 iYf1 39 l::t e xa6 ..t>f8 40 l::t b 6 �g7
41 l:tbb2 �g6 42 l::t f 2 �e 1 43 l::t a e2
it'd 1 ??
The decisive error which allows
White to manoeuvre his rooks into
the ideal position to attack the black
pawns and defend his own king.
44 :e5 �g7 45 l::t g 5+ �f8 46 l::t g 6! !
e5 47 :gf6 e4 48 l:txf7+ ¢>g8 49
l:te7 �d3 50 :f4 e3 5 1 :fe4 �e2+
52 ¢>h3 "e8+ 53 g4 "e 1 54 :lxe3
"h 1 + 55 ¢>g3 it'g 1 + 56 ¢>f4 iYf2+
57 �g5 �d2 58 �g6 1 -0
Game 2 1
Shirov-Morovic
Las Palmas 1994
1 d4 d 5 2 e4 e6 3 ltJe3 ltJf6 4 ltJf3
44
e6 5 �g5 dxe4 6 e4 b5 7 e5 h6 8
�h4 g5 9 ltJxg5 hxg5 1 0 .ltxg5
ltJbd7 1 1 g3 l::t g 8
With 1 1 ...11g8, Black seeks to ex­
ploit directly White's move order with
1 1 g3 . Since White has deferred the
'execution' of the knight on f6, it is
still alive Gust!) . If Black can break the
pin on the knight then he can move
the knight on d5 away and remain
material up.
1 2 h4 :xg5 1 3 hxg5 ltJd5 1 4 g6!
Using the doubled g-pawn to prise
open the black kingside.
1 4 . . .fxg6
Question 3: Help! What's happen­
ing?
Answer: After the exchanges White
now has unchallenged control of the
h-file for his rook on h I : the rook will
come to either h7, cutting across
Black's seventh rank, or h8, pinning
the black bishop on f8 to the king.
Black also has two isolated pawns on
e6 and g6 which are ideal targets for
the light-squared bishop on fl and
there is also a hole on d6 that would
make a perfect outpost for the white
knight on c3 via the e4-square.
Black has two pieces for the rook

and possession of the bishop pair, but
it is really Black's queenside which is
his major strength. It may seem that
Black can easily destroy White's cen­
tre with . . . c6-c5, but this weakens the
support of the knight on d5, which
requires the support of both e6- and
c6-pawns. Therefore Black often plays
... lDxc3, preventing lDe4-d6, and creat­
ing a new pawn base on c3 which
Black can then attack with . . . b5-b4
before following up with . . . c6-c5.
1 S �g4 �e7
15 . . :tWa5 has been practically re­
futed, as shall see in Game 25.
16 i.g2!
16 nh8 and 16 ft'xg6+ are less pre­
cise - see Games 23 and 24.
1 6 .. :ii'f7! 1 7 i.e4!
Attacking the weak g6-pawn ..
1 7 . . . lbe7
White must now activate his rooks
on the h-file and invade on the sev­
enth or eighth ranks (or both) ! Black's
main aim is to bring his queenside
pieces - the rook on a8 and bishop on
c8 - to active squares and move his
king to safety there. It may seem
strange to discuss positional factors
when White can simply win the ex­
change and a pawn with 1 8 lDxb5
cxb5 19 .llx a8, but in fact they are ex­
tremely relevant here. As Julian
Hodgson once remarked to me in a
similar position, by taking the rook
on a8, White has 'developed' this piece
for Black since he no longer needs to
worry about it! This manoeuvre also
gives Black two tempi ( . . . lDb6 and one
other) to carry out his queenside aims.
The popular 1 8 klh8 (see the next
game) by contrast gives Black no
B o t v in n ik Va ria tio n w i t h 1 1 9 3
tempi towards development of his
queenside, but of course it keeps the
material balance in Black's favour
(knight and bishop vs. rook) .
1 8 lbxbS cxbS 1 9 i.xa8 lbb6 20
i.e4
Shirov prefers White here and I feel
that he is right, although it must be
said that this position, with its very
lmusual material balance, is so rich in
possibilities that there is still plenty of
scope for experimentation on both
sides. Indeed, a little later Shirov of­
fered this line against Khalifman - as
Black!
20 . . . i.d7!
Shirov praises this improvement
over Kramnik's suggestion of
2o . . . lDbd5, when Shirov proposes 2 1
�h8! , intending 2 2 f3, 2 3 �f2 and 24
lhh1 , when 'White is clearly on top'.
2 1 l:th8 i.c6 22 f3!
Now Black will think twice about
exchanging bishops with . . . .llxe4 as
f3xe4 will 'fill in' the hole on d5 and
give White a massive centre.
22 . . . lbbdS?
A serious mistake which gives
White the time to carry out his ideal
plan. Shirov analyses 22 . . . 'itd7 23 'ite2
45
Th e S e m i - Sla v
(not 23 �f2 �xe4!, when White can
no longer recapture on e4 with the
pawn) 23 . . . �g7! (evicting the rook
from the eighth rank) and claims a
small advantage for White after 24
l:Ih7 CDbd5.
Black's dark-squared bishop IS
pinned to the queen on f7 and re­
stricted by the white pawns on d4 and
e5 . Moreover, the black queen is tied
to its protection and thus prevented
from becoming active. Black has two
plans: to attack d4, probably with
. . . CDf5, and to expand on the queenside
with . . . a7-a5-a4 and . . . b5-b4. Obvi­
ously, White must keep pounding the
kingside since the open h-file allows
him to use his big advantage: the pair
of rooks. The main idea is l:Ih6, at­
tacking g6, and iVg5-h6 is another
plan.
23 Wf2 lLlb4
As Shirov points out, 23 . . . �d7 now
allows 24 l:Iah 1 �g7 25 � 1h7! ! with
great play for White.
24 a3! lLld3+
24 . . . �xe4 25 axb4 .tf5 26 iVg5 CDc8
27 'iVh6! �e7 28 iVh4+! �d7 29 'iVf6! is
even more horrible according to Shi­
rov.
25 �xd3 cxd3 26 l:td 1 'iVf5 27 'iVxf5
gxf5 28 l:txd3
Shirov comments here that 'two
rooks and two extra pawns are too
much for three minor pieces' and in­
deed he makes the win look easy from
here.
28 . . . i.d5 29 l:tc3 lLlc6 30 We3 Wf7
31 l:th7+ �g7 32 l:th2 Wg6 33 l:thc2
lLla5 34 l:tc7 a6 35 l:ta7 lLlc4+ 36
We2 a5 37 b3 lLlxa3 38 l:tc8 �h6 39
l:tg8+ Wh5 40 l:th 7 1 -0
46
4 1 g4+ will be the end.
Game 22
Kalantarian-Yegiazarian
A rmenian Championship 1 994
1 d4 d5 2 c4 c6 3 lLlf3 lLlf6 4 lLlc3
e6 5 �g5 dxc4 6 e4 b5 7 e5 h6 8
�h4 g5 9 lLlxg5 hxg5 1 0 �xg5
lLlbd7 1 1 g3 l:tg8 1 2 h4 l:txg5 1 3
hxg5 lLld5 1 4 g6 fxg6 1 5 "iUg4 'iVe7
1 6 �g2 'iVf7 1 7 �e4 lLle 7 1 8 l:th8
1 8 . . . lLlb6! 1 9 We2!
Freeing the path for the queen's
rook to come into the action with
l:Iah 1-h7.
1 9 . . . b4?
This seems to be a mistake. Ak­
opian suggests 19 . . . �d7 20 a4 b4 2 1 as
CDbd5 (2 1 . ..bxc3 22 axb6 cxb2 23 l:Ixa7
l:Ib8 24 b7 �d8 25 'iVg5! is unpleasant
for Black) 22 CDa4 with an unclear po­
sition.
20 l:tah 1 Wd7
Akopian suggests instead that
20 . . . bxc3 2 1 :lh7 'iVxh7 22 :xh7 cxb2
23 'iVf3 �b7 24 �xg6+ �d7 25 'iVa3 is
unclear, but the threat of 'iVd6+ looks
unpleasant for Black.
21 l:t8h7! 'iVg8

2 1 . . .�g7 loses to 22 �xg6 ttJxg6 23
1I1h6!, when the knight on g6 and the
bishop on g7 are doomed to fall.
22 lbb5?
22 d5! bxc3 23 d6! was better ac­
cording to Akopian.
22 . . . cxb5 23 �xa8 lbxa8 24 'ii'f 3?
This is the fatal mistake. 24 d5! ttJc7
25 d6 ttJd5 26 �d4! would still have
been unclear according to Akopian.
24 . . . -tg7 25 �xa8 lbc6 26 �f1 �c7
27 d5 exd 5 28 e6 c3 29 bxc3 bxc3
30 �g2 c2 31 e7 �h3+ 32 �xh3
'ii'x h7+ 33 �g4 �xh 1 34 �xc6+
'it'xc6 35 e8� + �c5 36 �e7+ �c4
37 'ii'e 2+ '11o> c 3 0-1
Game 23
Khalifman-Shirov
Pardubice 1994_
1 d4 d 5 2 c4 c6 3 lbf3 lbf6 4 lbc3
e6 5 -tg5 dxc4 6 e4 b5 7 e5 h6 8
�h4 g 5 9 lbxg5 hxg5 1 0 �xg5
lbbd7 1 1 g3 �g8 1 2 h4 l:txg5 1 3
hxg5 lbd5 1 4 g6 fxg6 1 5 'ii'g4 'ii'e 7
16 l:th8 ? !
1 6 . . . lbxc3! 1 7 bxc3 'ii'a 3! ! 1 8 'ii' x g6+
18 �xe6+ <it>d8 is clearly better for
Black according to Khalifman.
B o t vin n ik Va ria tio n w i t h 1 1 g 3
1 8 . . . �d8 1 9 �d 1
Very sharp. Khalifman gives 19 �c2
lbb6 20 �c 1 <it>c7 as unclear.
1 9 . . :�xc3+ 20 '&te2 'ii' b 2+! 21 l:td2
'ii'b4 22 'ii'x e6 c3 23 a3!
23 . . . �xa3?
The decisive mistake according to
Khalifman. 23 . . . 1i'b 1 would have been
better, when 24 �d6 �e4+ 25 �d1
'i'g4+!? (25 . . . �b 1+ secures a perpetual)
26 l':te2 �g7 27 �xc6 is unclear!
24 �c2 �c7 25 �h7 ! b4 26 �a2
b3! ? 27 �xa3 c2 28 �a4 c 1 � 29
�c4 1 -0
The threat against c6 forces Black to
give up his queen (again) and
29 . . . 'i'xc4+ 30 'ti'xc4 b2 3 1 �a2 :b8
32 'i'b 1 �a6+ 33 <it>d2 �xf1 34 e6 �b3
35 .l::[xd7+ <it>b6 36 e7 �xe7 37 l:txe7
.1l.d3 38 Ite3 wins (Khalifman) .
I remember the next game well
since I was playing on the board next
to it!
Game 24
Mecking-San Segundo
Linares Open 1995
1 d4 d5 2 c4 c6 3 lbf3 lbf6 4 lbc3
47
Th e S e m i - Sla v
e6 5 �g5 dxc4 6 e4 b5 7 e5 h6 8
�h4 g 5 9 tZJxg5 hxg 5 1 0 .lixg5
tZJbd7 1 1 g3 �g8 1 2 h4 �xg5 1 3
hxg5 tZJd5 1 4 g6 fxg6 1 5 'iVg4 'iVe7
1 6 'iVxg6+ 'iVf7 1 7 'iVxf7+ \t>xf7 1 8
�g2
If White tries to avoid the exchange
on c3 with 1 8 ttJe4, then Black has
1 8 . . . i.b4+ followed by . . . c4-c3 with a
messy game.
1 8 . . . tZJxc3 1 9 bxc3 �b8 20 f4
20 �xc6 �b7 occurred in the game
Shirov-Stisis, London (Lloyds Bank
Masters) 1990, which continued 2 1
�h7+ �g6 2 2 .l:lxd7 i.xc6 2 3 �xa7,
and now 23 . . . b4 would have given
Black a good game according to Wells.
The text supports the white eS-pawn
in anticipation of the pressure that
Black is going to exert on d4 and c3 .
20 . . . b4 2 1 \t>d2 c5!
22 d 5 tZJb6! 23 dxe6+
23 d6 is critical, when Black can try
23 . . . ltJds or 23 . . . ttJa4, putting pressure
on the c3-pawn.
23 . . . �xe6 24 �e4 �d8+ 25 \t>e3
tZJd5+ 26 �xd5 �xd5 27 �hd 1 �xd 1
28 ttxd 1 �f5 29 \t>d2 �e7 30 a3
bxc3+ 31 \t>xc3 �d3 32 �xd3 cxd3
33 \t>xd3
48
Black should not really lose this po­
sition but he nervously ran himself
into time-trouble with disastrous re­
sults.
33 . . . <.t>g6 34 g4 �h4 35 \t>e4 .lig3
36 f5+ <.t>g7 37 g5 �h4 38 g6 c4 39
�d4 .lig3 40 e6 �d6 41 a4 �f6 42
�xc4 a5 43 <.t>b5 .lib4 44 \t>c6 1 -0
After the game, when San Segundo
was discussing the game his opponent
1 heard him say, 'I think 1 was better,
perhaps 1 could have won somehow.'
Mecking, an extremely devout Chris­
tian who credits God's intercession for
his recovery from a usually terminal
disease abruptly replied that 'God
would not have let it happen!'
Game 25
Knaak-Van der Wiel
Lugano 1989
1 d4 d5 2 c4 c6 3 tZJf3 tZJf6 4 tZJc3
e6 5 .lig5 dxc4 6 e4 b5 7 e5 h6 8
�h4 g5 9 tZJxg5 hxg5 1 0 �xg5
tZJbd7 1 1 g3 �g8 1 2 h4 �xg5 1 3
hxg5 tZJd5 1 4 g6 fxg6 1 5 �g4 �a5?
This aggressive counterattack
against the knight on c3 seems to lose
by force after White's splendid reply.

1 6 "xe6+ �d8 1 7 �g2! ! lbxc3 1 8
�f1 ! ! lbdS 1 9 �xc6 lbSb6
19 ... ltJc7 20 e6 is also very painful
according to Knaak.
20 :h8
Threatening 'iVd6.
20 . . . �b4 2 1 e6 l:tb8 22 exd7 lbxd7
23 a3! �e7
23 .. :iVxb2 24 liIxfS+! ltJxfs (24 .. .'itie7
25 :e l+) 25 'iVf6+ r:t;; c7 26 'iVf4+! is no
better according to Knaak.
24 �xg6 l:tb6 2S �hS l:tf6 26 lte 1
'i'd6 27 l:te8+ 'i;c7 28 �h4 lbb6 29
'i'e4 lbd7 30 Ith 7
30 . . . b4 3 1 axb4 �a6 32 "eS+ �d6
33 �h3 1 -0
A very impressive game by the
German grandmaster.
B o t vin n ik Va ria tio n w i t h 1 1 g 3
We shall now examine the popular
sideline with l 1 . . :i'aS. This line can
arise after either 1 1 g3 'iVaS 12 g3 or 1 1
exf6 �a5, when 1 2 g3 is the automatic
choice. Even if you prefer the 1 1 exf6
move order, this section will therefore
still be important for you.
l 1 . . :iVaS removes the queen from
the h4-dS diagonal and thus forces the
immediate capture on f6. As with
. . . 'iVb6, it also prepares queenside cas­
tling. However, unlike . . :iib6, .. :vWaS
does not protect the bishop on b7 and
thus the chances of a quick . . . c6-c5 are
very small. This move aims not at
White's weakness on d4, but to make
the most of Black's queenside
strength, supporting . . . bS-b4. This
would not only attack the knight on
c3, but also open the fifth rank for the
queen to attack the unprotected
bishop on gS and support . . . c4-c3 . If
Black can play both . . . bS-b4 and . . . c4-
c3 he will open the a6-fl diagonal to
which the black bishop on cS can
move m one go.
Game 26
Oll-Kaidanov
Kuibysev 1986
1 d4 lbf6 2 lbf3 dS 3 c4 e6 4 lbc3
c6 S �gS dxc4 6 e4 bS 7 eS h6 8
�h4 gS 9 lbxgS hxgS 1 0 �xgS
lbbd7 1 1 g3 �aS 1 2 exf6 b4 1 3
lbe4 �a6 1 4 "f3
By transposition, we would reach
the same position after 12 . . . .i.a6 1 3
'iWf3 b4! 14 ltJe4. This position is criti­
cal for the evaluation of the l 1 . . :iVaS
vanatlon. 14 b3 is seen in the next
game.
49
Th e S e m i- Sla v
1 4 . . . 0-0-0 1 S b3
Rather risky; this gives Black an ex­
tra means of opening up the queens ide
while White's king is in the centre.
15 iLe3 is the most dangerous move,
when Hertneck-Mueller, German
Bundesliga 1989, continued lS . . . 'iVdS
16 Ae2 i.b7 17 ltJgS! cS 18 'iVxds
.i.xdS, and now 19 0-0 eS 20 dxcS
would have given White the advan­
tage, according to Hertneck.
1 S . . . l2Jb6!
The text hits the d4-pawn, by un­
masking the attack of the rook on d8 .
1 6 l2JeS?
16 i.e3 .i.b7 17 i.g2 (17 bxc4 b3+!)
17 . . . cxb3 1 8 'iVe2 (18 0-0 bxa2) looks
very dodgy for White.
1 6 . . . �bS!
Protecting the pawn on c6.
1 7 �e3 l:tdS ! 1 8 a4 bxa3+! 1 9 iLd2
�xd2+! ! 20 �xd2 iLxeS 21 bxe4
l:txd4+
The rest must have been very pain­
ful for White.
22 �e3 l2Jxe4 23 �xe4 l:txe4+ 24
'1t>b3 l:tb4+ 2S �e3 l:tb2 26 l:tab 1
iLb4+ 27 �d4 eS+ 28 �eS l:te2+ 0-1
29 �d6 ':d2+ 30 �eS ':dS+ 31 �e4
i.c6 wins according to Kaidanov.
50
Game 27
Mecking-Matsuu ra
Sao Paulo Zonal 1 995
1 d4 dS 2 e4 e6 3 l2Jf3 l2Jf6 4 l2Je3
e6 S �gS dxe4 6 e4 bS 7 eS h6 8
iLh4 g S 9 l2JxgS hxgS 1 0 iLxgS
l2Jbd7 1 1 g3 �aS 1 2 exf6 b4 1 3
l2Je4 iLa6 1 4 b3
1 4 . . . l2Jb6
14 . . . 0-0-0 is considered stronger, as
15 'iVc2 ltJb6 16 .i.e3 eS! 17 dxeS 'iVxeS
gave Black good play in Khenkin­
Feher, Cappelle la Grande 1992.
1 S bxe4 l2Jxe4 1 6 �b3 �dS 1 7 f3
iLbS 1 8 l:te 1 l2Ja3 1 9 �xdS exdS 20
l2Jd2 iLd6 21 �f2
White is slightly better.
21 . . . �xf1 22 l2Jxf1 l2JbS 23 '1t>e3
'1t>d7 24 h4 l:tae8 2S l:txe8 l:txe8 26
hS l:te3+ 27 '1t>f2 l2Jxd4 28 h6 l:te8
29 l2Je3 as 30 iLf4 a4 31 iLxd6
�xd6 32 f4 b3 33 axb3 axb3 34 g4
b2 3S l:tb1 l:tb8 36 l2Jd 1 eS 37 l:txb2
l:th8 38 gS exf4 39 l2Je3 l2Je6 40
l:tbS d4 41 l2Je4+ �e6 42 l:teS d3 43
l:txe6+ fxe6 44 g6 l:txh6 4S g7 1 -0
We shall now consider less precise

move orders for Black: 12 . . . b4 13 lDe4
c3and 12 . . . �a6 13 'i'f3 l::t c S.
Game 28
Khalifman-Piket
A msterdam 1995
1 d4 d5 2 c4 c6 3 ttJc3 ttJf6 4 ttJf3
e6 5 �g5 dxc4 6 e4 b5 7 e5 h6 8
.th4 g5 9 ttJxg 5 hxg5 1 0 �xg5
ttJbd7 1 1 g3 'it'a5 1 2 exf6 b4 1 3
ttJe4 c3 1 4 bxc3 bxc3 1 5 'it'd3! �b 7
1 6 .te2 c5 1 7 O-O! c4 1 8 �e3 ttJb6?
Piket suggests that IS . . .'i'f5 19 f3
ttJb6 is worthy of attention.
1 9 ttJc5! �xc5 20 dxc5 ttJd5 2 1 'iVd4
l:tc8 22 l:tab 1 �a6 23 l:tfc 1 �xc5 24
�xc5 llxc5 25 l:tb8+ �c8 26 �e3
ttJxe3 27 fxe3 0-0 28 l:txc3 �a6 29
nxf8+ �xf8 30 g4 �g8 31 �f2 �h 7
32 h4 'ito>g6 33 g 5 �f5 34 �g3 �b5
35 �d3+ <t>e5 36 �c2 'itid6 37 'itig4
nd5 38 h5 l:td3 39 l:txd3+ cxd3 40
.td 1 1 -0
Game 29
K ramnik-Ehlvest
Riga 1995
1 ttJf3 d5 2 d4 ttJf6 3 c4 c6 4 ttJc3
B o t vin n ik Va ria tio n w i t h 1 1 g 3
e 6 5 �g5 dxc4 6 e4 b 5 7 e 5 h 6 8
�h4 g5 9 ttJxg5 hxg5 1 0 �xg5
ttJbd7 1 1 g3 �a5 1 2 exf6 �a6
1 3 �f3
This move attacks the point most
weakened by the light-squared
bishop's unusual development on a6
rather than b7 - the pawn on c6 - and
also prevents Black from castling
queenside immediately. By moving
the queen off the d-file White also
prevents his opponent from using a
pin on the d-file to transfer the knight
to d3 via e5 or c5. Moreover, the
queen supports the knight if it goes to
e4, protecting the bishop on g5 after
. . . b5-b4.
The alternative 13 a3 is sometimes
seen, but Black can then use the pin
on the d-file with 13 . . . 0-0-0 14 ..1g2
lDc5 15 0-0 lDb3 with an acceptable
game.
1 3 . . . l:tc8
Black's only reasonable choice here
is to transpose the line 12 . . . b4 13 lDe4
�a6 14 �f3 (Game 26) with 13 . . . b4 14
lDe4. White cannot play 14 'ili'xc6 as
both 14 . . . ItcS 15 'iWa4 'iWxa4 16 lDxa4
�b7 17 �gl �hh2 and 14 . . . �b7!? 15
'tixb7 ItbS! 16 ii'xbS+ {the only move:
51
Th e S e m i- Sla v
16 'i'e4 loses to 16 . . . bxc3) 16 . . . lbxb8
17 lbe4 (protecting the bishop on g5)
17 . . . b3+! 1 8 �d2 'i'xa2!! are good for
Black.
1 4 i.e2 b4 1 5 lLle4 c5 1 6 d5! exd5
1 7 �f5 ! !
The text indicates that White is will­
ing to sacrifice the knight on e4 in or­
der to exploit the main drawback of
13 .. J:k8: the uncastled black king.
1 7 . . . dxe4
17 . . . d4 loses to 1 8 .tg4 and 17 ... c3
18 i.g4 'i'b5 (18 . . . i.b5 19 'i'xd5 cxb2
20 :d1) 19 'i'xd5 cxb2 20 �d1 is also
good for White according to Kramnik.
1 8 0-0-0 �c7 1 9 i.g4 i.b5 20
�xe4+ �d8 21 .i.xd7 i.xd7 22 �he 1
Threatening 23 'i'e8 checkmate.
22 . . . i.h6 23 �a8+ �c8 24 �xd7+
�xd7 25 �d5+ 1 -0
25 . . . 'it>c7 26 :e7+ 'it>b6 27 'i'b7+
mate (Kramnik) .
Came 30
vanchuk-lII escas!
;> .. Linar;s 1 994
1 d4 lLlf6 2 c4 c6 3 lLlc3 d 5 4 lLlf3
e6 5 .i.g5 dxc4 6 e4 b5 7 e5 h6 8
.i.h4 g 5 9 lLlxg5 hxg5 1 0 .i.xg5
52
lLlbd7 1 1 g3 b4
Euwe's analysis of 1 1 . . .lbxe5 seems
convincing: 12 dxe5 'i'xd1+ 13 :xd1
lbd5 14 lbe4 i.b4+ 15 'it>e2 when
Black suffers from severe weaknesses
on the dark squares. The text gains
three pieces for the queen, but the
looseness of Black's structure and the
weakness of his king conspire against
Black.
1 2 lLle4 lLlxe4 1 3 .i.xd8 �xd8 1 4
�xc4 lLlb6 1 5 �d3 f5 1 6 .i.xe4 fxe4
1 7 �g4
1 7 . . . i.d7 1 8 �xe4 �c7 1 9 0-0 i.e7
Ivanchuk suggests that 19 . . . 'it>b7 20
a3 lbd5 would have been a better de­
fensive plan.
20 �ac1 Wb7 21 f4 �af8 22 �f3
lLld5 23 a3 �hg8 24 axb4 .i.xb4 25
�h7 �c7 26 �f2 �d8
26 . . . i.e7 was no better according to
Ivanchuk due to 27 :a1 'it>b7 28 b4
:b8 29 'i'd3 lbb4 30 :b2 winning.
27 �h 1 �h8 28 �d3 a5 29 �fc2
�h3 30 �xc6 �xf4 31 �c2 �f5 32
J::t c 8+ �e7 33 �g8 �h7 34 �g 1 a4
35 �b8 �hf7 36 �xb4 �f1 + 37 �xf1
�xf 1 + 38 �xf1 lLle3+ 39 Wf2 lLlxc2
40 �c4 lLla 1 41 h4 �e8 42 g4 lLlb3
43 h5 lLla5 44 �c7+ 1 -0
 B o t vin n ik Va ria tio n w i t h 1 1 g 3

Summary
1 1 g3 is my recommended move order for White players, when 1 1 . . .�b7 12 �g2
'i'b6 13 exf6 0-0-0 14 0-0 cS! is the best reply, leading to the main lines after 15 dS
b4 since Timman's 15 dxcS does not seem to be dangerous. The lines with
1 1 . . .'fiaS are still quite unexplored, but 12 exf6 b4 13 ttJe4, to meet 13 . . . �a6 with
14 'fif3, and Kramnik's 12 exf6 �a6 13 'ii'f3 seem very good ways to counter it.

1 d4 d5 2 c4 c6 3 liJf3 liJf6 4 liJc3 e6 5 �g5 dxc4 6 e4 b5 7 e5 h6 8 i.h4

g5 9 liJxg5 hxg5 1 0 �xg5 liJbd7 1 1 g3

1 1 . . .i.b7
1 1 . . .l1g8 12 h4 l:IxgS 13 hxgS ttJdS 14 g6 fxg6 15 'ii'g4 (D)
1S . . . 'iVe7
16 �g2 'ii' f7 17 �e4 ttJe7
18 ttJxbs Game 21
-

-18 l1h8 Game 22

- 16 �h8 Game 23
16 'ii'xg6+ Game 24
-

1S .. :ilaS Game 25
-

1 1 . . :iiaS 12 exf6 (D)

12 . . . b4 13 ttJe4
13 . . . �a6
-14 'iWf3 Game 26
- 14 b3 Game 27
- 13 . . . c3 Game 28
12 . . . �a6 Game 29
-

1 1 . . .b4 Game 30
-

1 2 i.g2 .b6 1 3 exf6 0-0-0 14 0-0 (D) c5

14 ... ttJeS Game 20
-

1 5 dxc5 Game 19
-

1 5 'fIg4 1 2 exf6 1 4 0-0

53
 CHA PTER FOUR
Botvi n n ik Variation : Early
Deviations after 5 �g 5 dxc4
1 d4 d5 2 c4 c6 3 tLlf3 tLlf6 4 tLlc3
e6 5 iLg5 dxc4
In this chapter we consider various
offshots in the Botvinnik variation for
White and Black.
The first two games deal with two
attempts by Black to deviate from the
main lines, neither of which is consid­
ered to be completely sound by open­
ing theory. In Game 33 we then take a
look a dubious double gambit by
White: 9 exf6 gxh4 10 tbeS 'iYxf6.
The rest of this chapter is devoted
to games in which White plays an
early a2-a4, with either 6 e4 bS 7 a4
(Games 34-37) or 6 a4 (Games 38-4 1) .
First we consider 10 . . . �e7 in the
main line Botvinnik system, which
has been practically put out of busi­
ness by the following variation.
Game 3 1
Demirel-Fridman
'RtEuropean JuniOr dh. 1992
1 d4 d 5 2 c4 c6 3 tLlf3 tLlf6 4 tLlc3
e6 5 iLg 5 dxc4 6 e4 b5 7 e5 h6 8
54
iLh4 g5 9 tLlxg5 hxg5 1 0 iLxg5 iLe7
1 1 exf6 iLxf6 1 2 iLe3!
It may seem strange to avoid the ex­
change of dark-squared bishops when
Black has so many dark-squared
weaknesses, but White wants to gain a
tempo in the future with tbe4!
1 2 . . . iLb7
12 . . . tba6 13 a4 tbc7 is the latest at­
tempt to resurrect this line. However,
14 'iYf3 (14 g3 cst? 15 dxcS .tb7 16
!lg1 b4 17 tbbs was unclear in Levitt­
Landero, Seville 1989) 14 . . . .td7 15
axbS cxbS 16 .tf4 !lc8 17 tbe4 tbdS 18
tbd6+ �e7 19 tbxc8+ followed by �e2
and 0-0 was good for White in Schoen­
Polajzer, Biel 1990.
1 3 �f3! �xd4
13 ... ii.e7!? 14 g3 tba6 15 .tg2 !lb8
16 0-0 cS enabled Black to unravel in
Schmidt-Hracek, Poznan 1987, and
seems to be Black's only hope in this
line.
1 4 0-0-0 iLxe3+ 1 5 fxe3 �e 7 1 6
tLle4 �f8 1 7 tLlc5!
Threatening 18 !ld8+ 'iYxd8 19
tbxe6+!, forking the king and queen.
 B o t vin n ik Va ria tio n : Ea rly D e via tio n s a f t e r 5 i. g 5 dx c 4
1 7 . . . �xc5 1 8 l:td8+ �e7 1 9 l:txh8
ttJd7 20 .l:xa8 �xa8 21 .lie2 ltJe5 22
�e4 ltJd3+ 23 ..t>b1 f5 24 �h4+ ..t>d7
25 �h7+ �e7 26 �h8 �e8 27
'i'd4+ 'it>c8 28 �xd3 cxd3 29 l:td 1 c5
30 �xc5+ �c6 31 �xc6+ .lixc6 32
l:td2 e5 33 ..t>c 1 e4 34 llf2 �d7 35
b4 'it>c7 36 h4 ..t>d6 37 g4 fxg4 38
h5 'it>d5 39 h6 g3 40 .l:[g2 .lif5 41
J:txg3 ..t>e6 42 l:tg7 'it>f6 43 l:txa7 1 -0
Though always fascinating, the line
with 9 .. .'�Jd5 never quite manages to
look totally convincing.
Game 32
Stefansson-I nkiov
Gausdal 1990
1 d4 ltJf6 2 c4 e6 3 ltJf3 d 5 4 ltJc3
c6 5 .lig5 dxc4 6 e4 b5 7 e5 h6 8
.lih4 g 5 9 ltJxg5 ltJd5
see follo wing diagram
1 0 ltJxf7! �xh4 1 1 ltJxh8 .lib4 1 2
a3!?
On the evidence of this game, this
looks almost like a forced win! Ionov­
Korneev, El Vendrell 1996, continued
12 �c1 c5 13 dxc5 �g5 14 ii.e2 i.. b7
15 .th5+ 'i.t>fs 16 iVf3+ 'i.t>gS 17 0-0
'it>xhS lS iVfS+ iVgS 19 iVxh6+ iVh7,
and now 20 "iWxh7+ �xh7 23 lLJxb5
would have been clearly better for
White according to Ionov. White has a
pleasant choice!
1 2 . . . ltJxc3 1 3 'i'f3! 'i'xd4
13 . . . it.a5 14 bxc3 �xd4 15 iVf7+
'i.t>dS 16 �d1 wins for White according
to T angborn.
14 �h5+ �d8 1 5 axb4 'i'e4+ 1 6
.lie2 liJxe2 1 7 �xe2 �h7 1 8 �d2+
'it>c7 1 9 "iVd6+ 'it>b7 20 l:td 1 �xh8 2 1
�e7+ .lid7 2 2 f4 �e8 2 3 �g7 ..t>c7
24 0-0 a5 25 f5 axb4 26 fxe6 �xe6
27 l:td6 �e8 28 l::t x h6 c3 29 bxc3 b3
30 e6 b2 31 c4 l:ta2 32 cxb5 cxb5
33 �e5+ 1 -0
We shall now consider the strange
line 6 e4 b5 7 e5 h6 S .th4 g5 9 exf6
gxh4 10 lLJe5 'ii'xf6 1 1 a4 .
Game 33
Alvarez-Antunes
Mondariz Balneario 1996
1 ltJf3 d5 2 d4 ltJf6 3 c4 c6 4 ltJc3
e6 5 .lig5 dxc4 6 e4 b5 7 e5 h6 8
.lih4 g5 9 exf6 gxh4 1 0 ltJe5 �xf6
55
Th e S e m i - Sla v
1 1 a4
Instinctively, I have always found
White's position repulsive here: he has
given up two pawns and the bishop
pair for seemingly only a negligible
lead in development. However, this
intuitive judgement completely over­
looks the most important positional
factor in White's favour: the ex­
tremely powerful knight on e5 which
exerts great influence on both sides of
the board. White's main field of influ­
ence is on the queenside, where his
pieces show excellent coordination.
The trio of black pawns on b5, c4 and
c6 are under intense pressure: the
knight on e5 attacks the pawn on c6,
preventing the knight on b8 from
moving; the knight on c3 and pawn
on a4 combine against the b5-pawn;
and the knight on e5 and the bishop
on f1 eye the c4-pawn. White's general
plan is to place his light-squared
bishop on to the h l-a8 diagonal via
�e2-f3 or g2-g3 and Jt.g2 to attack c6
and the rook on a8 beyond it. As a
little bonus, the knight on e5 attacks
f7 and when White plays �e2, he also
threatens Jt.h5, attacking f7.
Black's problems are entirely due to
56
the knight on e5. Therefore, by ex­
changing this piece, Black should logi­
cally be able to greatly reduce White's
activity. The natural carry out an ex­
change is to play . . . CLld7, but this is
impossible immediately due to
1 1 . . .CLld7 12 CLlxc6 �b7 13 axb5 (This
incidentally is the point behind play­
ing a2-a4 on move 1 1 . The old 1 1 �e2
allows 1 1 . . .CLld7 12 CLlxc6 .ib7 13 i.f3
a6 14 0-0 Jt.g7 15 a4 b4 16 CLle4 iVf4 17
iVc 1 iVc7 18 iVxc4 Jt.xc6 19 �ac1 0-0
with equality, as in Barlov-Karaklaic,
Yugoslavia 1987) . In order to play
. . . CLld7 in the game Antunes first pro­
tects the c6-pawn with 1 1 . . .�b7 and
then after 12 i.e2 plays 12 . . . CLld7.
1 1 . . . �b7 1 2 �e2 CDd7 ! 1 3 CDxd7
Perhaps 13 f4!? iVxf4 14 CLlxf7!?, in­
tending either 14 . . . iVxf7 15 �hS! or
14 . . . 'it>xf7 15 �f1 winning the queen in
both cases. Antunes mentions that 13
CLle4 is met by 13 ... iVf5.
1 3 . . . '1t>xd7
Although the black king has had to
move, it is in no real danger. I think
that Black is already better here.
1 4 �f3 a6 1 5 axb5?! axb5 1 6 l:txa8
�xa8 1 7 �a 1 �b7 1 8 �a7 '\te8 1 9
0-0
 B o t vin n ik Va ria tio n : Ea rly D e via tio n s a f t e r 5 jL g 5 dx c 4
1 9 lbxb5 i.b4+ is clearly better for
Black according to Antunes.
1 9 . . J�g8 20 ..t>h 1 'il'f4 2 1 ttJe4 h3!
22 ttJc5 �xc5 23 dxc5 hxg2+ 24
.txg2 1:[g5 25 'il'b6 "fic7 26 'il'a7 'il'e7
27 b4 cxb3 28 'il'a3 'il'xc5 29 'il'xb3
'i'c4 30 �a3 b4 31 'il'f3 �xf 1 +??
Oh no! ! Virtually anything would
win here, but not this!
32 .txf 1 c5 33 .ta6!
Ouch!
33 . . . .txa6 34 "fia8+ ..t>d7 35 'il'xa6
�e7 36 'il'c8 ..t>f6 37 "fih8+ ..t>g6 38
'i'g8+ �f6 39 �h8+ 'Ot>g6 40 'il'g8+
YZ - YZ
Black's position was s o good that
even after blundering his queen, he
could still hold the draw! In this line
he has a fortress position that White
cannot break down. In my opinion,
this idea marks the end of the road for
this interesting variation.
Game 34
D . G arcia-K ramnik
Pamplona 1 992
1 d4 d 5 2 ttJf3 c6 3 c4 ttJf6 4 ttJc3
e6 5 .tg5 dxc4 6 e4 b5 7 a4
Instead of the immediate 7 e5 White
keeps e4-e5 in reserve and attacks b5
instead. Black cannot protect the b­
pawn with 7 . . . a6 due to 8 axb5 cxb5 9
lbxb5!
7 . . . �b7
By protecting the rook on a8, Black
threatens . . . a7-a6, supporting the b5-
pawn. Note that the seemingly natural
7 . . . iLb4 can be met by 8 e5 h6 9 exf6
hxg5 10 fxg7 l:!g8 (the g-pawn isn't
defended any more!) 1 1 h4! with a
promising position for White. Black's
other alternatives here are discussed in
Games 36 and 37.
8 axb5
8 e5 is considered in the next game.
8 . . . cxb5 9 ttJxb5 �xe4
The alternative 9 . . . 'iWb6 10 'iWa4 was
played in P.Cramling-Galliamova,
Tilburg Women's Candidates 1994,
and now 10 . . . i.c6 1 1 'iWxc4 lbxe4 is
the critcial continuation. The immedi­
ate 9 . . . i.b4+ 10 lbc3 will most likely
transpose to the game after 10 . . . i.xe4
1 1 i.xc4.
1 0 �xc4
White has re-established material
equality at the cost of the dismantling
of his pawn centre. The resulting iso­
lated queen's pawn (IQP) structure
57
Th e S e m i - Sla v
with Black's b-pawn exchanged for
White's a-pawn favours Black in the
long run for two reasons: he has
gained the valuable b4-square for his
knight (since White cannot play a2-
a3); and has additional counterplay
along the b-file. However, in the short
term White has extra tactical chances
(for example along the a4-e8 diagonal) .
1 0 . . . -ltb4+ 1 1 tbc3 0-0 1 2 0-0 �b 7
12 . . . �xc3 1 3 bxc3 resembles the
Karpov variation of the Nimzo-Indian
(1 d4 ttJf6 2 c4 e6 3 ttJc3 i.b4 4 e3 0-0
5 �d3 d5 6 ttJf3 c5 7 0-0 dxc4 8 �xc4
cxd4 9 exd4 b6 10 i.g5 i.b7) , but
without the pawns on b6 and a2. In
that variation Black often plays
. . . �xc3, but here this is less effective
since the c-pawn is not restricted by a
pawn on b6 and can thus easily move
forwards to c5. Moreover, the pawn
on a7 is isolated.
1 3 �e2
The queen should move to b3, mak­
ing use of the exposed position of the
bishop on b7 and combining with a
knight on e5 and a rook on e 1 in a
typical IQP formation: 13 �e l ttJbd7
14 'i'b3 as 15 ttJe5! and now 15 . . . ttJb6
16 ttJxf7 �xf7 17 i.xe6 'i'e8 18 �e5
�f8 19 ..txf7 'i'xf7 20 'i'xf7+ �xf7 2 1
�b5 ttJfd7 2 2 ttJa4 ttJxa4 2 3 �xb7
ttJab6 24 d5 gave White good chances
in the endgame in Lutz-Zso.Polgar,
Rimaska Sobota 1994.
1 3 . . . -lte7 1 4 l:tfd 1 tbbd7 1 S dS
White simply plays for a draw
against his illustrious opponent . . .
1 S . . . exdS 1 6 tbxdS -ltxdS 1 7 1l.xdS
tbxdS 1 8 l:txdS �xgS 1 9 tbxgS h6
20 �d2 hxgS 2 1 l:txd7 �f6 % - %
. , .and gets it!
58
Game 35
Pogorelov-Korneev
Benasque Open 1 996
1 tbf3 dS 2 d4 tbf6 3 c4 e6 4 tbc3
c6 S �gS dxc4 6 e4 bS 7 a4 1l.b7 8
eS h6
Now 9 ..th4 g5 10 exf6 gxh4 is even
better for Black than in the main line,
but White has an interesting alterna­
tive.
9 1l.d2!?
This interesting idea is reminiscent
of the Slav Gambit (1 d4 d5 2 c4 c6 3
ttJf3 ttJf6 4 ttJc3 dxc4 5 e4!? b5 6 e5
ttJd5 7 a4) . White's knight will come
to e4 and eye the huge hole on c5,
while . . . h7-h6 has weakened Black's
kingside.
9 . . . tbdS 1 0 tbe4 a6 1 1 b3! cxb3 1 2
�xb3 tbd7 1 3 1l.d3 1l.e 7 1 4 0-0 0-0
1 S 1l.b1 !
15 �fe l allows Black the time to
play 15 . . . 'i'b6 and . . . !lfc8 .
1 S . . . l:te8 1 6 �c2 tbf8 1 7 tbcS
Maksimenko suggests 17 �e l f5 18
exf6 ttJxf6 (18 ... i.xf6!?) 19 ttJc5 with
compensation .
1 7 . . . -ltxcS 1 8 dxcS as!?
 B o t v in n ik Va ria tio n : Ea rly D e via tio n s a f t e r 5 Jt. g 5 dx c 4
18 . . .f5 19 exf6 �xf6 2 0 .ta2 �ad8
21 �fe l e5 22 .ta5 �d7 23 ttJd2 'ith8
24 ii.b l ttJf4 25 �a3 gave White a
slight edge in Maksimenko-Pinter,
Copenhagen 1995.
19 axb5 cxb5 20 c6 �c8 2 1 .ltxa5
%:txa5 22 c7 %:txa 1 23 cxd8'iV %:txd8
24 'iVb2 %:ta8 25 %:te 1 b4 26 .lte4 �b7
27 h3 l:tab8 % - %
Black is actually better here: White
has few targets, while Black can try to
push his b-pawn.
Game 36
Kallai-lukacs
Budapest 1995
1 d4 ttJf6 2 c4 e6 3 ttJf3 d5 4 ttJc3
c6 5 .ltg5 dxc4 6 e4 b5 7 a4 b4! ?
Although Black can drive the white
knight away with this move, this
move loosens his defence of the c­
pawn.
8 ttJb1 �e7 9 e5 ttJd5 1 0 �xe7
'iVxe7 1 1 .ltxc4 0-0 1 2 ttJbd2 c5! 1 3
dxc5? !
13 .txd5 exd5 14 dxc5 .ta6! pre­
vents White from castling and 13 ttJb3
ttJd7 14 0-0 .tb7 is fine for Black. 13
ttJe4 is therefore best, but after
13 . . . cxd4 14 'i'xd4 .tb7 Black has rea­
sonable chances according to Lukacs.
1 3 . . . .ltb7 1 4 ttJb3?
This is too risky. 14 0-0 ttJd7
(14 . . . ttJf4!?) 15 ttJb3 ttJxc5 16 ttJa5
�fd8 is fine for Black according to
Lukacs.
1 4 . . . ttJf4 1 5 'i'd6
15 0-0 ttJxg2 16 'itxg2 'i'g5+ 17 'ith 1
�d8! 18 'i'e2 �g4! wins according to
Lukacs.
1 5 . . . ttJxg2+ 1 6 We2 �xf3+ 1 7 <ot>xf3
'iWg5 1 8 l:thg 1 l:td8 1 9 l:txg2
A sad necessity. Lukacs shows that
19 �c7 ttJd7! 20 �xg2 ttJxe5+ 2 1 'ite2
�xg2 22 'i'xe5 'i/ig4+ wins for Black,
picking up the bishop on c4.
1 9 . . :�f5+ 20 <ot>e3 l:txd6 21 exd6
ttJd7 22 l:tag 1 'iWe5+ 23 Wd3 'iVxb2
24 l:txg7+ Wh8 25 We3 ttJe5 0-1
An attractive and theoretically im­
portant game.
Game 37
Bellon-Antunes
Platja d'Aro Barcino 1994
1 d4 d5 2 c4 c6 3 ttJc3 ttJf6 4 ttJf3
e6 5 .\tg5 dxc4 6 e4 b5 7 a4 'iVb6! ?
Black protects the pawn o n b5, sac-
59
Th e S e m i- Sla v
rificing a move in development to
keep hold of the gambit pawn. He
also unpins the knight on f6, drawing
the teeth from e4-e5.
8 �xf6 gxf6
8 i.xf6 doubled Black's f-pawns and
weaken his kingside. Although Black's
control of the central dark squares is
enhanced (the doubled f-pawn con­
trols e5 effectively) , he has less control
of the central light squares since there
is no longer a knight on f6 attacking
d5 or e4. White's general plan there­
fore is to play d4-d5, which will com­
bine with the pawn on a4 and the
knight on c3 to attack the black
queenside.
9 �e2 �b7 1 0 0-0 a6 1 1 d5?!
This produces a quite stunning
game, but my experience of such posi­
tions from both sides is that White
should not rush this move. 1 1 b3!,
opening the queenside, was played in
Lerner-Kaidanov, USSR 1985, and
gave White good play after 1 1 . . .cxb3
(or l 1 . . .b4 12 as Vlic7 13 ctJa4 c3 14
ctJb6 �a7 15 ctJc4 with compensation)
12 Vlixb3 ctJd7 13 d5! cxd5 14 exd5 .
1 1 . . . ttJd7 1 2 ttJd4 c5?
Two years later Antunes improved
significantly with 12 . . . cxd5 13 exd5
i.c5! 14 dxe6 fxe6, when 15 i.f3
i.xd4 16 i.xb7 :td8 ! 17 i.c6 O-O! 18
i.xd7 �xd7 19 Vlie2 'it>h8 gave Black a
large advantage in Campos-Antunes,
Mondariz Balneario 1996.
1 3 ttJc6 1:[g8 14 �h5 �xc6 1 5
dxe6 ! ! ttJe5 1 6 axb5 �xb5 1 7 �d5!
1:[a7 1 8 ttJxb5 �xb5 1 9 1:[fd 1 �b6 20
1:[xa6 1:[xg2+
20 . . . 'iYxa6 allows 2 1 'iYd8+ mate,
while 20 . . J�xa6 allows a supremely
60
artistic touch: 2 1 exf7+ 'it>e7 22 g8ctJ!!
mate (Bellon) .
2 1 �xg2 1:[xa6 22 .lixf7+ �e7 23
�d7+ 1 -0
23 . . . ctJxd7 24 �xd7+ is mate.
The immediate 6 a4 has developed a
small following in recent years.
Game 38
Stefansson-Tisdall
Reykjavik Zonal 1995
1 c4 c6 2 d4 d5 3 ttJc3 ttJf6 4 ttJf3
e6 5 .lig5 dxc4 6 a4 �b4
The immediate 6 . . . c5!? 7 e3 cxd4 8
exd4 i.e7 9 .txc4 0-0 10 0-0 ctJc6, as in
Ilundain-Korneev, Zaragoza 1995, is
not stupid either, as Black has a stan­
dard IQP position in which White has
helpfully conceded the b4-square for
his queen's knight.
7 e4 c5
Here 7 ... b5 transposes to the note to
Black's seventh move in Game 34.
7 . . . . i.xc3+ is somewhat greedy - see
Game 4 1 .
8 .lixc4 cxd4 9 �b5+
For 9 ctJxd4!? see Game 40.
9 . . . ttJc6
 B o t v in n ik Va ria tio n : E a r l y D e via tio n s a f t e r 5 � g 5 dx c 4
9 . . .lDbd7 is met by 10 'ii'xd4! ac­
cording to Makarov.
10 tDxd4 O-O ! ?
A very interesting pawn sacrifice.
The restrained 10 . . . �d7 is seen in the
next game.
1 1 tDxc6
11 .ixc6 bxc6 12 lDxc6 .ixc3+ 13
bxc3 "1Wc7 is fine for Black (Tisdall) .
1 1 . . .'iixd 1 + 1 2 l:txd 1 bxc6 1 3 .txc6
l::t b8 1 4 eS?!
This move is dubious according to
Tisdall.
14 . . . tDg4 1 S .tf4 .taS! 1 6 0-0 l:txb2
1 7 tDbS .tb6 1 8 .tg3 .ia6 1 9 l:tb 1
1:a2 20 l:ta 1 l:tb2 2 1 l:tab 1 Yz - Yz
Game 39
Cu H ansen-Tisdall
. �f4 21 g3 �gS 22 f4 tDxf4 23 l:txf4
Reykjavik Zonal 1995
1 c4 c6 2 tDc3 dS 3 d4 tDf6 4 tDf3
e6 S .igS dxc4 6 a4 .ib4 7 e4 cS 8
.txc4 cxd4 9 .ibS+ tDc6 1 0 tDxd4
.td7
In this game Tisdall studiously
avoids a reptition of his 10 . . . 0-0 from
the previous game, perhaps fearing a
prepared improvement.
1 1 0-0
1 1 lDxc6 bxc6 12 .id3 h6
(12 ... 'ii'a5!?) 13 �e3 e5 14 f3 'ii' a5 15
'ilfc 1 .ic5 16 .ic4 .ixe3 17 'ii'x e3 l:tb8
(17 .. :iVb6!?) 18 b3 was a tiny bit better
for White in Makarov-Bagirov, Pod­
olsk 1992.
1 1 . . . h6 1 2 .te3 0-0
The pawn grab is a little dangerous:
12 . . . .ixc3 13 bxc3 lDxe4 14 'iWg4 lDg5
15 h4 or 15 l:iadl is good for White
according to Tisdall.
1 3 f3 tDeS 1 4 tDc2 .ixc3 1 S bxc3
�c7 1 6 i.d4 l:tfd8 1 7 tDe3 a6? !
Tisdall criticises this move, which
weakens the queenside dark squares.
17 ... .ic6 immediately would have
been better.
1 8 .te2 .ic6 1 9 �b3 tDg6 20 .ib6
tDxe4 24 tDf1 �eS 2S .if3 tDxc3 26
.ixd8 %:txd8 27 l:tc 1 %:td3 28 �c4
l:txf3 29 l:bf3 tDe2+ 30 'it>f2 tDxc 1
3 1 �xc 1 Y2 - YZ
Game 40
Yermolinsky-Atal i k
Hastings 1995
1 d4 dS 2 tDf3 tDf6 3 c4 c6 4 tDc3
e6 S .tgS dxc4 6 a4 .tb4 7 e4 cS 8
.txc4 cxd4 9 tDxd4!? h6 1 0 .ie3 ! ?
10 �b5+ lDbd7 1 1 .ixf6 'ti'xf6 12
lDde2 a6 13 bxd7+ .ixd7 14 0-0 .ic6
15 'ili'b3 .id6 was poor for White in
Kiselev-Dreev, Helsinki 1992.
1 0 . . . 0-0
The safe option. 10 ... lDxe4 1 1 0-0
lDd6 12 .ia2 0-0 13 '1!i'f3 'ti'e7 14 �fdl
�h8 15 lDc2 .ixc3 16 bxc3 lDf5 17
.ic 1 �e8 18 �a3 'ii f6 19 lDd4 lDxd4
20 'ili'xf6 lDe2+ 2 1 '1t>f1 gxf6 22 Wxe2
led to an unclear ending in Sergeev-
61
Th e S e m i- Sla v
Savchenko, St Petersburg Open 1993 .
1 1 f3 'VJII e 7 1 2 0-0 CLlc6 1 3 CLlxc6
bxc6 1 4 �e2 a5!
Black's weak c-pawn is not too sig­
nificant because White has weakened
his queenside with a2-a4.
1 5 e5 CLld5 1 6 CLle4 CLlxe3 1 7 �xe3
l:[d8 1 8 l:[ad 1 .lta6 1 9 l:[xd8+ �xd8
20 b3 .ltxc4 21 bxc4 �c7 22 f4 l:[d8
23 ttJf2 .ltd2 24 �f3 �b6 25 l:[d 1
l:[d4 26 g3 �c5 27 �e2 .ltc3 28 �f1
�b4 29 l:[xd4 .ltxd4 30 �c2 �a3 3 1
�g2 �e3 32 h 4 �f8 33 h5 ri;e7 34
CLlh3 �d7 35 g4 �c7 36 �h7 �e2+
37 �g3 �d2 Y2 - Y2
Game 41
Cebalo-Palac
Croatian Ch., Slavonski Brod 1995
1 d4 d5 2 c4 c6 3 CLlf3 CLlf6 4 CLlc3
e6 5 .ltg5 dxc4 6 a4 �b4 7 e4
�xc3+ 8 bxc3 �a5 9 e5 CLle4 1 0
.ltd2 �d5
Defending the extra c4-pawn, but
after 10 . . . cs 1 1 .txc4 cxd4 (1 1 . . .lLlc6 12
ds lLlxd2 13 'iYxd2 exds 14 .txdS 0-0
15 0-0 .tfs 16 'i'f4 .tg6 17 c4 was bet­
ter for White in McCambridge­
Kaidanov, Las Vegas 1993) 12 cxd4
62
lLlxd2 13 lLlxd2 lLlc6 14 ..tbs 0-0 15
.txc6 bxc6 16 0-0 'i'dS 17 'i'c2 .ta6 18
lHd 1 the Donaldson suggestion o f
18 ... 'i'xd4 would have equalised 10
Lputian-Kaidanov, Lucern 1993.
1 1 a5!
Making it much harder for Black to
achieve . . . b7-bS.
1 1 . . . CLld7 1 2 �e2 0-0
Here the immediate 12 . . . bS 13 axb6
lLlxb6 14 0-0 also leaves Black's pawns
on the queens ide rather exposed.
1 3 0-0 c5 1 4 �c2 CLlxd2 1 5 �xd2
cxd4 1 6 cxd4 b5 1 7 axb6 CLlxb6
Black's dark squares are vulnerable
but he does still have his extra pawn.
1 8 CLlg5 �b7 1 9 .ltf3 �d7 20 �c2
g6 21 �xb7 �xb7 22 CLle4 l:[fd8 23
l:[fd 1 l:[ac8 24 CLld6 l:[xd6 25 exd6
�d7 26 �e4 �xd6 27 l:[xa7 CLld5!
The passed c-pawn is now ex­
tremely dangerous.
28 g3 c3 29 �f3 l:[c7 30 l:[a8+ �g7
31 �e4 CLlb4 32 l:[a3 l:[c4 33 ri;g2 c2
34 l:[c1 l:[xd4 35 �e3 l:[d 1 36 l:[aa 1
�c6+ 37 �h3 l:[xc 1 38 l:[xc 1 ttJa2
39 �d4+ �g8 40 �d8+ ri;g7 4 1
�d4+ e5 4 2 'iVxe5+ f 6 4 3 �e7+ �g8
44 'iVd8+ �g7 45 'iVe7+ ri;g8 46
�e8+ �g7 47 'iVe7+ Y2 - Y2
 B o t vin n ik Va ria tio n : Ea rly D e via tio n s a f t e r 5 iL g 5 dx c 4

Summary
6 a4 and 6 e4 b5 7 a4 are quite worth a try in the odd game, as they are quite
tricky in places. However, none of the other lines here really inspire much con­
fidence.

1 d4 d5 2 c4 cS 3 tDf3 tDfS 4 tDc3 eS 5 �g5 dxc4

6 e4
6 a4 ..tb4 7 e4
7 . . . c5 8 �xc4 cxd4 (D)
9 .ib5+ ttJc6 10 ttJxd4
10 . . . 0-0 Game 38
-

10 . . . ..td7 Game 39
-

9 ttJxd4 Game 40
-

7 . . . ..txc3+ Game 41-

6 b5 7 e5
. . .

7 a4
7 . . . ..tb7 (D)
8 axb5 Game 34 -

8 e5 Game 35
-

7 . . . b4 Game 36
-

7 . . . 'i'b6 Game 37
-

7 hS 8 �h4 g5 9 tDxg5 (D)

. . .

9 exf6 Game 33 -

9 hxg5
. . .

9 . . . ttJd5 Game 32 -

1 0 �xg5 �e7 Game 31 -

8 . . . cxd4 9 tDxg5

63
 CHA PTER FIVE
Moscow Variation with 7 e3
1 d4 d 5 2 c4 c6 3 tDf3 tDf6 4 tDc3
e6 5 �g5 h6 6 �xf6 �xf6 7 e3
The Botvinnik system is fascinating,
but sometimes you may feel you need
a break from all that dangerous living!
In recent years the Moscow variation
has become the most popular way to
try and defuse 5 .i.g5. By playing
5 . . . h6, Black seeks an improved Bot­
vinnik: if now 6 i.h4, then after
6 . . . dxc4 7 e4, Black can break the pin
one move earlier with 7 . . . g5! and after
8 .tg3 b5 Black has retained his extra
pawn on c4 without allowing White
an early e4-e5. Therefore, White play­
ers usually capture the knight on f6.
Question 1: But 6 i.xf6 'iVxf6 gives
up the bishop pair. What is White's
compensation?
A nswer: First, White buys himself a
tempo for development. Second, by
removing the knight on f6, White
weakens Black's control of e4; this
allows White to achieve the e2-e4
break which shows up the exposed
position of the black queen, as after
e2-e4 d5xe4, l2lxe4 White gains a
64
tempo on the queen on f6. Finally, by
removing the knight and diverting the
queen from d8, White weakens
Black's support of his d5-pawn. If
Black wishes to play his freeing breaks
... c6-c5 or . . . e6-e5, he must first release
the central tension with . . . d5xc4.
In this chapter we look at the most
common move 7 e3, preparing to de­
velop the bishop on f1 and thus allow­
ing kingside castling, to which Black
invariably replies 7 . . . l2ld7 supporting
both the . . . c6-c5 and the . . . e6-e5
breaks.
The fashionable way for Black to
handle this position is a kings ide fi­
anchetto with 7 ... l2ld7 8 .td3 dxc4
(Black releases the central tension to
facilitate his two breaks, but only after
White has played his bishop to d3, so
that it takes two moves to reach c4
instead of one) 9 .txc4 g6.
see following diagram
It may seem strange to play another
pawn move to develop Black's dark­
squared bishop when it has three posts

available on the a3-f8 diagonal. How­
ever, on b4 it is vulnerable to attack,
while on e7 it blocks the retreat of the
queen and on d6 it is vulnerable to e3-
e4 or perhaps lbe4 (if Black plays
... dSxc4) . On g7, the bishop is out of
range of White's knight on c3, sup­
ports . . . e6-eS, while it does not block
the retreat of the black queen. How­
ever, apart from the time involved, the
placement of the bishop on the al-h8
diagonal further weakens the d6-
square and slightly weakens the black
kingside. Furthermore, after h2-h4-hS,
Black will have to play ... g6-gS as
hSxg6 is an unpleasant threat.
Having seen this pOSItion many
times in tournament games, I have
accepted it as normal play. However,
looking at it afresh, I am struck by
how strange and amateurish Black's
position looks. If you were to just see
this position without any knowledge
of the opening, you might think that a
beginner was handling the black pieces
- why has Black made all these pawn
moves instead of developing his pieces
and why is the black queen on f6?
In fact, despite his seemingly eccen­
tric opening, Black's position is fully
M o s c o w Va ria tio n w i t h 7 e 3
playable. White has no concrete tar­
gets to aim at since Black is developing
only on his first three ranks, out of
the range of White's pieces and pawns.
Black reasons that if White opens up
the position, then although it may
cause him some danger, it will also be
to the advantage of Black's two bish­
ops. If White decides to build up care­
fully before opening the position, then
this gives Black extra time to develop
his pieces and prepare his position
both to carry out his own breaks and
to anticipate White's actions.
White has three basic approaches in
this variation:
1) Central systems
2) Manoeuvring systems
3) Queenside systems
Central systems
First we shall examine White's plans
involving the natural central push e3-
e4, starting with the immediate 10 0-0
�g7 1 1 e4.
Game 42
Khalifman-Akopian
Yerevan 1996
1 d4 d5 2 c4 c6 3 liJc3 liJf6 4 liJf3
e6 5 �g5 h6 6 �xf6 "it'xf6 7 e3 liJd7
8 �d3 dxc4
Black usually elects to capture on c4
either here or on the next move, since
after 8 . . . g6 9 0-0 �g7?! he has to
reckon with 10 e4 dxc4 1 1 eS! 'ii'e7 12
�xc4, when White has achieved his
desired advance in the centre and
Black is somewhat passively placed.
9 �xc4 g6 1 0 0-0 �g7 1 1 e4 e5
This is almost a reflex response to
65
Th e S e m i - Sla v
e3-e4 in this variation.
1 2 d5 tLlb6 1 3 �b3 �g4!
Taking advantage of the fact that his
queen stands on f6, Black creates a
weakness in White's position to com­
pensate for his own future weakness
on c6. The careless 13 . . . 0-0 allows
White to play 14 h3, preventing the
awkward . . . �g4.
14 h3 �xf3 1 5 'iVxf3 'ifxf3 1 6 gxf3
�e 7 1 7 dxc6 bxc6
This ending is a touch better for
White, since Black's weakness on c6 is
more vulnerable than White's on f3,
but it is eminently defensible for
Black.
1 8 l:tfc 1 l:thd8 1 9 tLld 1 l:td6 20 l:tc3
a5 21 tLle3 h5 22 l:tac 1 a4 23 �d 1
�d7 24 �f1 �h6 25 �e 1 �c7 26
�e2 'Oti>b7 27 l:t 1 c2 l:ta5 28 a3 �xe3
29 fxe3 f5 30 l:tc5 l:txc5 31 l:txc5
fxe4 32 fxe4 l:te6 33 l:ta5 �c7 34
�d2 �d6 35 l:ta6 �c5 36 l:ta5+ �d6
37 l:ta6 �c5 38 l:ta5+ � - �
White's lack of success after 13 �b3
�g4! has led to various experiments
with other moves. In the next game
White meets 12 . . . lbb6 with 13 lbd2!?
and in Game 44 with 13 dxc6 bxc6 14
�e2.
66
Game 43
Pi ket-Kramnik
Linares 1997
1 d4 d5 2 c4 c6 3 tLlf3 tLlf6 4 tLlc3
e6 5 �g5 h6 6 �xf6 'ifxf6 7 e3 tLld7
8 �d3 dxc4 9 �xc4 g6 1 0 0-0 �g7
1 1 e4 e5 1 2 d5 tLlb6 1 3 tLld2 ! ?
This offers an unusual position with
two knights against two bishops. In
general these positions are not very
active for White: once he gets his de­
sired set-up it is not easy to continue
to make progress.
1 3 . . . 0-0!?
The immediate 13 . . . lbxc4!? was also
possible.
1 4 a4! ?
1 4 �b3 !?, preserving the bishop,
might have been more prudent.
1 4 . . . l:td8 1 5 a5 tLlxc4 1 6 tLlxc4 'ifg5
1 7 'ifb3 �h3 1 8 tLle3 l:tab8 1 9 l:tac 1
�f8 20 'Oti>h 1 �d7 2 1 tLle2 cxd5 22
tLlxd5 �e6 23 l:tc7 l:tdc8 24 l:txc8
l:txcB.. 25 tLlec3 l:tc7 26 h3 �g7 27
'iVb5 l:td7 28 a6 bxa6 29 �xa6 �c5
30 �c6 �d4 31 tLlb5 'ifd8 32 tLlxd4
�xd5 33 'ifc5 �b6 34 �xb6 axb6
35 tLlf3 �xe4 36 tLlxe5 l:td5 37 tLlc4

b5 38 lLlb6 l:td2 39 b4 �d3 40 l:te 1
l:txf2 4 1 l:te3 �b 1 42 lLlc8 :b2 43
lLJd6 l:txb4 44 l:te 7 �a2 45 l:tb 7
l:tb1 + 46 �h2 b4 47 lLle8+ �f8 48
lLJf6 h5 49 g4 h4 50 �g2 �e6 0-1
Game 44
Lalic-Arduman
European Team Ch., Pula 1997
1 d4 d5 2 c4 c6 3 lLlf3 lLlf6 4 lLlc3
e6 5 �g5 h6 6 �xf6 'iVxf6 7 e3 lLld7
8 l:tc 1 g6 9 �d3 dxc4 1 0 �xc4 �g7
1 1 e4
Although in this game White has
played !te l instead of 0-0, this makes
no intrinsic difference to the position:
the strategies for both sides are identi­
cal to the two previous games.
1 1 . . .e5 1 2 d5 lLlb6 1 3 dxc6 bxc6 1 4
�e2 0-0 1 5 0-0 l:tb8 1 6 'iVc2 �e6
1 7 b3 'iVe7 1 8 lLlb1
Freeing the c-file for White's major
pieces and supporting the e4-pawn
from d2.
1 8 . . . f5 1 9 lLlbd2 fxe4 20 lLlxe4 �d5
21 lLlfd2 h 5 22 l:tcd 1 l:tbe8 23 �d3
h4 24 h3 :f4 25 l:tfe 1 l:tef8 26 f3
lLld7 27 lLlf2 lLlc5 28 �c4 �h7 29
lLlde4 lLlxe4 30 lLlxe4 �xe4 31 l:txe4
M o s c o w Va ria tio n w i t h 7 e 3
'iVc5+ 3 2 �h 1 'iVe7 33 �d3 l:t8f6 34
l:tde 1 l:txe4 35 l:txe4 :d6 36 l:tg4 1 -0
In the next four games White pref­
aces e3-e4 with both 10 0-0 and 1 1
l:tc l . As we shall see, this should not
be too dangerous for Black; usually he
will meet still e3-e4 with . . . e6-e5.
Game 45
Ivanchuk-Kramnik
Novgorod 1996
1 d4 d5 2 c4 c6 3 lLlc3 lLlf6 4 lLlf3
e6 5 �g5 h6 6 �xf6 'iVxf6 7 e3 lLld7
8 �d3 dxc4 9 �xc4 g6 1 0 0-0 �g7
1 1 l:tc 1
This is White's most non-committal
move. After 1 1 . . .0-0 he retains the op­
tion of playing in the centre (Games
45-48) , manoeuvring with ttJe4 and
.i.b3 (Game 50) or playing b2-b4
(Games 53-54) .
1 1 . . .0-0 1 2 l:te 1 'iVe7
A typical move, pulling the queen
back to a 'holding position' on e7
where it supports either . . . c6-c5 or
. . . e6-e5. However, as we shall see in
Game 47, 12 . . . !td8 is more accurate.
1 3 e4 l:[d8
13 . . . e5 is seen in the next game.
67
Th e S e m i- Sla v
1 4 e5!
This accomplishes three tasks: it
blocks the hS-a1 diagonal and thus
reduces the activity of Black's dark­
squared bishop on g7; it establishes an
outpost on d6 for a white knight; and
it deprives Black's knight of the defen­
sive square f6 and therefore makes
Black's kingside more vulnerable to
h2-h4-h5.
The most important factor here is
that White has protected e5 with the
e 1-rook, so that the undermining ... c6-
c5 can be powerfully met by d4-d5!
1 4 . . . b6
14 . . . b5 frees b7 for the light-squared
bishop with gain of tempo, but after
15 i-d3 �b7 16 t2Je4! the rook on c 1
combines with the knight o n e4 to
prevent . . . c6-c5.
1 5 "e2!
White aims to meet 15 . . .�b7 with
16 .ta6!, exchanging the light-squared
bishops.
Question 2: Why is this good for
White?
A nswer: A future . . . c6-c5 will open
the as-h 1 diagonal and threaten
. . . i-xf3, destroying the major defender
of White's centre and leaving it close
68
to collapse. By exchanging light­
squared bishops, White reinforces his
centre and maintains the advantages
that this gives him.
1 5 . . Jlb8 ! ?
A clever move, intending to play
... i.b7 and meet .ta6 with ... .taS .
1 6 iLd3!
White transfers the bishop to the
long diagonal to make sure that he
will exchange Black's light-squared
bishop once the as-h1 diagonal is
opened.
1 6 . . . �b7
16 ... 'iI8fS!? 17 h4 c5 1S d5 i.b7!? 19
h5 exd5 20 hxg6 occurred in Dautov­
Fridman, European Team Champion­
ship, Pula 1997, and now 20 .. .f6!? 21
e6 (2 1 exf6 'il8xf6) 2 1 . . .t2Je5!? would
have been interesting.
1 7 �e4 ttJf8 1 8 g3 l:1bc8 1 9 a3 l:1c 7
20 l:1ed 1 c5 21 �xb7 l:1xb7 22 ttJe4
l:1c7
22 . . .t2Jd7 23 t2Jd6 itbbS 24 'il8e4 cxd4
25 'il8xd4 t2Jc5 26 'il8f4 was Black's best
try according to Ivanchuk.
23 dxc5 l:1xd 1 + 24 l:1xd 1 bxc5 25 h4
ttJd7 26 ttJd6 ttJb6 27 ttJd2 l:1d7 28
f4 l:1xd6 29 exd6 �xd6 30 ttJe4 'i*'c6
3 1 l:1d8+ c;t?h7 32 h5 'VJIic7 33 hxg6+
fxg6 34 l:1d6 .ltd4+ 35 c;t?h2 ttJd5 36
l:1xe6 'VJIib7 37 'i*'g4 ttJe7 38 ttJd6
'VJIid7 39 'VJIie2 ttJg8 40 �e4 .ltf6 1 -0
Game 46
Pi ket-Novi kov
An,twerp Open 1996 k
1 d4 d5 2 c4 e6 3 ttJf3 ttJf6 4 ttJc3
c6 5 �g5 h6 6 .ltxf6 �xf6 7 e3 ttJd7
8 �d3 g6 9 0-0 dxc4 1 0 .ltxc4 �g7
Here we see a slightly different

move order to the previous game, but
the resulting position is the same.
1 1 .!:te 1 0-0 1 2 :e1 �e7 1 3 e4 e5
Preventing White from playing his
desired e4-e5.
1 4 d5 lLlb6 1 5 dxc6 ! ?
Piket suggests that 15 �b3! would
have been slightly better for White
here.
The insenion of the extra moves 1 1
lk1 0-0 1 2 :e 1 'ii'e 7 has drawn the
strength from . . . �g4 since White does
not end up with doubled pawns after
h2-h3. In this quiet position, Black's
bishops are fairly ineffective since
there are no raking diagonals on
which they can be activated. The d5-
pawn pressures Black's queenside and
cramps his pieces, but Black does not
want to capture on d5 since 15 ... cxd5
16 ttJxd5 ttJxd5 17 �xd5 �e6 1 8 �xe6
'i'xe6 19 'iWa4 gives White a nice pull
due to his superior minor piece, while
16 exd5, threatening d5-d6, may also
be dangerous.
1 5 . . . bxc6
Piket shows that 15 .. :i'c5 16 b3
bxc6 17 l:txc4 'iWd6 18 lhc6 wins, as
18 .. :ifxc6 Ioses to 19 ttJe7+!
1 6 i.b3 lId8 1 7 'iVc2 i.d7?!
M o s c o w Va ria tio n w i t h 7 e 3
1 7 ... �g4! 1 8 ttJd2 h5!, to activate
the dark-squared bishop with . . . �h6,
would have been stronger according to
Piket.
1 8 h3!
Piket assesses this pOSitIOn as
slightly better for White. The cramp­
ing d5-pawn has gone, but Black has a
weakness on c6 and his bishops don't
seem to have any good squares.
Question 3: Is this position very bad
for Black?
A nswer: It isn't disastrous - Black is
just slightly worse - but since Black
cannot create active play that easily, it
is much more fun for White to play.
1 8 . . . i.e8
This is a typical plan. On e8 the
bishop defends two weaknesses: f7 and
c6.
1 9 lLla4! lLlxa4 20 i.xa4 lIab8 2 1
:e3 �b4 2 2 b 3 .!:tbc8 2 3 .!:tc3 .!:td6
24 "iVe2
Piket suggests that the more dy­
namic 24 �c5 �e6 25 'ifc4! "1i'xc4 26
!t1xc4 f6 27 �a5 would have been bet­
ter than the quiet text.
24 . . :�b6 25 .!:tc5 %:te6 26 a3!? i.f8?
Black is tempted! Piket prefers
26 ... a5! 27 'ifd2 �a8, intending . . . �f8
69
Th e S e m i- Sla v
at a later stage.
27 l:txe5 .Jtxa3 2S l:txe6 fxe6 29 l:te2
.JtfS 30 e5 l:tdS 31 �e4 l:td5 32 �g4
�f7 33 l:txe6 l:td 1 + 34 'it>h2 �xf2 35
l:teS �b6 36 �h4 �b 7 37 l:txfS+
'it>xfS 3S �xh6+ 'it>e7 39 �h4+ 'it>fS
40 lDg5 �gS 41 �f4+ 'it>g7 42 �f6+
'it>h6 43 .JteS �g7 44 lDf7+ 1 -0
How should Black counter the plan
of :rc 1 , :re I and e3-e4 in this varia­
tion? The answer is to keep the black
queen on f6 for as long as possible.
Game 47
Timman-Gelfand
Yerevan Olympiad 1 �96
1 d4 d5 2 e4 e6 3 lDf3 lDf6 4 lDe3
e6 5 .i.g5 h6 6 �xf6 �xf6 7 e3 ttJd7
S .i.d3 dxe4 9 .i.xe4 g6 1 0 0-0 �g7
1 1 l:te 1 0-0 1 2 l:te1 l:tdS!
Black plays a useful move, pressur­
ing d4 rather than retreating his
queen. Now 13 e4 eS 14 dS is met by
the typical 14 . . . ttJb6 15 Ji.b3 Ji.g4!
White could now try 16 :re3 !?, prepar­
ing to play h2-h3 to drive the bishop
back. Timman didn't fancy the risk
and played:
1 3 'iVe2 �e 7 1 4 lDe4 a5 1 5 �b3 b6
70
1 6 lDe3 �b 7 Y2 - Y2
Game 48
Van Wely-Dreev
Wijk aan Zee 1996
1 d4 d5 2 e4 e6 3 lDe3 lDf6 4 lDf3
e6 5 .Jtg5 h6 6 �xf6 �xf6 7 e3 lDd7
S .Jtd3 dxe4 9 .i.xe4 g6 1 0 0-0 �g7
1 1 l:te 1 0-0 1 2 �e2
This is rather slow.
1 2 . . . �e7 1 3 l:tfd 1 a6 1 4 a3 b6 1 5
e4 �b7
Here Black doesn't need to play
. . . e6-eS as he has everything ready for
. . . c6-cS instead.
1 6 e5 e5 1 7 d5
Superficially this looks attractive,
but it is difficult for White to find a
follow-up. He cannot play dS-d6 as
after . . . 'iWd8, it is hard to meet . . . i.xB .
1 7 . . . b5 1 S �a2 e4 1 9 �b1 exd5 20
lDxd5 �xd5 21 l:txd5 l:tfeS 22 l:te 1
l:tadS 23 �d2 lDxe5 24 lDxe5 l:txd5
25 �xd5 .i.xe5 26 'it>f1 �f6 27 a4
<;t>g7 2S f3 bxa4 29 .i.a2 a3 30 bxa3
e3 31 �b3 l:tbS 32 l:te3 .Jtxh2 33
�e4 .i.e5 34 g3 �xg3 35 �xe3
�xe3 36 l:txe3 �d6 37 .Jte4 l:teS 3S
a4 a5 39 l:te2 <;t>f6 40 �d3 l:tdS 41

l:tc6 �g5 42 �c4 f6 43 �a6 .ltb4 44
f4+ �xf4 45 �xf6+ �g5 46 �a6 h5
47 �e2 �e8+ 48 �d3 �e7 49 �d4
h4 50 �d3 �d7+ 5 1 �e3 l:txd3+ 52
�xd3 h3 53 .l:!.a8 �g4 54 �e2 h2 55
.l:!.h8 �g3 56 �f1 �e7 0-1
Game 49
Ehlvest-Kharlov
Novosibirsk 1995
1 c4 c6 2 liJf3 d5 3 d4 liJf6 4 liJc3
e6 5 �g5 h6 6 �xf6 �xf6 7 e3 liJd7
8 �d3 dxc4 9 �xc4 g6 1 0 0-0 �g7
1 1 h3!?
I am surprised that this has not been
tried more often. White calmly pre­
pares to play e3-e4, but with the g4-
square under control. In the game
Black played along standard lines and
did not enjoy himself, so perhaps he
should adopt a different treatment.
1 1 . . . 0-0 1 2 e4 e5
Here 12 . . :�i'e7 makes sense, as 13 e5
is much less dangerous now that
White has spent a tempo on h2-h3 .
1 3 d 5 liJb6 1 4 �b3!
White reaches the desired forma­
tion.
1 4 . . J:td8 1 5 �e2 �f8 1 6 .l:!.ac 1 �d7
M o s c o w Va ria t i o n with 7 e 3
1 7 .l:!.fd 1 �e8 1 8 a3 �g7 1 9 �a2
.l:!.ac8 20 �e3 c5 21 �e2 c4? 22
liJd2 �c5 23 .l:!.f1 .l:!.c7 24 liJxc4
liJxc4 25 �xc4 .l:!.dc8 26 �b5 �xb5
27 �xb5 �f4 28 �e2 �d4 29 .l:!.c2
�c5 30 �d 1 a6 31 l:td3 h5 32 g3
�f6 33 h4 �d6 34 �f3 .l:!.8c7 35
�g2 .l:!.c4 36 �d 1 b5 37 �f3 �c5 38
�d 1 �d6 39 �f3 �c5 40 g4 hxg4
41 �xg4 �d6 42 �g3 b4 43 axb4
.l:!.xb4 44 .l:!.e2 .l:!.cc4 45 �f1 �b8 46
h 5 .l:!.xb2 47 hxg6 fxg6 48 �xb2
"iVxb2 49 liJe2 �xf2 50 �xf2 .l:!.xe4
5 1 .l:!.e3 �b6 52 liJc3 .l:!.f4+ 53 �g2
�b4 54 �e 1 :g4+ 55 �h3 'ii'f4 56
liJe4 �f5 57 �h2 g5 58 liJg3 .l:!.h4+
59 �g 1 �c2 60 �e2 �c5+ 6 1 .l:!.f2
�U4 62 liJh5+ 1 -0
M anoeuvring systems
The next three games deal with White
trying to play solidly in the centre,
avoiding any weakening pawn moves.
Game 50
Dautov-Dreev
Reggio Emilia 1995
1 d4 d5 2 c4 c6 3 liJf3 liJf6 4 liJc3
e6 5 �g5 h6 6 �xf6 �xf6 7 e3 liJd7
8 �d3 dxc4 9 �xc4 g6 1 0 0-0 �g7
1 1 .l:!.c1
In Games 5 1 and 52 White pre­
ferred to preface ttJe4 with 1 1 'iVc2
rather than 1 1 �c 1 .
1 1 . . . 0-0 1 2 liJe4 �e7 1 3 �b3!
White aims to prevent Black both
from developing his queenside with
. . . b7-b6 and . . . .ltb7, and from breaking
out with . . . c6-c5.
1 3 . . . .l:!.d8 1 4 �c2 liJf6 1 5 liJc5 b6?
This is a little careless.
71
Th e S e m i - Sla v
1 6 ttJe5! l:td6 1 7 ttJed3 it.b 7 1 8 it.a4
l:te8 1 9 b4!
This is something close to the ideal
for White as both Black's breaks have
been prevented and he is attacking a
black weakness on c6. However,
Dreev defends well and manages to
turn the game around.
1 9 . . . ttJd7 20 f4 ttJf6 21 .lib3 l:tdd8
22 a3 ttJd5 23 �e2 �a6 24 �e4
�b7 25 e4 ttJf6 26 f5 l:txd4 27 fxg6
fxg6 28 ttJxg6 �e8 29 ttJdf4 l:txe4
30 �a2 ttJd5 3 1 l:tee 1 l:txe 1 32 l:txe 1
�f7 33 l:txe6 l:te8 34 �e2 l:txe6 35
�xe6 .lid4+ 36 'it'h 1 �xe6 37 ttJxe6
.i.f6 38 ttJgf4 ttJxf4 39 ttJxf4+ 'it'f8
40 a4 'it'e 7 41 'it'g 1 �e3 42 ttJd3 e5
43 bxe5 bxe5 44 ttJf4 'it'd6 45 'it'f2
1i.e6 46 .i.b3 'it'e5 47 g3 'it'e4 48
ttJe6 .i.b4 49 'it'e2 �d7 50 h4 a5 5 1
�e2+ 'it'e5 52 ttJf8 .1g4+ 5 3 'it'e3
.lie1 54 ttJg6+ 'it'f6 55 ttJf4 �xg3 56
ttJd5+ 'it'e6 57 .i.b3 �xh4 58 �f4
�h3 59 ttJe3+ �d7 60 ttJe4 .1e 1 0-1
Game 51
Dal;ltov.:Dreev
Yerevan Olympiad 1996
1 d4 d5 2 e4 e6 3 ttJf3 ttJf6 4 ttJe3
72
e6 5 it.g5 h6 6 .lixf6 �xf6 7 e3 ttJd7
8 it.d3 dxe4 9 it.xe4 g6 1 0 0-0 olig7
1 1 ttJe4 �e7 1 2 �e2 0-0 1 3 �b3
White must be careful in these sys­
tems that he is ready to meet . . . e6-eS
by d4-dS. Here this is not so and Black
gets a comfortable game.
1 3 . . . e5!
The somewhat passive 13 . . J�d8 is
considered in the next game
1 4 l:tfe 1 �h8 1 5 ttJed2 l:te8 1 6 ttJe4
e4 1 7 ttJfe5 ttJxe5 1 8 dxe5?
This is just bad, 1 8 lbxeS .txeS 19
dxeS .tfS 20 e6 .txe6 2 1 'iWxe4 would
just have held the balance (Dreev) .
1 8 . . . i.f5 1 9 l:tad 1 l:tf8 20 ttJd6 .lixe5
21 ttJxf5 gxf5 22 l:td2 l:tad8 23 l:ted 1
�g7 24 �e4 l:txd2 25 l:txd2 l:te8 26
a4 l:te 7 27 g3 a6 28 �d 1 b5 29 �b3
�e6 30 �e2 e5 31 axb5 axb5 32 b3
e4 33 bxe4 bxe4 34 �a2 'it'f6 35
�a8 l:te8 36 �a5 e3 37 l:ta2 h5 38
h4 �d6 39 �a6 l:tb8 40 �xd6+
.i.xd6 41 i.e2 �e5 42 �f 1 �d5 43
�e 1 'it'e4 44 l:ta7 f6 45 l:ta6 l:td8 46
l:ta 1 i.b4 47 l:td 1 l:ta8 48 l:tb1 l:ta3
49 �e2 l:ta6 50 'it'e 1 l:ta3 51 'it'e2
l:ta7 52 �e 1 �e5 53 'it'e2 l:ta2 54
�d 1 'it'e4 55 �e 1 it.a3+ 56 �d 1
.lib4 57 'it'e1 l:ta5 58 .lib3+ 'it'd3 59

.ic2+ �c4 60 .Jtb3+ 'it>d3 6 1 .Jtc2+
�e2 62 l:txb4 'it>xf2 63 .Jtd 1 'it>xe3 64
�c2 'it'f2 65 .Jtxh5 'it>xg3 66 .Jtd 1 e3
67 'it'xc3 f4 68 l:tb6 l:td5 69 .Jte2 f3
70 .ixf3 'it>xf3 7 1 l:txf6+ ..t>g2 72
1:e6 'it'f2 73 l:tf6+ ..t>e 1 74 'it>c2 0-1
Game 52
Bareev-Dreev
Wijk aan Zee (match) 1995
1 d4 d5 2 e4 e6 3 lLlf3 lLlf6 4 lLle3
e6 5 .Jtg5 h6 6 .Jtxf6 �xf6 7 e3 lLld7
S .id3 dxe4 9 .Jtxe4 g6 1 0 0-0 .Jtg7
1 1 'i'e2 0-0 1 2 lLle4 �e7 1 3 .Jtb3
:dS 14 l:tad 1
14 l1ac1 would have transposed to
Game 50.
14 . . . a5 1 5 a3 l:ta6 1 6 l:td2 lLlf6 1 7
lLle5 l:ta7 1 8 lLle5
Finally, Black can also consider the
very solid option of developing his
bishop to the solid square e8, and
avoid weakening the queenside with
... b7-b6.
1 S . . . .Jtd7 ! 1 9 l:tfd 1 .Jte8 20 lLla4
l:taa8 21 e4 �e7 22 lLlf3 l:.ae8 23 e5
ttJd7 24 l1e 1 e5 25 �e4 b6 26 g3
'i'b8 27 d5 lLlxe5 28 lLlxe5 �xe5 29
'i'xe5 .Jtxe5 30 lLlxb6 l:.b8 31 lLle4
M o s c o w Va ria tio n w i t h 7 e 3
.Jtd4 3 2 lLlxa5 exd5 33 .Jta2 l:.xb2 34
l:.xb2 .Jtxb2 35 l:txe5 d4 36 lLle4
.Jte 1 37 lLle5 .Jtxa3 38 l:te7 �f8 39
lLlxf7 l:ta8 40 lLlxh6 .Jtd6 41 l:.e2 d3
42 l:td2 .Jtb4 43 l:txd3 l:txa2 44 lLlg4
.Jte6 45 lLle3 l:te2 46 l1d 1 .Jte5 47
'it>f1 .Jtf3 48 l:te 1 l:ta2 49 h4 �g7 50
l:te 1 .Jtb6 51 l:.b1 .Jtd4 52 l:te 1 'it>h6
53 l:tb 1 .Jtxe3 54 fxe3 l:tg2 0-1
Queenside systems
In the next four games we consider
ideas for White involving b2-b4. This
move stops Black from playing . . . c6-c5
and gives White the opportunity later
to open up the queenside with b4-b5.
Game 53
Beliavsky-Dreev
Novosibirsk 1995
1 d4 d5 2 e4 e6 3 lLlf3 lLlf6 4 lLle3
e6 5 .Jtg5 h6 6 .Jtxf6 �xf6 7 e3 lLld7
8 .Jtd3 dxc4 9 .Jtxe4 g6 1 0 0-0 .Jtg7
1 1 b4
1 1 . . .0-0 1 2 l:te 1 l:td8
The immediate 12 ... e5 can be met
by 13 �b3! exd4 14 exd4, when White
enjoys better cental control.
1 3 �b3 �e7 14 a4
73
Th e S e m i - Sla v
Perhaps this is premature. White
can also consider 14 �fd1!?, reserving
his options.
1 4 . . . aS!
As soon as White plays a2-a4, Black
strikes back with ... a7-aS to grab some
dark squares on the queenside.
1 S bxaS �xaS 1 6 �fd 1 b6!
Black then develops his queenside
by playing . . . b7-b6 and . . . i.b7 to pre­
pare . . . c6-cS.
1 7 i.e2
The immediate 17 lbd2 is consid­
ered in the next game.
1 7 . . . i.b7 1 8 tLld2 cS 1 9 tLlc4 �aa8
20 dS .Jtxc3!
This is a common idea: Black does
not want White to recapture with a
knight on ds and so gives up his dark­
squared bishop for the knight on c3.
2 1 dxe6 �xe6 22 �d6 'iHe7 23 'iHxc3
..ta6 24 �d2 .Jixc4 2S .Jixc4 tLleS 26
�xd8+ �xd8 27 .Jif1 tLlg4 28 h3 tLlf6
29 'iHb2 tLldS 30 'iVbS 'iVgS 31 as
tLlxe3 32 axb6 tLlxf1 33 �xf1 �b8 34
�b 1 �dS 3S b7 'iVd6 36 �e 1 �f8 37
'iVb2 �g8 38 'iHbS �f8 39 �b 1 �c7
40 'iVb2 �g8 41 �b6 �xb6 42 �xb6
'it>f8 43 �f1 �e7 44 'it>e2 �d7 4S
'it>d3 �c7 46 �f6 �f8 YZ - YZ
74
Game 54
Nikolic-Kramnik
Yerevan Olympiad 1996
1 d4 dS 2 c4 c6 3 tLlc3 tLlf6 4 tLlf3
e6 S .JigS h6 6 .Jixf6 'iVxf6 7 e3 tLld7
8 i.d3 dxc4 9 .Jtxc4 g6 1 0 0-0 i.g7
1 1 �c 1 0-0 1 2 b4 �e7 1 3 'iVb3 �d8
1 4 a4 as! 1 S bxaS �xaS 1 6 �fd 1 b6
1 7 tLld2 cS 1 8 dS .Jtxc3!
The other advantage of removing
the knight on c3 is that it makes it
harder for White to maintain his pawn
on dS.
1 9 �xc3 tLlf6 20 dxe6 .Jtxe6 21
�cc 1 l:ta7 22 �xe6 'iVxe6 Yz - Yz
White has sought to improve upon
these positions by playing a2-a3 and
b2-b4 before i.d3, reasoning that if
Black plays . . . dSxc4, then White will
be able to recapture on c4 in one
move, gammg a tempo.
Game 55
Pi ket-Dreev
Wijk aan Zee 1 996
1 d4 dS 2 c4 c6 3 tLlc3 tLlf6 4 tLlf3
e6 S i.gS h6 6 �xf6 �xf6 7 e3 tLld7

8 a3 g6 9 b4 �g7
Black waits for White to commIt
his bishop before taking on c4.
1 0 cxd5 exd5
White's aim is to provoke this re­
capture. Although the c8-h3 diagonal
is now opened for the light-squared
bishop, Black's . . . e6-e5 recapture is
taken away and this makes it harder
for Black to activate his bishop on g7.
White has the simple plan of the mi­
nority attack whereby White will iso­
late Black's c-pawn and leave it back­
ward on the half-open c-file by playing
b4-b5xc6. These positions are not ob­
jectively in White's favour, but since
they restrict Black's activity, they are
perhaps easier for White to play.
Now we can understand why
White doesn't play c4xd5 earlier: 9
cxd5 exd5 10 b4 is met by 10 . . . .id6!,
as in Van Wely-M.Gurevich, Ger­
many 1996, when 1 1 i.d3 iVe7 12 0-0
lDf6 allowed Black his ideal set-up in
these positions.
For 10 ... cxd5! see the next game.
1 1 �d3 0-0 1 2 0-0 lLlb6! 1 3 'ii' b 3
'i'd6! 1 4 .l::[ f c 1 �e6 1 5 lLld2 .l::[f b8 1 6
.l::[ a b1 a5!
This strike opens up the white
M o s c o w Va ria tio n w i t h 7 e 3
queens ide a little and gives Black active
play on the dark squares.
1 7 bxa5 lLld7 1 8 a4 .l::[x a5 1 9 'i'c2
l:ta7 20 lLle2 lLlf8 21 h3 �c8 22 lLlc3
lLle6 23 '1t>h 1 �d7 24 Ub6 'ii'c 7 25
.l::[ c b 1 .l::[ a 5 26 f4 lLlc5! ! 27 dxc5 .l::[ x c5
28 .l::[ 6 b3 b5 29 axb5 cxb5 30 lLlxd5
�xc2 3 1 lLlxc7 .l::[ x d2 32 �xb5 �f5
33 �c6 .l::[ x b3 34 .l::[ x b3 .l::[ c 2 35 lLld5
�h7 36 lLlb4 Uc 1 + 37 ..t>h2 �f8 38
�d5 .l::[ c 5 39 �xf7 ..t>g7 40 �e8
.l::[ c 2??
A terrible blunder: 40 .. J:tc8 would
have held the draw.
41 lLlxc2 1 -0
Game 56
Van Wely-Gelfand
Tilburg 1996
1 d4 d5 2 c4 c6 3 lLlc3 lLlf6 4 lLlf3
e6 5 �g5 h6 6 �xf6 'ii'x f6 7 e3 lLld7
8 a3 g6 9 b4 �g7 1 0 cxd5 cxd5!
This is an even safer way for Black
to play. By keeping the pawn struc­
ture fairly symmetrical, Black restricts
his opponent's active chances.
1 1 �d3 0-0 1 2 0-0 'i'e7 1 3 'ii' b 3
White's only superiority is a tempo
advantage on the queenside: he has
advanced his pawns to a3 and b4,
while Black's pawns are on a7 and b7.
1 3 . . . lLlb6 1 4 a4 �d7 1 5 lLld2 lLlc8
1 6 IUc 1 lLld6!
The knight is well-placed on d6,
eyeing c4 while defending b7.
1 7 b5 �fc8 1 8 a5 'ii'd 8 1 9 'ii' b 2 e5
20 lLlb3 exd4 2 1 lLlxd4 lLlc4 22
�xc4 dxc4 23 'ii'e 2 a6 24 b6 �c6
25 �d 1 'ii'e 7 26 �d2 �e4 27 l:tac 1
Uc5 28 lLla4 �xd4 29 exd4 .l::[ g 5 30
g3 �f3 3 1 .l::[ e 1 'iVd7 Yz - V2
75
Th e S e m i - Sla v

S um mary
The central lines considered in Games 41-49 offer White quite good chances of a
small structural edge in positions where e3-e4 . . . e6-e5, d4-d5 and then d5xc6 oc­
curs. White can also play in manoeuvring style {Games 50-52} but he must be
careful that he does not allow a quick . . . e6-e5 by Black; 1 1 l:!c1 0-0 12 ttJe4 'i'e7
13 i.b3 {Game 50} seems the best try. If White plays on the queenside with 1 1
�c1 0-0 1 2 b4, then Black should be fine as long as he adopts the . . . a7-a5 plan of
Games 53-54. 8 a3 {Games 55-56} does not seem dangerous with careful play.

1 d4 d5 2 c4 c6 3 lLlf3 lLlf6 4 lLlc3 e6 5 .Jig5 h6 6 .Jixf6 �xf6 7 e3

7 . Ji:Jd7 8 �d3
8 a3 g6 9 b4 i.g7 10 cxd5
10 . . . exd5 - Game 55; 10 . . . cxd5 - Game 56
8 . . . dxc4 9 �xc4 g6 1 0 0-0
10 l:ic1 - Game 44
1 0 . . . �g7 1 1 e4
1 1 �c1 0-0 (D)
12 l:tc1
12 . . . 'i'e7 (12 . . . �d8 - Game 47) 13 e4
- 13 . . . �d8 Game 45; 13 . . . e5 - Game 46
12 'i'e2 - Game 48; 12 ttJe4 - Game 50
1 1 h3 - Game 49
1 1 ttJe4 'iYe7 12 'i'c2 0-0 13 i.b3 (D)
13 . . . �d8 {13 . . . e5 - Game 51}
14 �ad1 - Game 52; 14 �c1 - Game 50 (by transposition)
1 1 b4 0-0 12 l:!c1 �d8 13 'ifb3 'ife7 14 a4 as! 15 bxa5 �xa5 16 �fd1 b6
17 i.e2 - Game 53; 17 ttJd2 - Game 54
1 1 . . .e5 1 2 d5 lLlb6 (D) 1 3 i.b3
13 ttJd2 Game 43
-

1 3 . . . �g4 - Game 42

1 1 . . . 0-0 1 3 �b3 1 2 . . . lLlb 6

76
1 d4 d 5 2 c4 c6 3 lDf3 lDf6 4 lDc3
e6 5 �g5 h6
In this chapter we consider ways in
which White can avoid the main line
of the Moscow variation, either by
playing for an early e2-e4 after 6 .lixf6
'i'xf6 (Games 57 and 58) , fianchettoing
(Game 59) or by gambitting the c­
pawn with 6 .lih4 (Game 60) .
Game 57
Timman-Gelfand
Belgrade 1995
1 d4 d5 2 c4 c6 3 lDc3 lDf6 4 lDf3
e6 5 .Jig5 h6 6 �xf6 �xf6 7 �b3
With this move White protects c4
and covers the b4-square in order to
allow a quick e2-e4 without a disrup­
tive check on b4.
The immediate 7 e4 dxe4 8 4Jxe4
does not pose any problems due to
8 . . . .lib4+ 9 �e2 'iWf4 10 'i'ic2 .lie7. In
Lautier-Kramnik, Paris (rapidplay)
1995, White now suffered from an
attack of misguided inspiration with
1 1 4Je5?!, a pawn sacrifice which pro-
vided little compensation after 1 1 . ..f5!
12 4Jd2 (12 4Jg6 'ilig4+!) 12 . . . 'ilixd4 13
4Jdf3 'iixc5 14 4Jg6 �g8 15 4Jxe7
'iWxe7 16 g3 e5!
7 . . . dxc4!? 8 �xc4 lDd7 9 g3
9 e4 is also met by simply 9 ... e5!
9 . . . e5! 1 0 0-0-0 �e7 1 1 lDe4 �f5
1 2 'ii'c 2 0-0 1 3 h4
13 �b 1 was seen in Korchnoi­
Dreev, Yalta 1995, when 13 . . . 4Jf6!
equalises at once (Timman) .
1 3 . . . exd4 1 4 lDxd4 �a5 1 5 �b1
lDf6 1 6 e3 lDd5 1 7 a3 .Jig4!
This equalises and now Black builds
up a slight initiative.
77
Th e S e m i - Sla v
1 8 �e2 �xe2 1 9 lLlxe2 lLlf6 20 lLld2
l:tfd8 21 lLld4 �f8 22 l:thg 1 lLlg4 23
lLlc4 'iVc5 24 lLlb3 'iVe7 25 e4 'iVe6
26 f3 l:txd 1 + 27 l:txd 1 b5 28 lLld4
�xc4 29 'iVxc4 bxc4 30 fxg4 c3 3 1
bxc3 �xa3 32 lLlxc6 l:tc8 3 3 lLle5 f6
34 lLlg6 l:txc3 35 h5 l:txg3 36 l:td8+
�f7 37 l:td7+ �g8 38 l:txa7 �c5 39
l:tc7 �d6 40 l:td7 �b4 41 l:tb7 % - %
Game 58 d
Atalik-Bacrot
Wijk aan Zee B 1 997
1 d4 d 5 2 c4 c6 3 lLlf3 lLlf6 4 lLlc3
e6 5 �g5 h6 6 �xf6 �xf6 7 'iVc2! ?
White again supports a quick e2-e4.
Now 7 . . . dxc4 8 e3 b5 9 a4!? is quite
risky for Black.
7 . . . lLld7 8 e4 dxe4 9 �xe4 g6 1 0
�d3 �g7 1 1 0-0 0-0 1 2 l:tfe 1 c 5 1 3
d 5 lLlb6 1 4 d 6 �d8!
1 5 �f4 �d7 ! 1 6 l:tad 1 �c6!
This plan really hits the spot! Black
intends . . . tDd7, solidly blocking the
white d-pawn and preventing tDe5.
1 7 �e4? lLlxc4 1 8 �xc6 lLlxb2 1 9
�xb7 l:tb8
It is always a horrid moment when
you realise that too many of your
78
pieces are hanging!
20 lLle5 lLlxd 1 21 lLlc6 'iVb6 22 l:txd 1
l:txb7 23 lLle7+ �h7 24 lLle4 f5 25
lLlg5+ hxg5 26 �xg5 l:txe7 27 dxe7
l:te8 28 h4 'iVc7 29 h5 �xe7 30
�xg6+ �h8 31 l:td3 0-1
Game 59
Petu rsson-Dreev
Yerevan Olympiad 1996
1 c4 c6 2 lLlf3 d 5 3 d4 lLlf6 4 lLlc3
e6 5 �g5 h6 6 �xf6 �xf6 7 g3
7 . . . lLld7 8 �g2 dxc4 9 0-0 �e7 1 0
lLle4 �f5! 1 1 lLled2 e5!
This plan of taking on c4 and play­
ing . . . e6-e5 seems very effective here.
1 2 lLlxc4
12 e4 was played in Komarov­
Renet, France 1996, but after 12 . . . 'iVe6
13 'iVe2 b5 14 a4 and now 14 . . . 0-0 15
d5 'iVd6 16 dxc6 'iVxc6 17 axb5 'iVxb5
18 tDxc4 Black would have equalised
quite easily according to Komarov.
1 2 . . . exd4 1 3 lLlxd4 �f6 1 4 e3
Komarov suggests the more aggres­
sive 14 e4, intending f2-f4 and e4-e5.
14 . . . 0-0 1 5 �h5 lLlb6 1 6 lLld2 �g5
1 7 'iVd 1 �g4 1 8 �c2 l:tad8 1 9 l:tac 1
l:tfe8 20 a3 �h5 2 1 lLle4 �h3 22
 M o s c o w Va ria tio n : Wh i t e 's 6 th a n d 7 th M o v e A l t e rn a t i v e s
�f3 �g4 23 .i.g2 .i.h3 24 �f3 % -%
Game 60
T opalov-Gelfand
Dortmund 1996
1 d4 d 5 2 c4 c6 3 lLlf3 lLlf6 4 lLlc3
e6 5 �g5 h6 6 �h4 dxc4 7 e4 g5 8
�g3 b5
Frustrated by the Black's solidity in
the 6 i.xf6 lines, White players have
recently turned back to this venerable
gambit, the only way for him to ob­
tain sharp play against S . . . h6.
As compensation for the pawn,
White has a strong pawn centre and
chances against the enemy king
(Black's . . . g7-gS has weakened f6 and
makes kingside castling a fraught af­
fair) . If Black develops his dark­
squared bishop to g7, then the d6-
square is extremely weak.
9 �e2 �b7
9 ... b4 is rather loosening but breaks
up the white centre: 10 tDa4 tDxe4 1 1
.txc4!? (1 1 i.eS tDf6 1 2 i.xc4 tDbd7
13 0-0 i.g7 14 'i'e2 tDb6 IS i.b3 0-0 16
tDcs gave White good compensation
in Relange-Giorgadze, Ubeda Open
1997) 1 1 .. .tDxg3 12 hxg3 tDd7 13 0-0
�g7 14 lie l 0-0 IS 1Ic1 , and now
Korchnoi's Is . . . tDb6! would have
equalised in Korchnoi-Timman, Wijk
aan Zee 1997.
1 0 0-0 lLlbd7 1 1 a4! ?
Also possible is 1 1 ds cxdS 12 exdS
tDxds 13 tDxbS, opening the centre.
1 1 . . . b4
Mikhail Gurevich preferred 1 1 .. .a6,
keeping the queenside solid, and ask­
ing White to make further efforts to
find some compensation, against Sher-
bakov in Niksic 1996. Sherbakov sug­
gests 12 axbS cxbs 13 dS tDcs 14 tDd4.
1 2 e5! lLlh5
12 ... bxc3 13 exf6 cxb2 14 l:tb l c3 IS
'ilfb3 with an edge (Gelfand) .
1 3 lLle4 c5 1 4 lLlfd2 lLlxg3 1 5 fxg3
�e7 !
IS . . . cxd4 16 i.hS! tDxeS 17 tDxc4!
tDxc4 18 !:txf7 wins for White! This
shows how White can strike seem­
ingly from nowhere in this line.
1 6 lLld6+ �xd6 1 7 exd6 c3 1 8 bxc3
bxc3 1 9 lLlc4 0-0 20 .tIc 1 f5 21 dxc5
.tIc8 22 �f3?
Missing Black's next. 22 .l::txc3 �xcS
was unclear according to Gelfand.
22 . . . �a6 23 c6 �xc4 24 c7 �f6 25
.tIe 1 lLlc5 26 l:te3 lLle4 27 h3 �d5 28
�xe4 fxe4 29 l:texc3 l:tf7 30 'it>h2
�e5 31 l:tc6 e3 32 �e2 �xc6 33
�xc6 �d5 34 �a6 l:tff8 35 l:tc3 ..tg7
36 l:td3 �e4 37 �xc8 l:txc8 38 d7
l::t x c7 39 d8'it' l:tf7 40 �d4+ �xd4
41 l:txd4 l:tb7 42 l:te4 ..tf6 43 l:txe3
.tIb4 44 l:tf3+ ..te7 45 l:tc3 l:txa4 46
l:tc7+ ..td6 47 l:th7 a5 48 l:txh6 l:tc4
49 l:th8 a4 50 h4 g4 5 1 h5 ..te7 52
h6 ..tf7 53 l:te8 l:tc5 54 l:ta8 l:th5+
55 ..tg 1 l:txh6 56 l:txa4 l:tg6 57 ..tf2
'it>g7 58 �f1 % - %
79
Th e S e m i - Sla v

S u m mary
It seems that Black has few problems in the offbeat lines after 6 �xf6. Only
Yermolinsky's 7 'iVc2 looks like it is worth further analysis. However, if you
like gambits and fancy having a go at Black's position, then the 6 �h4 line may
be for you, as it drags Black out of the solidity of the main line Moscow lines.

1 d4 d5 2 c4 c6 3 ltJf3 ltJf6 4 ltJc3 e6 5 �g5 h6 (D)

6 �xf6
6 �h4 Game 60
-

6 . . :iWxf6 (D) 7 'iVb3 (D)

7 'iVc2 Game 58
-

7 g3 Game 59
-

7 dxc4 Game 57
. . . -

5 . . h6
. 6. . . 'iixf6 7 'iVb3

80
 CHAPTER SEVEN

Meran Va riation : Main Line

1 d4 d5 2 c4 c6 3 liJf3 liJf6 4 liJc3 ously reduces the activity of his own

e6 5 e3 liJbd7 6 �d3 dxc4 7 �xc4 position (although Black's own light­
b5 8 �d3 �b7 9 0-0 a6 1 0 e4 c5 1 1 squared bishop has the same problem) .
d5 c4 1 2 �c2 �c7 After 5 e3 Black usually plays
In recent times this has become the 5 . . . liJbd7
crucial position in the Meran system Now White's main move is
and a major battleground in the Semi­ 6 �d3
Slav as a whole. First we need to get threatening the e3-e4 push. This
our bearings - a lot of things have will gain space in the centre and open
happened in just 12 moves! the c1-h6 diagonal for the dark­
With squared bishop, thus solving the only
5 e3 drawback to the white set-up: the
White takes a more relaxed ap­ bishop on c 1 .
proach to Black's fourth move than
with 5 �g5 . 5 e3 neutralises Black's
threat of . . . d5xc4 by protecting c4
with his light-squared bishop. In so
doing, he prepares to complete his
kingside development with �d3 and
castle his king to safety on the king­
side. Furthermore, White will soon be
in a position to play e3-e4.
Question 1: Sounds great! What's
the catch?
Answer: Unfortunately it also
blocks White's dark-squared bishop This i s a crucial situation for Black.
inside the pawn chain which obvi- White has developed his pieces

81
Th e S e m i- Sla v
smoothly and is on the verge of play­
ing e3-e4. Invariably Black now plays
6 . dxc4 7 �xc4 b5
. .
Question 2: What's the point?
A nswer: This marks the beginning
of Black's plan to deal with his
blocked-in bishop on c8. With
. . . d5xc4, Black removes one of his
pawns from the a8-h 1 diagonal, and
allows Black to follow up with . . . b7-
b5, freeing the b7-square for the light­
squared bishop. All Black has to do
after . . . ii.b7 is play . . . c6-c5 and the
bishop will be free!
Now after
8 �d3
returning to support the push e3-e4,
Black usually puts his bishop to its
best square with
8 . . . �b7
and after
9 0-0
Black can protect the pawn on b5
with
9 a6
. . .
Black is now ready to break against
the white centre with . . . c6-c5, attack­
ing the d4-pawn and opening the a8-hl
diagonal for the light-squared bishop.
White must therefore take the centre
with
1 0 e4 c5 1 1 d5
This move has a nice logic to it:
now that Black's light-squared bishop
has finally found a good post on b7,
White sets up a central pawn wedge
on e4 and d5 to block the diagonal!
Question 3: Doesn't White just lose
a pawn after 1 1 . ..exd5?
Answer: By taking on d5, Black
opens the e-file. With his king on e8
and no immediate chance of castling,
82
1 1 . . .exd5 is a huge risk. After 12 exd5
tDxd5 13 !!e 1+ ii.e7 14 'i'e2! Black
cannot castle due to 15 tDxd5 ii.xd5 16
'i'xe7 winning a piece.
Since 1 1 . . .exd5 is too dangerous,
Black usually plays
1 1 . . . c4 1 2 �c2 'fic7
reaching the starting position of this
chapter.
Question 4: Why does Black need to
play 12 . . .'i'c7 here?
Answer: White's attacking plan is to
play d5xe6 . . .f7xe6, and then e4-e5!
12 . . .'i'c7 anticipates this thrust by pro­
tecting the e5-square.
The critical position after 12 . . . 'i'c7
has developed very logically from the
plan that Black selected on his sixth
move. By exchanging his d-pawn for
White's c-pawn, Black simultaneously
obtained a queenside pawn majority
and removed the barrier to its expan­
sion - the white pawn on c4. How­
ever, the price of this queenside initia­
tive was a loss of central influence. By
relinquishing his pawn attack on e4,
Black freed White to play e3-e4. A
similar conflict was evident after
10 . . . c5: although this move gained
queenside space, it also loosened

Black's grip on the d5-square, allowing
White to increase his central territory
with 1 1 d5.
After 12 . . .'I'c7 Black's queenside
pawns on a6, b5 and c4 restrict
White's light-squared bishop on c2,
which is White's problem piece in this
line. These queenside pawns are also
extremely hard to dislodge - for ex­
ample, 13 b3 loses a piece to 13 . . . cxb3,
while 13 a4 b4 does not help White's
cause either. Once Black has com­
pleted his development and put his
king to safety, he will use his queen­
side initiative to cripple White's cen­
tre, chasing the white knight from c3
with . . . b5-b4 and thus weakening
White's support of e4 and d5. White
must act in the centre while Black's
king is uncastled and his lead in devel­
opment can make a difference.
The first two of White's major
plans in this position are introduced
by playing 13 dxe6 fxe6.
The exchange on e6 is a double­
edged decision, as it increases the po­
tential activity of Black's pieces.
Black's light-squared bishop has one
less pawn to bite against on the a8-h 1
diagonal and the d- and f-files are
M e r a n Va ria tio n : M a in L in e
opened for Black's rooks. The d-file is
particularly valuable, since with the
help of a rook on d8 Black can estab­
lish a knight on the strong d3-outpost
provided by the pawn on c4.
The strength of d5xe6 is that by
creating an isolated pawn on e6, White
loosens the protection afforded to the
black king. This gives him the chance
to cause Black some discomfort by
attacking e6 with liJg5 or liJd4, and
also furnishes him with the dangerous
idea of e4-e5, attacking the knight on
f6. If Black captures this pawn on e5,
this will allow the white major pieces
to directly attack the e6-pawn in front
of the black king. On the other hand,
if Black moves his knight, then Black's
kingside becomes vulnerable. The
white bishop on c2 attacks g6 and h7
along the newly-opened b1-h7 diago­
nal, while the white queen can now
check the black king from h5.
Incidentally, 1 1 d5 does rather
'wrong foot' Black, since an irony of
these lines is that Black would almost
rather have his bishop back on c8,
defending the weak pawn on e6!
We shall now (finally!) examine the
specific ways in which White has
sought to attack the black position.
The first four games in this chapter
deal with 13 dxe6 fxe6 14 liJd4, while
13 dxe6 fxe6 14 liJg5 is considered in
Games 65-69. Finally, the immediate
13 liJd4 is seen in Games 70-73.
Game 61
Lautier-Gelfand
Amsterdam 1996
1 d4 d 5 2 c4 c6 3 ttJf3 ttJf6 4 ttJc3
83
Th e S e m i - Sla v
e6 5 e3 tLlbd7 6 �d3 dxc4 7 �xc4
b5 8 �d3 �b7 9 0-0 a6 1 0 e4 c5 1 1
d5 c4 1 2 .ltc2 'iii c 7 1 3 dxe6 fxe6 1 4
tLld4
It took White players a long time to
realise the full strength of the knight
on d4. Obviously, it inconveniences
Black by attacking e6 and frees the
white f-pawn to move to f4, support­
ing e4-eS, but that is also true of 14
tZJgS, the old main line. But 14 tZJd4
has two advantages over 14 tZJgS. First,
the white knight on d4 combines with
the knight on c3 against Black's bs­
pawn, offering the possibility of a
knight sacrifice on bS if Black ever
castles queenside. Second, after 14
tZJgS, White's knight is easily removed
by . . . h7-h6; but this is not so easy after
14 tZJd4.
Note that after 14 'i'e2 Black can
play 14 . . . .td6, preparing to castle
kingside and ready to meet either 15
tZJgS or 15 tZJd4 with IS ... tZJcS.
1 4 . . . tLlc5
The best way to defend the e6-
pawn. 14 .. .'�JcS brings the knight in
contact with the outpost on d3 and
frees the d-file so that Black can bring
a rook to d8 to bear on the d4-knight.
84
1 5 �e3 !
This move completes White's set-up
by supporting the knight on d4. After
Black brings a rook to d8 , White can
move his queen to e2 or f3, side­
stepping the pin along the d-file. Black
will then be vulnerable to b2-b4, at­
tacking the knight on cS, the sole pro­
tector of the e6-pawn. Sooner or later,
Black will have to force the knight
from d4 and the only way to do so is
via the committal . . . e6-eS.
Question 5: But ... e6-eS just looks
good to me.
Answer: This move does indeed
have many positive points: Black
chases the white knight from d4 and
prevents the push e4-eS by occupying
the eS-square with a pawn himself;
removes the barrier on the d-file,
thereby making it possible to support
a knight on the d3-outpost with a
rook on d8; and creates a strong out­
post on d4 which the black knight on
cS can reach via e6. The negative side
to . . . e6-eS is that it weakens the light
squares, particularly fs and ds. After
. . . e6-eS, White can also envisage a2-a4
(softening up the black queenside)
. . . bS-b4, tZJdS - even as a pawn sacri­
fice. After Black captures on dS,
White can recapture with e4xdS, acti­
vating White's light-squared bishop by
opening the b I-h7 diagonal and open­
ing up the e-file for his major pieces.
1 5 0-0-0
. . .
Black gets his king out of the centre
and his rook to the d-file in one move,
and threatens to win a piece with . . . e6-
eS due to the pin on the d4-knight.
The drawback is that the black king is
not safe on the queenside. As we shall

see, Black can be a pawn up with no
immediate threats against him, but
because his king is exposed, White
always seems to have good chances.
The immediate lS . . . eS is dealt with in
Game 64.
16 'i'e2 e5 1 7 liJf3 !
This is the key to White's whole
concept. The first strength of 17 lbf3
is its calmness: White shows that he
understands that the black king on c8
is a long-term problem and will not
run away. Second, the knight is well­
placed on f3, where it ties the black
queen on c7 to the defence of the eS­
pawn, combines well with White's
idea of lbdS - since it leaves the b 1-h7
diagonal free for the bishop on c2 -
and is free to jump to gS, attacking the
e6-square.
1 7 . . . liJcxe4?
A bad decision. Although Black
wins a pawn, he does nothing to con­
tribute to his own activity. 15 . . . 0-0-0
was an active move and Black needs to
continue in the same vein. With
17 ... lbcxe4, Black is accepting a very
meagre price for his gamble of placing
the king in a vulnerable position. The
improvement 17 . . . lbe6 is considered in
Games 62 and 63.
1 8 liJxe4 liJxe4 1 9 a4! liJc5 20 axb5
axb5 2 1 b3!
Very simple and very effective -
White just opens lines on the queen­
side and lets Black do the worrying!
21 . . . cxb3 22 �f5+ �b8 23 'i'xb5 96
24 �h3 l:I.d5 25 l:Ub 1 ! 'i'c6 26 'i'c4
�e7 27 liJd2! l:hd8 28 liJxb3 'i'a4? ! ?
A fantastic saving attempt that Gel­
fand found with only a couple of
minutes left on his clock, but unfor-
M e r a n Va ria tio n : M a in L in e
tunately it is insufficient.
29 liJxc5! l:td 1 + 30 l:I.xd 1 lbd 1 +
3 0. . :iVxc4 loses to 3 1 �xd8+ .i.xd8
32 lbd7+ Wc7 33 lic1 picking up the
queen, according to Lautier.
3 1 l:txd 1 'i'xc4 32 liJd7+ 1 -0
32 . . . Wc7 33 �c 1 wins.
Game 62
Gelfand-Akopian
Yerevan Olympiad 1996
1 d4 d 5 2 c4 c6 3 liJc3 liJf6 4 e3 e6
5 liJf3 liJbd7 6 �d3 dxc4 7 i.xc4 b5
8 �d3 �b 7 9 0-0 a6 1 0 e4 c5 1 1 d5
c4 1 2 �c2 'i'c7 1 3 dxe6 fxe6 1 4
liJd4 liJc5 1 5 �e3! 0-0-0 1 6 'i'e2 e5
1 7 liJf3! liJe6
Gelfand was obviously impressed
with the previous game since he soon
decided to give it an outing from the
other side of the board! Black's 17th
move is a definite improvement, as
White must react to the threat of
... lbd4.
1 8 l:[ad 1 ! ?
Since White is planning to attack on
the queenside, it would seem more
natural to keep the rook on al and
play the rook on f1 to the d-file.
85
Th e S e m i - Sla v
However, Gelfand intends to first sof­
ten up the queenside with a2-a4 . . . bS­
b4, and then play tt:Jds . In this case,
the a-file will remain closed, so there is
no point in keeping a rook on a l . The
sacrificial 1 8 tt:JdS is the subject of the
next game.
1 8 . . . Sl.d6 1 9 a4! lLld4
19 . . . b4 20 tt:JdS tiJxdS 21 exdS is
clearly better for White, according to
Gelfand.
20 .lixd4! exd4 2 1 lLlxd4
This unstereotyped capture on d4
has given Black enormous problems:
the bS-pawn is attacked and the fork
tiJe6 is threatened.
21 . . . .lixh2+ 22 'ith 1 �f4 23 g3
.lixg3 24 fxg3 �xg3 25 lLlf5 �e5 26
�h2 �xh2+ 27 'itxh2 'itb8 28 axb5
White has a clear advantage, but
from here onwards he begins to let
things slip and this, together with a
sterling defensive performance from
Akopian, contributes towards a drawn
result.
28 . . . axb5 29 lLlxb5 lLlxe4 30 l:txd8+
l:txd8 31 lLlfd4 lLld6 32 lLle3 l:te8 33
l:tf2 h5 34 l:te2 l:txe2+ 35 lLldxe2 g5
36 lLld4 h4 37 .lid 1 lLlf7 38 lLlf5 lLle5
39 lLle3 .lie8 40 .lie2 Sl.e6 41 lLle4
86
g4 42 lLle5 g3+ 43 'itg 1 .lig8 44 lLlf5
'ite7 45 lLlxh4 'itb6 46 lLla4+ 'ita5 47
lLle3 �b4 48 lLlf5 'itb3 49 lLld 1 lLle6
50 �g2 Sl.e6 5 1 lLlfe3 lLld4 52 �h5
e3 53 bxe3 lLlb5 54 e4 .lixe4 55
'itxg3 lLle3 56 lLlf2 lLld5 57 lLlf5 lLlf6
58 �d 1 + 'itb4 59 lLle3 'ite5 60 �e2
'itd4 61 �f3 Sl.d5+ 62 �e2 �e4+ 63
'itd2 �e6 64 Sl.g6 lLld5 65 lLle2+
'ite5 66 lLld3+ �f6 67 .lie4 lLle7
%-%
The next game shows a crazy alter­
native path for White. I really don't
feel that it is necessary to play like
this, but it made for a great game!
Game 63
K rasenkov-Schandorff
Copenhagen (Politiken Cup) 1 996
1 d4 d5 2 e4 e6 3 lLle3 lLlf6 4 e3 e6
5 lLlf3 lLlbd7 6 �d3 dxe4 7 �xe4 b5
8 Sl.d3 Sl.b7 9 0-0 a6 1 0 e4 e5 1 1 d5
e4 1 2 Sl.e2 �e7 1 3 dxe6 fxe6 14
lLld4 lLle5 1 5 Sl.e3 0-0-0 1 6 �e2 e5
1 7 lLlf3 lLle6 1 8 lLld5 ! ?
Thematic.
1 8 . . . lLlxd 5 1 9 exd5 .lixd5 20 a4! b4
21 l:tad 1 g6
Stopping the bishop on c2 from ac-

tivating itself via f5.
22 l:txd 5 ! ?
A typical exchange sacrifice, freeing
a myriad of light squares for White's
own light-squared bishop and depriv­
in g Black of a crucial defensive piece.
22 . . . :txd5 23 �e4 l:td8 24 l:tc1 �b8
25 g3 i.c5?
A mistake according to Krasenkov
who prefers 2S . . . l::.c 8 26 �dS 'ifd6 27
.ixc4 with compensation for the sacri­
ficed material.
26 l:txc4 �d6 27 i.xc5 lLlxc5 28
l:txc5? !
An exchange sacrifice too far! 28
:txb4+ �c8 (28 . . .Wa7 29 'ii'e 3 wins) 29
'i'e3! lid7 30 lbxe5! liie 7 (30 .. :iVxeS 3 1
.ib7+ wins) 3 1 f4 would have given
Black terrible problems according to
Krasenkov.
28 . . :ii'xc5 29 ii'xa6 r3;c7 30 lLlg5
�d7 3 1 ii'e6+ '1t>c7 32 'it'a6 �d7 33
..-b7+ '1t>d6 34 lLlf7+ r3;e6 35 lLlxd8+
l:txd8 36 'it'xh7 l:tf8 37 ii'xg6+ r3;e7
38 �g5+ l:tf6 39 ii'g7+ l:tf7 40 ii'g5+
l:tf6 41 ii'd2 l:td6 42 �e2 ii'c 1 + 43
'1t>g2 l:td2 44 ii'h5 �xb2 45 ii'g5+
�e6 46 'i6'f5+ r3;e7 47 'i6'g5+ '1t>e6 48
-.f5+ r3;e7 49 h4 �d4 50 a5 b3 5 1
a 6 b 2 5 2 �h7+ '1t>d6 5 3 a 7 b 1 -. 54
M e r a n Va ria tio n : M a in L in e
i.xb 1 'i6'xf2+ 5 5 r3;h3 �g2+ 5 6 �g4
l:d4+ 57 �h5 'iWf3+ 58 '1t>h6 �xg3
59 h5 'it'e3+ 60 <;t>g7 ii'g5+ 61 -'g6+
�xg6+ 62 hxg6 l:ta4 63 i.e4 l:ba7+
64 '1t>f6 l:tc7 65 g7 Y2 - Y2 !
Recently, Black players have pre­
ferred to expend a little more time in
order to put the king to the relative
safety of the kingside.
Game 64
Krasenkov-Oll
Polanica Zdroj 1996
1 d4 d5 2 c4 c6 3 lLlc3 lLlf6 4 e3 e6
5 lLlf3 lLlbd7 6 i.d3 dxc4 7 i.xc4 b5
8 i.d3 i.b7 9 0-0 a6 1 0 e4 c5 1 1 d5
c4 1 2 dxe6 fxe6 1 3 i.c2 �c7 14
lLld4 lLlc5 1 5 i.e3 e5
Black chases the knight from d4
immediately.
1 6 lLlf5
Black wants to develop his dark­
squared bishop and then put his king
to safety. However, 16 . . . .ie7 leaves g7
hanging, while 16 . . . g6 is met by 17
lbh6, stopping the black king from
castling kingside. Although the knight
would then be precariously placed on
87
Th e S e m i- Sla v
h6, it does have a strong move back­
wards to g4, removing the knight on
f6 and thus increasing the strength of
ctJd5.
1 6 . . J:td8!
16 ... ctJcxe4 17 ctJxe4 ctJxe4 18 a4
'iVc6 19 axb5 axb5 20 Iha8+ i.xa8 2 1
'iVg4 gave White good attacking
chances in Lautier-Kramnik, Monaco
(rapidplay) 1996.
1 7 'iVf3 �d6 1 8 l::t a d 1 �c8?
This rather timorous move leads di­
rectly to disaster. The only consistent
move in this position is 1 8 . . . 0-0, when
Krasenkov suggests 19 'iVg3, putting
extra pressure on g7 and threatening
20 llxd6! �xd6 2 1 i.xc5 winning eve­
rything, as 2 1 . . .'iVxc5 loses to 22
'iVxg7+ mate. However, as Krasenkov
points out, Black can play 19 . . . ctJfxe4!
here.
20 ctJxe4 is met not by 20 . . . ctJxe4,
when 2 1 i.xe4 i.xe4 22 ctJxd6 l:1xd6
23 'iVxe5 i.d3 24 i.c5! �d7 25 'iVxc7
l1xc7 26 i.. xf8 i..x fl 27 .l:id8 wins for
White due to the double threat of 28
.td6+ and 28 'it>xfl, but by 20 . . . .txe4!
(Krasenkov) , when 2 1 i.. x c5 (not 2 1
ctJxd6 i.xc2; while 2 1 .l:ixd6 l:1xd6 22
�xc5 i.. x f5! 23 'iVxe5 .l:id7 24 'iVxc7
88
l::txc7 25 i.. xf8 .txc2 draws and 2 1
ctJh6+ 'it>h8 2 2 i..xc5 .txc5 2 3 i.. xe4
gxh6 is nice for Black) 2 1 . . .i.. xf5 22
i..xd6 �xd6 23 'iVxe5 :d7! holds for
Black. 20 i.. xe4 by contrast is met by
20 . . . ctJxe4 2 1 ctJxe4 :xf5! with a good
position for Black.
So it seems that 18 . . . 0-0 is sufficient
Gust!) for Black.
1 9 iLld5! iLlxd5 20 iLlxd6+ l::t x d6 21
'iWh5+!
A very uncomfortable move to face.
2 1 . . .g6 10ses to 22 'iVxe5+.
21 . . . l::t g 6 22 exd5 iLld3
Forced, to block the bishop's attack
on the rook on g6.
23 �xd3 cxd3 24 l::t c 1 'it'b8 25 �a7 !
1 -0
A nice finish! Krasenkov shows that
25 . . . 'iVa8 loses to 26 'iVxe5+ .te6 27
:c6! (White must be careful; 27 dxe6
loses to 27 . . . 'iVxg2+ mate!) 27 . . . 0-0 28
.l:ixe6.
We now move on to a line that be­
came popular after some Karpov
magIC.
Game 65
Karpov-Kramni k
Linares 1 994
1 d4 d5 2 c4 c6 3 iLlf3 iLlf6 4 iLlc3
e6 5 e3 iLlbd7 6 �d3 dxc4 7 i.xc4
b5 8 i.. d 3 a6 9 e4 c5 1 0 d5 c4 1 1
dxe6 fxe6 1 2 i.. c 2 �b 7 1 3 0-0 'iVc 7
1 4 iLlg5 iLlc5 1 5 e5!
The plan with 14 ctJg5 demands
much more urgency from White than
14 ctJd4, since after 14 . . . ctJc5, Black is
already threatening to neutralise
White's pressure on e6 by chasing

away the knight with . . . h7-h6. Kar­
pov's idea forces the game into a com­
plicated endgame.
1 5 .. .'iVxe5 1 6 l:te 1 'iVd6 1 7 �xd6!
.ixd6 1 8 .ie3
IS ttJxe6 ttJxe6 19 �xe6+ '1t>d7 is
harmless for Black, who is ready to
take advantage of his queenside pawn
maJonty
1 8 . . . 0-0
Karpov dismisses IS . . . ttJd3 due to 19
.ixd3 cxd3 20 �ad l , but in Kramnik­
Kuczynski, German Bundesliga 1994,
Black held the position quite easily
after 20 . . . 0-0 2 1 �xd3 .idS 22 ttJxe6
.ixh2+ 23 '1t>xh2 .ixe6 24 l:1d6 �feS. I
have not seen this idea repeated since.
1 9 l:tad 1 !
Gaining a tempo on the bishop on
d6.
1 9 . . . .ie 7 20 .ixe5 .ixe5 21 liJxe6
IUe8 22 h3! !
At first sight, White's play seems
unimpressive. Although he has some
activity - his two rooks occupy the
open central files, and the e6-knight
attacks the bishop on cS and the pawn
on g7, while covering the dS-square -
Black has the two bishops and a men­
acing queenside majority. However, a
M e r a n Va ria tio n : M a in L in e
closer analysis reveals that Black's
pieces lack coordination; it is hard for
Black to generate any activity. For
example, the bishop on cS attacks f2,
but Black is unable to utilise this:
... ttJg4 was prevented by 22 h3!! and
Black is unable to bring a rook to the
f-file. Moreover, by removing the e6-
pawn, White has Black's outpost on
dS, leaving the black knight stuck on
f6.
Question 6: Why does this matter?
Answer: This is the genius of Kar­
pov's concept. 22 h3 prepares g2-g4-gS
to drive the black knight away from
f6. If the knight can be forced to a pas­
sive or wayward square, then White
will be able invade the seventh rank
with l1d7 or play ttJdS, bringing
Black's position to breaking point.
22 . . . .if8? !
Although the text is solid, it does
nothing to interfere with White's
plan. 22 ... 1:abS is considered in Games
66-6S.
23 g4! h6 24 f4!
Reinforcing the threat of g4-gS.
24 . . . .if3 25 l:td2 .ie6
To give the knight a square on d7
after White attacks it with g4-gS.
89
Th e S e m i - Sla v
26 g5 hxg5 27 fxg5 tLld7 28 tLlxf8
tLlxf8 29 .t:rd6 b4 30 tLle4
30 CDd5 was possibly even stronger
according to Karpov.
30_ .. .Jie8 31 tLlg3 .t:rd8 32 tLlf5 .t:rxd6
33 tLlxd6 iLg6 34 iLxg6 tLlxg6 35
tLlxc4 .t:rd8 36 .t:re4 b3 37 axb3 .t:rd3
38 '5t>g2 .t:rxb3 39 h4 tLlf8 40 .t:re8 1 -0
In this very difficult posltIOn,
Kramnik lost on time.
Game 66
Gelfand-Shirov
Biel 1995 p port the advance of the b-pawn.
1 d4 d 5 2 c4 c6 3 tLlf3 tLlf6 4 tLlc3
e6 5 e3 tLlbd7 6 iLd3 dxc4 7 .Jixc4
b5 8 iLd3 a6 9 e4 c5 1 0 d 5 c4 1 1
.Jic2 'V/ic7 1 2 dxe6 fxe6 1 3 0-0 .Jib7
1 4 tLlg5 tLlc5 1 5 e5 'V/ixe5 1 6 .t:re 1
'V/id6 1 7 �xd6 iLxd6 1 8 iLe3 0-0 1 9
.t:rad 1 iLe 7 20 .Jixc5 .Jixc5 2 1 tLlxe6
.t:rfc8 22 h3! ! .t:rab8 !
A superb defensive concept! Black
fully understands the value of keeping
his dark-squared bishop on the a7-g1
diagonal. By pinning the f2-pawn to
White's king on g l , he prevents White
from advancing this and fully activat­
ing his kingside majority. Moreover,
90
the bishop restricts the movement of
White's rooks by covering the d4- and
e3-squares. Since 22 . . . .tb6 and
22 . . . .ia7 are both met by 23 CDxg7!
�xg7 (23 . . . ..txg2 24 �xg2 [24 CDf5
SLxh3 25 CDe7 + �f7 is more risky for
White] 24 . . . �xg7 25 l:re7+ is also diffi­
cult for Black) 24 l:re7+ winning the
loose bishop on b7, Black defends this
piece, preparing to retreat the dark­
squared bishop along its best diagonal
without incurring material loss. The
rook is also well-placed on b8 to sup­
23 tLlxc5
23 g4 (Game 67) is more critical,
while 23 a3 is seen in Game 68 .
23_ . . .t:rxc5 24 .t:rd6
With 23 CDxc5 .l::xc5, White aims
just for a safe, stable endgame edge.
The removal of the troublesome
bishop on c5 costs Black the bishop
pair and frees several important dark
squares for the white rooks to exploit.
24 l:rd6 floats the threat of l:rb6, pin­
ning the b7-bishop, and thereby hopes
to slow Black's counterplay on the
queenside. Of course, 24 a3 is just met
by 24 . . . a5, renewing the threat.
24 . . . b4
Although this move loosens Black's
position a little, it gains space on the
queenside and by chasing the knight
away from c3, allows Black to chal­
lenge White's control of the d-file by
. . . �d5.
25 tLla4 .t:rd5! 26 .t:rb6 .t:rb5
Yagupov criticises this move, claim­
ing an edge for White after 28 CDc5,
and I agree with this assessment.
Black's counterplay is based on an at­
tack on the backward pawn on a2

with a rook on a8, but before he can
do this, he must move his bishop from
b7. Unfortunately, the natural and
desirable 28 . . . .id5 loses a pawn to 29
ttJa6! !:ta8 30 lbxb4!
Black should therefore have played
Yagupov's 26 . . . a5, protecting the b4-
pawn and intending to drive the rook
from b6 with . . . lbd7. Yagupov gives
27 lie7 lbd7 28 l'lbe6 lbf8 29 .l::I b 6 lbd7
with a draw by repetition.
27 nxb5 axb5 Y2 - Y2
Now we turn our attention to the
most aggressive and consistent idea for
White, 23 g4.
Game 67
G reenfeld-Av. Bykhovsky
Beersheva 1996
1 d4 d 5 2 e4 e6 3 lLle3 lLlf6 4 e3 e6
5 lLlf3 lLlbd7 6 .Jtd3 dxe4 7 �xe4 b5
8 �d3 .Jtb7 9 0-0 a6 1 0 e4 e5 1 1 d5
c4 1 2 dxe6 fxe6 1 3 .Jte2 'fie7 1 4
lLlg5 lLle5 1 5 e 5 'ii'x e5 1 6 l:te1 �d6
1 7 'fixd6 �xd6 1 8 .Jte3 0-0 1 9 %:tad 1
�e7 20 �xe5 �xe5 2 1 lLlxe6 %:tfe8
22 h3! ! %:tab8! 23 g4
23 . . . �f3
This annoying intermediate move
M e r a n Va ria tio n : M a in L in e
disrupts the coordination o f White's
rooks.
24 %:td2 %:te8 ! ?
A new idea. Black accepts that his
knight will be driven to the side after
g4-g5. However, by pinning the
knight on e6, he hopes to be able to
activate the knight with . . . lbf4.
The alternative 24 . . . b4!? 25 lba4
�a7 26 g5 lbd5 has been the subject of
intensive high-level testing. After 27
g6! h6 (27 . . . hxg5 28 lbg5!) 28 lbd4 c3!
29 bxc3 bxc3 30 �d3 lbb4! 31 �xf3
.txd4 32 .tf5 �c7 33 a3 lbc6 34 \tg2
(and not 34 :f4?? �e5 when Black
was winning in Nikolic-Shirov, Hor­
gen 1994) 34 . . . lbe7 35 �c2 �f6, as in
Alterman-Akopian, Haifa 1995, 36
1:idl Wf8 37 l'le3 would have kept a
small edge according to Alterman.
25 �d 1
The calmest approach. 25 �fl .tb4!
26 a3 (26 g5 lbd5! is good for Black)
26 . . . �xc3 27 bxc3 �b6! 28 lbf4 (28
lbg5+ .tg2+! wins the exchange)
28 . . .l::txe l+ 29 <it'xe l g5! presented
Black with no problems in Filgueira­
Sorokin, Villa Balester 1996.
25 . . . �xd 1 26 l:tdxd 1 %:tbe8 27 g5
lLlh5 28 lLld5 �d6 29 l:td4 %:te6 30
%:tde4 %:tb8 31 h4 b4
White has an impressive looking
position, but it is not easy to do much
with it whereas Black has the simple
plan of creating a passed c-pawn.
32 %:te 1 %:te8 33 'it>f1 �f8 34 %:tee 1 e3
35 bxe3 bxe3 36 lLld4 %:txe4 37
lLlxe6 %:txh4 38 'it>e2 l:te4+ 39 lLle3
lLlf4+ 40 �f3 lLld3 41 <;t>xe4 lLlxe 1
42 lLld4 �e5 43 lLlb3 �a7 44 f4 e2
45 f5 .Jib6 46 lLld5 �e5 47 lLlf4 .Jta3
48 lLle2 .Jie 7 49 f6 gxf6 50 gxf6
91
Th e S e m i - Sla v
.lixf6 5 1 liJc5 a5 52 liJd3 liJg2 53
�f3 liJh4+ 54 �g4 h 5+ 55 �xh5
liJf3 56 liJdc 1 �b2 57 liJd3 liJd4 58
liJec 1 �c3 59 �g5 �d2+ 60 �f6
�f8 6 1 a3 a4 62 liJa2 liJe2 63 �e5
liJc3 64 liJxc3 .lixc3+ 65 �d5 �e7
66 �c4 �d2 67 �b5 Y2 - %
Game 68
Azmaigarashvili-Akopian
yerev�n Olympiad 1 996
1 d4 d5 2 c4 c6 3 liJc3 liJf6 4 e3 e6
5 liJf3 liJbd7 6 �d3 dxc4 7 �xc4 b5
8 �d3 .lib7 9 0-0 a6 1 0 e4 c5 1 1 d5
c4 1 2 �c2 VJkc7 1 3 dxe6 fxe6 1 4
liJg5 liJc5 1 5 e 5 VJkxe5 1 6 l:le 1 'i'd6
1 7 'iVxd6 �xd6 1 8 .lie3 0-0 1 9 l:lad 1
�e7 20 �xc5 �xc5 2 1 liJxe6 l:lfc8
22 h3 ! ! l:lab8! 23 a3
This prevents . . . b5-b4 and hopes to
stop any queenside counterplay.
23 . . . �b6! 24 l:ld6 �a5 25 l:le3 b4!
26 axb4 �xb4 27 l:ld4 l:le8 28
l:lxc4?
A mistake according to Azmai­
parashvili who recommends instead 28
{jjc7 �xe3 29 fxe3 .1i.c5 30 Ilxc4
.1xe3+ 3 1 'it>h 1 , assessing it as unclear.
28 . . . .lixc3 29 l:lcxc3 liJd5 30 �b3
92
liJxe3 31 liJc7+ liJd5 32 liJxe8 l:lxe8
33 l:lc 7 �a8 34 l:ld7 l:lb8
This leads to an easy draw and yet,
as pointed out by Ivanchuk, 34 . . .'it>f8!!
35 �xd5 Ile 1+ 36 'it>h2 We8 ! 37 �xg7
�xd5 38 1::[xh7 1::[e 2! would have given
Black the advantage.
35 �xd5+ �xd5 36 l:lxd5 l:lxb2 37
l:ld8+ �f7 38 l:ld7+ �g8 39 l:la7 l:lb6
40 �h2 l:lc6 41 f3 h6 42 �g3 Wh7
43 f4 �g6 44 �f3 l:lb6 45 g3 % - %
This is not the whole story of this
variation because, in typically inven­
tive fashion, Shirov has devised an­
other risky idea. To quote him,
' 15 . . . 'i'c6 has not been refuted yet, but
it looks dangerous!'
Game 69
Kamsky-Shirov
Madrid 1994
1 d4 d5 2 c4 c6 3 liJc3 liJf6 4 liJf3
e6 5 e3 liJbd7 6 .lid3 dxc4 7 �xc4
b5 8 .lid3 a6 9 e4 c5 1 0 d5 'VJIic7 1 1
0-0 �b 7 1 2 dxe6 fxe6 1 3 �c2 c4
1 4 liJg5 liJc5 1 5 e5 'i'c6 ! ?
Threatening mate o n g2, and thus
forcing White to block the queen's
path to the hS-square.

1 6 f3 lLlfd7 1 7 lLlxh 7
17 ";!Ve2 was played in I.Sokolov­
Shirov, Leon 1995. White claims that
achieving e4-e5 is more than enough
for an advantage and avoids risk with
l2Jxh7. However, 17 . . . lbd3!? (17 . . . �e7
18 �h 1 lbd3 19 �xd3 cxd3 20 'i'xd3
l2Jxe5 2 1 'i'd4 'iVc4 22 iYxc4 lbxc4 23
l2Jxe6 'it>f7 gave Black good compensa­
tion for the pawn in Krasenkov­
Luther, Tilburg 1994) 18 �xd3 cxd3
19 'ii'xd3 lbxe5 20 'iVe2 h6! 21 'iVxe5
'l'c5+ 22 'i'xc5 �xc5+ 23 'it>h l hxg5
24 �xg5 �d4! was only minutely bet­
ter for White.
1 7 . . . lLlxe5 1 8 lLlxf8 l:txf8 1 9 li'd4
lLlcd3!
Black is now threatening 20 . . Jhf3!
21 gxf3 lbxf3+ 22 lixf3 'ifxf3 followed
by mate!
20 �xd3 lLlxd3 21 lLle4 e5?!
A slight inaccuracy. In a later round
of the same tournament against Illes­
cas, Shirov played 2 1 . . J�d8 ! 22 'iVxg7
'i'b6+! 23 lbf2 lbxf2 24 fig6+ 1:1£7 25
'i'g8+ l:tf8 26 'i6g6+ !!f7 with a draw,
as 27 1:txf2 allows 27 . . . lId1+ mate.
22 .e3 0-0-0 23 b3 c3 24 �a3 b4
25 l:[fd 1 'ii' b 5 26 lLlxc3 llxf3 27 gxf3
bxc3 28 l:tac 1 lLlxc 1 29 "xc3+ �c6
30 'ii'x c6+ �xc6 31 l:txd8+ �xd8
YZ - YZ
And now the last of our 'big three'
main lines: 13 lbd4
Game 70
Gelfand-Dreev
Tilburg 1993
1 d4 d 5 2 c4 c6 3 lLlc3 lLlf6 4 lLlf3
e6 5 e3 lLlbd7 6 �d3 dxc4 7 �xc4
b5 8 i.d3 i.b7 9 0-0 a6 1 0 e4 c5 1 1
M e r a n Va ria tio n : M a in L in e
d5 c4 1 2 �c2 'ii'c 7 1 3 lLld4
By avoiding the preliminary d5xe6,
White forgoes two benefits: the loos­
ening of the pawn cover around the
black king; and the possibility of e4-e5
(which is now met by . . . lbxd5) . How­
ever, by maintaining the d5-pawn,
White aims to hold the knight on d4 -
and thus his pressure on e6 - indefi­
nitely. First, the d-pawn shields the
knight from attack along the d-file.
Second, . . . e6-e5 is met by lbf5, after
which Black cannot develop his
bishop from f8 as g7 will hang (unlike
after 13 dxe6 fxe6, when the queen on
c7 protects the g7-pawn), so he will
find it hard to put his king to safety
on the kingside.
1 3 . . . e5
13 ... exd5 14 lbxd5! lbxd5 15 exd5
0-0-0 16 a4 b4 17 lbc6 gives White a
powerful attack, while 13 . . . lbc5 is the
subject of the next game.
1 4 lLlf5 g6 1 5 lLlh6 lLlh5
I don't like Black's position here. 16
g3 would now have kept the knight
out of f4 and given White a definite
advantage.
1 6 �f3 lLlf4 1 7 lLlxf7 ! ?
Dreev suggests 17 lbg4!? here.
93
Th e S e m i - Sla v
1 7 . . . \t>xf7 1 8 g3 g 5 1 9 gxf4 gxf4 20
�h5+ �e7 2 1 �h4+ �f7 22 iLd 1 ? !
White would have done better to
repeat moves with 22 ifhS+. The rest
of the game is simply stunning!
22 . . .1::t g8+ 23 �h 1 ltJf6 24 iLh5+
lIg6 ! ! 25 iLxg6+ hxg6 26 lIg 1 iLe7
27 �h 6 lIg8 28 f3 b4 29 ltJe2
ltJxd 5 ! ! 30 exd 5 iLxd5 31 lIf 1 Jif6
3 2 �h7+ lIg7 3 3 �h3 iLe6 34 �g 2
g 5 3 5 a3 g4 36 axb4 iLd5 37 ltJc3
gxf3 38 �f2 iLb7 3 9 lIa5 �d 7 40
ltJd5 iLxd 5 41 �d2 iLc6 42 �xd 7+
Jixd7 43 lIxa6 il.h3 44 lIf2 il.h4 0- 1
Game 71
Sadler-Madwek w e
, London (Lloyds Bank) 1 994
1 d4 d 5 2 c4 c6 3 e3 ltJf6 4 ltJc3 e6
5 ltJf3 ltJbd7 6 il.d3 dxc4 7 il.xc4 b 5
8 il.b3 il.b7 9 0-0 a6 1 0 e 4 c 5 1 1 d 5
c4 1 2 Jic2 � c 7 1 3 ltJd4 ltJc 5 !
This is Black's best choice.
1 4 b4
For 14 iff3 see Game 73 .
1 4 . . . cxb3 1 5 axb3 b4!
Preventing White from chasing
away the black knight from cS with
b3-b4 and weakening the white centre
by forcing the knight offside to a4.
1 6 ltJa4 ltJxa4?
This is definitely wrong, although
the refutation is far from obvious. The
correct 16 . . . 4Jcxe4! is seen in the next
game.
1 7 lIxa4 exd 5 1 8 e 5 ! ltJe4
18 . . . ifxeS fails to 19 l:te 1 4Je4 20 f3
(20 l:taS!? ifc7 2 1 l:txdS .i.xdS 22 �xe4
0-0-0 23 ifg4+ is also very dangerous
for Black) 20 . . . .i.cS 2 1 .i.e3, when
2 1 . . .4Jc3 22 �f2 wins.
94
1 9 il.f4! ! il.c5
A desperate attempt to get his pieces
out. Instead 19 . . . 4Jc3 20 iff3 4Jxa4 2 1
e 6 ifb6 2 2 bxa4 gives White a huge
attack, as 22 . . . ifxd4 10ses to 23 �eS.
20 il.xe4 dxe4 2 1 ltJf5 f6?
Losing. 2 1 . . .0-0 22 ifg4 g6 23 e6! is
also pretty horrible, however.
22 e6! �c6 23 Jid6 1 -0
Game 72
Kasparov-Kramnik
Dos Hermanas 1996 .
1 d4 d 5 2 c4 c6 3 ltJc3 ltJf6 4 ltJf3
e6 5 e3 ltJbd7 6 il.d3 dxc4 7 il.xc4
b5 8 Jid3 Jib7 9 0-0 a6 1 0 e4 c 5 1 1
d 5 c4 1 2 �c2 �c7 1 3 ltJd4 ltJc5 1 4
b4 cxb3 1 5 axb3 b4 1 6 ltJa4 ltJcxe4!
1 7 il.xe4 ! ?
An earlier game Yakovich-Sorokin,
Calcutta 199 1 , continued 17 dxe6 �d8
18 exf7+ 'it>xf7 19 .i.e3 .i.d6 20 h3
l:the8 21 l:tc1 , and now 2L .. 4Jc3 22
4Jxc3 bxc3 would have been unclear
according to Yakovich.
1 7 . . . ltJxe4 1 8 dxe6 il.d6 1 9 exf7+! ?
The most forcing move. White
players could also consider the tricky
19 ifhS 0-0 20 .tb2.

19 :�xf7 !
. .
This is best and almost forced, as
19 ... 'it>xf7 is unpleasantly met by 20
'iYh5+ g6 2 1 �h3 !
20 f3! �h5! 2 1 g3
21 . . . 0-0!
Kramnik states that 2 1 . . .lbxg3!! 22
hxg3 (22 �e 1+ lbe4 23 .ua2 0-0 gives a
virulent attack) 22 . . . 0-0! 23 kla2 .ixg3
24 ltg2 �e5 25 lbc5 �ad8 26 .ie3 .ic8
was Black's best chance, when White
can only save himself with 27 f4!
'iYxdl 28 �xdl .ixf4 29 .ixf4 �xf4 30
�gd2 with reasonable drawing
chances. The rest of the game is thus
not theoretically imponant, but play
through it - it has to be enjoyed!
22 fxe4 �h3! 23 lLJf3? .i.xg3 24
lLJc5 l:txf3 25 l:txf3 �xh2+ 26 �f 1
�c6!
This is what Kasparov had missed
when calculating his 23rd move.
27 �g5 .i.b5+ 28 lLJd3 l:te8 29 l:ta2
'i'h 1 +
Winning, but as Kramnik shows,
29 . . . it.xd3+ 30 �xd3 "i'h 1 + 3 1 'it>e2
�xg2+ 32 �e3 lixe4 was checkmate!
30 �e2 l:txe4+ 31 �d2 'i'g2+ 32
�c 1 'i'xa2 33 l:txg3 �a 1 + 34 �c2
'ii'c 3+ 35 �b1 l:td4 0-1
M e ra n Va ria tio n : M a in L in e
A magical game!
Game l3
Tkachiev-Handoko
Jakarta (match) 1996
1 d4 d5 2 c4 c6 3 lLJc3 e6 4 e3 lLJf6
5 lLJf3 lLJbd7 6 �d3 dxc4 7 .i.xc4 b5
8 �d3 .tb7 9 0-0 a6 1 0 e4 c5 1 1 d5
c4 12 �c2 �c7 13 lLJd4 lLJc5 1 4
'i'f3! ?
White suppons d5 and prepares to
put extra pressure on e6 by 'iWh3.
14 . . . .td6
14 . . . 0-0-0 is interesting, looking to
win the d5-pawn without giving up
the light-squared bishop.
1 5 'i'h3 b4 1 6 lLJa4 lLJcxe4 1 7 dxe6
0-0 1 8 f3 �e5 1 9 �e3 lLJg5 20
exf7+ l:txf7 21 'i'h4 l:td8 22 lLJf5
lLJe6 23 .tb6!
Winning the exchange and essen­
tially the game.
23 . . . 'i'b8 24 �xd8 lLJxd8 25 l:tad 1
.i.d5 26 lLJe3 'i'a7 27 'i'f2 b3 28
axb3 cxb3 29 lLJxd5 bxc2 30 lLJxf6+
.i.xf6 31 'i'xa7 l:txa7 32 l:td2 l:tc7 33
l:tc1 l:tc4 34 b3 l:tb4 35 l:td3 lLJe6 36
l:txc2 lLJf4 37 l:tc8+ �f7 38 l:td7+
�g6 39 l:tc4 1 -0
95
Th e S e m i - Sla v

S u m mary
I would recommend the 13 dxe6 fxe6 14 liJd4 line to White players since it com­
bines solidity and aggression and does not require the refined endgame skills de­
manded by Karpov's 14 liJg5 'i'xe5 15 e5. Tkachiev's 13 liJd4 liJc5 14 'i'f3 is also
worth a try. For Black players, I would suggest that 13 dxe6 fxe6 14 liJd4 is best
met by 14 . . . liJc5 1 5 iLe3 e5 .

1 d4 d 5 2 c4 c6 3 lOf3 lOf6 4 lOc3 e6 5 e3 lObd7 6 iLd3 dxc4 7 iLxc4 b5 8

iLd3 iLb7 9 0-0 a6 1 0 e4 c5 1 1 d5 c4 1 2 iLc2 'ilc7
1 3 dxe6
13 liJd4
13 . . . e5 Game 70
-

13 . . . liJc5 (D)
14 b4 cxb3 15 axb3 b4 16 liJa4
16 . . . liJxa4 Game 71; 16 . . . liJcxe4 Game 72
- -

14 'i'f3 Game 73
-

1 3 . . . fxe6 1 4 1Od4
14 liJg5 liJc5 15 e5
15 . . . 'i'xe5 16 �e 1 'i'd6 17 'i'xd6 .txd6 18 i.e3 0-0
19 �ad1 .te7 20 .i.xc5 .i.xc5 2 1 liJxe6 �fc8 22 h3 (D)
22 . . . ..tf8 Game 65
-

22 . . . �ab8
23 liJxc5 Game 66; 23 g4 Game 67; 23 a3 Game 68
- - -

15 . . . 'i'c6 Game 69
-

1 4 . . . lOc5 1 5 iLe3 0-0-0

15 . . . e5 Game 64
-

1 6 'iVe2 e5 1 7 1Of3 (D) lOcxe4

17 . . . liJe6
18 �ad1 Game 62; 1 8 liJd5 Game 63
- -

1 8 lOxe4 Game 61
-

13 . . . lOc5 22 h3 1 7 1Of3

96
 CHAPTER EIGH T

Meran Variation :
Move Orders and Sidelines

1 d4 d 5 2 c4 c6 3 li::lf 3 li::lf 6 4 li::l c 3 'understand' than the lines arising after

e6 5 e3 li::l b d7 6 �d3 dxc4 7 �xc4 8 . .ib7 9 e4 or 9 a3.
. .

b5 8 �d3 Question 2: Is that all?

This chapter will be of particular Answer: I'm afraid not! Black has
value for players who like to confuse great scope for inverting moves in the
their opponents! Since we only reach lead up to the main line position. For
the critical main line position that example, when does Black play his
arose in the previous chapter after 12 queen to c7? After 8 . . . .ib7 9 0-0 a6 10
moves, there is plenty of scope for e4 cS 11 dS
trickery and treachery along the way!
There is no definitive path which
Black should adopt to reach the main
line position. Kramnik prefers 8 . . . .tb7
9 0-0 a6 (threatening . . . c6-cS) 10 e4 cS
1 1 dS c4 12 .tc2 'ii'c 7; Shirov's favour­
ite, however, is 8 . . . a6 (threatening
. . . c6-cS immediately) 9 e4 cS 10 dS c4
1 1 .tc2 .tb7 12 0-0 'iWc7.
Question 1: What is the difference?
A nswer: After 8 . . . .ib7, Black must
also be prepared against 9 e4 and 9 a3,
whereas after 8 . . a6, Black must be
. is 1 1 . . .'iWc7 possible instead of
prepared for 9 e4 cS 10 eS!? 1 1 . . .c4 here? And after 8 . . . a6 9 e4 cS 10
...8 .tb7 is more solid than 8 ... a6 dS c4 11 .tc2, can Black play 1 1 . . .'iWc7
since 8 . . . a6 9 e4 cS 10 eS leads to wild before he develops his bishop to b7 ?
positions with random pawn struc­ In fact there are really only two im­
tures and is more difficult to portant issues for move orders: Black

97
Th e S e m i- Sla v
can delay or avoid . . . c5-c4 or, via the
8 . . . a6 move order (Games 74-78) ,
Black can delay or avoid . . . .tb7
(Game 79) .
The rest of this chapter is devoted
to systems in which Black plays an
early . . . e6-e5 (Games 80-8 1) or takes
the white bishop on d3 in response to
d5xe6 (Games 82-83) .
Black delays or avoids . . . c5-c4
The most important position in this
section arises from either 8 . . . �b7 9 0-0
a6 10 e4 c5 1 1 d5 'iYc7 or 8 . . . a6 9 e4 c5
10 d5 �b7 (or 10 . . . 'iYc7 1 1 0-0 iLb7) 1 1
0-0 'iYc7.
Usually instead of l 1 . . .'iYc7 Black
plays 1 1 . . .c4 to gain space on the
queenside and to increase the activity
of Black's minor pieces. The . . . . c5-c4
advance frees c5 for the black knight
and also allows Black to play his
king's bishop to d6. Without it,
12 . . . �d6 for example would fail to 13
dxe6 fxe6 14 �xb5 axb5 15 liJxb5
forking the queen and bishop.
Question 3: So why avoid playing
1 1 . . .c4 in that case?
A nswer: Delaying . . . c5-c4 has two
advantages. Although it deprives Black
of . . . liJc5 to defend e6, it prevents
White from attacking e6 with liJd4.
Moreover, in the event of d5xe6, the
white bishop on d3 will be a useful
target for a black rook on d8 .
Game 74
Bareev-Dreev
Russian Ch., Elista 1 996
1 d4 d 5 2 c4 c6 3 tLJf3 tLJf6 4 tLJc3
e6 5 e3 tLJbd7 6 iLd3 dxc4 7 �xc4
98
b5 8 iLd3 iLb7 9 0-0 a6 1 0 e4 c5 1 1
d5 "VtIlc7 1 2 dxe6 ! ?
A committal decision. The prophy­
lactic 12 i.c2 is considered in Games
75-77 and 12 b3, preventing . . . c5-c4, in
Game 78.
1 2 .. .fxe6 1 3 iLc2
White retreats the bishop from its
exposed position on d3 and thus draws
the sting from Black's d-file ambitions.
Despite abandoning the attack on b5,
White does not give Black the chance
to play 13 . . . .td6 due to 14 liJg5!, as
Black has neither . . . liJc5 nor . . . 'iYc6 to
defend e6.
Question 4: Can't Black just trans­
pose into the main line with 13 . . . c4?
A nswer: Yes, that is the normal
move.
1 3 . . . 0-0-0! ?
The maximum actIvIty and maxi­
mum risk approach. Black gets his
king out of the centre and brings a
rook to the d-file in one move, but
hides his king behind his queenside
pawns, which are better placed for
attack than defence.
1 4 tLJg5 tLJe5! 1 5 �e 1 ! ?
The 'normal' 15 'iYe2 would allow
Black's knight on eS to transfer to d4
 M e r a n Va ria tio n : M o v e O r d e rs a n d Side lin e s
via c6 with tempo.
1 5 . . . �b6 1 6 f4 liJd3 1 7 'ii'g 3 c4+ 1 8
i.e3 �c5 1 9 �xc5 liJxc5 !
A very nice move, keeping the
queen on b6, where it defends the e6-
pawn, and supporting . . . b5-b4 by cov­
ering the a4-square.
20 'it'h 1 b4 21 liJa4 liJxa4 22 �xa4
h6 23 liJf7 liJxe4 Y2 - %
This was a strange place to offer a
draw, as Black seems to have very
good compensation for the exchange.
Question 5: So what about 12 �c2
without exchanging on e6 first?
Answer: This is the most flexible
move; in all the games I have seen,
Black has either transposed to the
main line with 12 . . . c4 or played
12 ... ..te7. However, Black could also
try 12 . . . 0-0-0!? to meet 13 lLlg5 with
13 ... lLlb6 (but not 13 . . . lLle5 14 f4! lLlc6
15 dxc6! �xd1 16 cxb7+ and 17 lilxd1) ,
ganging up on the d-pawn.
Game 75
• Lautier-Dreev
Linares 1995
1 d4 d 5 2 c4 c6 3 liJc3 liJf6 4 e3 e6
5 liJf3 liJbd7 6 �d3 dxc4 7 �xc4 b5
8 �d3 �b7 9 0-0 a6 1 0 e4 c5 1 1 d5
�c7 12 �c2 �e7
Black makes the most of White's
unforced retreat with 12 �c2 by opt­
ing for a developing move rather than
playing the routine 12 . . . c4, transpos­
ing to the previous chapter.
1 3 dxe6 fxe6 1 4 liJg5 'ii' c 6
The queen is well placed here, de­
fending e6 and preventing a e4-e5 due
to . . . �xg2 checkmate!
1 5 �f3! ?
White reasons that without . . . c5-c4,
Black cannot deal comfortably with
two attacks on e6. 15 �f3 aims for the
h3-square, where it will combine with
the knight against the e6-pawn. From
h3, the white queen also protects g2
and allows White to consider the e4-e5
push; for example 15 . . . c4 16 'i'h3 lLlc5
is met by 17 e5! However, Black's idea
is much more daring! (See Game 77
for 15 f4.)
1 5 . . . h6!? 1 6 �h3 hxg5 1 7 'ii'x h8+
""f7 1 8 �h3 g4
I find this exchange sacrifice hard to
believe for Black, but it has caused
White serious problems. Although
White's material advantage should pay
in the end, by winning the rook on h8
99
Th e S e m i - Sla v
for his king's knight, he has given up a
very valuable piece for attacking the
black position. In the short term he
now has no real way of creating play
against the black position, and thus he
will have to take on the chin whatever
Black can throw at him. Black will
have to make the most of a mixture of
factors: the half-open h-file; pressure
against the e4-pawn; and the vulner­
able white queen.
1 9 'it'h4
To stop the black rook from com­
ing immediately to the h-file.
1 9 . . . lZle5 20 f4
20 l:td1 is considered in the next
game.
20 . . . gxf3 2 1 gxf3 lZlg6! 22 'iVg5
22 'i6g3 is met by 22 . . . i.d6 accord­
ing to Dreev.
22 . . J:th8! 23 e5 lZlh4!
A very nice way of countering
White's attack. Now 24 exf6 gxf6 25
'iWe3 (to play �e4) 25 .. .f5! gives Black
huge play for the rook according to
Dreev.
24 �e4 lZld5 25 'it'g4 b4 26 lZle2
'it'c7 !
Dreev now assesses the position as
clearly better for Black.
1 00
27 'iVg3 l:th5 28 �f4 lZlf5 29 'ii'e 1 c4
30 �h 1 �h4 3 1 'iVd2 'fie7 32 l:tg 1
g5 33 �g3 lZlxg3+?
A blunder according to Dreev who
claims a win with 33 . . . liJde3, with the
idea of 34 ... .ixe4 35 fxe4 'iWb7!
34 lZlxg3 l:th8 35 lZlf5 exf5 36
�xd5+ �g6 37 �xb7 'ii'x b7 38
'iVd6+ �h5 39 l:taf1 g4 40 'ii'e 6 l:tf8
41 l:txg4 �g5 42 l:tg3 1 -0
Game 76
L. B . Hansen-lII escas
Moscow Olympiad 1924 @
1 d4 d5 2 c4 c6 3 lZlf3 lZlf6 4 lZlc3
e6 5 e3 lZlbd7 6 .id3 dxc4 7 .ixc4
b5 8 .id3 a6 9 e4 c5 1 0 d5 'it'c7 1 1
0-0 �b7 1 2 dxe6 fxe6 1 3 �c2 �e7
1 4 lZlg5 'iVc6 1 5 'iVf3 ! ? h6!? 1 6 'iVh3
hxg5 1 7 'iVxh8+ �f7 1 8 'it'h3 g4 1 9
"h4 lZle5 20 l:td 1 ! ?
20 . . . �d6 21 �e3? !
Illescas recommends instead 2 1 �f4
liJg6 22 �xd6 liJxh4 23 �xc6 .txc6,
giving back the material for a slight
endgame edge. This seems a reasonable
enough strategy, although in Jelen­
Pavasovic, Vienna 1996, the game
ended in a draw after 24 �e 1 �d8 25
 M e r a n Va ria tio n : Mo v e O r d e rs a n d Side lin e s
.igS ttJg6 26 eS b4 27 ttJe4 �xe4 28
lixe4 :dS 29 exf6 l:txgS 30 fxg7 ttJeS
31 i.b3 wf6 32 �f1 �xg7 33 f4 gxf3
34 gxf3.
21 . . . �c7! 22 l:[ac 1 l:[h8 ! !
Aah!!
23 'iix h8 ttJg6!
Suddenly the queen is trapped and
White is in trouble!
24 'ii'd 8 �xd8 25 l:[xd8 b4 26 ttJe2
liJxe4 27 l:tcd 1 ttJe5 28 l:[b8 ttJf6 29
liJf4 g5 30 l:[xb7+ 'ikxb7 31 ttJd3
liJxd3 32 �xd3 ttJe4 33 l:tc 1 a5 34
g3 'i'd5 35 �xe4 �xe4 36 l:txc5
'i'b1 + 37 �g2 a4 38 l:txg5 'ike4+ 39
'Otg 1 e5 40 l:th5 'ii' b 1 + 41 �g2 �xb2
42 l:[h7+ <;t>e6 43 �c5 �c2 44 l:th6+
<j;f7 45 �d6 �c6+ 0-1
Game ll
Gelfand-Dreev
Biel 1995
1 d4 d5 2 c4 c6 3 ttJc3 ttJf6 4 ttJf3
e6 5 e3 ttJbd7 6 �d3 dxc4 7 �xc4
b5 8 �d3 �b7 9 0-0 a6 1 0 e4 c5 1 1
d5 'i'c7 1 2 �c2 �e7 1 3 dxe6 fxe6
1 4 ttJg5 'ii' c 6 1 5 f4
With this move, White avoids the
time-consuming win of the exchange
and takes the opportunity instead to
support the e4-eS central push.
1 5 . . . h6
lS ... 0-0!? is interesting, as 16 eS fails
to 16 . . . 'i1Vxg2+ mate!
1 6 ttJf3 0-0-0 1 7 'ii'e 2 l:[hf8
Dreev suggests 17 . . . wb8 here.
17 . . . b4 18 eS bxc3 19 exf6 �xf6 20
bxc3 ttJb6 21 �d2 ttJdS 22 'ii'e l is then
slightly better for White according to
Gelfand.
1 8 e5 ttJd5 1 9 ttJxd5 "Vixd5 20 a4 b4
21 �e3?!
2 1 l:td I ! 'Wc6 22 as! gS 23 fxgS hxgS
24 �xgS �xf3 (24 . . . �xgS 25 ttJxgS
ttJxeS 26 �e4) 25 �xe7 11g8 26 �d6
'iWc7 27 �e4 ttJxeS 28 l:[xe6 is given by
Gelfand as clearly better for White.
21 . . . ttJb6 22 ttJd2 'i'c6 23 ttJb3?
ttJc4 Y2 - Y2
A strange draw offer as Black has a
wonderful position!
In the games so far in this section
White has tried to coax his opponent
Black back to the main line by offer­
ing him the chance to play . . . cS-c4. We
shall now consider the hard-line ap­
proach where White seeks to make
Black pay for omitting . . . cS-c4.
101
Th e S e m i - Sla v
Game 78
Kozul-Beliavsky
Slovenia 1995
1 d4 liJf6 2 c4 c6 3 liJc3 d5 4 e3 e6
5 liJf3 liJbd7 6 �d3 dxc4 7 �xc4 b5
8 1i.d3 a6 9 e4 c5 1 0 d5 �c7 1 1 0-0
iLb7 1 2 b3
I have changed the move order of
this game slightly (the actual sequence
was 1 1 b3 .tb7 12 0-0) to get a conven­
ient diagram! This move has two main
points: it ensures that a future . . . c5-c4
by Black will split his queenside
pawns after b3xc4 . . . b5xc4, and by
thus discouraging . . . c5-c4, it maintains
the white bishop on d3 and prevents
the active deployment of the black
king's bishop to d6.
However, by spending a tempo on
this consolidating move, White is
turning his back on the plan which
best suits his position: the central at­
tack via pressure against e6. Although
White intends instead to soften up the
black queenside with a2-a4, Black has
such an obvious and easy target in the
d5-pawn that I don't believe that
White can be successful.
1 02
1 2 . . . 1i.e7
Kozul-Lalic, Croatia 1995, saw the
risky 12 . . . c4 13 bxc4 bxc4 14 dxe6 fxe6
15 .tc2 i.b4, which seemed okay for
Black after 16 i.d2 (16 ctJa4!?)
16 . . . i.d6 17 ctJd4 ctJc5 18 f4 e5 19 ctJf5
0-0 20 'i'e2 ctJd3 2 1 i.xd3 cxd3 22
'i'xd3 .tb4.
1 3 �g5
The more testing 13 a4!? was played
in Krasenkov-Timman, European
Team Championship, Pula 1997. Af­
ter 13 . . . exd5 (13 . . . c4!? 14 bxc4 bxc4 15
.tc2 0-0 16 ctJd4 lUe8 17 dxe6 fxe6 18
ctJxe6 'i'e5! was unfathomable in Kra­
senkov-Se.Ivanov, Augustow 1996) 14
ctJxd5 ctJxd5 15 exd5 0-0 16 axb5 axb5
17 l::tx a8 i.xa8 18 �xb5 ctJf6 19 .tb2
ctJxd5 20 l::te 1 White was a little better.
1 3 . . . 0-0 1 4 1:1c1 1:1ad8! 1 5 �xb5! ?
A very clever spot as 15 . . . axb5 16
ctJxb5 'i'b8 17 d6! nets a pawn
(Beliavsky) . However, by eating up
White's centre, Black finds his way to
achieve even a slight edge.
1 5 . . . exd5 1 6 iLd3 dxe4 1 7 liJxe4
�xe4 1 8 �xe4 liJe5 1 9 �e2 liJxe4
20 iLxe7 liJxf3+ 2 1 gxf3 �xe7 22
�xe4 �g5+ 23 �g4 �d2 24 1:1xc5
�xa2 25 �a4 �e2 26 �e4 �d2 27
1:1 a 1 1:1fe8 28 1:1e5 1:1xe5 29 �xe5
�d3 30 �e4 �b5 3 1 �a4 �g5+ 32
�g4 �f6 33 1:1c 1 h5 34 �xh5 g6 35
�g4 1:1d4 36 �c8+ �h7 37 �h3+
�g7 38 1:1c8 �g5+ 39 �g3 �f5 40
1:1e8 1:1f4 41 �g2 g5 42 1:1e3 �g6 43
1:1e8 % - %
Black avoids o r delays . . . i. b 7
Question 6: Why delay . . . ..tb7? I
thought the whole point was to get
the bishop active on the long diagonal!
 M e r a n Va ria tio n : M o v e O r d e rs a n d Side lin e s
Answer: Black i s trying t o b e a
clever here: by switching the order in
which he develops his pieces, Black
hopes to prevent White from launch­
in g an early attack on e6 or making
the e4-es thrust. By leaving the bishop
on cS retains the protection of e6.
Plans such as dsxe6 followed by lbd4
or lbgs, or e4-es to open the e-file will
be easier for Black to defend against,
with e6 defended.
Note that this idea can only be used
via the S . . . a6 move order. The main
line using the delayed . . . i.b7 is S . . . a6 9
e4 cs 10 ds c4 1 1 �c2 'iic 7 12 0-0 �cs.
The bishop is very active on cS:
Black is anticipating dsxe6 f7xe6,
opening the f-file, and hopes to link
up with the bishop's attack on the f2-
pawn by castling and playing . . . lbg4.
Question 7: How will Black deal
with the threat of e4-es?
Answer: After White has taken on
e6 Black will block e4-es by playing
. . . lbes. Since the bishop on cs pins the
pawn on f2 to the king on g l , White
will not be able to play a quick f2-f4 to
chase away a black piece from es. If
White does not take on e6, then Black
can blockade in the centre, as we shall
see in the Game SO.
Game 79
Lautier-Piket
Monaco (match) 1996
1 d4 d 5 2 e4 e6 3 lLlf3 lLlf6 4 lLle3
e6 5 e3 lLlbd7 6 �d3 dxe4 7 �xe4
b5 8 �d3 a6 9 e4 e5 1 0 d5 "Vie7 1 1
0-0
1 1 dxe6 fxe6 12 lbgs was well met
by 12 . . . lbes! in Nikolic-Bareev, Euro-
pean Club Cup, Lyon 1994.
1 1 . . . e4 1 2 .ie2 .ie5 1 3 dxe6
After 13 'i'e2 es Black is okay - see
the next game.
1 3 . . . fxe6
Now Black would be fine after 14
'ilie2 lbes, ready to meet 15 �f4 with
Is . . . lbxf3+ 16 'it'xf3 es and 15 lbxes
'i'xes 16 �h l with 16 . . . 0-0 17 f4 'i'hS.
So White needs to act quickly.
1 4 e5!
Question 8: What? You said e4-es
wasn't dangerous with the bishop on
cS!
Answer: In fact e4-es is dangerous
here, but not for the usual reasons!
Since e6 is defended, White cannot
attempt anything on the e-file. How­
ever, since Black has not placed his
bishop on the as-h 1 diagonal, White
has many tactical ideas to gain a tempo
on the unprotected rook on as, and
this, coupled with White's lead in de­
velopment, is enough to give him a
powerful initiative. To tell you a little
secret, the key idea in this line is mine,
all mine! I had known about it for two
years, but never got a chance to play
it. While seconding Joel Lautier for
this match, I showed him this idea and
1 03
Th e S e m i - Sla v
10 and behold, he got to use it! 14 eS
had been played before, but it had
been dismissed as leading to equality.
However, I found something new that
gives Black a few headaches.
1 4 . . . liJxe5 1 5 �f4! �d6
1S . . .ttJxf3+ 16 1iVxf3 1iVb7 17 ttJe4! is
very unpleasant for Black.
1 6 �xe5 .ltxe5 1 7 liJxe5 'i'xe5 1 8
l:le 1 'i'e5
18 . . . 1iVc7 19 a4 b4 (19 . . . 0-0 20 axbS is
also good for White) 20 ttJdS! ttJxdS 2 1
1iVxdS �a7 2 2 1iVhS+! 1iVf7 2 3 1iVcS! is
very unpleasant for Black.
1 9 liJe4 liJxe4 20 �xe4!
Gaining a tempo on the rook on a8 !
20 . . . l:la7 2 1 b4! 'i'g5
2 1 . . .1iVxb4 22 1iVhS+ gives White ex­
cellent play for the two pawns.
22 f4! !
This was my discovery (hence the
' ! !') . 22 . . . 1iVxf4 is met by 23 1iVd4!
(threatening both 24 1iVxa7 and 24
.i.c6+) 23 . . . 1iVc7 24 11£1 !, preventing
Black from castling kingside. After
24 .. J�f8 2S �xf8+ �xf8 26 ..\ixh7, the
threat of �f1 + is horrific for Black.
Since the f4-pawn is immune, Black
cannot keep his queen on the fifth
rank to stop the disruptive 1iVhS+.
1 04
22 . . . 'i'h4!? 23 g3 'i'd8 24 'i'h5+ l:lf7
25 l:lad 1 'i'b6+ 26 �f 1 h6
Horrible, but Black will never get
castled otherwise.
27 �g6 0-0 28 �xf7+ l:lxf7
White has an overwhelming game.
However, the game was eventually
drawn after many adventures! If I had
known then what I knew at the end of
this match, I would not have been
surprised at this outcome, as the
match finished level at 4-4 after eight
hard-fought draws!
29 'i'e5 'i'xe5 30 bxe5 l:le7 3 1 l:le5
b4 32 l:ld6 a5 33 e6 a4 34 '\t>e 1 b3
35 axb3 exb3 36 �d2 '\t>f7 37 l:lb5
'\t>e7 38 l:ld4 l:lxe6 39 l:lxa4 l:le2+ 40
'\t>e3 l:lxh2 41 l:la7+ '\t>f6 42 l:lxb3
l:le2 43 '\t>d3 l:le6 44 l:lba3 l:ld6+ 45
�e3 .ltd7 46 l:l3a5 l:td 1 47 l:le 7 l:ld6
48 l:lb7 l:ld 1 49 l:laa7 �e7 50 l:la8
l:ld5 51 l:lg8 �f7 52 l:lh8 �e7 53
l:lb4 l:la5 54 l:ld4 �e6 55 l:le4 l:la3+
56 '\t>d4 �d7 57 l:le3 l:la4+ 58 '\t>e3
l:la5 59 l:ld3 e5 60 l:lb8 exf4+ 6 1
gxf4 l:la7 62 l:l d 5 �e6 6 3 l:ldb5 '\t>f6
64 l:l8b6 g6 65 l:le5 l:le7 66 '\t>d4
�f7 67 l:ld6 '\t>f6 68 l:la5 l:ld7 69
l:lxd7 .ltxd7 70 '\t>e4 h5 71 l:la6+
�e6 72 l:lb6 h4 73 l:la6 h3 74 '\t>f3
'\t>f7 75 '\t>g3 �f5 76 l:la8 '\t>f6 77
l:lh8 '\t>g7 78 l:lh4 �e8 79 '\t>f3 .ltf5
80 �e3 �e8 81 '\t>e4 �f5+ 82 '\t>e5
�e8 � - Y2
Black plays . e6-e5
. .
Game 80
Kharitonov-Ivanchuk
USSR 1 988
1 d4 d 5 2 liJf3 liJf6 3 e4 e6 4 liJe3
 M e r a n Va ria tio n : M o v e O r d e rs a n d Side lin e s
c6 5 e 3 ttJbd7 6 �d3 dxc4 7 �xc4
b5 8 �d3 a6 9 e4 c5 1 0 d5 c4 1 1
.te2 �c7 1 2 0-0 �c5 1 3 �e2 e5!
The ideal time for this move.
Black's bishop is well placed on c8,
covering the exposed f5-square and
White's queen is somewhat in the way
on e2, blocking the idea of liJe2-g3 .
14 ttJh4? !
A mistake according to Ivanchuk.
14 liJdl was better, preparing b2-b3.
14 . . . 0-0 1 5 'it>h 1 �d4! 1 6 ttJf5?!
ttJc5 1 7 �f3 �xf5 18 �xf5 �c8! 1 9
'i'f3 'it'g4 20 �xg4 ttJxg4 2 1 ttJd 1 f5!
Ivanchuk already claims a winning
advantage here!
22 exf5 ttJf6 23 ttJe3 l:tad8 24 a4
ttJxd5 25 axb5 axb5 26 l:ta5 ttJc7 27
�d2 l:ta8 28 g4 e4 29 'it>g2 ttJd3 30
l:tb 1 �xe3 3 1 fxe3 1:txa5 32 �xa5
ttJd5 33 �d2 l:ta8 34 <t>g3 b4 35
�xd3 cxd3 36 l:tc1 l:ta2 37 l:tc5
lixb2 38 l:txd5 l:txd2 39 'it>f4 l:tc2 40
�e5 b3 4 1 l:td8+ 'it>f7 0-1
Game 81
Podgaets-Muhametov
Moscow 1995
1 d4 d5 2 c4 c6 3 ttJc3 ttJf6 4 ttJf3
e6 5 e3 ttJbd7 6 �d3 dxc4 7 �xc4
b5 8 �d3 �b7 9 0-0 a6 1 0 e4 c5 1 1
d5 c4 1 2 �c2 e5
By releasing the central tension here
instead of playing the main line
12 . . . 'i'c7, Black accepts that his light­
squared bishop will not find activity
on the a8-h 1 diagonal. He reasons that
the piece activity that d5xe6 concedes
to him weighs less than the benefits it
provides White: the target of the e6-
pawn and the potential e4-e5 break.
By closing the centre with . . . e6-e5,
Black removes the point of tension
that was the basis of White's attacking
ambitions, forcing White to search
elsewhere for activity.
1 3 ttJe2!
The structure is similar to that of
the Chigorin Ruy Lopez (minus
White's c3-pawn and Black's d6-pawn)
and there are positional motifs com­
mon to both: the transfer of the
queen's knight to f5 and the under­
mining of the black queenside with
b2-b3 and even a2-a4. The text pre­
pares b2-b3 without allowing . . . b5-b4
and . . .c4-c3 with tempo, while taking
the knight towards the f5-square that
was exposed when Black went . . . e6-e5.
1 05
Th e S e m i - Sla v
1 3 . . . iLc5 1 4 b3! cxb3 1 5 axb3
In removing the c4-pawn, White
has gained some manoeuvring space
for his pieces on the queenside and
freed d3 for a white knight which is a
stepping stone towards two black
queenside weaknesses: c5 and c6 (via
b4, supported by a rook on c 1) . To see
the value of the c5-square, imagine
White's pawn on b4 and a white
knight on c5: from c5, the white
knight attacks a6 and the bishop on
b7, while allowing White to build up
his major pieces behind it on the c-file.
1 5 . . . 0-0 1 6 ltJg3 g6
Weakening the kingside, but the
threat of tbf5 is rather annoying for
Black. Perhaps 16 . . . tbe8 would have
been better, preparing to meet 17 tbf5
with 17 . . . tbd6.
1 7 iLh6 l:te8 1 8 h3 l:tc8 1 9 �d2 iLf8
20 iLxf8 ! l:txf8 21 l:tfc 1 !
2 1 . . .'�e7?
A bad mistake according to Pod­
gaets who recommends instead
2 1 . . .tbe8 22 ..td1 ! (to activate the
bishop via g4!) 22 . . Jixc1 23 �xc1
tbd6, though he still considers this to
be clearly better for White. This is a
typical example of what can happen in
1 06
this opening. White's pawn wedge on
e4 and d5 gives him a clear space ad­
vantage and reduces Black's minor
pieces to passivity - the bishop on b7
just defends the a6-pawn and prevents
White from occupying the c6 outpost
with a major piece; the knight on d7 is
tied to the e5-pawn; and the knight on
d6, while comfortably placed, has lim­
ited scope for manoeuvre. White's c­
file control also discourages Black's
queen from abandoning d8 due to the
possibility of lic7. Finally, since the
dark-squared bishops have been ex­
changed, Black's dark squares are vul­
nerable to attack or infiltration by the
white queen: White can challenge the
blockade on d6 via 'iVb4 or instigate
kingside threats with 'iVh6.
Clearly, the only piece that can
challenge the dark-square incursions of
the white queen is Black's own queen.
Now we begin to see the method that
White will adopt to develop an initia­
tive. By creating threats on the dark
squares with his queen, White will
draw the black queen from its protec­
tion of c7, allowing �c7. His play is
based on a combination of queen,
rook and space advantage; he will
normally seek the advantage in the
middlegame. Consequently, Black
must aim to exchange White's active
major pieces and head for the end­
game. However, it would be wrong to
assume that all endings are unfavour­
able for White. For example, after an
exchange of rooks, if White can fol­
low b3-b4 (preventing . . . a6-a5-a4) with
a transfer of a knight to c5, he will
enjoy a certain initiative. Thus White
can consider Podgaets's idea of 24 h4.
 M e ra n Va ria tio n : M o v e O rd e rs a n d Side lin e s
A preventive move such as 24 . . . h5
doubles the strength of 25 �h6, while
after 24 . . . �e7 25 h5 l::rc S 26 l:txcS+,
White can seek to implement the
above ending with the annoying h5
thrust as a bonus. It thus seems that
Podgaets's assessment is justified.
22 d6! �d8 23 b4!
Freeing b3 for the light-squared
bishop from where it attacks the sensi­
tive f7-pawn. 23 . . . 'ii'b 6 24 i.b3! lIcdS
24 . . . l:txc1 25 l:txc 1 lbxe4 26 lbxe4
.1xe4 27 lbg5 .taS 2S l:tc7 gives White
excellent compensation according to
Podgaets.
23 .. :i6'b6 24 .Ji.b3 l:tcd8 25 l:te 1 !
Calmly reinforcing the e4-pawn and
regrouping his rooks to active squares.
25 . . . lbe8 26 l:tad 1 lbdf6 27 lbxe5
lbxd6 28 'it'f4 �g7 29 l:td3 lbc4 30
.ixc4 bxc4 31 lbf5+!
31 . . . �h8 32 �h6 lbh5 33 l:td7 ! �f6
34 lbd6 �g8 35 lbxb 7 l:tde8 36 lbg4
"c6 37 lbc5 a5 38 e5 axb4 39 lbe4
l:te6 40 lbg5 1 -0
Essentially, as we have seen, the
plan of blocking the centre with . . . e6-
e5 is most successful when Black has
not yet committed his light-squared
bishop to b7, since then it guards f5
and can also emerge with effect either
to g4 or, after a black queenside pawn
advance with . . . a6-a5 and . . . b5-b4, to
a6. The most effective version of the
. . . e6-e5, for Black can be achieved via
S . . . a6 9 e4 c5 10 d5 c4 1 1 .tc2 e5.
Question 9: Can't White just avoid
. . . e6-e5 by taking on e6 as soon as
Black plays . . . c5-c4?
Answer: Aha! Read on!
Black takes the bishop on d 3
Game 82
Cu . Hansen-Chernin
Taastrup 1992
1 c4 c6 2 lbc3 d5 3 d4 lbf6 4 e3 e6
5 lbf3 lbbd7 6 .Ji.d3 dxc4 7 .Ji.xc4 b5
8 .Ji.d3 a6 9 e4 c5 1 0 d5 c4 1 1 dxe6
cxd3
A key resource if Black was hoping
to meet 1 1 �c2 with the blockading
1 1 . . .e5 rather than the main line
1 1 . . .i.b7 12 0-0 'ilVc7.
1 2 exd7+ �xd7 1 3 0-0 .Ji.b7
This standard position can also arise
from S . . . �b7 9 0-0 a6 10 e4 c5 1 1 d5
c4 12 dxe6 cxd3 13 exd7+ 'i'xd7. Kar­
pov considers that Black can only
107
Th e S e m i - Sla v
think of making a draw in this line.
However, in practice Black has tended
to hold his own comfortably. Gener­
ally he will generally sacrifice his
pawn on d3, but in contrast to normal
lines, he will have the two bishops and
a sound pawn structure.
1 4 .l:!.e 1
The drawback to this line is that if
White wants, he can force a draw with
14 lbe5 'iYd4 15 lbf3 �d7 16 lbe5.
1 4 . . . i.e 7 1 5 e5 tbd5 1 S tbe4 0-0 1 7
'iVxd3 'iVg4 1 8 tbg3
I S lbfg5!? is seen in the next game.
1 8 . . . f5! ?
This sharp move i s an idea o f the
famous Russian coach Mark Dvoret­
sky.
1 9 i.d2
19 exf6 i.xf6 20 h3 lbb4! gives
Black good counterplay (Chernin) .
1 9 . . . .l:!.ad8 20 �b3 �h8 2 1 h3 �gS
22 .l:!.ac 1 ! f4 23 tbe4 tbe3 24 i.xe3
i.d5!
24 . . . i.xe4 25 .i.c5 �d3 26 �xd3
.i.xd3 27 .i.xe7 would have provided
White with excellent play for the sac­
rificed queen according to Chernin.
25 i.c5! YZ - Y2
This amazing queen sacrifice (25
�c3 .i.xe4 is very good for Black) is
White's only way to play. Peter Wells
suggests that after Chernin's recom­
mendation of 25 . . . i.xb3 26 i.xe7 or
26 axb3!? .i.xc5 27 lbxc5, intending e5-
e6, White may even be slightly better.
1 d4 d5 2 c4 cS 3 tbc3 tbfS 4 tbf3
1 08
eS 5 e3 tbbd7 S i.d3 dxc4 7 SLxc4
b5 8 i.d3 i.b7 9 0-0 as 1 0 e4 c5 1 1
d5 c4 1 2 dxeS cxd3 1 3 exd7+ 'iVxd7
1 4 .l:!.e1 i.e7 1 5 e5 tbd5 1 S tbe4 0-0
1 7 'iVxd3 �g4 1 8 tbfg 5 ! ?
A n interesting Karpov idea, threat­
ening lbf6+!
1 8 . . . .l:!.fd8! 1 9 h3 'iVh5 20 tbg3
Piket also suggests 20 e6!? or 20
i.d2.
20 . . . �gS 21 'iVxgS hxgS 22 a3 .l:!.ac8
23 tbf3 b4? !
Perhaps a mistake. Piket suggests
23 .. .'it>fs or 23 . . . lbb6. White does de­
velop a certain amount of pressure in
the game.
24 Jt.g5 fS 25 exfS gxfS 2S Jt.d2
�f7 27 axb4 i.xb4 28 .l:!.a4 i.xd2 29
tbxd2 g5 30 tbc4 �gS 31 .l:!.eS �f7
32 .l:!.e 1 �gS 33 b3 tbf4 34 tbe3 .l:!.c5
35 .l:!.b4 .l:!.b5 3S .l:!.xb5 axb5 37 .l:!.c 1
.l:!.d3 38 .l:!.c7 i.xg2 39 tbxg2 tbxh3+
40 �f1 .l:!.xb3 41 tbe4 .l:!.b 1 + 42 �e2
.l:!.b2+ 43 �e 1 f5 44 .l:!.cS+ Yz - Yz
It seems therefore that White may
be able to keep a nagging plus in these
lines, although White players hoping
for a direct attack may be disap­
pointed by their endgame nature!
 M e ra n Va ria tio n : M o v e O r d e rs a n d Side lin e s

Summary
Dreev's queenside castling plans (Games 75 and 77) are particularly worthy of
attention, while 8 . . . a6 9 e4 c5 10 d5 c4 1 1 .i.c2 e5!? may also be worth a try if
Black wishes to establish a blockade in the centre.

1 d4 d5 2 c4 c6 3 lLlf3 lLlf6 4 lLlc3 e6 5 e3 lLlbd7 6 �d3 dxc4 7 �xc4 b5 8

�d3

8 .i.b7
. . .

8 ... a6 9 e4 c5 10 d5 (D)
10 . . . 'i'c7 1 1 0-0 c4 12 .i.c2 .i.c5
13 dxe6 - Game 79; 13 'ii'e2 - Game 80
10 . . . c4 1 1 dxe6 cxd3 12 exd7+ 'i'xd7 13 0-0 .i.b7 - Games 82 and 83
(by transposition)
9 0-0 a6 1 0 e4 c5 1 1 d5 'iic 7
1 1 . . .c4 (D)
12 .tc2 e5 - Game 81
12 dxe6 cxd3 13 exd7+ 'ifxd7 14 �e 1 .i.e7 15 e5 ttJd5
16 ttJe4 0-0 17 'i'xd3 'i'g4
1 8 ttJg3 - Game 82; 18 ttJfg5 - Game 83
1 2 dxe6
12 .i.c2 .i.e7 13 dxe6 fxe6 14 ttJg5 'i'c6 (D)
15 'ii f3 h6 16 'i'h3 hxg5 17 'i'xh8+ 'itt f7 1 8 'i'h3 g4 19 'i'h4 ttJe5
20 f4 - Game 75; 20 .l:td1 - Game 76
15 f4 - Game 77
12 b3 - Game 78
1 2 fxe6 - Game 74
. . .

1 0 d5 1 1 . . . c4 1 4 . . 'ii c 6
.

1 09
 , CHAPTER NINE

1 d4 d5 2 c4 c6 3 lLlf3 lLlf6 4 lLlc3 White. By a process of elimination we

e6 5 e3 lLlbd7 6 �d3 dxc4 7 i.xc4 therefore arrive at 9 . . . b4, which per­
b5 8 .ltb3 �b 7 mits . . . c6-c5 by freeing the c-pawn
Until recently, both of the lines ex­ from the job of defending the b-pawn.
amined in this chapter, 9 e4 and 9 a3, Since 9 . . . b4 attacks the knight on c3
might have been considered the main White has no time to use his central
lines of the Meran, but in the last cou­ trumps immediately. Another good
ple of years, they have both dropped side-effect is that White's knight is
out of fashion as White players have forced offside to the a4-square, weak­
concentrated on 9 0-0. ening his support of e4.
The first part of this chapter 1 1 e5 lLld5
(Games 84-90) is concerned with 9 e4
(Games 9 1-95 focus on 9 a3) so let us
start with an overview of this move.
9 e4
is a logical move: it achieves White's
basic aim of freeing his dark-squared
bishop by opening the c 1-h6 diagonal.
However, there is a drawback:
9 b4! 1 0 lLla4 c5
. . .

After 9 e4, Black had to act to chal­

lenge White's centre straightaway.
Black wants to play . . . c6-c5, but obvi­
ously the b-pawn would hang if this
were played immediately and 9 . . . a6 is
obvious but too slow: 10 e5! liJd5 1 1
liJxd5 cxd5 1 2 0-0 is very pleasant for
Question 1: What is White aiming
for here?
A nswer: White's strength is his cen­
tral pawn structure and in particular

1 10
 M e r a n Va ria tio n w i t h 8
his pawn on eS, which attacks Black's
weakened central dark squares and
provides an outpost for a white knight
on d6. This pawn also gives White
attacking chances on the kingside, as it
takes away the defensive square f6
from the black knights: thus g4 and hS
are free for the white queen, while h7
also lacks its usual protection. If White
plays on the kingside with lbgS and
'iWhs (attacking h7 and f7) he may be
able to cause Black grave danger.
However, White can only divert his
pieces to the wing if his centre is abso­
lutely secure. Black must therefore
keep the pressure on White's central
pawns. Thus his cS-pawn is very im­
portant: Black can play . . . cSxd4, de­
stroying the base of White's pawn
centre and preventing lbgS due to the
loose pawn on eS.
Question 2: What are Black's aims?
A nswer: Black's wants to develop
without allowing White to whip up
kings ide play. If he is successful in this,
then White's centre will switch from
being an attacking weapon to an easy
target for Black's pieces. An interest­
ing feature of this line is the position
of the pawn on a7.
Question 3: Well, what about it?
A nswer: Black's queenside structure
with the b-pawn on b4 is common in
queen's pawn openings, but usually it
arises after Black has played . . . a7-a6
and . . . b7-bS and White undermined
the queenside with a2-a4. Black never
normally plays . . . bS-b4 on his own
initiative without provocation by a2-
a4. Here, Black does have the com­
pensation of having forced the white
knight offside to a4, but White has
. . . i. b 7 : Wh i t e A l t e rn a t i v e s
kept his pawn on a2 and the black a­
pawn has not moved.
Question 4: Wow! Big deal!
A nswer: This may seem insignifi­
cant, but it plays a major part in every
single variation. White gains the pos­
sibility of a2-a3 to attack Black's
queenside, while the absence of a
pawn on a6 gives White .tbs to attack
the knight on d7 or embarrass the
black king on e8.
Now it's time to get down to spe­
cific moves. We shall first of all con­
sider the main line, 12 0-0 (Games 84-
88) before moving on to 12 lbxcs
(Game 89) and 12 dxcS (Game 90) .
Game 84
Yakovich-Giorgadze
Yerevan Open 1996
1 d4 dS 2 c4 c6 3 lLlc3 lLlf6 4 e3 e6
S lLlf3 lLlbd7 6 .id3 dxc4 7 .ixc4 bS
8 .id3 .ib7 9 e4 b4 1 0 lLla4 cS 1 1
eS lLldS 1 2 0-0 a6?
We examine this inferior line to
show the power of White's play if left
unchecked. The correct 12 . . . cxd4 is
considered in Games 8S-88.
13 lLlgS ! cxd4 14 lLlxe6! fxe6 1 S
�hS+ <;; e 7 1 6 .igS+ lLlSf6 1 7 l:tfe 1 !
�e8 1 8 exf6+ gxf6 1 9 :txe6+! !
<;;x e6 20 �g4+!
20 lite 1 would have been met by
20 . . . lbeS. The text forces the king to
move farther afield.
20 . . . �d6
20 . . .�f7 21 �c4+ Wg7 22 �d2+
'i'g6 23 'i'xd7 wins, as Yakovich
shows.
21 'it'xd4+ �c7 22 l:tc1 + ..t>b8 23
.if4+ lLleS 24 lLlb6!
111
Th e S e m i- Sla v
The key move in the attack, threat­
ening lbd7+ winning the black queen.
24 . . .1:;[a7 25 ttJd7+ 'ita8 26 ttJb6+
'itb8 27 ttJd7+ 'ita8 28 .1i.e3 1 -0
After a teasing little repetition, the
finish. Black resigned as 28 . . . lbc6 29
l:txc6 'ii'x e3 30 lbb6+ Wb8 3 1 'ii'd 8+
leads to mate. A gem of a game!
Game 85
'"., Piket-Kraf!lnik
, :,:/'Amst'erdam 1 993
1 d4 d5 2 c4 c6 3 ttJc3 ttJf6 4 ttJf3
e6 5 e3 ttJbd7 6 �d3 dxc4 7 .1i.xc4
b5 8 .1i.d3 �b7 9 e4 b4 1 0 ttJa4 c5
1 1 e5 ttJd5 1 2 0-0 cxd4 1 3 l::t e 1
For 13 lbxd4!? see Games 8 7 and 88.
1 3 . . . g6! !
After . . . cSxd4, White gained the ex­
tra possibility of lbxd4, allowing the
white queen to come to g4, attacking
g7 and e6, or hS, attacking f7. 13 . . . g6
sets up the ideal defensive structure
against a queen on g4 (the pawn on g6
and bishop on g7, defending the dark
squares), while also taking the hS­
square from White's queen. Further­
more, it also shields h7 from the
bishop on d3 along the b1-h7 diagonal,
1 12
which takes the sting out of lbgs.
1 4 �g5
14 .td2!? is seen in the next game.
1 4 . . :ii'a5 1 5 ttJxd4 a6 ! !
Despite their active pOSItiOnS,
White's pieces are not coordinating to
create one big threat, but rather a se­
ries of 'mini-threats' . For example,
lS . . . .tg7 would have been awkwardly
met by 16 .tbS! The calm lS . . . a6 pre­
vents .tbS and asks White to find an­
other idea.
1 6 l::t c 1 �g7 1 7 ttJc6 �xc6 1 8 l::t x c6
0-0 1 9 .1i.c4!?
White attacks the knight on ds now
that it has lost the support of the
bishop on b7. Now 19 . . . lbxeS loses
the queen to 20 l:tcS! (that knight on
a4 comes in useful at last!) and
19 . . . .txeS 20 lbcS! lbxcs 2 1 l:txes lbd7
22 l:te 1 l:tfc8 23 .txdS 'i'xdS 24 'ii'xds
exdS 2S l:td6 lbcs 26 ':'xdS lbe6 27
.tf6 was pleasant for White in Piket­
M.Gurevich, Belgium 1993. However,
Black has another resource.
1 9 . . . h6! 20 �xd5!
The sharpest attempt. 20 ii.h4
lbSb6 is about equal according to
Piket, while 20 .td2 lbxeS! 21 licS
'ii'xcs 22 lbxcs lbxc4 is a good version
 M e ra n Va ria tio n w i t h 8
of the queen sacrifice, which Piket
assesses as slightly better for Black.
20 .. :�xd5 21 'ii' x d5 exd5 22 .if6
22 . . . lLlxe5! !
Quite amazing! After 23 ':xe5,
Black stresses the weakness of White's
back rank with 23 .. J�acS!! 24 i.xg7
':xc6 25 i.xh6! (stopping . . J�c1+)
25 . . . g5! 26 i.xg5 f6! 27 llxd5 fxg5 2S
l1xg5+ '\th7 29 f3 �dS, when despite
White's temporary material edge,
Black has the better prospects due to
White's rather sad knight on a4. All
this analysis is by Piket.
23 i.xe5 l:tae8! 24 f4 f6 25 lLlb6
fxe5 26 lLlxd5 exf4 27 l:txe8 l:txe8
28 J:[xg6 �h7 29 l:txa6 l:te2 30 'it>f1
l:txb2 3 1 lLlxf4 Y2 - Y2
Game 86
Alterman-Dreev
Manila Olympiad 1992
1 d4 d 5 2 c4 c6 3 lLlf3 lLlf6 4 lLlc3
e6 5 e3 lLlbd7 6 .id3 dxc4 7 .ixc4
b5 8 .id3 i.b 7 9 e4 b4 1 0 lLla4 c5
1 1 e5 lLld5 1 2 0-0 cxd4 1 3 l:te 1 g6
14 i.d2! ?
Question 5: This looks rather
strange! What's the point?
. . . iL b 7 : Wh i t e A l t e rn a t i v e s
A nswer: Since Black is putting his
bishop on the hS-al diagonal, White
attacks the pawn whose protection the
bishop will abandon: the b4-pawn.
1 4 . . . i.g7 1 5 .ib5!
This move marks the start of a very
sharp tactical plan which pushes
Black's position to the limit. Again,
White uses the fact that Black's dark­
squared bishop is not covering the fs­
a3 diagonal any more: by pinning the
knight on d7 to the king on eS, White
threatens to dramatically activate his
knight on a4 with ltJc5!
1 5 . . . a6! 1 6 lLlc5! axb5 1 7 lLlxb7
-.b6 1 8 lLld6+ �e 7 !
The best choice, as I s . . . 'iit fs 19 ltJg5!
ltJxe5 20 l:txe5 i.xe5 (20 .. :i'xd6 2 1
'i'f3! i.f6 22 �xd5 exd5 2 3 i.b4! wins)
2 1 ltJdxf7 is rather grim for Black, as
Stohl points out.
Question 6: I'm sorry, but even after
IS . . . 'iite 7 this position looks losing!
A nswer: Stay calm! The knight on
d6 is a dangerous piece, but it is not
secure - Black is threatening to de­
stroy its support with . . . ltJxe5. White
also has few pieces in this attack - lit­
tle else apart from the knight and the
pawn on e5 - and not much time to
1 13
Th e S e m i - Sla v
bring up the reserves; he must attack
now or . . . lbxe5 will net another pawn.
Moreover, it is very difficult for White
to involve his major pieces as Black's
solid centre makes such a good barrier.
1 9 �g5+!
The most dangerous continuation,
forcing the black king back to the
back rank. 19 lbg5 was tried in an ear­
lier game between Stohl and Sakaev in
Dortmund 1992, but 19 . . . !lhfS ! 20
lbxh7 (20 lbxe6 fxe6 21 .tg5+ i.f6! 22
exf6+ \t>xd6 wins according to Stohl) ,
and now Stohl's 20 . . . lbxe5! 2 1 lbe4 (2 1
lbxfS 'i'xd6 22 lbh7 !lhS 23 lbg5 lbd3
wins) 2 1 . . .£5! 22 lbxfS (22 lbeg5 IlfdS
23 f4 lbe3 24 iLxb4+ \t>eS looks good
for Black) 22 . . . fxe4 23 lbxg6+ lbxg6 24
'i'g4 \t>f7 is slightly better for Black.
1 9 . . .'.t>fS 20 l:te 1 !
The most incisive continuation,
grabbing the open c-file and bringing
an extra unit into the attack. 20 lbxd4
lbxe5 2 1 lb5xb5 h6 22 iLh4 lbc4 23
lbxc4 1/2-1h occurred in Yakovich­
Novikov, Yerevan 1996.
20 . . . h6!
A vital strengthening move. The
immediate 20 . . . lbxe5 fails to 21 lbxe5
'i'xd6 22 !lc6 trapping the queen, as
22 . . . 'i'bS 23 lbd7+ forks king and
queen. This would not be a problem if
Black could give up his queen for good
material compensation with 22 . . . 'i'xe5
23 !lxe5 i.xe5. However, here 24
'i'e1 ! is nasty: 24 . . . iLg7 (24 . . . iLf6 25
iLxf6 lbxf6 26 'i'e5! \t>g7 27 g4 h6 2S
h4 wins, as does 24 .. .£6 25 !lxe6 \t>f7
26 llxe5 fxe5 27 'i'xe5) 25 'i'c1 ! ,
threatening !lcS+, when 2 5 . . J�eS fails
to 26 'i'c5+ \t>gS 27 IlcS! and 25 . . . jLf6
26 iLh6+ is also bad.
1 14
21 �h4
21 . . . 'it>gS?
Now was the time for 2 1 . . .lbxe5.
The difference is that after 22 lbxe5
'i'xd6 23 !lc6, the queen sacrifice
23 . . . 'i'xe5 24 ltxe5 iLxe5 was quite
promising for Black in Hjartarson­
Akopian, World Team Champion­
ship, Lucern 1993 . Without the threat
of iLh6, Black's king is perfectly safe.
Instead of 23 !lc6 Alterman suggests
23 i.g3 'i'e7 24 h4 to soften up the
black kingside, but after 24 . . . h5! 25
'i'f3 \t>gS Black has very good pros­
pects as 26 lbxg6 fxg6 27 'i'xd5 exd5
2S !lxe7 d3 29 �d7 iLxb2 30 !lc6 d2 is
by no means worse for Black.
The text is a bad mistake since it
gives White a crucial opportunity to
reinforce the pride of his position: the
knight on d6.
22 �g3 'it>h7 23 lbxf7 l:thfS 24 lbd6
lbf4 25 'iVd2? !
25 �e4! was cleaner and would have
given White a decisive advantage.
25 . . . lbh5 26 lbxd4 lbxg3 27 hxg3
�xe5 2S lbxe6 .lixd6 29 lbxfS+
l:txfS 30 l:te6 l:tf6 31 l:tee 1 'tWe5 32
l:t 1 e3 h5 33 l:te2 'it>g7 34 l:te 1 'it>f7
35 �e2 lbe5 36 l:txd6 l:txd6 37
 M e ra n Va ria tio n w i t h 8
'i'xe5 'it'xe5 38 l:txe5 lid2??
3S ... l:tdl+ 39 �h2 �d2 would have
made a draw according to Alterman.
39 b3! l:td 1 + 40 '1th2 l:td2 41 f3
l:txa2 42 l:txb5 l:ta 1 43 l:txb4 g5 44
J:tb8 1 -0
Game 87
I . Sokolov-Chernin
Wijk aan Zee 1991
1 d4 d 5 2 c4 c6 3 tLlc3 tLlf6 4 e3 e6
5 tLlf3 tLlbd7 6 �d3 dxc4 7 �xc4 b5
8 �d3 �b7 9 e4 b4 10 tLla4 c5 1 1
e5 tLld5 1 2 0-0 cxd4 1 3 tLlxd4!?
13 ... tLlxe5 14 �b5+ tLld7 1 5 11e 1
Question 7: What does White have
for the sacrificed pawn?
Answer: He seems to have a lot of
compensation: his rook is well-placed
on the half-open e-file; his bishop on
bS pins the knight unpleasantly; and
sacrifices such as lLlxe6 are in the air.
Black suffers from a considerable lag
in development and, to make matters
worse, has to deal with the threat of
lLlc6, attacking the black queen while
allowing �xdS!
And yet nothing seems to work for
White! The problem again is the
. . . 1J.. b 7 : Wh i t e A l t e rn a t i v e s
queen's knight which contributes
nothing to White's play in the centre
and on the queenside. This is an attack
where pieces will have to be sacrificed
to open up Black's solid structure;
there is an urgent need for reserves to
suppon the initial sacrifices and the
offside knight on a4 seems to ruin all
these attempts!
1 5 . . . l:tc8!
The first cool move, covering the
c6-square and thus preventing lLlc6.
1 6 �h5
The second big threat arises: by
pinning the pawn on f7 to the king on
eS along the hS-eS diagonal, White
instigates the threat of lLlxe6 or �xe6+!
The unsound but aesthetic 16
'ti'xdS?! gave Black an endgame plus in
Piket-M.Gurevich, Ostend 199 1 , after
16 . . . exdS 17 �e l+ iL.e7 IS lLlxe7 �dS !
19 .td2 'tWaS 20 iL.c6 l:!eS! 2 1 iL.xb7
l:txe7 22 �xe7 �xe7 23 .txaS ii'xa4.
1 6 . . . g6! 1 7 �e2
The aggressive 17 'ifeS is considered
in the next game.
1 7 . . . a6!
Amazingly enough, this is not the
only good defence in this position!
Dreev has played 17 . . . �e7!?, which
leads to an unclear ending after I S
.tgS 'ii'x gS 19 lLlxe6 'iVxg2+ 2 0 �xg2
lLlf4+ 2 1 �f1 lLlxe2 22 klad 1 fxe6 23
.ixd7+ ..t>f7 24 .ixcs .ixcs 25 �xe2
iL.a6+ 26 �f3 iL.b7+, as in Nadera­
Dreev, Manila Olympiad 1992.
However, the text gives Black a bet­
ter ending!
1 8 �xa6 �xa6 1 9 'it'xa6 �g7! 20
�g5 tLlc7 ! !
The position seems difficult, but
Chernin finds an amazing resource.
1 15
Th e S e m i- Sla v
2 1 �a5
21 'iVb7 l:tbS 22 �xdS l::tx b7 23
�xc7 .llxc7 24 ctJb5 .llc2 was very
pleasant for Black in Vyzmanavin­
Novikov, Moscow 1990, while the
violent 2 1 ctJxe6 fxe6 22 .llxe6+ wf71
23 .i.xdS ctJxa6 24 �e7+ wf6 25 .llxd7
�hxdS just wins for Black according
to Chernin.
2 1 . . J:ta8!
White cannot play 22 'ilxb4 as then
Black wins a piece with the tactic
22 . . . 'ilxg5.
22 �xd8 .l:txa5 23 �xc7 .l:txa4 24
.l:ted 1 liJf6 25 �e5 0-0 26 a3 liJd5
27 .i.xg7 'it'xg7 28 liJc2 bxa3 29
.l:txa3 .l:tc4 30 liJe3 liJxe3 3 1 .l:txe3
.l:tc2 32 b3 .l:td8 33 l:ta 1 .l:ta8 34 .l:tee1
l:txa 1 3 5 .l:txa 1 .l:tb2 36 h4 .l:txb3 37
g3 h6 38 .l:ta5 'it'f6 39 'it'g2 .l:td3 40
.l:ta7 l:td6 41 'it'f3 e5 42 .l:ta8 .l:td3+
43 'iPe4 .l:td4+ 44 �e3 �f5 45 .l:ta7
f6 46 .l:th7 h5 47 llg7 1:[g4 48 llh7
.l:tb4 49 'it'f3 e4+ 50 'it'e2 .l:tb2+ 5 1
'it'e3 llb3+ 5 2 'it'e2 l:tb2+ 5 3 'it'e3
.l:tb3+ 54 'it'e2 'it'g4 55 .l:th6 'it'h3 56
.l:txg6 .l:tf3 57 .l:th6 'iPg2 58 .l:txh5
.l:txf2+ 59 �e3 f5 60 .l:tg5 'it'f1 6 1 g4
.l:tf3+ 62 �d4 e3 63 .l:txf5 .l:txf5 64
<ot>xe3 llf8 0-1
1 16
Game 88
Wells-Kaidanov
Dublin 1991
1 liJf3 d5 2 d4 liJf6 3 c4 e6 4 liJc3
c6 5 e3 liJbd7 6 .lid3 dxc4 7 �xc4
b5 8 �d3 �b7 9 e4 b4 1 0 liJa4 c5
1 1 e5 liJd5 1 2 0-0 cxd4 1 3 liJxd4!?
liJxe5 14 �b5+ liJd7 1 5 .l:te 1 .l:tc8! 1 6
�h5 g6! 1 7 �e5
The most aggressive continuation,
keeping the pressure on e6 while at­
tacking the rook on hS.
1 7 .. :i'f6! 1 8 �e4? !
IS ctJf3 �g7 is the main line and
seems sufficient for Black: 19 .i.xd7+
Wxd7 20 'iVe2 only gave White a draw
in Akopian-Dokhoian USSR Cham­
pionship 199 1 , after 20 . . . 'ile7 2 1 �b5+
Wc7 22 a3 WbS 23 axb4 a6 24 'i'a5
'ilxb4 25 'ilxb4 ctJxb4 26 i.f4+ Wa7 27
.i.e3+ WbS 2S i.f4+.
1 8 . . . .l:tc7!
An excellent defensive move:
1S . . . �g7? was absolutely destroyed in
Mikhalchisin-Lanc, Trnava 19S5, by
19 ctJxe6 fxe6 20 .i.g51, when 20 .. :i!Vf7
(20 .. :i!Vxg5 2 1 'ili'xe6+ wfS 22 'i'xd7
ctJf4 23 .lle S+ wins) 2 1 'i'xe6+ 'ili'xe6 22
l1xe6+ Wf7 23 �xd7 �c7 24 .lld 61 gave
White a clear advantage.
1 9 'i'g4 �e7 20 �h6 a6!
see follo wing diagram
White's problem is that his oppo­
nent can easily chase back the white
pieces from their advanced positions .
21 .Jlxd7+ �xd7! 22 .l:tad 1 'it;>c8!
Putting the king to safety. White's
desperate sacrifice now flounders on
the weakness of his back rank.
 M e r a n Va ria tio n w i t h 8
23 lLlxe6 fxe6 24 l:txe6 -.fS 2S 'it'xfS
gxfS 26 l:txdS �xdS 27 lLlb6+ 'it>b7
28 lLlxd S l:td8 29 :LeS i.f8 30 i.gS
h6 31 h4 hxgS 32 lLlxc7 l:td 1 + 33
'1t>h2 i.d6! 0-1
Game 89
Bareev-Kramnik
Dortmund 1 995
1 d4 d S 2 c4 c6 3 lLlc3 lLlf6 4 e3 e6
S lLlf3 lLlbd7 6 �d3 dxc4 7 �xc4 bS
8 �d3 �b7 9 e4 b4 1 0 lLla4 cS 1 1
eS lLldS 1 2 lLlxcS
With this capture White rids his po­
sition of the offside knight in a4, while
clearing away some of the defences
around the black king. However,
these exchanges free Black's position.
1 2 . . . lLlxcs 1 3 dxcS �xcS 1 4 0-0
Question 8: Wait! Are you crazy?
White can play 14 .ibS+ here!!
A nswer: Aha, and now 14 .. .'�e7.
see follo wing diagram
Question 9: Yes, but I've forced the
black king to move! It's exposed, in
the centre, vulnerable to attack . . .
Answer: At first sight it feels bad to
move the king and lose the right to
. . . iL b 7 : Wh i t e A l t e rn a t i v e s
castle. However, the key thing about
Black's position is its solidity, espe­
cially in the central area: it will take
something special to break past the
knight on dS, supported by the bishop
on b7 and the pawn on e6. White also
has no pieces in the attack: there is the
bishop on bS, and . . . nothing else! In
fact the bishop is simply loose on bS,
and Black threatens .. :iVb6, hitting the
bishop and the f2-pawn. White has
only a temporary initiative; Black will
soon play . . . h7-h6, to stop anything
coming to gS, play his king's rook to
dS and then slip his king back to fs
and absolute safety. Since White has
no body of pieces to support his one
check, his temporary initiative IS
doomed to slip away without trace.
1 4 . . . h6!
Black must be very careful. As
Kholmov demonstrates, 14 . . . 0-0 is ask­
ing for disaster: 15 .ixh7+ �xh7 16
lbgs+ Wg6 (16 . . . �gS 17 'ii'h S l:.eS IS
'Wxf7+ �hS 19 'WhS+ �gS 20 '6'h7+
�fs 2 1 'iVhS+ �e7 22 'iVxg7+ is mate)
17 'Wc2+ fs 1 S exf6+ �xf6 19 'WxcS
!ics 20 'WxfS+!! 'iWxfs 2 1 lbh7+ neatly
finishes Black off. Consequently, with
his calm move, Black prevents .ixh7+
117
Th e S e m i - Sla v
tricks and prepares to castle.
1 5 tbd2 ! ? 0-0 1 6 tbe4 .td4!
The passive 16 . . . i.. e 7 would allow
White to attack the kingside with 17
'i'g4! By attacking the e-pawn that
White has left unprotected with his
knight manoeuvre, Black makes sure
that his opponent cannot simply
prosecute his ideas at his own pace.
1 7 tbd6 �c6 1 8 .lth7+ �xh7 1 9
'iVxd4 f6 ! 20 a3
20 i.. d2 fxeS 21 'i'xeS 'i'd7 was
played in Gagarin-Muhametov, Pots­
dam 1994, when Gagarin claims an
edge for White after 22 liJc4!? Jilfs 23
'i'g3 lIaf8 24 liJeS 'i'e8 25 lIfc1 i.bS
26 f3, but does not look terrifying.
20 . . . fxe5 21 'iVe4+ �g8 22 'iVxe5
�f6 � - Y2
Game 90,
YUSuRov-Kramnik
;H6rgenJ995
1 d4 d5 2 c4 c6 3 tbe3 tbf6 4 e3 e6
5 tbf3 tbbd7 6 �d3 dxc4 7 .ltxe4 b5
8 �d3 �b 7 9 e4 b4 1 0 tba4 c5 1 1
e5 tbd5 1 2 dxc5
This used to be considered a tricky
move order, but it seems now to make
1 18
no difference!
1 2 . . . tbxc5! 1 3 i.b5+
13 liJxcs i.. xcs transposes to Game
89 above.
1 3 . . . tbd7 1 4 i.g5 �a5!
14 . . . .te7 loses, as Wells points out,
to 15 liJcS! ..txgS 16 i.xd7+ \t>f8 17
i.xe6!
1 5 i.xd7+
15 liJd4 .i.a6! 16 i.. c6 lIc8 17 1Ic1
b3+! was horrific for White in C.Han­
sen-Shirov, Biel I992.
1 5 . . . �xd7
This position was originally assessed
as better for White, but as Kramnik
shows, Black can consolidate, as White
has no way to get at the black king.
1 6 0-0 .te7 1 7 b3 h6! 1 8 �xe7
�xe7 1 9 tbd2 tbf4 20 tbe4 'iVd5 2 1
�xd5 �xd5 2 2 tbe3 l:the8 2 3 l:tfe1
�e4 24 f3 �g6 25 tbc4 tbd3 26
l:ted 1 l:td8 27 a3 bxa3 Y2 - �
We shall now turn our attention to
the quieter 9 a3.
Game,91
Karpov-Kramnik
Dortmund 1995
1 d4 d5 2 e4 c6 3 tbe3 tbf6 4 e3 e6
5 tbf3 tbbd7 6 .ltd3 dxe4 7 .ltxe4 b5
8 i.d3 i.b7 9 a3
As we know, Black ideally wants to
play . . . c6-cS, but first he has to protect
bS with . . . a7-a6 or try . . . bS-b4. 9 a3 is
directed against both of these two
ideas. White will meet 9 . . . a6 with 10
b4, clamping down on cS; while by
attacking b4, he also hopes to make
. . . bS-b4 a little less tempting, since by
delaying e3-e4, he reserves the e4-
square for his queen's knight.
 M e r a n Va ria tio n w i t h 8
Question 10: Oh I see. It doesn't
sound earth-shattering!
Answer: No, it isn't, but this is one
of those annoying lines that can give
White a nice safe edge if Black is un­
wary or slightly careless.
9 . . . b4! 1 0 liJe4
After 10 axb4 i.xb4, Black easily
achieves . . . c6-c5. 10 ltJe4 is the move
which has revitalised this variation.
1 0 . . . liJxe4
10 . . . a5 is rather risky - see Games
92-95.
1 1 .li.xe4 bxa3! 1 2 bxa3 .li.d6
With a series of accurate moves
Black has solved the problem of his b­
pawn and he is now not far from play­
ing . . . c6-c5. The slight drawback to his
plan is that he has brought the white
light-squared bishop to e4, which
temporarily prevents . . . c6-c5 by pin­
ning the c6-pawn to the bishop on b7,
and that he has opened the b-file,
which gives White the chance to play
J:b 1 to harass the bishop on b7 along
the b-file, but these factors don't seem
to be sufficient for White to be able to
do any real damage.
1 3 i.d2 l:tb8 1 4 "a4!?
The sharpest continuation. 14 0-0
. . . iL b 7 : Wh i t e A l t e rn a t i v e s
0-0 15 .ib4 c5! was equal in Karpov­
Kramnik, Monaco (rapidplay) 1996.
14 . . . c5 1 5 i.xb7 l:txb7 1 6 dxc5
i.xc5 1 7 .li.a5 "b8 1 8 l:td 1 0-0 lh - lh
Chandler analyses 1 9 :xd7!? 'ii'e 8!
20 ltJe5 �b l+ (20 ... .lid6 2 1 0-0 .ixe5
22 �fd 1 �xd7 is equal and is safer) 2 1
�e2 �xh l 2 2 i.c3 �c8 2 3 �g4 g6 24
ltJxg6 'i'a6+! 25 '1itd2! hxg6 26 'ifxg6+!
fxg6 27 l::tg7+ �h8 28 lH7+ with a
draw by repetition!
The previous game was not too
thrilling, but it is a very effective
equaliser for Black. It is certainly pref­
erable to 10 . . . a5, which results in a
position that could arise from 8 . . . b4 9
ltJe4 (Chapter 1 1) but with extra
moves a2-a3 and . . . a7-a5 thrown in. As
we shall see, this should be nice for
White as he has the chance to open
the a-file with a3xb4 at some stage.
Game 92
Greenfeld-Khenkin
Israel 1995
1 liJf3 liJf6 2 c4 c6 3 liJc3 d5 4 d4
e6 5 e3 liJbd7 6 .li.d3 dxc4 7 .li.xc4
b5 8 .li.d3 i.b7 9 a3 b4! 1 0 liJe4 a5
1 1 liJxf6+
1 1 0-0 can be seen in Game 95.
1 1 . . .liJxf6 1 2 e4 .li.e7 1 3 ii'e2! liJd7
Black's sharp alternatives here,
13 . . . c5 and 13 . . . 'ii b 6, are considered in
Games 93 and 94 respectively.
1 4 0-0 0-0 1 5 l:te 1 ! ? ii'c7
15 .. J�e8 16 .lif4 c5 17 .lib5 is good
for White according to Greenfeld, but
15 . . . c5 is critical. After 16 d5, Green­
feld gives 16 . . . ltJb6 17 dxe6 fxe6 as un­
clear, but 18 .lie3 followed by consoli­
dation with l1ac1 and �ed 1 must give
1 19
Th e S e m i - Sla v
White a slight edge due to Black's ex­
posed queenside. 16 . . . exdS 17 exdS
l:ie8 18 ..1bS! and 17 . . . ..1f6 1 8 ..1f4! l:ie8
19 'i'c2 ..1xds 20 �xh7+ <;ith8 2 1
l:ixe8+ 'i'xe8 2 2 l:ie l both favour
White according to Greenfeld.
1 6 eS! l:tfe8 1 7 lLlgS SLxgS 1 8 SLxgS
lLlf8 1 9 �hS?
A mistake. 19 l:iac 1 ! , to prevent the
freeing . . . c6-cS, would have given
White an overwhelming position ac­
cording to Greenfeld.
1 9 . . . cS! 20 l:tac 1 'iVd7 2 1 l:txcS l:tec8
22 axb4 l:txcS 23 bxcS 'iVxd4 24
'iVe2 'iVxcS 2S h4 �dS 26 �f1 h6 27
�e 7 lLlg6 28 �xg6 % Y2 - lLleS!
To counter 13 'i'e2, Black players
have recently been trying some very
radical solutions.
Game 93
Karpov-Shirov
Linares 1 994 .
1 d4 dS 2 c4 c6 3 lLlf3 lLlf6 4 lLlc3
e6 S e3 lLlbd7 6 �d3 dxc4 7 �xc4
bS 8 �d3 -tb7 9 a3 b4! 1 0 lLle4 as
1 1 lLlxf6+ lLlxf6 1 2 e4 �e7 1 3 �e2!
cSt?
Thematic in a way, since it is part of
1 20
the Semi-Slav tradition not to mind
losing the right to castle. However,
this is usually done when White has
played e4-eS and Black has blocked the
central files with an immovable knight
on ds. Here, with the e4-pawn re­
stricting the knight on f6, Black's task
is much more difficult.
1 4 �bS+ Wf8 1 S dxcS �xcS
IS . . .i.xe4 16 i.e3 is slightly better
for White according to Karpov.
1 6 �d3 h6 1 7 0-0 g S
17 . . . g6 18 ..1f4 <;itg7 19 i.eS i s rather
nasty for Black according to Karpov.
1 8 �e3 �xe3 1 9 �xe3 Wg7 20
White is making good use of the eS­
square!
20 . . . lLld7 21 lLlc4 �e7 22 �d4+ eS
23 �d6 �xd6 24 lLlxd6 lLlcS 2S �c4
l:thd8 26 lLlfS+ wg6 27 f3 lLla4 28
l:tf2 lLlb6 29 �bS �c8 30 axb4 axb4
3 1 l:txa8 lLlxa8 32 lLle7+ Wg7 33
lLldS �e6 34 lLlxb4 l:tb8 3S �c6
l:txb4 36 �xa8 l:ta4 37 �b7 l:ta 1 + 38
l:tf1 l:ta4 39 l:tf2 l:ta 1 + 40 l:tt 1 l:ta4
41 SLc6 l:tb4 42 l:tt2 l:tc4 43 �dS
�xdS 44 exdS l:td4 4S b3 l:td 1 + 46
l:tf1 l:txdS 47 b4 'it>t6 48 l:tb 1 'it>e6 49
bS 'it>d7 SO b6 Wc8 S1 b7+ Wb8 S2
 M e ra n Va ria tio n w i t h 8
l:tb6 .::t d 4 53 �f2 h 5 54 �e3 g4 55
�f6 �d7 56 1:U5 �xb7 57 l:txh5
l:tb3+ 58 'it>f2 gxf3 59 gxf3 f6 60 h4
<l;c7 6 1 l:th6 l:tb6 62 h5 �d7 63 J:tg6
<l;e6 64 h6 l:tb7 65 'it>g3 l:tb1 66 'it>h2
.::t b 7 67 �h3 l:tb1 Yz - Yz
Game 94
Bareev-Dreev
Russia 1996
1 d4 d 5 2 c4 c6 3 ttJf3 ttJf6 4 ttJc3
e6 5 e3 ttJbd7 6 .Jtd3 dxc4 7 i.. xc4
b5 8 .i.d3 .i.b7 9 a3 b4! 1 0 ttJe4 a5
1 1 ttJxf6+ ttJxf6 1 2 e4 .i.e 7 1 3 �e2!
�b6 ! ? 1 4 0-0 c5 1 5 axb4! cxb4
lS . . . axb4 16 lIxaS+ ..txaS 17 ..tbS+
<;t>fs IS dS exdS 19 exdS ..txdS 20 l:te 1
'i'b7 2 1 liJeS is clearly better for White
according to Bareev.
1 6 d5 exd5 1 7 .Jte3 i.. c 5 1 8 .i.xc5
�xc5 1 9 l:tac 1 �b6 20 i.. b 5+ �f8
21 e5 ttJe8 22 l:tfe 1 ttJc7? 23 e6!
ttJxe6 24 ttJe5
With the threat of liJd7+, forking
king and queen.
24 . . . �g8 25 ttJxf7 �xf7 26 l:tc6!
26 . . . �d8 27 �xe6+ �f8 28 �d6 1 -0
The other plan with 0-0 and e4-eS
also looks very promising here!
. . . Jt.. b 7 : Wh i t e A l t e rn a t i v e s
Game 95
Gavrilov-Novikov
Riga Open 1995
1 d4 d5 2 c4 e6 3 ttJf3 ttJf6 4 ttJc3
c6 5 e3 ttJbd7 6 i.. d 3 dxc4 7 i.. x c4
b5 8 .Jtd3 i.. b 7 9 a3 b4! 1 0 ttJe4 a5
1 1 0-0 .i.e 7 1 2 ttJxf6+ ttJxf6 1 3 e4
0-0 1 4 e5 ttJd7 1 5 i.. e4 1:[b8 1 6
axb4! axb4 1 7 �c2 h6 1 8 .i.e3
Gavrilov shows that the freeing
lS ... cS fails tactically due to 19 ..txb7
l1xb7 (19 ... cxd4 20 jLxh6 J:!xb7 21 ..tf4
is best but still bad) 20 dxcS liJxcs 2 1
..txcS 'iVcs 2 2 I:taS! (the difference!)
1 8 . . . ttJb6 1 9 ttJd2 ttJd5 20 ttJb3
�b6?! 21 .i.d2 1:[fd8 22 1:[fc 1 1:[a8 23
1:[xa8 l:txa8 24 ttJc5 .i.c8 25 �b3
1:[a5 26 �g3 �h8 27 �h3 i.. f 8 28
i.. b 1 ttJe7 29 g4 �g8 30 �d3 ttJg6
31 ttJb3 1:[d5 32 h4 i.. e 7 33 h5 ttJf8
34 �c4 i.. a 6 35 �xc6 �xc6 36
�xc6 i.. e 2 37 i.. e4 l:td7 38 f3 .i.d 1
39 ttJa5 l:txd4 40 i.. e 3 l:td7 41 �c8
i.. a4 42 i.. c 6 l:td3 43 .i.f2 l:td 1 + 44
�g2 1:[c 1 45 l:txf8+ 'it>xf8 46 i.. xa4
1:[a 1 47 b3 l:ta2 48 �g3 f6 49 ttJc6
1:[xa4 50 ttJxe 7 'it>xe 7 51 bxa4 b3 52
i.. c 5+ 1 -0
12 1
Th e S e m i - Sla v

Summary
At the moment Black seems to be holding his own quite comfortably after both
9 e4 b4 10 lDa4 c5 and 9 a3 b4! 10 lDe4 lDxe4. One thing, however: don't touch 9
a3 b4 10 lDe4 a5 .

1 d4 d 5 2 c4 c6 3 liJf3 liJf6 4 liJc3 e6 5 e3 liJbd7 6 �d3 dxc4 7 �xc4 b5 8

�d3 �b7

9 e4
9 a3 b4 10 lDe4
10 . . . lDxe4 Game 91
-

10 . . . a5
I 1 lDxf6 lDxf6 12 e4 i.e7 13 'iYe2 (D)
13 . . . lDd7 Game 92
-

- 13 . . . c5 Game 93
1 3 . . . 'iYb6 Game 94
-

-1 1 0-0 Game 95
9 . b4 1 0 liJa4 c5 1 1 e5 liJd5 (D) 1 2 0-0
. .

12 lDxc5 Game 89; 12 dxc5 Game 90

- -

1 2 . . cxd4
.

12 . . . a6 Game 84
-

1 3 ];tel
1 3 lDxd4 lDxe5 14 i.b5+ lDd7 15 lIe 1 �c8 16 'iYh5 g6
17 'iYe2 Game 87; 17 'iYe5 Game 88
- -

1 3 . . . g6 (D) 1 4 �g5
14 iLd2 Game 86
-

1 4 . . :�a5 Game 85
-

1 3 'ike2 11 . . . liJd5 13 . . . g6

1 22
1 d4 d 5 2 c4 c6 3 lZJf3 lZJf6 4 lZJc3 10 cxd4
. . .

e6 5 e3 lZJbd7 6 �d3 dxc4 7 �xc4 Black counters the threat against his
b5 8 �d3 a6 9 e4 c5 1 0 e5 knight on f6 attacking White's knight
The variation with 9 e4 cS 10 eS is on c3. 1 1 exf6 bxc3 gives Black a good
the most aggressive way for White to game. Usually, White replies with
play, and is a direct attempt to refute, 1 1 lZJxb5
or at least exploit the inconveniences
of, Black's move order with 8 . . . a6.
Question 1: How is that?
A nswer: With 8 . . . �b7, Black only
plays . . . c6-cS once he has placed his
bishop on the a8-h 1 diagonal. Conse­
quently Black always has piece cover
of the dS-square; e4-eS can never really
inconvenience him since he can al­
ways play his knight to dS without
any bother. However, in the 8 . . . a6 9
e4 line, Black has to rush . . . c6-cS
without first developing his light­ Question 2: Why?
squared bishop. When White plays 10 A nswer: Rather than allow Black
eS attacking the knight on f6, 10 . . . ttJdS just to take White's knight on c3 ,
is poor since after 1 1 ttJxdS Black must White makes a 'desperado' sacrifice to
recapture with a pawn on dS, blocking get as much as he can for the knight
the a8-h 1 diagonal and leaving him before it succumbs to the inevitable.
with a rather shaky position. Now things can get a little confus­
Black's only chance is to fight fire ing. The main line here has histori­
with fire and play cally been 1 1 . . .ttJxeS, while both

1 23
Th e S e m i - Sla v
1 1 . . .CLlg4 and 1 1 .. .axb5 12 exf6 'ii'b 6
have also had a lot of theory devoted
to them. However in the last couple
of years, there have been virtually no
games played with these lines, since a
completely new main line has
emerged. It is so new in fact that Peter
Wells' magisterial The Complete Semi­
Slav (published in 1994) contains vir­
tually no mention of it!
Consequently, I have concentrated
almost exclusively on this new line
and will just give a summary (almost a
history lesson!) of the other lines.
1 1 axb5 1 2 exf6 gxf6 ! ?
. . .
12 . . . 'ii'x f6?? ( 1 2 . . . CLlxf6 13 �xb5+
followed by 14 CLlxd4 or 'ii'xd4 wins a
pawn) 13 .ig5 traps the queen.
Question 3: I can see that 13 �xb5
loses a piece to 13 ... 'ii' a 5+, but can't
White simply play 13 CLlxd4, regaining
his pawn and leaving him with a
clearly better pawn structure?
A nswer: This is a crucial question.
Let us take a look at the stem game for
this line.
Game 96
Altermari:' Chernin
Qroningen (PCA Qualifier)l993
1 d4 d 5 2 c4 c6 3 ti::l c 3 ti::l f 6 4 e3 e6
5 ti::lf 3 ti::l b d7 6 �d3 dxc4 7 �xc4 b5
8 �d3 a6 9 e4 c5 1 0 e5 cxd4 1 1
ti::l x b5 axb5
Black's alternatives here are consid­
ered in Game 100.
1 2 exf6 gxf6
The old 12 . . . iLb7 is the subject of
Game 99.
1 3 ti::l x d4 �b6 ! !
The key move that has revitalised
1 24
this variation. Alterman showed the
way to play against 13 . . . iLb7 in a game
against Har Zvi in Israel 1993: 14 iLe3!
(White's idea is not to castle, but to
develop by playing the king to e2.
This avoids giving Black any counter­
play against the white king by pressur­
ing g2 after castling) 14 ... .i.xg2
(14 .. J�g8 is simply met by 15 'ii'h 5!
.l::[xg2 16 �e2! with a clear advantage
for White according to Alterman) 15
1:lg1 �d5 16 �xb5 'ii'c7 17 'ii'd3 'ii'x h2
18 �xd7+ �xd7 19 'ii'b 5+ �d8 20
CLlc6+ �c7 2 1 iLb6+ with a very big
attack. White developed his play so
quickly in the game above because he
was able to maintain his knight on d4,
where it is is ideally placed on d4 be­
cause it attacks all the loose light
squares in Black's position: it can
threaten a sacrifice on e6, and it sup­
ports iLxb5.
Chernin's 13 ... 'ii'b 6 puts pressure on
the knight and intends to increase this
by playing . . . �c5, which forces White
to take early measures to bolster his
knight. The second and very crucial
point is that by leaving the light­
squared bishop on c8, Black robs the
force from iLxb5. In the 13 . . . �b7 line,
 M e r a n Va ria t i o n w i t h 8 .
once the black queen left d8, ..txbs
threatened ..txd7+, drawing the black
king into the firing line of the white
queen. Here i.xd7+ is simply met by
the recapture on d7 with the bishop.
13 .. :i!fb6 thus destabilises White's
strength (his d4-knight) and neutralises
one of his major threats (i.xbS, aim­
ing for .1i.xd7+) . Coupled with the
open lines this enables Black to
quickly whip up a ferocious initiative.
14 �e3
14 .1i.xbS is met by 14 . . . .1i.b4+ 15
�fl eS! 1 6 i.xd7 .1i.xd7 with a strong
initiative, while 14 ..te4 is met by the
surprising 14 . . . �a4! 15 .1i.e4 .1i.cs with
compensation according to Chernin.
Finally, 14 lbxbs i.b7 15 0-0 lilg8
gives a huge attack.
14 . . . �b4+! 1 5 'it>f1 �c5! 1 6 �xb5
e5 1 7 �xd7+ �xd7 1 8 �f3 �a6+
1 9 l2Je2 l::t g 8!? 20 �xc5 �c6 21 �h5
�xg2+ 22 '1t>e 1 �xh 1 23 'it'xh7 O-O-O!
Black is just winning but the finish
is very nice!
24 l:tc1 �b8 25 'it'xf7 �d5 26 �h5
�d3 27 'it'h4 �c4 0-1
28 'ii'xc4 'ifd2+ 29 �fl 'ii'd 1+! ! 30
�xd 1 �xd 1 + is mate!
Consequently, White players have
. . a 6 : O ld M a in L in e - 9 e4 c 5 1 0 e 5
shunned 13 lbxd4 in favour of 13 0-0,
putting the king into safety first!
Game 97
Gelfand-Shirov
, Linares 1997
1 d4 d5 2 c4 c6 3 l2Jc3 l2Jf6 4 l2Jf3
e6 5 e3 l2Jbd7 6 �d3 dxc4 7 �xc4
b5 8 �d3 a6 9 e4 c5 1 0 e5 cxd4 1 1
l2Jxb5 axb5 1 2 exf6 gxf6 1 3 0-0
'it'b6!
At the cost of a pawn and his pawn
centre, White has greatly loosened
Black's structure - e4-eSxf6 has dou­
bled Black's f-pawns, while lbxbS has
reduced Black's queenside to an iso­
lated b-pawn. Black's king, in contrast
to White's, lacks a haven on either
wing and will have to take its chances
in the centre behind the f7, f6, e6 clus­
ter. Due to his large range of weak­
nesses, it is difficult for Black to con­
solidate his extra pawn, so he should
seek instead to generate activity.
1 4 �e2!
Attacking the pawn on bs. 14 .1i.e4
..tb7 15 ..txb7 'i'xb7 16 lbxd4 1.ig8 17
f3 lbeS 17 ... lbeS 1 8 'iWe2 ..tcS 19 i.e3
lid8 20 :ad 1 lbc4 21 .1i.f2 'i'b6! was
equal, Kamsky-Kramnik, Linares 1994
1 4 . . . �a6
Black's wants to keep hold of some
queenside light squares by keeping the
b-pawn on bS. The alternative 14 . . . b4!
is seen in the next game.
1 5 l:td 1 !
This move has been causing Black
problems. White's removes the rook
from the fl-a6 diagonal and takes the
sting out of Black's idea of . . . lbxd3.
1 5 . . . �c5
1 25
Th e S e m i - Sla v
15 . . .'�jc5 is mentioned by Romanov
and now 16 ttJxd4 ttJxd3 17 'iYxd3.
Although Black has won the bishop
pair, it is at the cost of his extra pawn
and a couple of tempi. By exchanging
off the bishop on d3 , Black reduces his
opponent's ability to use the queen­
side light squares to attack the black
king. However, the exchange of
Black's knight further loosens his own
position. In particular, the kingside
dark squares - f6, for example - are
much easier to attack. Black's problem
is that although he has plenty of open
lines for his pieces, and can thus acti­
vate them easily, it is a difficult for
him to coordinate them.
After 17 'iYxd3, White intends to set
up his ideal attacking formation:
queen on the g-file (stopping Black's
counterplay with . . . �g8 and prevent­
ing . . . �e7 due to 'iYg7) , bishop to e3
and queen's rook to the c-file.
1 6 a3! .lib7 ! ? 1 7 .lixb5 1:1g8 1 8 b4? !
This seems a trifle rushed to me. 18
�f4!? looks interesting, preparing to
retreat the bishop to g3 and deal with
Black's pressure against g2.
1 8 . . . �e7 1 9 1:1xd4!?
19 �f4 is still worth a try here.
126
19 .. :�i'xd4! 20 4Jxd4 .lixg2!
Amazingly White has no way to get
out of this with a material advantage:
2 1 �xd7+ �xd7 22 'iYb5+ fails to
22 . . . �c6+!
21 �e3 �h3+ 22 'Ot>h 1 .lig2+ 23 �g 1
.lih3+ 24 �g3! ?
Very brave, sacrificing a pawn for
chances with his passed pawn.
24 . . . 1:1xg3+ 25 hxg3 .ltxb4 26 �b2
�e5 27 a4 e5 28 4Je6 �f8 29 a5
4Jb8 30 1:1e 1 .lid6 31 4Jxb8 1:1xb8 32
a6 1:1xb5 33 a7 1:1a5 34 1:1a 1 1:1xa7 35
1:1xa7 �g7 36 �h2 �e6 37 1:1a8 .lie5
38 'Ot>g 1 h5 39 .lie1 �g6 40 1:1h8
.lig4 41 'Ot>g2 �g7 42 1:1b8 .lie6 43
1:1b5 .lid4 44 1:1b8 % - %
Game 98
De Sousa-N . Eliet
French Championship 1996
1 e4 e6 2 d4 d5 3 4Je3 4Jf6 4 e3 e6
5 4Jf3 4Jbd7 6 �d3 dxe4 7 .ltxe4 b5
8 .lid3 a6 9 e4 e5 1 0 e5 exd4 1 1
4Jxb5 axb5 1 2 exf6 gxf6 1 3 0-0
'iVb6 1 4 �e2! b4!
This is Black's best chance: by play­
ing the pawn to b4, Black weakens his
grip over the queenside light squares
 M e r a n Va ria tio n w i t h 8
even more, but by occupying b4, he at
least creates the chance of defending
the d4-pawn with . . . .1i.c5 without hav­
ing the bishop driven away.
1 S l:td 1 �cS! 1 6 �f4
16 .ixh7 .1i.a6 (16 . . . llxh7 loses the
exchange to 17 'iVe4!) 17 .1i.d3 e5!? is
extremely murky. Black will retreat
the bishop to e7 and attack the light
squares with . . . tDc5. White should
stop this by transferring the king's
knight to b3 or e4 via d2.
1 6 . . . �b 7 1 7 .i.bS l:td8 1 8 l:tac 1 l:tg8
1 9 .i.g3 Wf8 20 'ilVc4 .i.dS 21 'ii'd 3
.i.xf3 22 gxf3 tLle5 23 'ii'e 2 d3 24
.i.xd3 tLlxd3 25 l:txd3 .i.xf2+ 26 Wg2
.i.d4 27 l:tc4 eS 28 'ii'd 2 fS 29 Wf1
l:ta8 30 l:txb4 �a6 3 1 'ii'e 2 f4 32
.i.f2 �g6 33 l:tbxd4 exd4 34 .i.xd4
l:te8 35 .i.c5+ Wg7 36 'ii'd 2 �fS 37
l:tdS 'iVh3+ 38 �g2+ �xg2+ 39
Wxg2 Wf6+ 40 Wf1 l:ta8 41 a3 �e6
42 l:td6+ We5 43 l:td2 'h - 'h
We shall now take a rapid look
through the older lines that used to
fashionable in this variation.
Game 99
Csiszar-Sploshrov
Budapest Open 1996
1 d4 d S 2 c4 e6 3 tLlc3 c6 4 e3 tLlf6
S tLlf3 tLlbd7 6 .i.d3 dxc4 7 .i.xc4 bS
8 .i.d3 a6 9 e4 c5 10 eS cxd4 1 1
tLlxb5 axb5 1 2 exf6 �b7 ! ?
A n interesting little idea.
1 3 fxg7
Very natural, but I wonder if 13
tDxd4 is possible here, hoping to trans­
pose to Alterman-Har Zvi in the notes
to Game 96 after 13 . . . gxf6 14 .1i.e3!
1 3 . . . .i.xg7 1 4 0-0 0-0 1 5 l:te 1
. . . a 6 : O l d M a in L in e - 9 e4 c5 1 0 e 5
The greedy 15 .1i.xb5 tDc5 16 .if4
e5!? 17 .1i.g3 tDe6!, intending . . . e5-e4,
gave Black good compensation III
Sakaev-Belikov, USSR 1990.
1 S . . . tLlcS ! ?
15 . . . 'iVb6 16 �e 1 .1i.d5 17 h4!, in­
tending h4-h5-h6, was dangerous for
Black in Shabalov-Kishnev, USSR
1988. The text is often dismissed as
bad, but it is not quite clear.
1 6 �xh7+! Wxh7 1 7 tLlg5+ Wg8! ?
This is the point! 1 7 . . .';t.tg6 1 8 'iNg4
f5 19 'iWg3! wf6 20 'i'e5+ Wg6 2 1 �xc5
is very unpleasant for Black.
1 8 'iVh5 l:te8 1 9 'iVxf7+ �h8 20 l:te5?
This is not the best. Unfortunately
for Black, White seems to be able to
obtian an advantage with 20 b4! 'iVd5
21 f3, when 2 1 . . .d3 22 bxc5 d2 23
�xd2 'iYxd2 24 'iVh5+ �g8 25 :tad1
'i'c2 26 �d7! "iixc5+ 27 'iti>h 1 is very
dangerous for Black.
20 . . . �e7 21 'ii'h S+ �g8 22 'iVh7+
Wf8 23 tLlxe6+ �xe6 24 l:tfS+ ji'xfS
2S �xf5+ Wg8 26 .i.f4 l:tac8 27 h4
d3 28 h5 .i.e4 29 'ii'g 4 Wh7 30 f3
1:[g8 31 �h3 .i.xf3 32 g4 .i.xb2 33
l:tf1 'h - 'h
On move 41 after a time scramble
that left both score-sheets illegible!
127
Th e S e m i - Sla v
The final idea is a line that theoreti­
cally always seems to be doing fine,
but is rather difficult to handle.
Game 100
Bareev-Yusupov
Linares 1993
1 d4 d 5 2 e4 e6 3 lDe3 lDf6 4 e3 e6
5 lDf3 lDbd7 6 i.d3 dxe4 7 i.xe4 b5
8 �d3 a6 9 e4 e5 1 0 e5 exd4 1 1
lDxb5 lDxe5
By eliminating White's pawn on e5,
Black makes sure that he keeps his
kingside pawn structure intact.
1 1 . . .ctJg4 12 ctJd6+! i.xd6 13 exd6
'i'a5+ 14 iYd2! iYd5 15 iYg5! 'i'xg5 16
.txg5 ctJc5 17 �e2 f6 18 �d2 e5 19 0-0
was good for White in Greenfeld­
D.Gurevich, Beersheva 1994.
1 2 lDxe5 axb5 1 3 i.xb5+ i.d7 1 4
lDxd7 'iVa5+ 1 5 Si.d2 'iVxb5 1 6 lDxf8
�xf8
Still the most popular, although
Lukacs's 16 .. Jixf8 deserves a mention
and brought him a good win against
Ibragimov in Budapest 1992 after 17 a4
'i'c4 1 8 b3 'i'd3 19 'i'e2!? 'i'xb3 20 0-0
lha4 2 1 'i'e5 ctJd7! 22 iYxg7 �xa1 23
�xa1 'iWb2! 24 lla8+ We7 25 �g5+
1 28
Wd6 26 l::ra6+ \t>d5! 27 h4 'i'b 1+ 28
Wh2 'i'b8+ and now White should
have played for the draw with 29 Wgl .
1 7 b3 lDd5
17 ... We7 was rehabilitated by Byk­
hovsky against Greenfeld in Kfar-Sava
1995, when 18 a4 'i'd5 19 0-0 llhc8 20
b4 ctJe4 2 1 b5 llc4! 22 b6 ctJxd2 23
'i'xd2 I!axa4 24 llxa4 llxa4 25 �b 1
fla8 26 'i'c2 'i'd6! 27 'iVxh7 flb8 29 b7
'iVd5 gave Black a fine game.
1 8 a4 'iVd3
To stop White from castling. 19
�b4+ fails because of 19 . . . ctJxb4.
1 9 'iVe2 'iVxb3 20 0-0 'tir'e2 21 life 1
d3 22 'iVe 1 'iVb2 23 lie4? !
23 as would have been better ac­
cording to Yusupov.
23 . . . lib8 24 a5 �e7 25 a6 lib5? !
25 . . . 11hc8 26 �xc8 llxc8 27 a7 :ta8
28 1:tb 1 'i'd4 29 �b7 wd6! would have
given Black a clear advantage accord­
ing to Yusupov.
26 liee1 'iVd4 27 lieb 1 lixb 1 28
'iVxb 1 Ub8 29 'iVe 1 lia8 30 lia5 lia7
3 1 'tir'a3+ �f6 32 lia4 'tir'b6 33 'tir'xd3
g6 34 .th6 �b2 35 h4 �e7 36 g3
'iVe3 37 'iVb5 �e7 38 �b2 lia8 39
a 7 f6 40 .tf4 lDxf4 41 l:lxf4 e5 42
lia4 �f7 43 'iVd2 �g7 44 'iVe3 �b7
45 'iVe5 lid8 46 lia 1 lia8 47 �h2
�g8 48 g4 'iVe4 49 lia4 �b7 50 g5
fxg5 5 1 hxg5 h6 52 gxh6 �h7 53
lia3 �f7 54 �e3 �b7 55 �g3 'iVe7
56 lia6 �f7 57 �h2 �b7 58 lia5
"ile7 59 lia6 'iVb7 60 'iVd3 'iVf7 6 1
"ile3 Y2 - Y2
White's main alternative to the cra­
ziness of 1 1 ctJxb5 is to play instead 1 1
ctJe4, an old move that was recently
resurrected by the Spanish player Pa­
blo San Segundo.
 M e ra n Va ria t i o n w i t h 8
Game 101
San Segundo-Vera
A lcobendas 1 994
1 d4 lDf6 2 c4 e6 3 lDf3 d5 4 lDc3
c6 5 e3 lDbd7 6 �d3 dxc4 7 �xc4
b5 8 �d3 a6 9 e4 c5 1 0 e5 cxd4 1 1
lLle4
1 1 . . . lDd5
This seems like Black's best reply.
1 1 ...ttJxe4 12 i.xe4 i.b4+ 13 i.d2
i.xd2+ 14 'iWxd2 �b8 15 'ilixd4 gives
White a small plus.
1 2 0-0 �c7 ! ?
This is the most logical way to meet
this line. Black makes use of the fact
that the knight on e4 blocks the e-file
to nip off White's e-pawn and go two
pawns up. But it's very risky! 12 . . . h6
13 a4 b4 14 i.. c4 (14 i.. c2!?, intending
�xd4) 14 . . . i.b7 15 'iYxd4 'ilfb6 16 l:tdl
l1c8 17 as 'i'xd4 1 8 l:txd4 ttJc5 19
ttJd6+ i.. xd6 20 exd6 0-0 21 .id2 was a
touch better for White in Korchnoi­
Gelfand, Madrid 1996.
1 3 �g5 lDxe5 14 lDxe5 'iix e5 1 5 tte 1
Vera suggests that 15 f4 'ii'b 8 16 f5
would have been dangerous.
1 5 . . ... b8
. . . a 6 : O ld M a in L in e - 9 e 4 c 5 1 0 e 5
At the Linares Open in 1995,
Dmitri Gurevich played 15 ... 'ilfc7
against San Segundo, when 16 �c1
'iVd7 17 .ib 1 i.. b4 18 'itxd4 0-0 19
ttJf6+ ttJxf6 20 'iWxb4 �d8 21 .ixf6
gxf6 22 11c3 'Wd4 23 'ita3 was very
dangerous but not conclusive. Vera
also suggests 15 . . . .ib7!? 16 ttJf6+ gxf6
17 ltxe5 fxe5, sacrificing the queen to
blunt White's initiative.
1 6 1:[c 1 �b6 1 7 'iig 4 �b7 1 8 a4
�b4 1 9 1:[e2 bxa4 20 �h4 'it>f8?
The decisive mistake after which
Black seems to be lost. 20 . . . 0-0, how­
ever, does not look too bad for Black:
21 ttJf6+ ttJxf6 22 i.xf6 g6 23 i.xd4
iVd8 24 iVf4 i.d6 25 'i'h6 e5 26 lIe3
exd4 27 �h3 l::.e 8 28 'iixh7+ �f8 29
i.xg6 'iYf6 'unclear' is Vera's analysis,
but this looks good for Black to me.
21 lDg5 1:[e8 22 �h5 g6 23 �xg6!
23 .. .fxg6 24 'iif 3+ 'ifo>g7 25 �f7+
'Oth6 26 ttxe6 �c6 27 h3 ttef8 28
1:[cxc6 1:[xf7 29 lDxf7+ 'it>g7 30 1:[xb6
'Otxf7 31 l::tx a 6 1:[c8 32 tte4 ttc 1 + 33
'Oth2 �d6+ 34 ttxd6 1 -0
This line thus deserves further in­
vestigation as a more sensible and
measured approach to the 8 . . . a6 9 e4
c5 10 e5 cxd4 variation!
1 29
Th e S e m i - Sla v

Sum mary
If you don't mind unbalanced pawn structures then this chapter is for you!
Black is not doing badly at all in general, but it obviously takes a special type of
player Oike Shirov!} to thrive in this sort of situation. The older lines with
1 1 . . .liJxeS are theoretically healthy, although it can be a little daunting in a prac­
tical game to face two connected passed pawns!

1 d4 d 5 2 c4 c6 3 tDf3 tDf6 4 tDc3 e6 5 e3 tDbd7 6 iLd3 dxc4 7 iLxc4 b5 8

.ltd3 a6

9 e4 c5 1 0 e5 cxd4 (D) 1 1 tDxb5

1 1 liJe4 Game 101
-

1 1 axb5
. . .

1 1 . . .liJxeS Game 100

-

1 2 exf6 (D) gxf6

12 . . . .tb7 Game 99
-

1 3 tDxd4
13 0-0 'iWb6 14 'iWe2 (D)
14 . . . .ta6 Game 97
-

14 ... b4 Game 98
-

1 3 . . . �b6 Game 96
-

1 2 exf6 1 4 �e2

1 30
 CHA PTER ELEVEN
Meran Variation : Systems
with an Early b 5-b4 . . .
1 d4 d 5 2 c4 c6 3 tiJf3 tiJf6 4 tiJc3
e6 5 e3 tiJbd7 6 �d3 dxc4 7 �xc4
b5 8 �d3
In this chapter we focus on lines
with an early . . . b5-b4 for Black.
Question 1: How does this system
arise?
Answer: There are two move or­
ders. 8 . . . b4 leads directly into the . . . b5-
b4 complex. By delaying the devel­
opment of the bishop on c8, Black
gains a variety of extra possibilities,
though it is not clear how good these
are! The most common move order,
however, is 8 . . . �b7 9 0-0 b4. We shall
discuss the significance of this move
order later, but suffice it to say that
this does give White a couple of extra
possibilities: he can play 9 e4 (when
9 . . . b4 was seen in Games 84-90) or 9
a3 (when 9 . . . b4 was the subject of
Games 9 1-95) .
Question 2: What is the point of
. . . b5-b4?
A nswer: With this advance Black
moves the b-pawn from its attacked
square on b5 to the safe b4-square. It
also gains a tempo on the knight on
c3, disturbing White's piece set-up and
allowing Black to play . . . c6-c5 with
the greatest possible speed. However,
by playing . . . b5-b4, Black loses some
control over the queenside light­
squares: he gives away the c4-square
(which is a particularly nice square for
a white knight) and the a4-square.
Moreover, unlike in the 8 . . . �b7 9 e4
b4 10 lLJ a4 line, Black does not force
White's queen's knight offside, as it
can come to the centre with lLJe4.
First, let us see what happens if
Black plays the straightforward but
rather inflexible 8 . . . �b7 9 0-0 b4 10
lLJe4 lLJxe4.
Game 102
Sadler-Bisby
Isle ofMan Open 1995
1 d4 d5 2 c4 c6 3 tiJc3 tiJf6 4 e3 e6
5 tiJf3 tiJbd7 6 �d3 dxc4 7 �xc4 b5
8 JiLd3 JiLb 7 9 0-0 b4 1 0 tiJe4 tiJxe4!?
1 1 �xe4
The exchange on e4 has blocked
131
Th e S e m i- Sla v
White from pushing e3-e4. Moreover,
by forestalling tLlxf6+, Black keeps his
queen's knight on d7, where it sup­
ports the . . . c6-cS break. The drawback
of the exchange is that it brings the
white bishop to the hI-aS diagonal,
making it less easy for Black to
achieve . . . c6-cS quickly.
1 1 . . . .ie7 1 2 �a4!? �b6 1 3 �d2? !
After this, White has no advantage.
During the postmortem, I was rather
confused as I had seemed to get noth­
ing whilst playing the only moves!
Then Peter Wells showed me 13 tLld2!
The knight is headed for c4, and as a
bonus, it will arrive with tempo by
attacking the queen on b6. Moreover,
after . . . tLlf6, the bishop on e4 can re­
treat to f3 and maintain White's pres­
sure along the h I-aS diagonal. White
will then develop his dark-squared
bishop on b2-b3 and .i.b2, making
sure of course that his queen does not
get trapped on a4! In fact, tLld2 would
also not be bad on the 12th move as
well.
In the game, I was struggling.
1 3 . . . 0-0 1 4 l:tfc 1 l:tfc8 1 5 l:tc4 lLlf6?!
A slightly nervous move. l S ... aS 16
�ac1 'iVa7!, threatening . . . tLlb6, would
1 32
have forced 17 'iVdl cS! with a slight
edge for Black.
1 6 .ixb4 lLlxe4 1 7 !iLxe7 �xb2 1 8
l:tf 1 a5? 1 9 l:tc2 �b5 20 �a3 �d5
21 l:tfc 1 l:tc7 22 !iLe5 !iLe8 23 lLle5
f6 24 lLlc4 lLlxe5 25 lLlb6 �e4 26
lLlxa8 l:ta7 27 lLlb6 1 -0
More often than not Black plays
this line with the immediate S ... b4.
Game 103
Ya kovich-Sveshnikov
Yerevan Open 1 996
1 d4 d5 2 c4 c6 3 lLlc3 lLlf6 4 e3 e6
5 lLlf3 lLlbd7 6 !iLd3 dxc4 7 !iLxc4 b5
8 !1J.. d 3 b4 9 lLle4 lLlxe4
9 . . . .i.b7 10 tLlxf6+ gxf6!? 1 1 e4 cS 12
i.e3 'iVb6 13 �c 1 O-O-O!? 14 0-0 �bS
was the rather outrageous attempt in
Yakovich-Filippov, Perm 1997. After
15 'iVe2 �gS 16 �fdl fS!? 17 il.f4+ i.d6
lS i.xd6+ 'iVxd6 19 dxcS tLlxcs 20
il.b l Black's position was very loose.
Black's other alternatives, 9 ... cS and
9 . . . it.. e 7, are considered in Games 104-
105 and 106-109 respectively.
1 0 !1J.. xe4 �b6!?
10 ... .i.b7 11 0-0 would have trans­
posed to the game above, but by using
 M e r a n Va ria tio n : S ys t e m s with a n Ea rly
the flexible (8 . . . b4) move order, Black
hopes to improve on that line. By pro­
tecting the attacked c6-pawn with the
queen, he frees his light-squared
bishop to come to the a6-fl diagonal
to cover the c4-square. His aim is then
to move his rook from a8 and break
with ... c6-cS.
1 1 'ifa4!?
11 0-0 seems sensible and if 1 1 . . .iLa6
then 12 �e 1 �c8 13 a3 is troublesome
for Black.
1 1 . . . .i.b7 1 2 lbd2 l:tc8 1 3 a3 c5!
This novelty seems to equalise quite
comfortably, so White definitely
needs something earlier.
14 dxc5 .i.xc5 1 5 axb4
Or 15 0-0 bxa3 16 bxa3 '1i'a6!
1 5 . . . .i.xb4 1 6 0-0 l:tc7 !
The key defensive move.
1 7 .i.xb7 'iVxb7 1 8 lbb3 0-0 1 9 l:td 1
l:tb8 20 .i.d2 h6 2 1 lba5 'iVb5 22 h3
.i.xd2 23 �xb5 l:txb5 24 l:txd2 lbf6
25 l::t a4 lbd5 26 lbc4 Wf8 27 Wf1
'li;e7 28 We2 f6 29 g3 l::t b c5 30 lba3
l:tb7 31 l:tc4 l:tbc7 32 l:txc5 l:txc5 33
lbc2 a5 34 h4 g5 35 hxg5 hxg5 36
e4 lbc7 37 lbe3 lbb5 38 Wd 1 lbd6
39 l::td 4 l:te5 40 f3 f5??
When this happens, it really is
heartbreaking.
41 l:txd6 fxe4 42 fxe4 l::t xe4 43 l:td3
�f6 44 �d2 l:tb4 45 Wc3 �e5 46
lbc4+ We4 47 l:te3+ Wf5 48 lbxa5
l:tb8 49 lbc4 Wg4 50 b4 �h3 5 1
tbd6 g 4 5 2 lbe4 Wg2 5 3 �c4 l:tc8+
54 �d4 l:tb8 55 l:tb3 l:tb5 56 lbc5
1 -0
A more aggressive and consistent
continuation for Black is 9 ... cS. He
has already safeguarded his b-pawn, so
why not play . . . c6-cS immediately?
. . . b 5-b4
Game 104
Ivanchuk-Oll
Biel lnterzonal 1993
1 c4 e6 2 d4 lbf6 3 lbf3 d5 4 lbc3
c6 5 e3 lbbd7 6 .i.d3 dxc4 7 .i.xc4
b5 8 .i.d3 b4 9 lbe4 c5
10 lbxf6+
After 10 0-0, Piket's 10 .. :iWb6! is
best. By putting pressure on d4, Black
stops White from setting up his attack­
ing structure after e3-e4 . . . cSxd4, liJxd4
as the queen on b6 defends the pawn
on d4. After 1 1 liJxf6+ gxf6 12 b3
cxd4! (the right time, as 12 . . . .ih7 13
iLh2 cxd4 14 .ixd4 is annoying for
Black) 13 exd4!? (13 liJxd4 .ics is fine
for Black) 13 ... .ib7 14 �e 1 iLd6 15
.ie4 .ixe4 16 lhe4 'i'b7 Black had
equalised in Bareev-Piket, Dortmund
1995.
1 0 . . . gxf6
10 . . . liJxf6 1 1 liJeS!, with ideas of
.ibS+ and 'i'f3, is difficult for Black.
1 1 .i.e4!?
11 0-0 ifb6! transposes to the note
to White's 10th move and 1 1 e4!? is
considered in the next game.
1 1 . . J:tb8 1 2 0-0 f5 1 3 .i.c6 ikc7
1 33
Th e S e m i - Sla v
Trying to drive the white bishop
from the long diagonal. 13 .. J�b6!? is
also possible.
1 4 d5
On 14 ctJe5, 011 gives 14 . . . �g7 15
iLxd7+ �xd7 16 ctJxd7 �xd7 17 dxc5
�xd1+ 1 8 �xd1 �c8 19 .td2 as with
equality.
1 4 . . . .lig7 1 5 e4 fxe4 1 6 l:te 1 ! ?
An interesting idea. Lugovoi­
Sveshnikov, Novgorod Open 1995,
continued instead 16 ctJg5 0-0 17 �xd7
�xd7 18 ctJxe4 c4! 19 �e 1 llb5 20
�h6!? �xh6 2 1 ctJf6+ 'it>g7 22 �d4
'it>g6 23 h4 �f4! 24 h5+ 'it>g7 25 ctJe8+
'it>g8 26 ctJf6+ 'it>h8 27 �xf4 �xf4 28
ctJxd7 �d8 29 dxe6 fxe6 and Black was
winning. I think he always had that
one under control!
1 6 . . . exd5 !
16 . . . 0-0 17 �xe4 exd5 1 8 iLxd5 with
�f4 to follow is very good for White
according to 011.
1 7 '1ifxd 5
17 �xd5 ctJf6! 1 8 �xe4 0-0 i s un­
clear according to 011.
1 7 . . . 0-0 1 8 '1ifxe4 l:tb6 1 9 .lta4 l:te6
20 i.f4 l:txe4 21 .1i.xc7 l:txe 1 + 22
l:txe 1 .ltxb2 23 i.d6 l:td8 24 !JLc7
l:tf8 25 i.d6 Yz - Yz
1 34
Neither side can avoid the repeti­
tIOn.
Game 105
Akesson-Ingbrandt
Stockholm (Rilton Cup) 1997
1 d4 d5 2 c4 c6 3 tiJf3 tiJf6 4 tiJc3
e6 5 e3 tiJbd7 6 !JLd3 dxc4 7 .ltxc4
b5 8 !JLd3 b4 9 tiJe4 c5 1 0 tiJxf6+
gxf6 1 1 e4! ? cxd4 1 2 tiJxd4 !JLc5 1 3
!JLe3! ?
This i s very risky. Black has not yet
committed his light-squared bishop,
which detracts from the strength of
�b5. 13 ctJb3 would have been safer.
1 3 . . JWb6! 1 4 i.c2!? .lia6
Stopping White from castling king­
side.
1 5 .lia4!
Now the fun begins! White's first
threat is 16 �xd7+ 'it>xd7 17 ctJb3+
. . .
wmnmg a plece.
1 5 . . . l:td8 1 6 l:tc 1 !
1 6 . . . 0-0
White threatened 17 �xc5 �xc5 18
ctJc6! �c4 19 ctJxd8 0-0 20 �g4+ \t>h8
21 iLh6! �g8 22 ctJxf7+ mate. 16 . . . b3!?
(intending . . . �b4+) is tempting, to
meet 17 l:lxc5, not with 17 ...�xc5
 M e r a n Va ria tio n : S ys t e m s w i t h a n Ea rly
when 1 8 tbc6 �c4 19 axb3! is strong,
but with 17 . . . bxa2!?
After 1 8 tbc2 'iix b2, I really don't
know what is going on!
1 7 i.xd7 l:txd7
Or 17 . . . b3!? 18 �xc5 'iixc5
(18 . . . bxa2!?) 19 'iVg4+ �h8 20 tbxe6!,
and now 20 . . :iYb4+ 2 1 .td2 fxe6!? (to
stop 'ti'g7+ mate; 2 1 . . :iixd2+!? 22 �d2
:xd7+ 23 \tie l seems better for White)
22 .txb4 bxa2. Here 23 'iWdl fails to
23 . . .lhd7 24 'iVaI �fd8, when 25
'iVxa2 �dl+ is mate, but 23 \t>d2
�xd7+ 24 �c3! (24 Wc2 .td3+ 25 �b3
.ib 1 ! ! followed by . . . a2-al'iV!)
24 . . . :c8+ 25 �b3 .tc4+ 26 Wa3 does
the trick for White.
1 8 l:txc5!
It took me a while to believe that
18 . . :iYxc5 loses to 19 'ii' g4+ Wh8 20
tbxe6! Black has no checks and cannot
deal with the double threat of 'i'g7+
mate and .ixc5.
1 8 . . . f5
Black's last chance was 18 ... b3.
1 9 exf5 �xc5 20 �g4+ �h8 21
tbxe6! �e5 22 tbxf8 l:td8 23 tbd7 !
1 -0
23 . . .lhd7 10ses to 24 f6!, threatening
'i'xd7 and 'iVg7+ mate.
. . . b 5-b4
Well it worked, but it certainly was
risky! There is still plenty of life in
this system for Black!
Game 106
Lautier-Piket
Leiden (match) 1995
1 d4 tbf6 2 c4 c6 3 tbc3 d 5 4 tbf3
e6 5 e3 tbbd7 6 i.d3 dxc4 7 i.xc4
b5 8 i.d3 b4 9 tbe4 i.e 7
This quiet move introduces the
main line of the 8 . . . b4 variation.
1 0 tbxf6+ tbxf6 1 1 e4 i.b7 1 2 "e2!
It is now time to discuss move or­
ders! Black players who play the main
line of this variation usually prefer to
reach it via 8 . . . .tb7 9 0-0 b4 10 tbe4
.te7 1 1 tbxf6+ tbxf6. The flexibility of
8 . . . b4 is of no use in the main line,
since Black wants his bishop on b7 in
all cases. Moreover, 8 . . . b4 9 tbe4 .te7
10 tbxf6+ tbxf6 1 1 e4 .ib7 gives White
the possibility of delaying castling
with 12 'ife2!
Question 3: Can't Black just castle
here?
Answer: 12 . . . 0-0 13 e5! tbd7 14 'iVe4!
(14 h4!?) 14 . . . g6 15 .th6 �e8 is not
pleasant for Black after 16 h4!? or the
1 35
Th e S e m i - Sla v
more sober 16 0-0. Black must there­
fore play 12 . . . ttJd7 to anticipate 13 eS.
However, after 13 0-0 0-0, Black has
lost all chance of playing the lines that
arise after 12 0-0 0-0 13 'iWe2 cst?
Question 4: How does the 8 ... .ltb7
move order help?
A nswer: After 8 . . . .1b7 9 0-0, White
has already castled. Consequently,
after 9 . . . b4 10 ttJe4 .1e7 1 1 ttJxf6+
ttJxf6 12 e4, he cannot play 'iiVe2 be­
fore Black castles, and after 12 . . . 0-0 13
'iiVe 2, Black can play the most active
lines with 13 ... cS.
1 2 . . . lLid7 1 3 e5?!
This, however, tries for too much.
13 0-0 was better - see the next game.
1 3 . . . c5!
Preparing castling by preventing
'iiVe 4.
1 4 dxc5 0-0 1 5 O-O? ! lLixc5 1 6 �c4
l:tcS 1 7 i.f4 lLia4!
A typical idea in this system. The
knight is impossible to remove from
a4, since b2-b3 would concede an out­
post on c3. White must live with the
annoying pressure against b2.
1 S l:tfd 1 �a5 1 9 l:td2 l:tfdS! 20
l:txdS+
The natural 20 llad 1 would have
1 36
lost a pawn to 20 ... ttJxb2! 2 1 ':xb2
llxdl+ 22 'iiVx dl ltxc4.
20 . . . l:txdS 2 1 l:tc 1 h6 22 h3 lLib6
Regrouping the knight now that it
has done its duty.
23 iob3 lLid5 24 iod2 'i'b6 25 l:tc2
a5 26 'i'c4 iofS 27 'i'g4 lLie7 ! 2S
i.e3 �b5 29 l:tc5 �e2 30 l:txa5 lLif5
31 �b6 l:tcS 32 Wh2 l:tc 1 33 ioe3
l:ta 1 34 lLid4 'i'f1 35 lLixf5 exf5 36
i.xf7+ WhS 0-1
A nice game from Piket.
Game 107
Levin-Antunes
Seville Open 1 994
1 d4 d5 2 c4 c6 3 lLic3 e6 4 lLif3
lLif6 5 e3 lLibd7 6 �d3 dxc4 7 �xc4
b5 S i.d3 b4 9 lLie4 �e7 1 0 lLixf6+
lLixf6 1 1 e4 iob7 1 2 'i'e2! lLid7 1 3
0-0 0-0
On d7, his knight is well-placed to
support ...c6-cS since d4xcS . . . ttJxcS
hits the bishop on d3 . However, Black
has less control over the dS-square, so
it is easier for White to meet ... c6-cS
with the central thrust d4-ds.
1 4 l:td 1
Very natural, but White could bor-
 M e r a n Va ria tio n : S ys t e m s with a n Ea rly
row from another variation and play
14 :e l ! ? to meet 14 . . . cS with 15 dS
exdS 16 exdS, hitting the bishop on e7.
For a more detailed discussion of this
idea, see the game Greenfeld-Khenkin
(Game 92) .
1 4 . . jj'e7 1 5 .i.e3 l:tae8 1 6 .i.a6?!
Very passive. 16 11ac 1 had to be bet­
ter.
1 6 . . . �b6 1 7 .i.xb7 'ii'x b7 1 8 l:tae 1
e5 1 9 dxe5 lLlxe5 20 e5 lLla4! 21
SiLg5 h6 22 SiLxe7 �xe7 23 h3 �b7
24 �d2 a5 25 a3 b3 26 'ii'd4 'it'b5
27 lLle 1 lLlb6 28 l:txe8 l:txe8 29 lLld3
lLld5 30 a4 l:te4 31 axb5 l:txd4 32
�f1 lLlf4 33 b6 lLlxd3 34 b7 l:tb4 35
l:txd3 l:txb 7 36 �e2 a4 37 l::t d 8+ �h 7
38 l:ta8 l:tb4 39 We3 g5 40 l:ta7 �g6
41 g3 h5 42 Wd3 h4 43 �e3 hxg3
44 fxg3 l:te4 45 �d3 l:txe5 46 l:txa4
l:te 1 47 l:tb4 l:tg 1 48 l::tx b3 l:txg3+ 49
�e4 l:txb3 0-1
The final two games deal with the
main line position that arises when
White has already committed himself
to early castling.
Game 108
Lautier-Piket
Monaco (match) 1996
1 d4 d5 2 e4 e6 3 lLlf3 lLlf6 4 lLle3
e6 5 e3 lLlbd7 6 SiLd3 dxe4 7 SiLxe4
b5 8 SiLd3 SiLb7 9 0-0 b4 1 0 lLle4
SiLe7 1 1 lLlxf6+ lLlxf6 1 2 e4 0-0 1 3
'it'e2 e5
see follo wing diagram
The most testing move. 13 . . . 4:Jd7
would have transposed to the previous
game.
. . . b 5-b4
14 dxe5!
This is inconvenient for Black, as
14 . . . .txcS? 15 eS! .txf3 (IS . . . 4:JdS 16
.txh7+!) 16 'i'xf3 4:JdS 17 'i'e4 g6 18
.th6 lIe8 19 .tbS :e7 20 .tgS! wins
the exchange.
1 4 . . . l:te8 ! ?
14 . . . 4:Jd7!?, t o recapture o n c S with
the knight, was slightly better for
White in Stohl-Novikov, Ostrava
1995, after 15 c6! .txc6 16 .te3 .tb7
17 l1ac 1 'iiaS 18 .tbS (18 4:Jd4!?)
18 . . . l:tad8 19 4:Jd4! (19 . . . .txe4 with 20
.tc6!) . This is nothing huge for White
but he is just a touch better.
1 5 SiLd2!?
This was improvised at the board
and is quite sneaky. Serper's sugges­
tion of 15 :dl !txcS 16 .te3 l:taS 17
4:Jd2! looks good, so I would love to
know what Piket had in mind!
1 5 . . . SiLxe5?
This falls into the same trap that
Black had previously avoided! IS . . . aS
16 a3 (16 !tac1!?) 16 . . . bxa3 17 :xa3
�xcS is suggested as unclear by Piket.
1 6 e5 SiLxf3 1 7 'it'xf3 lLld5 1 8 'it'e4
g6 1 9 SiLh6 l:te8 20 SiLb5 l:te7 2 1 SiLg5
�b6 22 .i.xe7 lLlxe7
Despite White's extra exchange, it is
137
Th e S e m i - Sla v
no easy matter to win this position
since Black is so solid.
23 .td7 .!:tdS 24 .!:tad 1 ttJf5 25 .ta4
.!:txd 1 26 i.xd 1 .td4 27 b3 �e5 2S
.tg4 ttJe7 29 i.f3 ttJd5 30 �h4 ttJe3
31 �dS+ '&t>g7 32 �f6+ '&t>gS 33
�dS+ % - %
Like the other seven games in this
match, the game was drawn!
Now the most aggressive try for
White, and the one favoured by most
of the top players: 13 eS.
Game 109
Beliavsky-Anand
\ Reggio Emilia 1 991
1 d4 d5 2 e4 e6 3 ttJe3 ttJf6 4 e3 e6
5 ttJf3 ttJbd7 6 .td3 dxe4 7 �xe4 b5
S i.d3 .tb 7 9 0-0 b4 1 0 ttJe4 .te 7
1 1 ttJxf6+ ttJxf6 1 2 e4 0-0 1 3 e5
ttJd7 14 i.e4
Preventing . . . c6-cS by pinning the
pawn to the bishop on b7.
1 4 . . . .!:tbS
1 38
Defending the bishop to allow ... c6-
cS.
1 5 �e2
The ECO recommendation of 1 5
�c2 i s rather mystifying here! How­
ever, 15 �e3 seems very sensible.
Jelen-Furlan, Bled 1992, continued
IS ... cS 16 �xb7 l:txb7 17 d5. White
had a slight edge after 17 . . . exd5 1 8
'i'xdS 'i'b6 19 �g5 �e8 2 0 �xe7 l:txe7
2 1 l:tac1 �c7 22 �fdl ttJf8 23 ttJd2.
1 5 . . . h6 1 6 i.e3 e5! 1 7 i.xb7 l:txb7
1 S dxe5
Gelfand has suggested that 18 'i'e4
'i'a8 19 l:tfdl is slightly better for
White here, but it's not that much.
Instead 1 8 l:tfdl 'i'c8 19 dxcS ttJxcS 20
Iiac1 �d8 ! 2 1 'i'c4 (2 1 l:txd8 iYxd8 22
�xc5 l:tc7! is equal according to Gel­
fand) 2 1 . . .l:txd 1 22 l:txd 1 as 23 l:tc1
�d7 was level in Karpov-Antunes Til­
burg 1994.
1 S . . . ttJxe5 1 9 .!:tfd 1 �bS 20 �e4
20 i.xcs l:tc8 21 i.xa7 l:txc2 22
.ixb8 llxb8 23 llab 1 11a8 24 ttJd4 Itcs
25 l:tal g5 gives Black good compensa­
tion for the pawn according to Anand,
who is a frequent advocate of the
Semi-Slav.
20 . . . ttJd7 21 �e4 .!:teS 22 .!:td2 .!:tbe7
23 .!:tad 1 .!:te4 24 .!:td4 ttJb6 25 �g4
'&t>fS 26 �e4 '&t>gS 27 �g4 .!:txd4 2S
.!:txd4 'Ot>fS 29 �h5 �e7 30 h4 i.e5
31 .!:td2 .txe3 32 fxe3 ttJd5 33 'Ot>f2
'iVe5 34 ttJd4 ttJf6 35 �f3 �xe5 36
.!:td 1 'Ot>gS 37 �f4 'iVd5 3S .!:ta 1 e5 39
�f5 l:te4 40 b3 exd4 0-1
 M e r a n Va ria tio n : S ys t e m s w i t h a n E a r l y . . . b 5-b4

Summary
These systems are still quite fresh and unexplored and they could well prove to
be a nasty surprise for an unprepared White player.

1d4 d5 2 c4 c6 3 lbf3 lbf6 4 lbc3 e6 5 e3 lbbd7 6 �d3 dxc4 7 �xc4 b5 8

�d3

8 �b7
. . .

8 . . . b4 9 ltJe4
9 . . . ltJxe4 Game 103
-

9 . . . c5 10 ltJxf6+ gxf6 (D)

1 1 i.. e4 Game 104
-

1 1 e4 Game 105
-

9 . . . i.. e 7 10 ltJxf6+ ltJxf6 1 1 e4 i.. b 7 12 'i'e2 ltJd7 (D)

13 e5 Game 106
-

13 0-0 Game 107

-

9 0-0 b4 1 0 lbe4 lbxe4

10 . . . i.e7 1 1 ltJxf6+ ltJxf6 12 e4 0-0 (D)
1 3 'ii'e2 Game 108
-

13 e5 Game 109
-

1 1 �xe4 Game 102

-

1 0. . . gxf6 1 2 . . lbd7
. 12 . . . 0-0

139
 CHA PTER TWEL VE

M e ran Variation :
Odds and Ends

1 d 4 d5 2 c 4 c 6 3 liJf3 liJf6 4 liJc3 The simplest. 1 1 .. .cxb2 12 fxg7

e6 5 e3 liJbd7 6 �d3
In this chapter, we take a brief look
at some of the unusual variations of
the Meran. The first three games with
White's other bishop retreats after
6 . . . dxc4 7 ..ixc4 bS, i.e. 8 .ie2 (Games
bxa1'iV 13 gxh8'iV looks fun, but actu­
ally it is just horrible for Black!
1 2 bxc3 �d6 1 3 �a4!?
13 ltJd2!? has not been scoring well
recently. Tunik-Savchenko, St Peters­
burg 1996, continued 13 . . 0-0 14 ltJc4
.

1 10 and 1 1 1) and 8 �b3 (Game 1 12) . .ie7 1 S ltJeS 'iVc7 16 .iO .ib7 17 �b 1
We then move on to a discussion of �a7 18 0-0 c5! 19 ..if4 'iVc8 20 dxcs
Black's solid 6 . . . ..id6 (Games 1 13-1 16) . ..txcS and Black was fine. 13 0-0 is
considered in the next game.
Game 1 1 0
Aleksandrov-Yagupov
. 'Russia 1996

1 d4 d 5 2 c4 c6 3 liJc3 liJf6 4 e3 e6
5 liJf3 liJbd7 6 �d3 dxc4 7 i.xc4 b5
8 �e2 a6 9 e4! ?
Piket's 9 0-0 worked successfully
against Kaidanov in Groningen 1993
after 9 ... cS 10 ds exdS 1 1 ltJxds ..ib7
12 ltJxf6+ 'iVxf6 13 a4 b4 14 e4 h6 15
.ic4 ltJb6 16 eS 'iVg6 17 �d3 with a
slight initiative for White. The idea of 1 3 . . . i.d7 1 4 liJe5 c5! 1 5 liJxd7
9 0-0 is to avoid the lines in the game, �xd7 1 6 �xd7+ �xd7 1 7 i.f3 l:tab8
meeting 9 . . . ..tb7 with 10 e4 and 1 1 eS. 1 8 dxc5 �xc5 1 9 �f4 l:tb6 20 0-0
9 . . . b4 1 0 e5 bxc3 1 1 exf6 liJxf6! liJd5!

1 40

This move equalises according to
Yagupov.
21 .l:!.fd 1 �d6 22 .ltxd6 .l:!.xd6 23 c4
�c3 24 .l:!.dc 1 .l:!.d3 25 .l:!.c2 .l:!.b8 26
'it>f1 ltJa4 27 c5 .l:!.d4 28 .l:!.e 1 'it>c7 29
g3 .l:!.b5 30 .l:!.e5 .l:!.b2 3 1 .l:!.ee2 .l:!.bb4
32 �g2 .l:!.bc4 33 c6 �c5 34 .l:!.b2
.l:!.c 1 35 l:tec2 �xc2 36 .l:!.xc2 �d3 37
'it>f1 �b4 38 .l:!.b2 .l:!.c4 39 �e2 .l:!.c 1 +
40 �g2 a 5 4 1 a3 �xc6 42 �a6 �d6
43 l:Ib7 �e5 44 �b5 .l:!.c7 45 l:Ib8 h5
46 a4 .l:!.c2 47 h3 .l:!.c5 48 .l:!.d8+ 'it>e7
49 .l:!.a8 �c4 50 h4 �d6 51 .l:!.a7+
�f6 52 l:txa5 .l:!.d5 53 f3 g6 54 �f2
'it>g7 55 �g2 e5 56 �f2 'it>h6 57 g4
f6 58 �g3 hxg4 59 fxg4 �e4+ 60
'it>f3 �c3 61 �e3 �xa4 Yz - Yz
Game 1 1 1
Gabriel-Slobodjan
Bad Homburg 1996
1 �f3 d 5 2 c4 c6 3 d4 �f6 4 �c3
e6 5 e3 �bd7 6 .i.d3 dxc4 7 �xc4
b5 8 .i.e2 a6 9 e4 b4 1 0 e5 bxc3 1 1
exf6 �xf6! 1 2 bxc3 1i.d6 1 3 0-0
1 3 . . :�i'c7 !
The most accurate, side-stepping the
pin with �g5 and preparing . . . c6-c5 as
quickly as possible.
M e r a n Va ria tio n : O dds a n d En ds
1 4 .i.d3 c5 1 5 .i.g5 �b7 1 6 .l:!.e 1
�d7 !
Preparing kingside castling when
Black will have no problems.
1 7 .i.e4 0-0 1 8 .i.xb7 'i'xb7 1 9 .l:!.b1
'i'c6 20 �e5 .i.xe5 21 dxe5 .l:!.fb8 22
.i.e7 .l:!.xb 1 23 'i'xb 1 .l:!.b8 24 'i'c 1
�xe5 25 .l:!.xe5 'i'c7 26 l:.xc5 'i'xe7
27 .l:!.c4 g6 28 h3 'i'd6 Yz - Yz
Game 1 12
I I incic-Kosic
Belgrade 1996
1 d4 d5 2 c4 c6 3 �f3 �f6 4 �c3
e6 5 e3 �bd7 6 .i.d3 dxc4 7 �xc4
b5 8 .i.b3
I tried this move once myself, but
didn't like it! The bishop is not well
placed here since dreams of sacrifices
on the a2-g8 diagonal are likely to re­
main just that.
8 . . . b4 9 �e2 �b7 1 0 �f4 �d6 1 1
�g5 �xf4! 1 2 exf4 0-0 1 3 0-0 c5
1 4 .i.e3?!
14 lie 1 was better, but White's posi-
. . . .
tlon is not impressive anyway.
1 4 . . . �a6 1 5 .l:!.e 1 c4!
Black now stands very well, but
loses it near the time control.
14 1
Th e S e m i - Sla v
1 6 iLe2 �b7 1 7 f5 exf5 1 8 iLxf5
lLlb6 1 9 �e2 g6 20 �h3 1;1e8 21
1;1ad 1 'iid 5 22 lLlf3 .lie6 23 b3 e3 24
lLle5 �b7 25 a3 a5 26 axb4 axb4 27
lLld3 'iVb5 28 lLle5 iLd5 29 i.g5
1;1xe 1 + 30 1;1xe 1 1;1e8 3 1 �e 1 lLle4 32
lLlxe4 1;1xe4 33 1;1xe4 iLxe4 34 'iVf4
�e8 35 'iVd6 lLld5 36 �d7 'iVa8 37
f3 �d3 38 h4 e2 39 i.e6 �a 1 + 40
�h2 e 1 'iV 4 1 i.xe 1 �xe 1 42 i.xd5
�e 1 43 �b8+ �g7 44 iVe5+ �xe5+
45 dxe5 f6 46 f4 fxe5 47 fxe5 g5
48 �g3 �g6 49 �f3 �f5 50 e6 iLb5
51 hxg5 �xg5 52 �e4 1 -0
!, Carne 1l3
Lalic-Nogueiras"
. Moscow Olympiad 1994'1
1 d4 d 5 2 e4 e6 3 lLle3 e6 4 e3 lLlf6
5 lLlf3 lLlbd7 6 i.d3 i.d6
7 e4! ?
Critical. A n early e3-e4 forces Black
to play lines involving the capture on
e4 and denies him the variations aris­
ing from 7 0-0 0-0 8 e4 dxc4 9 .i.xc4 e5
(Game 1 16) , as 7 e4 dxc4 8 .i.xc4 e5 9
dxe5 lbxe5 10 lbxe5 .i.xe5 1 1 'i¥xd8
'it>xd8 12 .i.xf7 is good for White.
7 . . . dxe4 8 lLlxe4 lLlxe4 9 �xe4 e5!?
1 42
9 ... .i.b4+ is seen in the next game
and 9 ... h6! in Game 1 15.
1 0 0 - 0 �e7 1 1 iLe2! b6? !
1 1 . . .0-0 was safer, but then perhaps
12 dxc5 to meet 12 . . . lbxc5 with 13 b4
and 14 c5, gaining queenside space.
1 2 d5! e5 1 3 1;1e 1 0-0 1 4 lLlg5 lLlf6
1 5 iVd3 g6 1 6 �f3 'iVe 7 1 7 �e3!
Fine play by Lalic: the black posi­
tion is now ripe to be opened by f2-f4
1 7 . . . lLlh5 1 8 f4 f6 1 9 fxe5 fxe5 20
lLle4 iLf5 21 �g5 iVd7 22 i.a4 �xa4
23 lLlxd6 e4 24 i.e7 1;1fe8 25 lLlxe8
1;1xe8 26 d6 lLlf4 27 'iVe5 lLld3 28
'iVd5+ �g7 29 1;1xe4 'i'e2 30 1;1f1 1 -0
Came 114.,
lII e scas-Prie "
Linares Zonal 19�h
1 d4 d5 2 e4 e6 3 lLlf3 lLlf6 4 lLle3
e6 5 e3 lLlbd7 6 i.d3 i.d6 7 e4! ?
dxe4 8 lLlxe4 lLlxe4 9 �xe4 �b4+
1 0 �d2 i.xd2+ 1 1 �xd2 e5!?
This may be too simplistic.
1 2 0-0-0 �e7 1 3 d5 exd5 14 exd5
'iVd6 1 5 i.e2 0-0 1 6 1;1he 1 lLlf6 1 7
lLle5!
This is nice for White: he has a large
spatial advantage and a passed d-pawn.

1 7 . . . �g4 1 8 f3 �h5 1 9 g4 �g6 20
�xg6 hxg6 2 1 g 5 tbd7 22 tbc4 �a6
23 b3 tbb6 24 'it'a5 l:Ud8 25 'iYxa6
bxa6 26 d6 �f8 27 tbe5 .:tac8 28
.:te4 tbd7 29 .:th4 'it>g8 30 f4 tbf8 3 1
d 7 .:tc7 32 l:d6 tbxd7 3 3 l:th3 f 6 34
gxf6 gxf6 35 tbxg6 �g7 36 tbe7
.:te8 37 tbd5 .:tb7 38 l:te3 .:txe3 39
tbxe3 tbb6 40 �d2 �g6 41 'it>e2 .:th 7
42 tbf1 l:th3 43 �f2 a5 44 l:tc6 a4
45 bxa4 .:ta3 46 .:txc5 tbxa4 � - �
Game 115
S herbakov-Shabanov
Russian Ch., Elista 1996
1 d4 d5 2 c4 c6 3 tbf3 tbf6 4 tbc3 e6
5 e3 tbbd7 6 �d3 �d6 7 e4 dxe4 8
tbxe4 tbxe4 9 �xe4 h6! 1 0 0-0 0-0
A key position. Black will play . . . e6-
eS to liquidate White's d4-pawn. The
prophylactic . . . h7-h6 was necessary as
9 . . . 0-0 10 0-0 eS? would have lost a
pawn to 1 1 dxeS liJxeS 12 liJxeS .txeS
13 .txh7+ Wxh7 14 li'hS+.
1 1 �c2 e5 1 2 .:te l �b4 1 3 �d2
�xd2 1 4 �xd2 exd4 1 5 -.xd4 'it'b6
1 6 �c3 a5 1 7 l:ad l tbf6 1 8 .:td6
'iYb4 1 9 'ii'e 5! 'iYxc4 20 �d3 "g4 21
h3 ""'h5 22 ""'g3! !
M e r a n Va ria tio n : O dds a n d En ds
Black is powerless against the twin
threats of ':xf6 and ':eS.
22 . . . tbd5 23 .:te5 f5 24 .:tg6 l:tf7 25
l:te8+ �h 7 26 tbe5 .:tc 7 27 f4 1 -0
Game 1 16
Lautier-Anand
London (rapidplay) 1995
1 d4 tbf6 2 c4 c6 3 tbc3 d5 4 e3 e6
5 tbf3 tbbd7 6 �d3 �d6 7 0-0 0-0 8
e4 dxc4 9 �xc4 e5
10 �g5 �e7 1 1 l:tel ! ?
1 1 dS is also possible, meetmg
1 1 . . .liJb6 with 12 .tb3.
1 1 . . . exd4 1 2 tbxd4?
This leads to disaster. White must
play 12 eS! liJxeS 13 liJe4. For the two
pawns, White has pins on the e-file
and the h4-d8 diagonal. Natural moves
all fail: 13 . . . .te6 loses to 14 liJxeS
.txeS 15 f4!; while 13 . . . .tfS 14 liJxd6!
liJxf3+ 15 'i'xf3 'i'xd6 16 'i'xfs and
13 . . . .tg4 14 'ii'xd4! liJxf3+ 15 gxf3 .teS
16 liJxf6+ gxf6 17 l:[xeS! fare no better.
1 2 . . . tbe5 1 3 �f1 �c5!
Suddenly f2 is looking really weak.
1 4 tba4 �g4 1 5 'it'd2 �b4 1 6 tbc3
l:ad8! 1 7 �e3 c5 1 8 tbf5 �xf5 1 9
""'c2 �g6 0-1
143
Th e S e m i- Sla v

Sum mary
S .te2 may be worth an occasional try as a surprise weapon, while Black players
in a solid mood may wish to give 6 ... �d6 a whirl.

1 d4 d5 2 c4 c6 3 lLlf3 lLlf6 4 lLlc3 e6 5 e3 lLlbd7 6 �d3

6 . . . dxc4
6 . . . i.d6
7 e4 dxe4 S liJxe4 liJxe4 9 �xe4 (D)
. . 9 . cS Game 1 13
-

. . 9 . .tb4+ Game 114

-

9 . . . h6 Game 115
-

7 0-0 0-0 S e4 dxc4 9 .txc4 Game 116

-

7 J.. x c4 b5 (D) 8 �e2

S �b3 Game 112
-

8 . . . a6 9 e4 b4 1 0 e5 bxc3 1 1 exf6 lLlxf6 1 2 bxc3 �d6 (D) 1 3 'iVa4

1 3 0-0 Game 1 1 1
-

1 3 ... �d7 Game 110

-

9 �xe4 7 . . . b5

1 44
 CHA PTER THIRTEEN

6 'iVc 2 ..td 6 :
7 ..te2 and 7 ..td3

1 d 4 d 5 2 c 4 cS 3 tiJf3 tiJfS 4 tiJc3 5 tiJf3 tiJgfS S �c2 .lidS 7 .lie2 0-0 8

eS 5 e3 tiJbd7 S 'iVc2 .lidS 0-0 .l:te8
6 .ltd3 is a straightforward move
with clear aims. By contrast, 6 'i!Vc2 is
a shadowy 'half-move' whose aims are
linked as much with Black's develop­
ment plans as White's own.
By delaying committing the light­
squared bishop, White discourages the
immediate 6 . . . dxc4, as after 7 .ltxc4 bS,
White has the useful extra move 'i!Vc2
in comparison with the 6 �d3 line.
6 . . . .ltd6 is the normal response.
This develops another piece, enabling
. . . 0-0, while supporting . . . e6-eS. Question 1: What is the idea here?
Now White faces a crucial choice Answer: White's basic aim remains
between two main moves, 7 �e2 and the same in all lines after S e3: he
7 .ltd3 (the less common 7 g4 and 7 b3 wants to push e3-e4 to gain central
are discussed in the (next chapter) . We space and free his dark-squared bishop.
shall first examine how Black has been S . . .l::t e S dissuades 9 e4 as 9 . ctJxe4 10
. .

neutralising 7 .lte2. ctJxe4 dxe4 11 'i'xe4 eS! is a good ri­

poste for Black.
Game 1 1 7 Now that Black has stopped the
J o . Horvath-Bareev early e3-e4, he will seek to solve his
Vienna Open 1 996 only positional problem: his inactive
light-squared bishop on cS . He can do
1 d4 d 5 2 c4 eS 3 tiJc3 cS 4 e3 tiJd7 this in two ways:

1 45
Th e S e m i - Sla v
1) He can release the central tension
with . . . d5xc4 and then play . . . e6-e5,
opening the c8-h3 diagonal for his
bishop. The rook on e8 supports the
e-pawn's advance to e4, attacking
White's knight on f3 .
2) He can fianchetto the bishop
with . . . b7-b6 and . . . �b7, and then
open the long diagonal with . . . c6-c5.
Note that 8 ... dxc4 9 .txc4 trans­
poses to the line 7 �d3 0-0 8 0-0 dxc4
9 �xc4 considered in Games 120-124.
9 l:td 1
By putting the rook on the d-file
opposite the queen on d8, White dis­
suades Black from playing . . . d5xc4 and
. . . e6-e5.
9 . . . 'i'e7
$0 Black removes his queen from
the d-file and supports . . . e6-e5 again.
1 0 a4
Question 2: What is White trying to
do here?
A nswer: White cannot play 10 e4
because of 10 . . . lbxe4 1 1 lbxe4 dxe4 12
'iVxe4 e5, so what is he to do? Think
of what I said earlier: that this system
often seems more concerned with an­
ticipating Black's development than
furthering White's own.
146
Question 3: Too subtle for me!
A nswer: White has spotted that
when Black tries to develop his bishop
with 10 . . . b6, then 1 1 e4! is possible as
1 1 ...dxe4 12 lbxe4 lbxe4 13 'Wxe4 e5 is
no longer a solution as c6 is hanging!
Question 4: Can't Black just play
13 ... �b7 with good chances?
Answer: White would have an edge
here as Karpov proved against Kam­
sky in their match in Elista 1996. Af­
ter 14 .tf4! .txf4 15 'Wxf4 c5 16 'Wc7!
!:tab8 17 b4! we can see that a2-a3 also
helps support this space-gaining push.
Karpov assesses the position after
17 . . . 1;!ec8 18 'Wf4 i.xf3 19 i.xf3 cxd4
(19 . . . cxb4 20 axb4 !:txc4 2 1 1lxa7 !:td8
22 b5 'Wb4 23 i.c6 lbf6 24 'iKc7 is
clearly better for White according to
Karpov. Fine preparation!) 20 'Wxd4 as
slightly better for White.
Question 5: Right, so Black should
break with 10 ... dxc4 1 1 i.xc4 e5 then?
A nswer: Wrong! Then 12 lbg5! is
annoying. But now you can guess why
Black plays his next move!
1 0 . . . h6! 1 1 h3
Question 6: Not again! Why?
Answer: Again White is anticipating
Black's plan: after ... d5xc4 and . . . e6-e5,
. . . e5-e4 unleashes the attack of the
black dark-squared bishop on the h2-
pawn. 1 1 h3 protects White's kingside
and so draws the attacking potential
from Black's plan. Remember that the
inclusion of h2-h3 and . . . h7-h6 does
not help Black to play . . . b7-b6!
1 1 . . . dxc4!
No more subtlety: Black goes for
his plan!
1 2 �xc4 e5 1 3 lbh4!
A typical idea. White tries to exploit

the weakened kingside light squares.
1 3 . . . liJf8 !
Preventing ltJg6 and preparing to
take the white knight if it lands on f5.
1 4 dxe5
14 ltJf5 iL.xf5 15 'iYxf5 e4! leaves the
white queen a little uncomfonable.
1 4 . . JWxe5 ! 1 5 liJf3 'ile7
The position is about equal.
1 6 �d2 �d7 1 7 l:tac 1 l:tad8 1 8 e4
liJg6 1 9 �e3 �b8 20 l:td2 �c8 2 1
l:txd8 l:txd8 22 l:td 1 l:te8 2 3 �f 1 b6
24 b4 Wf8 25 liJd4 �e5 26 g3 'iWh5
27 "e2 "xe2 28 �xe2 �d7 29 �g2
�c7 30 f3 liJe7 31 �f2 g5 32 l:tc 1
�e5 33 �d3 l:td8 34 liJce2 liJh5 35
�a6 �e8 36 l:td 1 l:td6 37 l:tc1 �d7
38 �c4 f5 39 exf5 liJxf5 40 liJxf5
�xf5 0- 1
This is Black's most reliable equal­
iser and has contributed to a loss of
faith in 7 iL.e2. 50 what about 7 i.d3 .
How does this help White?
Game 1 1 8
Dautov-Shirov
German Bundesliga 1996
1 d4 d 5 2 c4 c6 3 liJf3 liJf6 4 liJc3
e6 5 e3 liJbd7 6 'ilc2 �d6 7 �d3
6 'W c 2 iL d 6 : 7 iL e 2 a n d 7 iL d 3
0-0 8 0-0 �e7
The first point of 7 i.d3 is that the
fianchetto of the light-squared bishop
is not easy to achieve as 8 ... b6 9 e4
dxe4 10 ltJxe4 ltJxe4? loses a pawn to
1 1 i.xe4, forking the h7- and c6-
pawns. The text aims for a similar
build-up to the previous game.
However, 8 . . . h6!? is interesting,
simply removing the h-pawn from the
attack of the queen and bishop. In the
game 5praggett-Bacrot, Enghien 1997,
Black already stood well after 9 :dl
'iYe7 10 c5!? i.b8!? 1 1 e4 e5 12 cxd5
ltJxd5 13 iL.fl ltJxc3 14 bxc3 e4. An­
other try is 8 ... �e8 to meet 9 e4 with
9 ... dxc4 10 i.xc4 e5!? as 1 1 ltJg5 lU8
does not seem to lead anywhere.
Black's other major choices in this
position, 8 . . . e5 and 8 . . . dxc4, are con­
sidered in Games 1 19-124.
9 c5!
This, together with the next move,
gives the 7 i.d3 line its venom.
9 . . . �c7 1 0 e4 dxe4 1 1 liJxe4
The point of White's play is to re­
strict Black's choices and hinder his
development. This plan stops both of
Black's central breaks. By occupying
the e4-square, White prevents . . . c6-c5
14 7
Th e S e m i- Sla v
break. Unfortunately for Black, . . . e6-
e5 is also impossible: 10 . . . e5 1 1 exd5
cxd5 12 i.g5! e4 13 lDxd5! lDxd5 14
�xe7 exd3 15 �xd3 lDxe7 16 d5! was
a disaster in Dautov-Ribli, German
Bundesliga 1996, while 1 1 . . .e5 12
lDxf6+ lDxf6 (12 ... �xf6+ 13 .1xh7+)
1 3 dxe5 �xe5 14 lte 1 ! is also awful.
Question 6: Black does get the d5-
square though!
A nswer: That is true, but it is just
one square for one piece. It cannot be
used to launch an attack or to free
Black's game. The gains White makes
far outweigh the concession of d5.
1 1 . . . h6 1 2 l:te 1
Making sure that Black cannot stage
a breakout with . . . e6-e5.
1 2 . . . l:td8 1 3 a3 lbxe4 14 �xe4 lbf6
1 5 �d3 �d7 1 6 b4 .te8 1 7 �e2
lbd5 1 8 "e4! lbf6 1 9 �h4 l:td5 20
g4 �d8 21 g5 lbh7 22 �e4 lbxg5
23 lbxg5 l:txg5+ 24 oltxg5 "xg5+ 25
�xg5 hxg5 26 b5?
According to Dautov, 26 �ad 1 ! ltd8
27 lId3 �f6 28 ltad 1 ! a6 29 f4! would
have consolidated the white position,
leaving him with a clear advantage.
26 . . . �f6 % - Y2
So how should Black react to this
new threat? He has several options.
The most common is to play 8 . . . dxc4
9 �xc4, transposing back into the less
highly regarded lines of the .1e2 com­
plex (Games 120-124) . However, first
we take a look at an independent line.
Game 1 1 9
Karpov-Krainnik
, Las Palmas 1996
1 d4 d 5 2 c4 c6 3 lbf3 lbf6 4 lbc3
148
e6 5 e3 lbbd7 6 �c2 �d6 7 oltd3
0-0 8 0-0 e5
Question 7: Why does Black play
8 . . . e5?
A nswer: After this move, White can
give Black an isolated queen's pawn
with 9 cxd5 cxd5 10 dxe5 lDxe5 1 1
lDxe5 .1xe5. In return for this conces­
sion, Black frees his position and acti­
vates his pieces: he has opened the c8-
h3 diagonal for his bishop. Moreover,
White's kingside is rather short of de­
fensive pieces as both his light-squared
bishop and queen are in attacking po­
sitions on the b 1-h7 diagonal.
9 cxd5! cxd5 1 0 e4! exd4! ?
The sharpest move. Kramnik tried
10 . . . dxe4 1 1 lDxe4 lDxe4 12 .txe4 h6
against Karpov in Vienna 1996, but
after 13 i.e3! exd4 14 �h7+! 'i£i>h8 15
.i.xd4 lDf6 16 �f5! �xf5 17 �xf5,
White had some unpleasant pressure.
1 1 lbxd5 lbxd5 1 2 exd5 h6 1 3 lbxd4
�h4 1 4 lbf3 �h5 1 5 �h7+ �h8 1 6
�f5!
A very bold idea that was first
played in I.Sokolov-Piket, Nussloch
1996. In that game Black replied with
16 . . . �xf5 17 .txf5 lDf6, but after 1 8
�c2! �g4 ( 1 8 . . . lDxd5 19 litd1 lDb4 20

i.b3 i.c7 2 1 i.e3 lLlc6 22 .1cS! wins a
pawn as Sokolov shows) 19 i.b3,
White simply had an extra pawn in
the ending.
1 6 . . . g5! !
An outrageous reply that defies cal­
culation. Suddenly, White has prob­
lems with his bishop on h7 as the
threat is . . . �g7 and . . . lLlf6!
1 7 h4
17 i.e3!? is met by 17 . . . �g7 1 8
i.d4+ f6 19 'iWe6 i.b8! with the threat
of . . . lLleS and . . . gS-g4.
1 7 . . . lLlb6 1 8 "Wi'f6+ �xh7 1 9 �xd6
-tg4 20 lLlh2 l::t a d8 21 �b4 -tf5 22
hxg5 lLlxd5 23 �xb7 hxg5 24 �b3
l::t h 8 25 �f3 g4 26 "iVg3 lLlf6 27 f3
l::t d 3 28 lLlxg4 -txg4 29 fxg4 "iVg6?
The first mistake according to
Kramnik. 29 . . . "iVcS+ 30 'i'f2 'it'xf2+
was better, when the game Grabli­
auskas-Fridman, European Team
Championship, Pula 1997, finished 3 1
l:xf2 lLlxg4 3 2 lIf3 lIdl+ 3 3 l:tfl .l:tdS
34 b4 lh-1h .
30 �c7 'itg8 3 1 -tf4
3 1 gS would have been clearly bet­
ter for White according to Kramnik.
31 . . Jlh4 32 -tg3 l::t x g3 33 �xg3
�h6 34 �f3 l::t h 1 + 35 �f2 �d2+ 36
6 W c 2 iL d 6 : 7 iL e 2 a n d 7 iL d 3
�g3 �d6+ 37 �f2 �d2+ 38 �g3
�d6+ 39 �f2 Yz - Yz
We now turn our attention to the
lines after 8 . . . dxc4 9 i.xc4. The fol­
lowing game is a classic for this line
and shows the dangers that Black can
face.
Game 120
Karpov-Shirov
Biel 1992
1 d4 d5 2 c4 c6 3 lLlc3 lLlf6 4 e3 e6
5 lLlf3 lLlbd7 6 �c2 -td6 7 i.e2 0-0
8 0-0 dxc4 9 i.xc4 b5
With this move Black frees b7 for
the light-squared bishop with gain of
tempo. In typical Meran fashion he
will play his bishop to the long a8-h 1
diagonal and seek to quickly achieve
. . . c6-cS to activate it fully. This is the
most aggressive system against the 6
"iVc2 i.d6 7 i.d3 systems. The main
drawback of 9 . . . bS is that Black weak­
ens his queenside squares - cS in par­
ticular - a couple of moves before he
is ready to repair the damage with
. . . c6-cS. White's general strategical aim
must be to prevent . . . c6-cS in order to
keep the light-squared bishop passive
behind the c6-pawn. White has five
basic ways of achieving this:
1) To establish a knight on e4 via
lLlgS-e4, attacking cS.
2) To clamp down on cS by playing
b2-b4.
3) To keep attacking the bS-pawn,
meeting ... a7-a6 with a2-a4 so that
. . . c6-cS loses the bS-pawn.
4) To play b2-b3 and .1b2 so that
after . . . c6-cS, White can activate his
dark-squared bishop against the black
1 49
Th e S e m i - Sla v
king's position by d4xc5.
5) To play e3-e4, forcing ... e6-e5 and
creating central counterplay to distract
Black from achieving . . . c6-c5.
The solid 9 . . . 'i'e7 is considered in
Game 124.
1 0 ..ie2 1:e8
The more natural 10 ... .tb7 is the
subject of Games 12 1-123.
1 1 %:td 1 'iVc7 1 2 b3 e5 1 3 h3 i.b7 1 4
..ib2 a6 1 5 dxe5 lLlxe5 1 6 a4 %:tad8?
This natural move is a serious mis­
take. Arlandi-Illescas, Lisbon Zonal
1993, improved with 1 6 . . . l2Jg6! when
17 l2Jg5 can be met by 17 . . . .te5!, pre-
venting the white knights from com­
ing to e4.
1 7 lLlg5 ! 'V/ie7 1 8 lLlce4 lLlxe4 1 9
lLlxe4 i.b4 20 lLlg3!
A wonderful positional idea. The
teasing threat of l2Jf5 is quite irritating
and if Black anticipates it with
20 . . . 'i'e6 then 2 1 IhdS+ IhdS 22 'i'e4
is very unpleasant. Karpov sees things
that no one else can!
20 . . .f6 21 ..ixe5 �xe5 22 i.d3! h6
23 .i.g6!
White's advantage has been trans­
formed from a plus due to Black's
weak queenside dark squares to an
1 50
attack on Black's weak kingside light
squares! How did that happen?
23 . . . 1:f8 24 lLlf5 c5 25 axb5 axb5
26 %:ta7 'V/ic7 27 lLlh4 %:txd 1 + 28
�xd 1 %:ta8 29 �g4 �c6 30 %:txb7
�xb7 31 �e6+ 'it>h8 32 .i.e4 1 -0
A magnificent game.
Game 121
Gelfand-Kramnik
Dos Hermanas 1997
1 d4 lLlf6 2 c4 e6 3 lLlf3 d5 4 lLlc3
c6 5 e3 lLlbd7 6 �c2 ..id6 7 i.d3
0-0 8 0-0 dxc4 9 i.xc4 b5 1 0 i.e2
i.b7 1 1 1:d 1
The offbeat 1 1 .td2 is considered in
the next game and 1 1 a3 in Game 123,
while 11 e4 is simply met by 1 1 . . .e5.
1 1 . . . 'VJ#c7
Here too 12 .1d2 is possible, threat­
ening 13 b4 to clamp down on c5, as
13 . . . �xb4 14 l2Jxb5! is good for White.
1 2 b3
Black is ready to meet 12 e4 with
12 . . . e5.
1 2 . . . a6!
Equalising immediately according
to Gelfand, but 13 a4 seems the more
critical test here.

1 3 ttJe4 ttJxe4 1 4 �xe4 c5 1 5 �h4
ttJf6 1 6 �b2 'fie7 1 7 dxc5 Y2 - Y2
Game 122
Karpov-Gelfand
Dos Hermanas 1997
1 d4 d 5 2 c4 c6 3 ttJc3 ttJf6 4 e3 e6
5 ttJf3 ttJbd7 6 'iVc2 .)td6 7 �d3 0-0
8 0-0 dxc4 9 �xc4 b5 1 0 �e2 �b7
1 1 �d2 ! ?
A n unusual move order.
1 1 . . J�c8! ? 1 2 Itfd 1 b4 1 3 ttJe4
ttJxe4 1 4 'ii'x e4 'fie7 1 5 a3 bxa3 1 6
bxa3 ttJf6 1 7 'iVh4 c5
Black has equalised.
1 8 dxc5 �xc5 1 9 �b4 a5 20 oltxa5
�xa3 21 �d3 �c5 22 ttJe5 h6 23 h3
:ta8 24 �c3 ttJe4 25 'ii'x e7 �xe7 26
�d4 .)th4 27 g3 .)tf6 28 h4 Itxa 1 29
�xa 1 1:1a8 30 �d4 Ita2 31 Ita 1 Itd2
32 �xe4 �xe4 33 ttJc4 Itc2 34 ttJa3
l:c6 35 �xf6 gxf6 36 ttJb5 e5 37
ttJa7 Itc5 38 �h2 h5 39 g4 hxg4 40
�g3 f5 4 1 h 5 �h7 42 ::'a6 ::'c 1 43
�h4 Itf1 44 Itf6 Itxf2 45 ttJc8 �d5
46 ttJd6 nh2+ 47 'it>g5 g3 48 ttJxf5
�e6 49 ttJxg3 ::'g2 50 Wh4 Ith2+ 5 1
� g 5 Itg2 52 'it>h4 �g7 53 Itf1 Ith2+
54 �g5 Itg2 55 'it>h4 f6 56 ttJf5+
6 'fi c 2 i. d 6 : 7 i. e 2 a n d 7 i. d 3
�h7 57 e4 �d7 5 8 h6 Ite2 59 ttJd6
�g6 60 'it>g3 Ite3+ 61 Itf3 Ite 1 62
�g2 Itd 1 63 h7 �xh7 Y2 - Y2
Enough of all this subtlety; here is
one of my games!
Game 123
Grivas-Sadler
Cannes Open 1 995
1 d4 d5 2 c4 c6 3 ttJf3 ttJf6 4 ttJc3
e6 5 e3 ttJbd7 6 'fic2 �d6 7 �e2 0-0
8 0-0 dxc4 9 �xc4 b5 1 0 �e2 �b7
1 1 a3
By drawing the sting from . . . bS-b4,
White hopes to play e3-e4 and keep
his centre solid.
1 1 . . . Ite8 ! 1 2 e4 e5 1 3 �g5 h6 1 4
�h4 exd4 1 5 ttJxd4 'iVb8 !
16 �g3 is met by 16 . . . �xg3 17 hxg3
cS! when 18 'bdxbs a6! wins a piece,
as the white pawn on a3 prevents the
knight from retreating there!
1 6 ttJf3 a5 1 7 e5!? �xe5 1 8 ttJxe5
'iVxe5 1 9 �f3 b4 20 axb4 axb4 2 1
Itxa8 �xa8 2 2 ttJa4 c5 2 3 �xf6
�xf6 24 �xa8 Itxa8 25 b3 �c6 26
::'c1 Ite8 27 h3 Ite5 ! 28 'iVd2 1:1g5 29
f3 Itd5 30 'iVe3 lIe5 31 �f2 �d5 32
f4 Ite8 33 ttJxc5 ttJxc5 34 Itxc5
151
Th e S e m i - Sla v
�xb3 35 l:tb5 'iVc4 36 l:tc5 'iVd3 37
l:tc7 b3 38 l:tb7 l:te2 39 'iVb6 'iVg3
0- 1
We shall now move on to Black's
main alternative to 9 . . . bs: 9 . . . 'fIie7.
Game 124
Ruzele-Cifuentes
Groningen Open 1 996
1 d4 d 5 2 c4 c6 3 tbc3 tbf6 4 e3 e6
5 tbf3 tbbd7 6 'iVc2 .i.d6 7 i.e2 0-0
8 0-0 dxc4 9 iLxc4 �e7
9 . . . 'fIie7 is a flexible move. Although
its main purpose is to support . . . e6-eS
followed by a quick . . . eS-e4, Black
retains the option of using queenside
plans instead.
1 0 h3
This is the most critical test of
Black's plan.
Question 8: Why can't Black just
play 10 . . . eS?
Answer: Then 1 1 .1b3 is awkward
as 1 1 . . .e4 loses a pawn to 12 lbgS!,
while 1 1 . . . h6 12 lbh4!, intending lbg6
and lbfS, contains unpleasant threats.
1 0 . . . c5!? 1 1 dxc5 .i.xc5 1 2 e4 .i.d6
1 3 tbb5! tbe5 1 4 tbxe5 iLxe5 1 5 f4
�c5+ 1 6 ..t;h 1 ! ?
1 52
16 'it>h2 .1d7 17 b4!? was played in
Ruzele-Cifuentes, Groningen 1996,
and now 17 ... 'fIic8 18 lbxa7 �xa7 19
fxeS .1bS 20 .1e3 'fIixc4 2 1 'fIixc4 .1xc4
22 .1xa7 was slightly better for White
according to Ruzele.
1 6 . . . iLd7 1 7 b4 'iVc8 1 8 tbxa7 l:txa7
1 9 fxe5 iLb5 20 iLe3 'iVxc4 21 'iVxc4
i.xc4 22 iLxa7 iLxf1 23 l:txf1 tbxe4
24 l:tc1 h5 25 b5 l:td8 26 a4? !
26 'it>h2! first was better according
to Ruzele, when 26 .. J!d2 27 a4 h4 28
as gS 29 l:le 1 lbc3 30 a6 is very good
for White.
26 . . . h4! 27 \t>h2
27 l:lc4 �dl+ 28 'it>h2 l:ie l 29 as gS
30 a6 lbg3 is not very pleasant!
27 . . . g 5 28 l:te 1 tbc3 29 .i.e3 tbxa4
30 iLxg5 l:td4 3 1 l:tc1 l:tb4 32 iLf6
b6 33 l:tc8+ ..t;h7 34 l:th8+ ..t;g6 35
l:tg8+ \t>h6 36 l:th8+ ..t;g6 37 l:tg8+
\t>h6 38 l:tg7 tbc3 39 l:txf7 tbd5 40
l:tf8 tbxf6 41 l:txf6+ ..t;g7 42 l:txe6
l:txb5 43 l:te8 l:tb4 44 \t>g 1 l:tf4 45 e6
l:te4 46 e 7 ..t>f6 47 l:tb8 l:txe 7 48
l:txb6+ ..t;g5 49 l:tb5+ ..t;g6 50 ..t;f2
l:ta7 5 1 l:te5 l:tf7+ 52 \t>e3 l:tf6 53
l:te4 ..t>g5 54 l:tg4+ \t>h5 55 l:tf4 l:tg6
56 J:tf8 J:tg3+ 57 ..t;f4 \t>h6 58 J:th8+
..t;g7 59 l:[xh4 1:txg2 60 1:tg4+ Y2 - Y2
 6 � c 2 iL d 6 : 7 iL e 2 a n d 7 iL d 3

Sum mary
7 .i.d3 0-0 8 0-0 is the most critical test for Black after 6 �c2 .td6. If Bacrot's
8 . . . h6 does not fulfil its early promise then 8 . . .d.xc4 9 .i.xc4 bs is looking very
sound at the moment.

1 d4 d5 2 c4 c6 3 lUf3 lUf6 4 lUc3 e6 5 e3 lUbd7 6 'ii'c 2 .id6

7 .id3
7 .i.e2 0-0 8 0-0 (D)
8 . . J�e8 - Game 1 1 7
8 . . . d.xc4 9 bxc4 - Games 120-124 (by transposition)
7 0-0 8 0-0 dxc4
. . .

8 . . .'ife7 Game 118

-

8 . . . eS - Game 119
9 i.xc4 (D) b5
9 . . 'il.ie7 Game 124
. -

1 0 i.e2 l::t e 8
10 . . . .i.b7 (D)
1 1 :dl - Game 121
1 1 .td2 Game 122
-

1 1 a3 Game 123
-

1 1 l:td 1 - Game 120

8 0-0 9 i.xc4 10 . . . i.b 7

1 53
l_
CHA PTER FOURTEEN
Odds and Ends
1 d 4 d 5 2 c 4 c6 3 lbf3 lbf6 4 lbc3
e6
In this chapter, we briefly examine
lines I just couldn't fit in anywhere
else! The first two games deal with
White's rarer possibilities after 6 'iVc2
.id6 and the last two with 5 g3 and 5
'iVb3 respectively.
Game 125
+ Gelfand-Kramnik
European Club Cup, Berlin 1996
1 d4 d5 2 c4 c6 3 lbc3 lbf6 4 lbf3
e6 5 e3 lbbd7 6 �c2 .td6 7 94! ?
1 54
This amazing idea was the brain­
child of Alexander Shabalov. It is per­
haps surprising to see it relegated to a
tiny place in this book, but after this
game White is in desperate need of a
big improvement .
7 . . . �b4!
By pinning the white knight on c3,
Black provides a square on e4 for his
knight on f6 after White plays g4-g5.
Although Black loses a tempo with
. . . .ib4, he claims that g2-g4 is a greater
concession if White cannot launch an
immediate attack. Note that 7 . . . tLlxg4
is met by 8 :gl with awkward play
on the g-file.
8 �d2 "VIe7 9 a3 �xc3 1 0 iLxc3 b6!
1 1 �d3 �a6 !
Black prepares to exchange off one
of White's bishops and thus deprive
him of the two bishops.
1 2 "VIa4 dxc4 1 3 "VIxa6 cxd3 1 4
�xd3 0-0 1 5 9 5 lbd5 1 6 .Jid2 f5! 1 7
0-0-0 c 5 1 8 �b1 b5!
Despite the thrust g2-g4-g5, White
has no hint of a kingside attack. The
text sacrifices a pawn to open lines for

Black's pieces on the queenside.
1 9 �xbS :ab8 20 �aS l:tb3 2 1 'it>a2
l:tfb8 22 l:tb1 eS 23 l:the 1 �e6 24
'it>a 1 exd4 2S l:txeS lLlxeS 26 �xeS
lLle3 27 lLlxd4 l:txb2 28 l:txb2
28 .. :jja2+ 0-1
A very powerful game from Kram­
nik.
Game 126
Legky-M . Gurevich
Bruges 1995
1 d4 d S 2 e4 e6 3 lLlf3 lLlf6 4 lLle3
e6 S e3 lLlbd7 6 ft'e2 �d6 7 b3 0-0
8 �d3! ?
If White simply develops with 8
.i.e2, then after 8 . . . 'ife7 9 0-0 b6, we
are back into similar lines to the pre­
vious chapter where White has played
the passive b2-b3. The text is an at­
tempt to do something original.
8 . . . a6!
Preparing ... e6-eS by preventing
lbbS, which would be annoying after
8 . . . eS 9 cxdS cxdS.
9 0-0 eS! 1 0 eS!?
A risky attempt that turns out well
for Black. 10 cxdS cxds 1 1 dxeS lbxeS
was unclear according to Gurevich.
O dds a n d En ds
1 0 . . . �e 7 1 1 dxeS lLlxeS 1 2 lLlxeS
�xeS 1 3 f4 �e7 1 4 �b2 �e7 1 S
lLla4? ! �xe3+ 1 6 c.t>h 1 l:te8 1 7 �xf6
gxf6 1 8 :f3 �d4 1 9 �xh7+ ..t>f8 20
l:taf1 �g4 21 :3f2 :e3 22 �fS
�xfS 23 ft'xfS l:tae8 24 �hS fS 2S
lLlb6 �xb6 26 ft'h6+ �g7 27 �d6+
c.t>g8 28 exb6 tl.e 1 29 ft'd7 ft'f6 30
g4 l:t 1 e7 31 �xfS �xfS 32 gxfS c.t>g7
33 h4 l:th8 34 l:th2 ..t>f6 3S hS �xfS
36 h6 f6 37 l:tg 1 d4 38 l:thS+ ..t>xf4
39 l:tf1 + c.t>g4 40 l:th2 l:te3 41 l:tg2+
l:tg3 42 l:txf6 l:txg2 43 c.t>xg2 l:th7 44
l:td6 eS 4S c.t>f2 c.t>gS 46 l:tdS+ ..t>f4
47 l:thS ..t>g4 48 l:txeS l:txh6 49 �e2
l:txb6 SO ..t>d3 l:td6 'h - 'h
Game 127
T opalov-Kramnik
Linares 1997
1 d4 dS 2 e4 e6 3 lLlf3 lLlf6 4 lLle3
e6 S g3!?
This leads the game into the realms
of the Catalan. Kramnik states that
Black must capture on c4, otherwise
he will just stand worse.
S . . . lLlbd7 6 �g2 dxe4! 7 a4
Preventing . . . b7-bS.
7 . . . �e7 8 0-0 0-0 9 e4 eS! 1 0 dxeS
1 55
Th e S e m i - Sla v
tiJg4 1 1 �f4 'i¥aS 1 2 e6 fxe6 1 3
�e2 tiJgeS 1 4 tiJd4 tiJd3! ?
A novelty from Kramnik, giving up
the e6-pawn in order to activate his
light-squared bishop on c8 .
1 S tiJxe6 lU6 1 6 iLc 7 'i¥b4 1 7 tiJd4?
17 4Jg5 was better according to
Kramnik, when 17 .. :iWxb2 18 'iYxb2
4Jxb2 19 f4 with e4-e5 to follow leads
to an unclear ending.
1 7 . . . �cS 1 8 tiJc2 �xb2 1 9 tiJd 1
'i¥b3 20 tiJde3 tiJ7eS 2 1 h3 �e6 22
�h2 l:th6 23 l:tab 1 �g4! !
24 tiJxg4 tiJxg4+ 2S 'i¥xg4 'i¥xc2 26
�gS tiJxf2 27 l:txf2 'i¥xf2 28 l:tf1
'ii'd 4 29 �eS 'i¥d7 30 'i¥c 1 'i¥e6 3 1
.ltxg7 �xg7 3 2 'i¥gS+ l:tg6 3 3 'i¥xcS
'i¥d6 0-1
One of Karpov's favourite standby
ideas against the Semi-Slav has been 5
'iYb3 . However, Kasparov's treatment
in the next game seems very efficient.
,� Game 128
Karpov-Kasparov
Las Palmas 199�
1 d4 tiJf6 2 tiJf3 dS 3 c4 e6 4 tiJc3
e6 S 'ii' b 3 dxc4 6 �xc4
The actual move order in the game
1 56
was 4 . . . dxc4 5 'iYa4+ c6 6 'iYxc4, but I
have changed it to illustrate the Semi­
Slav sequence.
6. bS 7 'i¥d3 �b7
_ _
8 a3
8 il.g5 4Jbd7 9 e3 a6! 10 il.e2 c5 1 1
0-0 il.e7 was just equal in Karpov­
Timman, World Championship, Ja­
karta 1993.
8 . . . a6! 9 e3
After this, Black has no problems. 9
e4 c5 10 e5 is clearly better for White
according to Karpov, but I see no
problems for Black after 10 . . . cxd4 1 1
4Jxb5 4Jfd7!? 1 2 4Jbxd4 4Jxe5! 13
4Jxe5 'iYa5+.
9 . . . cS 1 0 dxcS SLXCS 1 1 �xd8+
�xd8 1 2 SLd2 �e 7 1 3 SLd3 tiJbd7 1 4
�e2 iLd6 1 S l:thd 1 l:tac8 1 6 l:tac 1
tiJb6 1 7 �e 1 tiJc4 1 8 l:tc2 �xf3+ 1 9
gxf3 tiJeS 20 h3 tiJxd3 2 1 l:txd3
l:thd8 22 l:tcd2 SLc7 23 l:tc2 SLb6 24
l:txd8 �xd8 2S l:td2+ �e7 26 l:td 1 g6
27 f4 l:tc4 28 f3 tiJd7 29 b3 l:tc6 30
tiJe4 l:tc2+ 31 l:td2 l:txd2+ 32 .i.xd2
�cS 33 tiJxcS tiJxcs 34 iLb4 �d6 3S
�d2 as 36 .txcS+ �XCS 37 �d3 f6
38 h4 �dS 39 b4 axb4 40 axb4 h6
41 e4+ �d6 42 �e3 eS 43 fxeS+
fxeS 44 �f2 �e6 4S �g2 Yz Yz -
 O dds a n d En ds

Sum mary
All these systems are in need of new ideas for White. For the moment, Black is
happy to face them!

1 d4 d5 2 c4 c6 3 ttJf3 ttJf6 4 ttJc3 e6 (D)

5 e3
5 g3 Game 127
-

5 'iWb3 Game 128

-

5 . . .ttJbd7 6 �c2 �d6 (D) 7 94

7 b3 Game 126
-

7 . . . �b4 (D) Game 125

-

4 . . . e6 7 . . . �b4

157
Akesson-Ingbrandt, Stockholm {R ifton Cup} 1997................................................... 134
Aleksandrov-Yagupov, Russia 1996 ......................................................................... 140
Alterman-Chernin, Groningen {PCA Qualifier} 1993 ............................................ 124
Alterman-Dreev, Manila Olympiad 1992 ................................................................ 1 13
Alvarez-Antunes, Mondariz Balneario 1996 .............................................................. 55
Atalik-Bacrot, Wijk aan Zee B 1997 ............................................................................ 78
Azmaiparashvili-Akopian, Yerevan Olympiad 1996 ............................................... 92
Bareev-Dreev, R ussia 1996 ......................................................................................... 121
Bareev-Dreev, R ussian Ch., Elista 1996 ...................................................................... 98
Bareev-Dreev, Wijk aan Zee {match} 1995 .................................................................. 73
Bareev-Filippov, Russia 1995 ....................................................................................... 35
Bareev-Kramnik, Dortmund 1995 ............................................................................ 1 1 7
Bareev-Yusupov, Linares 1993 .................................................................................. 128
Beliavsky-Anand, Reggio Emilia 1991 ...................................................................... 138
Beliavsky-Dreev, Novosibirsk 1995 ............................................................................. 73
Beliavsky-Illescas, Linares 1994 ................................................................................... 39
Bellon-Antunes, Platja d'A ro Barcino 1994 ................................................................ 59
Cebalo-Palac, Croatian Ch., Slavonski Brod 1995 ...................................................... 62
Csiszar-Sploshrov, Budapest Open 1996 ................................................................... 127
Dautov-Dreev, Reggio Emilia 1995 ............................................................................ 71
Dautov-Dreev, Yerevan Olympiad 1996 .................................................................... 72
Dautov-Shirov, German Bundesliga 1996 ................................................................ 147
De Sousa-Eliet.N, French Championship 1996 ........................................................ 126
Demirel-Fridman, European Junior Ch. 1992 ........................................................... 54
Ehlvest-Kharlov, Novosibirsk 1995 ............................................................................. 71
Gabriel-Slobodjan, Bad Homburg 1996 ................................................................... 141
Garcia.D-Kramnik, Pamplona 1992 ........................................................................... 57
Gavrilov-Novikov, R iga Open 1995 ......................................................................... 121
Gelfand-Akopian, Yerevan Olympiad 1996 ............................................................... 85
Gelfand-Dreev, Biel 1995 101
...........................................................................................

Gelfand-Dreev, Tilburg 1993 ....................................................................................... 93

Gelfand-Kramnik, Dos Hermanas 1997 ................................................................... 150

1 58
 In d e x o f C o mp le t e G a m e s

Gelfand-Kramnik, European Club Cup, Berlin 1996 .............................................. 154

Gelfand-Shirov, Biel 1995 90
............................................................................................

Gelfand-Shirov, Linares 1997 .................................................................................... 125

Greenfeld-Bykhovsky.Av, Beersheva 1996 ................................................................ 91
Greenfeld-Khenkin, Israel 1995 ................................................................................ 119
Grivas-Sadler, Cannes Open 1995 ............................................................................. 151
Hansen.Cu-Chernin, Taastrup 1992 ........................................................................ 107
Hansen.Cu-Tisdall, Reykjavik lonal 1995 ................................................................. 61
Hansen.L.B-Illescas, Moscow Olympiad 1994 .......................................................... 100
Horvath.Jo-Bareev, Vienna Open 1996 ................................................................... 145
Ilincic-Kosic, Belgrade 1996 ........................................................................................ 141
Illescas-Prie, Linares lonal 1995 ................................................................................ 142
Ionov-Popov, St Petersburg Open 1995 ....................................................................... 37
Ivanchuk-Illescas, Linares 1994 ................................................................................... 52
Ivanchuk-Kramnik, Novgorod 1996 ........................................................................... 67
Ivanchuk-Oll, Biel Interzonal 1993 ........................................................................... 133
Ivanchuk-Shirov, Novgorod 1994 ............................................................................... 24
Ivanchuk-Shirov, Wijk aan lee 1996 ......................................................................... 14
Kalantarian-Yegiazarian, A rmenian Championship 1994 ....................................... 46
Kallai-Lukacs, Budapest 1995 ........................................................................................ 59
Kamsky-Kramnik, Dos Hermanas 1996 ..................................................................... 29
Kamsky-Kramnik, New York (Candidates match) 1994 ............................................ 21
Kamsky-Shirov, Madrid 1994 ...................................................................................... 92
Karpov-Gelfand, Dos Hermanas 1997 ...................................................................... 151
Karpov-Kasparov, Las Palmas 1996 .......................................................................... 156
Karpov-Kramnik, Dortmund 1995 ........................................................................... 1 18
Karpov-Kramnik, Las Palmas 1996 ........................................................................... 148
Karpov-Kramnik, Linares 1994 ................................................................................... 88
Karpov-Shirov, Biel 1992 ........................................................................................... 149
Karpov-Shirov, Linares 1994 ..................................................................................... 120
Kasparov-Kramnik, New York (rapidplay) 1994 ........................................................ 27
Kasparov-Ivanchuk, Linares 1994 ............................................................................... 33
Kasparov-Kramnik, Dos Hermanas 1996 ................................................................... 94
Khalifman-Akopian, Yerevan 1996 ............................................................................ 65
Khalifman-Piket, A msterdam 1995 ............................................................................. 51
Khalifman-Shirov, Pardubice 1994 ............................................................................. 47
Kharitonov-Ivanchuk, USSR 1988 ........................................................................... 104
Knaak-Van der Wiel, Lugano 1989 ............................................................................. 48
Kozul-Beliavsky, Slovenia 1995 ................................................................................. 102
Kramnik-Ehlvest, R iga 1995 ........................................................................................ 51
Kramnik-Shirov, Monaco (blindfold) 1996 ................................................................. 28
Krasenkov-Oll, Polanica Zdroj 1996 ........................................................................... 87
Krasenkov-Schandorff, Copenhagen (Politiken Cup) 1996 ....................................... 86
Lalic-Arduman, European Team Ch., Pula 1997 ........................................................ 67
Lalic-Nogueiras, Moscow Olympiad 1994 ................................................................ 142
Lalic-Wilson.J, London 1996 ........................................................................................ 18
Lautier-Anand, London (rapidplay) 1995 .................................................................. 143
Lautier-Dreev, Linares 1995 ........................................................................................ 99

159
In d e x o f C o mp le t e G a m e s

Lautier-Gelfand, A msterdam 1996 .............................................................................. 83

Lautier-Piket, LeideJ(match) 1995 ............................................................................ 135
Lautier-Piket, Monaco (match) 1996 .......................................................................... 103
Lautier-Piket, Monaco (match) 1996 .......................................................................... 137
Legky-Gurevich.M, Bruges 1995 ............................................................................... 155
Levin-Antunes, Seville Open 1994 ............................................................................ 136
Mecking-Matsuura, Sao Paulo lonal 1995 ................................................................ 50
Mecking-San Segundo, Linares Open 1995 ............................................................... 47
Nikolic-Kramnik, Yerevan Olympiad 1996 ............................................................... 74
Nikolic-Shirov, Wijk aan lee (match) 1993 ................................................................ 22
Oll-Kaidanov, Kuibysev 1986 ....................................................................................... 49
Petursson-Dreev, Yerevan Olympiad 1996 ................................................................ 78
Piket-Dreev, Wijk aan lee 1996 .................................................................................. 74
Piket-Kramnik, A msterdam 1993 .............................................................................. 112
Piket-Kramnik, Linares 1997 ....................................................................................... 66
Piket-Novikov, Antwerp Open 1996 ........................................................................... 68
Piket-Shirov, A ruba (match) 1995 .............................................................................. 108
Podgaets-Muhametov, Moscow 1995 ....................................................................... 105
Pogorelov-Korneev, Benasque Open 1996 ................................................................. 58
Polugayevsky-Torre, Moscow 1981 ............................................................................. 38
Razuvaev-Filippov, Russian Championship 1995 ..................................................... 37
Ruzele-Cifuentes, Groningen Open 1996 ................................................................ 152
Sadler-Bisby, Isle ofMan Open 1995 .......................................................................... 131
Sadler-Madwekwe, London (Lloyds Bank) 1994 ........................................................ 94
San Segundo-Vera, Alcobendas 1994 ........................................................................ 129
Sherbakov-Shabanov, Russian Ch., Elista 1996 ...................................................... 143
Shirov-Morovic, Las Palmas 1994 ............................................................................... 44
Sokolov.I-Chernin, Wijk aan lee 1991 .................................................................... 115
Stean-Rivas, Marbella 1982 19
...........................................................................................

Stefansson-Inkiov, Gausdal 1990 ................................................................................ 55

Stefansson-Tisdall, Reykjavik lonal 1995 .................................................................. 60
Timman-Gelfand, Belgrade 1995 ................................................................................. 77
Timman-Gelfand, Yerevan Olympiad 1996 ............................................................... 70
Timman-Tal, Hilversum (match) 1988 ........................................................................ 41
Tkachiev-Handoko, Jak4rta (match) 1996 .................................................................. 95
Topalov-Gelfand, Dortmund 1996 ............................................................................. 79
Topalov-Kramnik, Dortmund 1996 ............................................................................ 25
Topalov-Kramnik, Linares 1997 ............................................................................... 155
Van Wely-Dreev, Bern Open 1993 .............................................................................. 31
Van Wely-Dreev, Wijk aan lee 1996 .......................................................................... 70
Van Wely-Gelfand, Tilburg 1996 ................................................................................ 75
Van Wely-Kramnik, Biel Interzonal 1993 .................................................................. 32
Wells-Kaidanov, Dublin 1991 .......... .......................................................................... 116
Yakovich-Giorgadze, Yerevan Open 1996 ............................................................... 111
Yakovich-Sveshnikov, Yerevan Open 1996 ............................................................. 132
Yermolinsky-Atalik, Hastings 1995 ............................................................................ 61
Yermolinsky-Kaidanov, USA Championship 1993 ................................................... 43
Yusupov-Kramnik, Horgen 1995 .............................................................................. 1 18

1 60