First published in 2013 by Gloucester Publishers Limited, Northburgh House,

10 Northburgh Street, London EC1V 0AT

Copyright © 2013 Sam Collins

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

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

retrieval system or transmitted in any form or by any means, electronic,
electrostatic, magnetic tape, photocopying, recording or otherwise,
without prior permission of the publisher.

British Library Cataloguing-in-Publication Data

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

ISBN Kindle: 978-1-78194-143-0

ISBN epub: 978-1-78194-172-0

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

All other sales enquiries should be directed to Everyman Chess, Northburgh House,
10 Northburgh Street, London EC1V 0AT
tel: 020 7253 7887 fax: 020 7490 3708
email: info@everymanchess.com; website: www.everymanchess.com

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

Everyman Chess Series

Chief advisor: Byron Jacobs
Commissioning editor: John Emms
Assistant editor: Richard Palliser

Typeset and edited by First Rank Publishing, Brighton.

Cover design by Horatio Monteverde.
About the Author
Sam Collins is an International Master with two Grandmaster norms, and a former Irish and Japanese
Champion. He has represented Ireland at seven Olympiads, winning an individual gold medal at Bled
2002. He has a wealth of teaching and writing experience, and has produced many books, DVDs and
magazine articles on chess.

Also by the Author

The French Advance
Gambit Busters
The Greatest Ever Chess Strategies
Contents
About the Author
Bibliography
Series Foreword

Introduction
Structural Introduction

1 9 Bg5 cxd4
2 9 Bg5 c4
3 9 Bg5 Be6
4 8 dxc5/9 dxc5
5 Other 8th/9th move options
6 Symmetrical Tarrasch
7 Tarrasch sidelines
8 Réti set-ups
9 In place of an epilogue ...

Index of Complete Games

Bibliography
Books
1 d4 Volume One, Boris Avrukh (Quality Chess 2008)
Chess Duels: My Games with the World Champions, Yasser Seirawan (Everyman 2010)
Garry Kasparov’s Greatest Chess Games Volume 1, Igor Stohl (Gambit 2005)
Garry Kasparov on My Great Predecessors Volumes 1-5, Garry Kasparov (Everyman 2003-06)
Garry Kasparov Part I: 1973-1985, Garry Kasparov (Everyman 2011)
I Play Against Pieces, Svetozar Gligoric (Batsford 2002)
Kasparov vs Karpov 1975-1985, Garry Kasparov (Everyman 2008)
My Best Games, Anatoly Karpov (Edition Olms 2007)
My Best Games Vol.2: Games with Black, Victor Korchnoi (Edition Olms 2001)
My Most Memorable Games, Boris Gelfand (Edition Olms 2005)
Opening Preparation, Mark Dvoretsky & Artur Yusupov (Batsford 1994)
Paul Keres: The Quest for Perfection, Paul Keres and John Nunn (Batsford 1997)
Revolutionize Your Chess, Viktor Moskalenko (New in Chess 2009)
Secrets of Grandmaster Chess, John Nunn (Batsford 1997)
The English Opening: Volume Two, Mihail Marin (Quality Chess 2010)
The Kaufman Repertoire for Black and White, Larry Kaufman (New in Chess 2012)
The Modern Réti, Alexander Delchev (Chess Stars 2012)
The Tarrasch Defence, Jacob Aagaard and Nikolaos Ntirlis (Quality Chess 2011)
Winning Chess Middlegames, Ivan Sokolov (New In Chess 2008)

Websites
www.chesspublishing.com
www.chessbase.com

Software/Databases
ChessBase 12
Megabase 2013

Magazines
The London Review of Books
Series Foreword
Move by Move is a series of opening books which uses a question-and-answer format. One of our main
aims of the series is to replicate – as much as possible – lessons between chess teachers and students.
All the way through, readers will be challenged to answer searching questions, to test their skills in chess
openings and indeed in other key aspects of the game. It’s our firm belief that practising your skills like
this is an excellent way to study chess openings, and to study chess in general.
Many thanks go to all those who have been kind enough to offer inspiration, advice and assistance in the
creation of Move by Move. We’re really excited by this series and hope that readers will share our
enthusiasm.

John Emms,
Everyman Chess
Introduction
First Thoughts
An excellent potted history of the Tarrasch is set out by one of its exponents in the 1970s, GM John Nunn:
“ ... the Tarrasch, a defence which has always been considered an uphill struggle for Black, but which
becomes fashionable for short periods when some great player attempts to revive and improve it. Spassky
used it for a while in the 1960s, and more recently Kasparov adopted it twice in his first match with
Karpov. Unfortunately, he lost on both occasions and then largely abandoned it. I played it occasionally in
the seventies, but after some poor results I gave it up. My experience indicated that Black’s winning
chances were slim since White could usually draw even if he played badly, while strong opponents
would generally win! Today [Nunn was writing in 1997 – SC] only a few players are willing to adopt it,
Murray Chandler being one of this vanishing breed.”

Dr. Nunn’s comments represent the consensus view and are certainly an accurate representation of
the historical development of the Tarrasch before 2000. They throw up a lot of interesting material for
discussion.

The Advantage of Playing White

On one analysis, White drawing if he plays badly and generally winning if he is a stronger player, is
something which happens in every opening. So what Nunn is describing isn’t so surprising. A look at the
statistics from Megabase 2013 (admittedly a rough and ready analysis since I haven’t corrected for rating)
bears this out. In the position after 1 d4 d5 2 c4 e6 3 Nc3, against 3 ... c5 White scores 54.8% from 8788
games (a lower percentage than he achieves after 3 ... Nf6 or 3 ... Be7). In the position after 1 d4 d5 2 c4
e6 3 Nf3 (while this signals the Catalan, it is also specifically an anti-Tarrasch move order since White
has more options when he hasn’t committed his knight to c3), White scores 60.6% from 6484 games
against 3 ... c5 (which is 0.1% less than he makes against 3 ... Nf6).
Looking at my own games provided interesting food for thought. I have played the Tarrasch against
all kinds of opponents, including some under 2000 and many strong GMs: Korchnoi, Khenkin, Baburin
(twice), Pert, Bischoff, San Segundo, Hebden, Burmakin, Socko. From what I can piece together, in 22
games with the line I scored 10 points. From a pure comparison of player’s ratings, my expected score
was 11 points. Given that White tends to score approximately 55% in most openings, this result is pretty
acceptable. The statistics are skewed in White’s favour by some other factors: first, if one looks at the
opening positions I got from those games, the results ought to have been even better (losses to Baburin,
Pert, Khenkin and Socko occurred in equal or very playable middlegames); second, in certain games (in
particular against Bischoff) I horribly mishandled the opening and genuinely played an awful game.
Accordingly, my own limited experience provides no reason to doubt the efficacy of the Tarrasch.

Fashion
Nunn is quite right to point to Kasparov’s two losses with the Tarrasch in his first World Championship
match against Karpov damaging the opening’s popularity in quite a permanent way.
Looking objectively at these two games, in the first (Game 7) Kasparov easily equalized and only
lost due to a late blunder. In the second (Game 9), against the best chess technician in history, Kasparov
blundered on move 46 and lost in 70 moves. To put it mildly, this was not a result of the opening. If
openings were shelved because arising endgames were lost against Karpov in the mid-80s, there
wouldn’t be a lot of openings played anymore.
It’s quite hard to find someone who justifiably dropped the Tarrasch. For instance, Alexander
Grischuk was the hero of this opening in the early 2000s (and laid the foundation for many of the modern
main lines, in particular through his excellent use of 9 Bg5 c4!). However, his loss to Gelfand (Russian
Team Championship 2004) seems to have been Grischuk’s last game with the Tarrasch at a classical time
control (admittedly he gave the line a spin at the Amber tournament in 2006, scoring ½/2 against Van
Wely and Ivanchuk). So did Gelfand show a good antidote against the Tarrasch? Actually, no. The game
followed the same first 13 moves as Sargissian-Halkias (covered later in this volume) where, as you will
see, 13 ... Nxd4! is a clean equalizer.
Moreover, let’s see the endgame which resulted in Gelfand-Grischuk:

How did Gelfand know how to exploit his advantage here? Well, let’s ask him!
“ ... I was very optimistic about the double-rook ending. In my first encounter with Anatoly Karpov
in Linares 1991 I had a similar ending, in which with Black I was defending a better pawn structure (my
pawn was on g6, not on f6) and where I was confident that we would soon agree a draw. But the 12th
world champion taught me a memorable lesson, allowing me no chances in such a harmless-looking
position. My hope of exchanging my a- and c-pawns for the a-pawn proved to be a mere illusion.”
Let’s have a look at the position Gelfand is referring to:
 And the opening in that game was a Grünfeld, which nobody (including Gelfand) ditched as a result
of the game.
Moreover, just how bad is this endgame? I’ve found the following examples of this or similar
endgames arising from the Tarrasch (many following the same line as in Gelfand-Grischuk):

V.Borovikov-W.Pereira Neto, Nettetal 2004 (½-½ in 60)

G.Pitl-E.Lie, Gausdal 2007 (½-½ in 38)
Wang Hao-S.Iuldachev, Asian Team Championship 2008 (1-0 in 64)
Zhou Jianchao-V.Akobian, Beijing (blitz) 2008 (½-½ in 66 – as Black is a rook up in the final position
I’m not sure about this result!)
Mi.Hoffmann-K.Kiik, European Cup, Ohrid 2009 (1-0 in 39)
A.Matthiesen-Kar.Rasmussen, Aarhus 2010 (½-½ in 61)
Zhu Chen-E.Kovalevskaya, FIDE Grand Prix, Nalchik 2011 (½-½ in 44)
I.Kragelj-T.Suc, Trieste 2012 (½-½ in 34)

So what looks like a forced win in the hands of Karpov or Gelfand starts looking pretty unclear at lower
levels (even “normal” GM level). Moreover, this endgame is close to the maximum White can hope for in
the 9 Bg5 c4 line (admittedly, as Gelfand observes, if White can keep the queens on Black may have some
additional problems to solve, as in S.Mamedyarov-V.Akobian, Gibraltar 2012) and ways to avoid it have
been found, including 13 ... Nxd4 in the Gelfand-Grischuk move order and other alternatives, such as
Aagaard and Ntirlis’ recommendation of an early ... h7-h6, not to mention the ninth move alternatives. So
let’s not throw our toys out of the pram just yet.
Looking at current representatives, it’s true to say that the Tarrasch doesn’t have too many regular
GM advocates (the patronage by Grischuk, T.L.Petrosian and Akobian appears to have cooled off
recently). That said, there are some players who have played it with success over the last few years.
Occasional GM practitioners with excellent results include Jacob Aagaard, Milos Perunovic, Ray
Robson and Julio Sadorra, while Vlad Jianu, Lucian Miron, Ivan Sokolov, Aaron Summerscale, Robin
Swinkels and Branko Tadic have all played it several times with good results. Alexander Berelowitsch,
Goran Todorovic and the rapidly-improving Spanish GM David Larino Nieto are probably the most
consistent GM practitioners at the moment, with good to excellent results. Ekaterina Kovalevskaya and
Natalia Zhukova seem to play it all the time, again with good results. Admittedly, the Tarrasch doesn’t
have a regular super-GM advocate at the moment and, accordingly, it is unlikely to become as popular as
when Spassky and Kasparov were beating everyone with it. But this is rather to our advantage, since
playing an opening at the height of fashion tends to lead to one’s opponents being better prepared and
more experienced in the resulting structures.

Begging the Question: What Happens When You Don’t Define Your Terms
– suggested title for popular science book (London Review of Books, March 2011)

Let’s see what we’re talking about:

1 d4
White occupies the centre. Classical theory would suggest that, if it were his turn to move again, 2
e4 would be his choice.
1 ... d5
Black responds in kind, stopping e2-e4 and gaining a foothold in the centre.
2 c4
White chips away at the black central pawn. While everybody knows that this isn’t a real gambit,
after 2 ... dxc4 3 Nf3 Nf6 4 e3 White can be made to work a little to regain the pawn following 4 ... Be6!?
or Romanov’s favourite 4 ... b5.
2 ... e6
This is an almost perfect move (holding the centre and accelerating kingside development), with
only one drawback: namely that the c8-bishop is blocked inside its pawns. Of course, in the Tarrasch this
drawback is quickly resolved when pawns are traded on d5.
3 Nc3
The choice between this move and 3 Nf3 is a matter of taste and the overall repertoire of the White
player. 3 Nc3 is necessary for players who use the Exchange Variation against the Queen’s Gambit
Declined. 3 Nf3, on the other hand, is indicated for Catalan fans.
Regarding the Tarrasch, it is fair to say that 3 Nf3 is slightly more flexible for White since he gains
certain options based on delaying the development of his queen’s knight (whereas Nf3 is going to be a
necessary move fairly early in almost every variation). In addition, 3 Nc3 allows a pretty interesting
gambit continuation which we’ll see in a few moves.
3 ... c5
 The move which defines the Tarrasch. Black plays the move he wants to play (striking at the white
centre, gaining space, preparing a guilt-free development of the knight to c6 etc). The only drawback from
a principled viewpoint is that his d-pawn is going to become isolated ...
4 cxd5
... hence this is the most principled continuation. White doesn’t have to go for this immediately, or
at all, and the alternatives will be discussed in plenty of detail later in the book.
4 ... exd5
Taking the pawn back and keeping a pawn in the middle of the board. Moreover, the c8-bishop is
looking at a clear diagonal (which it actually gets to use, as may be compared with, for instance, many
lines of the QGD Exchange Variation).
While a lot of chess writers point out that the Stonewall Dutch is one of the very few openings
where Black gains a space advantage, this is actually not correct. First, Black doesn’t technically have a
space advantage in the Stonewall (a white pawn is on the fourth rank, just like the d5- and f5-pawns, and
White can play Ne5, just like Black can play ... Ne4). Second, the best regular example of Black gaining a
space advantage is the Tarrasch. If the white d-pawn gets eliminated (through d4xc5 or ... c5xd4), Black
will have the only pawn on the fourth rank and, moreover, the only pawn in the centre. In some lines this
pawn even makes it to d4 which, while by no means always in Black’s favour, certainly marks out even
more central space. The lines where the white d-pawn survives are less common, and the most notable is
the 9 Bg5 c4 variation where central space is equal but Black gets a queenside space advantage.
None of the above has any real bearing on the assessment of this position, or of the whole opening.
However, what I’m trying to point out is that Black’s concept is an aggressive one. In effect, he’s trying to
play as White, albeit with a few tempi less. If White plays passively, Black will be able to mobilize
sufficiently to justify his central space and the commitment inherent in an IQP. Whether Black can justify
his concept in the face of accurate white play remains to be seen. But for the moment let’s put the pawn in
the middle. As Rear Admiral Grace Hopper put it, it’s easier to ask forgiveness than it is to get
permission.
4 ... cxd4!? is the gambit I mentioned a few moves ago. Black gets compensation for sure. How much
is the question ...
 5 Nf3
No real prizes for this one. Looking at the position I’m not even sure what other sensible moves
exist, though I’m sure that hasn’t stopped people writing books and articles on them.
5 ... Nc6
A rule of thumb I tend to follow in the Tarrasch is that, given the option, I develop this knight before
the king’s knight. One reason is because the queen’s knight doesn’t have a better square (as mentioned
above, after ... c7-c5 there’s no reason not to play ... Nc6), whereas the king’s knight, in a small number
of cases, might want to come to e7 or to avoid an early pin by Bg5, as is well illustrated by 6 Bg5 Be7 7
Bxe7 Ngxe7.
6 g3

An invention of the great Akiba Rubinstein. The fact that his many victories from this position were
achieved against players displaying as much resistance as wet tissue paper doesn’t change the assessment
of this idea as ahead of its time. Today it remains the critical test of the Tarrasch, for the simple reason
that the d5-pawn is now in considerable danger of just being taken and Black needs to find a way to avoid
this happening without lapsing into passivity.
 6 ... Nf6
Knights before bishops. The Tarrasch is a classical opening and rewards players who put their
pieces quickly on good squares. At least this is my theory behind Jacob Aagaard’s excellent successes
with the line, when many other titled players get caught up in more convoluted development schemes.
7 Bg2 Be7
The bishops go to their best squares. Obviously d6 isn’t a great square since it interferes with the
defence of the d5-pawn and doesn’t take any sting out of Bg5.
8 0-0 0-0
It’s pretty rare for the kings to go elsewhere in the Tarrasch – maybe in some of the d4xc5 endgame
lines, but otherwise there’s just too much happening on the c-file for either king to feel safe after long
castling.
9 Bg5

From a principled standpoint, a perfect move. White develops his final minor piece to a good square
and creates an immediate threat of some combination of Bxf6 and d4xc5. Black has to respond somehow,
by either 9 ... cxd4 (the classical main line), 9 ... c4 (the trendy modern main line) or 9 ... Be6 (the
discarded line). Notwithstanding what modern theoreticians may suggest, having analysed these lines I
think all three are perfectly playable – which, incidentally, leads to a pretty high analytical overhead for
White if he wants to play 9 Bg5, and may explain in part his preference for other systems such as d4xc5
(at move 9 or earlier).
We’ll get into specifics in a minute. In the meantime, I wish you best of luck with the Tarrasch!

Finally, I’d like to thank John Emms, Jonathan Tait and Byron Jacobs for their considerable assistance
with this book.

Sam Collins
Dublin,
October 2013
Structural Introduction
The main characteristic of an opening is the structures to which it leads, and for no opening is this more
true than the Tarrasch. The perceived (and actual) inferiority of Black’s structure compared to White’s is
the only reason for not playing an opening with such great development and, ordinarily, a central space
advantage.

In this section I introduce some of the main structures to which the Tarrasch leads. This is not a
comprehensive overview. Nor do I deal with the intricacies of each structure – entire books have been
devoted to the Isolated Queen’s Pawn alone.
For those learning the Tarrasch, there are two great additional resources. First, look up games in
other openings which feature the IQP. The ideas are applicable in all IQP positions regardless of the
initial sequence. Indeed, in some of the positions in the Symmetrical Tarrasch, after poor play by White
we end up in known variations where White has an IQP, with colours reversed.
Second, study specialist resources on the IQP and related structures. A real classic in this field is
Winning Pawn Structures by Alex Baburin, which focuses on the IQP and the related cases of hanging
pawns and the isolated pawn couple. In addition, Winning Chess Middlegames by Ivan Sokolov is an
absolute must for advanced players – while it deals with other pawn structures too, Sokolov’s knowledge
and handling of IQP positions is exemplary. Indeed, some of his games are featured in this text, with both
colours.

Structure No. 1

This is the basic Tarrasch structure, most commonly arising from the traditional main line 9 Bg5
cxd4, as well as some of the 9 dxc5 lines. White has fianchettoed his king’s bishop, which puts immediate
pressure on the d5-pawn (ordinarily this pawn is securely blockaded and cannot advance).
One of the things I realized when researching this book is how even experienced Tarrasch
practitioners don’t always realize just how vulnerable the d5-pawn is, especially at an early stage.
White’s c3-knight and g2-bishop already target the IQP, and the pressure can be increased with Qb3 (or
Qa4) and Rd1. This, when combined with the fact that the black queen is normally on the d-file (d8 or
d7), frequently leads to central shots where White takes on d5.
By far the most common way in which Black can (and does) endanger his IQP is with ... Bg4 and,
after h2-h3, retreating the bishop to h5. The bishop then seems active, targeting the e2-pawn, and is
immune from exchange. My conclusion (after a lot of work) is that, ordinarily, the bishop is just
misplaced on h5 in this structure. The pressure on the e2-pawn is of minor significance (White’s knight on
c3 protects e2 and, if necessary, e2-e3 can be played), but of more significance is that the bishop cannot
trade its counterpart on g2 and, importantly, cannot drop back to e6 to protect the pawn. Often a sequence
involving g3-g4 and Nxd5 can clip a pawn, even following a temporary piece sacrifice. Of course, if
White plays g3-g4 there may be some compensation in the form of weakened squares around White’s
king. However, just be sure, when putting the bishop on h5, that you can hold d5, or at least extract
sufficient concessions to form valuable compensation. If not – and in the majority of cases, Black doesn’t
get enough if d5 falls – retreat the bishop to f5 (if available) or e6.
The position is rich and normally the structure is changed into one of those discussed in the next few
pages.
If White keeps the structure as it is (and the choice to change the structure is entirely White’s, since
Black can’t force piece trades on c6 or e6), then develop your pieces in the middle and see what happens.
The knights will be well placed on c6 and f6 – the f6-knight normally needs to stay put to hold d5, while
the c6-knight can sometimes seek activity by ... Na5-c4 or ... Nb4 followed by ... a7-a5. The light-squared
bishop is well placed anywhere on the c3-h8 diagonal and, if a trade of light-squared bishops can be
arranged, this is normally in Black’s favour. Black’s dark-squared bishop has a less clearly defined role
in the early stages, and often is simply tucked out of the way after ... Re8 and ... Bf8. Black’s rooks are
well placed on any of the c-, d- and e-files – rooks on c8 and e8 seem the most common. The black queen
is a difficult piece to handle, and finding a good square for her often signals the solution to Black’s
opening problems; ... Qd7 followed by ... Bh3 is one common way, and sometimes ... Qb6 is seen instead
to put pressure on the queenside.

Structure No. 2

This is the most significant transformation of the basic Tarrasch structure, which occurs when White
trades knights on c6. Of course the change brings certain benefits to Black – the d5-pawn is protected and
the half-open b-file is available for a major piece.
When White goes into this structure he always has a lot of control over d4 and c5. Accordingly,
while Black could, in theory, go into a hanging pawns position with ... c6-c5, in the Tarrasch this happens
very rarely.
White’s plan is based on blockading the c- and d-pawns and, ultimately, winning one of Black’s
queenside weaknesses. Note, however, that c6 is more difficult to attack than d5 (for instance, the bishop
on g2 doesn’t hit the pawn, so White is normally limited to a frontal attack on the half open c-file; in
addition, a white knight or bishop frequently occupies the c5-outpost and, while such a piece is well
placed, it shields the pawn from attack). Therefore Black has time to try to build up his counterplay and
the risk of an early disaster is reduced, although Black needs to take care to avoid falling into passivity.
Normal methods include seeking counterplay on the half-open b- and e-files, while exchanging White’s
light-squared bishop with ... Bh3 is always a good plan, with potential for a kingside attack.

Structure No. 3

This is a structure which commonly arises when Black plays ... Be6, White plays Nxe6 and Black
recaptures with his f-pawn.
At first glance it seems that White must have a significant advantage here, with two bishops and a
superior pawn structure. In fact Black gets major trumps of his own. First, his pawn structure is improved
– the d5-pawn is now solidly defended and his rook will be active on the half-open f-file. Black’s central
control is now extremely strong, with more space and a 2-1 central majority. In addition, the pawns on d5
and e6 have a limiting effect on the white bishop on g2, which is no longer attacking an isolated pawn on
d5. This is a similar concept to that in the Stonewall Dutch, where a bishop on g2 is restricted by the
opponent’s light-squared central pawn chain. Incidentally, this also informs why an ... e6-e5 advance is
extremely risky, since the g2-bishop would come powerfully into play.
Note that, by contrast to the previous structure, it is very difficult for White to attack the e6-pawn,
which is not on a half-open file.
It is very important for the assessment of these structures that White cannot favourably make an e2-
e4 break. In order for Black to be okay, White either should not be able to play e2-e4, or Black should be
able to meet with e2-e4 with ... d5-d4, keeping the position closed. By contrast, ... d5xe4 would be a
strategic disaster, opening lines for the white bishop on g2 and exposing the weak pawn on e6.
I can’t be sure, but I suspect that a reluctance to go into this structure is the main reason that many
players retreat their bishop to h5 after White plays h2-h3 in the main lines after 9 Bg5 cxd4. As noted in
my discussion above, I really don’t like putting the bishop on h5 in most positions, but you can only avoid
it if you are happy to let White capture on e6.
This structure was defended with great success by Kasparov in his Candidates matches. He had no
problems in game two of his 1983 match with Korchnoi, or game two of his 1984 match with Smyslov,
and won a great game in this structure in game twelve against Smyslov.
The following comment of Sokolov, which will arise in the annotated games, is worth noting here:
“It is good to note that per definition Black is never worried about 13 Nxe6 fxe6, since the loss of the
bishop pair is fully compensated for by the improvement of his pawn structure”. I agree with this
assessment and believe that a key part of handling the Tarrasch well is being ready to go for this structure
as Black.

Structure No. 4

Question: Didn’t we look at this already?

Answer: Not really. The structure we examined previously is the main Tarrasch position with a
black pawn on d5, as normally arises from the 9 Bg5 cxd4 variations. In those positions White has far too
firm a grip on d4 for Black to be able to play ... d5-d4 himself. The structure currently on the board is
characteristic of a number of lines, including most of the major continuations after 9 dxc5. Advancing the
pawn to d4 affects the plans for both sides in several profound ways. On d4 the pawn is safer from attack
than it was on d5 – for instance, the g2-bishop isn’t hitting the pawn. The fact that Black has seized so
much central space means that it can be awkward for White to arrange an immediate attack on the pawn.
Rather he will rely on other advantages – the g2-bishop’s range of activity is extended, hitting the b7-
pawn, and a knight on d3 would be perfectly placed.
While the e2-pawn is rarely a real weakness in the 9 Bg5 cxd4 lines, here it is fixed on e2 (playing
e2-e3 will lead, at best, to an exchange of pawns) and its normal defender, the c3-knight, has been driven
away from its natural defensive square.
The plans for both sides are markedly different than with a black pawn on d5, so be sure not to play
on autopilot.

Structure No. 5

This structure is characteristic of the Symmetrical Tarrasch. The major difference compared to lines
we have looked at so far is that White has not gone for the more aggressive option of fianchettoing his
king’s bishop, but instead puts his bishop either on e2 or on b5 (intending to take a knight on c6).
Black is normally more comfortable in these lines than with a bishop on g2. The d5-pawn is under
less immediate pressure and, if White makes a couple of inferior moves, it is quite common to transpose
(with reversed colours) into other openings in which White gets a favourable IQP position. The pawn still
being on g2 means that Black can consider more direct kingside attacking methods, such as forming a
queen and bishop battery on the h2-b8 diagonal. Indeed, the ideal IQP formation – pawn on a6, knights on
c6 and f6, bishops on c7 and h3 (after White has been forced to play g2-g3), queen on d6, rooks on d8 and
e8 – is established by Black in a number of the examples we’ll examine in the chapter on the Symmetrical
Tarrasch. If Black gets these moves in, White will be the one who needs to equalize.
One effect of the pawn on e3 is that White’s dark-squared bishop can’t easily develop to g5, and
instead tends to go to b2 (after b2-b3 or a2-a3 and b2-b4). While the bishop is well placed here, it is at
risk of being irrelevant if Black succeeds in creating a strong kingside attack. In general White has fewer
kingside defenders than in the fianchetto lines, and Black must seek to exploit this.

Structure No. 6
 This structure is characteristic of the 9 Bg5 c4 line as well as the Swedish Variation (featuring an
early ... c5-c4, ... Bb4 and ... Nge7). Black gains additional queenside space and prepares, if allowed, a
massive advance with ... b7-b5-b4. In addition, the d5-pawn isn’t as weak as it would be if it sat on the
end of a half-open d-file.
As against this, playing ... c5-c4 takes all pressure off the white centre. Thus e2-e4 becomes an idea,
as does Ne5 and, if White wishes, Nxc6. We will also see examples where White seizes more kingside
space with f2-f4.
The positions in this structure tend to become quite sharp. White’s most reliable method is to seek to
destroy the black pawn chain by Ne5xc6 and b2-b3. Black, for his part after ... b7xc6, tries to arrange
counterplay with ... c6-c5. The game is unusual and unbalanced and I think that becoming closely
acquainted with this structure significantly enhances a Tarrasch player’s level, especially since a number
of quieter lines (for instance, after an early e2-e3 by White) can also be met by ... c5-c4.

Isolated Queen’s Pawn

One of the benefits of studying IQP positions is that they arise in such diverse contexts. As Ivan Sokolov
observes: “Isolated pawn structures are arguably the structures that arise from the most different openings
(Tarrasch Defence and Semi-Tarrasch, Queen’s Gambit Accepted, Queen’s Gambit Declined, Nimzo-
Indian Defence, Meran Variation, Ragozin Variation, Petroff Defence, etc) and are therefore very
important positions to understand, regardless of the opening preferences a player may have.”
There are many examples of players who have managed to exploit transpositions and structural
similarities to play a wide range of openings confidently. For instance, a number of GMs are happy to
defend an IQP with Black after 1 e4 e6 2 d4 d5 3 Nd2 c5 4 exd5 exd5, a line developed by Korchnoi and
Mikhail Gurevich amongst others. Bartosz Socko also feels comfortable in this line with Black and, in a
game against Gawain Jones in the German Bundesliga 2012, he answered 1 e4 c5 2 c3 with 2 ... e6 3 d4
d5 4 exd5 exd5, obtaining a very similar position and making a comfortable draw.
In the Politiken Cup 2013, I was interested to see Tarrasch enthusiast IM Jonathan Carlstedt
obtaining positions similar to his favourite opening in two consecutive rounds. Here he is in round six:

J.Carlstedt-J.Brorsen
Helsingor 2013
1 c4 c5 2 e3 Nf6 3 d4 g6 4 Nc3 Bg7 5 Nf3 0-0 6 Be2 d5 7 0-0
Carlstedt has used this set-up in several games as White, scoring heavily. After, for instance, 7 ... cxd4 8
exd4 Nc6 we have a Tarrasch with colours reversed and an extra move for White. While I don’t believe
this line is a serious attempt at an opening advantage, Carlstedt is able to draw on his experience with the
Tarrasch and plays such positions well.
The game continued:
7 ... dxc4 8 Bxc4 cxd4 9 exd4 Bg4 10 d5 Nbd7 11 Re1 Nb6 12 Bb3 Rc8 13 h3 Bxf3 14 Qxf3

Black hasn’t handled the opening in the most principled way and White has a comfortable advantage,
going on to win in 39 moves.

In round seven, the following happened:

J.Glud-J.Carlstedt
Helsingor 2013

1 c4 e5
As we will see, the Tarrasch is sufficiently universal that it can be used against the English: 1 ... e6
followed by 2 ... d5 and ... c7-c5. Carlstedt has played this way as Black a number of times, but decides
to surprise his opponent (an extremely strong IM from Denmark who made a GM norm in this
tournament).
2 g3 Nf6 3 Bg2 c6 4 d4 exd4 5 Qxd4 d5 6 Nf3 Be7 7 0-0 0-0 8 cxd5 cxd5 9 Nc3 Nc6 10 Qa4
 It’s easy to see the similarities of this position to the d4xc5 lines in the Tarrasch – indeed, 10 Qd1?!
Bc5?! would be a direct transposition. Carlstedt has a lot of experience in this structure and effortlessly
obtains a good game.
10 ... Qb6 11 Qb5 Bc5 12 Qxb6 axb6 13 Bf4 h6 14 a3 g5 15 Bd2 d4 16 Nb5 Ne4 17 b4 Be7 18
Rad1 Rd8

Black is very comfortable here. He went on to sacrifice a pawn and actually obtained decent
winning chances before the game ended in a draw at move 75.

I hope that these examples have demonstrated how universal Tarrasch-type positions are, and how
much is to be gained by learning to handle them well.
 Chapter One
9 Bg5 cxd4
1 d4 d5 2 c4 e6 3 Nc3 c5 4 cxd5 exd5 5 Nf3 Nc6 6 g3 Nf6 7 Bg2 Be7 8 0-0 0-0 9 Bg5 cxd4 10
Nxd4 h6

In this chapter we discuss the classical main line of the Tarrasch. I have to admit that Black’s
position looks a little loose, and he needs to play accurately in order to keep the d5-pawn alive without
making too many concessions. During a conversation with an elite player, he informed me that Black’s
position looked “stupid” to him. Certainly when we compare the position to other IQP positions
(especially those obtained by White), here Black seems several tempi down on his standard set-up.
Moreover, White has completed development and managed to fianchetto his king’s bishop, which is the
most aggressive anti-IQP approach (trying, as Larsen suggested, not to blockade the pawn but to capture
it). I think Black is okay here, but he needs to be accurate over the next 5-10 moves to avoid falling into a
passive and prospectless position.
11 Be3
By far the main continuation, which takes up the bulk of the chapter. White reinforces his control
over d4. The bishop did its job on g5 by forcing Black to release the central tension.
11 Bf4 is an annoying sideline in which Black doesn’t win many games but does lose a few. I’ve
covered two approaches for Black: the sound (Li Shilong-Braun, Game 6) and the enterprising (Socko-
Petrosian, Game 7). These games also cover 11 Bxf6 which Black should be happy to see appear on the
board.
11 ... Re8
This was Kasparov’s choice and accordingly is the only move most players are aware of in this
position.
11 ... Bg4 is a great alternative which has scored 75% in world championship matches. Recent
analysis (as demonstrated in practice) by Sokolov has shown that this doesn’t deserve to be a sideline.
See Movsesian-Sokolov (Game 5) for the details.
12 Rc1

The modern main line. White gets another piece into the game and prepares to strangle Black on the
dark-squares after Nxc6 and Na4 in some order.
12 Qb3 gets a game of its own for two reasons. First, this was Karpov’s weapon against Kasparov
in their first world championship match, and the 2/2 he achieved basically buried the Tarrasch for a
generation. Accordingly Black should make a point of being well prepared here.
Second, this is the recommendation in Lars Schandorff’s recent White repertoire series, so Tarrasch
players can anticipate to be tested here. 12 Qb3 is actually a line in which Black has a great new(ish)
idea, as demonstrated in Babula-Orsag (Game 3).
12 Qa4 and White’s 12th move alternatives are serious moves which have been played by serious
players. Bu Xiangzhi is one of them (see Game 4).
12 ... Bf8
Completing the regrouping commenced with 11 ... Re8. This can’t be a bad move but it’s fair to say
that White’s responses have been well mapped out. Sokolov-Petrosian (Game 1) is a nice demonstration
of White’s chances.
12 ... Bg4, just like on move 11, is a good alternative. See Kasimdzhanov-Berelowitsch (Game 2).

Game 1
I.Sokolov-T.L.Petrosian
European Cup, Kemer 2007

1 d4 d5 2 c4 e6 3 Nc3 c5 4 cxd5 exd5 5 Nf3 Nc6 6 g3 Nf6 7 Bg2 Be7 8 0-0 0-0 9 Bg5 cxd4 10 Nxd4 h6
11 Be3 Re8
 The traditional main line, which has fallen out of favour in recent years as Black has experimented
with different approaches, in particular 9 ... c4. John Nunn makes the following comment: “Technically,
we have a standard type of IQP position in which Black will look for kingside counterplay by occupying
the half-open e-file and perhaps the outpost at e4. White for his part has a choice of two plans: he may
continue to blockade d4 and build up pressure on the weak pawn, or he may exchange knights on c6 and
transfer his attack to the backward c-pawn and the blockade point c5.”
12 Rc1
This is by far the most popular move and the main line. White develops his rook and takes aim at the
knight on c6. Right now the rook occupies an open file and, in the event that White takes on c6, the rook
will be perfectly placed exerting pressure on the c6-pawn.
The immediate 12 Nxc6 is rarely played since it gives Black more information – now he knows that
he will be playing with an Isolated Pawn Couple rather than an IQP and can arrange his pieces
accordingly. An interesting example from one of the great Tarrasch experts of the 80s: 12 ... bxc6 13 Bd4
Nh7!? (aiming for g5 from where the knight supports ... Bh3 and can drop back to e6, which is a great
square for a knight in this structure) 14 Rc1 Ng5 15 Na4 Ne6!? (simply 15 ... Bd7 looks equal) 16 Rxc6
Bd7 17 Rc1 Nxd4 18 Qxd4 Bf6 19 Qd1 (Black is not worse in the endgame after 19 Qxd5 Bxa4 20 Qxa8
Qxa8 21 Bxa8 Rxa8) 19 ... Rb8 20 b3 Bxa4 21 bxa4 d4 and, while the computers love White here, his
extra doubled a-pawn isn’t felt in this opposite-coloured bishop position and a draw was soon agreed in
Y.Razuvaev-J.Nunn, London Lloyds Bank 1983 (½-½ in 29).
12 ... Bf8 13 Nxc6
This move is the most popular, though perhaps the trend has recently shifted to 13 Na4. In some
respects 13 Nxc6 is open to the same criticism as 12 Nxc6, namely that it shows White’s hand too early.
13 Na4 is a variation on the same theme – White is still likely to capture on c6, but brings his knight
to c5 directly, which can create additional problems by hitting the b7-pawn. Black has a wide choice:
a) 13 ... Bd7 14 Nc5 and then:
 a1) 14 ... Na5 15 b3 Rc8 16 Nxd7 Qxd7 17 Qd3 Nc6 18 Nxc6 bxc6 19 Rfd1 Qb7 20 Bc5 Bxc5 21
Rxc5 Qe7 22 Rc2 Qa3 23 e3 led to a classic win for White in V.Kramnik-M.Illescas Cordoba, Linares
1994 (1-0 in 38). Again Sokolov’s comments are instructive: “Black is facing a very difficult defence.
Due to his better pawn structure, White has a lasting advantage, while Black does not have dynamic
activity to nearly compensate for the positional problems related to his weak pawns. It is important to
note that transitions from an isolated pawn to two weak hanging pawns are a very common way to combat
the isolated pawn. Contrary to parallel hanging pawns in the centre, such weak connected pawns are by
definition not mobile, so there are no tactics that can be based on this mobility.”
a2) 14 ... Nxd4 15 Bxd4 (15 Qxd4 Bc6 16 Nd3 Bd6 17 Rc2 Bc7 18 Qh4 led to a typical exchange
sacrifice in B.Galstian-T.L.Petrosian, Stepanakert 2005: 18 ... Rxe3! 19 fxe3 and now 19 ... Bb6 would
have left Black with full compensation) 15 ... Bc6 16 Nd3 Bd6 17 e3 Ne4 18 a3 Qe7 19 Re1 a6 20 h4
Rad8 and Black had a solid position in R.Kasimdzhanov-S.Iuldachev, Tashkent 2009 (½-½ in 51).
a3) 14 ... Bxc5 15 Rxc5 Qe7 is interesting; for instance, 16 Qc1 Qe5! 17 Rd1 Qh5 with some
attacking chances in Se.Ivanov-A.Lugovoi, St. Petersburg 2005 (½-½ in 44).
b) 13 ... Qa5
 Exercise: Can White win a pawn by taking on c6?

Answer: Yes he can, and this is a key point of the entire line: 14 Nxc6 bxc6 15 Rxc6! (it has been
established that this is basically unplayable for Black, so 13 ... Qa5 is dubious) 15 ... Bd7 16 Bd2 Bb4
(16 ... Qd8 17 Rc1 Bxa4 18 Qxa4 Rxe2 19 Bc3 Ne4 20 Bd4 Qg5 21 Bf3 1-0 was a rapid defeat for a
strong player in A.Lupor-S.Galdunts, Bad Wörishofen 2008; while 16 ... Qb5 is met by 17 Rxf6! gxf6 18
Nc3 Qxb2 19 Nxd5 with extremely strong compensation for White) 17 Rc5 Qxa4 18 Qxa4 Bxa4 19 Bxb4
Rxe2 20 Ra5 with a very difficult endgame for Black, as in A.Greenfeld-M.Illescas Cordoba,
Thessaloniki Olympiad 1988 (½-½ in 42).
c) Some recent games have suggested that Black can take advantage of White’s omission of 13 Nxc6
with 13 ... Ne5!? 14 Nb5 and now:

c1) 14 ... Nc4 15 Bd4 Bg4 16 Nac3 Ne5 17 Nxa7 Qd7 (17 ... Nc6?! 18 Nxc6? bxc6 19 h3 Bf5 20 a3
Nd7 21 f4 c5 was agreed drawn in M.Sharif-V.Akobian, Abu Dhabi 2000 – Black has full compensation
for the pawn, with pressure on the e-pawn and Benko-style play on the queenside; in addition, his central
control is excellent, like in a Blumenfeld gambit; however, 18 Bxf6! followed by Nxd5 was a more
critical test) 18 Nab5 Bh3 19 Qb3 Rac8. Here White went on to lose in D.Anderton-O.Rodriguez Vargas,
European Senior Team Championship, Dresden 2008 (0-1 in 41), but he would have been winning after
20 Nxd5.
c2) 14 ... Nc6 15 Bc5 (15 Bf4! is slightly better for White) 15 ... Bg4 16 Re1 Bxc5 17 Nxc5 Qb6 18
Nd6 Re7 was level in T.Nyback-J.Ehlvest, Harjumaa (rapid) 2010 (½-½ in 34). The jury’s still out on
this 13 ... Ne5 idea.
13 ... bxc6 14 Na4 Bd7 15 Bc5 Bxc5 16 Nxc5
 Question: What’s your assessment of this position?

Answer: It seems to me that White has an edge. I found Ivan Sokolov’s comments on this position to
be extremely instructive: “White’s strategy is rather clear. The dark-squared bishops have been
exchanged, the knight on c5 is a strong, dominant piece and the black pawn on c6 is a newly created
weakness. Black has to create counterplay in order to achieve a dynamic balance. However, without the
dark-squared bishops and with the white knight on c5, counterplay is not easy to find. It should also be
mentioned that almost all further piece exchanges favour White.”
16 ... Bg4 17 Re1 Qa5
By far the most popular move here, activating the queen and targeting a2. Other moves such as 17 ...
Qb6 have been tried, but I think the text is best.

18 h3
Now Black faces a difficult decision of where to retreat the bishop.
18 ... Bh5
 Keeping an eye on the e2-pawn, but allowing the manoeuvre which White executes in the game.
18 ... Bf5 has been tested more often:

a) 19 Qd4 Rab8 20 a3 (an attempted improvement over 20 ... Qb5 21 b3 Ne4 22 b4 a5 23 Nxe4
Bxe4 24 Rc5 Qb6 25 bxa5 and White went on to win in G.Kasparov-M.Illescas Cordoba, Linares 1994;
Stohl suggests 21 ... a5!) 20 ... Qb6 21 b3 (21 Qf4 Be4 22 b3 Rbd8 23 b4 Re7 24 Qd2 Bxg2 was agreed
drawn in A.Bratchenko-A.Bezgodov, St. Petersburg 2000 ) 21 ... Ne4 22 Bxe4 dxe4 23 Kh2 e3!? 24 fxe3
Rbd8 25 Qf4 Bg6 with a fine position for Black in M.Dziuba-V.Genba, Pardubice 2008 (0-1 in 32).
b) 19 Qa4 Qd2 20 Qf4 Qxf4 21 gxf4 Rab8 22 b3 is an ending which has been defended by Akobian
a couple of times. Then 22 ... Rb5 (or 22 ... g5!? 23 fxg5 hxg5 24 Rc3 Kg7 25 Kf1 Nh5 26 e4 Bg6 27
exd5 Rxe1+ 28 Kxe1 Nf4 29 Bf1 cxd5 and, if anyone, Black was better in V.Filippov-V.Akobian, Las
Vegas 2002 ; while 22 ... Rb4 23 e3 g5 24 a3 Rb6 25 fxg5 hxg5 was level in T.Enhbat-V.Akobian,
Connecticut 2004) 23 e3 g5 24 a3 was Al.David-F.Manca, Cento 2011, and now I like 24 ... a5!, keeping
a target on b3.

19 Nd3 Bg6
Sokolov suggests 19 ... Qxa2 20 Rxc6 Bg6 21 Nb4 (21 Nf4 Be4!) 21 ... Qxb2 22 Nxd5 Nxd5 23
Bxd5 Kh7 24 e4 Rac8 “and a draw should be the most likely outcome”.
20 Nf4

20 ... Be4
Again Sokolov suggests the dynamic 20 ... Qxa2 21 Nxg6 fxg6 22 Rxc6 Qxb2, as after 23 Bxd5+
Nxd5 24 Qxd5+ Kh8 25 Rxg6 a5!, the a-pawn should give Black enough compensation to draw.
21 Rxc6 Bxg2 22 Kxg2 d4 23 Qc1

23 ... Nd5
23 ... Qxa2 24 Qc4! is similar to the next note, since Sokolov points out that 24 ... Qxb2?? fails to 25
Rc7 Rf8 26 Ng6.
24 Nd3 Nb4
Sokolov gives 24 ... Qxa2 25 Qc4 with advantage for White in view of his superior pawn structure.
25 Nxb4 Qxb4 26 Rc4 Qb7+ 27 Kg1 Qd7 28 Qf4 Rad8 29 Rd1! Qxh3 30 Rcxd4 Rxd4 31 Qxd4
a6
Not 31 ... Rxe2?? 32 Qd8+ Kh7 33 Qd3+, winning a rook.
32 e3
 White has a sound extra pawn. A consideration of the endgame would take us too far afield, but we
can conclude that the opening and middlegame have been a success for White.
32 ... Rc8 33 Qd7 Qxd7 34 Rxd7 Rc2 35 Rb7 g5 36 a4 a5 37 b3 g4 38 Rb5 Rb2 39 Kf1 Kg7 40
Rxa5 Rxb3 41 Rf5 Rb2 42 Ke1 Ra2 43 Rf4 h5 44 Kd1 Kg6 45 Kc1 f5 46 Kb1 Rd2 47 a5 Rd8 48 Kc2
Ra8 49 Ra4 Kg5 50 Kd3 h4 51 gxh4+ Kxh4 52 a6 Kh3 53 a7 Kg2 54 Ra2 Kf3 55 Kd4 f4 56 exf4 Kxf4
57 Ra3 1-0

Game 2
R.Kasimdzhanov-A.Berelowitsch
German League 2011

1 d4 d5 2 c4 e6 3 Nf3 c5 4 cxd5 exd5 5 Nc3 Nc6 6 g3 Nf6 7 Bg2 Be7 8 0-0 0-0 9 Bg5 cxd4 10 Nxd4 h6
11 Be3 Re8 12 Rc1 Bg4
While this move has not been historically as popular as 12 ... Bf8, it seems to be the recent trend. It
is notable that the move also has Kasparov’s endorsement, who writes: “Apparently 12 ... Bg4 13 h3 Be6
is better than 12 ... Bf8, as, for example, in the simultaneous game Züger-Kasparov (Switzerland 1987)
and in many later games.”
13 h3 Be6
 Question: Can’t White just win the bishop pair by taking on e6?

Answer: Yes he can, but this trade gives Black certain advantages too. Black’s structure is
improved – his d5-pawn is strongly supported and he has the half-open f-file which can be used by his
rook. Black’s central control is excellent. White would love to play e2-e4 to crack open the position for
his bishops but this isn’t so easy to achieve.
We’ll see later in the notes to Bu Xiangzhi-Li Shilong (Game 4) that Garry Kasparov successfully
defended this structure as Black in his Candidates matches against Smyslov and Korchnoi.
14 Nxc6
Scherbakov notes: “This approach can hardly pose problems for Black.” Nevertheless, 14 Nxc6 is
the most popular move in the position.
a) The immediate 14 Nxe6 fxe6 15 f4 is an ambitious approach, playing for a central bind.

However, pushing the f-pawn in this fashion means that Black will always have counterplay; for
instance, 15 ... Kh8 (or 15 ... Qa5 16 Bd2 Rad8 17 Qa4 Bb4 18 Qxa5 Bxa5 19 Rfd1 Bxc3 20 bxc3 e5 21
Rc2 and a draw was agreed in F.Levin-A.Berelowitsch, Düsseldorf 2011 ; Stohl suggests that 15 ... Bb4!?
16 Bf2 Qe7 “followed by ... Rad8 leads to an unclear position, as the scope of White’s bishops is
limited”) 16 Bf2 Bb4 17 Qb3 Qe7 18 a3 Bc5 19 Bxc5 Qxc5+ 20 Kh2 Qe3 with good play for Black in
Z.Rahman-S.Iuldachev, Asian Championship, Doha 2003.
b) 14 Qa4 Qd7 has also been extensively tested:
b1) 15 Kh2 a6 and now:

b11) 16 Rfd1 b5 17 Qc2 Rac8 has been well defended by Jeroen Bosch: 18 Qb1 (18 Nxe6 fxe6 19
Qb1 Bd6 and, if anyone, Black was better in J.Timman-J.Bosch, Dutch Championship, Hilversum 2006)
18 ... Bf8 19 Nxe6 fxe6 20 Ne4! was slightly better for White in C.Matamoros Franco-J.Bosch, Deizisau
2010 (½-½ in 82); but White’s exchanging operation could have been anticipated with the more active 18
... Bd6!, when Black is fine.
b12) 16 Nxc6 bxc6 17 Rfd1 Rab8 18 Rd2 Bf5 19 a3 Qe6 20 Bf4 Rbc8 and Black had no problems
in J.Djordjevski-G.M.Todorovic, Skopje 2011. The game is a good example of how Black’s kingside
build-up can become threatening if White is not accurate: 21 g4?! Bg6 22 Bg3 h5! 23 f3?! (White is trying
to hold his kingside structure, needlessly weakened by g3-g4?!, but is creating a lot of dark-square
weaknesses in the process) 23 ... Nd7 24 Re1? Nb6! 25 Qxa6 Nc4 26 Rd4 Nxb2 (26 ... h4 27 Bf4 Ra8 28
Qb7 Bc5 29 Rdd1 Bd6 is even stronger, when the white queen is trapped) 27 e4 h4 28 Bf4 Bg5 29 e5
Bxf4+ and White resigned.
b2) It is interesting that, after getting less than nothing with White in this line against Bosch in round
two of the Dutch Championship, Timman decided to try the black pieces in round 8! J.Werle-J.Timman,
Hilversum 2006, saw 15 g4 Bd6 16 Nxe6 fxe6 17 f4 Rad8 18 Rcd1 Qe7 19 Bf2 Bc5 20 e3 Rf8 with a
tense position in which Black’s chances were no worse (½-½ in 60).
c) 14 Qc2 Qd7 15 Nxe6 fxe6 was solid for Black in J.Hjartarson-M.Illescas Cordoba, Linares 1988
(½-½ in 88).
d) 14 Na4 was played in K.Chernyshov-G.M.Todorovic, Budapest 2002 (½-½ in 21), and now 14
... Nxd4 15 Bxd4 Rc8 gives Black good counterplay.
e) 14 Kh2 is the most flexible continuation, with a normal opening edge for White:
 14 ... Qd7 15 Qc2 (15 Qb3 Rac8 16 Rfd1 was slightly better for White in L.Van Wely-
M.Magomedov, Yerevan Olympiad 1996, and after the game’s 16 ... Ne5?? Ivan Sokolov notes that White
could have won on the spot with 17 Nxe6 fxe6 18 Nxd5!! exd5 19 Rxd5 Nxd5 20 Bxd5+ Kh8 21 Be6) 15
... Rac8 16 Rfd1 Red8 (o r 16 ... Bc5 17 Qa4 b6 18 Nc2 Red8 19 Nd4 and a draw was agreed in
D.Komarov-R.Ponomariov, Donetsk 1998) 17 a3 Bd6 18 Nxc6 bxc6 19 Bd4 Nh7 20 b4 h5 21 h4 a5 with
a draw in R.Ruck-G.M.Todorovic, Topola 2004.
14 ... bxc6 15 Na4 Qd7 16 Kh2
Not 16 Nc5? Bxc5 17 Bxc5 Bxh3 and White was a pawn down in B.Khotenashvili-N.Batsiashvili,
European Rapid Championship, Gaziantep 2012 (0-1 in 33).
16 ... Bf5 17 Bc5

17 ... Rab8
Instead:
a) 17 ... Be4 18 Bxe7 Bxg2 19 Kxg2 Qxe7 20 e3 Qe4+ 21 Kh2 Rac8 was equal in Y.Seirawan-
J.Piket, Merrillville 1997.
b) 17 ... Bd8!?, preserving the important dark-squared bishop, was Kasparov’s choice in his
simultaneous game (via a different move order), which continued 18 Bd4 Ne4 and then:

b1) 19 f3 Ng5 20 Nc5 Qe7? (20 ... Qc8! held the balance) 21 e4 dxe4 22 fxe4 Bg6?! was B.Züger-
G.Kasparov, clock simul, Switzerland 1987, when the champ would have been on the ropes after 23 h4!.
b2) 19 e3 Qd6 (19 ... Be7 20 Nc5 was M.Gurevich-R.Ponomariov, Belfort 1998, and now 20 ...
Nxc5 21 Bxc5 Bd8!? again retains the dark-squared bishops and promises interesting counterplay; note
that Black can defend his c6-pawn with the active ... Re6, building up on the e-file and preparing to swing
to the kingside) 20 Nc5 Nxc5 21 Bxc5 Qd7 22 Qa4 Be4 23 Rfd1 Qb7 24 b3 Bxg2 25 Kxg2 Re6 and
Black was fine in I.Cheparinov-G.M.Todorovic, Pancevo 2003 (½-½ in 54). The exchange of the light-
squared bishops has weakened the white king, while Black has met the pressure on c6 in a convenient
way, activating his rook with ... Re6!. This rook holds the c6-pawn and is well placed on the e-file, while
it retains options of moving to f6, g6 or (after ... h6-h5) h6 to menace the white king.
18 b3 Bd6 19 e3
19 Bxd6 Qxd6 20 e3 Nd7 21 Qd4 was A.Greenfeld-A.Frois, Novi Sad Olympiad 1990, and now 21
... a5 would have retained the balance.
 Exercise: Is there a way for Black to undermine White’s kingside pawn structure?

19 ... Be4
Answer: Instead of this sensible move, Scherbakov suggests the typical 19 ... h5!? 20 Bxd6 Qxd6 21
Qd4 Be4 22 Qc5 Qe5 23 Qc3 Qe7 “and the idea ... h5-h4 gives Black sufficient counter chances.”
20 Bxd6 Bxg2
A little too compliant. Again Scherbakov suggests 20 ... Qxd6 21 Qd4 h5.
21 Kxg2 Qxd6 22 Qd4 Ne4 23 Nc5
White has an edge, but Black remains solid.
23 ... a5

24 Rfd1

Question: What do you think of White’s chances in the rook and

pawn endgame after 24 Nxe4 Rxe4 25 Qc5 Qxc5 26 Rxc5?

Answer: This is a typical transformation of the position and one which we need to know how to
assess properly. In general, Black should not have any problems in the rook and pawn endgames arising
from the Tarrasch. However, he has to play actively. Here the move which fits the bill is 26 ... a4!,
eliminating the queenside. After 27 Rxc6 axb3 28 axb3 Rxb3, even if White wins the d5-pawn we would
have a theoretically drawish 4 vs. 3 position on the kingside (Black would prefer to trade one pair of
rooks before entering this structure). Scherbakov analyses 29 Rc8+ Kh7 30 Rd1 Re5 31 Rc5 Rb2 32
Rdxd5 Rxe3 33 Rc7, concluding: “White wins the f7-pawn and can certainly play for a win with four
rooks on the board, but Black should be able to defend.”
The rest is not so relevant for our purposes.
24 ... Ng5 25 e4 Rb4 26 Rc4 Ne6 27 Nxe6 fxe6 28 Rdc1 e5 29 Qd2 Rxc4 30 Rxc4 Rb8 31 Ra4
Rb5 32 Rxa5 Rxa5 33 Qxa5 dxe4 34 a4 Kf7 35 Qa7+ Kf6 36 a5 Qd5 37 Kh2 Qxb3 38 Qb6 Qa4 39
h4 e3 40 Qxe3 Qxa5 41 Qf3+ Ke7 42 Qxc6 Qa7 43 Kg2 Qd4 44 Qg6 Kf8 45 Qf5+ Ke7 46 Qg6 Kf8
47 Qe6 Qe4+ 48 Kh2 Qf3 49 Kg1 e4 50 Qd5 h5 51 Qe5 ½-½

Game 3
V.Babula-M.Orsag
Czech Championship, Ostrava 2010

1 d4 d5 2 c4 e6 3 Nf3 c5 4 cxd5 exd5 5 Nc3 Nc6 6 g3 Nf6 7 Bg2 Be7 8 0-0 0-0 9 Bg5 cxd4 10 Nxd4 h6
11 Be3 Re8 12 Qb3 Na5 13 Qc2 Bg4 14 Nf5 Rc8
14 ... Bb4 is the modern main line, after which 15 Bd4!? Bxc3 16 Bxc3 Rxe2 and now 17 Qd1 or 17
Qd3 gives White interesting compensation for the pawn.

15 Bd4
In the seventh game of the 1984/85 match Karpov opted for 15 Nxe7+. After 15 ... Rxe7! 16 Rad1
Qe8! Kasparov reached a position he had analysed before the match, but had taken an hour and a half on
the clock, which cost him dearly later in the game: 17 h3 (Black has several good responses to 17 Bxd5,
including Kasparov’s suggestion of 17 ... b5 18 a3 Bh3 “and the threat of ... Rxe3 forces White to seek a
way to maintain the balance”) 17 ... Bh5 18 Bxd5 (18 Qf5 Rxc3 19 bxc3 Bxe2 20 Bd4 Re6 was agreed
drawn in P.Harikrishna-S.Iuldachev, Commonwealth Championship, Sangli 2000 ) 18 ... Bg6 19 Qc1
Nxd5 20 Rxd5 Nc4 21 Bd4 Rec7 22 b3 Nb6 23 Re5 Qd7 24 Qe3 f6 25 Rc5 Rxc5 26 Bxc5 Qxh3 and
Black has a comfortable game, losing only because of a late blunder in A.Karpov-G.Kasparov, Moscow
1984 (1-0 in 42).
15 ... Bc5
 16 Bxc5
Not 16 e3?! Ne4 17 Nh4? Bxd4 18 exd4 b5! with a clear advantage for Black in S.Savchenko-
V.Ivanchuk, Odessa-Istanbul (rapid) 2006 (0-1 in 32).
16 ... Rxc5 17 Ne3
17 Nd4 Nc6! is suggested by Kasparov.
17 ... d4!

This is the move! Kasparov gives detailed analysis proving equality and observes:
“Unfortunately, I rejected the energetic 17 ... d4! – it seemed to me that the tactical features of the
resulting position were advantageous to White. This is clear evidence of the general slump in my play,
caused by the serious failures at the start of the 1984/85 match: after all, the obvious breakthrough in the
centre was both in the spirit of the Tarrasch Defence, and in my style!”
a) 17 ... Be6 and now:
 a1) 18 Rad1 Qd7 (the passive 18 ... Qc8 was Kasparov’s choice in his ninth game against Karpov in
Moscow 1984/85; the game, and its knight vs. bishop endgame, is an absolute classic and one of the most
analysed games in chess history – I encourage everyone to look it up in a database or, ideally, to examine
Kasparov’s commentary) 19 Qd3?! (Kasparov suggests 19 b4! Rc7 20 Ncxd5 Nxd5 21 Qb1 with a clear
advantage for White, or 19 ... Rcc8 20 bxa5 d4 21 Rd3 Rxc3 22 Rxc3 dxc3 23 Qxc3 and White has an
extra pawn) 19 ... Nc4 20 Nxc4 was A.Lastin-A.Bezgodov, Russian Championship, Moscow 1999, and
now 20 ... Rxc4! 21 e3 Rec8 22 Nxd5 Nxd5 23 Bxd5 Bxd5 24 Qxd5 Qxd5 25 Rxd5 Rc2 leaves Black
with excellent drawing chances in the rook and pawn endgame in view of his activity.
a 2 ) 18 Rfd1! (more accurate than 18 Rad1, according to Kasparov) 18 ... Nc4 was S.Shipov-
A.Bezgodov, Russian Championship, Elista 2001 (½-½ in 29), and now Kasparov suggests 19 Nexd5!
Nxd5 20 Nxd5 Bxd5 21 e4! (21 b4! is even stronger) 21 ... Qf6 (21 ... Qa5 is a better chance, with some
hope of a blockade after 22 exd5 Nd6) 22 Rxd5 Rxd5 23 exd5 “with an extra pawn, since in the event of
23 ... Qxb2 24 Qxb2 Nxb2 25 Bf1! Black has a difficult endgame.”
b) 17 ... Bh5 is “interesting, but hardly sufficient”, according to Kasparov:

b1) 18 Qa4 Nc6 19 Rfd1 Qb6 20 Qb3 Qxb3 21 axb3 was S.Temirbaev-A.Bezgodov, Russian Cup,
Samara 2002, and now Kasparov suggests 21 ... d4! 22 Bxc6 (or 22 Na4 Rb5) 22 ... Rxe3! 23 fxe3 dxc3!
24 Bxb7 cxb2 25 Rab1 Rc2 “with excellent compensation for the exchange”.
b 2 ) 18 Rad1 Qc8 was S.Iskusnyh-A.Bezgodov, Russian Team Championship 2005, and now
Kasparov suggests 19 Rfe1! b5 20 Qb1! Bg6 21 Qa1 “hiding the queen in the corner and nevertheless
picking up the d-pawn (or winning the exchange: 21 ... Rxe3 22 fxe3 Qe6 23 e4).”
18 Rad1
18 Rfd1 is the main alternative: 18 ... Nc6 19 Bxc6 Rxc6 20 Nxg4 Nxg4 21 e3 Ne5 22 exd4 Nf3+ 23
Kg2 Nxd4 24 Qa4 Rd6 25 Qxa7 Qd7 (25 ... Qc8!? 26 Rxd4? Ra6! or 26 f3 Red8 is suggested by
Kasparov, with full compensation for the pawn) 26 f3 Nc2 and Black had more than enough for the pawn
in H.Boshku-B.Ivanovic, Nea Makri 1990 (0-1 in 34).
18 ... Nc6
18 ... Nc4 19 Nxg4 Nxg4 was about equal in L.Mkrtchian-R.Pokorna, Khanty-Mansiysk Olympiad
2010.
19 Nxg4 Nxg4 20 e3 Nge5 21 Qb3

Exercise: How can Black justify his play?

Answer: 21 ... d3!

This novelty was suggested by Kasparov (in a book published 24 years after the match!). His
assessment of both the move, and the impact of his 9th matchgame against Karpov, is fascinating:
“To all appearances, 17 ... d4 gives Black sufficient counterplay and is the best response to
Karpov’s plan. But this move has still not won general recognition. The point is that our first match was
not properly analysed, and my defeats in the 7th and 9th games created the impression of being opening
failures, after which the Tarrasch Defence lost its popularity and the development of its theory was
retarded.”
Black sacrifices a pawn, but the passer on d3 provides at least sufficient compensation.
22 Qxb7 Qa5 23 Bxc6 Nxc6 24 Qd7
Kasparov gives 24 Rxd3 Rb8 25 b4 Qxb4 26 Qxb4 Nxb4 27 Rb1 Nxd3 28 Rxb8+ Kh7 29 Ne4 (or
29 Ne2 Rc2) 29 ... Rc1+ 30 Kg2 f5! 31 Nd6 Rc2 with equality.
24 ... Rd8 25 Qg4 Ne5 26 Qe4 Rc4 27 Qg2
 27 ... Rxc3?!
27 ... d2! would have left Black on top. Black still won this game against a good GM, which is yet
another advertisement for the benefits of familiarizing oneself with the classics!
28 bxc3 d2 29 Qe4 Qxc3 30 Rb1 Nd3 31 Qe7 Rc8 32 Rfd1 Qc2 33 Qd7 Ne5 34 Qd6 Qe4 35
Rb4 Nf3+ 36 Kh1 Nd4+ 37 Kg1 Rc1 38 Qb8+ Kh7 39 Rb1 Ne2+ 40 Kf1 Qh1+ 0-1

Game 4
Bu Xiangzhi-Li Shilong
Chinese Championship, Xinghua 2013

1 Nf3 c5 2 c4 Nf6 3 Nc3 e6 4 g3 Nc6 5 Bg2 d5 6 cxd5 exd5 7 d4 Be7 8 0-0 0-0 9 Bg5 cxd4 10 Nxd4 h6
11 Be3 Re8
In this game we will deal with White’s less popular 12th moves. Many of these are serious attempts
for an advantage, and Black’s response varies depending on White’s set-up, so good preparation is
required.
12 Qa4
By creating a cheap threat on c6, White gains time for Rad1. In addition, he hopes that the black
bishop will be passively placed on d7 (and interfere with the defence of the d5-pawn).
a) 12 Qc2 was another approach, used by Beliavsky in the sixth game of his Candidates match
against Kasparov in Moscow 1983. After 12 ... Bg4, White again has a wide choice:
 a1) 13 Rfd1 Bf8 (13 ... Qd7! is Kasparov’s suggested improvement, intending to meet 14 Nb3 with
14 ... Rad8; instead 14 f3?! Bh5 15 Bf2 Bg6 16 Qa4?! Bc5 was already better for Black in A.Graf-
G.Ginsburg, Nuremberg 2006) 14 Rac1 (14 Nxc6 bxc6 15 Bd4 Bh5 was level in M.Krasenkow-
M.Illescas Cordoba, European Team Championship, Batumi 1999) 14 ... Rc8 15 Nxc6 (Kasparov gives
15 Qa4! as stronger) 15 ... bxc6 16 Bd4 (16 Bxa7?! c5! – Kasparov) was A.Beliavsky-G.Kasparov,
Candidates (6th matchgame), Moscow 1983, and now 16 ... a5! (Kasparov) would have held the balance.
a2) Beliavsky tried the other rook some 22 years after his game with Kasparov: 13 Rad1 Qd7

14 f3 (14 Nb3 Be6 15 Nc5 Bxc5 16 Bxc5 was a better try for an advantage, as in E.Alekseev-
V.Yemelin, St. Petersburg 2006 ) 14 ... Bh5 15 Bf2 Bg6 16 Qa4 a6 17 Nxc6 bxc6 18 Rd2 Qe6 and Black
had no problems in A.Beliavsky-T.L.Petrosian, Moscow 2005 (½-½ in 32).
a3) 13 Rfe1 Qd7 14 Rac1 Rac8 15 Nxc6 bxc6 16 Bd4 was Zhao Jun-Tu Hoang Thong, Asian
Championship, Cebu 2007, and now 16 ... Bf5 17 Qd2 a5 would have kept chances level.
a4) 13 h3 has been tried by several GMs recently, giving Black a choice of where to put the bishop:
 a41) 13 ... Bd7 14 Rad1 Qc8 15 Kh2 Bf8 (15 ... Bc5 16 Nxc6 Bxe3 17 Nb4 d4 18 Nb1! was slightly
better for White in Zhang Zhong-S.Iuldachev, Manila 2008; 1-0 in 36) 16 Nxc6 bxc6 17 Rc1 Re5! 18 Bd4
Rh5 with good attacking chances in R.Chytilek-K.Malinovsky, Czech League 2013 (0-1 in 34).
a42) 13 ... Be6 14 Rad1 Qc8?! 15 Kh2 Bf8 16 Nxe6 fxe6 17 f4 and White had a favourable version
of a standard set-up in H.Stefansson-A.Pashikian, European Championship, Antalya 2004 (1-0 in 42),
since Black’s pieces are relatively passive. Instead of 14 ... Qc8?!, 14 ... Rc8 looks natural and good.
a43) 13 ... Bh5 14 Rad1 Bb4 was S.Volkov-S.Guliev, Dubai 2011, and now 15 Nf5! would have
given White a slight advantage.
b) 12 a3 was a move played five times against Garry Kasparov in Candidates matches – once by
Korchnoi and four times by Smyslov. The results of those games (four draws and one win for Black)
helped drive the popularity of the Tarrasch in the 1980s. The move is obviously quite slow (the point
seems to be to control b4 and allow Qb3-a2 in some lines, though Black could be forgiven for not quaking
in his boots at either of these ideas) and Black has a wide choice:

b1) 12 ... Bf8 13 Rc1 (or 13 Qb3 Na5 14 Qa2 Be6 15 Rad1 ½-½ D.Lekic-B.Tadic, Montenegrin
Team Championship 2007) 13 ... Bg4 14 Qb3 Na5 15 Qa2 Be6 16 Rfd1 Nc6 17 Qb1 Qd7 18 Nxe6 fxe6
19 Kh1 Qf7 20 f4 Rad8 and a draw was agreed in S.Bromberger-A.Lugovoi, Gausdal 2004.
b2) 12 ... Bg4 13 Qb3 (13 h3!? seems more typical) 13 ... Na5 14 Qa2 Nc4 15 Bc1 Rc8 and Black
was very comfortable in A.Yusupov-V.Akobian, Philadelphia 2002 (½-½ in 41).
b3) 12 ... Be6 was Kasparov’s selected defence, used in his Candidates matches against Korchnoi
and (four times!) against Smyslov:
b31) 13 Qb3 Qd7 14 Nxe6 fxe6 15 Rad1 Bd6! 16 Bc1 Kh8 17 Qa4 Qe7 18 e3 a6 19 Qh4 Rac8 was
comfortable for Black in V.Korchnoi-G.Kasparov, Candidates (2nd matchgame), London 1983 (½-½ in
31).
b32) 13 Kh1 was tested three times in the Smyslov-Kasparov match.

The following comments are based on Kasparov’s notes:

b321) 13 ... Qd7?! 14 Nxe6 fxe6 15 f4! Red8 16 Bg1 Rac8 17 Qa4 Kh8 18 Rad1 Qe8 19 e4 with a
White initiative in V.Smyslov-G.Kasparov, Candidates (2nd matchgame), Vilnius 1984 (½-½ in 41).
b322) 13 ... Bg4! was Kasparov’s improvement, achieving comfortable play after 14 f3 (14 h3
Bd7!) 14 ... Bh5 15 Bg1 (15 Nxc6 bxc6 16 Na4 Qc8! 17 Bd4 Qe6 18 Rc1 Nd7 was also comfortable for
Black in the 10th game of the match; ½-½ in 38) 15 ... Qd7! 16 Qa4 Bc5 17 Rad1 Bb6 18 Rfe1 Bg6! and a
draw was soon agreed in V.Smyslov-G.Kasparov, Candidates (8th matchgame), Vilnius 1984 (½-½ in
28).
b33) Finally, 13 Nxe6 fxe6 14 Qa4 Rc8 15 Rad1 Kh8 16 Kh1 was V.Smyslov-G.Kasparov,
Candidates (12th matchgame), Vilnius 1984 (0-1 in 40).
 Exercise: Kasparov now played 16 ... a6. Can you do better?

Answer: Kasparov recommends 16 ... Na5! giving the following variations:

b331) 17 Bxa7 Nc4, when White is in danger of losing his bishop, and gains some compensation
(but not more) after 18 Qb3 (or 18 Qb5 b6 19 Na4 Nd7 20 e4 Qc7 21 exd5 Qxa7 22 dxe6 Nf6) 18 ... b6
19 e4 Qc7 20 Nb5 Qb7!.
b332) 17 f4 Nc4 (“Black gains an important tempo and the entire rhythm of White’s attack is
disrupted” – Kasparov) 18 Bc1 Qb6! with good play for Black.
12 ... Bd7 13 Rad1
Here 13 Rfd1 Bc5 14 Ne6!? fxe6 15 Bxc5 wins the bishop pair but improves Black’s structure,
which should lead to balanced play; for instance, 15 ... b6 16 Bd6 Na5 17 Qd4 Nc4 18 Be5 Rc8 with
equality in Wang Yue-T.L.Petrosian, Tiayuan 2005 (1-0 in 49).
b) 13 Nxd5 Nxd5 14 Bxd5 Nb4 15 Qb3 Nxd5 16 Qxd5 Bh3 17 Qxd8 Raxd8 18 Rfd1 Bf6 gave
Black full compensation in G.Kaidanov-R.Robson, St. Louis 2011 (½-½ in 30).
 Exercise: How can Black activate his position?

Answer: 13 ... Nb4 14 Qb3 a5!

“Black holds on by tactical means, threatening by ... a5-a4 to drive the queen from its active
position”. (Kasparov)
I really like this plan, gaining space and stability on the queenside. The b4-knight is a monster which
defends the d5-pawn and almost traps the white queen, and White will need to spend time getting rid of it.

15 a4
The most principled move, cementing the b5-square, stopping Black’s queenside advance in its
tracks and safeguarding the queen on its active post. The b4-knight is happy about the situation though.
a) 15 Nxd5? Nbxd5 16 Bxd5 Nxd5 17 Qxd5 Bh3!, winning the exchange, has happened several
times, including R.Vaganian-B.Ivkov, USSR-Yugoslavia match, Odessa 1975.
b) In the stem game Beliavsky allowed the pawn to advance: 15 Rd2 a4 16 Qd1 a3 17 Qb1?! (17
bxa3 Rxa3 was also fine for Black in Y.Korsunsky-J.Arizmendi Martinez, Tanta 1998, ½-½ in 120 ;
Kasparov suggests 17 Qb3!, giving several variations to demonstrate White’s advantage, e.g. 17 ... Qa5
18 bxa3 Qxa3 19 Nxd5 Nbxd5 20 Bxd5 Nxd5 21 Qxd5 “winning a pawn and forcing the opponent to fight
only for a draw”) 17 ... Bf8 and now:

b1) 18 Nb3 Qc8 19 Rc1 Bf5 20 Qa1 axb2 (20 ... Nc6 is level) 21 Qxb2 Rxe3!? 22 fxe3 (22 Nxd5!
Rxb3 23 axb3 Qb8 24 Nxf6+ gxf6 25 Qxf6, with some advantage, was a better attempt) 22 ... Qe6 and
Black enjoyed full compensation for the exchange in a typical position in V.Kunin-E.Weinzettl, Oberwart
2006.
b 2 ) 18 bxa3 Rxa3 19 Qb2 Qa8 was A.Beliavsky-G.Kasparov, Candidates (2nd matchgame),
Moscow 1983 (0-1 in 38), and now Kasparov suggests 20 Nc2 Nxc2 21 Rxc2 with a level game. Instead,
Beliavsky played passively and Kasparov went on to score a famous victory.
15 ... Rc8 16 Nc2
Hurrying to exchange the powerful b4-knight. While White has tried other moves, I don’t believe he
has serious chances for an advantage if he leaves the b4-knight on the board. For instance, 16 Rc1 Bc5 17
h3? Qb6 18 Rfd1? runs into a standard combination: 18 ... Rxe3! 19 fxe3 Qd6! 20 Rf1 Qxg3 21 Rf3 Qg5
(21 ... Qe5 22 Qd1 Nh7 also led to a decisive attack in L.Ljubojevic-J.Bosch, Dutch Team Championship
1999; 0-1 in 41) 22 Ncb5 Ne4 and Black’s attack was too strong in F.Elsness-B.Byklum, Norwegian
Team Championship 2009 (0-1 in 36).
 Exercise: How would you continue here with Black?

16 ... Nxc2
Not bad, but Black had a more incisive option.
Answer: Kasparov suggests 16 ... b5! as “the most interesting reply”; for instance, 17 Nxb4 bxa4 18
Nxa4 Bxb4 19 Nc3 (White lost material after 19 Nb6? Rxe3! in V.Korchnoi-G.Kasparov, Herzeg Novi
blitz 1983) 19 ... Bxc3 20 bxc3 a4 “with good counterplay” (Kasparov), an assessment which has been
borne out by a number of games: 21 Qb2 (or 21 Qa2 Qe7 22 Rfe1?!, V.Majorovas-J.Ehlvest, USSR 1983,
when 22 ... Rxc3 23 Bd4 Ra3 is better for Black) 21 ... Qe7 22 Rfe1 was A.Lugovoi-A.Lastin, St.
Petersburg 2002, and now 22 ... a3! 23 Qa1 Bb5 24 Bd4 Ne4 would have left White struggling to
equalize.
The game continuation gives White chances of an advantage.

17 Bb6!
A nice intermezzo, as otherwise White has nothing. 17 Qxc2 Bb4 18 Bd4 Ne4 19 Qb3 Bxc3 20 Bxc3
Bc6 was agreed drawn in Wu Shaobin-C.Gokhale, Singapore 2002.
17 ... Bxa4! 18 Nxa4 Qd6 19 e3
19 Bxa5 Qa6 20 Bc3 b5 21 Qxc2 bxa4 22 Qd3 Qxd3 23 Rxd3 Ne4 24 Be5 Bf6 25 Bxf6 Nxf6 was
level in E.Ghaem Maghami-K.Yang, Subic Bay 2009 (½-½ in 80). The white pawn on b2 is as much of a
target as anything in the black camp.

19 ... Nb4?
19 ... Qb4 20 Qxb4 axb4 21 Bxd5 Nxd5 22 Rxd5 b3 should give Black sufficient counterplay against
b2 to hold, although I prefer White. 19 ... Bd8, trying to exchange the strong b6-bishop, also looks like a
better defensive attempt.
20 Bxa5 Ra8 21 Bxb4 Qxb4 22 Qxb4 Bxb4 23 Nb6 Ra2

24 b3
I don’t think that the endgame after 24 Nxd5 Nxd5 25 Bxd5 Rxb2 26 Rb1 Rxb1 27 Rxb1 Bf8 28
Rxb7 is a piece of cake for Black either with rooks on the board.
24 ... d4 25 exd4?!
25 Rxd4 Bc5 26 Ra4 gives more winning chances than the text.
 25 ... Ree2

Black has now rustled up serious drawing chances and held without much trouble. The last few
moves cannot be recommended from a theoretical perspective though and, in particular, 16 ... b5! seems
more promising.
26 Bf3 Reb2 27 Nd5 Nxd5 28 Bxd5 b6 29 Rc1 Rc2 30 Be4 Rxc1 31 Rxc1 g6 32 Rc7 Be1 33 f4
Bf2+ 34 Kh1 f5 35 Bd5+ Kh8 36 Bf7 g5 37 d5 Bc5 38 Be6 Ra1+ 39 Kg2 Ra2+ 40 Kf1 gxf4 41 gxf4
Rf2+ 42 Ke1 Rxf4 43 Ke2 Rh4 44 Bxf5 Rxh2+ 45 Kd3 Rf2 46 Be6 Rf4 47 Rc6 Kg7 48 Kc3 Kf6 49
Bc8+ Ke5 50 Rxh6 ½-½

Game 5
S.Movsesian-I.Sokolov
Sarajevo 2009

1 c4 e6 2 g3 d5 3 Bg2 Nf6 4 Nf3 c5 5 0-0 Nc6 6 cxd5 exd5 7 d4 Be7 8 Nc3 0-0 9 Bg5 cxd4 10 Nxd4 h6
11 Be3 Bg4
This move is played about three times less frequently than the most popular move, 11 ... Re8, but has
an impressive pedigree. It has been used by Spassky (several times, and in World Championship matches)
and Korchnoi. Modern GMs who have tried this more than once include Ivan Sokolov (who gave it a
good write-up in his wonderful book Winning Chess Middlegames), Daan Brandenburg and Michael
Feygin. The heroes of the line, however, are Spanish GM David Larino Nieto and German GM Alexander
Berelowitsch, both of whom have used this move extensively.
 12 Qa4
A logical reply. Since the black bishop has already committed itself to g4, Black needs to look for
something other than ... Bd7 (which worked well in Bu Xiangzhi-Li Shilong above).
a) 12 h3 Be6 and now:

a1) Sokolov makes the following instructive comment: “It is good to note that per definition Black is
never worried about 13 Nxe6 fxe6, since the loss of the bishop pair is fully compensated for by the
improvement of his pawn structure”. This was tested against me in a recent game: 14 Qc2 Rc8 15 Rad1
and now in V.Burmakin-S.Collins, Riga 2013, I played the careless 15 ... Bd6? and immediately panicked
about the consequences of 16 Bxh6!, which my opponent missed but would indeed have been very strong.
Later I even gained a winning position before messing up in time trouble (½-½ in 38). From a theoretical
perspective, 15 ... Qe8! is much stronger, for instance 16 f4 b5 with counterplay.
a2) 13 Rc1 Qd7 14 Kh2 and then:
 a21) 14 ... Ne5 (Sokolov’s recommendation) 15 b3 (after 15 Qa4 Qxa4 16 Nxa4 Nc4 17 Bf4 g5! 18
b3 Bd7, Black was not worse in B.Spassky-V.Korchnoi, USSR Championship, Moscow 1957; ½-½ in
30) 15 ... Rfc8 16 Qd2 was A.Czebe-A.Berelowitsch, Haarlem 2007, and now 16 ... Ba3!? is interesting,
when White can either agree to a repetition with 16 Rb1 Bb4 17 Rbc1 Ba3 or play a complex position
after 17 f4!? (17 Rc2 Nc6 also looks fine for Black, who has good prospects of gaining control of the c-
file thanks to the bishop on a3) 17 ... Bxc1 18 Qxc1 Nc6 19 f5 Nxd4 20 Bxd4 Bxf5 21 Bxf6 gxf6 with a
mess.
a22) 14 ... Rac8 15 Nxc6 bxc6 16 Na4 was what worried Sokolov,

but Berelowitsch seems to have no problems: 16 ... Rfe8 (16 ... Bf5 looks fine too, when 17 Nc5 can
be met by 17 ... Qd6) 17 Nc5 Bxc5 18 Bxc5 Bf5 19 Re1 Ne4 (Black has equalized and the game comes to
a logical conclusion) 20 Qd4 Ng5 21 h4 Ne6 22 Qa4 Nxc5 23 Rxc5 Be4 24 Bh3 Bf5 25 Bg2 Be4 26 Bh3
Bf5 27 Bg2 Be4 28 Bh3 Bf5 with a draw by repetition in S.Siebrecht-A.Berelowitsch, Dutch Team
Championship 2013.
b) 12 Nb3 Be6 13 Rc1 (13 Bc5 Bxc5 14 Nxc5 Qe7 15 Nxe6 fxe6 was harmless in M.Dambacher-
A.Berelowitsch, Dutch Team Championship 2007; ½-½ in 31) 13 ... Re8 was tested in two games of the
Petrosian-Spassky World Championship match, Moscow 1969:

b1) In Game 2 Petrosian opted for 14 Nb5 Qd7 15 N5d4 Bh3 16 Nxc6 bxc6 17 Qd3 Bxg2 18 Kxg2
a5! (“with equality” – Kasparov) 19 Rc2 a4 20 Nd2 (20 Nc5 Bxc5 is a favourable swap for Black,
according to Sokolov) 20 ... Qb7 21 Kg1 Rac8 22 Rfc1 Nd7! “transferring the knight to e5 and preparing
... c6-c5 at the same time. Black was better.” (Sokolov) The game was drawn in 61 moves.
b2) Kasparov suggests 14 Nc5 Bxc5 15 Bxc5 as leading to a small plus for White,

giving the following variations: 15 ... Qa5 (or 15 ... Ne4 16 Nxe4 dxe4 17 Qa4 Qd5 18 Bxe4 Qxa2
19 Qxa2 Bxa2 20 f3) 16 Bd4 Nxd4 (or 16 ... Ne4 17 a3 Rad8 18 e3) 17 Qxd4 Rac8 18 e3 Rc4 19 Qd3
Rec8 20 Rcd1 R8c5 21 a3.
b3) 14 Re1 was Petrosian’s choice in Game 4, which continued 14 ... Qd7 15 Bc5 (again Kasparov
suggests 15 Nc5!?) 15 ... Rac8 16 Bxe7 Qxe7 17 e3 and, as noted by Sokolov, “As a result of the
exchange of the dark-squared bishops, White has good control of the dark squares in the centre (d4 and
c5) and also enough space to manoeuvre his pieces. White should be slightly better here.”
Nevertheless, Black went on to win in 41 moves, which doubtless contributed to the Tarrasch’s
popularity (again, showing how fickle chess fashion is, since the result had nothing to do with the
opening!).

Exercise: How can Black activate his position on the queenside?

Answer: 12 ... Na5!

Starting a knight tour which will hit White’s bishop ( ... Nc4) and queen ( ... Nb6). While I don’t like
knights on b6 generally, here it has the plus side of securely defending the d5-pawn and not blocking
Black’s counterplay on the c-file. In addition, since it is difficult for White to play b2-b3 because of
pressure on the c3-knight, a knight on b6 will always have options of coming to c4.
12 ... Qd7 is not bad and is actually the most popular move in the position, but the stronger guys have
gone 12 ... Na5 and I think it’s a better move.
13 Rad1 Nc4 14 Bc1 Nb6!

As Sokolov notes in his book, this move is more precise than 14 ... Qc8 15 Qc2 Rd8 16 b3 Ne5 17
Bb2 Qd7 18 f3 Bh5 19 Qf5, when White was better in T.V.Petrosian-B.Spassky, World Championship
(12th matchgame), Moscow 1969 (½-½ in 49).
 15 Qb5
Keeping the queen active and trying to prevent ... Qd7.
a) 15 Qc2 Qd7 16 f3 looks odd to me, despite having been played by a number of GMs.

a1) 16 ... Bh3 17 Bxh3 Qxh3 18 Nf5 (18 Qf5 Qxf5 19 Nxf5 is better, as in A.Matnadze-L.Hodova,
European Junior Championships 2001) 18 ... Bc5+ 19 Kh1 was Y.Vovk-D.Brandenburg, Dieren 2011,
and Black would have had an excellent game after 19 ... Rfe8!.
a2) 16 ... Bh5 17 f4 (17 Qf5 Qe8 was G.Timoscenko-Z.Sturua, Pavlodar 1982, which Sokolov
assesses as better for Black) 17 ... Rac8 18 f5 Bd6 19 Qd3, H.Tikkanen-S.Brynell, Swedish Team
Championship 2010, and now 19 ... Be5!, threatening ... Bxd4+ followed by ... Rxc3 and ... Bxe2, would
have left Black on top.
b) 15 Qb3 and now:

b1) 15 ... Qd7 maintains the bishop on the h3-c8 diagonal and seems the natural response to White’s
refusal to play Qb5. Then 16 Rfe1 (instead, 16 a4 was met by the instructive 16 ... Rac8 17 Bf4 Rc4! 18
Nf3 Rb4 19 Qa2 Be6 with equality in J.Cuenca Jimenez-D.Larino Nieto, Spanish Junior Championships
2003; and the endgame after 16 Qb5 Rac8! looks acceptable for Black, who has better development to
compensate his potentially weak d5-pawn; but 16 Be3 gives some pressure, since 16 ... Bc5 17 Ne6! led
to a White advantage in L.Vadasz-P.Szilagyi, Hungarian Championship, Budapest 1974 ) 16 ... Rac8 (or
16 ... Rfe8 and if 17 a4, as in M.Horn-M.Breutigam, German League 2010, then 17 ... Rac8! with equality
looks like the most active set-up for Black) 17 Ndb5 Rc4 18 a3 was D.Larino Nieto, Madrid 2007, when
18 ... d4 19 e3 Be6 20 exd4 Rxc3 21 Qxc3 Qxb5 22 Bxb7 Nbd5 would have given Black a good game;
but in this line 18 Bxd5! is critical: 18 ... Rb4 19 Bxf7+ Rxf7 20 Rxd7 Bxd7 21 Qd1 Bxb5 22 a3 Ba4 23
Nxa4 Rxa4 and Black has enough material for the queen, but after 24 b4! his rook is sidelined and White
is much better.
b2) 15 ... Rc8 16 h3 and now:

b21) 16 ... Bh5 17 Nf5 Bg6 18 e4 Bxf5 19 exf5 Qd7 was A.Gupta-P.Konguvel, Paleochora 2009,
and now White could simply have played 20 Nxd5 with a clear advantage.
b22) 16 ... Bd7 is my preference. This is actually the position after 17 Qb3 in Movsesian-Sokolov
with White to play! Very surprisingly, White doesn’t seem to have a good way to exploit his free move.
For instance, 17 g4, as in E.Cordova-M.V.Santos, Havana 2007, can be logically met by 17 ... Bd6!,
taking control of the dark squares. After 18 Be3 Nc4 19 Nxd5 Nxd5 20 Bxd5 Qh4 21 Bxc4 Qxh3 22 Bf4
Qxg4+ 23 Bg3 Rxc4 24 Qxc4 Bxg3 25 fxg3 Qxg3+, the game ends in perpetual.
White’s alternatives on move 17 are no more impressive. 17 Bf4, as in R.Disconzi da Silva-
M.V.Santos, Guarulhos 2005, could be met by 17 ... a5!? 18 a3 a4 19 Qa2 Re8 with decent play for
Black; while 17 Rd3 Rc4 18 Qd1 Qc8 19 Kh2 Bb4 was comfortable for Black in De.Abel-D.Larino
Nieto, World Junior Championships, Chalkidiki 2003 (0-1 in 34) ; finally, Sokolov gives 17 Kh2 Rc4!
with good play for Black.
 15 ... Rc8!
The most solid move, developing a piece to a good square.
After the more aggressive continuation 15 ... a6 16 Qd3 Qd7, then 17 b3 Rfe8 18 Bb2 reaches a
position which has been played five times by Larino Nieto, scoring +2 =3 -0! In addition, he is the only
player to have played this way as Black, at least according to my database!

a) 18 ... Bc5 19 Rc1 Rac8 (19 ... Bh3 20 Rfd1 Re5 was a more aggressive approach in E.Llobel
Cortell-D.Larino Nieto, La Roda 2005, and ½-½ in 47, but it would not have looked convincing after the
simple 21 e3) 20 e3 (20 Rfd1 Bh3 21 Bh1 Ng4 22 e3 Ne5 was agreed drawn in A.Franco Alonso-
D.Larino Nieto, Elgoibar 2008) 20 ... Bh3 21 Nce2 Bxg2 22 Kxg2 Ne4 23 Nf4 Rc7 (23 ... Be7 24 f3 Nc5
25 Qd2 Bf6 26 Qb4 Qc7 was P.San Segundo Carrillo-D.Larino Nieto, Spanish Championship, Ceuta
2008, but here 27 Nh5! would have given White a decisive advantage) 24 f3 Nd6 25 a4 and Black’s
position looked ugly in V.Andrejic-D.Larino Nieto, European Championship, Budva 2009, though he
went on to win (0-1 in 38).
b) 18 ... Rad8 19 Rc1 Bh3 20 Rfd1 Bd6 21 e3 Be5 22 Nce2 Ne4 23 Bh1 Qe7 24 f3 Nf6 25 a4 Nbd7
26 Qd2 Bb8 27 Kf2 h5 28 Re1 h4 29 Nf4 Bxf4 30 exf4 Qd6 31 Bg2 Bxg2 32 Kxg2 hxg3 33 hxg3 Rxe1 34
Rxe1 Re8 35 Nf5 Qf8 was played in P.San Segundo Carrillo-D.Larino Nieto, Spanish Championship,
Ceuta 2008 (0-1 in 45). The notation starts going a little crazy over the next few moves, which are
irrelevant in any event, since instead of the game’s 36 Rxe8, White could have gained a decisive
advantage with 36 Rh1, when the difference in piece activity is too great.
16 h3 Bd7

17 Qb3
Another very strong player was easily neutralized after 17 Qd3 Re8 18 b3 Bb4 19 Bb2 Qe7 20 e3
Qe5 21 Nce2 Ne4 22 Ba1 Qe7 and Black’s activity fully compensated for his IQP in R.Kasimdzhanov-
A.Berelowitsch, German League 2006 (½-½ in 47).
17 ... Rc4 18 Nc2 Qc8 19 Ne3 Rc5 20 Nexd5 Nfxd5 21 Nxd5 Nxd5 22 Bxd5 Bxh3 23 Bxb7
Movsesian grabs a pawn. 23 Rfe1 b6 looks equal too.
23 ... Qc7 24 Bf4 Qa5 25 Qe3

25 ... Bg5!
25 ... Bxf1?? loses decisive material after 26 Qxe7, since 26 ... Bh3 27 b4! Qxb4 28 Bd6 is curtains.
26 Bxg5 Rxg5 27 Bg2 Bxg2 28 Kxg2 Qxa2 29 b4 ½-½
Game 6
Li Shilong-Ar.Braun
Wijk aan Zee 2008

1 d4 e6 2 Nf3 d5 3 c4 c5 4 cxd5 exd5 5 Nc3 Nc6 6 g3 Nf6 7 Bg2 Be7 8 0-0 0-0 9 Bg5 cxd4 10 Nxd4 h6
11 Bf4

A rather annoying sideline. The bishop is active on f4 and puts some early pressure on Black’s
position. In particular, White has the e2-e4 break in many lines (because the pawn isn’t blocked by a
bishop on e3). The tried and tested response is 11 ... Bg4 12 h3 Be6, usually leading to a tenable endgame
which, admittedly, is not a lot of fun for Black. He does have more enterprising alternatives, but Black
needs to be careful to hold his position together since, after Qd4 and Rfd1, White will already be creating
strong central pressure.
After writing the above comments, I found Korchnoi’s notes to his game against Lobron (discussed
below), which describe the perfect attitude for handling the endgame in this line:
“Here 11 Be3 is most often played. The bishop move to f4 is also known in theory. I think that here
my opponent committed his first inaccuracy. The bishop move to e3 leads to a complicated middlegame,
whereas after Bf4 White intends to gain an advantage in the endgame. And it is the endgame that I play
strongest of all!”
A great approach, and while it is admittedly adopted by one of the strongest endgame players in
history, it is worth striving to be able to approach endgames (in this and other lines) with the same
attitude.
By contrast, 11 Bxf6 Bxf6 leads to very comfortable positions for Black.
 Y.Seirawan-G.Kasparov, Niksic 1983, continued 12 Nb3 (12 Nxc6 bxc6 is an excellent hanging
pawn position for Black, who will easily develop with ... Re8 and ... Rb8, while White struggles to put
any pressure on the c6- and d5-pawns; 12 e3 is the most popular move, but 12 ... Nxd4 13 exd4 Be6 leads
to equality, for instance 14 Qb3 Bxd4 15 Bxd5 Bxc3 16 Bxe6 Bd4 17 Bd5 Qf6 18 Qf3 Bxb2 19 Rad1
Qxf3 20 Bxf3 Rad8 with total equality in B.Thorfinnsson-A.Pashikian, European Championship, Antalya
2004; ½-½ in 47) 12 ... d4 13 Ne4 (marked as “dubious” by Seirawan in his annotations; 13 Nd5 can be
comfortably met by 13 ... Be5!) 13 ... Be7 14 Rc1 Qb6

15 Nec5 (“The ‘big idea’ is revealed. Clearly, the plan is to stop the c8-bishop from developing
thanks to the long diagonal pressure. Also, simply to snap off the d4-pawn, sometimes at the costs of the
g2-bishop.” – Seirawan. A more recent game saw an impressive activation of Black’s queen’s rook after
15 Qd2 a5! 16 Nbc5 a4 17 Nd3 Ra5! 18 Qc2 Be6 and Black was better in M.Al Sayed-E.Senador,
Olongapo City 2010) 15 ... Rd8 16 Rc4?! (Seirawan had prepared 16 Qd3 Nb4 17 Qe4 Bxc5 18 Nxc5
Nxa2 19 Rc4 Qxb2 20 Rb1, which he thought was good for White, but decided against it at the last
moment – in fact Houdini thinks this position is equal; in addition, Black has other options such as 17 ...
Bg5 18 Rcd1 Nc6, and it isn’t clear what White has achieved) 16 ... Bxc5! 17 Nxc5 (“Already I have to
fight for equality. While it is typical in this line to play 17 Rxc5 Be6 18 Bxc6, giving up my pride and joy,
Black isn’t obliged to recapture my bishop. Instead, Black has 18 ... Bxb3! 19 Qc1 Bxa2 when he is a
pawn to the good. The text gives Black a smaller plus.” – Seirawan) 17 ... Qxb2 18 Qc2 Qxc2 19 Rxc2
Rb8 20 Rb2 Rd6 21 Rd1 b6 22 Nb3 Bb7 (it was also interesting to try 22 ... d3!? 23 Rxd3 Rxd3 24 exd3
Bd7, when the question is whether the d3-pawn will prove to be a strong passed pawn or a weak isolated
pawn) 23 Rbd2 Rbd8 24 Kf1

(“Although I will win back my pawn by capturing the c6-knight and then the d4-pawn, the resulting
position of rook and bishop versus rook and knight will favour Black, and I’ll be forced to fight for the
draw. Black will have the sunny side of the torture.” – Seirawan) 24 ... Ba6 25 Bxc6 Rxc6 26 Nxd4 Rc5
27 Nb3 Rxd2 28 Rxd2 Rc7 29 Rd8+ Kh7 30 Ke1 Bc4 31 Kd2 g6 32 Nc1 Kg7 33 a3 Kf6 34 e3 Ke7 35
Rd4 Rd7 36 Kc3 Rxd4 37 Kxd4 b5 38 Nd3 Bxd3 39 Kxd3 Kd6 and here, instead of the drawing 40 Kd4!,
Seirawan blundered with 40 e4? and went on to lose a famous endgame (0-1 in 54).
11 ... Bg4 12 h3
12 Qa4, as at move 13, can be met by 12 ... Qb6. Instead, 12 ... Qd7 13 Rfd1 Rfd8 14 Rac1 a6? 15
Nxd5! Nxd5 16 Bxd5 Nxd4 17 Qxd4 was clearly better for White in M.Kanep-J.Ehlvest, Harjumaa
(rapid) 2010.
12 ... Be6
The main line. This variation (as will be seen in the game and the notes) often gives rise to an
endgame (after White takes on c6 and plays e2-e4) where White is slightly better in view of his two
bishops but, in practice, Black has held the balance without too much trouble.
 Black’s 11th and 12th move alternatives will be covered in the next game.
13 Nxe6
This is the most direct, leading directly to the endgame.
a) 13 Ncb5 Qd7 14 Nxe6 fxe6 15 Rc1 e5 16 Bd2 Rad8 left Black with no problems in N.Dzagnidze-
A.Gara, French Team Championship 2013 (½-½ in 25).
b) 13 Rc1 Rc8 14 Nxe6 fxe6 15 e4 d4 16 e5 dxc3 17 exf6 Bxf6 18 bxc3 Qxd1 19 Rfxd1 Rfd8 also
leads to a slightly better endgame for White which Black ought to hold, as in I.Sokolov-V.Akobian,
Calatrava (rapid) 2007 (½-½ in 45) and A.Greenfeld-S.Iuldachev, Calicut 2007 (½-½ in 62). Sokolov’s
assessment of his game against Akobian was: “In the slightly better endgame that resulted after 11 Bf4, I
never had any real winning chances and the game ended in a dull draw.” We will get an idea of how
Black should handle these positions in the main game.
13 ... fxe6

14 e4!
The next few moves are forced. As noted in the Structural Introduction, taking on e4 would be a
disaster.
 14 ... d4 15 e5 dxc3 16 exf6 Bxf6 17 bxc3

17 ... Qxd1
The alternative is 17 ... Bxc3. As explained by Korchnoi: “The game is proceeding along a well-
trodden theoretical path. There were also games in which Black avoided the capture on c3 and in the end
equalized. But I thought: White already has the advantage of the two bishops, and if Black has to suffer,
then it should at least be for some material.”
White now has a choice:
a) 18 Rc1 Qf6 19 Bxc6 (19 Qb3 Nd4 and White has nothing better than retreating to d1) 19 ... bxc6
20 Qb3 Be5 21 Bxe5 Qxe5 22 Rxc6 (22 Rfe1 Qd5 equalized in M.Wiedenkeller-Al.Poulsen, Copenhagen
1986) 22 ... Rae8 and White has nothing, according to Korchnoi.
b) 18 Rb1 Qxd1 (18 ... Rf7 19 Rxb7 Rxb7 20 Bxc6 Qxd1 21 Rxd1 Rab8 22 Bxb8 was agreed drawn
in V.Anand-Comp Fritz 6, Frankfurt rapid 1999) 19 Rfxd1 and now:

b1) 19 ... Rf7 20 Rxb7 (20 Bxc6 bxc6 21 Rbc1 e5 22 Be3 Bd4 23 Bxd4 exd4 24 Rxd4 Rb8 25 Rc2
Rc7 was drawn in U.Krause-A.Ilgner, Kappeln 1990) 20 ... Rxb7 21 Bxc6 Rab8 22 Bxb8 and a draw was
agreed in W.Paschall-G.Kallai, Hungarian Team Championship 2012, but here 20 Rbc1 or 20 Rd6 were
better tries for an edge, so I prefer the move order Fritz used against Anand, given above.
b2) 19 ... Rad8 is also quite sound:
b21) 20 Be3 Bd4 21 Bxd4 Nxd4 22 Bxb7 Ne2+ 23 Kg2 Nc3 24 Rxd8 Rxd8 leads to a level
endgame, as in D.Hergott-R.Alonso Garcia, Linares, Holland 1992 (½-½ in 50) and J.Jurek-D.Florea,
Olomouc 2002 (½-½ in 50).
b22) After 20 Rdc1 Korchnoi gives 20 ... Bd2 21 Bxd2 Rxd2 22 Rxb7 Nd4 23 Rcc7 Rfxf2 24
Rxg7+ (24 Be4 Nf5 25 Bxf5 Rg2+ and Black has perpetual) 24 ... Kf8 with equality, which is confirmed
by the computer: 25 Bh1 Rf6, followed by checks by the rook on the d-file.
b23) 20 Rxd8 Nxd8 and now:

b231) 21 Bxb7 Nxb7 22 Rxb7 Bd4! is level.

b232) 21 Rd1 Nc6 leads to a further branch:
b2321) 22 Be3 Bb4 23 Rd7 Rf7 24 Bxc6 bxc6 25 Rxa7 Rxa7 26 Bxa7 Kf7 was roughly level in
D.Yevseev-A.Kirusha, Gatchina 2001. White has an outside passed pawn but Black has a passed c-pawn
and he really shouldn’t lose this position.
b2322) 22 Rd7 Rf7 23 Rd6 Re7 24 Bxc6 bxc6 25 Rxc6 Bd4 with equality in A.Panin-R.Bataev,
Murom 2000 (½-½ in 42).
b2323) 22 Rd6 e5 23 Be3 Bd4 24 Rd7 Bxe3 25 fxe3 (this endgame is assessed by Gurevich as
slightly better for White) 25 ... Na5 26 Bd5+ Kh7 (as Gurevich notes, Black plans ... b7-b5 and ... Nc4
with equality) 27 Re7 was M.Gurevich-J.Bosch, German League 2000. Now after Gurevich’s suggestion
of 27 ... Rd8 28 e4 (28 Be4+ Kg8 29 Rxe5 Nc6 is okay for Black), I like 28 ... Rd6!, intending to activate
the rook via b6. The position is quite unbalanced and, frankly, I still prefer White, but I think this is fully
playable for Black.
b233) 21 Be3!
 (“This subtle move, as I later learned, was an innovation: it forces an important weakening of
Black’s pawns on the queenside.” – Korchnoi) 21 ... b6 (“Now the black knight at d8 will not have a post
at c6, and much time will be required to bring it into play.” – Korchnoi) 22 Rd1 was played in E.Lobron-
V.Korchnoi, Bad Homburg 1998. Korchnoi assesses the position as follows: “White has definite
compensation for the pawn. His rook will invade Black’s position, and the two bishops will have their
say.” (Korchnoi notes that 22 Be4 has been played in subsequent games, but not ones which appear in my
database; I suggest continuing to bring the knight into play with 22 ... Nf7 23 Rc1 Nd6 24 Bd3 Be5 with a
complex fight ahead.) The game continued 22 ... Bf6 (the attempt to force an immediate draw with 22 ...
Nf7 23 Rd7 Ne5 24 Rxa7 Nc4 25 Rc7 Nxe3 26 fxe3 Bd2 fails to 22 Rc7! Bb2 23 f4! – Korchnoi) 23 Be4
Rf7 24 Bg6 Rb7 25 Bf4 Nf7 26 Rc1 Rd7 27 Rc8+ Bd8

when Korchnoi gives the following assessment: “White’s advantage is obvious – Black is obliged to
make forced moves. But perhaps White’s success is only temporary. If Black can regroup, he will evict
the bishop from g6 and neutralize the white rook that has invaded his position.”
Korchnoi has analysed the rest of the game in detail in Informator and, in particular, his wonderful
best games collection, the second volume of which is a fantastic lesson in how to win games with Black
(often featuring apparently dull or passive positions such as this one). I can’t recommend it highly enough.
For our purposes, this dull but playable position gives White full compensation but no more, so we
shouldn’t be afraid of it.
18 Rfxd1 Rfd8
18 ... Bxc3 has been played in several games, but if Black wants to go for this it is more accurate to
play it on move 17 (as Korchnoi did), when Rc1 can be met by ... Qf6. Here 19 Rac1 gives White strong
pressure.

19 Rab1
19 Be3 Bxc3 20 Rab1 was M.Dlugy-M.Vucic, New York Open 1992, and now I like 20 ... Bd4! 21
Rxb7 (21 Bxd4 Nxd4 22 Bxb7 Rab8 looks level too) 21 ... Bxe3 22 Rxd8+ Rxd8 23 fxe3 Rd1+ 24 Kf2
Rd2+ 25 Kf1 Na5 26 Rxa7 Nc4 with enough counterplay to hold.
19 ... Rxd1+ 20 Rxd1 Rd8 21 Rxd8+
The position after 21 Rc1 Rd3 22 Rb1 Rd7 23 c4 Bd4 has been defended twice by Czech IM Milan
Orsag: 24 Rb5 (24 h4 Kf7 25 Re1 was R.Akesson-M.Orsag, Aaland-Stockholm 1997, and now I suggest
25 ... Nb4!? 26 Bh3 Bxf2+ 27 Kxf2 Nd3+ 28 Ke2 Nxe1 29 Kxe1 Rd3, when Black is not worse in the
endgame) 24 ... Nd8 25 h4 b6 and Black held without difficulty in J.Nilssen-M.Orsag, Istanbul Olympiad
2012 (½-½ in 40).
21 ... Nxd8 22 Bd2
Golod gives 22 Be3 Bxc3 23 Bxa7 as slightly better for White, but 22 ... b6 is more solid, keeping
the position more closed.
22 ... Kf7 23 Kf1 Nc6 24 Ke2 Ke7 25 Be4 Kd6 26 h4 b6
 Black’s position is solid and, in the game, he drew without major problems. I’d be the first to admit
that such endgames aren’t a great deal of fun for Black, but sometimes we have to play positions like this.
In general, the entire endgame seems much closer to a draw than to a serious White advantage.
27 f4 Ne7 28 h5 e5 29 g4 exf4 30 Bxf4+ Be5 31 Bd2 Nd5 32 Kd3 b5 33 Bf3 a5 34 Be1 Nb6 35
Bf2 Na4 36 Bd4 Nc5+ 37 Bxc5+ ½-½

Game 7
M.Socko-T.L.Petrosian
Gibraltar 2008

In this game I’ve compiled a few ideas for Black should he want to avoid the endgame featured in the last
game. These are enterprising but have to come with a health warning since White seems to have at least
an edge in all lines.
1 d4 d5 2 c4 e6 3 Nc3 c5 4 Nf3 Nc6 5 cxd5 exd5 6 g3 Nf6 7 Bg2 Be7 8 Bg5 cxd4 9 Nxd4 0-0 10
0-0 h6 11 Bf4
 11 ... Bg4
Black has tried a lot of other moves here:
a) 11 ... Qb6 and now:

a 1 ) A game from the Women’s World Blitz Championship demonstrates a number of Black’s
attacking ideas in the Tarrasch: 12 Nxc6 bxc6 13 Na4 Qa6 14 Rc1 Bg4 15 Re1 Rfe8 16 h3 Bh5 17 Be3?
Bb4 18 Nc3 Rxe3! (this typical sacrifice works perfectly in this position) 19 fxe3 Qb6 20 Qd3 Bg6 21
Qd2 Re8 22 a3 Rxe3 23 Kh1 Ne4! 24 Bxe4 Bxe4+ 25 Kh2 Rxg3! 26 Qf4 Rg2+ 27 Kh1 Bxc3 (27 ... Rg1+
forced mate) 28 Rxc3 Rxe2+ and White resigned in Zhao Xue-V.Gunina, Batumi 2012.
a 2 ) 12 Nb3 Rd8 (or 12 ... d4 13 Na4 Qb4 14 Bxc6 bxc6 15 Qxd4) 13 Rc1 d4 14 Na4 was
Y.Pelletier-G.Meins, Biel 2009, and now Black could improve with 14 ... Qb5 15 Nac5 Nd5 with decent
play.
a3) The forced line after 12 Be3 Qxb2 13 Nxd5 Nxd5 14 Bxd5 Bh3 has scored heavily for White
following 15 Rb1 Qa3 16 Rxb7 Nxd4 17 Qxd4 Bxf1 18 Kxf1, as in M.Gurevich-A.Berelowitsch, Tanta
1997.
b) 11 ... Re8 is another possibility:
 b1) 12 Ncb5, as in I.Sokolov-G.M.Todorovic, Niksic 1991, can be met by 12 ... Qb6 13 Nc7 (13
Bc7 Qa6 14 Qb3 Nxd4 15 Nxd4 Bc5 is fine for Black) 13 ... Nxd4 14 Be5! Nc6 15 Nxa8 Qd8 16 Nc7
Nxe5 17 Nxe8 Qxe8 18 Bxd5 Nxd5 19 Qxd5 Nc6 with three minor pieces for two rooks and a pawn.
b2) 12 Rc1 Bg4 13 h3 Bh5 has been tested in a number of GM clashes:
b21) 14 Qa4 Rc8 (not 14 ... Bc5? 15 Nxc6 bxc6 16 g4! Bg6 17 Qxc6 and White was clearly better
in L.Fressinet-G.M.Todorovic, European Championship, Budva 2009) 15 Nxc6 (15 Be3 might be a better
try for an advantage) 15 ... bxc6 16 g4 Bg6 17 Qxa7 was Val.Popov-V.Yemelin, St. Petersburg 2005, and
now 17 ... Bd6 would have left Black with full compensation for the pawn.
b22) 14 Nb3 Bb4 15 g4 Bxc3 16 bxc3 Bg6 17 c4 Be4 and Black was not worse in D.Yevseev-
A.Lugovoi, St. Petersburg 2005 (½-½ in 41).
b23) 14 g4 Bg6 15 Ncb5 Qb6 16 Bc7 Qa6 17 a4, with an edge for White, was tested in B.Lajthajm-
G.M.Todorovic, Serbian Team Championship 2012 (½-½ in 37).
c) 11 ... a6 has also been tried:

c1) 12 Nxc6 bxc6 13 e4 d4 14 e5 dxc3 15 exf6 Bxf6 16 bxc3 Be6 17 Bd6 Be7 18 Bxe7 and a draw
was agreed in R.Vidonyak-G.Ginsburg, Nuremberg 2011.
c2) 12 Rc1 Re8 13 Na4 Bd7 and then:
c21) 14 a3 Bf8 15 Re1 Rc8 16 Nxc6 bxc6 17 Be3 Ne4 18 Nc5 and a draw was agreed in J.Lembak-
K.Malinovsky, Slovakian League 2013. Play could have continued 18 ... Nxc5 19 Bxc5 Bxc5 20 Rxc5 a5,
followed by ... Qb6 with counterplay, or 20 ... Rb8 21 b4 Rb5!?.
c22) 14 Nxc6 Bxc6 15 Qb3 a5 16 Rfd1 Bf8 17 Be3 was slightly better for White in A.Aleksandrov-
E.Magerramov, Dubai 2001 (1-0 in 57).
12 h3 Bh5
 An enterprising move. Having played over several hundred Tarrasch games while preparing this
book, I can say that very often players seem to like retreating their bishop to h5 after h3, presumably with
a view to pressuring the e2-pawn. In fact the pressure against e2 rarely becomes a significant factor, but
the fact that the bishop can no longer defend the d5-pawn renders Black’s position tactically shaky.
Whether he can get away with it depends on the specific features of a given position – in general you
should spend a couple of minutes before you play ... Bh5 in the Tarrasch and try to make sure that your d-
pawn isn’t dropping.
Here the move makes some sense since, without a bishop on e3 defending the d4-knight, Qb3 isn’t
available.
Instead, 12 ... Bd7 can lead to a forcing sequence after 13 Kh2 Qb6 14 Be3 Qxb2 15 Nxd5 Nxd5 16
Bxd5 Rad8 17 Rb1 Qa3 18 Rxb7! (improving on the 18 Qc1 of J.Lautier-A.Grischuk, Cannes rapid 2001;
½-½ in 29), and White went on to win in K.Sasikiran-A.Lugovoi, Moscow 2004 (1-0 in 41). Whatever
the objective merits of this line, it seems rather impractical to go in for a long forced variation, where a
single mistake might be fatal, in a basically harmless sideline like 11 Bf4.
13 Rc1
Many other moves have been tested against T.L.Petrosian. While the games were, admittedly, mainly
played at rapid or blitz time controls, they are instructive in demonstrating Black’s attacking potential:
a) 13 Nb3 Re8 14 g4 Bg6 15 Nxd5 Nxd5 16 Qxd5 Bf6 and Black had equalized in V.Golod-
T.L.Petrosian, Dos Hermanas (online blitz) 2007, since he is about to regain his pawn.
b) 13 Qa4 is dangerous.
 As indicated in my note to Black’s 12th, White aims simply to attack and win the d5-pawn: 13 ...
Qd7?! (13 ... Qb6! 14 Nf5 Bc5! is sounder, as in R.Backelin-L.Johannesson, Gausdal 1990; 0-1 in 54) 14
Rfd1 (already it is unclear how Black can neutralize the pressure) 14 ... Rfd8 15 Rac1 g5 16 Be3 Rac8
was L.Fressinet-T.L.Petrosian, France-Armenia rapid match, Paris 2009, and now both 17 g4 and 17
Nxd5 would have led to a decisive advantage for White.
c) 13 Nf5 Bg6 14 Nxe7+ Qxe7 and now:

c1) 15 Nxd5 Nxd5 16 Qxd5 Rad8 17 Qb5?! (17 Qc4 Nd4 18 Kh2 Be4 gives Black interesting
compensation; for instance, 19 Rfd1 Bxg2 20 Rxd4 Rxd4 21 Qxd4 Bc6 22 Bd6 Rd8 23 Qxg7+ Kxg7 24
Bxe7 Rd2 with a likely draw in view of the pieces remaining and Black’s activity) 17 ... Nd4! was
already better for Black in V.Golod-T.L.Petrosian, Dos Hermanas (online blitz) 2007.
c2) 15 e4 d4 16 Nd5 Nxd5 17 exd5 Nb4 18 Qa4 (18 Qxd4!? Nc2 19 d6 Qd7 20 Qd5 Nxa1 21 Rxa1
gives White enough compensation for the exchange) was G.Vojinovic-G.M.Todorovic, Skopje 2011, and
now 18 ... Nc2 19 Rad1 d3 would have offered Black counterplay based on his strong d-pawn.
13 ... Qd7 14 Qa4
 14 ... Rac8
After 14 ... Rfd8 15 Rfd1 a6 16 Nxc6, a draw was agreed in G.Gutman-T.L.Petrosian, Cappelle la
Grande 2008. But 14 ... a6 looks interesting; for instance, 15 Rfd1 b5 16 Qb3 Na5 17 Qc2 Bg6 with good
play for Black.

15 Rfd1
White has achieved an optimal set-up and the bishop on h5 just looks really misplaced.
15 ... Rfd8
15 ... Nxd4 16 Rxd4 b5 17 Qb3 was better for White in N.Kosintseva-L.Mkrtchian, Krasnoturinsk
2006 (1-0 in 61).
 16 Nxc6
Giving Black some major respite by solving his biggest tactical problem.
Instead, 16 g4! would have led to a clear advantage for White. 16 ... Bg6 17 Nxd5! is a typical
winning shot, so Black would have to go in for 16 ... Nxd4 17 Qxd7 Rxd7 18 Rxd4 Bg6 19 Nxd5 Rxc1+
20 Bxc1, when White is a solid pawn up and Black’s drawing chances don’t look so great.
Even the immediate 16 Nxd5 Nxd5 17 Bxd5 Qxh3 (17 ... Qxd5 18 Nxc6 wins) leaves Black
struggling to stay in the game after 18 Bg2 Qg4 19 Bf3 Qh3 20 Nxc6 bxc6 21 Qxa7.
16 ... bxc6

17 Be3
17 g4 Bg6 18 e4 would have led to unclear play after 18 ... d4.
17 ... Rb8
 Black has now equalized. The rest is not so relevant since we have covered the most important
theoretical moments. White lost the game in the end, which was partially a result of weaknesses she later
incurred on the light squares, but more a result of playing against a strong and inventive GM.
18 Rd2 Rb4 19 Qa5 Qe8 20 Bd4 Rdb8 21 b3 R4b7 22 Bc5 Bg6 23 e3 Nd7 24 Bxe7 Qxe7 25 Ne2
Ne5 26 Nd4 Rb6 27 Qc3 Rc8 28 f4 Nd7 29 Bxd5 Nf6 30 Bg2 Ne4 31 Bxe4 Bxe4 32 Qc5 Qd7 33 Kh2
Rb5 34 Qc4 Rd5 35 Rf2 c5 36 Ne2 Re8 37 Nc3 Rh5 38 Qf1 Bf5 39 h4 Rxe3 40 Rd1 Qe7 41 Nd5
Rxh4+ 42 gxh4 Qxh4+ 43 Kg1 Be4 44 Rg2 Bxg2 45 Qa6 Qh1+ 46 Kf2 Rf3+ 0-1

Key Notes
1. 9 Bg5 cxd4 leads to some of the strategically richest middlegames in the Tarrasch. Black needs a
lot of knowledge to handle these positions well and avoid falling into passivity. In the main line after 12
... Re8, White has a very wide choice, with several continuations which have been endorsed by leading
GMs. Black’s treatment differs for each of them.
2. In the lines with 12 Rc1 or 12 Nxc6, White tends to define the structure at an early stage. He will
aim to get a knight to c5 and solidly blockade Black’s c- and d-pawns. Black should seek counterplay on
the half-open b- and e-files.
3. Early queen moves by White allow Black to create active counterplay, for instance 12 Qb3 Na5
or 12 Qa4 Bd7 13 Rad1 Nb4! 14 Qb3 a5!. I have not seen a convincing route to an advantage for White in
these lines.
4. An underrated idea is ... Bg4, either on move 11 or after 11 ... Re8 12 Rc1. Black has been doing
well here of late.
5. After 11 Bf4, Black needs to adjust his plans. He has a choice between playing a pretty balanced
endgame, as in Li Shilong-Braun (Game 6), or fishing in troubled waters, as in Socko-Petrosian (Game
7).
 Chapter Two
9 Bg5 c4
1 d4 d5 2 c4 e6 3 Nc3 c5 4 cxd5 exd5 5 Nf3 Nc6 6 g3 Nf6 7 Bg2 Be7 8 0-0 0-0 9 Bg5 c4

This caused something of a stir when Alexander Grischuk started beating up various 2600+ GMs
with it in the early 2000s. Certainly the arising positions are unusual and unbalanced, and if White is not
well prepared he can quickly find himself worse. The same mystery elite GM I mentioned in the 9 ... cxd4
introduction confessed he had found “nothing” against 9 ... c4 after several hours' work. More recently,
Aagaard and Ntirlis devoted over 100 pages of computer analysis to proving equality in these lines
(generally featuring an early ... h7-h6 by Black). I don’t wish to parrot this work so my focus is on other
lines which also seem very playable.
10 Ne5
This is almost universally played.
10 ... Be6 11 b3
 This move, and the closely related 11 Nxc6, are, in my view, the best responses to 9 ... c4. White
sets about trying to dismantle the black pawn chain. Often complex endgames result where White has a
superior structure but Black has activity and counterplay (as well as the less noble drawing methods of
relying on simplified positions with the pawns on one side, and the peaceful properties of rook
endgames). See Sargissian-Halkias (Game 8) for the details.
White’s alternatives at moves 10 and 11 are covered in Onischuk-Akobian (Game 9). Of these the
most challenging is 11 f4. The mystery elite player just made a face when I suggested that this was
dangerous for Black, so I suppose we have a convincing antidote ...

Game 8
G.Sargissian-S.Halkias
European Championship, Aix-les-Bains 2011

1 d4 d5 2 c4 e6 3 Nc3 c5 4 cxd5 exd5 5 Nf3 Nc6 6 g3 Nf6 7 Bg2 Be7 8 0-0 0-0 9 Bg5 c4 10 Ne5 Be6
In this game we will examine White’s main attempts, 11 b3 and 11 Nxc6 (which often transpose).
White aims to break up the black pawn chain on c4 and d5 (often a pawn comes to c6 too).
11 b3
The alternative is 11 Nxc6 bxc6 and then:
 a) 12 Na4 Rc8 13 b3 Qa5 14 e3 Rfd8 15 Bxf6 has been tried by a couple of strong GMs. 15 ... gxf6!
(better than 15 ... Bxf6 16 Nc5 with a queenside bind in V.Laznicka-A.Givon, European Cup, Eilat 2012;
1-0 in 26) 16 Qh5 c5 17 dxc5 Bxc5 18 Nxc5 Qxc5 19 Rfd1 c3 with an unclear game where Black enjoys
strong counterplay with the c-pawn, as in H.Koneru-N.Zhukova, FIDE World Cup, Khanty-Mansiysk
2012 (0-1 in 37).
b) 12 b3 and now:

b1) 12 ... cxb3 13 axb3 h6 14 Bxf6 Bxf6 15 Na4 Bf5 16 Qd2 Qd6 17 Nc5 gave White some edge in
P.H.Nielsen-I.Rogers, Turin Olympiad 2006 (1-0 in 55).
b2) 12 ... h6 is an idea deeply analysed by Aagaard and Ntrilis. To take just one line: 13 Bxf6 Bxf6
14 bxc4 dxc4 15 e3 Qa5 16 Rc1 Rac8 17 Qa4 Qxa4 18 Nxa4 Be7 19 Nb2 (19 Rb1 Rc7 20 Nc3 Rd8 21
Rfd1 a5 22 Bf3 Bb4 looked balanced in Al.Ramirez-R.Skytte, Cappelle la Grande 2012) 19 ... Ba3 20
Rc2 Bxb2 21 Rxb2 Rc7 22 Rc1 was tested in A.Rasmussen-J.Aagaard, Helsingor 2012. Now Jacob
played 22 ... g5, forgetting his own recommendation 22 ... g6, which was given a couple of pages of
analysis in his book with detailed lines ending in equality.
b3) 12 ... Qa5 13 Qc2 (13 Qd2 Rad8 transposes to 13 Nxc6 bxc6 in the main game, although Black
might consider other options, such as 13 ... Qa6!?) 13 ... Rfd8 (13 ... Rad8 seems less logical, since the
rooks should aim for the queenside; after 14 Rfd1 Bb4 15 Na4 h6?! 16 Bxf6 gxf6 17 bxc4 dxc4,
Ax.Smith-E.Berg, Swedish Championship blitz play-off, Falun 2012, White could have taken a free pawn
with 18 Bxc6) 14 Rfd1 Rac8 15 Na4 c5 16 Bd2 Qb5 (16 ... Qc7 17 Bf4 Qa5 18 Nxc5 Bxc5 19 dxc5
Qxc5 20 Be5 d4 21 Bxf6 gxf6 is a playable alternative, as in I.Sokolov-B.Tadic, Bosnian Team
Championship 2009; ½-½ in 30) 17 bxc4 cxd4! has led to a couple of excellent Black wins:

18 Qxh7+ (or 18 cxb5? Rxc2 19 Be1 Ne4 20 Bf1 Bf6 21 a3 Rb8 22 Rab1 Nc3 23 Nxc3 dxc3 24
Rbc1 Rxc1 25 Rxc1 Rc8 and White resigned in M.Rodshtein-A.Kolev, Albena 2011) 18 ... Kxh7 19 cxb5
Ne4 20 Ba5 Rd7 21 Rac1 Rxc1 22 Rxc1 Rb7 23 Rb1 Bd7 and now, rather than 24 f3 with an unclear
game, White blundered with 24 e3?? Rxb5 25 Rxb5 Bxb5 26 Nb2 Ba3 27 Nd1 Be2 (White can save the
piece, but at too high a cost) 28 f3 Nd6 29 Nf2 Nc4 30 Be1 dxe3 31 Nh3 d4 32 Nf4 d3 33 Nxe2 d2 34
Bxd2 exd2 35 Nc3 Ne3 and White resigned in E.Bacrot-A.Grischuk, FIDE Rapid Grand Prix, Dubai
2002.
11 ... Qa5

This is the move we will focus on. 11 ... h6 is recommended by Aagaard and Ntrilis, which
normally transposes into the lines after 11 Nxc6 bxc6 12 b3 h6 above.
12 Qd2
12 Bd2!? was a fresh attempt from a former FIDE world champion: 12 ... Bb4 (12 ... Qb6!? looks
logical, trying to exploit the lack of the defence of the d4-pawn; for instance, 13 bxc4 Qxd4 14 cxd5 Nxd5
15 Nxd5 Bxd5 and Black must be close to equality) 13 Nxc6 bxc6 14 Qc2 Qa6 15 Nxd5! Nxd5 16 bxc4
with some advantage for White in R.Kasimdzhanov-S.Citak, Turkish Team Championship 2010 (½-½ in
38).

12 ... Rad8
Black brings a rook to the d-file. He chooses this rook not for strategic purposes (while one rook
certainly belongs on d8, its colleague would be better off on c8 or b8 rather than e8), but to avoid White
being able to take with the bishop on c6 with gain of time after b3xc4 and Nxc6.
After 12 ... Bb4 13 Nxc6 bxc6 14 Bxf6 gxf6 15 Rfc1 we reach a major position which has been
tested in several encounters between world-class players:

15 ... Rad8 (after 15 ... Rac8 16 bxc4 dxc4 17 a3 Bxc3 18 Qxc3 Qxc3 19 Rxc3 Rfd8 20 e3 c5 21 d5
Bxd5 22 Bxd5 Rxd5 23 Rxc4 Rb8, White has the better endgame, as reflected in Wang Hao-S.Iuldachev,
Asian Team Championship, Vishakapatnam 2008, and Mi.Hoffmann-K.Kiik, European Cup, Ohrid 2009 )
16 bxc4 dxc4 17 e3 and now Black has a choice:

a) 17 ... c5 18 d5! Bxc3 19 Rxc3 Bxd5 20 Bxd5 Rxd5 21 Qc2! Rc8 (or 21 ... Rb8 22 Rxc4 Rb4 23
Rb1 Rxb1+ 24 Qxb1 Qb6 25 Qc2 and White went on to win in S.Mamedyarov-V.Akobian, Gibraltar
2012; 1-0 in 104) 22 Rxc4 Qd2 23 Rc1 Qxc2 24 R1xc2 led to an absolutely model endgame performance
in B.Gelfand-A.Grischuk, Russian Team Championship 2004 (1-0 in 45). Anyone wishing to study this
ending in detail is referred to Gelfand’s wonderful best games collection for instructive coverage.
b) 17 ... f5!? 18 Bxc6 f4 is a much more aggressive approach. The only problem is that, as far as I
can see, it doesn’t quite work: 19 a3 fxe3 20 Qxe3 Bxc3 21 Rxc3 Rd6 was W.So-V.Akobian, Wijk aan
Zee 2010, and now 22 d5!? Bxd5 23 Bxd5 Qxd5 24 Qe2 would leave White a pawn up with good
winning chances.

13 bxc4
This move allows a shot which was thought to equalize completely. In this game Sargissian has a
new idea, but it seems rather tame.
Otherwise White mostly plays 13 Nxc6 (13 Rfd1?! Nxe5 14 dxe5 d4! was already good for Black in
H.Tikkanen-E.Berg, Gothenburg 2012) 13 ... bxc6 and then:

a) 14 Na4 Qxd2 15 Bxd2 was M.Brodsky-V.Ikonnikov, Le Touquet 2007, and now Black took on
b3, but I think 15 ... Bf5 would have equalized.
b) 14 Qc2 leaves White a tempo down on a line covered in the notes to 11 Nxc6 above. However,
as noted there, with the queen on c2 the most logical locations for the black rooks are d8 and c8, so Black
is perhaps slightly disorganized here. Nevertheless, after 14 ... h6 15 Bd2 Qa6 16 e3 c5 17 dxc5 Bxc5,
Black had at least equalized in Zhao Xue-E.Kovalevskaya, Women’s World Blitz Championship, Batumi
2012.
c) 14 bxc4 dxc4 15 Bxc6 Bb4 16 Rac1, as in M.Gurevich-S.Galdunts, French Team Championship
2006, is well met by 16 ... Qb6 17 d5 Rxd5!! 18 Bxd5 Nxd5 19 Be3 Qb7 20 Rb1 Qa8, regaining the
material with a decent game.
d) 14 Rfd1 Bb4 15 Bxf6 gxf6 16 Rdc1 leads to another position where Black is a whole tempo up
on a known line (see the notes to 12 ... Bb4). After 16 ... c5 (a logical attempt to make use of the extra
move; instead 16 ... Ba3 17 Rd1 Bb4 18 Rdc1 Ba3 was agreed drawn in Bu Xiangzhi-Wang Rui, Beijing
2005, but White could have played on with 19 Rc2 Bf5 20 e4, when I haven’t found a way for Black to
equalize) 17 bxc4 dxc4 18 d5 Bxc3 (18 ... Bxd5 19 Bxd5 Bxc3 was Zhu Chen-E.Kovalevskaya, FIDE
Grand Prix, Nalchik 2011, and now I like 20 Qxc3 Qxc3 21 Bxf7+ Rxf7 22 Rxc3 with a difficult endgame
for Black) 19 Qxc3 Qxc3 20 Rxc3 Bxd5 21 Bxd5 Rxd5 22 Rxc4 Rb8, a typical endgame arose in
A.Matthiesen-Kar.Rasmussen, Aarhus 2010. Black’s extra tempo isn’t keenly felt but it has enabled him
to start his counterplay earlier with ... Rb8. All the same, White still seems a bit better.
 Exercise: Black to play. Try and find a way to destroy White’s centre.

Answer: 13 ... Nxd4!

A great shot which seems to equalize. 13 ... Bb4 transposes into the note to Black’s 12th (after Nxc6
and Bxf6) and was in fact the move order used in several of the games featured there.
14 cxd5!?
A new attempt by the Armenian Super-GM.
Instead, 14 Qxd4 dxc4 15 Qe3 Bc5 16 Qf4 Qxc3 17 Bxf6 gxf6 18 Qxf6 leads to an interesting
position, but interesting only for Black! Now, rather than 18 ... Bd4?! as in V.Malakhatko-M.Meinhardt,
Paris 2005 (1-0 in 39), Black has two better options: forcing an immediate draw with 18 ... Bd6 (when
White has nothing better than delivering perpetual from g5 and f6); or playing for a win by 18 ... Be7!? 19
Qxe7 Qxe5 with an unclear game, but one where the black c-pawn looks like a major trump.
14 ... Bb4
14 ... Qxc3 15 Qxc3 Nxe2+ 16 Kh1 Nxc3 17 dxe6 fxe6 18 Bxb7 gives White something to play for,
with the bishop pair and a superior structure.
15 Bxf6 gxf6 16 Qxd4 Bxc3
 17 Nc4
White’s only attempt for advantage. 17 Qh4 Bxd5 18 Qxf6 leads to a draw after 18 ... Bxg2 19 Qg5+
Kh8 20 Qf6+.
17 ... Bxd4 18 Nxa5 Bxa1 19 dxe6 Bc3 20 e7 Bxa5 21 exd8Q Rxd8 22 Rb1
A nice touch, trying to interfere with Black’s control of c7.
22 ... b6 23 Rc1 Rd7

Given that Sargissian’s day job is analysing Marshall endgames with Levon Aronian, he must have
had some view on his winning chances here. Certainly White can never lose and can play for a long time.
However, it is hard to see the final result being anything other than a draw – a result which would be
immediately agreed after an exchange of rooks.
24 Bh3 Re7 25 e3 Bb4 26 Rc4 Bc5 27 Bf5 h6 28 Kg2 Re5 29 Rg4+ Kf8 30 Bc2 Rd5 31 Kf1 Rd2
32 Rc4 Rd5 33 Ke2 Rh5 34 Be4 Re5 35 Bd3 Rh5 36 Ra4 a5 37 Rf4 Kg7 38 h4 Rd5 39 a4 Rd8 40 Rf5
Rd7 41 g4 Be7 42 f4 Rc7 43 Rd5 Kf8 44 Bb5 Rc2+ 45 Kf3 Rc3 46 Rd7 Ra3 47 Ke4 Bc5 48 Kf5 Rxe3
49 Kxf6 Be7+ 50 Kf5 Bxh4 51 Rb7 Re6 52 Bc4 Rf6+ 53 Ke4 Bg3 54 f5 Bh4 55 Bd5 ½-½
Game 9
A.Onischuk-V.Akobian
Merida 2008

1 c4 e6 2 d4 d5 3 Nf3 c5 4 cxd5 exd5 5 g3 Nc6 6 Bg2 Nf6 7 0-0 Be7 8 Nc3 0-0 9 Bg5 c4 10 Ne5 Be6
11 f4

This is perhaps the most ambitious move at White’s disposal, but Black seems to get decent
counterplay. Moving pawns in front of your own king means that you need to be accurate for the rest of the
game.
a) 11 e3 seems to give Black several ways to equalize:

a1) 11 ... Nd7 12 Nxc6 bxc6 13 Bxe7 Qxe7 leaves Black with a very solid pawn chain and no
problems: 14 b3 (14 Na4 Rab8 15 Qc2 Rb4 16 b3 Rfb8 gave Black excellent queenside counterplay in
H.Stefansson-V.Akobian, Lubbock 2008; 0-1 in 37 ) 14 ... Nb6 15 Qc2 Rfc8 16 Rfc1 Rab8 17 bxc4 (17
Rab1 was V.Burmakin-E.Kanter, Dubai 2013, and now I like 17 ... g6, preparing ... Bf5 ) 17 ... Nxc4 18
Na4 g6 19 Bf1 Rb4 20 Nc5 was A.Adly-V.Kotronias, World Team Championship, Bursa 2010, when
Black can hold the balance with 20 ... Rb2 21 Qc3 Qf6 22 Rc2 Rxc2 23 Qxc2 Bf5 24 Qc3 Nd6, reaching
a comfortable Carlsbad structure – the black knight is excellently placed on d6, as known from the classic
games in the Queen’s Gambit Declined Exchange Variation: M.Bobotsov-T.V.Petrosian, Lugano
Olympiad 1968 (0-1 in 41), and L.Portisch-G.Kasparov, World Cup, Skellefteå 1989 (0-1 in 62).
a2) 11 ... h6 12 Bxf6 Bxf6 and now:

a21) A great demonstration of Black’s chances is 13 Nxc6 bxc6 14 Qc2 Rb8 15 Rfd1 Qa5 16 Qa4?!
Qxa4 17 Nxa4 Rb4 with excellent queenside play for Black in A.Beliavsky-I.Jelen, Slovenian
Championship, Sentjur 2013.
a22) 13 f4 Ne7 14 Qd2 (14 b3, as in D.Flores-P.Barrionuevo, Moron 2012, can be met simply by 14
... Rc8) 14 ... Qa5 15 Qc1 Rad8 and Black had no cause for complaint in A.Schenk-M.Petr, German
League 2012 (½-½ in 45).
b) The principled 11 e4? is in fact a serious mistake.

Exercise: How should Black respond?

 Answer: 11 ... Nxe4! 12 Nxc6 bxc6 13 Bxe7 Nxc3 14 Bxd8 Nxd1 15 Be7 (15 Raxd1 Raxd8 simply
leaves Black a pawn up) 15 ... Nxb2! 16 Bxf8 Kxf8 (in this position Black has two pawns for the
exchange and a serious advantage: the protected passed pawn on c4 is a huge asset, and the white rooks
struggle for scope) 17 Rfc1 Rb8 18 Bf1 (or 18 Rc3 a5 19 Ra3 a4 20 f4 Ke7 21 Kf2 Kd6 22 Ke3 c5 23
Bf3 Re8 24 dxc5+ Kxc5 and White resigned in J.Burnett-P.Kiriakov, British League 2005) 18 ... Ke7 19
Rc3 a5 20 f3 Kd6 (the immediate 20 ... c5 21 dxc5 d4 was even stronger) 21 Rb1 Bf5 22 Rbc1 c5 23
dxc5+ Kxc5 and Black was winning in H.Callam-S.Collins, European Cup, Rogaska Slatina 2011, since
White has no long-term defence to the plan of carefully pushing the c- and d-pawns.
11 ... Ng4

This is the consistent choice of the strongest players with Black. Alternatively, 11 ... h6!? is again
recommended by Aagaard and Ntirlis and has scored well.
12 Nxg4
12 Bxe7 Nxe7 13 Qd2 doesn’t pose serious problems for Black; for instance, 13 ... f6 14 Nxg4 Bxg4
15 b3 cxb3 16 Rfb1 Qd7 17 Rxb3 was about level in M.Krysztofiak-K.Miton, Polish Championship,
Warsaw 2011 (½-½ in 24).
12 ... Bxg4 13 Bxd5
As Scherbakov notes, the problem with 13 Nxd5 is 13 ... f6! (as yet unplayed; all games which
reached this position proceeded with 13 ... Bxg5?!) 14 Bh4 (14 h3! Bd7 15 Nxe7+ Qxe7 16 d5 is better,
although after 16 ... Qc5+ 17 Kh2 fxg5 18 dxc6 Bxc6 Black is certainly not worse) 14 ... g5 15 h3 Bd7 16
Nxe7+ Qxe7 17 d5 and now Black is better after, for instance, 17 ... gxh4 18 dxc6 Qc5+.
13 ... Bxg5 14 fxg5 Qxg5 15 Rf4!
 White now has a 2-0 central majority but Black can get rapid counterplay down the central files.
15 ... Rad8
15 ... Be6 is a common alternative, but I prefer to bring a rook into play than move a well-placed
bishop.
16 Qd2 Bh3!?
This has become established as the main line. Black threatens the d5-bishop and aims to provoke e2-
e4, after which the white king has even less pawn cover. The position is sharp but I think Black has a
good share of the chances.
17 e4 Qe7
This looks sensible. 17 ... Be6 is also possible (since 18 Bxe6, trying to win material, fails to 18 ...
Nxd4!), but after 18 Raf1 Black probably has nothing better than 18 ... Qe7 anyway.

18 Qf2
18 Rd1! is a good move, overprotecting the key d4-pawn. After 18 ... Be6 19 Bxc6 bxc6 20 d5 (20
Na4 is interesting, when Black can try 20 ... f6 21 Nc5 Bf7) 20 ... Qb7 21 Qf2 Bd7, White had some
advantage in N.Javanbakht-A.Ismagambetov, Al Ain 2012 (½-½ in 52).
 18 ... Be6 19 Rf1
A brave decision, leaving the dangerous knight on the board. 19 Bxc6 bxc6 was sounder, though
Black will always have counterplay against the white d-pawn – the black pawn moving from b7 to c6 has
taken a lot of dynamism out of the white centre.
19 ... Nb4! 20 Bxe6

Exercise: Black to play!

20 ... Qxe6?
Answer: 20 ... fxe6! would leave White struggling to hold his centre together.
21 d5
Attacking the queen and hitting the a7-pawn. Black should probably try 21 ... Qb6, though the
endgame is no picnic. In the game he sacrificed a pawn for activity, but it wasn’t enough.
21 ... Qe7 22 Qxa7 Nd3 23 Rf5 g6 24 Rf6 Nxb2 25 Qd4 Nd3 26 Qxc4 Ne5 27 Qd4 Rd6 28 Rxd6
Qxd6 29 Nb5 Qd7 30 Qxe5 Qxb5 31 Qd4 Rc8 32 d6 Qd7 33 Rd1 h5 34 Kg2 h4 35 Qd3 h3+ 36 Kg1
Qa4 37 d7 Rd8 38 Rd2 Qc6 39 Kf2 Qb6+ 40 Kf3 Qf6+ 41 Ke3 Qg5+ 42 Ke2 Qg4+ 43 Kf2 Qe6 44
Qd5 Qf6+ 45 Ke3 Qb6+ 46 Kf3 Qg1 47 Kg4 Qf1 48 Qd3 Qg1 49 Kxh3 Kg7 50 Qd4+ f6 51 Qxg1 1-0

Key Notes
1. 9 Bg5 c4 leads to fundamentally different positions to the Classical Tarrasch line 9 ... cxd4. Black
has found a lot of resources here and White needs detailed preparation to have a chance at obtaining an
advantage.
2. In Sargissian-Halkias (Game 8) White uses the safest and most classical plan, aiming to break the
black queenside and central pawn chain. Often an endgame results where White has a superior structure
but Black has a lot of activity and, on occasion, relies on the drawish properties of rook and pawn
endgames.
3. Onischuk-Akobian (Game 9) shows a more ambitious approach by White, playing for a kingside
space advantage and a 2-0 central majority. While this plan is very dangerous, we can see from the game
and notes that Black always has counterplay against the white centre and, sometimes, the white king.
 Chapter Three
9 Bg5 Be6
1 d4 d5 2 c4 e6 3 Nc3 c5 4 cxd5 exd5 5 Nf3 Nc6 6 g3 Nf6 7 Bg2 Be7 8 0-0 0-0 9 Bg5 Be6

Black develops a piece while protecting a central pawn, a concept known to be good from the
moves 1 e4 e5 2 Nf3 Nc6 (and some other openings). So why does nobody play this anymore? Well,
because everyone “knows” that the line is refuted by a long forcing sequence, resulting in an endgame
which Spraggett lost against Yusupov in their Candidates match in 1989. (The same endgame was played
by Spassky to make a very easy draw against Petrosian in their 1969 World Championship match, but
maybe people have forgotten about that.)
The various nuances which might affect the above assessment – Yusupov is a great player,
especially strong in the endgame, and played almost perfectly, which included deploying a strong novelty;
Spraggett’s defence could have been considerably improved; and most of us aren’t facing World
Championship Candidates when deciding what line to play in the next club match – don’t appear to have
been sufficiently influential to rescue this line. So I hope that Kunte-Aravindh (Game 10) is some good
PR for this line, as well as a demonstration that the technical “conversion” of this heavily simplified
position is no walk in the park for a normal GM. Incidentally, White is not forced to go into this endgame
(I’m not sure if I’d do so myself, given the chance) and his alternatives are also covered.

Game 10
A.Kunte-C.Aravindh
New Delhi 2013

1 Nf3 d5 2 c4 e6 3 g3 Nf6 4 Bg2 Be7 5 d4 0-0 6 0-0 c5 7 cxd5 exd5 8 Nc3 Nc6 9 Bg5 Be6
 Question: What are the pros and cons of this move?

Answer: Pros: Black develops his last minor piece and defends the crucial d5-pawn, thus taking the
sting out of White’s threats of Bxf6 and d4xc5 (in some order), followed by capturing on d5. In so doing,
Black maintains the tension, rather than resolving it with 9 ... c4 or 9 ... cxd4. As a general principle,
maintaining the tension is often the most ambitious continuation, since resolving it always comes with
concessions (for instance, in this particular position 9 ... c4 takes the pressure off the white d4-pawn,
while 9 ... cxd4 brings the white knight to a strong post on d4 and unleashes the g2-bishop on the long
diagonal). However, the big question when seeking to maintain the tension is whether a player can get
away with it or not! Whether Black can get away with such ambition is something we will discuss in the
further course of the game.
Cons: The bishop on e6 could be classed as passive, in that it has a defensive function (as opposed
to a more active deployment on g4, from where it would increase the pressure on White’s d4-pawn). In
addition, the bishop can be a target for exchange – if White gets his knight to d4, Black will need to
consider the consequences of Nxe6. This exchange, as we have seen, is by no means only in White's
favour – he gains the two bishops, but Black’s centre is strengthened and he gets the half-open f-file to aid
his attack.
The last con in the position was hinted at above: maintaining the tension always begs the question as
to whether a player can get away with it. Here, White has a liquidating sequence which leads to an
endgame in which he has scored 75% over a large number of games – although, as we will see, I believe
this position is quite playable for Black.
10 dxc5
Australian GM Ian Rogers notes that, although the Encyclopaedia of Chess Openings gives White a
small advantage after 10 Rc1, 10 ... c4! leaves the rook misplaced on c1. This looks like an improved
version for Black of the 9 ... c4 variation, since ... Be6 is necessary whereas Rc1, although sometimes
useful, is not invariably required.
10 ... Bxc5
 11 Bxf6
Starting a forced sequence resulting in a slightly favourable endgame for White. However, there are
several alternatives:
a) 11 Rc1 Be7 (11 ... Bb6?! 12 b3!, threatening Bxf6 and Nxd5, leads to a White advantage, as in the
stem game A.Rubinstein-Em.Lasker, Berlin 1918, and in V.Korchnoi-J.Piket, 3rd matchgame, Nijmegen
1993) 12 Nd4 h6 13 Be3 Qd7 14 Qa4 Rfc8 (or 14 ... Bh3, as in V.Korchnoi-J.Nunn, Hastings 1975/76,
“with an acceptable game” – Kasparov) 15 Rfd1 Ng4? was Kir.Georgiev-E.Berg, Gibraltar 2008, and
now 16 Nxe6! would have won material, but 15 ... Bh3 would have been fine.
b) 11 Ne1 d4 12 Bxf6 Qxf6 13 Ne4 Qe7 14 Nxc5 Qxc5 led to success for White in a game between
the inventors of both the Tarrasch and the g3-system against it: A.Rubinstein-S.Tarrasch, Carlsbad 1923
(1-0 in 45). Nevertheless, Black looks fully equal here. We will see more examples of how to handle this
structure in the chapter on 9 dxc5.
c) 11 Na4 was played in M.Tal-P.Keres, Bled/Zagreb/Belgrade Candidates 1959, where Keres won
a famous game after 11 ... Bb6 12 Nxb6 axb6. On 11 ... Be7 12 Be3,

Keres suggested that “White could have created strong pressure on the c5-square and secured an
enduring initiative without any risk”. In fact this position more commonly arises with White to move (in
the variation 9 dxc5 Bxc5 10 Na4 Be7 11 Be3 Be6), so it is not surprising that Black, a whole tempo up
on a known variation, has few difficulties here. After 12 ... Re8 (12 ... Bg4 gets Kasparov’s endorsement:
13 Rc1 Re8 14 Nc5 Bxc5 15 Bxc5 Ne4 16 Be3 Qd7 with level chances in E.Geller-B.Spassky, 7th
matchgame, Riga 1965; ½-½ in 41) 13 Nd4 Bg4 14 Nxc6 bxc6 15 Rc1 Qd7 16 Re1 and Black had no
problems in L.Guidarelli-D.Brandenburg, Vlissingen 2010 (½-½ in 31). Kasparov also gives the
variations 12 Nd4 Nxd4 13 Qxd4 Qa5 14 Rfc1 Rfc8 or 14 Nc3 Rac8! and Black is fine.
11 ... Qxf6 12 Nxd5

Neither player has a decent way to avoid the following forcing line.
12 ... Qxb2 13 Nc7 Rad8 14 Qc1 Qxc1 15 Raxc1 Be7
Alternatively, 15 ... b6 16 Nxe6 fxe6 is an endgame which has been extensively defended by German
GM Lutz Espig. According to my database, he has made a remarkable +2 =9 -0, which is rather amazing
since Black should never win such a position! The most famous example here is the 16th game of the 1969
Petrosian-Spassky World Championship match in Moscow, where after 17 Rc4 (several commentators,
including Yusupov, have suggested that White retains better chances of an edge with 17 e3!? or 17 Bh3!?)
17 ... Nd4! 18 Nxd4 Rxd4 19 Rxd4 Bxd4, the trade of a pair of rooks and knights had considerably eased
Black’s defence. The game concluded 20 e3 Bc5 21 Rd1 Rf7 22 Be4 g6 23 h4 Rc7 24 Kg2 Be7 25 Kf3
Kg7 26 Rd2 Bb4 27 Rc2 Rxc2 28 Bxc2 Kf6 29 Bd3 h6 30 Be4 g5 31 h5 Be1 and a draw was agreed.
16 Nxe6 fxe6
 Question: What’s your assessment of this endgame?

Answer: One thing is clear – Black can’t be better! White has the superior pawn structure which
guarantees him a small but enduring advantage. The stats from this position are strongly in White’s favour,
suggesting that his position is considerably easier to play.
On the other hand, the game has been considerably simplified. While Black, admittedly, has a
weakness on e6 his position is otherwise quite solid, and the “principle of two weaknesses” suggests that
White will need to create another target before generating serious winning chances. The presence of
opposite-coloured bishops is also a potentially significant factor: if the rooks and knights are traded the
players would agree to an immediate draw. However, it is important not to overestimate the drawing
tendencies of the opposite-coloured bishops – if White manages to generate an initiative on the light
squares it will be hard for Black to defend. Black has no f7-pawn, and if the pawn on e6 is lost or moved
to e5, Black will be severely weakened on the a2-g8 diagonal which the white bishop could likely occupy
(for instance, Bd5, anchored by a white pawn on e4, is possible). Conversely, White would need to be
very careless to allow Black similar chances on the a7-g1 diagonal, since white pawns on f2 and e3
would stunt the black bishop’s activity.
We should also mention that White has a kingside majority and might consider mobilizing it with f2-
f4 and e2-e4, although, for the reasons given above, I would be very careful to check that Black can’t
generate play on the dark-squares before moving my f-pawn.
17 Rc4!
 Question: Why is this a strong move?

Answer: Nimzowitsch defined a strong positional move as one which advances your own plans
while preventing your opponent’s. This move ticks both boxes. The rook is well placed on c4, where it
can move along the fourth rank to target the black pawns (by Re4, for instance). The way is also cleared
for White’s king’s rook to come to b1 or c1. In the meantime, the rook prevents one of Black’s ideas,
namely further simplification with ... Nd4, after which he would be one trade closer to a drawn opposite-
coloured bishop endgame.
The most famous game to feature 17 Rc4! is probably Yusupov-Spraggett, Candidates (3rd
matchgame), Quebec 1989, which ended in a win for White after 42 moves. However, Yusupov has
tremendous technique even by GM standards, and the line hardly deserves to be consigned to the dustbin
just because he won a good game. Chess fashion is rather fickle and variations are often discarded simply
because, for instance, Karpov won a precise technical game. Positions which are unplayable against the
elite can often be used with success against lesser mortals and even, as in this game, against good
grandmasters.
17 ... Bf6
 18 Bh3
Instead:
a) 18 Rb1 is one of the most popular moves in this position. However, after 18 ... Rd7 19 h4, Black
can follow Spassky’s idea from his match against Petrosian and simplify with 19 ... Nd4!?, as in
L.Ftacnik-N.Minev, Bucharest 1978. 18 ... Rd6!? is also possible since the pawn on b7 is poisoned.
Black has good equalizing chances here, as for instance in A.Beliavsky-M.Illescas Cordoba, Linares
1988 (½-½ in 50).
b) 18 e3! was the novelty Yusupov unleashed against Spraggett in Game 3 of their 1989 Candidates
match. Yusupov explains the idea behind the move as follows: “Realizing the disadvantages of
exchanging rooks, White doesn’t hurry to play Rb1. In fact, it is not yet clear whether this move is needed
at all (the rook may turn out to be more useful on c1). The move played, which I had prepared to use
against Illescas [in Linares 1988 – SC], proved to be a theoretical novelty, although hardly anything could
be more logical – White takes control of the d4-square, on which unwelcome exchanges could otherwise
take place.”
The game continued 18 ... Rd6 19 h4 h6 20 Re4 Rfd8 (a more recent attempt also ended badly for
Black: 20 ... b5 21 Rc1 a6 22 Bh3 Kf7 23 Rxe6 Rxe6 24 Bxe6+ Kxe6 25 Rxc6+ Kd5 26 Rxa6 and White
was winning in S.Brunello-D.Contin, Italian Championship, Martina Franca 2007; 1-0 in 46) 21 Bh3 Kf7
22 Kg2 Re8 23 Rc1 Re7 24 Rc2
 24 ... b6? (this move significantly weakens the c6-knight, and it is indicative of how difficult it is to
play such passive positions that someone of Spraggett’s class would make such a mistake; instead, Rogers
suggests 24 ... g6, intending to mark time with ... Kg7-f7 and challenge White to find a way to break
through) 25 Rf4! (creating an immediate threat of Rxc6 and Ne5+) 25 ... Kg6 (Rogers says that moving the
king to e8 or g8 would have been better, although White is still on top after 26 g4) 26 g4! Ba1 (Yusupov
gives 26 ... Ne5 27 g5 hxg5 28 hxg5 Nd3 29 gxf6 Nxf4+ 30 exf4 with a clear advantage to White, for
instance 30 ... gxf6?! 31 f5+!) 27 Rc1 Bb2 28 Rc2 Ba1 29 a4 Ne5 30 Nxe5+ Bxe5 31 Rf8 (White now has
a devastating attack on the black king) 31 ... Rdd7 32 f4 Rc7 33 Rd2 Bc3 34 Rd6 Kh7 35 g5 hxg5 36 hxg5
Bb4 37 Rdd8 (Yusupov notes that 37 g6+! was immediately decisive: 37 ... Kxg6 38 Bf5+ or 37 ... Kh6
38 Bf5!) 37 ... Kg6 (37 ... g6 is better) 38 Kf3 Rf7 39 Rh8 e5 40 Bg4 exf4 41 Rd5 fxe3+ 42 Kg3 and
Black resigned in view of imminent mate.
18 ... Rd6 19 Rb1 Rf7 20 Re4 Re7 21 a4 g6 22 e3 Kg7 23 Bf1 Nd8

That’s some solid defence of e6! As mentioned in previous notes, White needs to generate another
target to create genuine winning chances.
24 h4 h6!
 There is no need to allow the white knight access to g5.
25 Rc4 Nc6 26 Re4 Nd8 27 Rc1 Rc6 28 Rxc6 bxc6
Rather a committal decision, albeit one which worked out well in the game. I don’t see anything
wrong with the straightforward 28 ... Nxc6. Every trade favours Black and he still seems rock-solid here,
even if White can play on with no risk.
29 Bd3 Nf7 30 Rc4 Rc7

Exercise: How can White improve his position?

31 Nd4?!
As will be seen in the game, this allows a pretty clear equalizing sequence.
Answer: Pretty much any neutral move would be okay, but White had a couple of ideas to maintain
the pressure:
a) 31 Rg4 g5 32 hxg5 hxg5 33 Bc4 and White is obviously better here, although in view of the
simplified position and the drawish properties of the remaining pieces, my money would be on Black
holding on for a draw.
b) 31 Rc5!?, taking control of some useful squares on the fifth rank and preparing Be4. (The
immediate 31 Be4 isn’t so convincing because of 31 ... c5, and Black is fine.)
 Exercise: How can Black bring the position to safety?

Answer: Part of the technique of defending passive, slightly worse positions is being able to
accurately calculate when you can transform the position, for instance through a trade or a bid for
counterplay.
31 ... Bxd4!
Black is alert and realizes that he can either get counterplay or trade off his weak e6-pawn.
Admittedly, in view of the fork by the knight on d4, Black didn’t have very much choice since 31 ... Nd8
was not attractive.
32 exd4
32 Rxd4 is the move White wants to play, but then Black gets counterplay with the c-pawn; i.e. 32 ...
c5! 33 Re4 Kf6 and Black is fine, given his ideas of ... g6-g5, ... Nd6 and, if appropriate, ... c5-c4 (though
this last move needs to be considered carefully, since the additional threats created by pushing the passed
pawn must be balanced against the pawn’s increasing vulnerability and the weakening of the d4-square).
White has nothing better than the repetition available after 34 Rf4+ Kg7 35 Re4 Kf6.
32 ... e5
 33 Be4
The sensible move. After 33 dxe5 Nxe5 34 Rc3 Nxd3 35 Rxd3 c5, if anyone has chances in the
resulting rook and pawn endgame, it’s not White.
33 ... exd4 34 Rxc6
If White wanted to play on, his last chance was 34 Rxd4. The computer proposes that White still has
an edge, though I think Black isn’t worse at all, given his passed pawn and the simplified nature of the
position. After 34 ... g5 35 hxg5 hxg5 White has no reason to claim any advantage.
34 ... Rxc6 35 Bxc6 g5!

Trading off more pawns brings the position closer to safety.

36 hxg5 hxg5 37 f4 gxf4 38 gxf4 ½-½
A very solid draw with Black, despite the 150+ rating point gap, and a good advertisement for the
solidity of this system.
Key Notes
1. 9 Bg5 Be6 is considerably better than its reputation!
2. If White chooses to go for the forced line, resulting in an endgame, Black needs to be patient to
hold his slightly inferior position. It isn’t obvious to me that Black’s game is any worse than those
enthusiastically defended by the world elite on a daily basis.
3. White has a range of alternatives on move 11 which keep more tension. Black should be familiar
with the main themes from the 9 Bg5 cxd4 variation, as similar positions often arise here.
 Chapter Four
8 dxc5/9 dxc5
1 d4 d5 2 c4 e6 3 Nf3 c5 4 cxd5 exd5 5 g3 Nc6 6 Bg2 Nf6 7 0-0 Be7

Systems with d4xc5 at move 8 or 9 have a number of attractions from White’s perspective. White
defines the structure and saddles Black with an IQP, and accordingly does not need to worry about
alternative structures such as those which arose in the 9 Bg5 c4 line. Similarly, Black’s choice is
somewhat limited and so the theoretical workload required to play 9 dxc5 is much less than that involved
in 9 Bg5. White can take on c5 at any stage after move three.
8 Nc3
The alternative 8 dxc5 Bxc5 can result in a transposition after 9 Nc3 0-0, but also has independent
significance after White’s alternatives such as 9 a3, which is covered in Manolache-Jianu (Game 15).
8 ... 0-0 9 dxc5 Bxc5
9 ... d4!? is an enterprising gambit continuation which I have tried before and might try again. See
Chatalbashev-Sokolov (Game 14) for the details.
10 Bg5
This was historically the main continuation but now, following recommendations of authors like
Avrukh, you are just as likely to face an alternative here such as 10 Na4 or 10 a3. See Akopian-Jianu
(Game 13) for the details.
10 ... d4 11 Bxf6
11 Ne4 (and the similar 11 Bxf6 Qxf6 12 Ne4) are examined in Baburin-Collins (Game 12).
11 ... Qxf6 12 Nd5
 This is the old main line and is covered in Mulyar-Perunovic, the first game below.

Game 11
M.Mulyar-M.Perunovic
Chicago 2012

1 d4 d5 2 c4 e6 3 Nf3 c5 4 cxd5 exd5 5 g3 Nc6 6 Bg2 Nf6 7 0-0 Be7 8 dxc5 Bxc5 9 Nc3 0-0 10 Bg5 d4
11 Bxf6 Qxf6 12 Nd5 Qd8
12 ... Qf5!? is a pet line of Narciso Dublan and has been adopted by current Tarrasch exponents like
Aagaard, Swinkels and Carlstedt. However, I think 12 ... Qd8 is more reliable.
13 Nd2
An idea of Jan Timman, as far as I’m aware. White intends Rc1, putting the question to the c5-
bishop. Black has a choice of three squares – f8, b6 and a7. Each option seems fully playable to me.

13 ... Re8
 The rook is always good on this square and particularly so in this line, since there is no knight on c3
to hold the e2-pawn. Tactics against e2 are much more common here than in the 9 Bg5 variations.
13 ... a6 14 Rc1 Ba7 is the main alternative, after which White has a wide choice:

a) 15 Nb3 Qd6 16 Qd2 Re8 17 Nf4 Bf5 with level chances in J.Plachetka-J.Nunn, European Team
Championship, Skara 1980 (½-½ in 39).
b ) 15 Nc4 Bg4 16 Qd2 Rc8 17 h3 Be6 18 Nf4 Bf5 19 a4 was P.Wells-J.Cooper, British League
2006, and now the simple 19 ... Re8 20 b4 Qd7 equalizes.
c ) 15 a3 Bg4 16 Nf4 Rc8 17 Nd3 Re8 18 Bf3 Qd7 and Black had no problems in E.Matsuura-
C.Umetsubo, Maringa 2012 (½-½ in 41).
d) 15 b4 Bg4 16 Ne4 Re8 17 Re1 Re5! 18 Nf4 Qe7 and White was struggling to meet all the threats
in J.Leiva Rodriguez-C.J. Ruiz, Lima 2013 (0-1 in 40).
e) 15 Ne4 Kh8 16 Nf4 Qe7 17 Nd3 Bf5 18 Nec5 Rac8 19 Qd2 Rfe8 20 Rfe1 Nd8 21 b4 a5 and a
draw was agreed in R.Khusnutdinov-L.D.Nisipeanu, Pardubice 2012.
f) 15 Nf4 is the main line and has scored well for White in practice. However, the computer thinks
this position is completely equal:
 f1) 15 ... Bg4 16 h3 (16 Re1 Rc8 17 Ne4 was V.Epishin-G.Schnur, Werther 2000, when 17 ... Bb8
18 Nc5 Bxf4 19 gxf4 Qe7 20 a3 a5 leaves Black no worse – White also has to take care of his own
weaknesses here; instead, 16 Bf3 Bxf3 17 Nxf3 has been played twice by Petr Haba – in both games
Black opted for the pseudo-active 17 ... Qa5; a beginner would put rooks on the open files with 17 ... Re8
18 Qb3 Qd7 19 Rfd1 Rac8 which, fortunately, completely equalizes) 16 ... Bf5 17 Qb3 Rb8 18 g4 (18
Qa3 was P.Brodowski-M.Chlost, Olomouc 2010, and now it seems Black can get away with 18 ... g5!?
19 Nd3 Re8 20 Rfe1 Qe7 21 Qxe7 Rxe7 with no problems in the endgame) 18 ... Qg5 19 Qg3 Be6 20
Nxe6 fxe6 21 Ne4 Qb5 22 Qb3 Qe5 23 Kh1 Rf4 and Black was no worse in R.Kasimdzhanov-A.Mallahi,
Asian Championship, Tehran 1998.
f2 ) 15 ... Qe7 led to a draw by repetition after 16 Nd5 Qd8 17 Nf4 Qe7 in A.Moiseenko-
R.Ponomariov, Ukrainian Team Championship 1998.
f3) The prophylactic 15 ... Rb8 has been played numerous times and should be fine for Black.
However, I prefer 15 ... Bg4 since the queen’s rook might find a better square on c8 or d8.

14 Rc1
One of my happiest memories with the Tarrasch proceeded 14 Nb3 Bf8 15 Nf4 Bg4 16 Bxc6 bxc6
17 Qxd4 Bxe2 (17 ... Qxd4 18 Nxd4 g5 19 h3 Bg7 20 Rad1 Bc8 21 Nh5 is level) 18 Nxe2 Rxe2 19 Qc3
Qb6 20 Rad1 Rae8 21 Rd7 R8e7 22 Rd4?! (22 Rd6 held the balance) 22 ... g6 23 Rfd1 h5! (White’s
position is unpleasant since Black intends 24 ... Bg7; I turned down a draw somewhere around here but
was surprised by how quickly White’s game went downhill) 24 Qf3 Rxb2 25 Rd8 Ree2 26 R1d7 Rxf2 27
Rxf8+ Kxf8 28 Rxf7+ Kg8 29 Qxf2 Rxf2 30 Rxf2 Qe3 and Black soon won in A.Baburin-S.Collins, Dun
Laoghaire 2010 (0-1 in 35).
14 ... Bb6
14 ... Bf8 is also fully playable and has been endorsed by Grischuk amongst others.
15 Nc4
15 Re1 is another logical move:
 a) 15 ... Bg4 16 Nf4 (16 Nc4 Ba5 transposes to the main game) 16 ... d3!? 17 Nxd3 was tested in
two online games L.Van Wely-T.L.Petrosian in the European Blitz Championship 2004. Game 3
proceeded 17 ... Bxe2 (an improvement over the first game’s 17 ... Rxe2 18 Rxe2 Qxd3 when White was
clearly better; but 17 ... Nd4! 18 Bf1 Qe7 seems stronger, with a good game for Black) 18 Rxe2 Qxd3 19
Rxe8+ Rxe8 20 Nc4 Qxd1+ 21 Rxd1 and White enjoyed a small edge in the endgame (though 0-1 in 51).
b) In a game in California I mixed up my move order with 15 ... Ba5?! 16 a3! Be6 17 Nf4 Bg4 18 h3
(18 b4 Bc7 19 Nd3 seems better, with an edge for White, although this should be playable for Black) 18
... d3? (simply 18 ... Bf5 would have led to a normal position) 19 hxg4 dxe2 20 Rxe2 Rxe2 21 Qxe2
Qxd2 22 Qxd2 Bxd2.

Question: How do you assess this endgame?

Answer: In my analysis I had assumed that the endgame, with equal (and reduced) material and
opposite-coloured bishops, must be level. In fact Black is in for a difficult defence, since all the white
pieces are more active than their black counterparts. This game was a salutary lesson for me in the
Tarrasch. When you trade the isolated d-pawn, there is a natural tendency to relax since Black’s position
has no formal weaknesses anymore. However, it is vital to check whether the opponent’s activity can be
neutralized. In particular, and especially with a white bishop on g2, the b7-pawn is a key target which can
be difficult to defend. D.Zilberstein-S.Collins, Berkeley 2011, continued 23 Rc2 Rd8 (23 ... Bxf4 24 gxf4
Rc8 was a better try) 24 Bxc6 bxc6 25 Rxc6 g6 26 Ra6 Rd7 27 Kf1 with an extra pawn and a clear
advantage for White. I eventually lost a painful rook endgame (1-0 in 65).
15 ... Bg4 16 Re1

16 ... Ba5
16 ... Bc5 is an interesting alternative, preserving the dark-squared bishop after 17 a3 a6 18 b4 Ba7.
Now 19 Qd2 Rc8 20 a4 Be6 (20 ... h5! 21 Nf4 h4 22 Nd3 Be6 gave counterplay in D.Yevseev-
V.Yemelin, St. Petersburg 2006; ½-½ in 49 ) 21 Nf4 Bxc4 22 Rxc4 Ne5 23 Rxc8 Qxc8 24 Rc1 Qd7 is a
logical sequence leading to a position which has been successfully defended by a couple of Tarrasch
experts: 25 Qc2 d3 (25 ... Bb8 26 h3 b5 27 axb5 axb5 was L.Ftacnik-A.Berelowitsch, German League
2003, ½-½ in 50, but now 28 Qc5 would have been better for White) 26 exd3 (and here 26 Nxd3! Nxd3
27 Qxd3 Qxa4 28 Rc7 keeps more pressure) 26 ... Bd4 27 h3 g6. This position is tough for White to win,
especially against Grischuk, and a draw eventually resulted in L.Van Wely-A.Grischuk, Enghien les Bains
2001 (½-½ in 53). Nevertheless, Black was under pressure in these games after 20 ... Be6, and I prefer
Yemelin’s more purposeful approach with 20 ... h5!.
17 Nxa5 Qxa5
 Exercise: How can White fight for an advantage?

Answer: 18 b4!
This shot is a key resource. The white pieces gain a lot of activity and Black needs to be careful not
to lose his whole queenside.
18 ... Nxb4
18 ... Qxa2? was refuted the first time it was played, which hasn’t stopped this error being repeated
in over a dozen games since. A.Chernin-G.Giorgadze, Tallinn 1984, continued 19 Nc7 d3 20 Qxd3 Rad8
21 Bd5! Rxe2 22 h3!! Rxe1+ (Scherbakov gives 22 ... Ne5 23 Qc3 Rxd5 24 Nxd5 which wins, since
Black has no time to recapture in view of his porous back rank) 23 Rxe1 Be6 24 Qe4 Rxd5 25 Nxd5 Qb2
26 Nf4! Qxb4 27 Qxb4 Nxb4 28 Rd1! Nc6 29 Nxe6 fxe6 30 Rd7 and after a series of incredibly accurate
moves, White went on to win the endgame (1-0 in 49).
19 Qd2
Perhaps not the most dangerous move, though Black needs to be well prepared here.
19 Qxd4 Nxd5 20 Qxd5 Qxd5 21 Bxd5 Rad8 22 Bxb7 Rxe2 23 Rxe2 Bxe2 led to easy draws in
A.Miles-P.Konguvel, Sakthi 1996 (½-½ in 28) and E.Prokopchuk-V.Potkin, Russian Cup, Moscow 1999
(½-½ in 38), but 20 Qxg4 leads to a position which most GM commentators seems to agree is good for
White.
 Milov: “This position is slightly better for White due to the superiority of bishop over knight. [The]
knight is not stable in the centre and soon will have to retreat.”
Atalik: “Although theory is counting on this position like it is equal, in reality White is clearly better
due to the better piece and the liquidability of [the] queen’s flank pawns.”
Black usually plays 20 ... Rad8 (20 ... Nf6 21 Qc4 Rab8 led to a quick draw in E.Ghaem Maghami-
S.Mamedyarov, Baku 2003, ½-½ in 26, but this approach is passive and White must be better here) and
then:
a) 21 Bf3 Nf6 22 Qc4 b6 23 Red1 g6 led to draws in E.Inarkiev-V.Potkin, Elista 2001 (½-½ in 39)
and I.Saric-O.Zierke, Pula 2007 (½-½ in 44).
b) 21 Red1 Nf6 (not 21 ... Qxa2?? 22 Qd4 with a fatal pin; while Milov claims an edge for White
after 21 ... Nc3 22 Rxd8 Rxd8 23 Qc4) and now:

b1) 22 Qc4 b5 23 Qc2 is a position which should certainly be tenable for Black, but isn’t much fun.
Many strong GMs have happily gone for this as White and have caused a lot of problems. For example:
23 ... h6 (23 ... h5 24 Bf3 h4 25 Kg2 hxg3 26 hxg3 Rxd1 27 Rxd1 was better for White in V.Milov-
M.Haag, Suncoast 1999; 1-0 in 37) 24 Bf3 Rxd1+ (here 24 ... a6 25 Rxd8 Rxd8 26 Kg2 Qa3 27 h4 g6 28
Qc7 Qd6 29 Qxd6 Rxd6 30 e3 Kg7 31 Rc2 Nd7 32 Be2 Nb6 was agreed drawn in J.Nogueiras Santiago-
R.Vera Gonzalez Quevedo, Linares, Holland 1993, but White would have been clearly better had he
avoided the queen exchange with 29 Qb7 or 29 Qa7) 25 Rxd1 Rd8 (or 25 ... b4 26 Rd6 Rd8 27 Rc6 Qe5
28 Kg2 and Black was in for a difficult defence in S.Halkias-V.Akobian, World Junior Championships,
Yerevan 2000) 26 Rc1! (as explained by Atalik in a different variation, White should keep the rooks on in
this position) 26 ... b4 (26 ... Qd2 is worth considering) 27 Qc7 Qxc7 28 Rxc7 Rd7 29 Rc8+ Kh7 30 Rc5
Rd1+ 31 Kg2 Ra1 32 Rc2 g5 with equality in R.Kasimdzhanov-J.Bosch, Hoogeveen 1999.
b2) The direct 22 Rxd8 Rxd8 23 Qc8 gave White an extra pawn after 23 ... h6? in J.M.Hodgson-
A.Summerscale, French Team Championship 1999 (1-0 in 95) and 23 ... g6? in S.Atalik-V.Akobian, Los
Angeles 2003 (1-0 in 26). Instead, 23 ... b5 is better, keeping equal material, though I still prefer White in
this position since (amongst other reasons) I prefer to capture queens when they land on my back rank!
c) 21 Qc4 gives Black a choice:

c1) 21 ... g6 was successful in a recent game. 22 Qb3 (22 Red1 is a better attempt) 22 ... Nb4 23
Kf1 Nxa2 24 Ra1 Rd2 25 Qxb7 Qb6 26 Qxb6 axb6 27 Rab1 Rexe2 28 Rxe2 Rxe2 29 Rxb6 and a draw
was agreed in L.Ftacnik-P.Dias, Portuguese Team Championship 2012.
c2) 21 ... b5 22 Qb3 Nf6 23 Bc6 Re5 and now:
 c21) 24 Red1 Rd2 (Atalik assesses 24 ... Rxd1+ 25 Rxd1 h6 26 Bf3 Rc5 27 Kg2 as good for White)
25 Rxd2 Qxd2 26 Rd1 Qxe2 27 Rd8+ Ne8 28 Kg2 h6 (Atalik gives the nice variation 28 ... a5 29 Qc3
Re6 30 Bd7 Re7 31 Qc8 Qe4+ 32 Kh3 Kf8 33 Qc5! Kg8 34 Qxe7 Qxe7 35 Rxe8+ and wins) 29 Ra8 and
White was better in K.Spraggett-K.Gentes, Canadian Championship, Brantford 1999 (1-0 in 41).
c22) 24 Qb2 Rc5 25 Bf3 was S.Atalik-J.Prizant, Vladimir 2002, and now Atalik notes that 25 ...
Rdc8! would have equalized, since Black has no problems after all the rooks are exchanged: 26 Kg2
Rxc1 27 Rxc1 Rxc1 28 Qxc1 g6 with level chances.
19 ... Nc6

20 Qxa5
By far the most popular, but probably not the best. Instead, 20 Qg5!? is an underrated idea: 20 ... d3
21 Nf6+ Kh8 (after 21 ... Kf8, as in A.Fominyh-B.Miljanic, Budapest 1990, White should have interposed
22 Nxh7+ Kg8 23 Nf6+ Kf8 before trading queens) 22 Qxa5 Nxa5 23 Nxe8 d2 24 Nd6 Be6 25 a3 Rd8 26
Nxb7 Nxb7 27 Bxb7 left White with an extra pawn in M.Hörstmann-J.Blauert, German League 2004.
Black has reasonable saving chances after 27 ... dxe1Q+ 28 Rxe1 Rd2 29 Ba6 Ra2 30 Rd1 g5 31 Rd3,
but only White can try to win this position.
 20 ... Nxa5 21 Nc7
21 f3 d3 22 Red1? (after 22 Ra1 Bd7 a draw was agreed in A.Yusupov-M.Petursson, Reykjavik
1985 – a pretty easy draw with Black against the future world number three) 22 ... dxe2 23 Re1 Rac8 24
Rc7 Be6 25 Ne7+ Rxe7 26 Rxe7 Kf8 27 Rxe6 fxe6 28 Rxe2 with an extra pawn for Black in A.Dreev-
B.Ivanovic, Manila Interzonal 1990 (½-½ in 50).

Exercise: How can Black stay in the game?

Answer: 21 ... d3!

The only move, but fully sufficient.
22 exd3
Alternatively:
a) Ftacnik gives 22 Rb1 d2! with counterplay. Then 23 Red1 Bxe2 24 Rxd2 Rad8 25 Rbb2 Rxd2 26
Rxd2 Nc4 27 Nxe8 (after 27 Rd4 Rb8 White should bail out with 28 Re4 Bd3 29 Rd4 Be2) 27 ... Nxd2
was drawn in K.Spraggett-M.Leski, San Francisco 1987.
b) 22 Nxe8 d2 23 Nd6 has also been played. After 23 ... Rd8 24 Nxb7 Nxb7 25 Bxb7 Bxe2 26 Red1
dxc1Q 27 Rxc1, the players can shake hands with a clear conscience.
22 ... Rxe1+ 23 Rxe1 Rd8 24 h3
24 d4 Kf8 25 d5 was played in T.Fiedler-N.Michaelsen, German League 2002, when the cleanest
draw is 25 ... Rc8! 26 d6 Nc4 27 Bxb7 Rb8 28 Re4 Nxd6 29 Rxg4 Rxb7 30 Rd4 Rxc7 31 Rxd6; while
after 24 Nb5 g6 25 Nxa7 Rxd3 Black was at least equal in A.Frois-M.Illescas Cordoba, Saragossa 1991
(½-½ in 47).
24 ... Bd7
 The ending is level. In the game White loses the thread and goes on to lose.
25 d4 Kf8 26 d5?! Nc4! 27 Bf1 Nd6 28 Ne6+ Bxe6 29 dxe6 Re8 30 Rc1 Rxe6
White has decent drawing chances despite his pawn deficit, but didn’t manage to hold on.
31 Bg2 Re2 32 a4 Ke8 33 Rb1 b6 34 a5 bxa5 35 Rb8+ Kd7 36 Ra8 Re1+ 37 Bf1 Nc8 38 Kg2
Rxf1 0-1

Game 12
A.Baburin-S.Collins
Bunratty 2011

1 d4 e6 2 c4 d5 3 Nf3 c5 4 cxd5 exd5 5 g3 Nc6 6 Bg2 Nf6 7 0-0 Be7 8 Nc3 0-0 9 dxc5 Bxc5 10 Bg5 d4
11 Ne4

This is a quiet sideline, but one which has appealed to a number of GMs who are simply looking for
a normal, playable game against the Tarrasch. Being able to handle the quiet, level, simplified positions
arising in this variation is a key skill for Tarrasch players, since if Black is inaccurate then he might end
up worse in view of his compromised structure. Accordingly we will devote some attention to how Black
should play here.
11 Bxf6 Qxf6 12 Ne4 Qe7 13 Nxc5 Qxc5 is another quiet way for White to play:

a) 14 Qb3 Re8 15 Rfe1 Re7 16 Rac1 Qa5 17 a3 Bg4 18 h3 Bh5 19 g4 Bg6 20 Nh4 Rd8 21 Red1
Qg5 22 Nxg6 hxg6 23 Qd3 Ne5 24 Qg3 Qf6 was fine for Black in A.Onischuk-V.Akobian, Montreal 2009
(½-½ in 54).
b ) 14 Qc1 Qb6 15 Rd1 Rd8 16 Qc2 h6 17 Rd2 Be6 18 b3 Rd7 19 Rad1 Rad8 was balanced in
N.Vasovski-G.M.Todorovic, Skopje 2011 (0-1 in 57).
c) 14 Qd2 was tried against me in a very recent game: 14 ... Rd8 (perhaps 14 ... Bg4 15 Rfd1 Rad8
is simpler, as in P.Bailet-V.Jianu, Rennes 2013; ½-½ in 40) 15 Rfd1 Be6 16 Qf4 h6 17 a3 a5 18 Rac1
Qb6 19 Rd2 Rd5 (19 ... a4 is sensible) 20 Ne1 Rf5 21 Qe4 and now, instead of the 21 ... Rd8 of
B.Socko-S.Collins, Riga 2013 (which still left me with a decent position), I should have played 21 ...
Rb5 – the b2-pawn is falling and White needs to generate something; for instance, 22 Nd3 Bf5 23 Qf4
Bxd3 24 Rxd3 Rd8 25 Be4 Rxb2 26 Rf3 f6 27 Qf5 Ne5 28 Qh7+ Kf8 and White’s compensation is only
enough for a draw.
d) 14 Rc1 Qb6 and now:
 d1) 15 Qc2 h6 16 Rfd1 (16 a3, as in O.Cvitan-T.L.Petrosian, European Championship, Warsaw
2005, is best met by 16 ... Be6 17 b4 a5 with equality; while 16 Nd2 Be6 17 Nc4 Qa6 18 b3 Rfd8 was
already equal in D.De Vreugt-A.Berelowitsch, Bussum 2008, ½-½ in 40) 16 ... Be6 (16 ... Re8 17 Rd2
Be6 18 b3 Rad8 was level in B.Kurajica-H.Velchev, European Championship, Plovdiv 2010; ½-½ in 46 )
17 b3 Rfd8 (or 17 ... Rad8 18 Rd2 Nb4 19 Qd1 a5 20 h3, as in F.Berkes-A.Graf, Leipzig 2002, and now
20 ... a4! 21 bxa4 Nxa2 22 Rb1 Qa7 gives Black good counterplay, while 20 Nxd4 Nxa2 is just equal) 18
Rd2 a5 (or 18 ... Rac8 19 Qb2 Rc7 20 Ne1 Bd5! 21 Bxd5 Rxd5 22 Nf3 Rcd7 and Black was very solid
in B.Lajthajm-R.Sertic, Paracin 2011; ½-½ in 65) 19 h3 Rac8 20 Qb2 Rc7 and Black was fine in
I.Radulov-A.Cioara, Sunny Beach 2012.
d2) 15 Qd2 Be6 16 b3 h6 17 Rfd1 Rad8 18 Qb2 Rd7 19 Rd2 Rfd8 20 Rcd1 a5 and Black had
established his ideal defensive set-up, with no problems in R.Barcenilla-J.Sadorra, Philippines
Championship, Boracay Island 2012 (½-½ in 65).
e) 14 a3 Re8 (alternatively, 14 ... Bf5 15 Rc1 Qb6 16 Nd2 Rfe8 17 Nc4 Qd8 18 b4 Qd7 19 Re1
Bh3 20 Bxh3 Qxh3 21 Qd3 Rad8 was fine for Black in B.Kurajica-D.Larino Nieto, Malaga 2009, 0-1 in
62; or 14 ... Rd8 15 Ne1 Qe7 16 Nd3 Bg4 17 Re1 Rd6 18 Rc1 Re8 and Black had no problems in
N.Kalesis-S.Halkias, Greek Team Cup 2004, ½-½ in 27) 15 b4 Qb6 16 Qd3 g6 17 Rfd1 Bf5 18 Qd2 was
A.Astaneh Lopez-S.Collins, Bunratty 2012.
 Exercise: How should Black continue?

Answer: Instead of the premature 18 ... d3?, after which I had to struggle to make a draw (½-½ in
65), I could have obtained a pleasant position with the simple 18 ... Rad8.
At the risk of repeating myself, the Tarrasch is a classical opening and responds well to classical
ideas like rapidly and actively developing pieces. If you look at Aagaard’s games, this is basically all
he’s doing, which is sufficient for a 2700 performance. 18 ... d3? is a typical error, a result of looking for
a tactical continuation to justify Black’s inferior structure, when in fact the justification for my structure is
that I have tons more space in the centre and all my pieces are more active than their counterparts.
11 ... Be7

12 Bxf6
12 Nxf6+ Bxf6 13 Bxf6 (13 Qd2 h6 14 Bxf6 Qxf6 15 Rfd1 Rd8 16 b4 a6 17 Qb2 Bf5 was very
comfortable for Black in J.Murey-S.Vysochin, Cappelle la Grande 2003; 0-1 in 39 ) 13 ... Qxf6. White
can’t even pretend to have an advantage in this position. 14 b4 (or 14 Ne1 Re8 15 Nd3 Bf5 16 Rc1 Be4
17 Bxe4 Rxe4 18 Nc5 Re7 19 Qc2 Rae8 and Black was at least equal in A.Wirig-M.Kazhgaleyev, French
Team Championship 2003; 0-1 in 27) 14 ... Rd8 15 Qd2 Bg4 16 h3 (spotting the threat: 16 Rac1?? d3 and
White resigned in A.Wohl-M.Kanep, Gibraltar 2011 ) 16 ... Be6 17 Rfd1 Bc4 18 Qb2 was R.Swinkels-
A.Berelowitsch, Dutch Team Championship 2013, and now Black has several good ways to play – I like
18 ... b5, cementing Black’s control over c4, when he is certainly not worse.
12 ... Bxf6

13 Ne1
White has a wide choice here. In all lines Black should remain confident in his position and try to
develop his remaining pieces.
a) 13 Rc1 Re8 14 Ne1 Bf5 15 Nc5 Qb6 16 Ned3 Rac8 17 a3 (“Two bishops and active pieces
slightly favour Black in this position; White has no active plan.” – Ftacnik) 17 ... Bg5 18 Rc4 Qb5 19 b3
Be7 20 Qc1 was P.Nikolic-A.Grischuk, French Team Championship 2003, and now Ftacnik gives 20 ...
Ne5 21 a4 Qb6 22 Nxe5 Rxc5 23 Qf4 Be6 24 Rxc5 Bxc5 25 Nd3 Be7 as good for Black.
b) 13 Nc5 Qb6 14 Nd3 Re8 15 Qd2 Bf5 16 Rac1 Rad8 shows Black quickly reaching an optimal
set-up with full development. I like the idea of putting the bishop directly on f5, from where it is well
placed and has ideas of trading on d3 or trading the g2-bishop after ... Be4. Then 17 Rfd1 h6 18 b3 Ne7
19 Nf4 Nc6 20 Nd3 Ne7 21 Nf4 Nc6 22 Nd3 Ne7 23 Nf4 was agreed drawn in I.Enchev-M.Nikolov,
Sofia 2010, but in fact 23 ... g5! 24 Nh5 (or 24 Nd3 Nd5 followed by ... Nc3) 24 ... Bh8 leaves Black on
top – taking on g5 doesn’t work and otherwise Black will tidy up with ... Bg6 and ... Nd5.
c) 13 Qd2 Bf5 14 Nxf6+ Qxf6 15 Qf4 Rfe8 is equal and has led to draws in several games between
good players; for instance, 16 Nh4 Be4 17 Qxf6 gxf6 18 e3 Rad8 19 Rad1 dxe3 and a draw was agreed in
C.Foisor-Al.David, Milan 2009 – although it may have been more accurate to play 19 ... f5.
d) The forcing 13 b4 needs to be carefully considered. After 13 ... Nxb4 (13 ... a6 14 a4 Nxb4 15
Nxf6+ Qxf6 16 Qxd4 is similar) 14 Nxf6+ Qxf6 15 Qxd4 (or 15 Nxd4 Rd8 16 Qa4 a5 17 e3 Bd7 18 Qb3
Rab8 and Black had no problems in P.Prohaszka-J.Horvath, Austrian League 2012; ½-½ in 31) 15 ...
Qe7,
 White has liquidated Black’s only weakness and the position is drawish. Nevertheless, Black needs
to be careful, since he has not yet completed his development. The transition to such positions can be
psychologically uncomfortable for Tarrasch players and it is important to remain attentive:
d1) 16 Qb2 Nc6 17 Rfd1 (or if 17 a4, as in P.Maletin-A.Belozerov, Tomsk 2009, then 17 ... Rd8
followed by 18 ... Be6 leads to equal play) 17 ... Bf5 18 Nd4 Nxd4 19 Rxd4 was N.Vitiugov-V.Yemelin,
Tallinn (rapid) 2006, and now the active 19 ... Rae8 20 e3 b6 levels the chances, since Black has solved
the problem of his b7-weakness.
d2) 16 Qe5 Qxe5 17 Nxe5 leads to a further branch:

d21) After 17 ... Be6?! 18 Bxb7 Rab8 19 Bf3, Black didn’t have enough for the pawn in J.Hammer-
V.Akobian, Lubbock 2009 (1-0 in 46). It is interesting to see the mindset of Akobian (a Tarrasch expert),
preferring active counterplay to passive defence. In this particular position, however, Black could have
equalized with more cautious play.
d22) 17 ... Re8 18 Rfd1 Bf5 19 Nf3 was D.Fridman-M.Feygin, Belgian Team Championship 2009,
and now 19 ... Nc6 was the most accurate, with equality. Alternatively, 19 f4, as in A.Pashikian-
J.Bejtovic, European Championship, Plovdiv 2012, could have been well met by 19 ... f6 20 Nc4 Rxe2
21 Bxb7 Rb8 with no problems; for instance, 22 Nd6 Bc2 23 Rd4 a5 and Black’s pieces are active and
secure.

13 ... Re8
13 ... Bf5 14 Nd3 Bxe4!? 15 Bxe4 was agreed drawn in M.Palac-R.Zelcic, Croatian Championship,
Split 2008. I don’t think this is entirely necessary, but trading down to opposite-coloured bishops is worth
bearing in mind – as is well known, an advantage of the two bishops is that you can always trade one of
them.
14 Nd3 Be7 15 Rc1 Qb6
15 ... Bf8 16 Re1 Qb6 17 Qd2 Bf5 18 Nec5 Rac8 looks like the ideal set-up. After 19 h3 g6, a draw
was agreed in M.Suba-R.Vazquez Igarza, Collado Villalba 2008. Black had other options too, such as 19
... Ne5 when his position is slightly preferred by the engines.

16 Qd2
A.Belozerov-A.Shomoev, Russian Cup, Tomsk 2002, saw 16 a3 Bg4 17 h3 Bh5. It’s always a tough
decision where to retreat this bishop: while h5 always looks the most active (targeting e2), in fact the
bishop can end up slightly passive on g6 (after g3-g4). In addition, Nf4 is likely to be annoying. In the
circumstances 17 ... Bf5 seems like a more active approach. Black plans simply ... h7-h6 and ... Bf8 with
a harmonious position in which his chances are no worse. Instead, after 17 ... Bh5 the game continued 18
b4 Bf8?! (a little passive; 18 ... a5 is a standard way to seek to break down the white queenside – here it
seems dubious in view of 19 Nf4 Bg6 20 Nc5 and White is better since 20 ... axb4 21 Nd5 Qa7 22 Nxb7
looks threatening, but after 22 ... Rab8 23 Rxc6 Rxb7 24 axb4 Bxb4 25 Rc4 Bf8 26 Qxd4 Qxd4 27 Rxd4,
Black has decent drawing chances with all the pawns on one side of the board; alternatively, 18 ... Rac8
19 Nf4 Bg6 20 Nc5 Bxc5 21 Rxc5 Be4 also keeps White’s advantage within manageable limits) 19 Nf4
Bg6 20 Nc5 Bxc5 21 bxc5 Qa6 22 Bxc6 bxc6 23 Qxd4 Qxa3 24 Ra1 and White had the better prospects
(1-0 in 43). However, I’ve outlined how the defence could have been improved.
16 ... h6 17 Nec5 a5
It was interesting to consider 17 ... Bg5!? 18 f4 Be7, trying to loosen the white position.
18 a3 Bd6
A logical move, preparing ... Ne5. Again 18 ... Bg5!? 19 f4 Be7 creates some weaknesses and seems
more enterprising.
19 Qc2 Ne5 20 Nf4 Nc6 21 Nd5

Exercise: Assess the endgame after 21 ... Qxc5

22 Qxc5 Bxc5 23 Nc7 Rxe2 24 Nxa8 Bf8.

Answer: Black has a pawn (the d4-passer) and the bishop pair to compensate for his exchange
deficit. Black really doesn’t seem at all worse here and I think I made the mistake of trusting my GM
opponent.
21 ... Qd8?!
This compliant move lets White get an edge.
22 Rfe1 Re5 23 Nd3 Re8 24 Qb3 a4 25 Qb6
 25 ... Ra5?
Carelessly conceding control of b6. 25 ... Ra6 26 Qxd8 Rxd8 looks like a level endgame.
26 Qxd8 Rxd8

27 Nb6
Now the queenside pressure effected by the Nb6, Bg2 and Rc1 is unpleasant. Nevertheless, Black
still has a lot of defensive resources.
27 ... Be6
27 ... Bf5 was more active; for instance, 28 Bxc6 bxc6 29 Rxc6 Ra6 with a better version of the
game continuation.
28 Bxc6 bxc6 29 Rxc6 Ra6 30 Rec1 g5!
 Even after a number of inferior decisions by Black, the resulting position is far from unplayable.
Black has the bishop pair to compensate his pawn deficit, and Houdini assesses the position as only
slightly better for White.
31 Kf1 Kg7 32 Ke1 Bf8
32 ... Be7 was slightly more active.
33 Kd2 Re8 34 Nc5 Bxc5 35 R1xc5 Rb8
I had placed my hopes on the double pin, but White can deal with this without too many problems.
35 ... Bg4 was more enterprising: 36 f3 Bh3 37 Nc4 Rxc6 38 Rxc6 Bf1 39 e4 dxe3+ 40 Nxe3 Rd8+ 41
Ke1 Bd3 with drawing chances.
36 Rb5 Ra7 37 Rb4 Rab7 38 Kd3 Bb3 39 Kxd4?
Houdini recommends 39 h4 with a clear advantage for White.

Exercise: Find a way for Black to stay in the game.

39 ... Re7??
 Certainly not this!
Answer: 39 ... Rd8+ 40 Kc3 Rd1 gives Black active play for the two pawn deficit and it’s actually
not even clear that White is better here.
40 e4 Rd8+ 41 Ke3 1-0
The game score ends here, and I presume I lost after some time scramble.

To be honest, I only seriously analysed this game in preparation for this book. The way I had
remembered the game – “got a weak d-pawn in the Tarrasch, was ground down in a passive position” –
turns out to be completely incorrect. In fact, a more accurate version is:
1. Effortlessly got an equal, playable position out of the opening with Black against a strong GM.
2. Played the position passively (which probably involved being slightly intimidated as Black
against a strong GM who was out for revenge, having been beaten in a Tarrasch we played the previous
year), which gradually led to difficulties.
3. Was given a reprieve on move 39, but was too influenced by the trend of the game (which looked
like a standard GM grind) to see it.
These lessons inform some general psychological points about handling such positions (in addition
to the specific moves and motifs you can see from the above variations, such as the nice ... Be7-g5-e7
idea). I think it is vital to have an active mindset. The position naturally gives rise to counterplay (for
instance, the e2-pawn is always a target) and it’s important to be alert to Black’s dynamic possibilities. If
necessary, Black should be prepared to sacrifice material to stay active – the variation at Black’s 21st
move is an excellent example.

Game 13
V.Akopian-V.Jianu
European Championship, Legnica 2013

1 d4 d5 2 c4 e6 3 Nf3 c5 4 cxd5 exd5 5 g3 Nc6 6 Bg2 Nf6 7 0-0 Be7 8 dxc5 Bxc5 9 Nc3 0-0
Here we will look at a couple of lines recommended in recent repertoire books. Accordingly, these
lines are important since they are very likely to arise in practice.
10 Na4
 This is Boris Avrukh’s selection.
10 a3 is recommended by Larry Kaufman in his repertoire book. Analysis by Aagaard and Ntirlis
showed that Black needs to be very careful here. They recommended the enterprising pawn sacrifice 10 ...
Ne4! (not mentioned by Kaufman) which has subsequently been tested in several games between strong
players:

a ) 11 Nxd5 Be6 12 Nc3 (12 Ne3 was tested in E.Vorobiov-T.Suc, Trieste 2012, and now Black
should have played 12 ... f5! 13 Qxd8 Raxd8 14 b4 Bd4 with full compensation thanks to his much more
active pieces) 12 ... Nxc3 13 bxc3 Qxd1 (or 13 ... Re8 14 Qa4 Qa5 15 Qxa5 Nxa5 16 Nd4 Bc4 17 Be3
Rac8 18 Bf3 Ba6 and it was hard to find a way for White to make progress in A.Greenfeld-R.Swinkels,
Amsterdam 2013 – that Black went on to win can’t be considered a result of the opening, but does
perhaps suggest that White’s technical task is far from straightforward) 14 Rxd1 Rfd8 (14 ... Rad8 15 Bb2
Na5 16 Nd4 Bb3 17 Re1 Ba4 18 e3 Nc4 19 Bc1 b5 and Black had a dream Nimzo-Indian position – even
with a pawn less – and full compensation in E.Bacrot-A.Delchev, European Championship, Plovdiv
2012; ½-½ in 31) 15 Bf4 Rxd1+ 16 Rxd1 Bxa3 17 Nd4 Rd8 18 Bxc6 bxc6 19 Ra1 Bc5 20 Nxc6 Rc8 21
Nd4 was A.Giri-M.Kopylov, German League 2012, and now 21 ... Bd7 would have kept full
compensation for the pawn. I don’t see how White can make progress while avoiding drawn opposite-
coloured bishop endgames.
b) 11 Qxd5 Nxc3 12 Qxc5 Nxe2+ 13 Kh1 Nxc1 14 Raxc1 Qf6 15 b4
 was R.Khusnutdinov-R.Bulatov, Pavlodar 2012, and now Black equalizes with 15 ... Bg4; while
after 15 Rc3 Re8 16 Qb5, as in I.Tokarev-S.Bezgodova, Kazan 2010, Aagaard gives 16 ... h6 17 Rd1
Rb8 as leading to equality.
c) 11 Qc2 Nxc3 12 Qxc3 Bb6 (12 ... Be7 13 Nd4 Nxd4 14 Qxd4 Be6 15 Bf4 Bf6 16 Be5 Bxe5 17
Qxe5 gave White an edge in K.Lahno-N.Zhukova, Ukrainian Women’s Championship, Kharkov 2012 ) 13
b4 Bf5 (13 ... d4 is also playable: 14 Qd2 Qe7 15 Bb2 Rd8 16 Rfd1 Be6 17 b5 Na5 18 Qb4 Qxb4 19
axb4 Nb3 20 Ra3 Rac8, when Black had no problems and went on to score an upset in A.Rasmussen-
B.Byklum, Stockholm 2012; 0-1 in 45) 14 Bb2 d4 (14 ... f6 also deserves attention) 15 Qd2 Be4 16 a4 a6
17 b5 Na5 18 Qb4 axb5 19 axb5 Qd5 20 Bxd4 Bxf3 21 Bxb6 Bxg2 22 Rfd1 Nc6 (22 ... Qe6 23 Bc5 Nb3
is okay, but I prefer Moranda’s treatment) 23 Qxf8+ (23 bxc6 Rxa1 24 Rxa1 Bh3 25 f3 Qxc6 is balanced)
23 ... Rxf8 24 Rxd5 Bxd5 25 bxc6 Bxc6 and the long forcing sequence led to a drawn endgame in
V.Potkin-W.Moranda, European Championship, Legnica 2013 (½-½ in 41).
10 ... Be7

Retreats to d6 and b6 have also been tested, but the text seems the most logical and is by far the most
popular.
 11 Be3
While (as will appear) I don’t agree with his assessment of certain key variations, Avrukh’s
explanation of the strategic plans here is crystal clear: “In general we have clear rules about how to play
either side of a position with an isolated pawn. In this specific case exchanging certain minor pieces
would clearly favour White. The most desirable exchanges are: trading dark-squared bishops and a pair
of knights (f3 for c6), which would help White to establish full control over the d4-square. Obviously, the
endgame positions are always better for the side playing against the isolani. On the other hand, Black
should strive for activity keeping as many pieces as possible on the board.”
After 11 Bf4 Bg4 12 Rc1 Re8 13 a3 Ne4 14 b4 Rc8 15 Qd3 Bf6, Black was perfectly mobilized in
R.Salvador-J.Sadorra, Philippines Championship, Boracay Island 2012 (0-1, 36).

11 ... Re8
One of three popular moves here.
a) 11 ... Ne4 was tested in a game between two young American GMs: 12 Rc1 Re8 13 Nd4 Bd7 (13
... Bf6 14 Nxc6 bxc6 15 Rxc6 d4 16 Bf4 didn’t give Black quite enough for the pawn in V.Iordachescu-
Y.Kuzubov, European Championship, Kusadasi 2006 ) 14 Nb3 (Avrukh’s suggestion of 14 Nc3 Nxc3 15
Rxc3 is probably a better try) 14 ... Bf5 15 Nac5 Bxc5 16 Bxc5 Qd7 17 Re1 Bh3 18 Bh1 h5 19 Be3 h4 20
Nc5 Nxc5 21 Bxc5 Rad8 and Black had no problems in A.Ramirez Alvarez-R.Robson, US
Championship, St. Louis 2012 (0-1 in 63).
b) 11 ... Bg4 12 Rc1 (12 h3 Bf5 13 Rc1 Rc8 14 Nd4 Be4 15 Nxc6 bxc6 16 Bxa7 Bxg2 17 Kxg2 Ne4
looked like a comfortable version of a standard sacrifice in V.Artemiev-A.Nadanian, World Blitz
Championship, Khanty-Mansiysk 2013 – Black has potential play down three half open files and a
reasonable game)
 and now, as well as 12 ... Re8 (transposing into the main game), Black has a choice between two
good moves:
b1) 12 ... Rc8 13 h3 (or 13 Nc5 Bxc5 14 Bxc5 Re8 15 h3 Bh5 16 g4 Bg6 17 e3 Be4 18 Qa4?! h5!
19 g5 Nd7 and Black was much better in Y.Shulman-V.Akobian, blitz match, Upper Lake 2009; 0-1 in 27 )
13 ... Bh5 14 Nc5 Bxc5 15 Rxc5 with some advantage for White in E.L’Ami-V.Akobian, Wijk aan Zee
2010 (1-0 in 52).
b 2 ) 12 ... Qd7 13 Nc5 Bxc5 14 Rxc5 Rfe8 15 Re1 (15 Bg5 Qf5 16 Bxf6 Qxf6 was level in
D.Fridman-A.Berelowitsch, Dutch Team Championship 2011; ½-½ in 36) 15 ... h6 16 Nd4 Bh3 17 Nxc6
bxc6 18 Qa4 Bxg2 19 Kxg2 Rac8 looked reasonable for Black in D.Antic-V.Kotronias, Greek Team
Championship 2011 (0-1 in 55).
12 Rc1 Bg4
12 ... Bf5 is a pet variation of Larino Nieto. He tested the position after 13 Nd4 Nxd4 14 Bxd4 Ne4
15 e3 b6 16 Nc3 Qd7 17 Qe2 Rad8 twice against strong opposition in Don Benito 2012 and had no
problems, even using the same plan in both games:

a ) 18 Rfd1 Bg4 19 f3 Nxc3 20 Rxc3 (Black could also retreat to f5 or e6 since the g2-bishop is
passive, but trying to weaken the white king is logical) 20 ... Bh3 21 Bxh3 Qxh3 22 Rc7 Rd7 23 Rxd7
Qxd7 24 Rc1 h5!? 25 Kg2 Bd6 26 Qd3 Bb8 was L.Rubio Mejia-D.Larino Nieto, Don Benito 2012 (½-½
in 62). White has a hard job co-ordinating an attack on the c-file and minding his king at the same time.
b) 18 a3 Bg4 19 f3 Nxc3 20 Rxc3 Bh3 21 Rfc1 Bxg2 22 Kxg2 Bd6 23 b4 h5!? was A.Alonso
Rosell-D.Larino Nieto, Don Benito 2012 (½-½ in 57). Although computers like this for White, I think
Black is not worse in a practical game, since his king is safer and White has no entry points on the c-file.

13 h3
Instead:
a) 13 Bc5 Qd7 14 Re1 Rad8 15 Nd4 was M.Ragger-S.Brenjo, Bosnian Team Championship 2011,
and now 15 ... Rc8 16 Bxe7 Qxe7 17 Nxc6 Rxc6 would have been equal.
b) 13 Nc3 Qd7 14 Qa4 h6 15 Rfd1 was P.Tregubov-V.Akobian, FIDE World Cup, Khanty-Mansiysk
2009, and now 15 ... Bb4 or 15 ... Rad8 leads to good play for Black.
c) 13 Nd4 Qd7 14 Nxc6 (14 f3 Bh5 15 Bf2 Rac8 16 Re1 Ne5 17 Nc3 Bc5 18 Qa4 was agreed
drawn in R.Markus-S.Halkias, Serbian Team Championship 2010; while 14 Nc5 Bxc5 15 Rxc5 was
A.Dreev-V.Yemelin, Tallinn rapid 2009, and now I like 15 ... Bh3, exchanging White’s best minor piece)
14 ... bxc6 leads to a further branch:
 c1) 15 Bd4 Rac8 (15 ... Rab8 16 Re1 Bb4 17 Bc3 Bd6 18 a3 was M.Kekelidze-V.Akobian, Las
Vegas 2010; now Akobian launched his h-pawn forward, but 18 ... Be5 is solid and equalizes) 16 f3 Bh3
17 Qd2 Bxg2 18 Kxg2 Qe6 19 Rc2 Nd7 20 Rfc1 h5! with good kingside counterplay in Y.Drozdovskij-
K.Kiik, Tromsø 2009 (0-1 in 42). In particular, Black gets good compensation if White grabs on a7.
c2) 15 Re1 Bb4 16 Bd2 Bd6 17 Nc5 Qf5 18 Nb7 was M.Ragger-M.Kopylov, German League 2013,
and now Scherbakov recommends 18 ... Be5! which leads to good play for Black.
d) 13 Nc5 is Avrukh’s suggestion.

Exercise: Find a good response for Black.

Answer: 13 ... Qb6! is one of the first “novelties” I found on preparing the Tarrasch a few years ago
(if you count Houdini’s first choice as a novelty). Certainly the opposition of the black queen and the e3-
bishop creates an odd impression but White simply doesn’t have any useful discoveries:
d1) 14 Na4 Qa6 15 a3 Rad8 16 Re1 Ne4 17 Nd4 Ne5 18 f3? (18 Qb3 is equal) 18 ... Bd7 19 fxe4
Bxa4 and Black was clearly better in Mi.Novkovic-V.Jianu, European Championship, Plovdiv 2012 (0-1
in 37).
d2) 14 Nd3 d4 15 Bg5 (15 Rxc6 bxc6 16 Nxd4 c5! is not worse for Black) 15 ... h6 16 Bf4 Nd5 17
Bd2 Rad8 and Black was very comfortable in Ki.Georgiev-B.Byklum, European Cup, Eilat 2012.
d3) The position after 14 Nb3 Qa6 15 a3 Rac8 has been extensively tested:

d31) 16 Qd3 Qxd3 17 exd3 h6 was agreed drawn in A.Ushenina-N.Zhukova, Ukrainian Women’s
Championship, Kharkov 2012.
d32) 16 Re1 h6 17 h3 (after 17 Nbd4, as in P.Jaracz-J.Asendorf, German League 2012, I like 17 ...
Na5, aiming for c4) 17 ... Bf5 18 Bc5 Be4 19 e3 Bxc5 20 Nxc5 Qb6 21 b4 a5 22 bxa5 was also agreed
drawn in D.Reinderman-R.Swinkels, Dutch Championship, Amsterdam 2012. With Black I might be
tempted to play on here, since his pieces are very active and he runs no real risk of losing once the white
queenside is liquidated.
d33) 16 h3 Bh5 (normally I prefer to retreat the bishop to f5, and here too 16 ... Bf5 looks logical)
17 Re1 h6 18 Qd3 Qxd3 19 exd3 with a symmetrical and equal endgame in V.Laznicka-V.Jianu, European
Championship, Plovdiv 2012.
 Question: Where should Black retreat his bishop?

Answer: He shouldn’t!
13 ... Bxf3!?
A rather striking approach which has been played with success by Vlad Jianu in several games.
Black gives his opponent the bishop pair but manages to organize his own position fully. Retreats to e6
and f5 have been tested too, but we will concentrate on the text.
14 Bxf3
14 exf3 d4!? (14 ... Qd7 15 Nc5 Bxc5 16 Rxc5 h6 17 Qd3 Rad8 was also equal in A.Giri-V.Jianu,
European Championship, Plovdiv 2012) 15 Bg5 d3 16 Rc3 Qa5 17 Bxf6 Bxf6 18 Rxd3 Rad8! 19 Nc3
Bxc3 20 bxc3 Qxa2 and Black had equalized in O.Barbosa-Wei Yi, Jakarta 2012 (½-½ in 31).
14 ... Qd7

15 Kg2
15 Bg2 Rad8 16 Nc5 Bxc5 17 Bxc5 Ne4 18 Bd4 Qf5 19 e3 Rc8 20 Bc3 h5!? 21 Qf3 Qxf3 22 Bxf3
Nxc3 23 bxc3 (after 23 Rxc3 d4 24 exd4 Nxd4 25 Bxh5 Rxc3 26 bxc3 Ne2+ 27 Bxe2 Rxe2, Black should
be able to hold the endgame due to his active rook and the white queenside weaknesses) 23 ... Re5 24 c4
was R.Wojtaszek-A.Delchev, French Team Championship 2013, and now simply 24 ... dxc4 25 Rxc4 g6
looks solid enough for Black.
15 ... Rad8
 16 Qc2
16 Nc5 Bxc5 17 Bxc5 was D.Arngrimsson-V.Jianu, Arad 2012, and now Black can create
interesting play with 17 ... Ne4 18 Qc2 Ng5 19 Bg4 Ne6 20 Rfd1 g6, followed by ... f7-f5 or ... h7-h5.
Black has some (albeit faint) attacking hopes, whereas it’s not quite clear how White should continue.
16 ... Ne4 17 Bxe4 dxe4 18 Rfd1 Qe6 19 Rxd8 Rxd8 20 Qc4 Qxc4 21 Rxc4 f5

The endgame is equal. While this is no guarantee of a favourable result against Akopian, Jianu
managed to draw with ease, playing actively and even sacrificing a piece.
22 Nc5 b5! 23 Rc1 Nd4! 24 Nb7 Rd7 25 Na5 Nxe2 26 Rc8+ Kf7 27 Rc2 Nd4 28 Bxd4 Rxd4 29
Rc7 Rd2! 30 Nc6 Ke6! 31 Nxe7 e3 32 Kf1 Rxf2+ 33 Ke1 Rxb2 34 Rxa7 g5 35 Ng8 f4 36 Re7+ Kd5
37 gxf4 gxf4 38 Nf6+ Kd4 39 Rd7+ Kc4 40 Rc7+ Kd3 41 Rd7+ Kc4 42 Rc7+ Kd3 43 Rd7+ Kc4 ½-½

Game 14
B.Chatalbashev-I.Sokolov
Jakarta 2012
1 d4 d5 2 c4 e6 3 Nc3 c5 4 cxd5 exd5 5 Nf3 Nc6 6 g3 Nf6 7 Bg2 Be7 8 0-0 0-0 9 dxc5 d4!? 10 Na4
Otherwise Black captures on c5 with no problems.
10 ... Bf5

Question: What does Black have for the sacrificed pawn?

Answer: Frankly, Black’s compensation is not completely obvious, and it’s far from clear that it’s
sufficient to balance the pawn deficit. First and foremost, the white knight is rather out of play on a4 and
will take several moves to get back into the game. Black wants to continue with moves like ... Be4, ...
Qd5, ... Rfe8 and ... Rad8 with an impressively centralized position, and possible tactical shots with ...
d4-d3, ... h7-h6 and ... g7-g5 also come into consideration.
White doesn’t need to do much, since if he simply consolidates he will be a pawn up. However, one
subtlety of the position should be noted: Black’s only weakness is on b7 which is actually a bit awkward
to defend after Qb3; but if White plays the most natural plan of consolidating c5 and bringing his knight
back into play (a2-a3, b2-b4 and Nb2), then the b7-pawn can no longer be attacked.
11 b4!?
 A dynamic approach. White returns the pawn in order to activate his rook on the b-file. Giving
White b2-b4 “for free” (without a2-a3) would be too much of a concession, so Black needs to take the
pawn, but can play an intermezzo first.
11 Bf4 Be4 12 Rc1 gives Black a wide choice:
a) 12 ... Re8 and now:

a1) 13 a3 Qd5 14 b4 Rad8 15 Re1 (15 b5 Na5 16 Bc7 Nb3 17 Bxd8 Rxd8 is an important tactical
point, with equality in E.Vegh-D.H.Toth, Hungarian Team Championship 2007; but 15 Nb2 g5 16 Bd2 g4
17 Ne1 Bxg2 18 Nxg2 was good for White in M.Stojanovic-A.Profumo, Lugano 2006, since Black had
given a lot of squares away, in addition to his pawn) 15 ... h6 16 b5 (16 Nb2 looks more sanguine) 16 ...
Na5 17 Bc7 Nb3 18 Bxd8 Rxd8 19 Rc2 Bxc2 20 Qxc2 Nxc5 21 Nxc5 Bxc5 and Black had a good
Tarrasch-type position in A.Barsov-E.Magerramov, Abu Dhabi 1999 (0-1 in 40).
a2) 13 b4 Nxb4 14 Qxd4, as in M.Hörstmann-R.Östreich, Bad Lauterberg 1994, should be met by 14
... Nxa2 15 Rcd1 Qxd4 16 Nxd4 Rac8 with an approximately level endgame.
a3) 13 Qb3 and now, after the ridiculous 13 ... Na5? 14 Qb4 Nc6 15 Qxb7 Qd5 16 Rfd1!, I lost
without a fight in L.Cernousek-S.Collins, Teplice 2011. Instead, 13 ... Qd7 14 e3 Rad8 was more
sensible, with decent compensation.
a4) Houdini’s recommendation of 13 e3! doesn’t seem to have been tested yet.
b) 12 ... h6 13 a3 g5 has been essayed by Ekaterina Kovalevskaya, whose Tarrasch expertise has
been praised by none other than Viktor Bologan. After 14 Bd2 Qd5 15 b4 Ne5,

it is rather hard to believe that Black enjoys full compensation here, but the position is messy and
Black’s practical chances shouldn’t be underestimated. While White has basically consolidated on the
queenside (her pawn chain is strong and Nb2-d3/c4 is in the pipeline), Black doesn’t have to worry about
her b7-weakness which, as will be seen, features in some other lines:
b1) 16 h4 Nh7 17 c6 b6 18 Bc3! was much better for White in A.Galliamova-E.Kovalevskaya,
FIDE World Cup, Khanty-Mansiysk 2012 (1-0 in 44). Since the g5-pawn wasn’t really threatened, Black
could improve with 16 ... Ng6, though it remains to be seen whether she has enough compensation.
b2) 16 Nb2 d3 (simply continuing the build-up with 16 ... Rfe8 17 Nd3 Nxf3+ 18 Bxf3 Rad8 looks
sensible – I don’t see how White is going to unravel) 17 Nxe5 (17 Rc3! would have given Black serious
problems) 17 ... Bxg2 18 Nbxd3 (18 Re1 Bh3 19 Nf3 dxe2 20 Qxe2 might be a better try for an
advantage, though players who go for this line should be delighted at the long-term weakness of the light
squares on the white kingside) 18 ... Bxf1 19 Qxf1 Ne4 and White had good compensation for the
exchange, but Black had no reason to complain in V.Cmilyte-E.Kovalevskaya, European Cup, Eilat 2012
(½-½ in 56).
c) 12 ... Qd7 is a speciality of Serbian GM Goran M Todorovic (not to be confused with Serbian IM
Goran N Todorovic). Black simply puts his pieces in the centre and challenges White to break out. One
feature of this line is that Black hangs on to his b7-pawn, which rather holds his position together. After
13 a3 Rad8 14 b4, there are several examples:
 c1) 14 ... Rfe8 15 Nb2 (or 15 Ne1 Bxg2 16 Kxg2 Nd5 and Black had pleasant compensation in
V.Radonjanin-G.M.Todorovic, Serbian Team Championship 2005; 0-1 in 37) 15 ... Nd5 16 Bd2 Qf5 17
Qb3 a6 18 Nd3 Bf6 19 Nf4? was a typical blunder in S.Krivoshey-G.M.Todorovic, Dos Hermanas
(online blitz) 2004. Now the game’s 19 ... d3 was alright (0-1 in 49), but 19 ... Nxf4! 20 gxf4 (or 20 Bxf4
d3) 20 ... Re6 would have been even better.
c2) The speculative 14 ... Nd5 15 Bd2 Qe6 16 Qe1 Bf6 17 b5 Rfe8 18 bxc6 Qxc6 19 Nb2 didn’t
give Black enough for the piece in Ki.Georgiev-G.M.Todorovic, Topola 2004 (1-0 in 40). Black has
improvements such as 17 ... Nce7 and 16 ... a6; but earlier White has the annoying 15 Bd6, when I don’t
think Black has enough compensation.
d ) 12 ... Qd5 13 Qb3 Qxb3 (trying to bail out) 14 axb3, as in I.Zugic-I.Seitaj, Calvia Olympiad
2004, and now 14 ... Nd5 15 Bd6 Rfe8 16 Bxe7 Rxe7 leaves Black with an uphill struggle to draw. But
13 ... Qh5 seems dubious to me:

d1) 14 Qxb7! looks very strong: 14 ... d3 (Black doesn’t have to play this, but it’s one of the main
ideas of the whole line and it’s not clear what else he would play to justify his two-pawn deficit) 15 Rfe1
(Houdini prefers 15 Nd2) 15 ... dxe2 16 Rxe2 (again 16 Nd2! was stronger) 16 ... Bxf3 17 Bxf3 Qxf3 18
Rxe7 Ne4 19 Rxe4 Qxe4 20 Nc3 Qf3 with an edge for White in D.Fridman-S.Halkias, European Team
Championship, Novi Sad 2009 (1-0 in 52).
d2) 14 Rfd1 Rad8 (after 14 ... d3 15 Rxd3! Bxd3 16 exd3, White had two pawns for the exchange
plus much better pieces in J.Gonzalez Valero-S.Pozo Vera, Spanish Team Championship 2009; note how
difficult it is to defend the b7-pawn) 15 Bd6? (Ftacnik gives 15 Qxb7 Rfe8 16 Ng5 Bxg2 17 Kxg2 Nd5 18
Qxc6 Bxg5 19 Bxg5 Qxg5 20 Rxd4 Qxc1 21 Rxd5 as clearly better for White) 15 ... b5! (taking advantage
of the bishop on d6; either capture of the black b-pawn – 16 cxb6 or 16 Qxb5 – drops a piece) 16 Bxe7
bxa4 17 Qxa4 Nxe7 18 Rxd4 (or 18 Nxd4 Bxg2 19 Kxg2 Ned5 20 Qb3 Rfe8; Ftacnik gives 21 Qf3 Qg5,
but modern engines spot 21 ... Ng4! immediately: 22 h3 Nde3+! 23 fxe3 Nxe3+ 24 Kg1 Qxf3 25 exf3
Nxd1 26 Rxd1 Re5 with a big advantage) 18 ... Rxd4 19 Nxd4 Bxg2 20 Kxg2 Qd5+ 21 f3 Qg5 was
slightly better for Black in S.Matveeva-A.Lugovoi, Solin 2005 (½-½ in 58).
d3) 14 Ng5 featured in one of the most exciting games I ever played. Having ruined an excellent shot
at a GM norm by losing with White to Sune Berg Hansen in the previous round, I needed to win with
Black against a strong player to keep a realistic norm shot. I got some excellent advice (unfortunately not
fit for publication) before the game from Aussie GM Dave Smerdon, and selected this gambit variation.
B.Gundavaa-S.Collins, Khanty-Mansiysk Olympiad 2010, continued 14 ... Ng4 15 h3 Bxg2 16 Kxg2
Nxf2!? (16 ... Bxg5 is sounder, when Black is at least not worse) 17 Qxb7? (17 Rxf2 Bxg5 18 Qd5 h6 19
h4 is balanced) 17 ... Bxg5 18 Qxc6 Bxf4 19 gxf4 Qxh3+! 20 Kxf2 Rae8 (starting a caveman attack a full
knight down) 21 Qf3 Qh4+ 22 Kg2 Re3 23 Rh1 Qe7 24 Qh5 Qe4+ 25 Kf1.

Now I went on to win a crazy game after 25 ... Qxf4+ (0-1 in 101). However, 25 ... d3! was a better
move: 26 Rc4 (I didn’t want to let White “bail out” with 26 Qxh7+, but in fact 26 ... Qxh7 27 Rxh7 Kxh7
is obviously very good for Black – clearly a little too much adrenaline!) 26 ... dxe2+ 27 Kf2 e1Q+ 28
Rxe1 Qxc4 29 Rxe3 Qxf4+ 30 Qf3 Qxa4 and, while the game isn’t over yet (especially in view of
White’s c-pawn), the three connected passers give Black cause for optimism.
11 ... d3!
 This seems best. 11 ... Nxb4 12 Nxd4 Be4 13 Bxe4 Nxe4 was Z.Li-I.Seitaj, Istanbul Olympiad
2012, and now 14 Nf5! returns the pawn with a much better endgame after 14 ... Bxc5 15 Nxc5 Nxc5 16
Be3 Qxd1 17 Rfxd1.
12 e3
12 Ba3 dxe2 13 Qxe2 Bd3 14 Qb2 Bxf1 15 Rxf1 was unclear in Zhu Chen-E.Kovalevskaya, FIDE
Grand Prix, Shenzhen 2011 (½-½ in 29).

12 ... Nxb4
12 ... b5 13 cxb6 axb6 targets the a4-knight in an attempt to fuel Black’s counterplay. Unfortunately,
14 b5 Nb4 15 Nd4 Be4 16 Bb2 seems better for White.
13 Nd4 Be4 14 Bxe4 Nxe4 15 Rb1
 15 ... Nxa2?
15 ... Qa5 gives good equalizing chances.
16 Qxd3 Ng5 17 Ba3 a5
17 ... Nf3+ 18 Kh1 Qd5 19 Qf5! keeps White on top.
18 Nb6 Ra6 19 Rfd1 Nb4 20 Bxb4 axb4 21 Nc6 Nf3+ 22 Kh1 Ra2 23 Qxd8 Bxd8 24 Kg2 bxc6
25 Nd7??
A horrible mistake, throwing away his advantage and then some. Instead, after 25 Kxf3 Be7 26 Nd7
Rc8 27 Rxb4, Sokolov could only beg for a draw.

Exercise: Black to play and win!

Answer: 25 ... Nd2!

The knight blocks the d-file and shields the d8-bishop from the rook on d1.
26 Nxf8 Kxf8 27 Rxb4 Ke7
White should still hold this, but it’s unpleasant – and obviously nothing to do with the opening.
 28 Rd4 Ba5 29 g4 g6 30 g5 Bc3 31 Rh4 Ke6 32 Rc1 Be5 33 Rb4 Kf5 34 Rb7 Ne4 35 Rxf7+ Ke6
36 Rf3 Nxg5 37 Rf8 Ke7 38 Rc8 Ne4 39 Kf3 Nxf2 40 Rf1 Nh3 41 Ke4 Bxh2 42 Rxc6 Rd2 43 Kf3
Ng5+ 44 Kg4 Rg2+ 45 Kh4 h5 0-1

Game 15
M.Manolache-V.Jianu
Baia Sprie 2012

1 d4 d5 2 c4 e6 3 Nf3 c5 4 cxd5 exd5 5 g3 Nc6 6 Bg2 Nf6 7 0-0 Be7 8 dxc5 Bxc5 9 a3
This is another attempt which has been played against me by a strong GM.
9 ... 0-0
Aagaard and Ntirlis recommend 9 ... Ne4.

We have yet another example of Jacob slightly varying from his own recommendation and getting a
good result:
a) 10 Nfd2 (another novelty suggested by Aagaard and Ntirlis which has subsequently been tested in
practice) 10 ... f5! (after 10 ... Bf5 11 Nxe4 Bxe4 12 Bxe4 dxe4 13 Qc2 Qe7 14 Nc3, Black had some
problems in V.Georgiev-M.Meinhardt, German League 2012; 1-0 in 51 ) 11 Nc3 Be6 12 Nb3 Bb6 13 Na4
d4 14 Nd2 Nxd2 15 Bxd2 0-0 16 Rc1 Bd5 and Black had equalized in W.Ju-N.Zhukova, FIDE World
Cup, Khanty-Mansiysk 2012.
b) 10 b4 Bb6 (here 10 ... Be7 11 Bb2 Bf6 12 Bxf6 Qxf6 13 Ra2 0-0 14 Rc2 Rd8 with equality in
Y.Kruppa-V.Potkin, St. Petersburg 2000, was the route suggested in Jacob’s book ) 11 Bb2 0-0 12 Nc3
Nxc3 13 Bxc3 d4 14 Bb2 a6 15 Qd2 Be6 16 Qf4 Bc4 17 Rfe1 Re8 18 Bf1 a5 19 Rad1 (19 Rac1 is
another try, but after 19 ... Bc7 20 Qd2 b5 Black’s pieces look too well placed for him to be worse; a
natural line might continue 21 e3 axb4 22 axb4 dxe3 23 Rxe3 Qxd2 24 Nxd2 Rxe3 25 fxe3 Nxb4 26 Nxc4
bxc4 27 Bxc4 with what should be a drawn endgame) 19 ... Bb3 20 Rd3 Bc4 21 Rdd1 Bb3 22 Rd3 Bc4
23 Rdd1 and a draw was agreed in L.Johannessen-J.Aagaard, European Team Championship, Porto
Carras 2011. Another example of Jacob’s exemplary handling of the Tarrasch – nothing fancy, just putting
his pieces on good squares.
 10 b4 Bb6 11 Nc3 h6 12 Bb2 Re8 13 Rc1
13 Na4 Bc7 looks okay for Black.

13 ... a6
13 ... Bf5 was my choice when I reached this position in N.Pert-S.Collins, British League 2012. For
the moment I find a good plan, bringing my pieces to decent squares. Nick’s idea of Na4xb6 will weaken
his control of the e4-square. The game continued 14 Na4 Be4 15 Nxb6 axb6, when I seem to remember
being dissatisfied with my position, but unfortunately I had forgotten my chess history.
 A similar position (same structure and with White having the bishop pair) was reached in M.Tal-
P.Keres, Bled/Zagreb/Belgrade Candidates 1959 (0-1 in 79). Although it looks as though White must be
better, in fact it’s hard for him to find an active plan. My bishop on e4 is the best minor piece on the board
and if it’s traded for its white counterpart, Black will always have some attacking chances. Also the e2-
pawn is weak and doesn’t want to move since this would weaken a bunch of light squares.
After 16 Nh4 and then 16 ... Ne5 17 Bxe4 Nxe4!, the knight is threatening to come to c4, while if
White trades on e5 the position should be level since White has his own weaknesses on a3 and e2.
Alternatively, 16 ... Bxg2 17 Nxg2 Qe7 followed by ... Rad8 is a sound continuation, with equal chances.
Instead, I played 16 ... Qe7?, which indicates a mistaken plan.

One of the things to be careful of when playing the Tarrasch is structural changes and, in particular, a
diminished sense of danger regarding lines where White “corrects” Black’s pawn structure. To take a
basic example, if the d5- and e2-pawns get traded in a normal Tarrasch position, we have a symmetrical
and open position in which the better mobilized side will have the better chances. Often the better
mobilized side will be White, a natural consequence of getting to move first.
Similarly here, I barely gave any thought to Nick sorting out my pawn structure by taking on e4.
However, in the resulting position, the bishop on b2 is extremely powerful (it can never be blocked by ...
d5-d4, for instance) and White’s weakness on e2 is covered. The dark-squared bishop can combine well
with the knight on h4, particularly if this knight gets into f5.
After 17 Bxe4!, I could in fact have kept the game unclear with 17 ... Qxe4 here, since the capture on
f6 isn’t as dangerous as it looks – though this is still worse than my options on the previous move, and
when I played 16 ... Qe7 I was of course planning to take back on e4 with the pawn (to the extent that I
examined this line at all – as outlined above, I probably devoted more time to continuations which left my
IQP on the board). So 17 ... dxe4 18 Qb3 Rad8 (more playing by rote; 18 ... Qe6 was the only attempt to
keep the knight on h4) 19 Nf5 (White is already clearly better) 19 ... Qe6 20 Qxe6 Rxe6 (the computer
recommends the pawn sac with 20 ... fxe6 21 Nxh6+ Kh7) 21 Rfd1

(around here I was probably looking with pride at my half-decent pawn structure, not noticing how
White’s pieces have become several times more active than their counterparts) 21 ... Rb8 22 Rd6 Kf8 23
b5 Ne7 24 Rxe6 fxe6 25 Nxe7 Kxe7 26 Rc7+ Kd6 27 Rxg7 and White is now winning and converted
without too much hassle (1-0 in 46).
14 Na4 Ba7 15 Nc5
The knight transfer to c5 brings no advantage after Jianu’s accurate reaction.
15 ... Ne4!
 16 Rc2 Nxc5
16 ... a5!? is a good alternative.
17 bxc5 Qe7?!
A rather speculative pawn sacrifice. After 17 ... Bf5 18 Rd2 Be4 19 Nh4 Bxg2 20 Kxg2 d4 21 Bxd4
Nxd4 22 Rxd4 Qc7, Black regains his pawn with equality.
18 Qxd5 Be6 19 Qd2 Rad8 20 Qc3 Qf8 21 Ba1?
21 e4! would keep White on top.
21 ... Bf5 22 Rd2 Bxc5 23 Rxd8 Rxd8

Black has fully equalized. The rest of the game sees Jianu convincingly outplaying a good GM.
24 Qb2 Rd7 25 Ne5 Nxe5 26 Qxe5 Be6 27 a4 b6 28 Bc6 Re7 29 Qb2 Bc4 30 Bf3 Rd7 31 Qc2
b5 32 axb5 axb5 33 Bb2 Qd8 34 Qf5 Bf8 35 Ra1
35 Be4 was a better move, when Black maintains equality with 35 ... g6 36 Qe5 f6!, since he is
certainly not worse after 37 Qxf6 Qxf6 38 Bxf6 Kf7 and 39 ... Bxe2.
35 ... Rd1+ 36 Kg2 Rxa1 37 Bxa1 Qe7 38 Be5 g6 39 Qf4 Bg7 40 Bxg7 Kxg7
 Black is now solidly better. White’s defence can be improved, but in the game he didn’t last much
longer.
41 Qd4+ Qf6 42 Qd2 Qe5 43 h4 Kh7 44 g4 Qc5 45 Qa5 Qe7 46 Kh3 b4 47 h5 gxh5 48 gxh5
Be6+ 49 Kg3 Qg5+ 50 Qxg5 hxg5 51 Be4+ Kh6 52 f4 f6 53 Bb1 Kxh5 54 e4 b3 55 Bd3 b2 56 e5 fxe5
57 fxg5 Kxg5 0-1

Key Notes
1. Capturing on c5 is a principled decision and forces Black into an IQP structure.
2. In the main line with 10 Bg5 d4 11 Bxf6 Qxf6 12 Nd5, White hopes to obtain a slight advantage
after solidly blockading the d4-pawn and playing on the queenside. Black has a wide choice as to how to
respond and, in particular, whether to retreat his bishop to b6, a7 or f8.
3. 11 Ne4 or 11 Bxf6 Bxf6 12 Ne4 leads to quiet and balanced positions. Black should be sure to be
well prepared here since, as my game with Alex Baburin demonstrates, thoughtless play will lead to a
passive and inferior position.
4. 10 Na4 is a recently popular try. I like Jianu’s simple solution of ... Bg4 and taking on f3, which
has scored well against very strong players.
5. a2-a3 on move 9 or 10 has the idea of queenside expansion with b2-b4; here ... Ne4! is a good
response, even as a pawn sacrifice.
6. I’m not sure of the objective merits of the gambit variation in Chatalbashev-Sokolov (Game 14). From
a practical perspective, however, the line is fully playable and gives Black dangerous compensation.
Chapter Five
Other 8th/9th Move Options

1 d4 d5 2 c4 e6 3 Nf3 c5 4 cxd5 exd5 5 g3 Nc6 6 Bg2 Nf6 7 0-0 Be7

After the previous chapters I think we’re pretty well prepared for this position. White has a couple
of logical alternatives which we’ll discuss.
8 b3
This queenside fianchetto, and the closely related 8 Nc3 0-0 9 b3, are the subject of Giri-Swinkels
(Game 16).
8 Be3 has been the traditional choice of the legendary Victor Korchnoi. See how I got on in
Korchnoi-Collins (Game 17).

Game 16
A.Giri-R.Swinkels
Dutch Championship, Amsterdam 2012

1 c4 e6 2 g3 d5 3 d4 c5 4 cxd5 exd5 5 Nf3 Nc6 6 Bg2 Nf6 7 0-0 Be7 8 b3

White can also start with 8 Nc3 0-0 9 b3. This was played against me in a tournament where I made
my second GM norm. After 9 ... Ne4 10 Bb2 Bf6, White needs to deal with the pressure on d4.
 a) 11 Na4 b6 has proven resilient for Black; for instance, 12 Rc1 Re8 13 dxc5 Bxb2 14 Nxb2 bxc5
15 Nd2 Qf6 16 Na4 (16 Nxe4 dxe4 17 Rxc5 Bg4 and ... Rad8 promises reasonable compensation) 16 ...
Ba6 17 Nxe4 dxe4 18 Qd2 Rad8 19 Qe3 Nd4 gave Black good chances in J.Speelman-T.Luther, German
League 2002 (0-1 in 33).
b) 11 e3 Bg4 12 Na4 cxd4 13 Bxd4 Nxd4 14 exd4 b5 15 Nb2 (15 Nc5 is more active) 15 ... Qb6 16
Qd3 Bxf3 17 Bxf3 Bxd4 18 Bxe4 dxe4 19 Qe2 with a decisive advantage for Black, which I eventually
(and following several major mistakes) converted in R.Goetzee-S.Collins, Tromsø 2010 (0-1 in 43).
8 ... 0-0 9 Bb2

White has not attempted to put any immediate pressure on the d5-pawn, so Black has a wide choice.
9 ... Re8
Alternatively:
a) 9 ... Be6 10 Nc3 b6 11 Rc1 Rc8 is solid, but a little passive, and White has a good score here
(64% in the database).
b) 9 ... Bg4 is one of the more aggressive moves, immediately pressurizing d4.
 b1) 10 e3?! is entirely misconceived: 10 ... cxd4 11 exd4 Qd7 12 Nbd2 Ne4 and Black’s pieces
were much better placed in the symmetrical pawn structure arising in I.Teodorescu-G.Timoshenko,
Rumanian Team Championship 2009 (0-1 in 77). The b2-b3 and Bb2 idea makes no sense in this structure
– the dark-squared bishop is biting on its own pawn, and the c3-square is weak.
b2) 10 dxc5 Bxc5 11 Nc3 a6! 12 Rc1 Ba7 and Black had comfortably established his bishop on its
ideal diagonal. This idea of ... a7-a6 and ... Ba7 (or, with White, a2-a3 and Ba2) arises very often in IQP
positions. In addition to placing the bishop on a great square where it is safe, targets important pawns and
squares on the g1-a7 diagonal, and doesn’t get in the way of Black’s pieces, the set-up has an important
benefit in preventing White’s Nc3-b5-d4 manoeuvre. I already prefer Black, who scores very well from
this position (62% in the database); GM Spartak Vysochin actually has 4/4 here as Black.
b3) 10 Ne5 and now Black has a choice:

b31) 10 ... Bh5 11 dxc5 Bxc5 12 Nd3 Bd6 13 Nd2 Re8 14 Nf3 leads to an interesting position
where the slightly unusual feature is the white knight on d3. On one hand, this knight shields the d5-pawn
from attack (much as a knight would on d4); but it is also controlling important dark-squares and limiting
the activity of Black’s pieces. Probably the position is dynamically balanced, as suggested by a couple of
examples:

14 ... Bg6 (14 ... Qe7 15 Re1 Rad8 16 Rc1 Bg6 17 Nd4 Nb4 18 Nxb4 Bxb4 19 Rf1 a6 was level in
W.Ju-Zhao Xue, FIDE Grand Prix, Shenzhen 2011 ) 15 Rc1 Qe7 16 Re1 Ne4 17 Nf4!? Bxf4 18 gxf4 Rad8
19 e3 led to an interesting struggle in J.Tompa-D.Berczes, Hungarian Team Championship 2012. Black
has weakened the white kingside – at the cost of (probably) his best minor piece. If Aronian was handling
the white pieces the result could have been quite different, but in the game White rapidly allowed his GM
opponent a devastating initiative: 19 ... Bh5 20 Rc2?! (the rook is misplaced on this square) 20 ... Qe6 21
a3? (White is justifiably worried about ... Nb4 shots, but it was better to simply move the rook back) 21
... Qg6! 22 Kf1 and now 22 ... Na5 23 b4 Nc4 would have left Black completely on top.
b32) 10 ... cxd4 11 Nxg4 Nxg4 12 e3 Nf6 13 exd4, and in this symmetrical pawn structure the two
bishops are not really felt – in fact, White needs to be careful that his bishop on b2 is not locked out of the
game. After 13 ... Re8 14 Nc3 Bf8 15 Qd3 Qd7 16 Rae1 Rxe1 17 Rxe1 Rd8 18 Bc1 h6 19 Bf4 Bb4 20
Rd1 Bd6 21 Be3 Ba3 22 Bf4 Bd6 23 Be3 Ba3 24 Bf4 a draw was agreed in M.Gurevich-L.Aronian,
European Team Championship, Heraklio 2007.
c) By far the most popular move is 9 ... Ne4 and now:
 c1) 10 Nc3 transposes to the 8 Nc3 0-0 9 b3 variation in the first note to this game.
c2) 10 Nbd2 Bf6 11 e3 is a tame set-up. Here 11 ... Bg4 12 h3 might be a little better for White;
instead, I like 11 ... Bf5!? as in B.Svendsen-S.Brynell, Gausdal 2005, with active play for Black (0-1 in
23).
c3) 10 dxc5!? Bxc5 is critical:
c31) 11 Nc3 Nxc3 12 Bxc3 d4 13 Bb2 Bg4 leaves Black comfortable. After 14 Rc1 Bb6 15 h3 Bf5
16 Qd2 Qd7 and a draw was agreed in I.Morovic Fernandez-A.Bachmann Schiavo, Asuncion 2010.
c32) 11 e3 has scored well for White.

c321) Black has been suffering after 11 ... Bg4 12 Nc3!; for instance, 12 ... Qa5 (12 ... Nxc3 13
Bxc3 Qd7 14 Qd2 Rfd8 was A.Galkin-S.Kozhuharov, Turkish Team Championship 2010, and now 15
Bxg7! would have been good for White in view of the forced variation 15 ... Bxf3 16 Bxf3 Qf5 17 Bb2
Qxf3 18 Qc3 Nd4 19 exd4 Qxc3 20 Bxc3 with an extra pawn) 13 Nxd5 Rad8 14 Qc2 and Black didn’t
demonstrate sufficient compensation in V.Zvjaginsev-Xu Jun, FIDE World Championship, Moscow 2001
(1-0 in 22).
c322) Perhaps Black should try 11 ... Qb6!? 12 Re1 (12 Qxd5? Nxf2!! is excellent for Black) 12 ...
Rd8 13 Nc3 Nxc3 14 Bxc3 Bg4 with equality, as in Mi.Marin-M.Thesing, Rumanian Team Championship
2009 (½-½ in 47).
10 Nc3 Bg4
 11 dxc5
Drawing the bishop to the c5-square where it will be vulnerable after White plays Rc1. By contrast,
the immediate 11 Rc1 seems to save some time for Black: 11 ... cxd4 12 Nxd4 Qd7 13 Re1 Rad8 and
Black had established his desired set-up in B.Chatalbashev-Y.Solodovnichenko, Val Thorens 2008 (0-1
in 32).
11 ... Bxc5 12 Rc1

12 ... Bf8
Keeping the bishop out of harm’s way looks logical. However, an alternative approach was tried in
a rapid game between two top GMs: 12 ... Bb4!? 13 h3 Bh5 14 a3 Bxc3 15 Bxc3 Ne4 16 Bb2 Qb6 17 g4
Bg6 18 Nd4 Rac8 with comfortable play for Black in R.Mamedov-S.Ganguly, Moscow (rapid) 2013.
13 h3
Probably a useful move to include. Trying to do without it by 13 Re1 Qd7 14 Nd4 allows Black to
threaten the white king with 14 ... Bh3 15 Bh1 (otherwise the kingside light squares would be weak, but
now the white king is short of breathing space) 15 ... Rad8 with a dream IQP position for Black in Z.Ilic-
T.Willemze, Haarlem 2012 (0-1 in 32).
 13 ... Bh5 14 Nb5 Ne4 15 Nbd4 Rc8

Question: How would you assess this position?

Answer: Black has every reason to be happy. He has active development with all the pieces still on
the board (so a long way from a depressing IQP endgame) and White doesn’t have any pressure on the d5-
pawn. In addition, while the blockade on d4 looks very solid, Black has some pressure – an exchange of
rooks on c1 could draw away a white defender, while the bishop on h5 and the rook on e8 combine well
to pressurize the e2-pawn. In short, a full and complex middlegame is in prospect, where the better player
will win. In the game the 200+ rating point gap proves too much.
16 Rc2
White can also opt for a more direct approach: 16 Nxc6 bxc6 17 Ne5 Qa5 (perhaps Black should try
the more solid 17 ... Qf6 18 Nd3 Qd6 19 Nf4 Bg6 20 Nxg6 Qxg6 with approximately level chances) 18
Nxc6 Qxa2 was played in H.Hoang Canh-Tu Hoang Thong, Subic Bay 2009, and now White took
immediately on d5, missing the strong intermezzo 19 Qd4! followed by 20 Qxd5(+) with a clear
advantage.
 16 ... Nb4!?
A good move, aiming for queenside counterplay. Houdini likes the bizarre-looking 16 ... f6!?. The
idea seems to be to overprotect g7 (and so allow ... Nxd4 in some lines) while allowing ... Bf7 to cover
the d5-pawn if needed.
17 Rxc8 Qxc8 18 a3

18 ... Na6?!
Aiming to put the knight on the tempting c5-square, but it turns out not to be doing much there. The
natural 18 ... Nc6 seems perfectly acceptable for Black. As so often in the Tarrasch, the obvious,
centralizing move proves to be best.
19 Qc1
19 g4 Bg6 20 Nh4, followed by taking on g6 and playing b3-b4, gives White some advantage.
19 ... Nac5 20 Nh4 Be7
Again 20 ... f6 is Houdini’s recommendation, for reasons which will become clear on the next
move.
 21 Ndf5!
Giri has managed to obtain co-ordination in an unusual way. Now the weakness of g7 is apparent.
21 ... Bxh4
Allowing a nice tactic. Swallowing his pride and retreating with 21 ... Bf8 would have kept White’s
advantage within manageable bounds.

Exercise: Find a shot for White.

Answer: 22 Nxg7!
That’s a big pawn to lose.
22 ... Bg5
22 ... Bxe2 23 Nxe8 Bg5 24 Nf6+ Bxf6 25 Bxf6 is similar – White has a strong initiative since his
bishops are so powerful.
23 Qd1 Bg6 24 Nxe8 Qxe8 25 Qxd5
 Black’s position collapses.
25 ... Qe6 26 Rd1 Qxd5 27 Rxd5 Be7 28 b4 Na4 29 Ba1 Nec3 30 Rd7 Bf6 31 Kf1 b5 32 Rxa7
Kg7 33 g4 h6 34 Bc6 Bc2 35 Ra6 Bd4 36 e3 Bd3+ 37 Ke1 Bf6 38 h4 Bc4 39 g5 hxg5 40 hxg5 Bxg5 41
Rxa4 1-0

Game 17
V.Korchnoi-S.Collins
San Sebastian 2011

1 d4 d5 2 c4 e6 3 Nf3 c5 4 cxd5 exd5 5 g3 Nc6 6 Bg2 Nf6 7 0-0 Be7 8 Be3

A favourite of Korchnoi, who explains the ideas as follows: “One of the best moves in this position
is 8 Be3, and if 8 ... c4 9 b3 cxb3 10 Qxb3, when the d5-pawn is weak and still requires defending. In
addition, White has pressure on the half-open b-file. ” Korchnoi goes on to note that, while he has not
won all his games with this setup, he has always had an advantage in the middlegame.
Looking ahead, I can say that this passage perfectly describes the current game!
 8 ... c4
8 ... cxd4 9 Nxd4 0-0 10 Nc3 leads to a position from the 9 Bg5 cxd4 variation, but with the black h-
pawn on h7 rather than h6. As ... h7-h6 is generally a useful move, I can’t recommend this line for Black.
9 b3
9 Ne5 0-0 10 b3 (10 Nxc6 bxc6 11 b3 Qa5 12 bxc4 dxc4 was fine for Black in J.Glud-R.Alonso
Garcia, Catalonian Team Championship 2012, since 13 Bxc6 Bh3 14 Bxa8 Rxa8 gives full compensation
for the pawn; White can’t hold on to the exchange with 15 Re1? a s 15 ... Ng4! leads to a very strong
attack) 10 ... cxb3 11 Qxb3 transposes to the main game, while 11 axb3 Nb4 12 Nd3 Bf5 13 Nxb4 Bxb4
was very comfortable for Black in J.Glud-S.Brynell, Danish Team Championship 2010 (½-½ in 31).
9 ... cxb3 10 Qxb3 0-0 11 Ne5
11 Rc1 h6 12 Ne5 Na5 13 Qa4 Be6 14 Bd2 Nc4 15 Nxc4 dxc4 16 e3 Qc8 17 Qb5 Rb8 18 e4 a6 19
Qb2 b5 was level in V.Epishin-V.Akobian, Minneapolis 2005 (½-½ in 32).

11 ... Na5
Instead:
a) The computer’s preference 11 ... Bd6 has been tested in a game: 12 Nc3 Nxe5 13 dxe5 Bxe5 14
Rad1 Bxc3 15 Qxc3 Be6 16 Bd4 b6 and White had full compensation in Se.Ivanov-C.Jepson, Swedish
Team Championship 2012 (1-0 in 54) – of course, he can also just take on f6 and d5 with an endgame
which should probably be drawn.
b) 11 ... Qb6 has a great lineage: 12 Rc1 (Kasparov gives 12 Qxb6 axb6 13 Rc1 Nb4 14 Nc3 Be6
15 a3 Nc6 16 Nb5 Rfc8 as unclear) 12 ... Qxb3 13 axb3 Nb4 with an endgame which Black went on to
win in V.Korchnoi-G.Kasparov, Candidates (6th matchgame), London 1983 (0-1 in 78). Worth
investigating if you don’t like the main line, but I think White should be a bit better in this endgame (his
rooks are doing a good job pounding the queenside), so I prefer 11 ... Na5.
12 Qd3
12 Qa4 Be6 (12 ... a6 13 Bd2 Nc4 14 Nxc4 b5 15 Qa5 bxc4 16 Qxd8 Rxd8 17 Ba5 Rd7 18 Nc3
Bb7 19 Rab1 is some old analysis of Kasparov’s, which the great man thought led to equality – modern
computers prefer White here, but 17 ... Re8! equalizes) and now:
 a) 13 Bd2 Nc4 14 Nxc4 dxc4 15 e3 was J.Norri-K.Kiik, Finnish Team Championship 2013. Now
Black played 15 ... Qc8, seeking to trade light-squared bishops with ... Bh3. The simple 15 ... Bd5 looks
fine too.
b) 13 Rc1 was S.Baumegger-P.Pisk, Austrian League 2013. Here Black played 13 ... b6, but he
could have been more principled: 13 ... a6!, preparing ... Nc4 with a good game. Note that, with the queen
on a4, this can be played immediately since Nxc4 can be met by the intermezzo ... b7-b5. Otherwise, just
play ... b7-b5 and then ... Nc4.
12 ... Be6 13 Nc3 Rc8

14 Rab1!
A strong move, preventing Black’s idea of planting his knight on c4. Instead, 14 Bg5 Nc4 15 Rab1
Nd6 16 Rbc1 Qa5 17 Bd2 Qd8 18 Bg5 was drawn in Y.Kruppa-V.Potkin, Kiev 2001.
14 ... g6 15 Qd2
 15 ... Nc4
This is the move Black wants to play, but I didn’t assess the resulting position properly. The
alternatives are sounder:
a) 15 ... a6 16 Na4 (16 a4 seems more principled) 16 ... b5 17 Nc5 Bf5 18 Rbc1, as in L.Ftacnik-
V.Vojtek, Slovakian Team Championship 2012, could be well met by 18 ... Nc4 19 Nxc4 when either
capture on c4 is fine for Black. I prefer 19 ... bxc4!, keeping a stake in the centre and, provided we can
prevent White from playing e2-e4, giving Black every reason to play for a win.
b) 15 ... b6 is more solid. Then 16 Nb5 Nc4 17 Nxc4 Rxc4 is equal, since 18 Nxa7 Qd7 19 Nb5
Rfc8 gives Black full compensation for the pawn.
16 Nxc4 Rxc4

Exercise: How should White play?

Answer: 17 Rxb7!
This move, which is based on an exchange sacrifice, was confidently cracked out by Korchnoi and
Black’s position immediately becomes critical.
17 ... Bb4 18 Rxb4! Rxb4 19 Bg5

Now we can see White’s compensation. He already has a pawn for the exchange, plus a much better
structure. The dark squares around my king are just horrible, and the pin of the f6-knight is extremely
annoying.
Still, we have to make a move I suppose ...
19 ... Rc4 20 f4?
Going for the knockout punch. This was the move I had feared but, in fact, it squanders White’s
advantage. The straightforward 20 Nxd5? doesn’t work either: 20 ... Bxd5 21 Bxd5 Qxd5 22 Bxf6 Qf5!
and Black defends his king while preparing his counterplay with ... Rc2.
Instead, the calm 20 Rd1 would have kept some advantage for White; for instance, 20 ... Re8 21
Nxd5 Bxd5 22 Bxf6 Qxf6 23 Bxd5 Rc7 and White has two pawns for the exchange plus a monster bishop
on d5. Black should be able to draw with careful defence (an exchange sacrifice on d5, going into a
possibly tenable pawn down position with major pieces, might be a resource), but White is having all the
fun.
 Exercise: Find a way for Black not to lose on the spot.

Answer: 20 ... h6!!

Probably the best move I’ve ever played, especially considering the calibre of my opponent and
how the position turns 180 degrees.
a) 20 ... Qa5 21 f5! is not playable for Black: 21 ... Qxc3 22 Qxc3 Rxc3 23 fxe6 Ne4 (23 ... fxe6 24
Bxf6 followed by Bh3 is horrible) 24 Bxe4 dxe4 25 d5 wins for White in view of 25 ... fxe6 26 d6.
b) Partial credit for 20 ... Re8!?. White is better after 21 f5! (21 Bh4 is dangerous and perhaps
stronger) 21 ... Bxf5 (21 ... gxf5 22 Bh4 again looks horrible for Black) 22 g4 Rxc3 23 Bxf6 Rc2 24 Qxc2
Bxc2 25 Bxd8 Rxd8 26 Kf2 is a slightly favourable endgame for White, but Black should hold.

21 Bxh6
21 Bh4 Qa5 22 f5 Qxc3 23 Qxc3 Rxc3 24 fxe6 g5! is the difference compared to 20 ... Qa5. After 25
e7 Re8 26 Bxg5 hxg5 27 Rxf6 Rxe7 28 Kf2 Rc2, Black is the one with winning chances.
21 ... Qb6
 This, of course, is the point (otherwise Black’s last move would have been poor). The counterplay
on the a7-g1 diagonal based on ... Rxd4 and, if necessary, ... Ng4 is actually quite difficult to deal with.
22 h3??
Taking away the g4-square and so preparing Qe3 in response to ... Rxd4, but it wastes crucial time.
Instead:
a) 22 e3?? also loses to 22 ... Rfc8 with perfect co-ordination – see how the white queen can never
reach the dark squares on the kingside.
b) 22 Rb1 is logical, aiming for a queen trade, but after 22 ... Rxd4 23 Rxb6 Rxd2, the black rook
has broken through and the white king is still uncomfortable. On 24 Rd6 I was planning to continue in
dynamic style with 24 ... Ng4!? (24 ... Rc2! is even stronger) 25 Bxf8 d4, when the mating net forces
White to search for a draw with 26 Rxd4 (after 26 h3 dxc3 27 hxg4 Kxf8, the passed c-pawn looks too
strong) 26 ... Rxd4, which will be difficult to achieve since he is losing a pawn after 27 Ba3 Bxa2!.
c) 22 Bxf8! Rxd4 23 Qe1! was the only defence, which I’m pretty sure I had missed. Luckily Black
has more than enough compensation after 23 ... Rc4+ 24 e3 Kxf8 in view of his extremely active pieces
and the continuing weakness of the white king, but White should have gone for this.
22 ... Rfc8!
Actually 22 ... Rxd4 23 Qe3 Re8 was winning too, but my move is stronger.
23 Kh2
 23 ... Qxd4
The computer prefers taking on c3 (or even the bizarre 23 ... Bf5!?!?), but I was delighted to trade
into an endgame where White’s queenside will be demolished by my rooks.
24 Qxd4 Rxd4 25 Nb5 Ra4 26 f5
The only attempt at counterplay, but it’s not sufficient.
26 ... gxf5 27 Nd6 Rc2 28 Nxf5 Bxf5 29 Rxf5 Rxe2
The threat of doubling on the seventh means that Black saves his knight.
30 Kg1 Raxa2 31 Bf1 Re6

Now only basic care is required to bring home the full point.
32 Bf4 a5 33 Be5 Ne4 34 h4 Rg6 35 Rf3 a4 36 Rd3 Rd2 37 Ra3 Re6 38 Bf4 Rd4 39 Kg2 Rb6 40
Ra2 Rb3 41 Kh3 a3 42 Rc2 Rc3 43 Ra2 f6 44 h5 Rb4 45 Bg2 Rb2 46 Ra1 a2 47 h6 Rcc2 0-1

Key Notes
1. The double fianchetto with b2-b3 is a playable way to get a complex position but is not a serious
try for an advantage. Black was doing just fine for a long time in Giri-Swinkels (Game 16).
 2. Korchnoi’s Be3 leads to positions which have more in common with 9 Bg5 c4 than with the 9 Bg5
cxd4 lines. If Black can get a knight to c4 he tends to be doing fine.
 Chapter Six
Symmetrical Tarrasch
1 d4 d5 2 c4 e6 3 Nc3 c5 4 e3 Nf6 5 Nf3 Nc6

The Symmetrical Tarrasch is fundamentally different to the fianchetto lines we have just discussed at
length. White abandons hope of creating early pressure on d5, since his light-square bishop no longer
attacks the pawn, while its dark-squared counterpart can’t come to g5 anymore. Indeed, it’s no longer
clear if there will be any IQP to attack, and in fact White is often the side which ends up with the IQP in
this line (albeit not in the variations I recommend for Black).
The black side of this position (in contrast to the regular Tarrasch) is regularly endorsed by 2700+
GMs via a Symmetrical English move order (for instance, 1 c4 c5 2 Nf3 Nc6 3 Nc3 Nf6 4 e3, when
Black often plays 4 ... e6 5 d4 d5). Accordingly we are on slightly firmer ground from a theoretical
standpoint here. My main concern with Black’s position in the previous lines was hanging on to the d5-
pawn without making major concessions, and so I (and most Tarrasch players, I expect) automatically feel
more comfortable when the white pawn lands on e3. Nevertheless, this is a serious line and demands
serious work.
Black has a wide choice from a structural perspective, and many lines involve early ... c5xd4 or ...
d5xc4 trades. My choice is more consistent with the Tarrasch, keeping the central tension for as long as
possible and inviting White to give us an IQP. This choice is only partially motivated by principle; if it
were otherwise, then all sorts of c4xd5 and e2-e3 move orders could throw us out of book.
White has many reasonable moves here, of which the mysterious-looking ...
6 a3
... is by far the most popular. Control of the b4-square is useful either if White gets an IQP (after ...
c5xd4, e3xd4 ... d5xc4) or if White expands on the queenside with d4xc5 and b2-b4. See Aronian-
Melkumyan (Game 18) for some of the latest developments.
White’s many alternatives, including the popular and principled 6 cxd5, are covered in Carlsen-
Radjabov (Game 19). As always, don’t let the result of the game fool you into thinking that Magnus got
anything out of the opening.

Game 18
L.Aronian-H.Melkumyan
German League 2013

1 c4 c5 2 Nf3 Nc6 3 Nc3 Nf6 4 e3 e6 5 d4 d5 6 a3

A useful waiting move. White controls the b4-square and, in some lines, prepares d4xc5 and b2-b4
followed by Bb2. In addition, by keeping his bishop on f1, he ensures that ... d5xc4 can be met by Bxc4
without loss of time.
6 ... a6
This is my recommendation, as it seems most consistent with the main ideas of the Tarrasch. Black
is still happy to play with an IQP. By contrast, 6 ... cxd4 7 exd4 Be7 leads to a position where White
could end up with an IQP himself – although 8 c5!? has become fashionable, gaining queenside space and
playing for a bind.
7 dxc5
White prepares to expand on the queenside. Alternatively:
a) 7 Bd3 is, on one level, a ridiculous move, for a reason which will become apparent; i.e. 7 ... dxc4
8 Bxc4 b5. If you want ideas on how to play this position, check the main game, since White’s loss of
tempo means that you now have the same position as after Aronian’s 9th move, but with colours reversed!
Still, it can be confusing playing a white opening with the black pieces, so here are a couple of examples:
9 Ba2 Bb7 10 0-0 Qc7 11 Qe2 Rd8 12 Rd1 Be7 13 dxc5 (13 d5 exd5 14 Nxd5 Nxd5 15 Bxd5 was
agreed drawn in D.Lemos-K.Mekhitarian, Villa Martelli 2012) 13 ... Rxd1+ 14 Nxd1 Ne5 15 Ne1 Bxc5
was good for Black in A.Beliavsky-R.Ruck, Slovenian Team Championship 2012 (½-½ in 60).
b) 7 h3 cxd4 (7 ... h6 8 Bd3 Be7 9 0-0 0-0 was also decent in S.Ernst-C.Daly, Dublin 2012) 8 exd4
dxc4 9 Bxc4 Be7 10 0-0 0-0, as in R.Kempinski-K.Piorun, Polish Championship, Chorzow 2013, looks
like a comfortable IQP position for Black, since h2-h3 isn’t especially useful.
c ) Keeping the tension with the rare 7 b3 is interesting and can lead to numerous different pawn
structures depending on whether Black captures on c4 and/or d4. I suggest the simple 7…Be7 followed
by castling, after which a complex middlegame results where Black has a good share of the chances..
d) 7 cxd5 exd5 8 Be2 (8 Ne5 Bd6 9 Nxc6 bxc6 10 dxc5 Bxc5 was at least not worse for Black in
D.Lintchevski-A.Smirnov, Kazan 2013) 8 ... c4 (8 ... Be7 9 dxc5 Bxc5 10 0-0 0-0 11 Qc2 Bd6 12 Rd1
Be6 was equal in Zhao Xue-B.Kovanova, Russia-China rapid match, St. Petersburg 2012) 9 Ne5 Bd6 10
f4 was L.Lenic-J.Bejtovic, European Championship, Plovdiv 2012, and now 10 ... Na5!, aiming for the
b3-square, is good for Black.
7 ... Bxc5 8 b4

8 ... Ba7
Again Black has a choice of bishop retreats.
a) 8 ... Bb6 seems less logical to me since Black can no longer play ... b7-b5 easily. All the same, 9
Bb2 dxc4 10 Qxd8+ Bxd8 11 Bxc4 Bd7 12 Ke2 Rc8 13 Bd3 0-0 14 Rac1 Be7 led to a pretty level
endgame in Z.Ribli-J.Pinter, Austrian League 2010 (½-½ in 32).
b) 8 ... Bd6 9 Bb2 and now:

b1) 9 ... a5 was tried a couple of times by Akobian at the 2008 World Mind Sports Games, but with
poor results: 10 c5 axb4 (10 ... Be7 11 b5 Nb8 12 Na4 was already good for White in J.Markos-
V.Akobian, Beijing rapid 2008; 1-0 in 55 ) 11 axb4 Rxa1 12 Qxa1 Be7 13 Qa3 0-0 14 Bb5 Bd7 15 0-0
(White has an edge due to his queenside space advantage) 15 ... b6?! (15 ... Qc7 at once is preferable) 16
Qa4 Qc7 was A.Korobov-V.Akobian, Beijing blitz 2008, and now, instead of 17 Nd4, White should have
played 17 Bxc6 Bxc6 18 b5 with a clear advantage.
b2) 9 ... 0-0 10 cxd5 exd5 11 Be2 Be6 12 0-0 and Black has a decent position. For example, 12 ...
Rc 8 (12 ... Qe7 13 Rc1 Rfe8 14 h3 Rad8 15 Nd4 Ne5 was fine for Black in R.Laxman-N.Short,
Commonwealth Championship, Gauteng 2011; 0-1 in 45) 13 Rc1 Qe7 (13 ... Bb8 14 Na4 Qe7 15 Nc5
Ne4 16 Qd3 Nxc5 17 Rxc5 a5 18 Rb5 axb4 19 axb4 f6 20 Nd4 was agreed drawn in V.Potkin-E.Bacrot,
Moscow 2009) 14 Qd3 Rfd8 15 Rfd1 b5 and Black had his full share of the chances in L.Le Quang-
Y.Pelletier, French Team Championship 2012 (½-½ in 55).

9 Bb2
9 cxd5 exd5 10 b5 (10 Be2 0-0 11 0-0 d4 12 exd4 Nxd4 13 Nxd4 Qxd4 14 Qxd4 Bxd4 is equal,
S.Mamedyarov-A.Istratescu, European Team Championship, Gothenburg 2005; ½-½ in 23) 10 ... d4 (also
possible is 10 ... axb5 11 Nxb5 Bb6, intending ... Ba5+) 11 bxc6 dxc3 12 Qxd8+ Kxd8 13 Ne5 Kc7 14
Bc4 Re8 15 f4 (15 cxb7 Bxb7 16 Nxf7 was better, but Black is still for choice after 16 ... Re4) 15 ... bxc6
16 Bxf7 Re7 17 a4 Ng4 18 Nxg4 Bxg4 with a clear advantage for Black in A.Aleksandrov-V.Gashimov,
Warsaw (rapid) 2009 (0-1 in 64).
9 ... 0-0
 10 Qc2
The best move, but if you think White is messing around while Black is developing pieces and
seizing space in the middle, you’re basically right.
a) 10 Be2 dxc4 11 Bxc4 Qxd1+ 12 Rxd1 Rd8 13 Ke2 (13 Rxd8+ Nxd8 14 Ke2 b5 15 Bd3 Bb7 was
level in A.Onischuk-G.Kaidanov, US Championship, St. Louis 2012) 13 ... Bd7 14 Bd3 Rac8 15 Rd2 Be8
16 Ne4 Nxe4 17 Bxe4 Rxd2+ 18 Nxd2 Ne7 19 Rc1 Rxc1 20 Bxc1 Bb5+ 21 Kd1 Ba4+ 22 Ke2 Bb5+ 23
Kd1 Ba4+ 24 Ke2 and a draw was agreed in Zhou Jianchao-Bu Xiangzhi, Chinese Championship,
Xinghua 2013.
b) The following game is a great demonstration of the attacking potential of the IQP: 10 cxd5 exd5
(this already looks good for Black since White can’t conveniently stop ... d5-d4) 11 Na4? (Moskalenko
suggests 11 Be2 d4 12 exd4 Nxd4 13 Nxd4 Bxd4 14 0-0 Be6, limiting Black to an edge)

Exercise: How should Black continue?

Answer: 11 ... d4! (a thematic breakthrough, opening the e-file while the white king is stuck in the
centre) 12 exd4 (Moskalenko suggests 12 Nxd4 Nxd4 13 Bxd4 Bxd4 14 Qxd4 Qxd4 15 exd4 Re8+ 16
Be2 Bg4 with some initiative for Black) 12 ... Bg4 13 Be2 Bxf3! (eliminating a defender of the d4-pawn
and the e-file) 14 Bxf3 Re8+ 15 Kf1 (Moskalenko notes that 15 Be2 fails to 15 ... Nxd4 16 Bxd4 Bxd4 17
Rc1 b5 – and 17 ... Bxf2+ 18 Kxf2 Ne4+ 19 Kg1 Qg5, with a decisive attack, is even stronger) 15 ...
Nxd4 16 Bxd4 (perhaps White should have tried 16 Nc5 Bxc5 17 bxc5 Nxf3 18 Qxd8 Rexd8 19 gxf3 with
some saving chances, but not 18 Qxf3?? Ne4! and wins – Moskalenko) 16 ... Bxd4 17 Rc1 (17 Ra2 was
better, though Black is still on top after 17 ... Ne4) 17 ... Ne4 (the black pieces dominate and it’s hard to
offer White any constructive advice) 18 Bxe4 Rxe4 19 g3 Re6 20 Rc2 Qd5 21 Rg1 Rae8 22 Nc5 Rd6 23
Qg4 h5 24 Qf4 Bxc5 25 Rxc5 Re1+! and White resigned in S.Halkias-V.Moskalenko, Chalkida 1998.

10 ... d4
The immediate central breakthrough has proven very reliable in recent games, though Black’s choice
here is a matter of taste.
a) 10 ... Qe7 (chosen by Anand against Aronian in the 2011 Amber rapid event) 11 Rd1 (11 Bd3
dxc4 12 Bxc4 was successful in both T.Nyback-D.Tjiam, European Cup, Plovdiv 2010, and O.Cvitan-
O.Kurmann, Swiss Team Championship 2012, but for some reason in both games Black avoided the
natural 12 ... b5 13 Bd3 Bb7 followed by ... Rac8 with equality; 11 Be2 h6 12 0-0 dxc4 13 Bxc4 b5 14
Ba2 Bb7 was nothing for White in R.Vasquez Schroder-M.Cornejo, Cochabamba 2013, 0-1 in 45 ) 11 ...
Rd8 12 Be2 and, frankly, it’s hard to see how White can hope for an advantage playing in such a way.
 a1) 12 ... dxc4 13 Rxd8+ Nxd8 14 Ne4 (14 Bxc4 b5 15 Bd3 Bb7 was equal in T.Georgescu-
D.Dumitrache, Cap Aurora 2013; ½-½ in 31) 14 ... Nxe4 15 Qxe4 Bd7 16 Bxc4 Rc8 and Black had no
problems in K.Miton-B.Macieja, Polish Championship, Warsaw 2012 (½-½ in 31).
a2) 12 ... d4 13 exd4 Nxd4 14 Nxd4 Bxd4 15 0-0 e5 16 Na4 Be6 17 c5! gave White an edge in
L.Aronian-V.Anand, Monte Carlo (rapid) 2011 (1-0 in 59).
a3) 12 ... h6?! is a little slow: 13 0-0 dxc4 14 Rxd8+ Nxd8 15 Ne4 Ne8 16 Bxc4 and White had the
initiative in H.Banikas-J.Borisek, Istanbul Olympiad 2012 (½-½ in 64).
b) 10 ... dxc4 11 Bxc4 Qe7 12 Ba2 Bd7 13 0-0 Rac8 14 Rad1 Rfd8 also looked solid and sound in
N.Vitiugov-K.Sakaev, FIDE World Cup, Khanty-Mansiysk 2007 (0-1 in 92).
11 exd4 Nxd4 12 Nxd4

12 ... Bxd4!
12 ... Qxd4 is less precise. 13 Bd3 Qh4 14 g3 Qh6 15 Ne4 Nxe4 16 Bxe4 and Black’s pieces were
poorly placed in T.Nyback-J.M.Degraeve, German League 2009 (1-0 in 39).
13 Be2 e5 14 0-0 Be6 15 Na4
15 Rad1 Qc7 16 c5 Rad8 17 Bf3 h6 18 Rfe1 Rfe8 didn’t offer White much in T.Sanikidze-E.Bacrot,
Baden-Baden 2013 (½-½ in 49).

15 ... Bxb2
Instead of this move, Black can try to keep the tension. 15 ... Rc8 16 Rad1 Re8 17 Nc5 Qe7 18 Bxd4
exd4 19 Rxd4 didn’t give him enough for the pawn in V.Tkachiev-T.Markowski, European Blitz
Championship, Warsaw 2012; but 15 ... Qc7 16 Bxd4 exd4 17 h3 Rad8 was fine for Black in L.Le Quang-
Zhou Weiqi, Asian Cup, Zaozhuang 2012 (½-½ in 47).
16 Qxb2 Qc7 17 Nc5 Bf5 18 Rfe1 Rfe8 19 Rad1 a5 20 Bf1 axb4 21 axb4 b6 22 Nb3 h6

White has nothing here.

23 h3 e4 24 c5 bxc5 25 Nxc5 Qb6 26 Rd4 Rad8 27 Red1 Rxd4 28 Qxd4 Rc8 29 Ba6 Rb8 30 b5
Bg6 31 Qd6 Qa7 32 Rd4 Kh7 33 Na4 Re8 34 Nc5 Rb8 35 Rb4 Rb6 36 Qe5 Qa8 37 Rd4 Rb8 38 Rd6
Qa7 39 Qd4 Qc7 40 Nd7 Nxd7 41 Rxd7 Qc1+ 42 Qd1 Qc3 43 Qd4 Qc1+ 44 Kh2 Qf4+ 45 g3 Qf3 46
Qe3 Qf5 47 Rd6 Qe5 48 Qf4 Qb2 49 Qe3 Qb4 50 Rd5 Qc4 51 Rd2 Qf1 52 g4 Re8 53 h4 h5 54 Rd5
Re6 55 g5 Qc4 56 Rc5 Qb4 57 Rd5 Qc4 58 Rc5 Qb4 59 Rd5 ½-½

Game 19
Ma.Carlsen-T.Radjabov
Sandnes 2013

1 c4 c5 2 Nf3 Nc6 3 Nc3 Nf6 4 e3 e6 5 d4 d5 6 cxd5

The most critical move. White’s attempts to delay the development of his king’s bishop just let
Black develop in comfort, as we saw in the last game, so releasing the central tension is the natural way
of trying to make use of White’s extra tempo.
6 ... exd5!

The natural reaction for a Tarrasch player. 6 ... Nxd5 is a position from the Semi-Tarrasch which is
a completely different opening, where White often ends up with an IQP himself.
7 Bb5
By putting pressure on the c6-knight White discourages ... c5xd4. The resulting positions are popular
with strong GMs who presumably think that White has chances of an edge, but I really don’t see why
Black should have anything to worry about.
Instead, 7 Be2 has been tested in over a thousand games. Going into similar positions to the main
game with 7 ... Bd6 (or 7 ... Be7) 8 dxc5, although fully playable, leaves White’s bishop better placed on
e2 than on b5. I would suggest instead 7 ... cxd4 8 Nxd4 Bd6 as a simple and reliable solution.
 a) 9 Bf3 0-0 10 0-0 Be5 is fine for Black:
a1) 11 Qd3 Bg4 (11 ... Bc7 12 Nxc6?! bxc6 13 e4 dxe4 14 Bxe4 Nxe4 15 Qxe4 Qf6 and Black was
already better in T.Banusz-A.Volokitin, European Championship, Rijeka 2010 ) 12 Bxg4 Nxg4 13 Nf3
Qd6 14 h3 Bxc3 15 Qxc3 Nf6 was equal in Z.Ribli-J.Borisek, Slovenian Team Championship 2008 (½-½
in 60).
a2) 11 Be2?! a6 12 Nf3 Bc7 13 b3 Re8 14 Bb2 Qd6 15 g3 Bh3 16 Re1 Rad8 gave Black an ideal
IQP formation in G.Izsak-P.Tregubov, German League 2013 – in fact, with White’s loss of a tempo (Be2-
f3-e2 spent two extra moves, compared to one extra move on ... Bd6-e5-c7), I think we have directly
transposed into an IQP position commonly arising from the 2 ... d5 line of the c3-Sicilian, and in which
the side with the IQP scores heavily.
b) 9 0-0 0-0 is logical.

b1) 10 Re1 Re8 11 Nxc6? bxc6 12 b3 was K.Drozdowski-M.Parligras, European Championship,

Legnica 2013, and now 12 ... Qc7 13 g3 h5! would have left Black clearly better since the h-pawn is
untouchable. As often happens in IQP positions, after a couple of passive moves White now seems to be
playing “as Black”, with less space and facing a strong initiative.
 b2) The danger inherent in these positions for White is well illustrated by the following game: 10
Qd3 Re8 11 Nxc6? (why does everyone keep taking on c6 in these positions? – the black pawn structure
is considerably improved and the d4-knight can no longer drop back to cover the kingside) 11 ... bxc6 12
b3 Ng4 (12 ... Qc7 13 g3 Qd7, aiming for h3, looks dangerous as well) 13 Bxg4 (strategic capitulation;
White had to try 13 h3) 13 ... Bxg4 14 Bb2 Re6? (14 ... Qh4 15 f4 Re6 was much stronger) 15 f4 (15 Ne2
kept chances to defend) 15 ... Qe7 16 Rae1 Bc5 17 Bc1 g6! 18 Bd2 Bf5 19 Qa6 Bxe3+ and White
resigned in J.Lopez Martinez-A.Fier, Sabadell 2010.
b3) 10 b3 Nxd4 11 Qxd4 Re8 12 Bb2 Be5 13 Qd2 was D.Khismatullin-S.Ionov, Russian Team
Championship 2010, and now Black could have equalized comfortably with 13 ... Bg4 14 Bxg4 Nxg4 15
h3 Nf6 16 Rfd1 Rc8 17 Qd3 Qa5.
7 ... Bd6 8 0-0 0-0 9 dxc5 Bxc5 10 b3 Bg4 11 Bb2

Question: How do you assess this position?

Answer: I think any Tarrasch player should be very happy with a position like this. He is fully
developed, all the pieces are on the board, and the d5-pawn isn’t under serious pressure. Black’s
challenge is to mobilize his major pieces, and if he can generate a kingside initiative, so much the better.
11 ... a6
Black has tried a range of moves in this position. We will only focus on a few of the more logical
ones.
11 ... Bd6 has been extensively tested. We will confine ourselves to one example: 12 h3 Bh5 13 Be2
a6 14 Nh4 Bxe2 15 Nxe2 Re8 16 Rc1 Ne4 17 Nf5 Be5 18 Bxe5 Rxe5 was level in V.Kramnik-J.Polgar,
Baku (rapid) 2010 (½-½ in 67).
The most popular line is 11 ... Rc8 12 Rc1 and then:
 a) 12 ... a6 13 Be2 (here 13 Nxd5! seems good for White; for instance, 13 ... Qxd5 14 Qxd5 Nxd5
15 Rxc5 Ncb4 16 Rxc8 Rxc8 17 Bc4 and Black needs to prove his compensation for the pawn) 13 ... Ba7
14 h3 Bh5 15 Nh4 Bg6 16 Nxg6 hxg6 17 Bf3 d4 18 exd4 Bxd4 19 Na4 Bxb2 20 Nxb2 Nd4 21 Rxc8
Nxf3+ 22 Qxf3 Qxc8 was level in V.Topalov-S.Karjakin, Stavanger (blitz) 2013 (0-1 in 64).
b) 12 ... Qd6 has also been tested in grandmaster games:
b1) 13 Be2 a6 14 Nd4 (14 h3 Bxf3 15 Bxf3 Rfd8 16 Qc2 was Z.Ribli-A.Adorjan, Riga Interzonal
playoff, Budapest 1979, and now 16 ... Ne5 would have equalized) 14 ... Bxd4 15 Bxg4 Nxg4 16 Qxg4
Be5 17 Rfd1 Bxh2+ 18 Kf1 Qe6 19 Qf3 Be5 20 Ba3 Bd6 21 Bxd6 Qxd6 and a draw was agreed in
J.Speelman-L.Portisch, World Cup, Brussels 1988.
b2) 13 Na4 Ba3 14 Bxa3 Qxa3 15 Bxc6 Bxf3 16 gxf3 Rxc6 17 Rxc6 bxc6 was equal in T.Hillarp
Persson-S.Brynell, Swedish Team Championship 2006 (½-½ in 56).
c) 12 ... Bd6 leaves White with a wide choice:

c1) 13 Ne2 Bxf3 14 gxf3 Be5! was level in S.Kalinitschew-M.Thesing, German League 2003, and
B.Golubovic-B.Vuckovic, European Championship, Rijeka 2010. A draw by repetition after 15 Ba3 Bd6
16 Bb2 Be5 would be a logical finish.
 c2) 13 Bxc6 bxc6 14 Ne2 Bxf3 15 gxf3 Re8 16 Ng3 left Black with no problems in I.Rozum-
D.Ayupov, Kazan 2013 (1-0 in 60).
c3) 13 h3 Bh5 (13 ... Be6, avoiding the exchange of light-squared bishops, is also sensible, as in
Cu.Hansen-B.Gulko, Malmö 2001) 14 Be2 a6 15 Nh4 Bxe2 16 Nxe2 Be5 17 Bxe5 Nxe5 18 Qd4 Re8 19
Nf5 Rxc1 20 Rxc1 Qd7 21 Nfg3 h6 with a typical position in B.Grachev-N.Chadaev, Russian Team
Championship 2013. White has managed to exchange several pieces and definitely enjoys an edge (1-0 in
55).
c4) 13 Be2 is more critical. Now Black has tried a number of moves but we will focus on the
caveman approach 13 ... Bb8 which has been endorsed by Kasparov and Spraggett.

c41) 14 h3 Bh5 15 Nh4 (a typical idea, aiming to trade Black’s light-squared bishop) 15 ... Qd6 (15
... Bxe2 16 Nxe2 Re8 17 Nf5 Qd7 18 Neg3 Be5 basically equalized in M.Voiska-I.Farago, Bolzano 2000;
White was actually a bit worse after 19 Ba3?!, while on something like 19 Bxe5 Rxe5, if White does have
an advantage, it is minimal) 16 g3 Bxe2 (16 ... Bg6 17 Nb5 Qe6 18 Bxf6 Qxf6 19 Nxg6 hxg6 was
Y.Seirawan-N.De Firmian, US Championship, Key West 1994, when White could have simply taken the
pawn) 17 Nxe2 (this position is assessed as better for White by both Ftacnik and Curt Hansen) 17 ... Rfd8
(17 ... Qe6 18 Kg2 Be5 19 Bxe5 Qxe5 20 Nf3 Qe4 21 Nc3 Qe6 22 Qd3 Ne5 23 Nxe5 Qxe5 24 Rc2 was
beginning to look like a trademark Karpovian squeeze in A.Karpov-J.Lautier, Monte Carlo blindfold
rapid 2000; 1-0 in 93) 18 Nd4 Ne5 19 Rxc8 Rxc8 20 Ndf3 Qe6 21 Ng5 Qd7 22 f4 Ng6 23 Nxg6 hxg6 24
Qd3 a6 25 Bd4 Ne4 26 Nxe4 dxe4 27 Qxe4 Qxh3 28 Qg2 Qg4 29 Qf3 Qh3 30 Qg2 Qg4 31 Qf3 Qh3 and
a draw was agreed in V.Zvjaginsev-C.Lutz, Essen 2002.
c42) 14 Re1 Re8 15 g3 h5!? showed a typical response to g2-g3, increasing Black’s control over g4
and preparing to soften up the white kingside pawns with ... h5-h4. Black went on to lose a mammoth
game in T.V.Petrosian-B.Ivanovic, Bar 1980 (1-0 in 123), but it had nothing to do with the opening.
c43) 14 Nb5 and now:
 c431) 14 ... Ne4 15 Nbd4 (15 Nfd4 Bxe2 16 Qxe2 was D.Svetushkin-D.Hummel, Canarias online
blitz 2004, when 16 ... Qg5 would have been active and strong) 15 ... Re8 16 h3 Bxf3 17 Nxf3 Qd6 18
Qd3, and now Black went on to win with the direct 18 ... Ng5!? in J.Sunye Neto-G.Kasparov, World
Student Team Championship, Graz 1981. Instead, 18 ... Rcd8 with equality is sounder; while the pawn
sacrifice 18 ... d4!? 19 exd4 Rcd8 with full compensation is interesting as well.
c432) 14 ... Re8 has been endorsed by Tarrasch specialists Murray Chandler and Kevin Spraggett:
15 Qd3 (or 15 Nbd4 Qd6 16 g3 Bh3 17 Re1 Ne4 with good play for Black in T.Markowski-K.Spraggett,
Cannes 1994) 15 ... Ne4 16 Nbd4 Qd6 17 g3 h5 18 Nxc6 bxc6 19 Nh4 Bh3 20 Rfd1 Qh6 21 Bf3 g5 22
Ng2 h4 and, while Black went on to lose in A.Karpov-M.Chandler, Bath 1983 (1-0 in 36), this again had
more to do with the strength of his opponent than with any defect in his position. Black has active and
promising play here.

12 Bxc6
Changing the structure and, with it, the plans. Taking the knight means the ... d5-d4 advance is not
really an issue for White anymore. He aims to build pressure on the c6-pawn after Rc1, and restrain any
freeing ... c6-c5 break.
 Alternatively, White can keep the pawn structure as it is with 12 Be2, after which 12 ... Ba7 is
logical to get the bishop out of the way.

a) 13 Nd4 Bd7 14 Nxc6 Bxc6 15 Bf3 was R.Loncar-L.Totsky, Zadar 2000, and now Black retained
the tension with 15 ... Qd6, which was okay (½-½ in 37), but had he wanted to he could have equalized
immediately with 15 ... d4 16 exd4 Bxd4 17 Bxc6 bxc6 18 Na4 c5.
b) 13 h3 Be6 14 Rc1 Qd6 15 Bd3 (15 Qd3 Rad8 16 Rfd1 Bb8 17 Nb1 Ne4 18 Ba3 Nb4 19 Qd4 a5
20 Nc3 was A.Ornstein-S.Brynell, Swedish Championship, Haparanda 1994, when 20 ... Nxc3 21 Rxc3
Rc8 would have retained the balance) 15 ... Rad8 16 Ne2 Ne4 17 Ned4 Ne5? (instead, 17 ... Nxd4 18
Bxd4 Bxd4 19 Nxd4 Bd7 is just an edge for White) 18 Nxe5 Qxe5 19 Nc6! Qxb2 20 Nxd8 Rxd8 21 Bxe4
was clearly better for White in L.Portisch-G.Ligterink, Wijk aan Zee 1985 (1-0 in 29).
c) 13 Rc1 Qd6 and now:

c1) 14 Nd4 Ne5? (Ftacnik gives 14 ... Bxd4 15 Bxg4 Be5 16 h3 d4 17 exd4 Nxd4 with equality) 15
f4! Bxe2 16 fxe5 Bxd1 17 exd6 Bg4 18 Rxf6!! Bxd4 19 exd4 gxf6 20 Nxd5 and the d-pawn was too
strong in V.Filippov-E.Romanov, Russian Team Championship 2005 (1-0 in 37).
c2) 14 h3 Bh5 (retreating to e6 or d7 would have avoided the exchange, with a balanced game) 15
Nh4!? Bg6 16 Nxg6 hxg6 17 Bf3 Rad8 18 Ne2 was J.Horvath-J.Borisek, Meissen 2013, and now simply
18 ... Rfe8 would have kept level chances.
12 ... bxc6 13 Rc1

13 ... Ba7
13 ... Bd6 has been more popular in practice: 14 Ne2 (moving the knight away from the centre with
14 Na4 inevitably gives Black some attacking chances; 14 ... Rc8 15 Qd3 was J.Sunye Neto-
L.Christiansen, World Student Team Championship, Mexico 1980, when simply completing development
with 15 ... Re8 would have left Black with a fine game) and:

a) 14 ... Bxf3 (a logical attempt to punish Ne2) 15 gxf3 Rc8 16 Kh1 (16 Qd3 Nd7 17 Ng3 has high-
level endorsement, albeit in a simul: G.Kasparov-Jo.Wilson, London simultaneous 2003; 1-0 in 56) 16 ...
Re8 17 Rg1 g6! (very logical, dealing with the pressure on the g-file, taking f5 under control, and
preparing ... Bd6-f8-g7) 18 Qd2 Bf8 19 b4 a5! 20 a3 (moving the b-pawn forward – to b5 or a5 – allows
20 ... c5 with active counterplay) 20 ... axb4 21 axb4 Bg7 and Black was fine in M.Ashley-N.De Firmian,
New York 1996 (0-1 in 75).
b) 14 ... Rc8 15 Ng3 Re8 16 Qd3 Ne4 (16 ... a5 is sensible too, as in I.V.Ivanov-M.Ashley, New
York 1994; ½-½ in 21 ) 17 Nd2 (here 17 Rc2 Nxg3 18 hxg3 Be6 19 Qxa6 Ra8 20 Qd3 Rxa2? 21 Bf6!
was A.Moiseenko-P.Leito, Bled Olympiad 2002, 1-0 in 45; but 20 ... c5 brings about a position where
White’s extra pawn isn’t really felt – see the rest of this note for more examples) 17 ... Nxg3 18 hxg3 h5

19 Nf3 (19 Qxa6 Qd7 gives full positional compensation for the pawn; White’s only active plan is
to push the a-pawn, which will leave him with queenside weaknesses – even Houdini wants to repeat
with 20 a4 f6 21 Qd3 Bf5 22 Qe2 Bg4 23 Qd3) 19 ... Bxf3 20 gxf3 Qg5 21 f4 Qg4 22 Kg2 h4 23 Rh1 was
F.Peralta-R.Swinkels, Wijk aan Zee 2013 (1-0 in 32), but Black could have played 19 ... Qd7! with
equality since, as noted before, taking on a6 leaves Black with full positional compensation.

14 Ne2
Keeping the knight within touching distance of the kingside and d4 seems sensible.
14 Na4 is more ambitious, but I would be happy to see such a move with Black:
 14 ... Qd6 15 Rc4?! Qe6 (15 ... Bxf3 16 gxf3 Rad8 17 Rc1 c5 was good too) 16 Nc5 Qc8 17 Rc1
Ne4 18 b4 (18 Nd3 is equal) 18 ... Bxc5! 19 bxc5 f6! changed the pawn structure in Black’s favour in
U.Nielsen-G.Sax, Vejle 1994 (0-1 in 46). The c6-pawn is protected by White’s c5-pawn, and the d5-
pawn is a rock. Black’s last move takes control of the e5-square and blunts the b2-bishop on the long
diagonal. After a few more moves Black’s position got even better: 20 Qd4 Bxf3 21 gxf3 Ng5 22 Qg4
Rb8 23 Bc3 Rb5! and White struggled to cover all his weaknesses.
14 ... Qd6 15 Be5 Qe7 16 Ned4

16 ... Bxf3?!
After this slightly passive move White has chances of an edge. Instead, 16 ... c5! seems to work
well:
 17 Nc6 Qe8 18 Bxf6 Qxc6 19 Be7!? Rfc8 and the hanging pawns are alive. White seems to have no
route to an advantage; for instance, 20 Ne5 (20 h3 Bf5 21 Nd4 Qd7 22 Nxf5 Qxf5 looks comfortable for
Black) 20 ... Bxd1 21 Nxc6 Bxb3 22 axb3 (22 Nxa7 Rxa7 23 Rxc5 Rxc5 24 Bxc5 Rc7 25 Bd6 Rc6 26
axb3 Rxd6 is another drawn rook and pawn endgame) 22 ... Rxc6 23 Rfd1 Rc7 24 Bd6 Rb7 25 Bxc5
Bxc5 26 Rxc5 Rxb3 27 Rcxd5 g6 with a draw.
17 Nxf3 Rfc8 18 Qd3

18 ... a5
18 ... Ne4 was another option, which White could have avoided by taking on f6 a move earlier.
19 Bxf6 Qxf6 20 Rc2 Rd8 21 Rfc1 c5 22 e4
22 Qb5 was tempting. Then 22 ... h6 (22 ... Bb6 23 Ne5! is good for White) 23 Qxa5 (after 23 h3 d4
24 exd4 cxd4 Black has decent counterplay) 23 ... Bb6 24 Qb5 Ra5 25 Qe2 Ra3 is probably better for
White, but it won’t be easy for him to mobilize his a-pawn.
22 ... Qg6 23 Re1 dxe4 24 Qxe4 Qxe4 25 Rxe4 Rd1+ 26 Re1 Rxe1+ 27 Nxe1 Rd8
 This endgame is equal. As so often, the fact that Magnus goes on to win does not suggest that he got
anything in the opening. In fact, this whole game strongly reminds me of Carlsen’s crucial win against
Radjabov in the 2013 London Candidates Tournament.
28 Kf1 a4 29 bxa4 Rd4 30 a5 Ra4 31 Rd2 Kf8 32 Nd3 f6 33 Nb2 Rxa5 34 Nc4 Ra4 35 Rc2 Ke7
36 Ke2 Ke6 37 Kd3 Kd5 38 a3 h5 39 h3 h4 40 Rc1 g6 41 Rc2 g5 42 Rc1 Ra6 43 Re1 Bb8 44 Re7 Bf4
45 Kc3 f5 46 Kb3 g4 47 a4 gxh3 48 gxh3 Rg6 49 a5 Rg1 50 a6 Rb1+ 51 Kc3 Rc1+ 52 Kd3 Rd1+ 53
Ke2 Ra1 54 Nb6+ Kd6 55 Rg7 Kc6 56 Rg6+ Kb5 57 Nd5 Be5 58 Rb6+ Kc4 59 Ne3+ Kc3 60 f4 Bd4
61 Nxf5 c4 62 Rc6 Rh1 63 Nd6 Rh2+ 64 Kf3 Kd3 65 Rxc4 Rxh3+ 66 Kg4 Rh1 67 Ra4 Bf2 68 Ra3+
1-0

Key Notes
1. The Symmetrical Tarrasch is a completely different concept to Rubinstein’s kingside fianchetto.
The resulting positions have more in common with other IQPs (such as those from the Nimzo-Indian),
since White’s bishop is developed more passively, normally to e2, and doesn’t put immediate pressure on
the d5-pawn.
2. 6 a3 is a move which has been popular for a long time. The equalizing approach in Aronian-
Melkumyan (Game 18) looks very reliable – if White has anything here, it’s not much. All the lines where
Black gets an IQP look more than fine for the second player.
3. 6 cxd5 exd5 leaves Black with an IQP and good chances. If, as in Carlsen-Radjabov (Game 19), White
captures on c6, a different structure arises, but one where Black is solid and has good chances of
counterplay.
Chapter Seven
Tarrasch Sidelines

1 d4 d5 2 c4 e6 3 Nf3 c5

As I hope has become clear by now, the Tarrasch is a principled opening and White should not be
able to hope for too much by avoiding the main lines. That said, main lines change over time and this
chapter contains one very serious attempt.
This is probably a good time to mention that, when confronted by a sideline (in the Tarrasch or
elsewhere), please do NOT rush your response. Sidelines by their nature have slightly offbeat ideas
which, if well neutralized, can leave Black with a pleasant game. However, often on seeing an early Bg5
or e2-e3 in the Tarrasch (I speak from bitter experience here, in particular from my game against
Bischoff) the natural tendency is to relax and play a quick move. Black still needs to be precise in the
sidelines as in every other Tarrasch variation – his structure is fundamentally more vulnerable than
White’s and we don’t have the luxury of second-rate responses. Apologies if this seems blatantly obvious,
but I have personally scored worse against sidelines than against main lines in the Tarrasch, and want to
save you from a similar experience.
Coincidentally, all the games in this section are played by Jacob Aagaard. Jacob has a great feel for
the Tarrasch and, in particular, is comfortable developing his pieces to good squares rather than searching
for ways to “justify” his compromised structure. The sidelines tried here have not troubled him.
4 cxd5
4 g3 Nf6 5 Bg2 Nc6 6 0-0 Be7 7 dxc5 Bxc5 leaves Black with a big pawn in the middle of the
board, and no weaknesses. Black should not be worse here and Schandorff-Aagaard (Game 22) shows
why.
4 ... exd5 5 Nc3 Nc6
 6 dxc5!?
This move has become one of White’s most serious and popular attempts, following Aagaard and
Ntirlis’ analysis which showed that Black needs to be precise to equalize. See L’Ami-Aagaard (Game
20) for a good example.
The less threatening 6 Bg5, together with 6 Bf4 and 6 e3, is covered in Rewitz-Aagaard (Game 21).

Game 20
E.L’Ami-J.Aagaard
Helsingor 2012

1 d4 d5 2 c4 e6 3 Nc3 c5 4 cxd5 exd5 5 Nf3 Nc6 6 dxc5!?

This was believed to be a harmless sideline until the monumental analytical work done by Aagaard
and Ntirlis demonstrated that Black’s traditional responses led to difficult positions. In fairness, the move
had already been discovered by Krasenkow, who tends to score 70-80% against a lot of Black’s main
defences to 1 d4 and so is a guy worth following.
 6 ... d4
Seizing space in the centre is the most principled response.
6 ... Nf6 leaves White with some advantage after 7 Be3!, whereas the most popular 7 Bg5 should be
met with the excellent move 7 ... Bxc5! (not yet played by anyone over 2000, as far as I can see) with very
comfortable play for Black, since 8 Bxf6 Qxf6 9 Qxd5 (the complications after 9 Nxd5 Qxb2 are in
Black’s favour) 9 ... Bb4 10 Rc1 Be6 11 Qe4 0-0 leaves White struggling to equalize. If he tries to hang
on to his extra material with 12 a3, then 12 ... Bxc3+ 13 Rxc3 Rad8 and White will need to show
something special to get out of the opening in one piece.
7 Na4
7 Ne4 was tried several times by Ivan Ivanisevic, but he seems to have shelved it after the following
defeat: 7 ... Bf5 (7 ... Qd5 8 Nd6+ Bxd6 9 cxd6 Qxd6 was also playable in I.Ivanisevic-G.M.Todorovic,
Serbian Team Championship 2005) 8 Ng3 Bg6 (an earlier game saw 8 ... Bg4 9 h3 Bxf3 10 exf3 Bxc5 11
a3 Nf6 12 Bd3 0-0 13 0-0, I.Ivanisevic-G.M.Todorovic, Serbian Team Championship 2004, and I prefer
White) 9 h4 h6 10 h5 Bh7 11 e4 Bxc5 12 Bd3 was I.Ivanisevic-R.Zelcic, Cannes 2006, and now apart
from 12 ... Bd6 (0-1 in 81), simply 12 ... Nge7 13 0-0 0-0 looks fine for Black.

7 ... Bxc5
Definitely the safest move. The alternatives are considerably more complicated but should lead to
some advantage for White, it seems:
a) 7 ... Bg4 8 a3 Be7 9 h3 Bh5 10 b4 and White was en route to a rapid victory in A.Dreev-
V.Yemelin, Tomsk 2006 (1-0 in 17). It’s hard to believe that Black has enough for the pawn here.
b) 7 ... b5 8 cxb6 axb6 9 e3 Bb4+ 10 Bd2 is clearly better for White, as most recently demonstrated
in F.Elsness-E.Lie, Norwegian Championship, Lillehammer 2013 (1-0 in 47).
c) I don’t trust 7 ... Nf6 8 e3 Bg4 9 Be2 either, since Black will struggle to regain his pawn.
8 Nxc5 Qa5+
Now White has a major decision to make.
9 Bd2
9 Qd2 Qxc5 is another branch:
 a) 10 e3 leaves Black with a pleasant choice:
a1) 10 ... Nf6!? is an interesting gambit, if not 100% sound: 11 Nxd4 (or 11 exd4 Qd5 12 Be2 0-0
13 0-0 Re8 with some compensation – but probably not enough – in M.Misojcic-N.Somborski, Vrnjacka
Banja 2008) 11 ... 0-0 12 Be2 (here 12 Qc3! looks like a better shot) 12 ... Ne4 13 Qc2 Qa5+ 14 Bd2
Nxd2 15 Qxd2 Qxd2+ 16 Kxd2 Rd8 17 Bf3 Nxd4 18 exd4 Rxd4+ and a draw was soon agreed in
G.Serper-V.Akobian, Reno 2004 (½-½ in 26).
a2) 10 ... dxe3 11 Qxe3+ Qxe3+ 12 Bxe3 Be6 (12 ... Nf6 13 Bc4 Be6 14 Bxe6 was agreed drawn in
R.Ris-A.Demuth, Wijk aan Zee 2012, but this looks like an edge for White) 13 Bb5 (or 13 Nd4 Nxd4 14
Bxd4 f6 15 Bb5+ Kf7 16 Rc1 Ne7 17 a3 Rac8 18 0-0 a6 19 Be2 Rxc1 20 Rxc1 Rd8 with approximate
equality in M.Narciso Dublan-J.Gros Fernandez, Catalonian Team Championship 1995; ½-½ in 61) 13 ...
Nge7 14 Nd4 Bd5 15 0-0 0-0 16 Rfd1 Rfd8 17 h3 h6 18 b3 a6 19 Bf1 Rac8 and Black’s superior piece
placement balanced the bishop pair in V.Ivanchuk-A.Grischuk, Monte Carlo (blindfold rapid) 2006 (½-½
in 59).
b ) 10 a3 is more ambitious, but Black gets good compensation by sacrificing the d4-pawn: 10 ...
Nge7 11 b4 Qb6 12 Bb2 0-0.
 This novelty was suggested by Aagaard and Ntirlis, and played in a game which followed their
analysis for a long time: 13 Bxd4 Nxd4 14 Qxd4 Nc6 15 Qb2 (Aagaard and Ntrilis give 15 Qc5 Qxc5 16
bxc5 Na5 with good compensation for Black) 15 ... a5 16 b5 a4 17 e3 Na5 18 Be2 Nb3 19 Rd1 Qa5+ 20
Kf1 Be6 21 h4 (White is the first to play a move not analysed by Aagaard and Ntirlis) 21 ... Rfd8 22 Nd4
Nxd4 23 exd4 Rac8 and Black had more than enough compensation in F.Urkedal-I.Jelen, Pula 2013.
9 ... Qxc5

10 Rc1
Instead:
a) 10 e3 should be met by 10 ... dxe3 11 Bxe3 Qb4+ 12 Qd2 (or 12 Bd2 Qxb2) 12 ... Qxd2+ with a
balanced endgame in E.Grünfeld-S.Tarrasch, Teplitz-Schönau 1922 (½-½ in 26).
b) 10 b4!? Nxb4 11 Rc1 Qd6 12 e3 was an interesting sacrifice in F.Graf-J.Krassowizkij, Hanover
2012, when 12 ... Ne7 13 Nxd4 0-0 seems to be safest response, with good equalizing chances.
10 ... Qb6
10 ... Qd6 was another successful Tarrasch outing for the German IM in J.Timman-J.Carlstedt,
Helsingor 2012 (½-½ in 25), but he would have been suffering had Timman found 11 Bc3!.
11 e3 Nf6
11 ... dxe3?! is now too compliant and has scored badly in practice; for instance, 12 Bxe3 Qxb2 (not
12 ... Qb4+ 13 Bd2! Qxb2? 14 Bc3 and wins) 13 Bc4 Qb4+ 14 Nd2 with a strong initiative for the pawn
in E.L'Ami-T.Willemze, Haarlem 2012 (1-0 in 24). It tends to be advisable to develop your own pieces
rather than the opponent’s.
 12 Bc4
After 12 Qa4, another recent game followed Aagaard and Ntirlis’ analysis for quite some time: 12 ...
0-0 13 Nxd4 Qxb2 14 Nxc6 bxc6 15 Qc2 Rb8 16 Bc4 Qxc2 17 Rxc2 Rb1+ (Aagaard and Ntirlis
recommend 17 ... Bf5) 18 Rc1 Rxc1+ 19 Bxc1 Be6 20 Bxe6 fxe6 21 f3 Nd5 22 e4 (Aagaard and Ntirlis
had analysed 22 Kd2 Rb8 23 e4 Nb4 24 a3 Na2 25 Kc2 Nxc1 26 Rxc1 Kf7 with a very slight edge to
White in the rook ending) 22 ... Nc3 23 Be3 Rd8 24 0-0 a6 25 h4 Kf7 26 Kh2 Rd3 27 Bf4 Nd1 28 Re1 c5
29 Re2 c4, when White had had enough excitement and decided to force a draw with 30 Rc2 Rc3 31 Rd2
Rd3 32 Rc2 Rc3 33 Rd2 Rd3 in A.Indjic-S,Brenjo, Banja Koviljaca 2013.
12 ... dxe3
Yet another novelty from the Aagaard/Ntirlis laboratory. After 12 ... 0-0 13 0-0 Bg4 14 exd4 Rad8
15 Bc3, Black had problems restoring material parity in C.Sandipan-M.Vucicevic, Lugano 2013 (1-0 in
31).
13 Bxe3 Qb4+

14 Bd2
14 Kf1 Be6 15 Bxe6 fxe6 16 Qe2 Qa5 17 Ng5 Qxa2 18 Qb5?! (18 b4 Qxe2+ 19 Kxe2 e5, intending
20 ... Nd4+, is equal) 18 ... 0-0 19 Qxb7 Rac8 20 Kg1 was M.Krasenkow-J.Koch, French Team
Championship 2012, and now 20 ... Ng4 would have given Black a serious initiative.
14 ... Qe7+ 15 Be2 0-0 16 0-0 Rd8
Departing from his own recommendation of 16 ... Be6. Perhaps he should have stuck to this, since in
the game he gets into trouble.
17 Re1 Ne4 18 Bb5 Bg4 19 Bxc6 bxc6 20 Qe2?
Losing his way. 20 Qc2! would have left Black on the ropes.
20 ... Re8 21 Qe3 Qd7 22 Bc3 Nxc3 ½-½

White no longer has any advantage and a draw is a fair result. For instance, 23 Qxc3 Bxf3 24 Qxf3
Qd2! levels the chances.

Game 21
P.Rewitz-J.Aagaard
Danish Team Championship 2011

1 d4 d5 2 c4 e6 3 Nc3 c5 4 cxd5 exd5 5 Nf3 Nc6

 6 Bg5
This move does have some point – after all, trading dark-squared bishops should be in White’s
favour in most IQP positions. Having lost a bad game in this variation (see the note with 8 ... 0-0 below),
I know how important it is to retain concentration when an ‘innocuous sideline’ is used; Black’s
development is accelerated but he needs to avoid ending up with no counterplay in an inferior structure.
a) 6 Bf4 Nf6 7 e3 is also well met by 7 ... c4!, as played by Tarrasch himself. White doesn’t have
much here; for instance, 8 Be2 Bb4 9 0-0 (the most popular move but Black has no problems after it;
instead, 9 Ne5 Ne4 10 Nxc6 bxc6 11 Qa4 Qb6 12 Rc1 was Vl.Georgiev-T.L.Petrosian, Chicago 2009,
½-½ in 43; while 9 Nd2 Qa5 10 Qc2 0-0 11 0-0 g6 12 Bg5 Bf5 13 Qc1 Be7 gave Black a comfortable
game in P.Harikrishna-I.Sokolov, Sarajevo 2009 ) 9 ... 0-0 10 Qc2 (10 Ne5 Bxc3 11 bxc3 Ne4 was equal
in V.Tkachiev-V.Yemelin, Russian Team Championship 2006 ) 10 ... Bg4 11 Rfc1 Re8 12 h3 Bh5 13 Nb5
Rc8 14 Qa4 a6 15 Na7 Nxa7 16 Qxb4 b5 17 g4 Bg6 18 Ne5 a5 19 Qe1 Nc6 and Black had promising
queenside play in M.Meyer-A.Berelowitsch, German League 2011 (0-1 in 48).
b) 6 e3 gives rise to Symmetrical Tarrasch positions where White has shown his hand by capturing
on d5 early. For instance, after 6 ... Nf6 7 Bb5 Bd6 we have transposed into Game 19 (Carlsen-
Radjabov) in the previous chapter.
6 ... Be7 7 Bxe7 Ngxe7 8 e3
By taking control of the d4-square, White prepares to capture on c5.
The immediate 8 dxc5 is met by 8 ... d4 9 Ne4 (9 Nb1 0-0 10 g3 Nf5 11 Bg2 Qa5+ 12 Qd2 Qxc5 13
0-0 Be6 14 Na3 Rfd8 15 Rfd1 Rac8 was comfortable for Black in S.Cicak-S.Galdunts, German League
2003) 9 ... 0-0 and White needs to be a little careful:
 a) 10 g3 Qd5 11 Ned2 Qxc5 12 Bg2 Be6 13 0-0 (if 13 Nb3 Qb6 14 0-0 Rfd8 15 Ne1 Nb4 16 Nd3,
as in S.Collas-G.M.Todorovic, Skopje 2011, I like 16 ... Rac8 which covers c5 and gives Black good
play) 13 ... Qb5 (on a principled basis I prefer 13 ... Rfd8 followed by ... Rac8 – Black hasn’t completed
development yet) 14 Nb3 Bxb3 (again 14 ... Rfd8 would be my preference) 15 Qxb3 Qxe2 16 Rfe1 Qa6
17 Bf1 Qa5 18 Qxb7 Rab8 19 Qa6 was T.Polak-F.Ljubicic, Split 2008, and now 19 ... Qxa6 20 Bxa6
Rxb2 21 Nxd4 g6 would have led to a draw.
b) 10 Ned2 Re8 11 g3 is asking for trouble.

Exercise: How can Black take advantage of White’s lack of development?

Answer: 11 ... d3! (of course!) 12 a3 (not an attractive move, but White was already in a bad way:
12 e4 Nb4 is much better for Black, while after 12 Rc1 Nd5 White will do well to get out of the opening
in one piece) 12 ... dxe2 13 Bxe2 Bh3 14 Ng5 (14 Nc4 is a better try, but Black’s initiative continues in
the endgame after 14 ... Nf5! 15 Qxd8 Raxd8) 14 ... Bg2 15 Rg1 Nf5 16 Rxg2 Qxg5 (a knight is coming to
d4 and White is losing decisive amounts of material) 17 b4 Ncd4 18 f4 Qf6 19 Nc4 Rad8! (nice attacking
technique, bringing in the final piece) 20 Qb1 Nc2+ (other wins were easier but it doesn’t matter at this
point) 21 Qxc2 Qxa1+ 22 Kf2 Nd4 and White resigned in V.Bernadskiy-K.Malinovsky, Krakow 2013.

Question: How would you respond with Black?

Answer: To a certain extent, what you play here should be informed by your approach after 9 Bg5. If
you delight in the IQP positions after 9 ... cxd4, you can get similar structures here by taking on d4 or
castling. If you go for 9 ... c4, you should push your pawn here too.
8 ... c4
In general, I think this is the best move. The 9 Bg5 c4 lines require White to play very purposefully
in order to break the black d5/c4 pawn chain and, if White doesn’t achieve this goal, he can easily end up
worse. The current position seems to me like a very favourable version of this structure for Black – there
is no annoying bishop on g5 threatening to disrupt the kingside structure by taking on f6, while White’s
light-squared bishop is not actively supporting the e4-advance from g2, but rather will sit passively on e2.
Instead:
a) I played 8 ... cxd4 9 Nxd4 0-0 a couple of years ago.
 The game is a decent example of how Black can gradually overtake the initiative if White is
careless: 10 Be2 Be6 11 0-0 Qb6 12 Na4 Qa5 13 a3 (13 Nxe6 fxe6 14 Rc1 was better, when I prefer
White) 13 ... Nxd4 14 exd4 (14 Qxd4 Nc6 15 Qd1 d4! is fine for Black) 14 ... Rad8 15 b4 (bringing a
rook to an open file seems more purposeful) 15 ... Qc7 16 Qd2 Nf5 17 Nc5 Bc8 18 Rac1 Qe7 19 Bg4
Nd6 20 Rfe1 Qh4 21 Bf3 Nb5 22 Nb3 Nxa3 23 Qd3 Nc4 24 Bxd5 Rxd5 25 Rxc4 Be6

(an unusual structure has been reached; Black certainly doesn’t have any problems here, but nor
should White) 26 Rc5 (26 Rc7 is more active) 26 ... g6 27 Rxd5 Bxd5 (now the black bishop is
dominant) 28 Nc5 Rd8 29 Re5? Kg7? (letting White back in the game; 29 ... Bc6 was stronger, since 30
d5? Qxb4 doesn’t work for White) 30 f3? (in trying to restrict the bishop on d5, White severely weakens
his king’s position and creates some tactical weaknesses; 30 Qe3 would have held the balance) 30 ... b6
31 g3 (the white knight doesn’t have a good retreat, since going to e4 allows the black queen to capture on
b4 after a check on e1) 31 ... Qh6 32 Ne4 (32 Nb3 was a better try, but Black is now well on top after 32
... Bxb3 33 Qxb3 Qc1+ 34 Kf2 Qd2+ 35 Kf1 Rc8!) 32 ... Qc1+ 33 Kg2 (losing immediately, though
alternatives failed to save White either) 33 ... Bc4! – an elegant finish: mate on f1 is looming if the queen
moves, so White resigned in L.Klement-S.Collins, Teplice 2011.
 b) 8 ... 0-0 9 dxc5 Qa5 (not 9 ... Bg4? 10 Be2 Qa5? 11 Qa4 Qxa4 12 Nxa4 and I had no
compensation for the pawn in K.Bischoff-S.Collins, Dun Laoghaire 2010; 1-0 in 29 – this is a classic
example of Black, thinking he needs to “justify” his compromised pawn structure, trying to be too clever)
10 Be2 Qxc5 11 0-0 and Black has a comfortable game.

Here’s a great example of how a Tarrasch expert can play such positions for a win: 11 ... Be6 (11 ...
Rd8 12 Rc1 Bf5 was also reasonable in H.Teske-L.B.Hansen, German League 2000; ½-½ in 25) 12 Nb5
Nf5 13 Nfd4 Nfxd4 14 Nxd4 Rac8 15 Rc1 Qb4 16 Nxe6 (16 Nxc6 Rxc6 17 Rxc6 bxc6 is level as well)
16 ... fxe6 17 Bg4

(White has created some imbalance by capturing on e6, but – as always – this gives Black chances
too, based on his strong centre and potential play on the f-file) 17 ... Rf6 18 Rc2 Rcf8 19 f4?! (White
loses patience and creates permanent weaknesses; simply holding the position was more appropriate) 19
... Qb6 (the immediate 19 ... d4 was fine but, interestingly, Akobian prefers to prepare this move ) 20 Re2
Kh8 21 Rff2 d4 (while this position is level according to the computer, Black’s position is much easier to
play) 22 Qb3? was played in Y.Lapshun-V.Akobian, Philadelphia 2002 (the counter-intuitive 22 exd4
was apparently the best move) and now 22 ... Qxb3 23 axb3 dxe3 24 Rxe3 e5 would have been good for
Black.
9 Be2

9 ... 0-0
9 ... Rb8 10 0-0 b5 has also been tested in a number of games, with good results for Black. My
personal preference is to complete development and offer a trade of b-pawns, as Jacob does in this game.
10 0-0
10 Ne5 b5 11 Nxc6 Nxc6 12 Bf3 Be6 13 a3 a5 14 0-0 (here 14 Nxb5 Rb8 15 a4 Nb4 16 0-0 gives
Black full compensation, or simply a draw after 16 ... Bf5 17 Rc1 Na2 18 Ra1 Nb4) 14 ... Rb8 15 Re1
Qd6 16 Qd2 f5 and Black had more space at little cost in J.Svatos-S.Svoboda, Czech League 2013 (½-½
in 44).

Exercise: How can Black get going on the queenside?

Answer: 10 ... b5!

Black can also prepare this advance with 10 ... Rb8 but, although this has been played by Tarrasch
experts like Akobian, I don’t think it’s necessary, since the trade of b-pawns favours Black provided he
can stop White playing e3-e4.
11 Nxb5
The more sedate 11 b3, as played in M.Bartel-M.Petr, Czech League 2012, can be well met by 11 ...
Qa5 12 Qc1 Rd8. Again Black seems to have a great version of the 9 Bg5 c4 lines.
11 ... Rb8 12 Nc3 Rxb2

Question: Who’s better here?

Answer: I already prefer Black in this position, since he has a protected passed pawn and it’s
entirely unclear what trumps White has. In such a structure there’s no pressure on d5, so the chances for
White to break the d5/c4 pawn chain (his favourite plan in the 9 Bg5 c4 line) are close to zero.
Note the bishops. It’s a trite point, but while Black technically has a “bad bishop” (since his central
pawns are on the same colour) and White has a “good bishop”, just looking at the position reveals that the
e2-bishop is passive and completely restricted by the black pawn chain, while its counterpart has bright
prospects on f5, or even just e6, holding the key d5-pawn. If White were able to simplify and get his
bishop to f3 the assessment might change, but for the foreseeable future the black bishop is going to have
more fun. This concept of a formally “good” bishop actually being no better than its counterpart, since it
is restricted by the opponent’s pawns, is known from several openings, in particular the Stonewall Dutch.
13 Na4
Two alternatives have been tried against the German IM Jonathan Carlstedt, who is a faithful
Tarrasch practitioner and has very good results with the opening. Indeed, one of his biggest upsets (a
miniature win with Black against Milov) occurred in this position:
a) 13 Qc1 Rb4 14 Rd1 Bf5 15 Nh4 Be6 16 g3 Qa5 17 Ng2
 17 ... Rfb8 (since e3-e4 is only White’s idea, it is worth using prophylactic measures against it: 17
... Rd8 18 Nf4 Bf5 keeps White bottled up, and even 17 ... f5!? looks like a promising idea) 18 Nf4 Bf5
19 Rd2 R4b7 20 Bf3 Nb4 21 e4 dxe4 22 Nxe4 Nbd5 23 Nxd5 Nxd5 (through inspired play Milov has
managed to achieve the e3-e4 break, only to make an amazing mistake ... ) 24 Nc5?? Rb1. Whoops. White
resigned in V.Milov-J.Carlstedt, Odense 2011.
b) 13 Ne5 Qa5 (13 ... Nxe5 14 dxe5 Bf5 is good too) 14 Nxc6 Nxc6 15 Qc1 Rb6 16 Rd1 Bf5 17
Bf3 Rd8 18 Qd2 Bd3 19 h3 f5! and Black was better against a strong GM in R.Buhmann-J.Carlstedt,
Deizisau 2010 (½-½ in 25).
13 ... Rb4 14 Nc5 h6
14 ... Qd6 is also strong, since 15 e4 can be met by 15 ... Qg6 with some advantage for Black.
15 h3
White would like to play 15 e4, but he is insufficiently prepared for the opening of the game after 15
... dxe4 (or 15 ... Bg4) 16 Nxe4 Bg4.
15 ... Bf5 16 Qc1 Qa5
 Question: How would you assess this position?

Answer: White is busted. He has one good piece (the knight on c5) which is not co-ordinating with
any of his other pieces. In the meantime, Black is preparing to penetrate down the b-file. White decides on
a rather desperate kingside break, though it is already hard to offer him good advice.
17 g4 Bh7 18 g5 hxg5 19 Nxg5 Bf5 20 Bg4 g6
20 ... Qd8 was more accurate, taking aim at the knight on g5.
21 Qd1 Qd8 22 Qf3 Qd6 23 h4 Kg7 24 Kg2
Jacob can now improve the position as he chooses but goes instead for a trap.
24 ... c3!?
24 ... Rh8 is good, just attacking on the kingside.
25 Rac1?
Falling for the trap, but White’s position was already terrible.

Exercise: Black to play and win.

Answer: 25 ... Bxg4 26 Qxg4 Qxc5!

Oh dear. I tend to resign when I’m a piece down for nothing against a GM but I suppose not
everyone feels the same way. In fairness Jacob soon returns the piece (probably informed by his own
sense of humour rather than the objective merits of the move) and takes all White’s pawns instead.
27 a3 Rc4 28 Qf4 Nxd4 29 exd4 Qxd4 30 Qxd4+ Rxd4 31 Rxc3 Rxh4 32 f4 Nf5 33 Rc7 Ne3+
34 Kg3 Rg4+ 35 Kf3 Rxg5 36 Kxe3 Rg3+ 37 Kd4 Rxa3 38 f5 Ra4+ 39 Kxd5 Ra5+ 40 Ke4 Re8+ 41
Kd4 Rd8+ 42 Ke4 Re8+ 43 Kd4 Rxf5 44 Rxf5 gxf5 45 Rxa7 Kg6 46 Kd3 f4 47 Ra4 Kg5 48 Ra7 f6 0-
1

Game 22
L.Schandorff-J.Aagaard
Danish Championship, Odense 2011
1 d4 d5 2 c4 e6 3 Nf3 c5 4 g3 Nf6 5 Bg2 Nc6 6 0-0 Be7 7 dxc5 Bxc5 8 a3
This is another approach (which, again, can arise from multiple move orders). White aims for b2-b4
without taking on d5, thus keeping the bishop on c8 bottled up, but this quiet approach allows Black to
complete development comfortably. If White takes on d5, Black will have the option of favourable
Tarrasch positions where White has weakened key squares on the c-file. If White does not take on d5,
practice has shown that Black has sufficient defensive resources, as illustrated by the present game.

8 ... 0-0 9 b4 Bb6

It is more common to retreat the bishop to e7, but 9 ... Bb6 is more active and the move I prefer.
10 Bb2 Qe7 11 Nbd2 Rd8 12 cxd5
This is White’s last chance to create an IQP. Instead, 12 Qc2 allows Black to seize the centre with
12 ... e5.

White is not worse here but his play is a little too subtle for me.
a) 13 cxd5 Nxd5 and now:
a1) 14 Nc4 e4 15 Nxb6 Nxb6 16 Nd2 was R.Buhmann-R.Strohhäker, Böblingen 2006, when 16 ... f5
would have been quite playable for Black.
 a2) 14 e3 a6 15 Ne4 Bf5 16 Nh4 Be6 17 Rac1 Rac8 was very comfortable for Black in E.Bricard-
S.Galdunts, French Team Championship 2006 (0-1 in 52).
a3) 14 b5 Nd4 15 Nxd4 exd4 16 Nb3 Bg4 17 Nxd4 Ne3! 18 fxe3 Qxe3+ 19 Kh1 Bxd4 20 Bxd4
Rxd4 21 Bxb7 Re8 with at least equal chances for Black in J.Magem Badals-M.Illescas Cordoba, Leon
1990 (½-½ in 29).
a4) 14 Rfd1 f6 (the immediate 14 ... Be6 is also possible, since 15 Bxe5?? Nxe5 16 Nxe5 Bd4 wins,
and 15 Nxe5? Nd4 16 Qd3 Nf6 is good for Black: ... Nb3 is a threat) 15 Nc4 Be6 16 e4 Nc7 17 Nxb6
axb6 18 Bf1 was O.Romanishin-S.Silva, Figueira da Foz 2008, and now 18 ... Ne8! was good, aiming to
bring the knight to d6 and (after ... b6-b5) perhaps to c4.
a5) 14 Rad1 f6 15 Nc4 Be6 16 e4 Nc7 17 Nxb6 axb6 18 Qe2 b5 19 Rxd8+ Rxd8 20 Rd1 Bc4 was
level in O.Romanishin-M.Meinhardt, Hockenheim 2006 (½-½ in 41).
b ) 13 b5 Nd4 14 Nxd4 exd4 15 c5 Bxc5 16 Nb3 Bb6 17 Bxd4 Bg4 18 Rfe1 Rac8 19 Qb2 was
A.Khalifman-A.Grischuk, German League 2003. It is remarkable to see how Grischuk, from an equal
position against a world-class player and former FIDE world champion, manages to whip up a dangerous
attack and win the game in 10 moves.

Exercise: How should Black start to build an initiative?

Answer: 19 ... h5! – always a standard move against the kingside fianchetto. Grischuk wants to put
his pawn on h4, from where it can either weaken the white structure ( ... h4xg3) or threaten the white
king’s breathing space ( ... h4-h3). Here Ribli recommends 20 h3 as equal, but Khalifman makes a couple
of careless moves: 20 a4 h4 21 a5 Bxd4 22 Nxd4 Qe5 23 Rab1?! (23 Rac1 maintains the balance) 23 ...
Ne4 24 Nf3 Qxb2! 25 Rxb2 h3! 26 Bf1 d4 27 Rd1?? (an outright blunder; 27 Ne5 Be6 28 Nd3 is still
only slightly better for Black).
 Exercise: Black to play and win.

Answer: 27 ... d3! (a standard Tarrasch breakthrough: White sheds material) 28 Rb4 (or 28 Rxd3
Rxd3 29 exd3 Bxf3 30 dxe4 Rc1, followed by ... Bg2) 28 ... Bxf3 29 exf3 Nc3! and White resigned, in
view of 30 Rxd3 Ne2+ 31 Bxe2 Rc1+ 32 Bf1 Rxd3 (Ribli).
12 ... exd5

Black has no problems. His position is active, all the pieces are still on the board, he has several
targets (the e2-pawn, the c4-outpost, and possibilities of undermining the white queenside with ... a7-a5),
and there is zero pressure on the d5-pawn.
13 Nb3 Ne4 14 b5 Na5 15 Nxa5 Bxa5 16 Qd4 f6 17 Qa4 Bb6 18 Bd4 Bd7 19 Rfc1 Qe8 20 Qb4
Rac8
20 ... Bxb5 21 Bxb6 axb6 22 Nd4 regains the pawn, though the position remains equal.
21 a4 Rxc1+ 22 Rxc1 Rc8 23 Qa3 Kf7 24 h3 ½-½
 The players agreed a draw in this balanced position. I really like Jacob’s play in this game – simple
development, putting his pieces on good squares. People often get too clever in the Tarrasch and end up
putting their pieces on the wrong squares.

Key Notes
1. Against all Tarrasch sidelines, take your time and handle the first few moves carefully. Some of
these variations have dangerous ideas and Black needs to respond accurately.
2. 6 dxc5 is probably the most dangerous sideline, going straight for positions with an open centre
where White relies on his bishop pair. Nevertheless, Black has excellent development and can expect to
equalize with accurate play.
3. 6 Bg5 Be7 7 Bxe7 Ngxe7 is a quiet sideline which shouldn’t trouble Black. After 6 e3, 6 ... c4!
gives Black a great version of the 9 Bg5 c4 variations.
4. Delaying the capture on d5, as in Schandorff-Aagaard (Game 22), would inconvenience Black if
he had to take on c4 or fianchetto his queen’s bishop, but he can solve his problems in a more direct way
by simple development. If White insists on not taking on d5, Black can seize central space with ... e6-e5;
then if White takes on d5, the resulting IQP position is fine for Black.
 Chapter Eight
Réti Set-ups
1 Nf3 d5 2 c4 e6 3 g3 c5

One of the attractions of the Tarrasch is that it can be played against everything except 1 e4. Come to
think of it, Tarrasch’s “refutation” (3 ... c5) of his own variation (1 e4 e6 2 d4 d5 3 Nd2) does bear a
strong stylistic resemblance to some lines of our opening, but I’ll leave coverage of this line to more
experienced French players.
White can play a broad range of flank openings and we can’t possibly cover them all here.
Fortunately we don’t need to, since many of these openings have a remarkable tendency to reach the same
position:
4 Bg2 Nf6 5 0-0 Nc6 6 b3 Be7 7 e3 0-0 8 Bb2 b6 9 Nc3 Bb7
This position is covered in Reinderman-Irwanto (Game 23) and knowledge of this game should be
sufficient against most English/Réti set-ups. If White goes more offbeat, your first few moves involved
gaining central control and developing pieces, so I suggest you keep doing that.
Game 23
D.Reinderman-S.Irwanto
Jakarta 2011

1 Nf3
The games in this section began through various move orders. For example, Kramnik-Jones, covered
in the notes to Black’s 10th, in fact began 1 Nf3 c5 2 b3 d5 3 e3 Nf6 4 Bb2 e6 5 g3 Nc6 6 Bg2 Be7 7 0-0
0-0 8 c4 b6 9 Nc3 Bb7. Gawain was aware that this was a known position (albeit one not in his
repertoire, which normally features a kingside fianchetto for Black, such as in the Sicilian Dragon and
King’s Indian defences in which he is a leading expert) and was spending a lot of time in the opening
trying to find a way to deviate.
The fact that he didn’t find such a way, to me, indicates a couple of points about this variation. The
first is that Black has developed his pieces in a natural way, to active squares, and has seized space in the
centre. This fully conforms to opening principles and such a logical approach simply cannot be bad. The
second is that Black’s set-up is universal. Unlike fianchetto set-ups (where the 1 Nf3 c5 2 b3 move order
is actually a bit of a pain), to say nothing of lines like the Benko which require White to play a very
specific series of moves in order for Black to reach his desired formation, the Tarrasch can be played
against basically everything except 1 e4. This is of great benefit since it means we have fewer move order
issues after 1 c4 and 1 Nf3, and also means that we can gain more experience in the Tarrasch structures
since they will arise in more of our games.
1 ... Nf6

Question: This move certainly isn’t bad,

but is it the best choice for a Tarrasch player?

Answer: I don’t think so. If Black is comfortable with a Tarrasch set-up ( ... d7-d5, ... c7-c5, ... e7-
e6), he is best to get these moves on the board as soon as possible. Since both 1 ... c5 and 1 ... e6 allow
transpositions to major king’s pawn openings after 2 e4, this suggests that 1 ... d5 is the best option.
 Question: But what is the problem with 1 ... Nf6?

Answer: Let’s see a possible continuation: 2 d4 d5 3 c4 e6 4 Nc3, and now if Black wants to get
Tarrasch-type positions he needs to play 4 ... c5 here, but this is a Semi-Tarrasch.

Question: Can we get back to the positions we’ve studied?

Answer: Well, yes if White plays 5 e3 (which is a Symmetrical Tarrasch), but White’s main option
5 cxd5 creates more problems. The normal Semi-Tarrasch move is to recapture with the knight, which
leads to entirely different positions from those we’ve looked at after both of White’s main moves (namely
6 e4 and 6 e3). Thus Black, if he wants to play a regular Tarrasch, needs to meet 5 cxd5 with 5 ... exd5.

But now we’ve been tricked into a position which could arise after 1 d4 d5 2 c4 e6 3 Nc3 c5 4 cxd5
exd5 5 Nf3 where, instead of playing the recommended 5 ... Nc6 (when 6 Bg5 can be met by 6 ... Be7),
we have developed our king’s knight first. Black is not lost here, or even much worse, but White makes a
nice score with the pin 6 Bg5 and Black needs to learn a different defensive set-up.
One of the main points of 1 Nf3 is to trick the opponent into positions outside his repertoire. It is
always worth spending a few minutes on the initial moves in order to make sure we have the
transpositions covered.
My game with Socko from Riga 2013 (covered in the notes to Game 12) followed a more usual
move order: 1 ... d5 2 g3 c5 (since there is no pressure on the centre, there is no reason for Black to
commit to ... e7-e6 yet) 3 Bg2 Nc6. Now Black is basically threatening 4 ... e5, so 4 d4 is logical, and
then 4 ... e6 5 0-0 Nf6 6 c4. Here Black can transpose into Catalan lines with 6 ... dxc4, but the most
logical choice for a Tarrasch player is the simple 6 ... Be7. After a few minutes’ thought, my opponent
found nothing better than 7 dxc5 Bxc5 8 cxd5 exd5 9 Bg5 0-0 10 Nc3 d4 with a transposition into Chapter
Four.
2 c4 e6 3 g3 d5 4 Bg2 Be7 5 0-0 0-0 6 b3 b6
As noted above, a true Tarrasch devotee will set up the c5/d5/e6 formation earlier to avoid getting
tricked into positions from other openings.
7 Bb2 Bb7 8 e3 c5 9 Nc3 Nc6
 10 cxd5
The only serious move.
a) 10 d3 is sometimes played just to get a position; for instance, 10 ... Qd7 11 Qe2 Rad8 12 Rfd1 d4
13 exd4 Nxd4 14 Nxd4 cxd4 15 Nb5 Bxg2 16 Kxg2 Bc5 17 b4 Bxb4 18 Nxd4 Be7 with equality in
V.Gashimov-Wang Yue, FIDE Grand Prix, Astrakhan 2010 (½-½ in 62).
b) 10 d4 dxc4 equalizes on the spot, since Black is even better after 11 bxc4?! Na5! with a rapid
attack on c4.
10 ... Nxd5
This is the main line. The move I wanted to make work was 10 ... exd5, which is the most consistent
Tarrasch response, keeping pieces on the board and not being afraid of having a potential weakness on
d5. Unfortunately, after 11 d4 White seems to have the better chances:

a) 11 ... Re8 12 Rc1 Bf8 (12 ... Ne4 13 dxc5 Nxc3 14 Bxc3 bxc5 is similar to 11 ... Ne4 below and
looks a little better for White) and now:
 a1) 13 Ne2 Ne4 14 Nf4 Ba6 15 Re1 c4 16 bxc4 Bxc4 17 a3 Rc8 was balanced in N.Vitiugov-
D.Frolyanov, Russian Rapid Championship, Olginka 2011.
a2) 13 dxc5 bxc5 14 Na4 Ne4 15 Nd2 Ba6 16 Nxe4 dxe4 17 Qxd8 Raxd8 18 Rfd1 f5 and Black had
sufficient activity to compensate for his queenside weaknesses in G.Welling-C.Sandipan, Gibraltar 2008
(½-½ in 49).
a3) 13 Re1 Rc8 (13 ... Qd7 14 dxc5 Bxc5 15 Na4 Bb4 16 Re2 Ne4 17 Rec2 Rac8 18 Bf1 Red8 19
Bb5 is already good for White, and after 19 ... Qd6?? 20 a3 Black resigned immediately in V.Papin-
A.Rychagov, Voronezh 2012) 14 Bh3 Rb8 15 Ne2 Ne4 (15 ... c4 looks okay for Black) 16 Nf4 cxd4 17
Nxd4 Nxd4 18 Qxd4 was better for White in V.Ivanchuk-L.Ljubojevic, Monte Carlo (rapid) 2002 (1-0 in
54).
b) 11 ... Rc8 12 Rc1 (after 12 Ne2 Ba6 13 Re1 c4 14 bxc4 Bxc4 15 Nc3 b5 16 Nd2 b4 17 Na4 Na5,
Black had active queenside play in J.Cuenca Jimenez-R.Ponomariov, Spanish Team Championship 2009;
0-1 in 54) 12 ... Ba6 13 Re1 cxd4 14 Nxd4 was slightly better for White in V.Akopian-V.Iakemov,
European Championship, Rijeka 2010 (1-0 in 34).
c) 11 ... Ne4 12 dxc5 Nxc3 13 Bxc3 bxc5 14 Qe2 was V.Kramnik-G.Jones, London 2012.
 As Kramnik explained in the press conference after the game, the knight belongs on d7 in such
positions, from where it can come to f6 and keep everything defended. Regarding White’s ideas, Kramnik
mentioned three: Qb2, Qb5, and Rd1.
The game continued 14 ... Re8 (Gawain had overlooked that after 14 ... Qb6 15 Qb2! he is forced to
play the weakening ... f7-f6, and 14 ... Qd7 15 Bb2! is similar; while after the logical 14 ... Qd6 15 Rfd1
Rfd8, Kramnik thought that the excellent 16 Ne1! would put the d5-pawn under intolerable pressure, e.g.
16 ... Qe6 17 Nd3! and 18 Nf4 is threatened) 15 Rfd1 Bf8 16 Qb5 (or 16 Ne1 Ne7 17 Nd3 Qb6 18 Nf4
Rad8, when Kramnik noted that Qg4, Qh5 and Qe1, threatening Ba5, were all promising; or if 17 ... Ng6
then 18 h4! and h4-h5 is coming) 16 ... Qb6 (16 ... Rb8 17 Ne1 Ba8 18 Qa4 is clearly better for White) 17
Qxb6 axb6 18 Rxd5 (going for a forcing line; 18 Ne1 Ne7 19 Nd3 Red8 20 a4! is also quite pleasant for
White according to Kramnik) 18 ... Nd4 19 Nxd4 Bxd5 20 Bxd5 cxd4 21 Bxd4 (as Kramnik noted after
the game, if Black could trade a pair of rooks he would get good drawing chances; but with rooks on the
board, White has excellent winning prospects) 21 ... Ra5! (a good move which sets a trap; 21 ... Rab8 is
met by 22 a4!) 22 e4 (the tempting 22 Bc4 runs into 22 ... b5! 23 Bb6 Ra3, when 24 Bxb5?? loses to 24 ...
Rb8) 22 ... Bc5 (Kramnik analysed 22 ... b5 23 Rc1 Rxa2 24 Rc7 Re7 25 Rc8, intending Bc5, as winning
for White) 23 Bc3 Ra3! (23 ... Ra7 24 a4 Rc7 25 Bd2 is easier for White) 24 Bb2 Ra7 25 a4 Rc7 26
Ra2!

(an excellent move overlooked by Gawain, defending against ... Bxf2+ and keeping rooks on; and
stronger than 26 Rc1 Ree7!?, trying to take on c7 with the rook after 27 Kg2 Bb4, when the black bishop
is more active) 26 ... h6 (or 26 ... Bb4 27 Bd4 Rc1+ 28 Kg2 Bc5 29 Bxc5 Rxc5 30 b4 Rc7 31 a5 bxa5 32
bxa5 Ra7 36 a6 and White wins by slowly bringing his king to the queenside) 27 Kg2 Kh7 28 f4
(Kramnik later suggested the “pensioner style” 28 g4, killing all counterplay) 28 ... f6 29 Kf3 Rd7 30 a5
bxa5 31 Rxa5 Rc8 32 Rb5 Rd6 (Kramnik gave 32 ... Ra7 as a better drawing chance) 33 Rb7 Rb6 34 Rf7
Rf8 35 Rc7 Bd6 36 Rd7 Ra6 37 Bd4 Bb8 38 Bc5 Re8 39 Kg4 Ra2 40 h4 Rc8 41 b4 Ra3 42 h5 Rc3 43
Bd4 R3c7 44 Rxc7 Bxc7 45 Kf5 Bd6 46 b5 Rc1 47 b6 Bb8 48 Ke6 Rh1 49 Bc5 and Black resigned.
 11 Nxd5 exd5
The main line is 11 ... Qxd5, after which 12 d4 leaves Black with a number of continuations but no
route to easy equality. In fact I had this position myself with Black against an English Opening specialist:
12 ... Na5 13 Ba3 Rfe8 14 dxc5 Bxc5 15 Qxd5 Bxd5 16 Bxc5 bxc5 17 Rac1 c4 18 bxc4 Bxc4 19 Ne5
Bxf1 20 Bxa8 Ba6 21 Be4 Rc8 22 Rxc8+ Bxc8 23 f4 and a draw was agreed in A.Kosten-S.Collins,
Kilkenny 2009. However, there are of course improvements and, in general, the positions after 12 d4
strike me as slightly inferior and passive for Black, which is not what Tarrasch players normally want.
12 d4

12 ... a5!?
This move has not been played very often but has been tried by a number of players for whom I have
a lot of respect – in particular, Andrei Sokolov, who would simply not play an illogical variation with
either colour. Black takes advantage of White’s slightly passive approach (after all, Black is not behind in
development and his central pawns are not under any immediate pressure) to generate counterplay on the
a-file. Having examined all the games played here my conclusion is that the line is quite playable, and
White can easily get into trouble if he doesn’t know what he’s doing.
 Another useful and little-played idea is 12 ... Ba6 13 Re1 c4!?.

Black aims for an unusual structure where he has good chances of developing active queenside
counterplay by pushing his pawns and bringing a knight to c4. 14 bxc4 (after 14 Bc3!? Rc8 15 Qd2 Qd6
16 Ne5 rather than the 16 ... cxb3 of I.Csom-L.Jakobetz, Hungarian Team Championship 1993, Marin
proposes the novelty 16 ... b5 17 a3 cxb3 18 Reb1 with the assessment: “White has the more compact
structure and an active position. It should be noted that the c3-bishop prevents the strong manoeuvre ...
Na5-c4 which, if it were possible, would immediately turn the tables.”) 14 ... Bxc4 15 Nd2 (15 a3 Rc8
16 Nd2 Na5 17 Bc3 Be2 18 Qxe2 Rxc3 was J.Sadorra-M.Kravtsiv, Beijing blitz 2008, and now 19 Qb5
would have been a little better for White so perhaps Black should try 17 ... Bb5) 15 ... Na5 16 Bc3 b5 17
Nxc4 Nxc4 18 Rb1 Rb8 19 Qh5 (Kosten suggests 19 e4, opening the position for the bishops) 19 ... b4 20
Ba1 Rb5 21 e4 Na3 22 Qe2 was A.Kosten-K.Landa, French Team Championship 2008, and now Kosten
suggests 22 ... Nxb1 23 Qxb5 Nc3 24 Bxc3 bxc3, when the opposite-coloured bishops make the position
quite drawish.

13 dxc5
White has tried a bunch of moves here without demonstrating a clear route to an edge. In general, it
seems logical to respond to a move like ... a7-a5 by opening the centre, since Black’s last move did
nothing to protect his d5-pawn. However, it turns out that Black has sufficient dynamic resources to meet
this direct approach.
a) 13 Qd2 a4 14 Rfd1 axb3 15 axb3 Rxa1 16 Bxa1 Qa8 17 Bc3 Rd8 18 Qb2 Bf8 19 Ne1 cxd4 20
Bxd4 Nxd4 21 Qxd4 Qa5 22 Nd3 Qb5 23 Qb2 Bd6 24 Nf4 Bxf4 25 gxf4 h6 was level in J.Hickl-
A.Sokolov, Swiss Team Championship 2007 (½-½ in 48). The pawn on d5 is a target but it will be
difficult for White to go after it without exposing his king, especially now that he no longer has a g-pawn.
b) 13 Re1 a4 14 bxa4 Na5 15 Rb1 Nc4 16 dxc5 Nxb2 17 Rxb2 Bxc5 was slightly better for White in
M.Sorokin-E.Romanov, Russian Team Championship 2006, but Black could have kept the balance by
simply developing with 13 ... Qd7.
c) 13 Ne5 Nxe5 14 dxe5 Qd7 15 Qh5 a4 16 Rad1 (16 Rfd1 Qb5! is also preferable for Black) 16 ...
axb3 17 axb3 Ra2 18 Bc3 Qe6 gave Black the advantage in A.Srebrnic-I.Bulmaga, European Team
Championship, Porto Carras 2011 (0-1 in 59).
d) 13 Qb1 a4 14 Rd1 axb3 15 dxc5? (15 axb3 is equal) 15 ... bxc5 (why not take the pawn with 15
... bxa2?) 16 axb3 Nb4 17 Rxa8 Qxa8 and Black had an active position with strong hanging pawns in
P.Blatny-J.Pelikian, Groningen 1997 (½-½ in 42).
e) 13 a4, as in C.Mena Crespo-Z.Corrales Jimenez, Havana 2009, concedes the b4-square without a
fight; i.e. 13 ... Nb4 with a wonderful position for Black.
f) 13 Rc1 Ba6 14 Re1 Nb4 15 Ne5 f6 16 a3 fxe5 17 axb4 cxd4 18 exd4 e4 19 bxa5 bxa5 20 f3 Bb4
21 Bc3 Bd3 was agreed drawn in J.Forgacs-L.Sapi, Hungarian Team Championship 1988, but Black is
slightly better in the final position and should have played on.
13 ... bxc5

14 Nh4!?
White goes for a forcing approach, radically altering the character of the position. Black can’t allow
the knight to come to f5 so the following sequence is mandatory.
Alternatively, White can try a quieter build-up with 14 Ne1 Ba6 15 Nd3 Nb4 (Delchev analyses 15
... Rc8 16 Re1 c4 17 Nf4 Bb4 18 Nxd5 Bxe1 19 Qxe1 cxb3 20 axb3 Re8 21 Qd1 with great compensation
for the exchange, and in fact in this line 19 Nf6+! wins on the spot; instead, I like the thematic 15 ... a4!,
creating queenside counterplay – depending on White’s response, the pawn can capture on b3 or advance
to a3) 16 Nxb4 axb4 was E.Pigusov-I.Lysyj, Russian Team Championship, and now White should have
tried 17 Qxd5 Qxd5 18 Bxd5 Rad8 19 Rfd1 Be2 20 Re1 Rxd5 21 Rxe2 Ra8 with some winning chances
in view of his extra pawn.
14 ... Bxh4 15 gxh4 Qxh4 16 Qxd5 Nd8!
16 ... Rad8 was tried in S.Makarichev-A.Panchenko, Moscow 1981, but here simply 17 Qxc5 would
have left Black with insufficient compensation for the pawn.

Exercise: Does Black have more than a draw after 17 Qxc5?

Answer: He certainly does: 17 Qxc5?? Bxg2 18 Kxg2 Qg4+ 19 Kh1 Qf3+ 20 Kg1 Ra6! 21 Be5
Rg6+ 22 Bg3 Ne6 with a decisive attack.
17 Qe5
Wisely trading the queens, though White has nothing in the endgame.
17 ... f6 18 Qg3
Keeping the queens on (for instance, with 18 Qf5) seems risky to me. Queen and knight often co-
operate better than queen and bishop and, while the knight on d8 doesn’t look threatening at the moment,
when it comes into play White will regret having traded off his g-pawn and light-squared bishop.
 18 ... Qxg3
While I extolled the dangers of the middlegame for White in the last note, trading queens is Black’s
best option here since the queen on g3 is very active and covers the king very well.
18 ... Qh5 19 Rad1 Bxg2 20 Qxg2 seems like an edge for White.
19 hxg3 Bxg2 20 Kxg2

20 ... Ne6
20 ... a4, as in L.Timurova-S.Bezgodova, Russian Women’s Championship, Elista 2000, is slightly
less accurate since White can open lines for the b2-bishop with 21 Ba3 Ne6 22 b4.
21 Ba3 Rfb8
 Black has no problems in this endgame and, in fact, I slightly prefer his position since he has the
pawn break ... a5-a4, whereas White has no active play and struggles to make use of his bishop on a3.
22 Rab1 Rb5 23 Rfc1 a4 24 b4 Rab8 25 Rc4 f5 26 Rb2 Kf7 27 g4 fxg4 28 Rxg4 R8b7 29 Rc4
Rb8 30 f4 g6 31 Kf3 h5 32 Re4 R8b7 33 Kg3 Ng7 34 Kf3 Nf5 35 Rc4 Re7 36 Re2 cxb4 37 Bxb4 Rd7
38 a3 Rbb7 39 Rc3 Rd1 40 Rec2 Rbd7 41 Rc1 Rxc1 42 Rxc1 Rd3 43 Rc3 Rxc3 44 Bxc3 Ke6 45 e4
Nd6 46 Bb4 Ne8 47 Bc3 Nd6 48 Bb4 Ne8 49 Bc3 ½-½

Key Notes
1. One of the great advantages of the Tarrasch is that it can be played against everything except 1 e4.
I suggest setting up the d5/c5/e6 formation as soon as possible, when the best White has is a transposition
into regular Tarrasch lines.
2. In the main Réti line (as in Reinderman-Irwanto), Black has a number of options which might appeal to
Tarrasch players. I have recommended two lines: 10 cxd5 Nxd5 11 Nxd5 exd5 12 d4 a5!? and 12 ... Ba6
13 Re1 c4!?. In both, I rather like Black’s active and purposeful queenside play.
Chapter Nine
In Place of an Epilogue

1 d4 d5 2 c4 e6 3 Nc3 c5

As a bonus for getting this far, I’ve decided to offer coverage of two lines which, while arising from
a Tarrasch move order, are actually different openings. Both give rise to dynamic play which I think
would suit Tarrasch players well and, while I’ve only tried one of these lines on one occasion, it resulted
in a nice win. So here we are:
4 cxd5 exd5
4 ... cxd4 is an aggressive gambit continuation. Enough for the pawn? Maybe. See Wang Yue-Wang
Chen (Game 24).
5 Nf3 Nc6 6 g3
Of course we’re familiar with this position. The new concept (actually an old and forgotten concept)
I’d like to cover is the Swedish Variation ...
6 ... c4
 ... followed by a quick ... Bb4 and ... Nge7. See Parligras-Miron (Game 25).
There are a lot of lines in the closed openings which resemble the Tarrasch, and I encourage you to
look them out and test them. A good understanding of Tarrasch structures just makes someone a better and
more dangerous player. For example, Ivan Sokolov (whose model games with Black and White have
featured in this book) has given 1 d4 d5 2 c4 e6 3 Nf3 Nf6 4 Nc3 c5 (the Semi-Tarrasch, which can lead
to quite different positions to the regular Tarrasch) 5 cxd5 cxd4!? a whirl twice (so far) in 2013. Clearly,
in order to play this, one needs to feel comfortable with an IQP. As it turns out, neither of his games
resulted in an IQP for Black, either against L’Ami at Wijk aan Zee in January (½-½ in 84) or against
Jakovenko in the French Team Championship in June (0-1 in 34). But for any 1 d4 d5 players, Sokolov’s
games are compulsory viewing, so keep an eye on him.

Game 24
Wang Yue-Wang Chen
Chinese Championship, Xinghua 2012

1 d4 d5 2 c4 e6 3 Nc3 c5 4 cxd5 cxd4

This dynamic gambit is a speciality of the very strong Chinese IM Wang Chen, who has used it to
good effect against a number of Chinese Super-GMs. It has been used several times by the very strong
Romanian IM Miron as well. I expect that at least one of these players will be a GM by the time this book
goes to print, so you could do worse than follow their games!
 5 Qa4+!
The most precise move order. 5 Qxd4 Nc6 gives Black an extra option (see note ‘b1’):
a) 6 Qa4 exd5 7 Nf3 d4 8 Ne4 (or 8 Nb5 Bb4+ 9 Bd2 Bxd2+ 10 Nxd2 Nge7 11 g3 0-0 with decent
play for Black in D.Harika-P.Cramling, Beijing blitz 2012) 8 ... Bf5 9 Ng3 Be6 10 e4 a6 11 Bd3 Bb4+ 12
Bd2 Bxd2+ 13 Nxd2 Nf6 14 0-0 0-0 with good play for Black in V.Malakhov-Wang Chen, Chinese Team
Championship 2008 (½-½ in 41).
b) 6 Qd1 exd5 and now:

b1) 7 Qxd5 allows Black to transpose into the main game with 7 ... Bd7, but it is also interesting for
him to try the endgame after 7 ... Be6!? 8 Qxd8+ Rxd8. Black has a number of threats including ... Nb4
and White has to be careful. This position has little theoretical significance since Black can elect to play
the main line after 7 ... Bd7 in any event, so 5 Qxd4 (unless followed by the relatively harmless 6 Qa4)
just seems to be an inaccurate move order.
b2) 7 e3 was played against me once: 7 ... Nf6 8 Nf3 Bc5 (if Black’s bishop is aiming for c7 then 8
... Bd6 seems to be a quicker route: 9 Be2 0-0 10 0-0 Bg4 11 Nb5 Be7 12 b3 Qd7 13 Bb2 and a draw
was agreed in a complex position in B.Posedaru-G.M.Todorovic, Belgrade 2012) 9 Be2 0-0 10 0-0 a6
11 a3 Bb6 12 b4 Bc7 13 Bb2 Qd6 14 Rc1 Re8 15 g3 Bh3 16 Re1 Rad8 (this looks like an ideal IQP
formation and White had to try to disrupt Black’s game with 17 b5) 17 Nd4?! Nxd4 18 Qxd4 Bb6 19 Qh4
Bd7 20 Bf3 Bc6 21 Red1 Qe6 (already 21 ... d4 works well) 22 Ne2.

Exercise: Find a thematic breakthrough for Black.

Answer: 22 ... d4! 23 Bxc6? (I was expecting 23 Rxc6 bxc6 24 Nxd4 Qa2 with only a slight
advantage to Black; I had seen the variation 23 ... dxe3 24 Rxd8 exf2+ 25 Kf1 Rxd8 26 Rxb6! Qxb6 27
Bd4 and White is better) 23 ... dxe3, and now the key point is that after 24 Bxe8 (24 f3 Rxd1+ 25 Rxd1
Qxc6 is good for Black) 24 ... exf2+ 25 Kf1 (or 25 Kg2 Qxe2) 25 ... Rxe8! White has no defence; e.g. 26
Rd2 Ng4 27 Rc5 Qe4 and wins. Instead, White opted for 24 Nd4, but after 24 ... exf2+ 25 Kg2 Bxd4 26
Bxd4 bxc6 27 Bxf6 Rxd1 28 Rxd1 Qe1, my French IM opponent resigned in Nguyen Chi-Minh-S.Collins
Vandoeuvres 2013.
5 ... Bd7 6 Qxd4 exd5 7 Qxd5 Nc6

8 Nf3
 White can also delay the development of this knight:
a) 8 e3 Nf6 9 Qd1 Bc5 10 a3!? (a similar idea to the main game) 10 ... Qe7 11 Nf3 and now Black
needs to decide where to place his king: 11 ... 0-0-0 (11 ... 0-0 was the choice of the talented American
GM Ray Robson – after 12 Be2 Rfd8 13 0-0 Rac8 14 b4 Bd6 15 Nb5 Bb8 16 Bb2 Ne4 17 Nbd4 Bg4 18
Rc1 h5 19 g3 a6 20 Qe1 Nxd4 21 Nxd4 Ng5 22 Rxc8 Bxc8 23 h4 Ne4 24 Kg2 g6 25 Bf3 and a draw was
agreed in A.Sharma-R.Robson, Las Vegas 2012, but White is just a pawn up for nothing and the result
reflects the rating difference and not the position) 12 Qc2 Kb8 can lead to aggressive play with opposite
side castling. A recent upset occurred in the following game: 13 Be2 g5 14 b4 Bb6 15 Bb2 Rc8 16 0-0 g4
17 Nh4 Rhg8 18 Rad1 Bc7 19 Nb5 Be5 20 Qb1 Bxb2 21 Qxb2 Ne4 22 Bd3 Ne5 23 Bxe4? (23 Nxa7! is
very good for White) 23 ... Bxb5 24 Nf5 Qf6 25 Nd4 Bxf1 26 Rxf1 Nc4 and Black went on to win in
M.Turov-I.Karimov, Tashkent 2013 (0-1 in 34).
b) 8 Bg5 Nf6 9 Qd2 h6 10 Bh4 g5 11 Bg3, as in I.Ivanisevic-M.M.Ivanov, Cannes 2006, can be met
in a range of ways. Personally I like the look of the endgame after 11 ... Bf5!? 12 Qxd8+ Rxd8 with some
initiative for Black.
8 ... Nf6

9 Qd1
Instead:
a) Wang Chen has also defended Black’s cause after 9 Qd3 Qb6 10 a3 Be7 11 e3 0-0 12 Be2 Rfd8
13 0-0 Rac8 when he again enjoyed the rapid and comfortable development characteristic of this line.
After 14 Rd1 Na5 15 Qc2 Qb3! 16 Qb1 (16 Qxb3 Nxb3 17 Rb1 Bf5 or 17 Ra2 Be6! wins material) 16 ...
Qb6 17 Nd4 Bd6 18 b4 Nc4 19 Qb3 Qc7 20 g3 a6 21 Bxc4 Qxc4 22 Qxc4 Rxc4 23 Bb2 Be5, Black’s
bishop pair game him reasonable compensation in Zhou Jianchao-Wang Chen, Chinese Team
Championship 2011 (½-½ in 64).
b) 9 Qb3 looks more natural, but in fact leaves Black with a number of reasonable responses since
the threat to the b7-pawn is illusory.
 b1) 9 ... Bc5 10 e3 0-0 11 Be2 Be6 12 Qa4 a6 13 0-0 b5 14 Qc2 Nb4 15 Qb1 Rc8 16 Rd1 Qe7 17
Ng5 (Houdini likes White after 17 h3, but I don’t clearly see how he is going to unravel) 17 ... Bc4 18 a3
Bxe2 19 Nxe2 Rfd8! with full compensation in Ki.Georgiev-N.Sedlak, Montenegrin Team Championship
2010 (½-½ in 30).
b2) 9 ... Be6 is interesting too: 10 Qa4 (10 Qxb7 Nb4 forces a draw; for instance, 11 Qb5+ Bd7 12
Qe5+ Be6 13 Qb5+ Qd7 or 11 Nd4 Rb8 12 Qxa7 Ra8 13 Qb7 Rb8 etc) 10 ... a6 11 Be3 b5 12 Qd1 Qc7
13 Bg5 Be7 with advantage to Black who has gained several tempi on the main line, Ge.Grigore-L.Miron,
Bucharest 2011 (0-1 in 28).
9 ... Bc5
Black has tried other moves, such as 9 ... Bb4 and 9 ... Qa5, but there seems no reason to avoid
putting the bishop on its most active square.

10 a3!?
A fresh and interesting attempt to pose Black some problems, and advocated by Kaufman in his
repertoire book.
The traditional main line runs 10 e3 Qe7 11 Be2 0-0-0 12 0-0 and now:
 a) I like the prophylactic move 12 ... Kb8! – the black king needs to step off the c-file in any event,
so Black takes care of this at once while retaining all his aggressive options. Now White has not yet tried
13 b4 in practice, perhaps because Black’s prospects seem better than with 12 ... g5 13 b4, as after 13 ...
Bxb4 14 Bb2 Black isn’t weak on the a1-h8 diagonal, while 14 Nb5 only hits a well-guarded pawn on a7.
However, 13 b4 may still be worth investigating, since the quieter moves seem to give Black good play.
For instance, 13 a3 (how quickly White’s game can go downhill is demonstrated by the following game:
13 Bd2 g5 14 b4 Bxb4 15 Nb5 Bc5 16 Bc3 g4 17 Nfd4 Ne4 18 Bb2 Rhg8 19 Nxc6+ Bxc6 20 Nd4 Bxd4
21 Bxd4 Nc3 22 Bxc3 Qe4 and White resigned in P.Amoyal-L.Miron, Saint-Lo 2012 – not model play by
White, it has to be said!) 13 ... g5 14 b4 g4! 15 bxc5 (15 Nd2 Bb6 was agreed drawn in D.Berczes-
G.M.Todorovic, Budapest 2012, which suggests that White didn’t like his position too much; 15 Nd4, as
in B.Dubessay-L.Miron, Fourmies 2013, looks like a better try) 15 ... gxf3 16 Bxf3 Ne5,

when Black is temporarily two pawns down but this is fully compensated for by his initiative – in
addition, the position is much harder to play with White, who immediately goes wrong) 17 Be2? Bh3! 18
Qb3 Rhg8 19 g3 Bxf1 20 Bxf1 (20 Kxf1 Qxc5 looks a little better for Black too) 20 ... Nf3+ 21 Kh1 (21
Kg2 improves, but after 21 ... Nh4+ 22 Kg1 Qxc5 White doesn’t have enough for the exchange) 21 ... Ng4
with a material advantage and an attack in J.Richardson-R.Pert, London 2011 (0-1 in 32).
b) The direct 12 ... g5 launches into complications with pawn storms and opposite-side castling. It
seems well established that White’s best approach is not to nurse his extra pawn but to strive to open files
on the queenside with the aggressive 13 b4!. White has scored heavily after 13 ... Bxb4 (Black hasn’t
done any better with 13 ... g4 14 Nd4) and now:

b1) 14 Qb3 is the most popular move, but Wang Chen has demonstrated Black’s resources here: 14
... Kb8 15 Rb1 Be6 16 Qa4 Bf5 17 Bb2? (Krasenkow suggests 17 e4 with unclear play) 17 ... g4 18 Nb5
Bxb1 19 Rxb1 gxf3 20 Bxf3 Ne4 21 Bxh8 Rxh8 22 Nxa7 Nxa7 23 Rxb4 Rc8 and White didn’t have
enough for the piece in Yu Yangwi-Wang Chen, Qinhuangdao 2011 (0-1 in 66).
b2) 14 Bb2 g4 15 Nd4 Kb8! has also been tested, with promising play for Black in an unclear
position, for instance V.Shishkin-T.Fodor, Baia Sprie 2012 (0-1 in 69).
b3) 14 Nb5!? is an aggressive recent attempt:

14 ... Ne4 15 Qc2 Kb8 16 Bb2 Rhe8 17 Ne5! Nxe5 (Krasenkow gives 17 ... f5 18 Rad1 Nxe5 19
Qc7+ Ka8 20 Bxe5 Qxe5 21 Rxd7 Rxd7 22 Qxd7 Nf6 23 Qf7 Rf8 24 Qc4 or 21 ... Qxc7 22 Nxc7+ Kb8
23 Rxh7 Re7 24 Rxe7 Bxe7 25 Ne6 Rd2 26 Bc4 with a slight advantage for White in view of his extra
pawn, though Black retains drawing chances with his active pieces and opposite-coloured bishops) 18
Qxe4 Bc5 19 Rfd1 f6 20 Rac1 Bc6 21 Qf5 Rxd1+ 22 Rxd1 a6 23 Nd4 and Black had regained his
material, but the much more active white pieces (with outposts on d4 and f5) left White clearly better in
I.Ivanisevic-S.Sulskis, European Team Championship, Porto Carras 2011 (1-0 in 61).
10 ... 0-0
Wang Chen deviates from his seventh round game against the very strong GM Zhao Jun, which saw
10 ... Qe7 11 Bg5 (11 e3 would transpose to the note with 8 e3 above) 11 ... h6 12 Bxf6 Qxf6 13 Rc1
Bb6 14 e3 Rd8 (Kaufman analyses 14 ... 0-0-0 15 Bb5 Kb8 16 Qe2 “and Black has very little for the
pawn”, but 14 ... Rd8 is better and more consistent with how Wang Chen handles the position) 15 Qc2
Bf5 16 Qa4 0-0 17 Be2.

Question: What does Black have for his pawn deficit?

Answer: Black has two pluses in this position. First, he has the bishop pair, which can never be
underestimated, especially in an open position such as this. In addition to their inherent ability to create
and support active counterplay, the bishop pair often gives rise to good drawing chances in other ways –
for instance, if the queens, rooks and a pair of knights were exchanged, if Black could trade his light-
squared bishop for the remaining white knight, then opposite-coloured bishops would be left on the board,
which exhibit strong drawing tendencies in heavily simplified positions.
Second, he has a lead in development, and his pieces are more actively placed. However, he needs
to act quickly since White is ready to castle and neutralize Black’s development advantage.

Question: How would you continue with Black?

Answer: 17 ... g5! – I really like this move! Because White has no weaknesses, it is hard to see how
Black makes progress after more normal continuations; for instance, after 17 ... Rfe8 18 0-0 Black has no
immediate threats and White is preparing to fight for control of the d-file. Wang Chen’s aggressive pawn
move aims to disturb the white knight on f3, with a number of benefits:
1. White’s kingside loses a key defender.
2. Black gains access to the d2-square, which dramatically enhances the value of his control of the
d-file.
3. Black’s knight can occupy a strong square on e5.

It should be noted that White is still perfectly fine in this position – as mentioned before, he has no
weaknesses and his development is about to be completed. But Wang Chen’s aggressive play means that
he keeps sufficient compensation for the pawn.
The game continued 18 0-0 g4 19 Ne1 Ne5 20 Ne4 Qg6 21 Ng3 Be6 22 Qf4 Qg5! 23 Qe4 (Black is
happy to play the endgame after 23 Qxg5+ hxg5, as the white pieces are in a bind and, in addition to other
plans, Black can consider mobilizing his kingside pawns with ... f7-f5-f4) 23 ... Bd5 24 Qf4 Be6 and if I
were White, I would be tempted to take the repetition here. Instead, he played on with 25 Ne4, but after
25 ... Qxf4 26 exf4 Ng6 27 g3 f5 Black had full compensation and went on to win in Zhao Jun-Wang
Chen, Chinese Championship, Xinghua 2012 (0-1 in 48).

11 Bg5
Kaufman advocates 11 Bf4 Qb6 12 e3. Frankly, after 12 ... Rfd8 this looks like Black’s normal
compensation for the pawn in this line, namely better development. In the absence of white weaknesses I
would suggest that Black has about half a pawn’s worth of compensation, with some additional practical
chances based on his position being easier to play. Certainly Wang Chen’s games have demonstrated that
these positions are not easy to handle, even for very strong GMs.
11 ... Qb6 12 e3 Be7
Not 12 ... Qxb2? 13 Bxf6, since 13 ... gxf6? 14 Na4 wins.
13 Bd3

13 ... Bg4
The b-pawn is poisoned, because of the standard idea 13 ... Qxb2?? 14 Na4, trapping the black
queen.
14 Qc2 h6 15 Bh4 Rfd8
Perhaps Black should look for an improvement around here. For instance, the immediate 15 ... Bxf3
was interesting: 16 gxf3 Ne5 17 Be2 Rac8 18 0-0 Ng6 19 Bg3 h5!?, attempting to provoke a further
weakness in the white kingside structure.
16 0-0 Bxf3 17 gxf3 Nd5 18 Nxd5 Rxd5 19 Bg3

White now has a clear advantage: his two bishops control a lot of squares and prevent Black from
asserting that he has real compensation for his pawn deficit. The only “weakness” Black might point to is
the doubled f-pawns in front of the white king, but these are hard to attack and, in fact, cover a number of
useful squares.
19 ... Bf6 20 Rab1 Rad8 21 Rfd1 h5 22 Bh7+ Kh8 23 Rxd5 Rxd5 24 Be4 Rd7 25 h3 h4 26 Bf4
Qa5 27 Rd1 Rxd1+ 28 Qxd1 g5 29 Bh2 Kg7 30 Qd7 Bxb2 31 Qxb7 Be5 32 Bxc6 1-0
I think this gambit is underrated, and it seems several top Chinese GMs might now agree with me!

Game 25
M.Parligras-L.Miron
Rumanian Championship, Sarata Monteoru 2011

1 d4 d5 2 c4 e6 3 Nc3 c5 4 cxd5 exd5 5 Nf3 Nc6 6 g3 c4

The Swedish Variation. Here’s an introduction from Tarrasch expert, Svetozar Gligoric. “The so-
called Swedish variation of the Tarrasch Defence, where Black relaxes the tension in the centre, creates a
pawn majority on the queenside and closes the d-file in order not to be left with an isolated d-pawn.
There is a price for this – some delay with the development of the pieces.”

7 Bg2
Gligoric notes that the double pawn sacrifice with 7 e4 dxe4 8 Ng5 Qxd4 is risky:
 a) 9 Bf4 Bb4 10 Qc2 (not 10 Ngxe4?? Qxe4+ and White resigned in S.Khalilbeili-S.Furman, Tbilisi
1956 – no comment required!) 10 ... Bg4 and Black was well on top in V.Winz-G.Stahlberg, Mar del
Plata 1941 (0-1 in 40).
b ) 9 Be3 Qxd1+ 10 Rxd1 h6 (Scherbakov suggests 10 ... Bg4!?, since 11 Bxc4 Bxd1 12 Nxf7 is
much better for Black) 11 Nd5 Bb4+ 12 Nxb4 Nxb4 13 Nxe4 Nc2+ 14 Kd2 Nxe3 15 Kxe3 Be6 16 Bxc4
Bxc4 17 Nd6+ Ke7 18 Nxc4 Nf6 with a level endgame in A.Zaitsev-V.Mikenas, USSR Championship,
Yerevan 1962 (½-½ in 29).
7 ... Bb4 8 0-0 Nge7

9 e4
This principled break (now or in the next few moves) is White’s only real attempt at an advantage.
The immediate 9 a3 often transposes after 9 ... Ba5 10 e4 dxe4 11 Nxe4 0-0, but here Black can try 9
... Bxc3 10 bxc3 0-0 11 Nd2 Be6 12 Re1 b5 13 Rb1 Rb8 14 e4 a5 15 Nf1 dxe4 16 Bxe4 Nd5 with an
active position in A.Onischuk-V.Akobian, Merida 2008 (½-½ in 43).
White can insert 9 Ne5 0-0 10 Nxc6 bxc6 before breaking with 11 e4, but then Black has a number
of sound continuations; for instance, 11 ... Be6 12 a3 Bxc3 13 bxc3 dxe4 14 Bxe4 Bd5 with a solid
position in P.Cooksey-A.Kosten, British League 2004 (½-½ in 46).
9 ... dxe4
Black can also let White capture on d5 after 9 ... 0-0, but I prefer the text.
10 Nxe4 0-0

11 a3
11 Qc2 looks critical; for instance, 11 ... Bg4 12 Qxc4 Bxf3 13 Bxf3 Qxd4 was Nguyen Thi Thanh
An-N.Zhukova, World Team Championship, Mardin 2011, and now, rather than taking on d4, all queen
retreats promise White a slight advantage.
11 ... Ba5 12 Qa4 Bb6

13 Be3
Instead:
a) 13 Qxc4 was Gligoric’s choice. He commented: “This complying with the opponent’s intentions
is actually the result of the better arrangement of White’s pieces.” S.Gligoric-Bo.Kostic, Yugoslav
Championship, Ljubljana 1947 (1-0 in 81), continued 13 ... Nxd4 14 Nxd4 Qxd4 15 Qe2 – “The key
move, after which it is clear that Black’s position is difficult. The black queen is exposed and the b7
square is weak” (Gligoric). However, modern engines instantly find 13 ... Be6!, followed by 14 ... Nxd4
with equality.
b) 13 Bg5 h6 14 Bxe7 Qxe7 15 d5 Qxe4 16 dxc6 Qxc6 (16 ... bxc6 17 Nd2 was M.Euwe-G.Stoltz,
Zaandam/Wormerveer 1946, and now Euwe suggests 17 ... Qe2 with equality) 17 Qxc6 bxc6 18 Ne5 Ba6
19 Nxc6 Rae8 and Black had no problems in D.Hergott-O.Rodriguez Vargas, Manila Olympiad 1992 (½-
½ in 27).
13 ... Nf5
Alternatively:
a) 13 ... Be6 14 Nc5 (14 Rad1 Nd5 15 Neg5 was L.Bernal Moro-O.Rodriguez Vargas, Spanish
Championship, Lleida 1991, when 15 ... Bf5! 16 Qxc4 Rc8 would have given Black interesting
compensation – he plans to take on e3 and build up on the e-pawn, trying to open the game for his
unopposed dark-squared bishop) 14 ... Bxc5 15 dxc5 was V.Korchnoi-O.Rodriguez Vargas, Salamanca
1991, and now I think 15 ... Nf5 equalizes. Gligoric believed that this had to be preceded by ... h7-h6 in
view of 16 Ng5, but in fact 16 ... Bd5 is completely fine for Black. Instead, 16 Bf4 is a better try, but then
16 ... Qe7 (targeting c5) maintains the balance.
b) 13 ... Bf5 14 Nc5 Bxc5 15 dxc5 Bd3 16 Rfd1 Nf5 17 Bf4 Qf6 was also fine for Black in J.Gomez
Esteban-O.Rodriguez Vargas, Salamanca 1991 (½-½ in 36), but here the intermezzo 17 Bg5! f6 18 Bf4
would have been better for White.
14 Qxc4 Qe7

15 Nc5!?
Returning the pawn is a practical choice, but White can play more ambitiously:
a) 15 d5!? leads to complicated play, where it seems that Black’s chances are no worse after 15 ...
Nxe3 16 fxe3 Bf5 17 Nfd2 Ne5!.
b) 15 Qd3 Rd8 16 Nc3 (since the d-pawn is poisoned: 16 ... Ncxd4? 17 Nxd4 Bxd4 18 Nd5 Qe5 19
Bf4 or 18 ... Rxd5 19 Bxd5 Nxe3 20 fxe3 with insufficient compensation for the exchange; note that Black
still needs to develop his queenside pieces) 16 ... Be6! gives Black decent compensation for the pawn.
Perhaps White’s best is to return it immediately; i.e. 17 Rad1 Ncxd4 18 Nxd4 Bxd4 19 Bxd4 Nxd4 20
Qe3 with an edge in view of his dominant g2-bishop.
15 ... Ncxd4 16 Bxd4 Nxd4 17 b4
 This is a critical position, and one of a type which frequently occurs in the Tarrasch after the white
e-pawn and the black d-pawn have been traded. From a structural point of view, this suggests that Black
should have equalized, since he no longer has any formal weaknesses and the pawn structure is
completely symmetrical. However, Black needs to be extremely careful over the next few moves. White,
by virtue of moving first, has a lead in development and, in particular, his bishop on g2 is much stronger
than its counterpart on c8. Black needs to find a decent way to develop his c8-bishop without losing the
b7-pawn.

Exercise: What would you play here with Black?

17 ... Nxf3+
Going in for a slightly speculative sacrifice; however, I think it would have been simplest to take on
c5 immediately.
Answer: 17 ... Bxc5! 18 bxc5 (18 Qxc5 Qxc5 19 bxc5 Nb3 collects the c5-pawn, though after 20
Rab1 Nxc5 21 Nd4 Black will have to return the pawn – on b7 – in order to untangle) 18 ... Nxf3+ 19
Bxf3 Be6 20 Rfe1 Qf6 21 Qd3 Rac8 22 Qe3 Qe7 and the b7 and c5-pawns will be exchanged, with a
draw.
18 Bxf3 Rb8
It is too late for 18 ... Bxc5 in view of 19 Qxc5! with advantage.
19 Rfe1 Qf6 20 Bd5
 This is now slightly unpleasant for Black, who needs to find a way to complete development. Miron
comes up with an enterprising pawn sacrifice which seems to me to level the chances.
20 ... Bh3
20 ... h5!? was logical too.
21 Ra2 Rbe8 22 Rae2 Rxe2 23 Qxe2 g6 24 Bxb7 Rd8
24 ... Qc3 is more active, since the endgame after 25 Qe3 Qxe3 26 Rxe3 should be an easy hold for
Black.
25 Qe3
25 Rc1 retained some advantage.
25 ... Kg7 26 Bg2 Bf5 27 h4 Rd3 28 Qf4 Rxa3 29 Re8 Ra1+ 30 Kh2 Rd1 31 h5 g5 32 Qe3 Rd8
33 Re5 h6 ½-½

The game score in my database shows that a draw agreed was after 34 Qf4??, which I find hard to
believe for obvious reasons! Instead, I guess the players shook hands here. Even now I would play on
with Black, who has the bishop pair if nothing else. The bishop on b6 is contained by the knight on c5 for
the moment, but possibilities of ... a7-a5 or ... Bc7 (or both) mean that only Black has winning chances
here, even if White should be able to hold.

Key Notes
1. The gambit variation 4 ... cxd4!? leads to interesting complications. Wang Chen has demonstrated how
enduring Black’s development-based compensation can be, even against very strong GMs, and despite
White’s lack of weaknesses.
2. The Swedish Variation, 6 ... c4, leads to positions analogous to the 9 Bg5 c4 lines. As there, White
needs to play actively in order to avoid falling into a worse position. I think this is a perfect surprise
weapon, especially for players who like 9 Bg5 c4.
Index of Complete Games
Akopian.V-Jianu.V, European Championship, Legnica 2013
Aronian.L-Melkumyan.H, German League 2013
Babula.V-Orsag.M, Czech Championship, Ostrava 2010
Baburin.A-Collins.S, Bunratty 2011
Bu Xiangzhi-Li Shilong, Chinese Championship, Xinghua 2013
Carlsen.Ma-Radjabov.T, Sandnes 2013
Chatalbashev.B-Sokolov.I, Jakarta 2012
Giri.A-Swinkels.R, Dutch Championship, Amsterdam 2012
Kasimdzhanov.R-Berelowitsch.A, German League 2011
Korchnoi.V-Collins.S, San Sebastian 2011
Kunte.A-Aravindh.C, New Delhi 2013
L’Ami.E-Aagaard.J, Helsingor 2012
Li Shilong-Braun.Ar, Wijk aan Zee 2008
Manolache.M-Jianu.V, Baia Sprie 2012
Movsesian.S-Sokolov.I, Sarajevo 2009
Mulyar.M-Perunovic.M, Chicago 2012
Onischuk.A-Akobian.V, Merida 2008
Parligras.M-Miron.L, Rumanian Championship, Sarata Monteoru 2011
Reinderman.D-Irwanto.S, Jakarta 2011
Rewitz.P-Aagaard.J, Danish Team Championship 2011
Sargissian.G-Halkias.S, European Championship, Aix-les-Bains 2011
Schandorff.L-Aagaard.J, Danish Championship, Odense 2011
Socko.M-Petrosian.T.L, Gibraltar 2008
Sokolov.I-Petrosian.T.L, European Cup, Kemer 2007
Wang Yue-Wang Chen, Chinese Championship, Xinghua 2012