First published in 2017 by Gloucester Publishers Limited, London

Copyright © 2017 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 405 9

ISBN Epub: 978 1 78194 406 6

Distributed in North America by National Book Network,

15200 NBN Way, Blue Ridge Summit, PA 17214. Ph: 717.794.3800.

Distributed in Europe by Central Books Ltd.,

Central Books Ltd, 50 Freshwater Road, Chadwell Heath, London , RM8
1RX.

All other sales enquiries should be directed to Everyman Chess.

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
Typeset and edited by First Rank Publishing, Brighton.
Cover design by Horatio Monteverde.
Printed by TJ International Limited, Padstow, Cornwall.
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
The Tarrasch Defence: Move by Move
Karpov: Move by Move
Contents
About the Author
Bibliography
Introduction

1 The Fianchetto Variation

2 The Classical Variation
3 The Sämisch Variation
4 Systems with h2-h3
5 Systems with Be2
6 Other Lines

Index of Complete Games

Bibliography
Books
A Practical Black Repertoire with Nf6, g6, d6, Volumes 1&2, Alexei Kornev
(Chess Stars 2016)
Bologan’s King’s Indian: A Modern Repertoire for Black, Victor Bologan
(New in Chess 2017)
Chess Structures, Mauricio Flores Rios (Quality Chess 2015)
Grandmaster Versus Amateur, Jacob Aagaard & John Shaw (Quality Chess
2011)
Grandmaster Repertoire 1 d4, Volume Two, Boris Avrukh (Quality Chess
2010)
King’s Indian Warfare, Ilya Smirin (Quality Chess 2016)
Kotronias on the King’s Indian, Volumes 1-4, Vassilios Kotronias (Quality
Chess 2013-16)
Modernized: The King’s Indian Defence, Dejan Bojkov (Metropolitan Chess
2014)
Most Stunning Victories of 2016, Arkadiy Naiditsch, Csaba Balogh &
Sébastien Mazé (Chess Evolution 2017)
Opening Repertoire: The English, David Cummings (Everyman Chess 2016)
Opening Repertoire: The Fianchetto System, Damian Lemos (Everyman
Chess 2014)
Playing 1 d4 – The Indian Defences, Lars Schandorff (Quality Chess 2012)
Play the King’s Indian, Joe Gallagher (Everyman Chess 2004)
Positional Masterpieces of 2012-2015, Arkadiy Naiditsch & Csaba Balogh
(Chess Evolution 2016)
The Dynamic English, Tony Kosten (Gambit 1999)
The English Opening: Volume 2, Mihail Marin (Quality Chess 2010)
The Ultimate Anti-Grünfeld: A Sämisch Repertoire, Dmitry Svetushkin
(Chess Stars 2013)

DVDs & Databases

A World Champion’s Guide to the King’s Indian (2nd edition), Rustam
Kasimdzhanov (ChessBase 2010)
King’s Indian: A Modern Approach, Victor Bologan (ChessBase 2015)
GM Repertoire Against the King’s Indian Defence, Aleksander Delchev
(Modern Chess 2016)
Mega Database 2017 (ChessBase)
Solid and Safe against the Wild Indians, Nicholas Pert (ChessBase 2015)

Periodicals
ChessBase Magazine
New in Chess
Websites
chess24.com
chessbase.com
chesspublishing.com
twitter.com
youtube.com
Introduction
1 d4 Nf6 2 c4 g6 3 Nc3 Bg7 4 e4 d6

The King’s Indian (the main position of which is shown in the diagram,
but which also includes some other systems) is an opening that needs little
introduction for most players It was Fischer and Kasparov’s main weapon
with Black in their peak years and continues to be featured in games at the
highest level.

Why Play the King’s Indian?

The King’s Indian has many obvious and subtle advantages, which I hope
will become apparent from reading this book. I’d like to highlight two in
particular:

1) Black plays for three results

“No white player playing 1 d4 likes to get mated. They like long endings
where they’re slightly better, they like to attack the isolated pawn, they like to
have two bishops in the Nimzo-Indian, what they do not like is to get mated.
And still, every now and again, they do, although by now we have 100 years
of experience in this opening.” – GM Rustam Kasimdzhanov.
In his instructive series on How to build your 1 d4 repertoire GM Jan
Gustafsson describes the King’s Indian as his “personal nemesis amongst
chess openings”, noting that top players use it to play for a win against
weaker opponents and that he has suffered more defeats in the King’s Indian
than in all other openings combined.
Of course there are systems within the King’s Indian where White can
play with less risk (such as the Fianchetto Variation, which is dealt with in
the first part of the book), but the fact remains that strong players can get
quickly mated with White in this opening if they make a few inaccuracies.
Moreover, as we will see, Black is not merely relying on a rapid attack on the
king, but can play a more positional game, for example by exploiting the
weakness of d4, or by exerting pressure on White’s queenside (perhaps in
combination with a Benko-style ... b7-b5 sacrifice). Overall, despite the very
broad range of systems at White’s disposal, it is quite difficult to identify a
safe line where White plays for an edge with no risk. This alone makes the
King’s Indian a very attractive opening, and one which may be played in
most or all of your black games against everything except 1 e4.

2) The King’s Indian is universal

“The King’s Indian is an opening you employ with Black against 1 d4, or
basically against any other move which belongs to the so-called ‘closed
openings’, so basically anything which is not 1 e4. 1 c4, 1 Nf3, 1 d4, 1 g3, 1
b3, anything” – GM Rustam Kasimdzhanov
“The main advantage from a black point of view is that the King’s Indian
you can employ against pretty much anything and there’s not much cunning
White can use regarding the move order.” – GM Jan Gustafsson
“I have more frequently favoured queen’s pawn openings, and have found
that a healthy mix of 1 d4, 1 Nf3 and 1 c4 works well. Many players have the
basis of a decent repertoire, but have failed to pay sufficient attention to the
subtleties of different move orders.” – GM Peter Heine Nielsen
A trend I have noticed in my own games, and in those of the players I
follow, is that 1 d4 players often start their games with 1 Nf3 and 1 c4. This
certainly seems to be the case in elite tournaments, where White has
struggled to demonstrate anything against the Grünfeld, and lower rated
players have copied the trend. One of the practical advantages of the King’s
Indian, compared with the Grünfeld or the Nimzo-Indian, is that it is a
complete system against closed openings and is not heavily dependent on
White employing a particular move order. This cuts down on the amount of
preparation required before a particular game – although, it must be admitted,
there’s plenty to study in the King’s Indian main lines in any event. Perhaps
the main benefit, if a less tangible one, is that Black knows that he can expect
a similar type of position against all these first moves, and can build
experience in these structures rather than, for example, learning a bespoke
line against the English which he only plays every couple of months.

A (Very) Brief Structural Overview

Partly because Black doesn’t immediately contest the centre with pawns,
White has a very broad range of possible ways to set up. I had to dispense
with the idea of a detailed “structural introduction” section due to the sheer
number of distinct pawn structures which can arise from this opening, in
particular depending on whether Black challenges the centre with ... e7-e5 or
... c7-c5, each of which gives White different options in reply. Even listing
the main structural possibilities takes some time.

1) Black plays ... e7-e5

1. White answers with d4-d5.
 This is the pawn structure most characteristic of the King’s Indian,
though it can also arise from other openings such as the Breyer Ruy Lopez
and the Old Indian (which is basically an inferior King’s Indian). Generally
speaking, White will play on the queenside (typically with c4-c5), while
Black will arrange kingside counterplay with ... f7-f5 and, if White plays f2-
f3, then a further ... f5-f4 and ... g6-g5-g4. This is the pure race situation,
though many subtleties can be introduced in the systems where either side
plays prophylactically ( ... a7-a5 or ... c7-c5 by Black, an early g2-g4 by
White).

2. White takes on e5 and Black recaptures with the pawn.

 This structure is outwardly symmetrical but the d4-square is an outpost.
The combination of these two factors means that many inexperienced players
rush into this structure with White (for example, in the Exchange Variation),
hoping for an easy draw against a stronger player, only to be outplayed in
convincing fashion. The structure is not inherently bad for White – for
instance, when Black plays ... c7-c6 (which is often necessary), White has
good plans based on c4-c5 and bringing a knight to d6 – but it does demand
energetic play.

3. Black takes on d4 (or White takes on e5 and Black recaptures with a

piece, resulting in the same structure).
 I haven’t recommended many lines resulting in this structure, with the
exception of the Gligoric Variation. Exchanging Black’s only pawn on the
fourth rank is, in positional terms, a concession, and heightens Black’s
responsibility to generate active counterplay (a concept I discuss further at
the end of the Introduction), relying on the King’s Indian bishop on g7 which
really shines in this structure. If you like these positions then I can
recommend Dejan Bojkov’s instructive book, listed in the Bibliography,
where he built a repertoire largely around this structure.

2) Black plays ... c7-c5

1. White answers with d4-d5.
 This leads to a family of structures known from the Benoni and the
Benko. Black will typically aim for ... e7-e6 and ... e6xd5 when, as you can
see, three different structures are possible depending on whether White takes
on d5 with the c-pawn, the e-pawn or the knight. Black will also aim for
queenside counterplay with ... b7-b5, sometimes sacrificing a pawn to
achieve this end.

2. White takes on c5.

 This is not a good structural trade from White’s perspective and is almost
exclusively connected with the idea of winning the pawn on c5, as we will
see in the Sämisch. I like Black’s compensation and I hope you will too.

3. Black takes on d4.

 This leads to a Maróczy Bind structure, characteristic of many variations
of the English and the Accelerated Dragon. Generally it tends to lead to
positions which are slightly too passive for my taste, but in some lines I have
recommended going into favourable versions.

Model Players
Many strong players have adopted the King’s Indian throughout their careers
and it’s a great idea to follow their games. Of course, the games of Fischer
and Kasparov must be studied by anyone who wishes to improve, or just to
enjoy classical chess heritage, and these include many wonderful King’s
Indians.
Of the modern elite, the King’s Indian is regularly essayed by GMs
Nakamura, Grischuk, Ding Liren, and Radjabov.
Several other GMs are worth following if you adopt this opening, of
whom I would mention in particular Ilya Smirin (whose excellent recent book
on his games in this opening is required reading), Arkady Naiditsch, Etienne
Bacrot, Ivan Saric, Sebastian Mazé, and Gawain Jones. I am particularly
impressed by Gawain’s recent games which, despite a lot of variation with
White, and with Black against 1 e4, show an astonishing faithfulness to the
King’s Indian against all levels of opposition, with excellent results.

Two Final Thoughts on the King’s Indian

This section is slightly unusual for an opening book, and players interested in
the theory of the opening can probably skip it, but these concepts are central
to my understanding of the King’s Indian and the spirit in which it should be
played.

1) Responsibility
One of the best descriptions I’ve heard of the nature of the positions arising
in the King’s Indian was that of my friend and coach GM Jacob Aagaard:
namely that the level of responsibility in this opening is very high. In more
concrete terms, from the opening moves Black concedes central space and,
unless he can quickly generate counterplay, will find himself in a
prospectless position. In many lines of the Classical Variation, White is
winning on the queenside, so Black will likely lose the game if his kingside
attack fails. As we will see, Black often sacrifices material for positional
gains or to seize the initiative, and such sacrifices also carry their own
responsibilities, both in evaluating them correctly beforehand and in playing
the position well thereafter.
I would put a gloss on this point by saying that the level of responsibility
is high for both sides (as set out above, it is difficult for White to play without
risk in this opening). One of the unusual features of the King’s Indian is that
many positions which are objectively lost for Black still retain prospects for
counterplay, and White often draws or even loses such positions in practice.
One of the first examples in GM Ilya Smirin’s recent book on the King’s
Indian is the game between Yannick Pelletier and Andreas Skytte Hagen in
Legnica 2013, where Black overcame a 200 point Elo deficit and won what
was, objectively, a completely lost position against an experienced GM. I
also like following the games of GM Mark Hebden in this opening, who very
frequently gets half or full points from positions which are, objectively, dead.

2) Complexity
The King’s Indian is staggeringly complex. As you can see from the
Bibliography, even in the last couple of years several large works have been
published on this opening, including repertoires for both colours. GM
Kotronias has produced a five-volume repertoire work on the opening (the
first volume of which is 720 pages). The opening has been played in
hundreds of thousands of games, and is a darling of correspondence players
who are pushing forward theory with increasingly powerful engines.
This touches on another unusual aspect of the King’s Indian, namely that
computers are of limited assistance (they tend to assess most positions in the
Classical Variation as winning for White because Black’s standard attack
takes so many moves to execute that the mate does not appear on the engines’
horizons). As explained by Ilya Smirin in his wonderful book, the King’s
Indian “is a big enough opening to have enough dimensions not to get
boring”. As with responsibility, complexity cuts both ways, so let’s get to
work and bring White players out of their comfort zone.

Dedication
I’m not normally in the habit of dedicating my chess books. However, I’d
like to dedicate this one to up and coming Irish junior players, many of whom
have risen through the ranks with astonishing speed, thanks to the efforts of
great teachers like GM Alex Baburin and Daniel Lynch and the support of the
Irish Chess Union. I strongly suspect that some of these talents will surpass
previous generations of Irish players and obtain titles, even the coveted GM
title. If they are to succeed in this goal, they had better know a good defence
to 1 d4, so this book is written for them.
Sam Collins, Dublin, July 2017
Chapter One
The Fianchetto Variation
Introduction

1 d4 Nf6 2 c4 g6 3 Nf3 Bg7 4 g3 0-0 5 Bg2 d6 6 0-0

The starting point for this book is the Fianchetto Variation. Most King’s
Indian books start with Black’s knockout attacks in the main lines of the
Classical Variation, and we’ll get to those shortly. However, in my
experience, the Fianchetto is the toughest challenge for a KID player to face.
White doesn’t aim for much, but completely changes the character of the
play. Instead of being able to execute the standard plans, Black has a wide
choice of lines where White runs little risk.
At the start of 2015 I lost four consecutive games in the Fianchetto
Variation, against two IMs and two GMs. The fact that my positions in the
early phases of these games were perfectly playable, and that during the same
period I had convincingly beaten an experienced GM by 2½/3 in a training
match on the same territory, meant little by comparison and my enthusiasm
for the King’s Indian waned, driving me to other openings. One of my aims
in this book is to rekindle my own appetite for the King’s Indian, and a big
part of that is developing a good weapon against this line.
Black has a very wide range of options in the Fianchetto Variation (and
I’ve tried almost all of them), from going into English/Benoni/Benko territory
with an early ... c7-c5, to a dynamic Grünfeld with ... d7-d5, to a symmetrical
Grünfeld with ... c7-c6 and ... d7-d5. The pure King’s Indian responses are
based around an early ... d7-d6, but even here Black has a wide choice,
principally between the classical ... Nbd7 (aiming for ... e7-e5) and the more
modern ... Nc6 (which can be combined with various plans, including
queenside expansion with ... a7-a6, ... Rb8 and ... b7-b5).
My selected line for this book is a relatively recent trend, based on
putting the knight on c6 and pushing ... e7-e5, which has shown up repeatedly
in games of strong KID experts like GMs Arkady Naiditsch, Alexander
Onischuk, and Sebastian Mazé. Black obtains a similar set-up to some lines
of the Classical Variation (in particular, the Petrosian Variation) and, in my
view, the means of counterplay are quite typical for the KID (often based on
... f7-f5) and so should be relatively easy to play.

Game 1
A.Naumann-A.Naiditsch
German League 2016

1 d4 Nf6 2 c4 g6 3 Nf3 Bg7 4 g3

This move signals the Fianchetto Variation (which can be reached by
various move orders, including 3 g3). White adopts the same fianchetto set-
up as Black, claiming that his extra space with d2-d4 and c2-c4 should
promise some edge.
4 ... 0-0 5 Bg2 d6 6 0-0 Nc6
 The knight might look exposed here, but if White attacks it with d4-d5
(which is a decent move), he will weaken some squares and lose some
flexibility in his pawn chain while opening the long diagonal for the black
bishop. In the meantime the knight on c6 serves several functions including,
as we will see, supporting the ... e7-e5 push, one of Black’s standard moves
in the King’s Indian.
7 Nc3
By far the most popular move, developing the knight to its best square.
White can also kick the knight immediately with 7 d5, after which I think
Black should opt for a different set-up with 7 ... Na5 (7 ... Nb8 8 Nc3 e5!?
transposes to the main lines after 9 e4, but 9 dxe6 retains some pressure) and
now White has a couple of options:
a) 8 Nfd2 c6 9 Nc3 (9 Na3 was tried in A.Cherniaev-M.Cornette, London
2016, when 9 ... cxd5 10 cxd5 e6!? 11 dxe6 Bxe6 gives Black active play;
while 9 a3 Nd7 10 Ra2, Pr.Nikolic-D.Stellwagen, Amsterdam 2007, and now
10 ... Ne5 11 b3 cxd5 12 cxd5 Bd7 looks fine for Black) 9 ... cxd5 10 cxd5
Bd7 11 Nb3 Nc4 and Black had no problems in B.Esen-M.Annaberdiev,
Turkish League 2015 (½-½ in 35).
b) 8 Qa4 c5 sets up a very typical pawn structure where, in general, Black
aims to play on the queenside with ... a7-a6, ... Rb8 and ... b7-b5, while
White argues that the knight on a5 is misplaced and seeks to attack in the
centre and on the kingside. In this particular position, because White has put
his queen on a4, his plan is slightly different and involves directly targeting
the a5-knight with 9 Bd2 b6 10 Bc3. This has been tried by several strong
players including Mamedyarov. In reply I like 10 ... e6!? which exploits the
fact that the queen is not defending the d5-pawn:

b1) 11 Nfd2 exd5 12 cxd5 is a Benoni structure in which Black has

reasonable chances, not least because White will need to spend time
developing his queenside: 12 ... Ba6 13 Re1 Re8 14 Bf3 Qd7 (offering the
trade of queens, since most Benoni endgames are good for Black; White
more often aims for a direct attack and so typically avoids the exchange) 15
Qh4 Qe7 was fine for Black in B.Tadic-S.Brenjo, Serbian Team
Championship 2004 (½-½ in 43).
b2) 11 dxe6 fxe6 (now Black’s pawns control a lot of important squares
and there is no immediate way to exploit the h1-a8 diagonal) 12 Nbd2 (12
Bxa5 bxa5 looks dubious to me, even though the black structure is slightly
compromised, since White gives up the bishop pair while eliminating Black’s
worst piece and allowing him to use the half-open b-file, e.g. 13 Nc3 Rb8 14
Qc2 Bb7 15 Rad1 Qe7 with excellent counterplay in E.Scarella-P.Zarnicki,
Villa Martelli 1998, ½-½ in 42; while 12 Ng5 Rb8 13 Nd2 Ng4 14 Bxg7
Kxg7 15 h4 h6 16 Ngf3 Qe7 led to an unclear position where Black had no
problems, P.Vavrak-J.Markos, Slovakian Championship, Banska Stiavnica
2007, ½-½ in 19) 12 ... Rb8 13 Rad1 was J.Veng-I.Smirin, Helsingor 2015,
and now 13 ... Bb7 14 Ng5 Bxg2 15 Kxg2 (not 15 Nxe6?? Bc6 and wins) 15
... Qe7 is fine for Black, who has an active position with no weaknesses
(again, Bxa5 is not something to fear in my view).
7 ... e5

The most thematic move in the King’s Indian, fighting for control of the
central dark squares. Typically this gives White the choice between
maintaining the tension, exchanging or advancing. Here keeping the tension
leaves Black with a very comfortable position, so White’s choice is between
8 d5 and 8 dxe5.
8 d5
The most ambitious response, gaining space in the centre and preparing to
play on the queenside. 8 d5 by far the most popular move here (with over
1800 games in my database), but frankly this is what we’re hoping for; 8
dxe5 is a more annoying move, covered in the next game.
Instead, 8 h3 is quite playable but not a serious try for the advantage in
my opinion. This line was contested between two of my favourite players at
one of my favourite tournaments: 8 ... exd4 9 Nxd4 Nxd4 10 Qxd4 Re8 11
Qd3 Nd7 12 b3 Nc5 13 Qd2 Bf5 14 Bb2 Ne4 15 Bxe4 Bxe4 16 f3 Bc6 17 e4
Qd7 and having exchanged two sets of minor pieces, so his disadvantage in
space isn’t felt, Black had reached equality in Mi.Adams-G.Jones, Bunratty
2012 (½-½ in 64).
8 ... Nb8!?
This is the new idea, or at least the new fashion. Black aims to deploy his
pieces as follows: ... a7-a5, ... Na6, ... Nfd7, and occupying c5 with a knight.
A nice practical benefit is that this move isn’t mentioned in Avrukh’s 1 d4
repertoire series, which recommends 8 d5, though I have no doubt that the
second edition will cover this system in detail. Lemos doesn’t deal with the
move either.

Question: But Black has lost two whole tempi!

Answer: It’s true that Black has invested two moves in this operation, by
bringing his knight to c6 and returning home; but in order to gain time, White
has defined the central structure by pushing d4-d5. This is by no means a bad
move, gaining central space, but it means that play will be directed to the
wings: White will tend to play on the queenside, and Black will counterattack
on the kingside. And this is where the attraction of this line lies for Black.
A number of other systems against the Fianchetto Variation offer Black
prospects of counterplay while running some strategic risk; for example:
1. The classic ... Nbd7, ... e7-e5 and ... e5xd4 lines, as featured in the
stunning classic wins by Bronstein against Zita (Moscow 1946) and
Reshevsky (Zurich 1953) – if you haven’t seen these games, put this book
down and look them up immediately!
2. The Panno System – with the move order in this game, continue 7 ... a6
8 d5 (of course there are alternatives) 8 ... Na5 9 Nd2 c5. Black will follow
up with ... Rb8 and ... b7-b5 and try to create queenside counterplay.
3. The Gallagher Variation, based on ... Nbd7, ... e7-e5, ... e5xd4 and ...
a7-a6 when, having placed his rooks on e8 and b8, Black can create chaos
with ... c7-c5 and ... b7-b5, often sacrificing his d6-pawn in the process.
While these systems are all playable, for the most part they aim for
counterplay in the centre or on the queenside. The traditional kingside pawn
storm, which Black so enjoys in the Classical Variation, was believed simply
not to work against the Fianchetto. This 8 ... Nb8 variation is the only line
where I have seen these kingside pawn storms being launched (and landing!)
on a regular basis.
I might also note that this concept of losing time with the knight to
provoke d4-d5 is known from other openings, such as the Classical Pirc: 1 e4
d6 2 d4 Nf6 3 Nc3 g6 4 Nf3 Bg7 5 Be2 0-0 6 0-0 Nc6 7 d5 Nb8 (a favourite
of Mamedyarov, which he’s used to great effect in blitz games but also in
important events like the World Cup).
The traditional move here was 8 ... Ne7, with over 2000 games in my
database. I’ve wanted nothing to do with it since seeing A.Karpov-J.Polgar,
Las Palmas 1994 (1-0 in 28).
 9 e4

Question: How should we respond if White seizes queenside space with 9

b4 - ?

Answer: Black has the typical 9 ... a5!, when White can’t maintain his
structure with 10 a3? since 10 ... axb4 would win material due to the pin on
the a-file.
Instead, 10 bxa5 was played by the very strong Hungarian GM, Zoltan
Almasi (a good King’s Indian player himself), in a game against my friend
and clubmate GM Sebastian Mazé. Black won a great game which we will
follow: 10 ... Rxa5 11 Nd2 Nfd7! 12 Nb3 Ra8 13 a4 Na6 14 Ba3 f5 (Mazé
correctly assesses this position as slightly better for Black: note how
effectively he has regrouped his knights to prevent the c4-c5 break; this was a
much better plan than playing ... b7-b6, which would have given White a
chance to open lines on the queenside with a4-a5) 15 e4 (Mazé gives 15 a5 e4
16 Qc2 Ne5 17 Nb5 Bd7 as “also pleasant for Black; the bishop on g7
shines”) 15 ... f4 16 Nb5

Exercise: Black has a standard kingside regrouping here, improving his

co-ordination and increasing his control over important points like d6 and c5.
See if you can find it.

Answer: 16 ... Rf7! – an idea we will see in many variations, in particular

in the Classical Variation. Mazé explains it perfectly: “A typical plan in the
King’s Indian Attack. The rook goes to f7 to protect the seventh rank, but
also has the possibility to aid the attack on the g- or h-files. Then the bishop
drops back to f8 to help with this strategic idea and also to protect the d6-
pawn.”
The game continued 17 Qd3 Bf8 18 Rfd1 h5! (preparing to open the h-
file; 18 ... g5 is a common reaction to gain space, in particular where a white
pawn is on f3, and also seems to work well in the present position – at first I
was concerned that it gave White the chance to block the kingside with 19 g4
h5 20 h3, but the engine quickly points out that Black is simply winning after
the standard 20 ... hxg4 21 hxg4 Nf6 22 f3 Nxg4! 23 fxg4 Bxg4 24 Rd2 Bd7!
and rolling the kingside pawns with ... g5-g4 etc) 19 Qc3 h4 20 g4 (Mazé
gives 20 gxf4 exf4 21 N3d4 Bg7 as good for Black; again, the strength of the
g7-bishop is obvious)

Exercise: Time for some prophylactic thinking. What move does White
want to play? And how should Black prevent it?

Answer: White would like to support his g4-pawn and block the kingside
with 21 h3. Despite Black’s advantage in space on that side of the board, he
has no pawn breaks and it will take a long time to arrange a piece sacrifice to
open the position (compare this with the variation given at move 18), during
which period White can advance his play on the queenside. So Mazé strikes
first with an excellent move: 20 ... h3! (now the h-file will be open and the
g4-pawn is likely to fall) 21 Bxh3 (Mazé gives the instructive variation 21
Qxh3 Nf6 22 Bf3 Rh7 23 Qg2 Rh4 24 h3, when White seems for the moment
to have consolidated his kingside, but after 24 ... Nh7! 25 Rd3 Ng5 26 Be2
Qe7, Mazé correctly assesses the position as clearly better for Black: he has
several ideas including doubling – or tripling! – on the half-open h-file, while
White is condemned to passive defence; even computers prefer Black’s
position, and in a practical game the defence appears quite impossible) 21 ...
Qh4 22 Bg2 Qxg4 23 f3 Qh5 (now Black attacks with equal material and a
superior structure) 24 Rd2 Rh7 25 Bf1 Nf6 26 Raa2? (another instructive
moment: Mazé assesses this move as “too passive” and suggests that Almasi
had to seek counterplay with 26 c5! Nxc5 27 Nxc5 dxc5 28 Rf2 b6 29 Bb2
Bd6 30 a5 bxa5 31 Rxa5 Rb8 “and Black has a decent advantage”; at least
White has some hopes of counterplay here, whereas in the game he is
crushed) 26 ... g5 (as Mazé notes, “Black’s plan is very easy, just push the g-
pawn and then create some powerful threat with all the pieces and the f-pawn.
White can’t stop it – it’s funny to see all of White’s pieces on the queenside”)
27 c5 g4 28 Rg2 Kh8 29 Be2, Z.Almasi-S.Mazé, Stockholm 2016. The
players were now in huge time trouble and traded inaccuracies, though Mazé
never let the win slip (0-1 in 40). Here the most clinical finish was 29 ... g3!
30 Bf1 (30 hxg3 fxg3 quickly leads to mate or decisive material loss) 30 ...
Bh3 and White collapses on the kingside. A great King’s Indian game!
 Exercise: In this line Black has the opportunity to establish a temporary
blockade on the queenside. How should he start?

Answer:
9 ... a5!
A very typical idea in response to d4-d5. White wants to push b2-b4 and
c4-c5 so, to prevent or delay this, Black moves his pawn to a5 and his knight
to a6. White will then need to spend time on a2-a3 and Rb1 to arrange b2-b4,
while Black will have ... a5xb4 in response, reducing potential targets on the
queenside.
10 Ne1!
A very typical manoeuvre in the Fianchetto and Classical Variations. The
knight doesn’t have much to do on f3 and reroutes to d3 where it supports the
c4-c5 break. In the meantime, White’s f-pawn is freed to support or attack the
centre, by playing f2-f3 or f2-f4 respectively.
10 ... Na6 11 Nd3 Nd7
 Controlling c5 and preparing ... f7-f5.
12 Be3 f5

The majority of GM games in this line seem to gravitate towards this

position, which isn’t surprising since the last few moves have been very
natural. Now White has a wide (and important) choice of how to continue
and, in particular, which kingside structure he’s willing to play.
13 Qd2
Developing, connecting the rooks, and discouraging ... f5-f4.
Alternatively:
a) 13 exf5 gxf5 14 f4 is a typical reaction to Black’s ... f7-f5 push in the
fianchetto systems. Here I like 14 ... e4 (14 ... Nb6 was tried in
M.Mchedlishvili-V.Onischuk, World Blitz Championship, Dubai 2014, 0-1
in 57) 15 Nf2 Nb6 (15 ... Nac5 16 Rc1 Nf6 17 Bd4 Bd7 was also reasonable
in J.Vakhidov-K.Kalavannan, Telford 2016, 1-0 in 66) 16 Bd4 Nc5 17 Bxg7
Kxg7 18 Qd4+ Qf6 19 Rfd1 Bd7 20 Qxf6+ Kxf6 and Black has no problems
in the endgame, H.Melkumyan-V.Onischuk, World Blitz Championship,
Berlin 2015 (1-0 in 74).
 b) 13 Rc1 Nac5 14 Nxc5 Nxc5 15 Qd2 b6 16 f4 was agreed drawn
already in I.Stohl-V.Rasik, Czech League 2016. Black would be very
comfortable after 16 ... exf4 17 Rxf4 fxe4 18 Nxe4 Rxf4 19 Bxf4 Nxe4 20
Bxe4 Qf6 21 b3 a4 with promising queenside play (I can’t even call it
“counterplay” since I don’t see where White’s play is coming from!).
c) 13 a3 Ndc5 (more enterprising types might be tempted by the
promising pawn sacrifice 13 ... f4!? 14 gxf4 Qh4 with decent chances) 14 f3
was D.Paunovic-R.Castellanos Rodriguez, Llucmajor 2016, and now I like 14
... b6 with the same dark square blockade Naiditsch achieves in the main
game.
d) 13 f4 was played in M.Askerov-K.Shevchenko, St Petersburg 2015: 13
... exf4 14 gxf4?! (14 Nxf4 Ndc5 15 Bd4 Qe7 16 Bxg7 Qxg7 looks about
equal)

Exercise: What should Black play?

Answer: 14 ... Bxc3! (rather than the game’s 14 ... fxe4) 15 bxc3 fxe4 is a
typical motif, especially once White has pushed f2-f4. It seems like sacrilege
to give up the King’s Indian bishop for a mere knight, but the white pawn
structure and centre are very badly damaged. In addition, it’s far from easy to
organize an attack on the black king, since the f4-f5 push lacks adequate
support and otherwise White’s pieces aren’t aggressively placed. This
sequence would give Black a very comfortable game in any event, but here
Black is clearly better, since 16 Bxe4 runs into 16 ... Qe7.
13 ... Nac5 14 Nxc5 Nxc5

15 Rae1
White is preparing for play in the centre. Instead:
a) 15 b3 b6 16 exf5 gxf5 17 Rab1 would transpose to B.Esen-G.Popilski,
Cesme 2015, which was another instructive GM encounter. 17 ... f4! (I really
like this move: Black should of course understand that conceding the light
squares and giving White an outpost on e4 could never work ... except that it
does!) 18 Bxc5 bxc5 19 Qe2 Qg5 20 Ne4 Qg6 21 Bf3 Bh3 and here White
chose to sacrifice the exchange with 22 Kh1? (22 Rfd1 was better, although
22 ... fxg3 23 hxg3 Rf4!? leaves White’s king looking considerably less
secure than its black counterpart; I’m also pretty unimpressed with White’s
queenside “attack”), but his compensation is insufficient after 22 ... Bxf1 23
Rxf1 a4, when the black rooks can find useful work on the a-, b-, f- and g-
files.
b) 15 Bxc5 dxc5 leads to unclear consequences. Black’s pawn structure is
slightly compromised, but he has the bishop pair, and while the bishop on g7
looks unimpressive at the moment, it can come into play in many lines. For
example, if the white queen leaves d2, then ... Bh6 would be possible, or if
White goes for the typical 16 exf5 gxf5 17 f4 to fix the structure, the bishop
would come to life after 17 ... e4.
15 ... b6 16 f4
Creating maximum tension, which Naiditsch maintains with his next
move. An alternative move order is 16 exf5 gxf5 17 f4, when Black should
respond with 17 ... e4.

Question: How would you assess this structural change?

Answer: It might seem as if White has made a mistake, giving Black a

protected passed pawn on e4 and freeing the dark-squared bishop.
Nonetheless, this combination of e4xf5 and f2-f4 is one of White’s most
common ideas in the Fianchetto Variation and serves several purposes. First,
the black pawns in the centre are no longer flexible, and White no longer
needs to calculate ... f5xe4 or ... f5-f4 on every move. Secondly, White can
aim to establish a good blockading position by rerouting his knight to e3 via
d1. Once this is done, White can consider undermining the black pawns (or
even starting a kingside attack) by pushing g3-g4. In this particular position,
Black is well placed and is certainly not worse. However, please bear in mind
the idea of e4xf5 and f2-f4. For an excellent example of White’s plan, see the
game R.Wojtaszek-A.Fedorov, Czech League 2012 (1-0 in 31).
16 ... Ba6
Naiditsch starts playing ambitiously. I’m quite impressed by how he
consistently refused to release the tension, both now and over the next several
moves.
Instead, 16 ... exf4 was tried in a classic encounter over 60 years before
the text game: 17 Bxf4 fxe4 18 Nxe4 Nxe4 19 Rxe4 Bf5 20 Bg5, G.Lisitsin-
A.Lilienthal, USSR Championship, Kiev 1954 (1-0 in 43). Here Black,
perhaps noticing that 20 ... Qd7?? loses to 21 Re7 (followed by Bh6 with a
decisive attack), panicked and gave up his queen by taking on e4. But there
was no need: 20 ... Qc8 21 Re7 (otherwise Black is comfortably equal) 21 ...
Re8 22 Rfe1 (22 Rxg7+ Kxg7 is only enough for equality; if needed, Black
can block the long diagonal and return the exchange with ... Re5) 22 ... Rxe7
23 Rxe7 Bf8 24 Re1 Bg7 and White has nothing better than repeating. If
Black manages to plays ... Qf8 and ... Re8 (which are his next two moves),
White’s queenside will look more vulnerable than Black’s.
17 b3 Qf6 18 Rc1 Rae8

Black is perfectly developed and well prepared for any opening of the
centre. If anything, White needs to be careful to maintain equality while
resolving the tension.
19 exf5!
The right capture. 19 fxe5?! Qxe5 is already a bit better for Black, since
White can’t play exf5 for obvious reasons, while 20 Bxc5 would forfeit the
dark squares.
19 ... Qxf5
Here 19 ... gxf5 is the most principled recapture, aiming to keep mobile
pawns on f5 and e5. Unfortunately, after 20 fxe5, Black probably has to
recapture with a piece (with approximate equality, although he must play
actively in order to compensate for the potentially weak f5-pawn), since 20 ...
dxe5 gives White some promising options, including the thematic 21 g4!?.
This looks bizarre but, after 21 ... f4 22 Bxc5! bxc5 23 Ne4, it’s tough to
break the light square blockade now that Black has already pushed his f-pawn
into White’s half of the board. Coupled with the weaknesses in Black’s
structure, this looks like a pleasant advantage for White.
20 fxe5 Qxe5
The pawn tension has been resolved and the black pieces are active.
21 Rxf8+
21 Rfe1 is another possibility, though again I slightly prefer Black after
21 ... Qf5. Note that there’s no repetition since 22 Rf1? is just weak after 22
... Qd3!, infiltrating White’s position.
21 ... Rxf8 22 Nb5!?
A radical decision, trying to change the character of the game. 22 Kh1 is
more conservative, but Black is at least equal since his pieces are more active
and his king is safer.
22 ... Bxb5 23 cxb5
 Exercise: How should Black proceed?

23 ... Qe8?
Naiditsch targets the b5-pawn, but this can be covered quite easily.
Answer: 23 ... Re8! was the best move, forcing White into an unpleasant
queenless middlegame with heavy pieces and opposite-coloured bishops after
24 Bxc5 (24 Bf2 Qe2 is no improvement; of course 24 Re1?? Bh6 25 Kf2
fails to several moves, most simply 25 ... Qf5+ followed by ... Nd3) 24 ...
dxc5.
It’s taken me a long time to learn how to assess this sort of position.
When I was less experienced, I used to think that such positions were simple
draws. Certainly, if the major pieces come off the board, a handshake would
follow immediately and it’s natural to suppose that Black can’t make much
progress without offering such exchanges. Then I played a tournament in
Budapest in 2007. In my first two white games, against GMs Ilincic and
Kosic, I reached inferior positions with rooks and opposite-coloured bishops.
I was lucky to score one draw from these two games.
The problem for the defender is that it’s really hard to force the trade of
rooks and, because of the opposite-coloured bishops, king safety can be an
issue even in very simplified positions (since you can’t compete on the
squares controlled by the enemy bishop). Even where there is no immediate
breakthrough, in practical terms these positions are really unpleasant to
defend, since the superior side can keep pressing and there is little to be done
to resolve the tension.
In short, while this position might appear innocuous, I think it’d be a
tough defensive slog for White.
24 Bf1 Ne4 25 Qd3 Nc5 26 Qd2 Rf5 27 Bh3 Re5 28 Bxc5 dxc5 29
Be6+
This is a different story to the version at move 23. The strong white
bishop means that the black major pieces can only become active by taking
risks.
29 ... Kh8 30 Re1?!
If a pawn is to be jettisoned, I’d prefer 30 Rf1! Qxb5 31 Rf7, when c7 is
weak and there’s no reason for White to be any worse here.
30 ... Qxb5 31 Rxe5 Bxe5
 For some of the reasons given in the lengthy note to move 23, this isn’t a
trivial draw for White, but I think his defensive chances are better with
queens than with rooks, since it is harder for Black to use his king as an
attacker without offering a trade of major pieces (when the drawing margin
dramatically increases). Naiditsch tries for a while, but Naumann defends
coolly and makes his draw.
32 Kg2 Kg7 33 Qf2 Bf6 34 Qf4 Qe2+ 35 Qf2 Qe4+ 36 Qf3 Qc2+ 37
Qf2 Qb1 38 h4 c4 39 bxc4 Qe4+ 40 Qf3 Qxc4 41 g4 Bxh4 42 g5 Qxa2+ 43
Kh3 Qf2 44 Qc3+ Kf8 45 Qa3+ Ke8 46 Qa4+ Kf8 47 Qa3+ Ke8 48 Qa4+
Kf8 ½-½

Game 2
A.Matnadze-A.Ushenina
European Women’s Championship, Mamaia 2016

1 d4 Nf6 2 Nf3 g6 3 g3 Bg7 4 Bg2 0-0 5 c4 d6 6 0-0 Nc6 7 Nc3 e5 8 dxe5

 Question: Surely White can’t hope for anything with such an
unambitious move? After all she effectively exchanges her d4-pawn for the
d6-pawn.

Answer: This simple approach contains some hidden poison and was GM
Nick Pert’s recommendation on his DVD outlining a White repertoire against
the King’s Indian and Grünfeld, based on the Fianchetto Variation. Many
theoretical sources dismiss this line as equal (as with comparable exchange
variations in the Classical King’s Indian, the Slav or the French). However, it
is precisely these lines where I think detailed preparation is required by
Black, for a number of reasons:
1. White, having an extra move, retains some initiative in exchange
variations, even when the structure is symmetrical and/or the position is
heavily simplified. This initiative needs to be carefully extinguished before it
results in something tangible.
2. Many players automatically relax with Black in symmetrical or
simplified positions, or positions which are assessed as equal, sometimes
assuming that their opponents are playing for a draw. This results in
diminished vigilance and careless play, which can get you into real trouble.
3. The character of the play in exchange variations is often quite different
from the main lines. Certainly White is running less risk here and a direct
attack on her king will be very rare. Therefore this type of position is often
unwelcome to a King’s Indian player, who is seeking more exciting and
unbalanced play.
8 ... dxe5
Both this move and 8 ... Nxe5 have their supporters. In general, trading a
pair of knights simplifies the position, which should increase the drawing
margin for both players. It also enables Black to play ... c7-c6, which is an
effective way of fighting for the important d5-square and seeking to limit the
activity of the bishop on g2. In the end I decided to concentrate on 8 ... dxe5,
since it is the choice of such experts as Naiditsch and Bacrot, whose opening
selections are worth considering carefully.
9 Bg5
The best square for the bishop, pursuing a light square strategy. If
pressed, White is normally happy to exchange this bishop for the f6-knight,
after which her c3-knight can occupy a great square on d5.
9 ... Be6
 Completing development and covering d5.
10 Qa4
This was Nick Pert’s recommendation on his DVD, although he only
considered 10 ... h6 here. Black needs to be precise over the next five to ten
moves in order to equalize.
Another logical move is 10 Nd2, which has scored well, including against
KID experts like Radjabov and Onischuk. After 10 ... Qc8 11 Nd5 Nd7 12 b4
h6 13 Be3 Re8 14 Nb3 (14 Rc1 was played in T.Banusz-T.Radjabov, Spanish
League 2016, and now I like 14 ... f5 followed by ... Nd8-f7 with counterplay
on the kingside) 14 ... Nd8 15 Qd2 Kh7 16 Rac1 f5 17 Rfd1 c6 18 b5,
A.Adly-F.Libiszewski, Francophonie Championship, Menton 2016, Black
can continue his kingside build-up with 18 ... Nf7 19 a4 e4, which also limits
the g2-bishop. After something like 20 Nf4 Nde5 21 Nxe6 Qxe6, I think
Black has a decent position.
10 ... Qc8!
This (admittedly, slightly awkward-looking) move makes the most sense
of Black’s position in my view. It has several aims:
 First (as often happens after ... Qd7 or ... Qc8), Black is ready to play ...
Bh3 with the aim of trading White’s best piece (both her most active piece,
and her king’s best defender).
Secondly, Black anticipates having to move her queen after a white rook
lands on d1, and avoids losing a tempo when this happens.
Finally, Black is preparing to manoeuvre her knight from f6 to b6. This
would significantly improve her position, since the knight on b6 would
control d5 and attack c4, while gaining time on the queen on a4. It also
unblocks the a1-h8 diagonal, meaning that the g7-bishop can become very
active (especially if Black pushes ... e5-e4). Of course the knight was pinned
to the queen while she remained on d8, so some move was essential.
We should also note that, from c8, the queen protects the b7- and c7-
pawns, which could otherwise prove vulnerable.
11 Rfd1
This move seems the most logical to me.
a) 11 Rad1 Nd7 12 Nd5 e4 13 Nd2 was played in L.Pantsulaia-
V.Onischuk, Nakhchivan 2015. Here I would opt for 13 ... Nb6 14 Qxc6
Bxd5 15 Qc5 Bc6, by analogy with the main game. White does have the extra
option of 16 b3 (as we’ll see below, 16 Nxe4 Bxe4 17 Bxe4 Re8 18 f3 Re5
19 Qe3 Nxc4 is fine for Black), but after 16 ... f5 Black has limited the g2-
bishop and should be quite satisfied with this position, as White has no real
pressure on the queenside.
b) 11 Bxf6!? Bxf6 cuts across Black’s plan in a very direct way. This has
scored well for White, so Black needs to be precise:
 b1) 12 Rad1 Nd4 13 Nd2 c6 14 e3 Nf5 15 b4 was Bu Xiangzhi-Changren
Dai, Chinese Blitz Championship, Shenzhen 2016. Here I like 15 ... a6,
holding up White’s queenside play, when Black looks quite solid.
b2) 12 Nd5 Bd8 has been seen in several games, but in none of them has
Black found the best plan, which involves ... e5-e4 and ... f7-f5 against most
moves. For instance, after 13 Rfd1, as in S.Farago-At.Schneider, Budapest
1991, rather than the solid but passive 13 ... f6, I prefer 13 ... e4! 14 Nd4 (14
Nd2 f5 is also nice for Black, who can follow up with ... Ne5) 14 ... Nxd4 15
Rxd4 f5. White’s play has missed the mark and, in addition to his more
standard ideas, Black can try to make use of the bishop on d8 with ... c7-c6
and ... Bb6 in some lines.
b3) 12 Nd2 was played in G.Froewis-P.Zelbel, Sibenik 2012. White’s
control of the light squares is impressive, and I prefer his position after the
game continuation of 12 ... Nd4 13 e3 Nf5 (1-0 in 43). Therefore I would
challenge the queen on a4 with 12 ... Bd7 13 Nde4 Bg7 14 Nc5 Be8, when
Black looks pretty comfortable. A sample line might run 15 Rad1 Kh8 16 b4
a6 17 Rd2 f5 with counterplay.
 11 ... Nd7!
As noted above, bringing the knight to b6 is one of Black’s key ideas
against this Qa4 set-up. The idea of ... Nfd7-b6 is known from the Grünfeld
(1 d4 Nf6 2 c4 g6 3 Nc3 d5 4 Nf3 Bg7 5 Qb3 dxc4 6 Qxc4 and so on),
although I’ve rarely seen it played in a King’s Indian position.
12 Nd5
Occupying the weakened d5-square is very natural.
12 ... e4

The position is becoming concrete and both sides are creating threats.
13 Nd2
Instead:
a) 13 Qxc6?? bxc6 14 Ne7+ is a typical idea in these lines, aiming to
simplify and spoil Black’s structure, but here it fails for straightforward
tactical reasons: 14 ... Kh8 15 Nxc8 exf3 and White loses material.
b) 13 Ne1 Nb6 14 Qc2 (compared with the main game, 14 Qxc6? Bxd5
15 Qc5 fails to 15 ... Bxc4) 14 ... f5 15 Rd2 was V.Akopian-F.Nijboer, Wijk
aan Zee 1993. Black proceeded logically with 15 ... Kh8 16 Rad1 Ne5 17 b3
c6 18 Nf4, but this was a little better for White because he is fully mobilized
while Black struggles to bring his queenside pieces (his a8-rook in particular)
into the game (1-0 in 38). Accordingly I would like to suggest 15 ... Rf7! 16
Rad1 Rd7 as an improvement. Black will continue with ... Qe8 and perhaps
... Qf7, when his rook on a8 is ready to join the play and simplifications on
the d-file are likely. I don’t see why Black should be worse here at all.
13 ... Nb6 14 Qxc6!
After 14 Qa3 f5, I already prefer Black’s position since her pieces (in
particular, both bishops) seem better placed than their counterparts. The text
move is therefore the only real attempt at an advantage. Black would lose
material with a horrible structure after 14 ... bxc6?? 15 Ne7+ etc, but she has
a much stronger option at her disposal.

Exercise: How should Black continue?

Answer:
14 ... Bxd5!
 Regaining the piece and transferring the bishop to c6, where it will be
well placed.
15 Qc5 Bc6 16 Nxe4 Na4
I think this is a good move but, since the arising position might not be to
everyone’s taste, I should point out that Black can essentially force a draw
with 16 ... Bxe4 17 Bxe4 Re8 (White’s bishops are exposed to tactical shots)
18 f3 Re5 19 Qe3 Nxc4 20 Qb3, when Black can repeat with 20 ... Na5 21
Qe3 Nc4 or play on after 20 ... Rxg5.
17 Qe7 Nxb2
This looks extremely dangerous, since Black has to live with a white
knight on f6, but in fact it seems like quite a good move.
Black could have won the exchange with 17 ... Bxe4 18 Qxe4 Bxb2, since
19 Rab1 can be met by 19 ... Nc3. However, after 19 Qc2 Bxa1 20 Rxa1 (not
20 Qxa4? Bg7, when Black retains her most important piece and White’s
compensation looks insufficient), Black has no advantage. For instance, 20 ...
Nc5 (20 ... Qf5 21 Qxf5 gxf5 22 Bxb7 is at least not worse for White, who
has a pawn for the exchange, a far superior structure, and a powerful bishop
pair) 21 Be7 wins back the exchange with equal play. White could well have
alternatives, since I believe that she has full positional compensation for the
exchange in any event, with the bishop pair and Black’s weak dark squares
on the kingside.
18 Nf6+ Kh8!
Not 18 ... Bxf6? 19 Qxf6 Qf5 20 Bxc6 bxc6 21 Qxf5 gxf5 22 Rd4, which
is horrible for Black, who has a horrendous structure and a poorly placed
knight on b2.
19 Rd4
Here 19 Bxc6 Nxd1 20 Bd7 Qd8 21 Qxd8 Raxd8 22 Rxd1 h6 23 Bf4
Bxf6 24 Bxc7 Ra8 25 c5 looks like a positional draw. The bishops do a great
job keeping Black’s rooks at bay for the moment, but it is hard for White to
improve her position since, if the game opens, the black rooks will come into
their own.
19 ... Qf5 20 Rh4 h5 21 f4
 We have an odd construction on the kingside where White has trapped
several of her own pieces, but Black can’t remove the defenders because a
mating attack would be threatened.
21 ... Qe6!?
I quite like this decision, reducing the risk of mate and trying to exploit
the absence of the h4-rook from the queenside.
22 Qxe6 fxe6 23 Bxc6 Bxf6
23 ... bxc6 24 Nd7 and 25 Ne5 looks more promising for White.
24 Bxf6+ Rxf6 25 Bxb7 Rd8 26 Rc1 c5
Black has decent compensation for the pawn since the h4-rook remains
out of play.
27 Bf3 Kg7 28 Rc2 Nd1 29 e4
 29 ... Ne3?
The prophylactic 29 ... Rf7! was better, when White has nothing better
than 30 Rc1 Nb2 31 Rc2 Nd1 with a draw.
30 Rc3 Nd1 31 Rc1?
White missed a late chance: 31 Ra3! Rf7 32 f5! exf5 33 exf5 Rxf5 34
Re4! and White is slightly better since she can bring her bishop to d5.
31 ... Nb2 32 Rc2 Nd1 33 Rc1 ½-½

Game 3
I.Efimov-S.Mazé
Francophonie Championship, Menton 2016

In this game I want to discuss some of the move orders White can use to
avoid the main lines of the Fianchetto Variation.
1 d4 Nf6 2 Nf3
Botvinnik tried the system 2 c4 g6 3 Nc3 Bg7 4 g3 0-0 5 Bg2 d6 6 e3 in
his 1954 match with Smyslov, and it has recently cropped up in elite GM
games, especially at faster time controls. White wants to put his knight on e2,
leading to a different set of problems for Black to solve. I think it’s a good
idea to respond consistently with 6 ... e5 7 Nge2 (taking on e5 now leaves
Black very comfortable since White has closed in his bishop with e2-e3) Nc6
8 0-0 (8 h3 exd4 9 exd4 Ne7 was similar in E.Cekro-T.Piceu, Belgian League
2006; ½-½ in 59) 8 ... exd4 9 Nxd4 (or 9 exd4 Ne7 10 Bf4 Nf5 11 Qc1 Re8
12 h3 c6 13 g4 Ne7 14 Bh6 d5 15 c5 b6 16 cxb6 axb6 and Black was fine in
B.Gelfand-A.Giri, Moscow 2016; 0-1 in 36) 9 ... Bg4 10 Nxc6 bxc6 11 Qc2
was Wang Yue-P.Svidler, FIDE Grand Prix, Astrakhan 2010, and now 11 ...
Qd7 followed by ... Bh3 would have been comfortable for Black.
2 ... g6 3 g3 Bg7 4 Bg2 0-0 5 0-0
White can also play 5 b3 d6 6 Bb2. Here too I recommend 6 ... a5 and
then:

a) 7 c4 a4 8 0-0 (8 b4 c5 9 bxc5 a3! 10 Bc3 dxc5 is fine for Black, or 9

dxc5 dxc5 10 Qxd8 Rxd8 11 b5 Nbd7 with an unclear position where Black
is not worse in P.Cramling-A.L’Ami, Stockholm 2012; 0-1 in 54) 8 ... axb3 9
axb3 (if 9 Qxb3 Nc6 10 Nbd2, as in Bu Xiangzhi-N.Kabanov, Moscow 2012,
I like 10 ... e5! 11 e3 Nd7 when Black has good counterplay) 9 ... Rxa1 10
Bxa1 c6 11 Nbd2 Na6 12 Ne1 d5 13 Nc2 Bf5 14 Ne3 Be6 was comfortable
for Black in R.Fridman-Y.Visser, Dutch League 2007 (0-1 in 70).
b) 7 a4 sees White better prepared for ... e7-e5 than in the main game, so I
suggest 7 ... Ne4!? when the trade of a pair of knights should be in Black’s
favour, since he has less space; for example, 8 Nfd2 (8 Nbd2 looks more
logical, but 8 ... Nxd2 9 Qxd2 c5 10 e3 Nc6 11 0-0 Bg4 should be okay for
Black; while 8 0-0 f5 allows Black to get a favourable version of the Dutch,
and 8 ... Nc6 and ... e7-e5 is also possible) 8 ... Nxd2 9 Qxd2 was played in
S.Drazic-K.Shanava, Mersin 2016. Here I like going into a symmetrical
Grünfeld structure with 9 ... d5 and ... c7-c6, when White is a long way from
occupying the e5-square. This should be very close to equal.
5 ... d6 6 b3
This is one of the sidelines Kramnik has been using recently to terrorize
opponents. He won an impressive game against Vocaturo in the last round of
the 2016 Baku Open where Black opted for the immediate 6 ... e5. This leads
to quite forcing lines where Black needs to be well prepared to equalize. I
think the inclusion of ... a7-a5 and a2-a4 eases Black’s defence.
Other moves:
a) 6 Nc3 is another try, preparing e2-e4. I think the simplest way to
respond to this is with 6 ... d5!, when the lost tempo is compensated for by
White’s slightly misplaced pieces. In particular, it is difficult for him to put
pressure on the centre, since c2-c4 is currently impossible and e2-e4 will lead
to exchanges.
b) 6 Re1 is a more subtle way of preparing e2-e4, since after 6 ... d5 7 c4
White can aim for a Fianchetto Grünfeld a tempo up. Black often goes into a
Fianchetto Pirc position here with 6 ... Nbd7 7 e4 e5; while this is perfectly
playable, it seems to give White the type of position he’s looking for. Instead,
I recommend the interesting 6 ... Ne4!? as essayed by Emil Sutovsky and
Nidjat Mamedov.
 The knight can be exchanged or supported by ... f7-f5. I don’t think Black
is worse here. For example: 7 Nbd2 (7 c3 c5 8 Nbd2 was played in
A.Aleksandrov-E.Sutovsky, Baku 2010, and now the simplest solution seems
to be to take on d2 and d4 and develop the other knight to c6, when the
exchanges have eased the effects of Black’s disadvantage in space and left
the g7-bishop very active) 7 ... Nxd2 8 Bxd2 (8 Qxd2 a5 9 e4 a4 10 h3 c5 11
Rb1 cxd4 12 Nxd4 Qb6 13 c3 Nc6 gave Black a very active position in
B.Takyrbashev-E.Sutovsky, Tromsø Olympiad 2014; 0-1 in 33) 8 ... c5 9 c3
Qb6 10 dxc5 dxc5 11 Qa4 Nc6 12 b3 Rd8 13 Rac1 and a draw was agreed in
A.Galkin-Nid.Mamedov, Turkish League 2011.
c) 6 a4 has also gained a following. White wants to gain space or define
the structure on the queenside, after which he can aim either for c2-c4 or Pirc
positions with e2-e4. Here I like the ambitious 6 ... Na6 (preparing ... c7-c5;
the immediate 6 ... c5 leaves Black under unpleasant pressure after 7 dxc5
dxc5 8 Nc3)
 7 Nc3 (or 7 Re1 c5 8 e4 cxd4 9 Nxd4 Bd7 10 h3 Rc8 11 Be3 Nc5 with a
comfortable Sicilian-type position in A.Stefanova-A.Beliavsky, European
Championship, Plovdiv 2012, 0-1 in 37; similarly, 7 a5 c5 8 Nc3 Bd7 9 dxc5
Nxc5 10 Nd4 Rc8 11 e4 a6 gave Black an excellent version of the Sicilian
Dragon in B.Bogosavljevic-N.Theodorou, Novi Sad 2016, 0-1 in 47; while 7
c3 c5 8 Re1 cxd4 9 cxd4 d5 10 Nc3 Nb4 11 Qb3 a5 was also very
comfortable for Black in R.Rapport-G.Jones, Helsingor 2014, ½-½ in 24) 7 ...
c5 8 e4 cxd4 9 Nxd4 Bg4 10 f3 Bd7 11 Be3 Nb4 12 Qd2 Nc6 and Black had
no problems in M.Hebden-D.Gormally, Telford 2016 (½-½ in 25).
d) 6 c4 returns to the main line, where 6 ... Nc6 was examined in Games 1
and 2. Although it’s not my recommended variation, I want to show the
following model game which I saw live, again featuring Mazé with the black
pieces: 6 ... c6 7 b3 (7 Nc3 is the critical move) 7 ... e5 8 dxe5 dxe5 9 Ba3
Qxd1 10 Rxd1 Re8 (after the game Sébastian told me that he already viewed
his position as easier to play around this point) 11 Nc3 e4 12 Nd4 a5!
(controlling the queenside dark-squares and preparing the typical ... Na6-b4)
13 h3 h5! (gaining kingside space and preventing expansion with g3-g4;
Black also gains options of playing ... h5-h4 in some positions) 14 e3 Na6 15
Na4 Nd7 16 Rac1 Nb4!

(a very comfortable square for the knight once White’s bishop is on a3,
since it attacks a2 and can’t be kicked away for the moment – of course Bxb4
would be a significant concession; still, White has done nothing wrong yet
and should be equal) 17 Rd2 Bf8 18 Rcd1 Nc5! 19 Nxc5 (many players
would have shied away from Black’s last move in view of 19 Nb6 Ra6 20
Nxc8 Rxc8 when White gains the bishop pair, but the black knights have
outstanding squares and the white bishops struggle to find useful roles;
nevertheless, this was probably the better choice since White is unlikely to be
worse in this line, while in the game Mazé quickly seizes the initiative) 19 ...
Bxc5 20 Ne2 a4! (generating play on the a-file; White’s penetration on the d-
file is easily dealt with) 21 Rd8 Kf8 22 Rxe8+? (22 Bb2 was more stubborn,
with the point that 22 ... Nxa2? 23 Bg7+ wins the exchange) 22 ... Kxe8 23
Bxb4 (now 23 Bb2 Nxa2 leaves Black with an extra pawn and a better
position) 23 ... Bxb4 24 Nd4 axb3 25 axb3 Ra2, when Black was clearly
pressing and went on to win in A.Baburin-S.Mazé, Kilkenny 2016 (0-1 in
70). Kilkenny is one of the world’s strongest weekend events, and these two
players shared first place ahead of several GMs, including Mickey Adams.
Alex Baburin is very experienced on the white side of the fianchetto King’s
Indian and the fact that he could get outplayed in such a quiet position shows
that the line can be dangerous for both sides.
6 ... a5!?

Aiming to create play on the queenside with ... a5-a4. If White plays a2-
a4 himself then both the b4- and b5-squares becoming tempting posts for,
respectively, a black and a white knight.
7 a4
Otherwise 7 Bb2 a4 8 b4 d5 9 Ne5 c6 10 a3 Nfd7 11 Nd3 Nb6 12 Nd2
Na6 13 e3 Nc7 was balanced in B.Kurajica-M.Illescas Cordoba, Seville
(rapid) 2006 (½-½ in 77).
7 ... e5 8 dxe5 dxe5 9 Ba3
Instead:
a) 9 Bb2 e4 10 Qxd8 Rxd8 11 Ng5 Bf5 12 Na3 h6 13 Nh3 Nc6 was
already excellent for Black in L.Barczay-J.Pinter, Hungarian Championship,
Budapest 1975 (1-0 in 41).
b) 9 Nxe5 is possible, but after 9 ... Ng4 White should give the exchange:
10 Nxg4 Bxa1 11 Ne3 Qxd1 12 Rxd1 Nc6 13 Nd5 Be5, when the position is
unclear and playable for both sides.
9 ... Qxd1
Here 9 ... Re8 has been played more often (see, for example,
H.Gretarsson-A.L’Ami, Reykjavik 2014; 1-0 in 35), but Sébastian’s move is
more straightforward.
10 Rxd1 Re8 11 Nc3 e4

12 Nd4
After 12 Nd2 e3!? 13 fxe3 Na6, Black has decent compensation for the
pawn. A long, sample, computer variation goes 14 Rac1 Bh6 15 Nde4 Nxe4
16 Nxe4 Bxe3+ 17 Kh1 Bxc1 18 Bxc1 Bf5! (returning the exchange
immediately is the safest approach) 19 Nf6+ Kg7 20 Nxe8+ Rxe8 21 Bb2+
(21 Bxb7 Bxc2 22 Rd2 Nb4 is level, with the bizarre repetition 23 Bd5 Bd3
24 Bf3 Bc2 25 Bd5 a distinct possibility) 21 ... Kg8 22 e4 Bc8 with
approximately balanced play.
 12 ... e3 13 f4 c6
Black is already very comfortable. Mazé goes on to demonstrate Black’s
key ideas in this structure.
14 Rac1 Na6 15 h3 Nb4
The knight is very annoying on b4 (as in Mazé’s game with Baburin
above), pressing on the c2-pawn and limiting the a3-bishop. Meanwhile,
White has no targets for his own play.
16 Kh2 h5

Both restraining g3-g4 and, in some positions, preparing ... h5-h4 himself.
17 Bxb4?
This is a huge concession, but obviously White was struggling to find
constructive ideas.
17 ... axb4 18 Na2 Bf8 19 c3 bxc3 20 Rxc3 h4!
 21 Nc2
Or 21 g4 Bd6 and the f4-pawn falls.
21 ... Nh5 22 gxh4 Bh6 23 Bf3 Bxf4+ 24 Kg2 Ra5 25 Nab4 Bh6 26 Nd3
Bg7 27 Rc4 Be6
 On an open board and with so many targets, the bishop pair is far too
strong.
28 Rb4 Bc8
Here 28 ... Ra7 was better.
29 Bg4! Bf8 30 Rd4 c5 31 Rc4 f5 32 Bxh5 gxh5 33 Nf4 b5 34 axb5
Rxb5 35 Rd3 Bb7+ 36 Kg3?
Too ambitious; White would have almost equalized after 36 Kf1!.
36 ... Be4! 37 Rd7
Now 37 Rxe3 Bd6 will collect material.
37 ... Re7
37 ... Rxb3 was even stronger.
38 Rxe7 Bxe7 39 Nxh5? Bd6+ 40 Nf4 Rxb3 41 Ne1 Rb2 42 Nf3 Rxe2
43 Rxc5 Rf2 0-1
 A striking position. With huge simplifications and nominal material
equality, White resigned because he’s losing a piece and then getting mated.

Game 4
V.Zviaginsev-V.Akopian
FIDE World Rapid Championship, Khanty-Mansiysk 2013

As discussed in the Introduction, the King’s Indian is a universal opening

which can be used against 1 d4, 1 c4 or 1 Nf3 with equal effectiveness. In
this game we look at the systems available to White in the English Opening
(i.e. without d2-d4) where Black adopts a King’s Indian set-up.
1 Nf3
After 1 c4 Black can adopt several move orders to get into the King’s
Indian. 1 ... g6 is one of the more flexible, but I recommend the
straightforward 1 ... Nf6, when White has many transpositional options. One
important line to consider is the Botvinnik System which arises after 2 Nc3
g6 3 g3 Bg7 4 Bg2 0-0 5 e4 d6 6 Nge2. This system became quite popular,
having been promoted in repertoire books by GMs Kosten and Marin
(although Kosten only recommended it after Black had committed to ... e7-
e5). In this line I suggest going for 6 ... c5 rather than ... e7-e5. It’s important
to note that White’s attempts to transpose to Accelerated Dragon or English
positions with d2-d4 don’t work very well.

a) 7 d3 Nc6 8 0-0 gives Black a choice of good systems. I prefer 8 ... Rb8
9 h3 a6 10 a4 (against neutral moves ... b7-b5 comes with excellent play on
the queenside; 10 f4 stops ... b7-b5 in view of e4-e5, but Black can play 10 ...
Nd4 here too, e.g. 11 Nxd4 cxd4 12 Nd5 Nxd5 13 cxd5 Qb6 14 Qe2 Bd7 15
e5 Rbc8 16 e6?! fxe6 17 dxe6 Bc6 and Black was already more comfortable
in O.Brendel-M.Kanep, European Rapid Championship, Wroclaw 2014, 0-1
in 38) 10 ... Nd4! (this move was not covered by Marin in his English
Opening trilogy) 11 Nxd4 (11 Rb1 Bd7 12 b4, as in D.Harika-A.Dreev,
Almaty blitz 2016, could be comfortably met by 12 ... cxb4! 13 Rxb4 Qa5 14
Ba3 Nh5 when Black’s pieces are very active) 11 ... cxd4 12 Ne2 Nd7 13 a5
b5 14 axb6 Qxb6 and Black was very comfortable in D.Flores-L.Fressinet,
World Blitz Championship, Doha 2016 (0-1 in 54).
 b) 7 d4 cxd4 8 Nxd4 has been tried as a surprise weapon, but Black is
doing well after 8 ... Nc6! since White isn’t in time to support his knight on
d4 and retreating it will leave Black with a very comfortable position. The
trappy 9 Nxc6 bxc6 10 e5 (otherwise Black has just gained control of d5 with
a half-open b-file for counterplay) can be met by 10 ... Ng4! 11 exd6 exd6 12
0-0 Ne5 and Black is simply better. Note that Bxc6 over the past couple of
moves would have failed badly to ... Qb6, forking c6 and f2.
c) 7 0-0 Nc6 8 d4? is not a good idea and leads to a tactical exercise.

Exercise: How should Black continue?

Answer: After the standard trick 8 ... cxd4 9 Nxd4 Nxe4!, Black already
has the better game.
1 ... Nf6 2 c4 g6 3 Nc3
The bizarre-looking 3 b4 is referred to as the “Speckled Hen” by some
annotators. Despite its unusual appearance, the move makes a good deal of
sense. In many lines of the King’s Indian we see White playing aggressively
on the queenside, where he has more space, and b2-b4 can be seen as
consistent with that aim. One of the Irish players chasing the IM title, FM
Stephen Jessel, plays this all the time with White and with decent results. GM
Victor Bologan, a great champion of the King’s Indian from the black side,
has recently shown a fondness for this move too, although he was
convincingly crushed by GM Ilya Smirin in a game we’ll look at now: 3 ...
Bg7 4 Bb2 0-0 5 g3 (5 e3 is a sideline of a sideline, but White can afford to
take some liberties in the opening: 5 ... d6 6 Be2 e5 7 d3 a5 8 a3 was
L.Lenic-G.Jones, Italian League 2012, where Gawain took on b4 but I prefer
the immediate 8 ... e4 9 dxe4 Nxe4 10 Qc2 Bxb2 11 Qxb2 Qf6 12 Qxf6 Nxf6
and Black has a comfortable game, as the queenside tension will force White
to give up some squares) 5 ... d6 6 Bg2 e5

a) 7 d3 a6 8 0-0 Nc6 9 a3 h6 10 Nc3 Re8 11 Nd5 Nxd5 12 cxd5 Ne7 13

Nd2 c6 14 dxc6 Nxc6 15 Rc1 d5 16 Rc5 Be6 17 Qa1 Qd6 18 Rfc1 Rad8 19
Nb3 h5 20 b5 and a draw was agreed in B.Medak-R.Markus, Croatian Team
Championship 2016. Black had a 250+ rating point advantage and normally
such a result suggests that the stronger player has offered a draw in an
inferior position, but I think Black has quite a reasonable game here, for
example, 20 ... Nb8 21 Rc7 Rd7 22 Rc8 Rde7 23 Rxe8+ Rxe8 24 Nc5 axb5
25 Nxb7 Qe7 26 Nc5 Bh6 27 e3 d4 and Black is doing fine.
b) 7 0-0 a5! 8 a3 (a useful inclusion from Black’s perspective, since he
now has the option of taking on b4 on every move) 8 ... e4 9 Ng5 (this
doesn’t work well but other knight retreats aren’t convincing either – it’s hard
to believe that Black has overextended himself by putting his pawn on e4
with tempo, rather he has seized central space and disturbed White’s co-
ordination) 9 ... Re8 10 d3 exd3 11 exd3? (this was played in V.Bologan-
I.Smirin, Poikovsky 2016; 11 Qxd3 is a sounder decision and in fact this
structure is often favourable for White – for instance, a knight on d5 could
only be driven away by ... c7-c6, weakening the d6-pawn – but here White’s
poor co-ordination means he has no chance to be better, e.g. 11 ... axb4 12
axb4 Rxa1 13 Bxa1 Nc6 14 b5 Ne5 15 Qc2 Nh5!? and Black can follow with
... h7-h6 and ... f7-f5 with active play), and now it’s time for a tactical
exercise.
 Exercise: How can Black win a pawn in this position?

Answer: 11 ... Nh5! (or 11 ... Ng4!) 12 Bxg7 Nxg7 13 Nf3 axb4 and
White loses material.

3 ... Bg7 4 g3 0-0 5 Bg2 d6 6 0-0 Nc6 7 d3

Of course White could have transposed into standard Fianchetto lines

with d2-d4, now or on any of the previous moves.
7 ... e5 8 Rb1
Intending to gain space on the queenside with b4-b5. 8 Bg5 tends to
transpose after 8 ... h6 9 Bxf6 Bxf6 10 Rb1 a5 11 a3 Re8.
8 ... a5
By inserting this move, Black aims to reduce the number of queenside
pawns (so that the a7-pawn won’t be a weakness). It’s tempting to point also
to his control of the a-file in the event of a2-a3 and b2-b4, but in reality
White is more likely to seize control of this file.
9 a3
 Of course White should be consistent and continue his plan to play b4-b5.
9 ... Re8

A useful, flexible move with a tactical point.

10 Bg5
This is by far the most common continuation. It might look
counterintuitive, giving up the bishop pair, but the move’s popularity can be
explained by reference to two factors:
1) In general, the English can be characterized as a “light square”
opening, where White competes for control of the light squares (especially
d5) with moves like c2-c4, Nc3 and Bg2. Trading the f6-knight for the dark-
squared bishop is fully consistent with these objectives.
2) More concretely, White gains time for his operations on the queenside
since, as we will see, the immediate 10 b4 runs into a tactical sequence which
leaves Black very comfortable. If the position is thought of as a race (where
White attacks on the queenside, and Black seeks counterplay in the centre
and, perhaps, on the kingside), we can see how valuable gaining one or two
tempi might be.
 Instead:
a) 10 b4 allows Black to show one of the ideas behind 9 ... Re8, playing
10 ... axb4 11 axb4 e4! 12 Nxe4 Nxe4 13 dxe4 Rxe4, which is well known to
be comfortable for Black. Mamedyarov has more than held his own here
against elite opposition: 14 Nd2 (or 14 b5 Nd4 15 Nxd4 Rxd4 16 Qb3 Be6 17
Bd5 Bxd5 18 cxd5 Re4 19 Be3 b6, V.Topalov-S.Mamedyarov, World Blitz
Championship, Astana 2012; ½-½ in 37) 14 ... Re8 15 Bb2 Bxb2 16 Rxb2
Qf6 17 Rb1 (or 17 Rb3 Bf5 18 b5 Nd4) 17 ... Bf5 18 e4 (18 b5 Nd4 19 e4
Be6 is also excellent for Black) 18 ... Be6 19 Qc2 Ra3 and Black had a huge
initiative in P.Svidler-S.Mamedyarov, Almaty (blitz) 2016 (0-1 in 30).
b) 10 Nd2 Nd4 has been championed by Alexei Shirov: 11 b4 (or 11 e3
Ne6 12 b4 axb4 13 axb4 c6 14 Qb3 h5 15 h4 Ng4 16 Bb2 f5 and Black was
no worse in V.Artemiev-A.Shirov, Moscow rapid 2014; ½-½ in 43) 11 ...
axb4 12 axb4 c6 13 Nde4 (13 e3 was more prudent, with a likely
transposition to Artemiev-Shirov) 13 ... Nxe4 14 dxe4 Be6!? (14 ... h5 was
also logical) 15 Qd3 (15 e3 Bxc4 16 exd4 Bxf1 17 Qxf1 exd4 18 Ne2 leads
to a position with mutual chances: Black has a rook and two pawns for two
pieces, which is at least enough material, though White could develop strong
threats if he manages to activate his bishop pair) 15 ... Qd7 16 Rd1 (16 e3 is
also well met by 16 ... b5!) 16 ... b5 17 cxb5 cxb5 18 Nd5 (here 18 e3 Bc4 19
Qd2 Nb3 20 Qxd6 Qxd6 21 Rxd6 Bf8 22 Rd1 Nxc1 23 Rdxc1 Bd3 is equal:
Black’s bishop pair fully compensates for the extra doubled pawn) 18 ...
Bxd5 19 exd5 Qg4 20 e4? Ne2+ 21 Kh1 Nc3! and Black won material in
M.Klekowski-A.Shirov, Czech League 2016 (0-1 in 36).
10 ... h6
“Putting the question” to the bishop (or, as GM Victor Bologan prefers,
“checking the documents”). It makes sense to force White to clarify his
intentions, in particular before Nd5 lands (which would make it much more
difficult to break the pin).
11 Bxf6 Bxf6
 12 b4
Again, by far the most popular move; White gets on with his queenside
play. Others:
a) 12 e3 Bg7 13 b4 axb4 14 axb4 Ne7 15 b5 Be6 16 Nd2 c6 17 bxc6 bxc6
18 Ra1 Rxa1 19 Qxa1 d5 20 cxd5 cxd5 21 Qa4 Qa8! was model play from
Black in A.Timofeev-Rau.Mamedov, European Cup, Bilbao 2014 (½-½ in
41). White already needs to play precisely to draw.
b) 12 Nd2 Bg7 13 e3 (13 b4 axb4 14 axb4 e4 15 Nd5 exd3 16 exd3 Nd4
was excellent for Black at this point in A.Zubarev-V.Onischuk, Lutsk rapid
2016; 1-0 in 71) 13 ... f5 14 b4 axb4 15 axb4 e4 16 d4 Ne7 17 Ra1 Be6 was
level in B.Heberla-E.Gleizerov, Jurmala (blitz) 2015 (½-½ in 54).
c) 12 Nd5 was tried in a Serbian League match. GM Atalik laconically
observed that “12 b4 is more played but I guess not on the days with 3-3.”
Indeed, 12 ... Bg7 13 b4 was already agreed drawn in S.Atalik-I.Miladinovic,
Kragujevac 2016. Play might continue 13 ... axb4 14 axb4 Ne7 15 Nc3 Be6,
aiming for ... d6-d5 (with or without ... c7-c6).
12 ... axb4 13 axb4 Bg7 14 b5 Ne7 15 Qc2
 15 Ne1 certainly isn’t an improvement, since the knight’s placement
breaks the connection between White’s rooks. After 15 ... c6 16 Qb3 Be6 17
Nf3 d5, Black was more than comfortable in M.Bezold-V.Akopian, World
Blitz Championship, Berlin 2015 (0-1 in 31).
15 ... c6
This creates a potential target for White but is consistent with Black’s aim
of playing in the centre with ... Be6 and ... d6-d5.
16 Nd2 Be6 17 bxc6 bxc6

18 Rb7
White doesn’t get anywhere with 18 Ra1 d5 19 Rxa8 Qxa8 20 Rb1 Nf5
21 cxd5 cxd5 22 Qa4? (allowing a thematic shot, although I already prefer
Black’s position) 22 ... e4! (the c3-knight is hanging) 23 Qxa8 (23 dxe4 Bxc3
24 Qxa8 Rxa8 25 exf5 Bxf5 transposes) 23 ... Rxa8 24 dxe4 Bxc3 25 exf5
Bxf5 26 Rd1 Ra2 27 Nf3 d4 (after 27 ... Rxe2 28 Rxd5, Black can continue
pressing for a long time, but there are always drawing chances with reduced
material and all the pawns on one side of the board; I prefer Mamedov’s
treatment) 28 Kf1 Kg7 and it requires precision and patience for White to
hold this position; Black went on to win in A.Timofeev-Rau.Mamedov,
Khanty-Mansiysk 2014 (0-1 in 45).
18 ... Bc8
As we will see, the game quickly ended in a repetition after this move.
Black has tested two more ambitious alternatives in practice.
a) 18 ... Rb8 19 Rxb8 Qxb8 20 Rb1 Qa7 was played in A.Timofeev-
V.Onischuk, Abu Dhabi 2015 (½-½ in 49). I prefer White slightly here, since
his structure is better. The c6-pawn is weaker than the e2-pawn, and pushing
... d6-d5 or ... c6-c5 is likely to create weaknesses. If Black is to play
aggressively and seek to establish a central space advantage, I’d prefer to
keep more pieces on the board in order to profit from the extra territory.
Therefore ...
b) 18 ... d5!? is an ambitious and logical move, seizing central territory.
The rook on b7 looks threatening, and can of course be dangerous, but might
also constitute a tactical liability; for example, Black might gain a tempo on
the rook with ... Qc8, in order to prepare a favourable exchange of light-
squared bishops with ... Bh3. In response 19 Rfb1!? is always played,
bringing in the last piece and dominating the b-file by doubling rooks.
 Now GM opinion is divided on how to handle Black’s position.
b1) 19 ... e4 is extremely ambitious. Since the English is primarily aimed
at central light square control, if Black could comfortably achieve ... d6-d5
and ... e5-e4 on consecutive moves this would basically refute the opening!
Indeed, I haven’t been able to make the tactics work for White here. After 20
cxd5 cxd5 21 Nb5 Rc8 (Black argues that the empty a- and b-files are
irrelevant to the position, and he might be right!) 22 Rc7 (maybe 22 Qa4 is
the toughest defence, though after 22 ... exd3 23 exd3 Nc6 24 Nf3 Bf5 25 Bf1
Re7 26 Rxe7 Qxe7, Black was pressing in U.Andersson-V.Akopian, Ubeda
1997; ½-½ in 33) 22 ... exd3 23 exd3 Qd7!! (not too hard for a player of
Shirov’s calibre, but a very elegant way to complete development) 24 Nf3
Rxc7 25 Qxc7 Rc8 26 Qxd7 Bxd7, even the Swedish GM’s legendary
technique was insufficient to hold this position in U.Andersson-A.Shirov,
Monte Carlo (rapid) 1997 (0-1 in 46).
Perhaps White should try 20 dxe4 dxc4 21 e5!? (slow play would leave
White facing a considerable risk, since the doubled c-pawns are dangerous,
passed and supported by an active bishop pair) 21 ... Bxe5 22 Nce4 c3! (a
typical desperado motif – the pawn was dropping anyway, so Black disturbs
White’s co-ordination by forcing the recapture on c3 instead of c4) 23 Nxc3,
though I don’t think Black is worse at all in this interesting position. His c-
pawn is isolated, but is also passed. White’s rooks are impressive, but the
black bishops are awesome. If I had to choose, I’d take Black here.
b2) If that wasn’t enough, Black opted successfully for 19 ... Nf5 in
J.Cox-M.Hebden, British Championship, Southport 1983 (½-½ in 39) and
K.Kuzmicz-K.Miton, Polish Team Championship 2012 (½-½ in 34), which I
simply note as a good alternative for further investigation.
19 Rb3 Be6 20 Rb7
20 Rfb1 d5 21 Rb7 transposes the note with 18 ... d5!? above. This isn’t
forced, but I don’t know what other ideas White has in this position.
20 ... Bc8 21 Rb3 Be6 22 Rb7 ½-½
A reasonable end to a logically-played game, but certainly Black is the moral
victor. If a draw isn’t enough for you, I’ve pointed out some ambitious
alternatives to test at move 18.
Chapter Two
The Classical Variation
Introduction
1 d4 Nf6 2 c4 g6 3 Nc3 Bg7 4 e4 d6 5 Nf3 0-0 6 Be2

This introduces the Classical systems of the King’s Indian. White

continues his development in the most straightforward manner, aiming to
castle short and exploit his space advantage.
The first decision Black has to make is whether to challenge the centre
with ... c7-c5 (either immediately or after ... Bg4 and, possibly, ... Nfd7) or ...
e7-e5. Here 6 ... e5 is the traditional main line and is also our choice. It is also
by far the most popular move, especially if one includes games with the
move orders 6 ... Na6 7 0-0 e5 and 6 ... Nbd7 7 0-0 e5, which are directed
against the Exchange Variation (6 ... e5 7 dxe5). However, these move orders
won’t be available to us, since in the main line after 6 ... e5 7 0-0, I’m
recommending the most aggressive approach with 7 ... Nc6.
6 ... e5 7 0-0
This is the Mar del Plata Variation. White has some significant
alternatives:
a) 7 dxe5 is the Exchange Variation, which isn’t any kind of theoretical
problem for Black but can lead to quieter positions than he may be used to.
It’s worth being well prepared here, both theoretically and psychologically,
ready for a long game.
b) 7 d5 is the Petrosian Variation, which has never been the most popular
but has often been used by very strong players as a surprise weapon. It’s
possible that this will become more popular since its endorsement in recent
repertoire works by Cummings and Delchev.
c) 7 Be3 is the Gligoric Variation, which is another serious line.
We will look at these possibilities in Games 13-15.
7 ... Nc6
This is the main line and the favourite of both Fischer and Kasparov.
Black actually has a wide choice here, which suggests a heavy workload for
players wishing to include 7 0-0 in their repertoires.
For instance, 7 ... Nbd7 is an older line which is used with success by
strong GMs like Gawain Jones and Zdenko Kozul; while 7 ... Na6 is a more
flexible approach, which was developed by GM Igor Glek and has remained
reasonably popular and playable.
Black can also change the structure with 7 ... exd4, which was
endorsed by GM Dejan Bojkov in a repertoire book a few years ago. In
purely structural terms, this is a concession, since White is left with a
significant central space advantage and it becomes more committal for
Black to play aggressively on the kingside in view of the open centre.
That said, the King’s Indian bishop comes alive and Black gains some
definite possibilities for counterplay.
8 d5
Gaining space with tempo. One of the disadvantages of 7 ... Nbd7 and 7
... Na6 (relative to 7 ... Nc6) is that they put less pressure on the centre and so
do not force White to resolve the tension. Here, however, the pressure on d4
is quite real.
Other moves are less critical:
a) 8 dxe5 is a delayed Exchange Variation, which has gained some
traction amongst GMs looking to play a quiet game for two results. This is
covered in the notes to Game 15.
b) 8 Be3 is also worth knowing about, although it isn’t a serious attempt
at an advantage (see the notes to Game 13).
8 ... Ne7

Question: White has closed the centre. What changes does this make to
both sides’ plans?

Answer: With a closed centre, play typically migrates to the flanks. I

think there are two reasons for this. Firstly (and most simply), there are no
central pawn breaks available anymore, and putting the rooks on the d- and e-
files isn’t effective for either side. Secondly, play on the flank cannot be
effectively met by a strike in the centre. With his pawn still on d4, White
would need to think twice before pushing his a-, b- and c-pawns, since the
g7-bishop would exert unpleasant pressure on the queenside, and Black could
seek to arrange central counterplay, possibly with ... c7-c6 and ... d6-d5.
Similarly, with White’s pawn still on d4, if Black pushed his f-, g- and h-
pawns up the board, he would risk being with left with weaknesses or being
overrun in the centre (with d4xe5, or perhaps c4-c5 to soften up the central
structure) and would expose his king; for example, unpleasant threats might
be created on the a2-g8 or a1-h8 diagonals.
In the current structure, it’s very attractive for White to play on the
queenside and, in particular, to achieve the c4-c5 break. He has far more
space there (in particular, his c- and d-pawns are more advanced than their
counterparts), while Black has few defenders but many potential targets. For
example, if White penetrates to c7 with a rook or a knight, Black’s position
could become critical.
Black therefore needs to find counterplay, and the most reliable way of
doing so is to aggressively play on the kingside (and against the white king)
with ... f7-f5 followed, in some cases, by a massive kingside pawn storm.
This is one of the few openings where this is a good plan for Black, and one
of the only openings after 1 d4 where White can get checkmated without
making an obvious blunder. Therein lies the attraction of the King’s Indian
from Black’s perspective. Nevertheless, White’s queenside plan is extremely
dangerous and Black needs to play precisely here.
White has three main options at move nine. Two are directed against ...
Nh5, while rerouting the knight to the queenside (9 Ne1 and 9 Nd2). The
third, the so-called “Bayonet Attack” with 9 b4, is a more modern approach
which was popularized by Kramnik and, according to legend, is the reason
Kasparov stopped playing the King’s Indian towards the end of his career.

Game 5
P.Maghsoodloo-M.Tabatabaei
World U-16 Championship, Khanty-Mansiysk 2016

1 d4 Nf6 2 c4 g6 3 Nc3 Bg7 4 e4 d6 5 Nf3 0-0 6 Be2 e5 7 0-0 Nc6 8 d5 Ne7

9 Ne1
 Question: What’s the idea of this move?

Answer: As we discussed in the introduction, White aims to play c4-c5

and Black aims to play ... f7-f5. 9 Ne1 helps White in both respects, by
preparing Nd3 (to support c4-c5) and preventing ... Nh5 (which would
prepare ... f7-f5 in its most aggressive version). White also gains more
flexibility with his kingside structure: in particular, he can play f2-f3
(supporting his e4-pawn, and creating a square on f2 for a bishop or a knight)
and even, in some variations, g2-g4 (an apparently outlandish move, but one
which aims to avoid the typical suffocating KID attack by giving more space
to the white king and his defenders).
The alternative retreat, 9 Nd2, is examined in Game 12. The line I’ve
chosen against 9 b4 generally transposes, so is covered there as well.
9 ... Nd7

Question: Again, what’s the idea?

 Answer: Black wants to play ... f7-f5 and needs to get the knight out of
the way. The h5-square has been ruled out by White’s last move. In addition,
from d7 the knight fights against the c4-c5 break, which is why I prefer it to
the alternative 9 ... Ne8 (which is also a very serious move, played in
hundreds of games, including by guys like Kasparov, Topalov and Smirin).
10 Be3

Question: Surely this is asking for it? Black aims for a kingside pawn
storm and White invites it with tempo?

Answer: 10 Be3 does indeed lose time to the forthcoming ... f5-f4, and so
this line is the most critical in the whole King’s Indian. It’s also the line
where White is most likely to get checkmated by move 30. However, the
bishop (which will soon drop back to f2) has both a defensive function
(covering important squares on g3 and h4, although again this can encourage
a ... g4-g3 sacrifice by Black) and a very important attacking function
(aiming at key queenside dark squares and, in particular, a7). Black is now
completely lost on the queenside, since White’s firepower on that part of the
board is overwhelming. Accordingly, he needs to put all his energy into
prosecuting his kingside attack.
White sometimes prefers to do without this move: 10 Nd3 f5 11 f3 f4 is
seen in Game 10, while the disruptive 10 f3 f5 11 g4 is the subject of Game
11.
10 ... f5 11 f3
White effectively committed to this with his last move, as otherwise the
bishop would remain vulnerable to an ... f5-f4 push without any good retreat
squares.
11 ... f4 12 Bf2 g5
 The position has closed even further. Compared with that after move 9, it
looks like Black has made more progress than White. However, c4-c5 is now
coming quickly and, as mentioned above, the bishop on f2 puts unbearable
pressure on Black’s queenside.
13 Rc1
One of White’s most logical and dangerous moves. He develops a piece,
supporting his key break. Moreover, if he achieves c4-c5, this rook will be
perfectly placed, supporting a Nb5-c7 invasion, for instance.
The main alternatives, 13 Nd3 and 13 a4, are examined in Games 7-9,
while 13 g4 features in the notes to Game 11.
13 ... Ng6
Black continues with his standard regrouping. Note that this move is
more logical than 13 ... Nf6, after which White achieves 14 c5 for free.
Black has tested 13 ... Rf7 as well, but this gives White the unpleasant
option of 14 c5!? Nxc5 15 Bxc5 dxc5 16 Bc4. In W.So-F.Perez Ponsa, Baku
Olympiad 2016, Black was already much worse after 16 ... Kf8?! 17 d6 cxd6
18 Bxf7 Kxf7 19 Nb5 (1-0 in 44).
 14 c5!?
An enterprising pawn sacrifice, aiming to accelerate White’s play on the
queenside. We will look at 14 Nb5 in the next game.
If White continues to prepare the c4-c5 advance with 14 b4, he risks
falling under a strong attack. The following game is an excellent example: 14
... Nf6 15 c5 Rf7 16 cxd6 (16 a4 Bf8 17 cxd6 Bxd6!? led to success for Black
in the classic game A.Miles-H.Ree, Amsterdam 1982; 0-1 in 43 – this is an
extra option for Black in many lines where White delays taking on d6) 16 ...
cxd6 17 Nb5

Black has two good moves here:

a) 17 ... a6 18 Nc3 (18 Na3 b5! keeps the knight passive and is a typical
reaction to Nb5 without a2-a4, as shown in the wonderful game between So
and Nakamura which we’ll examine later) 18 ... Bd7 is a decent plan, when
White has spent two tempi to provoke ... a7-a6 and has no clear breakthrough
on the queenside.
b) 17 ... g4! is my preference; Black needs this move to get his kingside
attack going and it’s a very typical response to Nb5, since f3xg4 can now be
met by ... Nxe4. R.Knechtel-R.Schmidt, Wunsiedel 2014, continued 18 Qc2
(18 Nxa7? g3! wins material since defending the knight with 19 hxg3? fxg3
20 Be3? runs into 20 ... Nxd5!, when the queen is headed for h4, while 18
Bxa7? allows 18 ... g3 with a huge attack for the pawn)

Exercise: How should Black continue his kingside attack?

Answer: 18 ... g3! (a thematic sacrifice which works perfectly; you’ll see
plenty of examples of this idea over the coming games) 19 hxg3 fxg3 20
Bxg3 (20 Bxa7 Nh5 and ... Qh4 will be quickly decisive) 20 ... Bh6 21 Rb1
Nh5 22 Bf2 Ngf4 (Black’s pieces flow effortlessly to great squares) 23 g4
Nh3+ 24 Kh1 Bxg4! (a great shot, exploiting the bishop on f2; White doesn’t
respond optimally, but his game is hopeless in any event) 25 Nxd6 Qxd6 26
Bc5 Qf6 27 fxg4 Ng3+ 28 Kh2 Nxf1+ and Black soon won (0-1 in 38).
14 ... Nxc5!
Eliminating a dangerous attacker. 14 ... Nf6? 15 cxd6 cxd6 16 Nb5 gives
White far too strong an initiative on the queenside, and at no cost.
 15 b4 Na6

One of the points of the 14 c5 sacrifice is that this awkward-looking

retreat is forced, since the queenside would collapse after 15 ... Nd7? 16 Nb5.
However, the knight on a6 sometimes finds useful work in the King’s Indian,
and here it at least attacks b4 and defends c7.
16 Nd3
After best play, 16 Nd3 and 16 Nb5 will generally transpose. For
instance: 16 Nb5 Bd7 17 Nxa7 h5 18 a4 Rf7 19 Nb5 (or 19 Nd3 Bf8) 19 ...
Bf8 and White has nothing better than 20 Nd3, transposing to the main game.
In this line 18 Bxa6 is an independent option, but one which Black should
be happy to see. After 18 ... bxa6 19 Qc2 Rf7 20 Qxc7 Qf8, Black follows
with ... g5-g4 with very dangerous compensation. This position has been
debated in several correspondence games, where Black has done okay; over
the board, I would certainly take Black. Let’s follow one game to see some
typical ideas: 21 Qb6 (the only OTB game in this line in my database
continued 21 Qb7? Bb5 and Black won material in S.Kesten-J.Jorczik,
German Championship, Bad Wörishofen 2008; 0-1 in 53) 21 ... g4 22 Qxa6
g3 23 Bb6 (if 23 hxg3 fxg3 24 Bxg3 then 24 ... Be8!, preventing Rc7,
followed by ... Bh6 and ... Rg7 gives Black typical and excellent
compensation on the kingside dark squares) 23 ... gxh2+ 24 Kxh2 Bf6 25 Bc7
Bh4 26 Bxd6 was F.Cottegnie-R.Amico, correspondence 2011. Here I like 26
... Qg7! followed by ... Bg3+ and ... Nh4 (or ... h5-h4), when Black has a
strong attack.
White often leaves the h-pawn as a shield, and 24 Kh1 was worth
considering. I’d be minded to continue blasting open the kingside with 24 ...
h4, when a sample variation runs 25 Bc7 h3 26 b5 Bxb5 (Black doesn’t want
to give up his light-squared bishop, but otherwise the b-pawn would be too
dangerous) 27 Qxb5 Rxa7 28 Qb6 Rxa2 29 Bxd6 hxg2+ 30 Nxg2 Qa8. The
engines evaluate this position as clearly better for White, but I think they
continue to pay too little attention to the vulnerability of the white king. It
seems to me that all three results are possible.
16 ... Rf7 17 Nb5 Bf8!
I prefer giving the pawn back to complete development rather than
accepting light square weaknesses on the queenside with 17 ... b6, as in
W.So-Ding Liren, Bilbao 2015 (1-0 in 60). This sequence of ... Rf7 and ...
Bf8 is a very typical regrouping in the King’s Indian. It helps defend the
queenside (covering the weak c7-pawn) and, as we will see, support’s
Black’s kingside aspirations.
18 Nxa7 Bd7!
To an experienced King’s Indian player this move is obvious, but it’s
based on a subtle idea. Black’s light-squared bishop is his most important
attacking piece, for reasons which will be discussed in the next game. 18 ...
Rxa7? 19 Bxa7 b6 might look tempting but this rarely, if ever, works since it
is difficult for Black to catch the bishop on a7 – and even if he did, White
would build up too much queenside pressure in the interim with Qa4 and
doubling rooks on the c-file.
19 a4
Black was threatening 19 ... Qb8, catching the knight on a7.
 Exercise: What pawn break should Black be trying to achieve, and how
can he regroup his pieces to support it?

Answer: Obviously ... g5-g4 is what Black needs, and the best way to
prepare it is:
19 ... h5! 20 Nb5 Rg7!
Giving the rook access to g7 or h7 is one of the aims of the ... Bf8 retreat.
21 Be1
A typical regrouping, which prepares Nf2, aimed against Black’s main
idea of ... g5-g4.
Alternatively, 21 Kh1 is a logical move, getting the king off the g-file
before it opens: 21 ... Nh4 22 g3 (22 Rg1 g4 23 fxg4 hxg4 24 g3 f3 25 gxh4
fxe2 26 Qxe2 Be7 has been tested in over a dozen correspondence games;
Black is objectively fine and, in a practical game, has outstanding play) 22 ...
fxg3 23 Bxg3 Ng6 (GM Van Kampen assesses this position as unclear in his
chess24 lectures) 24 Rg1 was A.Yermolinsky-J.Becerra Rivero, Chicago
2008, and now I like 24 ... Be7, followed by ... Bf6 with unclear play.

Exercise: How can further Black support ... g5-g4 - ?

Answer:
21 ... Nh8!
An outstanding manoeuvre! The knight clears the g-file for the rook and
prepares to support ... g5-g4 itself via ... Nf7-h6. Once the g-pawn lands
there’s always counterplay for Black.
22 Nf2 Nf7 23 h3
I’m not sure about this move. White makes ... g5-g4 harder to achieve in
the short term but, once it lands, it will come with greater effect.
23 ... Nh6 24 Rc4 Be7 25 Rc3
Another great example of Black’s play in such positions continued 25
Qc2 Rc8 26 Na7 Rb8 27 Nb5 Rc8 28 Na7 Ra8 29 Nb5 Bf6! (declining a
draw by repetition) 30 Bd2 g4 31 fxg4 hxg4 32 hxg4 Bh4 33 Nxc7 Nxc7 34
Rxc7 Bxa4 35 Rxg7+ Kxg7 36 Qb2 Rc8 (White’s extra pawn isn’t felt and I
prefer Black’s chances; objectively 36 ... Kg8 might have been even stronger,
since the move in the game gives White a chance) 37 Rc1??

Exercise: How can Black win the game?

Answer: 37 ... Bxf2+ 38 Kxf2 Qb6+ 39 Kf1 Rh8! (the key move: the
threatened ... Nxg4 can’t be prevented) 40 Be1 and White resigned in
H.Molvig-S.Bekker Jensen, Borup 2012.
Going back, 37 Bxf4!! was the best chance, but required White to see a
very subtle move deep in the variation: 37 ... Rc2 38 Bxh6+ Kxh6 39 Qa3
Rxe2 40 g3! (40 Qxa4? Bxf2+ 41 Rxf2 Qb6 leads to a rapid mate) 40 ... Bb5
41 gxh4 Qxh4 and White should hold this position (of course he has no
winning chances, despite his extra pawn, since the black pieces are far too
active).
25 ... Qf8
Black could also play 25 ... g4 immediately, with similar play to the
previous note.
 26 a5
Threatening to capture on c7; Black is more than fine after 26 Nxc7 Nxc7
27 Rxc7 Bxa4.
26 ... c6 27 dxc6 bxc6 28 Na3 Nxb4
Black has now won a pawn, but the position remains dynamically
balanced.
29 Nc4 Be6 30 Rc1 c5 31 Bxb4 cxb4 32 a6?!

32 ... Rxa6?!
Here the thematic 32 ... g4! would have been very strong: 33 fxg4 hxg4
34 Nxg4 (34 hxg4 Rxa6 35 Nxe5 dxe5 36 Bxa6 Nxg4 gives Black a much
better version of the game) 34 ... Nxg4 35 Bxg4 (35 hxg4 d5! followed by ...
Bc5+ is terminal) 35 ... Bxg4 36 hxg4 Rxa6 and White has an uphill struggle
to make a draw, facing an extra pawn and having an exposed king.
33 Nxe5 dxe5?
Moving the rook would have left the position unclear. Now White gains
an advantage (for the first time!), though the game was drawn after further
adventures. For our purposes, I think we’ve seen enough.
 34 Bxa6 g4 35 hxg4 hxg4 36 Rc6 Qf7 37 fxg4 Nxg4 38 Bc4 Bxc4 39
Nxg4 Bg5 40 Rc8+ Kh7 41 Nxe5 Qa7+ 42 Rf2 Bf7 43 Nxf7 Qxf7 44 Rc6
Bh6 45 e5 Qf5 46 Qc2 Qxc2 47 Rcxc2 Rb7 48 Rb2 Kg6 49 Rfc2 Kf5 50
Rc4 Bf8 51 Kf2 Kxe5 52 Kf3 Bd6 53 Rxf4 b3 54 Re4+ Kd5 55 Ra4 Be5 56
Rb1 b2 57 Ra5+ Ke6 58 Ke4 Rb4+ 59 Kd3 Rb3+ 60 Ke4 Rb4+ 61 Kd3
Rb3+ 62 Ke2 Bf4 63 Rc5 Bc1 64 Rcxc1 bxc1Q 65 Rxc1 Kf7 66 Kf2 Rb8
67 Rc3 Kg7 68 g4 Rb4 69 Kg3 Ra4 70 Kh4 Ra5 71 Rc6 Rd5 72 Re6 Ra5
73 Re4 Rb5 74 Re1 Ra5 75 Rf1 Rb5 76 Kg3 Rb3+ 77 Kf4 Rb5 78 Re1 Kf6
79 Re4 Ra5 80 Rb4 Rc5 81 Rb6+ Kg7 82 Re6 Kf7 83 Re5 Rxe5 84 Kxe5
Kg6 85 Kf4 Kf6 86 g5+ Kf7 87 Kf5 Kg7 88 g6 Kg8 89 Kg4 Kh8 90 Kh4
Kg8 91 Kg5 Kg7 92 Kf5 Kg8 93 Kf6 Kf8 94 g7+ Kg8 95 Kg6 ½-½

Game 6
R.Kasimdzhanov-H.Nakamura
FIDE Grand Prix, London 2012

1 d4 Nf6 2 c4 g6 3 Nc3 Bg7 4 e4 d6 5 Nf3 0-0 6 Be2 e5 7 0-0 Nc6 8 d5 Ne7

9 Ne1 Nd7 10 Be3 f5 11 f3 f4 12 Bf2 g5 13 Rc1 Ng6 14 Nb5
 Perhaps the most natural move in the position. Rather than giving up a
pawn with 14 c5, White supports the c4-c5 break and creates the immediate
threat of Nxa7 (which is unfortunate) followed by Nxc8 (which is absolutely
fatal). Although we will touch on this theme several times in this book, now
is a good time for a pop quiz:

Question: Which is Black’s most important piece in the King’s Indian, in

particular when he launches a kingside pawn storm?

Answer: Somewhat counterintuitively, the answer is the c8-bishop. This

piece (which is undeveloped and currently has no legal moves) has two vital
functions:
The first is supporting the ... g5-g4 break, without which Black doesn’t
have any real attack.
The second looks like a tactical fluke, but is actually a key strategic
theme: once Black plays ... g4-g3, if this is met by h2-h3, then ... Bxh3!! is
the key sacrifice without which the King’s Indian, in its modern form, would
simply not be playable, since the white king would sit safely behind a
completely blocked pawn chain. Once ... Bxh3 lands, in very many positions
White is completely lost. You’ll need to see a few examples of this theme
(and there are several in this book) to believe it, but for now just take my
word for it – in the King’s Indian generally, and the Mar del Plata line in
particular, Black cannot allow his light-squared bishop to be exchanged.

Exercise: White has an immediate threat of Nxa7. How should it be

parried?

Answer:
14 ... b6!
An unusual reaction, but by far the strongest way to meet the queenside
threats. The routine 14 ... a6 is well met by 15 Na7! – Black can’t allow the
exchange of his light-squared bishop, and trying to catch the minor pieces
with 15 ... Rxa7 16 Bxa7 b6 fails to 17 b4 followed by c4-c5. Black can try to
attack on the kingside while White extracts his dark-squared bishop, but I
don’t think this can compensate for a full exchange.
15 b4
Threatening c4-c5 again.
15 ... a6!
 Gaining time against the knight, which must reroute to enable c4-c5.
Normally in these variations with Ne1 and Be3, Black avoids touching his
queenside pawns (which make it easier for White to open files), but here
these moves are justified concretely (there are no good alternatives) and by
the fact that White must lose a lot of time.
16 Nc3
16 Na3 is well met by 16 ... a5! 17 c5 axb4 18 cxd6! cxd6 (after 18 ...
bxa3 19 dxc7, the connected passed pawns give White excellent
compensation for the piece) 19 Nb5 Nc5 and I don’t think Black can be
worse. The game might continue 20 Nd3 g4 21 Nxb4 Bd7 22 Nc6 Qf6 and I
would prefer Black’s position.
16 ... Rf7
Starting the standard regrouping. The rook defends the c7-pawn, allows
the bishop to come to f8, and prepares to attack the white king from g7.
17 Nd3
White can also start with 17 c5. To play consistently, I would suggest 17
... bxc5 18 bxc5 Bf8, when 19 Nd3 h5 transposes to the game, and if White
takes on d6 Black can recapture with the bishop.
17 ... h5 18 c5 bxc5
Nakamura deviates from a previous win against a top grandmaster, where
he played 18 ... Bf8, A.Giri-H.Nakamura, Reggio Emilia 2011/12 (0-1 in 59).
I think that move is fine too, but in the game we get to see a nice concept
which can be applied when White delays c5xd6.
19 bxc5 Bf8 20 cxd6
Pushing on with 20 c6 Nf6 helps Black more than White, since although
the queenside isn’t completely blocked, closing the c-file makes a
breakthrough much harder. The engines assess this position as only slightly
better for White which, given their general evaluation of the Classical
Variation, suggests that Black is clearly better.
20 ... Bxd6!?

Here it is! Black retains control of b6, defends e5, and puts his (normally
passive) bishop to good use.
21 Na4 Nf6 22 Ndc5
Here 22 Kh1 Rg7 transposes to a correspondence game which showed
some interesting tactical themes: 23 Nac5 g4 24 Rg1 g3 25 hxg3 Ng4!? 26
Be1 (not 26 fxg4? hxg4, when White must give up the queen with 27 Bxg4
Bxg4 28 Qxg4 Rh7+ 29 Qh3, since 27 Re1 Qg5 gives Black a crushing
attack) 26 ... Ne3 27 Qd2 Qg5 28 gxf4 exf4 with another impossibly complex
position in S.Cade-K.Bronnikov, correspondence 2014 (½-½ in 47).
22 ... Rg7 23 Ne6 Bxe6 24 dxe6 Qe7 25 Nc5 Kh8!
A good prophylactic move, especially because Nakamura wants the
option of playing ... Qxe6 in several lines.
26 Nb7 Ba3 27 Rc3 Bb4 28 Rd3
White can also try 28 Rc6, when Black gets counterplay with 28 ... g4!?
etc.
28 ... Qxe6 29 Qb3 Qxb3 30 Rxb3

White has enough compensation for the pawn in the form of his bishop
pair, but Black can’t have any problems in such a position. The rest of the
game features very impressive play from Nakamura, who keeps applying
pressure.
30 ... a5 31 Rc1 g4!
 Even without queens the white king can be a target, as Nakamura
excellently demonstrates.
32 Kf1 g3!?
The alternative is 32 ... gxf3 33 gxf3 Nh7, intending ... Ng5.
33 hxg3 fxg3 34 Bxg3 h4 35 Bf2 Nh5
Nakamura has executed the standard kingside pawn sacrifice and now has
good play on the dark squares.
36 f4!
This is the typical counter-sacrifice by White, improving his light-squared
bishop and gaining control of important squares on the third rank. After 36
Rc6? Ng3+ 37 Bxg3 hxg3, the g3-pawn is both a threat to the white king and
an eventual queening risk – amazingly, White is already lost!
36 ... Nhxf4 37 Bg4

Kasimdzhanov has played actively and well. The bishop pair fully
compensates for the pawn deficit in such an open position.
37 ... Rf8 38 a3 Bd6 39 Nxa5 Ne7
Here 39 ... Nxg2 is also possible, leading after 40 Kxg2 Ne7 41 Bxh4
Rxg4+ 42 Kh3 Rxe4 43 Bxe7 Bxe7 44 Rxc7 to similar play as in the game.
In either case White’s king remains insecure and he will need to be precise to
draw.
40 Bf3 h3 41 gxh3 Nxh3 42 Ke2 Ng5 43 Nc4 Ng6 44 Rh1+ Kg8 45 Nd2
Nf4+ 46 Kd1 Ra8 47 Nc4 Ra4 48 Nxd6 cxd6 49 Rh6 Rf7 50 Rxd6 Nfh3 51
Ke2 Rc4 52 Bb6 Rc2+ 53 Kd1 Rh2 54 Be2 Nxe4 55 Rd8+ Kg7 56 Re3
Nhg5 57 a4 Rf6 58 Rd7+ Kg6

59 a5?
Extremely natural, but this move is a mistake. White could have retained
the balance with 59 Ba5, but the mere fact that he has to find such moves to
maintain equality shows that Black’s position is much easier to play.
59 ... Rc6 60 Bd3?
Walking into mate in two. Instead, 60 Ba6 Nf2+ 61 Kd2 Nge4+ 62 Rxe4
Nxe4+ 63 Ke3 Nc5 leaves White a full exchange down, but there are always
drawing chances with such limited material.
60 ... Rd2+ 0-1
Since 61 Ke1 Rc1 mate follows.
 I find this game rather inspiring, for two reasons. Firstly, as he has done
in the past, Nakamura trusted his King’s Indian in one of the most important
competitions (moreover, against Kasimdzhanov, who is himself a King’s
Indian expert). Secondly, despite the exchange of queens on move 30,
Nakamura spent the next 30 moves attacking the white king, ending in mate.

Game 7
W.So-H.Nakamura
Sinquefield Cup, St Louis 2015

When I started working on this book, this game was the first I wanted to
include. It’s a perfect King’s Indian game by Nakamura.
1 d4 Nf6 2 c4 g6 3 Nc3 Bg7 4 e4 d6 5 Nf3 0-0 6 Be2 e5 7 0-0 Nc6 8 d5
Ne7 9 Ne1 Nd7 10 f3 f5 11 Be3 f4 12 Bf2 g5 13 Nd3

Now we look at alternatives to the immediate 13 Rc1. This one is

extremely logical. Black can no longer prevent c4-c5 since 13 ... b6?? would
be horrible, only helping White to open lines; i.e. 14 b4 and White is close to
winning.
13 ... Ng6
13 ... Nf6 comes to the same thing, since Black can’t stop c4-c5 anymore.
14 c5 Nf6
There is no good way for Black to resolve the tension, so taking on c5
isn’t to be recommended. White has achieved his break, but Black has
aggressively posted his knights on the kingside and is looking for an
opportunity to push ... g5-g4.
15 Rc1 Rf7!

Question: What’s the idea?

Answer: The same one we’ve seen before: this is the start of a typical
KID regrouping ( ... Rf7 and ... Bf8), which has the effect of securing key
squares and pawns on the queenside (c7 and d6 in particular), while
improving Black’s co-ordination for kingside operations; for example, ... Rg7
can be played to support ... g5-g4 and ultimately aim at the white king.
16 Kh1
This prophylactic move is preferred by Shirov. We will look at 16 a4 in
the next game.
16 ... h5
I like this aggressive approach. 16 ... Bf8 is another thematic move which
has been contested between strong players, such as A.Shirov-D.Vocaturo,
European Team Championship, Reykjavik 2015 (0-1 in 55). In the game, we
see Nakamura opting for a different regrouping.
 17 cxd6?!
I think this is where So mixed up his move order. The immediate 17 Nb5!
is better and leads to a critical position for the assessment of the whole line:
17 ... g4 (17 ... a6 18 Na3 Bf8, as in L.Bruzon Batista-D.Arenas Vanegas,
San Salvador 2016, is probably what So was aiming for: Black can’t play ...
b7-b5, so the knight is coming to c4 with strong pressure) 18 cxd6 cxd6 19
Qc2! (this move is endorsed by Krasenkow, who is an expert on the white
side of the King’s Indian; 19 Nxa7 Bd7 20 Qb3 was A.Merry-M.Hebden,
Jersey 2016, when Krasenkow suggests 20 ... h4 as unclear) 19 ... Ne8 (I
think Black should take time out for this move; 19 ... g3 is what Black wants
to play but it’s well met by 20 Nc7! Rxc7 21 Qxc7 Qxc7 22 Rxc7 gxf2 23
Rfc1, which left White with dominant rooks in A.Shirov-E.Bacrot, European
Cup, Bilbao 2014 – even Bacrot, one of the best King’s Indian players,
couldn’t save the game; 1-0 in 51) 20 Nxa7 Bd7 21 Qb3 g3 22 Bg1 was
T.Batchuluun-A.El Jawich, Abu Dhabi 2016; here I like 22 ... Bf8 followed
by ... gxh2 with an unclear position.
17 ... cxd6 18 Nb5 a6!
 This is based on a very concrete idea, because White has not yet played
a2-a4.
19 Na3
After 19 Na7 Bd7 20 Qb3 g4 21 Qb6 Qxb6 22 Bxb6 h4!, Black was
pressing in G.Kjartansson-V.Kovalev, Riga 2015 (0-1 in 48). The knight on
a7 looks like more of a liability than an asset.

Exercise: What is White’s threat and how should Black respond?

Answer: The threat is Nc4, aiming at the weak dark squares on the
queenside. And the response:
19 ... b5!
This is why White normally starts with a2-a4 before commencing his
knight tour. He will now need to spend several tempi getting his queenside
play going, during which time Nakamura will attack the white king.
20 Rc6 g4 21 Qc2 Qf8 22 Rc1 Bd7 23 Rc7
It’s quite striking how little White has achieved on the queenside. Despite
the resources devoted to that side of the board (including tripling heavy
pieces on the c-file), he has provoked no weaknesses and has no targets and
no pawn breaks. In the meantime, Black’s kingside attack is just as dangerous
as it looks.
23 ... Bh6!
An excellent regrouping. We often see ... Rf7 in combination with ... Bf8
so that the rook can operate on the g- and h-files. Here the f8-square is
occupied by the queen but h6 is a great alternative. An added bonus is that
the bishop can find work on the c1-h6 diagonal, if and when the kingside
opens, and will attack the rook on c1.
24 Be1 h4 25 fxg4 f3 26 gxf3

Exercise: How should Black continue?

Answer:
26 ... Nxe4!!
A beautiful shot, which it takes the computer a long time to assess
properly.
27 Rd1!
Nakamura’s concept is based on an amazing move after 27 fxe4 Rf1+ 28
Kg2.

Exercise: Can you find it?

Answer: 28 ... Be3!! is the bone crusher. (The immediate 28 ... h3+ would
only have led to a draw.) Black’s threat, which is impossible to meet without
decisive material loss, is 29 ... Rg1+ 30 Kh3 Qf1+!! 31 Bxf1 Bxg4 mate. In
my view, this is far better than anything which occurred in the Immortal or
Evergreen games.
 Exercise: So has managed to find a move posing Black practical
problems. There’s only one good move, what is it?

Answer:
27 ... Rxf3!
Black is winning, since 28 Bxf3 Qxf3+ 29 Kg1 (29 Qg2 Bxg4 wins the
house) 29 ... Be3+ 30 Bf2 Nxf2 31 Nxf2 Nf4 leads to a quick mate.
28 Rxd7 Rf1+ 29 Kg2 Be3
Here 29 ... h3+! 30 Kxh3 Rf2!! was even better, but Nakamura’s
approach is more than good enough to win.
30 Bg3 hxg3 31 Rxf1 Nh4+ 32 Kh3 Qh6
The white king is facing decisive threats.
33 g5 Nxg5+ 34 Kg4 Nhf3 35 Nf2 Qh4+ 36 Kf5 Rf8+ 37 Kg6 Rf6+!
 Wonderful – even if 37 ... Ne6 or 37 ... Nf7 forced mate one move faster.
38 Kxf6 Ne4+ 39 Kg6
Or similarly 39 Kf5 Nd4+ 40 Kg6 Qg5 mate.
39 ... Qg5 mate
The most beautiful game of the 21st century. Nakamura won surprisingly
easily against such strong opposition.

Game 8
D.Jakovenko-D.Fedoseev
Taganrog 2011

1 Nf3 Nf6 2 c4 g6 3 Nc3 Bg7 4 e4 d6 5 d4 0-0 6 Be2 e5 7 0-0 Nc6 8 d5 Ne7

9 Ne1 Nd7 10 Be3 f5 11 f3 f4 12 Bf2 g5 13 Nd3 Nf6 14 c5 Ng6 15 Rc1 Rf7
16 a4
 A very useful move, gaining space and supporting Nb5. We can see just
how badly White needed this move in So-Nakamura.
16 ... h5
Again, 16 ... Bf8 is a major alternative, which has been used by
Namakura and Ding Liren. However, I have decided to recommend the more
direct move, aiming to force through ... g5-g4.

Question: What do you think about stopping White from gaining further
space on the queenside with 16 ... a5 - ?

Answer: We will see a few examples of meeting a2-a4 with ... a7-a5 in
the next game (including at move 9). However, this only makes sense when
Black can block the queenside with a quick ... b7-b6. In this position, where
White has already achieved c4-c5, the weakening of the b5- and b6-squares is
fatal: 17 cxd6 cxd6 18 Nb5 and Black is close to lost.
17 a5
White can also play more directly with 17 Nb5 (or 17 cxd6 cxd6 18 Nb5
g4, transposing) 17 ... g4 18 cxd6 cxd6 and then:

a) 19 Bxa7 g3 20 Kh1 (20 Qe1 was seen in J.Zhou-Y.Wen, Chinese Team

Championship 2011, when I like 20 ... Bh6, improving Black’s bishop while
keeping his options of ... g3xh2+ and ... Rg7 open; note that 21 hxg3 Rg7!
offers excellent compensation) 20 ... Ne8 21 Bg1 Bd7 (21 ... gxh2?! has been
tried in a few correspondence games, but this approach seems wrong in
principle: after 22 Bf2! the pawn on h2 shields the white king and there is no
easy way to continue the black attack) 22 Qb3 Nh4 23 Ne1 Nxg2!! 24 Nxg2
h4 25 Rfd1 h3 26 Ne1 and the kingside pawn roller is worth a piece. Here
Black dropped his bishop back to f8 in R.Bondars-A.Kural, correspondence
2013 (½-½ in 40), whereas I prefer the more active 26 ... Bf6, followed by ...
Rg7 and ... Bh4, with a thoroughly unclear position.
b) 19 Qc2 creates the strong positional threat of Nc7-e6, which I think
Black should take time to prevent with 19 ... Ne8. This has been extensively
tested in correspondence games, with no clear verdict. The same position was
reached in Ju Wenjun-I.Chelushkina, Khanty-Mansiysk Olympiad, which
continued 20 Kh1 Bd7 21 Nxa7 g3 22 Bg1 Nh4?! 23 Ne1 Nxg2 24 Nxg2 h4
25 Rfd1 h3 26 Ne1 and White went on to win (1-0 in 55). The theme of a ...
Nxg2 sacrifice followed by a slow pawn storm is familiar to us from line ‘a’,
but here White has an extra resource in Bb5!, aiming to exchange Black’s
key attacker, when an invasion on c8 might follow (even the knight on a7
contributes to this). The difference is subtle, but I haven’t been able to make
Black’s sacrifice work in this version, so it is probably better to resort to 22 ...
gxh2. This is often rejected on sight as, unless Black has a concrete follow-
up, the pawn exchange takes a lot of dynamism from his kingside attack.
Here, however, Black seems to be quite fast after 23 Bf2 Bf8 24 Nb5 Rg7,
followed by ... Qg5 and perhaps ... h5-h4, and his position seems quite
promising.
17 ... g4

We can see that, in this line, both sides simply push on “their” respective
sides of the board, without expending any time or energy on defence.
18 cxd6
Other moves:
a) 18 c6 was played in D.Kigel-I.Cheparinov, Golden Sands 2012, and
now I agree with Van Kampen’s recommendation of 18 ... bxc6 19 dxc6 Be6,
when Black has a good game.
b) 18 a6 seems inconsistent with the rook on c1. After 18 ... bxa6! 19 Ra1
g3 20 hxg3 fxg3 21 Bxg3 h4 22 Bf2 Nh5, Black had the standard kingside
pressure in M.Martys-S.Marton Bardocz, correspondence 2006 (0-1 in 35).
c) 18 Nb5 is well met by 18 ... a6!, an improvement suggested by
Krasenkow in his notes to A.Volokitin-D.Reinderman, European
Championship, Aix les Bains 2011 (1-0 in 34), where 18 ... g3 was played.
Matters aren’t so clear after that move either, but I think 18 ... a6 is a strong
idea. The standard piece sacrifice of 19 cxd6 axb5 20 dxc7 clearly doesn’t
work, so the knight has to go back: 19 Na3 (the most ambitious retreat,
aiming to bring the knight to c4 from where it exerts pressure on b6 and d6;
instead, 19 Nc3 was tried in A.Ponomarev-Y.Ilyushchenko, correspondence
2012, when I recommend continuing consistently 19 ... g3!? 20 hxg3 fxg3 21
Bxg3 h4, followed by ... Nh5 with good compensation) 19 ... g3! (19 ... h4
was Krasenkow’s original recommendation, when he assessed 20 cxd6 g3 21
Bc5 h3 as unclear; in fact Black’s game seems to be overwhelming here, so
the piece sacrifice 21 dxc7! etc is a better defensive attempt, and I’m not a
fan of the resulting position for Black: for the piece White has two pawns,
including a very strong passed pawn on d5, and good prospects on the
queenside, whereas it’s not clear how Black can make progress on the
kingside) 20 hxg3 fxg3 21 Bxg3 h4 22 Bh2 Bh6 23 Rc3? (the exchange
sacrifice 23 Nc4 is favoured by the engines and was tested in F.Retamoza-
J.Simmelink, correspondence 2008, where Black took on c1 – I think it is
more dynamic to maintain the tension with 23 ... Rg7 with a complete mess)
23 ... Be3+ 24 Kh1?? (now Black has a winning attack, which is unsurprising
since he has tremendous play on the dark squares; White had to block on f2
with the knight or – probably better – the rook, though Black would have had
great play in any case) 24 ... Nh5! (let’s look at the attacking ratio here: Black
is attacking with two knights, his bishop on e3, and his pawn on h4; in
addition, the Rf7, Qd8 and Bc8 are all contributing and can be mobilized in a
single move) 25 Rf2 Qg5 26 Nc2 Ng3+ 27 Kg1 Bxf2+ 28 Nxf2 Nf4 29 Bf1
Nge2+ (this is more than good enough, though 29 ... h3! was probably the
most devastating win) and here White took on e2 with the queen in
R.Alarcon Casellas-K.Oliva Castaneda, Sitges 2014 (0-1 in 35), which was
equivalent to resignation, but otherwise Black captures on c3 with a
continuing attack.
 18 ... cxd6 19 Nb5

Exercise: How should Black continue?

Answer:
19 ... g3!
A typical pawn sacrifice, which we will assess in a few moves. Note that
19 ... a6?? simply loses to 20 Bb6, followed by Nc7, either winning an
exchange (after ... Rb8 and Ba7) or killing Black’s attacking chances with
Ne6. I can’t emphasize enough that positions of this type, where Black is
collapsing on the queenside but has a far advanced kingside pawn storm, are
almost always unplayable without the light-squared bishop, and taking a
knight on e6 (even winning a pawn in the process) isn’t a good enough
outcome for this piece.
20 hxg3
Not 20 Bxa7? Qxa5! and Black wins material.
20 ... fxg3 21 Bxg3
 Question: What has Black obtained for his pawn sacrifice?

Answer: Several things:

1. Black now has two half-open files (g- and f-) towards the white king.
In particular, the g-file will be a very promising avenue for a rook (on g7)
and perhaps the queen (on g5).
2. Black has obtained possibilities of using several excellent squares close
to the white king. The c1-h6 diagonal has opened, and can be occupied by
Black’s dark-squared bishop. The squares on g3 and f4 can be occupied by
the black knights (after ... h5-h4 and ... Nh5).
3. Black obtains a potential pawn break, namely ... h5-h4-h3, to weaken
White’s structure further.
4. White’s kingside structure has been weakened by the removal of the
h2-pawn. Note that a standard sequence of ... g5-g4-g3xh2 can often be met
by Kh1!, sheltering behind the black pawn which can block its own forces.
Here the white king is more open.
 5. Finally, Black gains some time and initiative. The white bishop has
been deflected from its attacking duties on the queenside (from f2 it targeted
a7, which couldn’t easily be dealt with), and is exposed to further attack
(after ... h5-h4 or ... Rg7, it will likely have to move again).
Against this, White has some reasons to be happy too. He is now a pawn
up, and often relies on this by returning it with f3-f4 (which gives his pieces
more space to operate on the kingside and can dampen Black’s attack there).
In addition, White now knows the type of game Black is playing. With the
pawn on g4, Black threatens a gradual protected push to g3 (say after ... Bh6,
... Rg7 and, perhaps, moving the knight from g6) when the pawn controls key
squares immediately around the white king, and h2-h3 can often be met by
the key ... Bxh3! sacrifice. By committing to the pawn sacrifice, Black loses
some flexibility in his attack.
21 ... Bh6
I like this move; Black activates his “problem” dark-squared bishop with
tempo and vacates the g7-square for the rook.
Alternatively, 21 ... h4 has been tried by Cheparinov; for example, 22 Be1
(22 Bf2 Nh5 gives White the option of grabbing the a-pawn with 23 Bxa7,
since 23 ... Qxa5 doesn’t win a piece in view of 24 Bf2 Qxb5 25 Nf4, but
after 23 ... Ng3! Black has excellent play on the kingside) 22 ... Nh5 23 f4!
(this is the correct reaction in 90% of such positions. White’s situation
becomes extremely dangerous if he permits the kingside to be blocked by a
black knight on f4, generating threats against the white king) 23 ... Nhxf4 24
Bh5 was A.Volokitin-I.Cheparinov, World Rapid Championship, Dubai
2014, where 24 ... Qg5 25 Nxf4 Rxf4 led to an unclear position in which
Black eventually prevailed (0-1 in 51), but at this moment he had an
interesting shot with 24 ... Nh3+!? 25 gxh3 Rxf1+ 26 Kxf1 Bxh3+ 27 Ke2
Qg5 28 Bf3 Bg2. Despite his extra piece, it’s tough for White to defend his
king while dealing with the passed h-pawn.
 22 f4?
A standard reply in such positions (as we saw in the previous note),
returning the pawn to reduce the pressure on the kingside, and had Black
played 21 ... h4 22 Bh2 Bh6, this would have been a good move.
Unfortunately, there is a key difference here. Instead, two rook moves have
been tested in correspondence games:
a) 22 Rc3 Rg7 (22 ... a6 has also been tried, but I prefer the aggressive
rook move, preparing play down the g-file) 23 Qc2 (23 Bf2 Bh3 24 Rc7 Rxc7
25 Nxc7 Qxc7 26 gxh3 Qxa5 was unclear in C.Oger-A.Ciciotti,
correspondence 2011; ½-½ in 49) 23 ... h4 24 Be1 Ne8 25 Bd2 Bxd2 (25 ...
Nf4 is also interesting, as in K.Wapniewski-B.Philippeit, correspondence
2012; ½-½ in 39) 26 Qxd2 Qxa5 and Black was no worse in B.Philippeit-
A.Kazoks, correspondence 2013 (½-½ in 36).
b) 22 Rc2 h4 23 Bf2 Nh5 24 f4 Nhxf4 25 Bf3 b6 26 axb6 axb6 27 Rc6
Bf8 28 Kh2 Qf6 29 Ne1 Rb8 and Black’s attacking chances compensated for
his queenside weaknesses in K.Ederer-V.Astroukh, correspondence 2012 (½-
½ in 43).
 22 ... Nxe4!
Regaining a key central pawn with tempo. This is the first difference
based on the omission of 21 ... h4.
23 Bh2
Perhaps Jakovenko was relying on the intermezzo 23 Bxh5?, but after 23
... Nxg3 24 Bxg6? Rf6, both the g6-bishop and f1-rook are hanging, so White
loses material.
23 ... Nh4!
This is the second difference. Without ... h5-h4 having been played,
Black can use the h4-square for a piece. His knights now create dangerous
threats to the white king.
24 Nf2 Bf5

25 Bd3?
Understandably, Jakovenko tries to trade some of the dangerous attacking
pieces but allows Black fresh attacking resources.
The computer stubbornly insists that 25 Rc4 is equal, based on the line 25
... Rg7 26 Nxe4 Nxg2 27 Kh1 Bxf4 28 Nexd6 Bg6, when it claims that
Black’s compensation is only good enough for a draw. Assuming you tend to
play humans rather than engines, here are two points about this position:
1. No human player would voluntarily go for this, because;
2. The vast majority of players would simply get mated with White.
Nonetheless, it does seem to be a draw, the cute main line running 29
Rg1! Qh4 30 Rxg2 Be4! 31 Nxe4 Rxg2 32 Kxg2 Qxh2+ 33 Kf1 Qh1+ 34
Kf2 Qh4+ and Black delivers perpetual check.
25 ... Nxf2 26 Rxf2 e4
26 ... a6! was also strong, but Fedoseev is playing an excellent game and
his move retains some advantage.
27 Be2
27 Qxh5 Rh7 doesn’t help.
27 ... e3 28 Rf1 Rg7
Black has a wonderful attack at no material cost. Jakovenko does well
even to stay in the game.
29 g3 Bh3 30 Nd4
Now if Black takes on f1, the knight will land on e6, but Fedoseev is in
no rush to trade his best attacker.
30 ... Qf6! 31 Ne6 Nf5?!
Playing consistently for the attack, but this seems to be an inaccuracy.
The computer strongly favours 31 ... Bxe6 32 dxe6 Nf5, when Black will
play ... Kh8, ... Rag8, and collect the e6-pawn at leisure.
32 Nxg7 Qxg7 33 Kh1 h4
 34 Rg1??
Understandably, Javokenko (presumably in time pressure) cracks under
the pressure. 34 g4! was the only move. Then Black is no worse after 34 ...
Ng3+ 35 Bxg3 hxg3 36 Rf3! Qxg4 37 Qd3 but nor is he any better, despite
the dangerous appearance from White’s perspective. Black doesn’t seem to
have more than the draw he can force by 37 ... g2+ 38 Kh2 Bxf4+ 39 Rxf4
Qxf4+ 40 Kxh3 Qh6+ 41 Kxg2 Qg7+ etc.
34 ... hxg3 35 Bxg3 Nxg3+ 36 Kh2 Bxf4 37 Kxh3 Kh8 38 Bg4 e2
Here 38 ... Ne4 was stronger.
39 Qd3 Qh6+?
I suspect both players were in severe time trouble. 39 ... Re8 would
maintain a winning position.
40 Kg2 Bxc1 41 Qc3+ Qg7 42 Qxg7+ Kxg7 43 Kxg3 Bxb2 44 Bxe2
Bc3 45 a6 bxa6 46 Bxa6
 How should we assess this position? Endgames with rooks and opposite-
coloured bishops are notoriously difficult to evaluate. I would have thought
that White has good drawing chances here, but given the failure of a 2700+
GM like Jakovenko to hold, White is obviously facing practical problems.
46 ... Kf6 47 Rc1 Ba5 48 Ra1 Bb6 49 Re1 Bc5 50 Kf4 Rg8 51 Re6+
Kf7 52 Be2 Rg1 53 Bh5+ Kg7 54 Re7+ Kh6 55 Be2 Kg6 56 Re6+ Kg7 57
Re7+ Kf8 58 Rc7 a5 59 Bb5 Rb1 60 Bd3 Rd1 61 Bf5 a4 62 Kg5 Rg1+ 63
Bg4 Be3+ 64 Kf6 Bd4+ 65 Kg5 Bb6 66 Rc8+ Kf7 67 Ra8 Be3+ 68 Kf5
Rf1+ 69 Ke4 Bc1 70 Be6+ Kg7 71 Kd3 a3 72 Kc2 Bb2 73 Rg8+ Kh7 74
Ra8 Rf3 75 Kb1 Bd4 76 Ra4 Bc5 77 Rh4+ Kg6 78 Rh2 Kf6 79 Re2 Rg3
80 Ka2 Ke7 81 Bf5+ Kd8 82 Bc2 Kc7 83 Bb3 Kb6 84 Rc2 Ka5 85 Rc4
Rg2+ 86 Rc2 Rg7 87 Rc4 Kb5 88 Rc2 Bd4 89 Rc4 Bb2 90 Re4 Rg2 91
Ba4+ Kc5
 Black has clearly made huge progress, but I’m still not sure whether he
can force a win after 92 Bb3. Instead, exhausted by a prolonged defence,
Jakovenko blunders.
92 Bc6? Bc1+ 93 Kb3 Rb2+! 94 Kc3 Bd2+! 95 Kd3 a2 96 Rc4+ Kb6
97 Ra4 Bh6 0-1
Black moves his rook along the second rank, plays ... Bg7, and queens the
pawn.

Game 9
J.Krassowizkij-R.Van Kampen
German League 2016

1 d4 Nf6 2 c4 g6 3 Nc3 Bg7 4 e4 d6 5 Nf3 0-0 6 Be2 e5 7 0-0 Nc6 8 d5 Ne7

9 Ne1
If White likes systems with a2-a4 he can enter them as early as right now
with 9 a4. As in the main game, I recommend meeting this with 9 ... a5!,
stopping b2-b4 and denying White a natural pawn break on the queenside.
Play can continue 10 Ne1 Nd7 11 Nd3 (after 11 Be3 f5 12 f3, Black can
either transpose to the main game with 12 ... f4 13 Bf2 g5 or vary with 12 ...
Nc5!?, as Kasparov did in a famous victory against Korchnoi in Barcelona
1989) 11 ... f5 12 f3 Kh8 and then:

a) 13 Bd2 c6!? (the alternative is to play ... b7-b6 and ... Nc5 as in line
‘b’, but Sasikiran’s plan is also valid) 14 Kh1 Nf6 15 dxc6 (I’d prefer
retaining the tension with 15 Rc1 or 15 Bg5) 15 ... bxc6 16 c5 d5 17 Nxe5 d4
18 Qb3 Nxe4! (perhaps Bruzon was counting on 18 ... dxc3 19 Bxc3, when
White has quite decent compensation for the piece) 19 Nxe4 fxe4 20 fxe4?
(20 Nf7+ had to be tried, though 20 ... Rxf7 21 Qxf7 e3 gives Black a
wonderful position) 20 ... Bxe5 21 Rxf8+ Qxf8, when White had insufficient
compensation for the piece and went on to lose in L.Bruzon Batista-
K.Sasikiran, Danzhou 2015 (0-1 in 38).
b) 13 Be3 is more common, but the plan with 13 ... b6 followed by ... Nc5
is very solid. For example, 14 Kh1 Nc5 15 Nxc5 bxc5 16 Nb5 Ng8! (a useful
regrouping, made possible by ... Kh8) 17 Ra3 Bh6! (trading the dark-squared
bishops and avoiding the risk of being left with a passive bishop on g7) 18
exf5 Rxf5 19 Qd2 Bxe3 20 Rxe3 Rf7 21 f4 exf4 22 Qc3+ Qf6 23 Qxf6+
Nxf6 24 Rxf4 Kg7 with a roughly level endgame in N.Nikcevic-N.Djukic,
Montenegrin Championship, Podgorica 2014 (0-1 in 88).
9 ... Nd7 10 Be3 f5 11 f3 f4 12 Bf2 g5 13 a4

Question: What’s the idea of this pawn advance?

Answer: We see in other lines in the Classical just how useful it can be
for White to include this move. The pawn protects the b5-square and supports
a knight on b5, both for a raid into enemy territory (Nb5xa7-b5) and a retreat
(meeting ... a7-a6 with Na3, when ... b7-b5 as in So-Nakamura isn’t
possible). The pawn can also advance to a5, threatening to go to a6, while if
Black plays ... a7-a6 himself, the b6-square will be weak after c4-c5 and an
exchange of pawns on d6.
13 ... a5!
This seems the most logical response to me, blocking the a-pawn and
emphasizing that White’s b2-b4 push will create weaknesses in his own
camp.
Players willing to take more risk can consider 13 ... Rf6 14 a5 Rh6, a very
direct approach which was successful in M.Swayams-A.Pichot, Moscow
2016 (0-1 in 67). The problem is that, if the kingside attack fails, White has
established a dominant position on the queenside thanks to his pawn on a5.
14 Nd3

Exercise: What should Black play here?

Answer:
14 ... b6!
More or less essential. If White is allowed to get c4-c5 in, the inclusion of
a2-a4 and ... a7-a5 will give him a decisive advantage compared to other 9
Ne1 lines, since Black is so compromised on the queenside and the weakness
of the b5- and b6-squares will tell.
The structure which we now have is an interesting one. Black has played
much more aggressively on the queenside than in other lines, but it can be
seen that White’s pawn formation is compromised quite badly by the
inclusion of a2-a4, and that he’ll need to trade a lot of pawns to achieve his
desired c4-c5 break. Accordingly, Black runs considerably less risk of being
overrun in this line, since the entire queenside tends to get exchanged.
15 b4
This is the only sensible plan. White still needs to seek his chances on this
side of the board.
15 ... axb4 16 Nxb4
White has also experimented with 16 Nb5, aiming to recapture the b4-
pawn with the bishop. He’s scored awfully with this idea, since Black gets
very rapid play on the kingside: 16 ... Nf6 17 Be1 g4! (of course; Black takes
advantage of the knight’s absence from c3, meaning that the e4-pawn will
hang if White plays f3xg4) 18 Bxb4 g3 19 h3

Exercise: Which standard device does Black have here?

Answer: 19 ... Bxh3! (Black can also prepare this move – 19 ... Ng6, for
instance, is a good move – but the standard bishop sacrifice works fully: as
we’ve mentioned before, without this resource the white kingside would be
quite secure, but now White needs to play precisely to stay in the game) 20
gxh3 Qd7 21 Qc2 (21 Nf2! was the only chance, although after 21 ... Ng6
Black’s attack continues at no material cost) 21 ... Ng6! (excellent technique
and a typical follow-up to ... Bxh3; Jacob Aagaard talked about this concept
wonderfully under the heading “Revolution/Evolution” in his Attacking
Manuals; instead, 21 ... Qxh3 22 Bd1 enables the white queen to come to g2,
and while this still looks good for Black, there’s absolutely no reason to
allow it) 22 Bd1 Nh4 (this is the difference: White can’t hold the kingside
and Black wins) 23 Ne1 Qxh3 24 Nxc7

For a surprisingly long time my engine thinks White is fine here, which
shows just how poorly they cope with King’s Indian positions (compared to
their prowess in every other type of position). To most experienced King’s
Indian players it’s immediately obvious that White can’t cover the key
squares around his king, no matter how many extra pieces he collects.
 Exercise: What’s the most efficient execution?

Answer: 24 ... Nh5! (the rook on a8 isn’t important; getting a knight to g3

is) 25 Nxa8 g2 26 Nxg2 (26 Qxg2 is the sort of move which must always be
considered in such positions, since Black often gives up lots of material to get
here; in this case, after 26 ... Nxg2 27 Nxg2 Rxa8, Black retains a decisive
material and strategic advantage) 26 ... Ng3 and White resigned in M.Dziuba-
J.Czakon, Koszalin 2005 – another example of a strong player (rated 2460)
getting mated with White before move 30, which just doesn’t happen in the
Queen’s Gambit Declined.
16 ... Nf6
Black actually played 16 ... h5 17 Nb5 Nf6 here, but I’ve changed the
move order to 16 ... Nf6, as Van Kampen himself recommended in his
chess24 lectures.

17 Nb5
Improving his queenside prospects while clearing the third rank to enable
the rook to defend the kingside laterally. Instead:
 a) 17 Be1 h5 18 h3 Rf7 19 Nd3 Bf8 20 Nf2 Rg7 and Black was fine in
K.Friedrichs-J.Koscielski, German League 2004 (1-0 in 39).
b) 17 Nc6 has led to decent results for White, who runs less risk by
opening the position. A couple of good examples for Black: 17 ... Nxc6 18
dxc6 Qe8 19 Nd5 Rf7 20 a5 (20 Qb3 was played in F.Perez Ponsa-D.Stets,
Rochefort 2015; here I like 20 ... Bf8 followed by ... h7-h5 and ... Kh7) 20 ...
bxa5 21 Qa4 g4 22 Rfb1 was J.Hammer-K.Berbatov, Reykjavik 2011, and
now 22 ... g3! is a good recommendation by GM Van Kampen; e.g. 23 hxg3
Nxd5 24 cxd5 fxg3 25 Bxg3 Bh6, followed by ... Rg7 with excellent play for
the pawn.
17 ... h5
White’s last move made the immediate 17 ... g4!? possible since the e4-
pawn would drop after 18 fxg4. However, Van Kampen’s move is also
strong.
18 Ra3
A new move, but quite a logical one, supporting the kingside along the
third rank. Instead:
a) 18 a5 Rxa5 19 Rxa5 bxa5 has been played a couple of times, when
Black is doing well. White will lose time regaining the pawn on a5, while
Black will push his g-pawn up the board.
b) 18 Nc6 Nxc6 19 dxc6 Be6 20 Qd2 was N.Stanec-R.Leitao, Yerevan
Olympiad 1996, and now I favour the immediate 20 ... g4 with standard
kingside counterplay.
c) 18 Nd3 g4 19 a5 bxa5 20 c5 was M.Bucher-A.Baryshpolets, Albena
2012. Here I like 20 ... gxf3 21 Bxf3 Bg4 with an unclear game. Again, the
time White spends regaining the pawn may be invested in Black’s kingside
attack.
18 ... g4 19 Bh4 Qd7 20 fxg4?!
I don’t like voluntarily opening the kingside in this fashion, though Black
already had a comfortable game.
20 ... hxg4 21 Nc6 Ng6 22 Bxf6 Rxf6 23 Nca7
 Exercise: How should Black proceed?

Answer:
23 ... f3! 24 gxf3 g3!
A fantastic double pawn sacrifice by Van Kampen, opening the white
king.
25 hxg3?
White had to find 25 f4!! (a thematic move, but not easy to see in this
position) 25 ... exf4 26 hxg3, when the game is quite unclear, including after
the inspired 26 ... Rxa7!? (removing a defender of the a3-rook) 27 Nxa7 Qh3
with a continuing attack.
25 ... Qh3 26 Kf2 Qh2+ 27 Ke3 Bh3!
Van Kampen continues playing with energy and precision.
28 Re1 Bg2!
A remarkable “fianchetto”, putting the white kingside under enormous
pressure.
 29 Kd3 Bh6 30 Kc2 Rxf3 31 Rxf3 Bxf3 32 Kb3 Bxe2 33 Rxe2 Qxg3+
34 Ka2 Nf4!

White has skilfully evacuated his king from the danger zone, but he
remains with the inferior structure and poorly placed pieces.
35 Rc2?
The wrong square. 35 Rd2 would have kept White in the game.
35 ... Nd3!
Bringing the knight, with tempo, towards the vulnerable dark squares on
the queenside.
36 Qb1 Qe3 37 Rc3 Qe2+ 38 Ka3 Nc5 39 Qh1 Bg7 40 Qf3 Qe1 41 Rc2
Qa1+ 42 Ra2 Qc1+ 43 Kb4 Qb1+ 44 Ka3 Nd3 45 Nc6? Qc1+ 46 Kb3
Nc5+ 47 Kb4
 Exercise: How can Black give a rapid mate?

Answer:
47 ... Rxa4+! 48 Rxa4 Qb2+ 49 Qb3 Nd3 mate
An excellent game by Van Kampen, finishing with a beautiful mate.

Game 10
M.Ragger-A.Grischuk
European Cup, Skopje 2015

1 Nf3 Nf6 2 c4 g6 3 Nc3 Bg7 4 e4 d6 5 d4 0-0 6 Be2 e5 7 0-0 Nc6 8 d5 Ne7

9 Ne1 Nd7 10 Nd3 f5 11 f3
White often plays 11 Bd2 first, when 11 ... Nf6 12 f3 f4 is another
common move order, transposing to the game.
11 ... f4 12 Bd2
 Question: What is White’s idea?

Answer: In the previous games in this chapter, we saw White putting his
bishop on f2. This poured gasoline on both sides’ attacks – Black was under
severe pressure on the queenside, but could gain time against this bishop for
his kingside pawn storm with ... f5-f4 (and, on a good day, ... g4-g3, even as a
pawn sacrifice). White’s aim in the current variation is rather different. He
plans to use the f2-square for a knight rather than the bishop. This knight is
an excellent defender of the white king (for a start, ... g5-g4 will be harder to
achieve), but conversely Black is under much less queenside pressure than in
the variations with Be3. Although this set-up appears more modest than the
Be3 lines, it still gives rise to deeply complicated tactical play with attacks on
opposite sides of the board, which is exactly what we’re looking for when we
play the King’s Indian!
12 ... Nf6
Via this specific move order Black can start with 12 ... g5 to delay the c4-
c5 break, although it turns out that, after 13 Rc1 Ng6 14 c5, taking on c5
would give White excellent compensation (for example, 14 ... Nxc5 15 Nxc5
dxc5 16 Na4 b6 17 b4), so Black has nothing better than transposing to the
game with 14 ... Nf6.
13 c5
Trying to block the kingside with 13 g4 h5 14 g5 Nh7 15 h4 allows the
surprising resource 15 ... Nxg5! 16 hxg5 Nf5!. Taking on f5 gives Black at
least a draw, while 17 Rf2 Ng3 18 Rg2, as played in A.Khalifman-U.Eliseev,
Russian Team Championship 2016, is best met by 18 ... Qxg5, when the
monster knight on g3 and the rolling kingside pawns give Black full
compensation for the piece.
13 ... g5

14 Rc1
White can take on d6 at any point over the next several moves. Black’s
response is similar in most of the lines: 14 cxd6 cxd6 15 Nf2 Ng6 16 Qc2
Rf7 17 Rfc1 Bd7 (completing the standard regrouping with 17 ... Bf8? would
be met by 18 Nb5! Ne8 19 a4 with similar play to the game, except that
Black’s knight on e8 is more passively placed; note that 19 Nxa7 is met by
the typical trick 19 ... Rc7, but even so 20 Ba5! Rxc2 21 Bxd8 Rxe2 22 Nxc8
Rxb2 looks unpleasant for Black) 18 a4 h5 19 h3 Bf8 20 Nb5 Rc8 (Black has
other good options including 20 ... a6, compromising the queenside but
gaining time, and 20 ... g4, playing aggressively on the kingside; Fedorov
plays directly for a pawn sacrifice which gives excellent compensation, at
least in a practical game) 21 Qb3 Nh4!? 22 Rxc8 Bxc8 23 Nxa7 Bd7! (Black
must of course retain the light-squared bishop; I would not want to play
White here, since the position is probably objectively balanced and far easier
to play for Black, and even computers, which hate the King’s Indian for
Black, only give White a slight advantage) 24 Be1 Rg7 25 Ra3? (a natural
and thematic move, improving the rook and defending the kingside along the
third rank, but a big mistake; 25 Rc1 was better, although Black retains the
initiative after 25 ... g4 26 hxg4 hxg4 27 fxg4 Nh5 with huge complications)

Exercise: How should Black continue?

 Answer: 25 ... g4! (this thematic break opens the kingside and exposes
the g2-pawn, in addition to the king sitting behind it) 26 hxg4 hxg4 27 Nc6
(trying to change the character of the game; after 27 fxg4 Nxg4 28 Bxg4
Bxg4 29 Nxg4 Rxg4 30 Bxh4 Rxh4 31 Qxb7 Rh7 32 Qa6 Qh4, Black has
overwhelming compensation for the two pawns and only White can be
worse) 27 ... Qc8 (one of several good options) 28 Nxg4 Nxg4 29 fxg4 Bxg4
30 Bf1 Nxg2! (subtlety isn’t Fedorov’s strong suit) 31 Bxg2 f3 32 a5 Bh5 33
Kf1 Rxg2 and White finally threw in the towel in A.Aleksandrov-A.Fedorov,
Belarus Championship, Minsk 2016. A brilliant KID attack.
14 ... Ng6 15 Nb5
Increasing the pressure on the queenside, supported by some nice tactical
ideas. Sometimes White opts to exchange on d6 first: 15 cxd6 cxd6 16 Nb5
Rf7 17 Qc2 Ne8, after which it looks like Black has been forced into
passivity, but ... h7-h5 and ... g5-g4 are coming, while there are no
breakthroughs for White on the queenside (with the bishop on f2 instead of
d2, Black would probably have to resign). For example:
a) 18 a4 h5 19 h3?! (I don’t like this decision as it makes ... g5-g4 even
more powerful when it lands) 19 ... g4 20 fxg4 hxg4 21 hxg4 was
A.Jakubiec-A.Fedorov, Czech League 2014. Now Fedorov put his bishop on
f8, which was okay, but I prefer 21 ... Bf6!, followed by ... Bh4, which looks
really unpleasant for White.
b) 18 Nf2 h5 19 a4 Bd7 20 Rfd1 Bf8 21 a5 g4! 22 fxg4 hxg4 (22 ... Nh4
is also very promising) 23 Nxg4 Qh4 24 Be1 Qg5 25 h3
 Exercise: Yet another tactical exercise. How should Black continue?

Answer: 25 ... f3!! (beautifully exploiting the knight on b5 and the

enduring weaknesses in White’s kingside) 26 Bxf3 Bxb5 with a clear
advantage, or if 26 gxf3, as in P.Acs-Z.Almasi, Paks 2009, then 26 ... Nf6
followed by ... Nf4 gives Black a brutal attack, more than compensating for
the two pawns.
 15 ... Rf7
Not 15 ... a6?, since 16 cxd6! axb5 17 dxc7, as in F.Berkes-M.Pavlovic,
Serbian Team Championship 2009, is an extremely dangerous piece sacrifice
to meet. Such sacrifices crop up in many lines of the Classical Variation, so
it’s worth treating White’s mobile central majority with a lot of respect.
Black will be forced onto the defensive and have to find a lot of good moves,
whereas White has a great position and much easier play.
16 Ba5!?
An artificial-looking idea, but one with the purpose of inducing
weaknesses in Black’s queenside structure, while offering yet another piece
sacrifice along the way. Instead, 16 cxd6 cxd6 transposes to the 15 cxd6 line
above.
16 ... b6 17 cxd6 cxd6
Again, please don’t get involved in 17 ... bxa5? 18 dxc7, when White has
overwhelming compensation for the piece.
18 Be1
18 Bb4 Ne8 19 Rc6 looks active, but after 19 ... Bf8 20 a4 a6 21 Na3 Bd7
22 Rc1, White had been pushed back in G.Meier-S.Feller, French League
2012, and now I like 22 ... a5 23 Be1 Nf6, when Black proceeds with his
normal kingside counterplay.
18 ... a6
Here 18 ... g4 is an enterprising alternative, but I’ve decided to stick to
the main line.
19 Nc3 a5!
Otherwise Nb4-c6 would be unpleasant.

20 Bf2
Grischuk had defended this variation previously:
a) 20 a4 Bf8 21 Nf2 h5 22 h3 Rg7 23 Nb5 Nh4 24 Rc3 Rb8 25 Qc2
(otherwise it’s hard to see how White can make progress, but now Black
executes his standard break, equalizing immediately) 25 ... g4! 26 fxg4 hxg4
27 Nxg4 Nxg4 28 Bxg4 Bxg4 29 hxg4 Rxg4 30 Bxh4 Qxh4 31 Rff3 Qe1+ 32
Rf1 Qh4 33 Rff3 Qe1+ 34 Rf1 Qh4 and a draw was agreed in A.Giri-
A.Grischuk, Monte Carlo (blindfold rapid) 2011.
b) 20 Nb5 Bf8 gives White a choice of squares on the c-file for his rook.
For example: 21 Rc4 (or 21 Rc3 g4 22 Qc2 gxf3 23 Bxf3 Ng4 24 Bf2 Nh4 25
Bxg4 Bxg4 26 Bxh4 Qxh4 27 Ne1 Qh5 28 Nf3 and a draw was agreed in
H.Banikas-T.Radjabov, European Team Championship, Porto Carras 2011)
21 ... h5 22 Qc2, and now 22 ... Ba6 has been tested in a couple of GM
games, including E.Postny-E.Bacrot, European Cup, Rogaska Slatina 2011
(0-1 in 42), but I struggle to make the kingside attack work with the bishop
on this diagonal. Instead, I would suggest 22 ... g4 with good counterplay. In
some lines Black can play ... g4xf3 and ... Ng4 (as in the Radjabov game),
while if White tries to neutralize the light-squared bishop with 23 Nc7 Rb8
24 Ne6!?, the position is unclear after 24 ... Qe8.
20 ... Rb8 21 a4
G.Meier-R.Polzin, Austrian League 2010, saw 21 Nb5 Bf8 22 Rc6, and
now I like 22 ... g4 with similar play to the main game. An immediate ... g5-
g4 can be considered whenever the white knight has moved from c3.
21 ... Bf8 22 Nb5 g4 23 Rc6

The engine strongly prefers White’s position, just like in every main line
of the Classical Variation! To me, Black’s kingside threats look at least as
real as White’s play on the queenside.
23 ... Rg7 24 Qc2 Bd7
It’s important to understand that this was a reaction to the threatened
Bxb6. Black has no interest in taking the exchange at this point, since this
would give up his most important attacking piece (the light-squared bishop).
25 fxg4
Ragger correctly didn’t have any appetite for 25 Rc1 g3! 26 hxg3 Nh5 27
g4 Ng3, when the white kingside is collapsing.
25 ... Nxg4 26 Bxg4 Bxg4 27 Ne1 Nh8!
This is a typical idea when Black wants to prepare ... g5-g4, since the
knight unleashes the g7-rook. It makes a lot of sense here too, since the
knight can cover the d6-pawn from f7, or else go on to h6 to support the g-
pawn push.
27 ... Be7 is also logical and might be safer, since the knight will cover
important light squares (especially e6) from f8.
28 Kh1 Nf7 29 h3 Bd7 30 Nf3 Ng5!
To reiterate the point, I don’t think a King’s Indian expert of Grischuk’s
class would even consider 30 ... Bxc6?? 31 dxc6, since the extra exchange is
no comfort against White’s queenside dominance, especially as there’s no
kingside attack any more.
31 Bh4 h6
 32 Nxd6!?
Changing the character of the game. 32 Rc1 would keep the position
sharp and unclear.
32 ... Bxc6!
Effectively forced.
33 Nf5 Be8!?
Again demonstrating an instinct to keep his light-squared bishop (I have
no doubt that Grischuk was down to seconds at this point).
Instead, 33 ... Rc7! was reasonable since Black collects the c6-pawn after
34 dxc6 Qf6 35 Rc1 Rbc8, for example 36 Bxg5 hxg5 37 Qb3+ Kh8 38 Qxb6
Rxc6 39 Rxc6 Qxc6 40 Qb3! and the game is balanced. Indeed, the knights
are dangerous as long as White can keep the queens on, so a repetition with
40 ... Qc4! 41 Qb6! Qc6! is indicated.
34 Nxh6+ Kh8 35 Nf5 Rc8 36 Qd3 Rg8 37 Nxe5
Here 37 d6 looks more dangerous: 37 ... Bxd6 38 Nxd6 (38 Bxg5 Be7 is
fine for Black, while 38 Qxd6 Qxd6 39 Nxd6 Rc6 is only dangerous for
White) 38 ... Rg6 39 Rd1 Rxd6 40 Qxd6 Qxd6 41 Rxd6 Nxf3 42 gxf3 Kg8
43 Rxb6 Bxa4, which is a tough position to assess since all the pawns on the
e- and f-files are vulnerable. Black is probably okay, but certainly only White
can win, so 33 ... Rc7 was probably a safer option.
37 ... Qf6?
This gives White a chance; 37 ... Qc7 was better, with a very sharp
position.
Time for another exercise!

Exercise: What resource did both players miss?

38 Nf3?
Missing a great opportunity.
Answer: 38 Bf2!! was extremely strong, transferring the bishop to the
long diagonal. After 38 ... Bc5 39 Bd4 Bxd4 40 Ng4!, Black is suffering.
38 ... Bd7 39 Qd2?
39 d6! is still unclear.
39 ... Bxf5 40 Qxf4 Bh6?
 40 ... Qg7! 41 Nxg5 Be7! is good for Black.
41 exf5 Rcd8 42 Ne5?
And here 42 Rd1! levels the chances.
42 ... Rxd5 43 Ng4 Qd6 44 Qxd6 Rxd6 45 Bg3 Rc6 46 Be5+ Kh7 47
Nf6+ Rxf6 48 Bxf6 Ne4! 49 Be5

White has three connected passers for a piece, but they aren’t sufficiently
advanced to create threats. Grischuk converts confidently.
49 ... Bg7 50 f6 Re8! 51 fxg7 Rxe5 52 Rf3 Nc5 53 Rf6 Re6 54 Rf4 Re4
55 Rf6 Re6 56 Rf4 Kxg7 57 Rg4+ Kf7 58 Rf4+ Ke7 59 Kh2 Re4 0-1

Game 11
L.Ftacnik-E.Bacrot
German League 2015

In this game we examine White’s attempts to gain some kingside space with
f2-f3 and g2-g4. It might look like White is playing on the side where he is
weaker, but these systems have proved to be viable, if not as critical as the
lines where White immediately tries to obliterate Black’s queenside.
1 Nf3 Nf6 2 c4 g6 3 Nc3 Bg7 4 e4 d6 5 d4 0-0 6 Be2 e5 7 0-0 Nc6 8 d5
Ne7 9 Ne1 Nd7 10 f3
White can also try a related system with 10 Be3 f5 11 f3 f4 12 Bf2 g5 13
g4 Ng6 (White is in time to reinforce the kingside after 13 ... fxg3 14 hxg3
Ng6 15 Ng2, when an ... h5-h4 push will be met by g3-g4 and bringing the
knight to f5 via e3) 14 Nd3 h5 15 h3 Rf7 16 c5 (in his chess24 video lectures
Van Kampen gives 16 Kg2 Bf8 17 b4 Rh7 18 Rh1 a5 19 a3 b6 with good
play for Black, for example 20 c5 axb4 21 axb4 Rxa1 22 Qxa1 dxc5 23 bxc5
bxc5 and the bishop comes to d6 with a compact position) 16 ... hxg4 17
hxg4 dxc5 18 Nxc5 Nxc5 19 Bxc5 Bf8 and Black is fine. After 20 b4 Rh7 21
Kg2, as in Le Quang Liem-T.Radjabov, Beijing (rapid) 2013, Van Kampen
points out that 21 ... Bxc5! 22 bxc5 Qf8 would give Black a huge advantage
based on the combined pressure on the c5-pawn and the h-file.
10 ... f5 11 g4

Question: This looks like madness! Isn’t White horribly weakening his
own king and playing on the side of the board where he’s weaker? Is he
trying to checkmate the black king?

Answer: We can see from other variations of the Classical that, in the
“race” lines where White plays on the queenside and Black pushes his pawn
to g3, the white king has very few squares and it’s hard to bring defenders to
the kingside since White has so little space. The general idea behind g2-g4 in
the Classical is to gain more space so as to make it easier to defend the king.
A related and very important idea is that, if the kingside becomes
blocked, Black will have no counterplay and White will simply win the game
on the queenside.
11 ... Nf6
Another option is 11 ... Kh8!?, followed by ... Ng8 in many lines, but I
am advocating an approach played by Bacrot, Ding Liren and other
specialists: Black puts his knight on f6, plays ... c7-c6, and then fianchettoes
his light-squared bishop. This can sometimes be achieved by an immediate ...
b7-b5, but I prefer the more restrained approach in this game (and the notes)
of starting with ... b7-b6 and ... Bb7, followed by ... a7-a6 and ... b6-b5 in
many lines. Once the tension between the pawn chains is established, tactical
possibilities can easily arise, and it is notable that the white king is typically
less safe than his black counterpart in these variations.

Question: What do you think of 11 ... f4 in this position?

Answer: It’s a huge strategic mistake. After 12 h4, followed by Ng2,

Black can’t generate meaningful play on the kingside since ... h7-h5 will
always be met by g4-g5, and ... g6-g5 will be met by h4-h5. Blocked
structures are acceptable when Black has a pawn on g3 and can land a ...
Bxh3! sacrifice, but there are no tactics working in Black’s favour here, so he
has nothing to oppose White’s natural attack on the queenside.
12 Nd3 c6 13 Be3 Kh8
 A useful move, vacating g8 for a rook or a knight while moving the king
to a safer square.
14 a4
This physically prevents ... b7-b5, though as we will see Black often
plays ... b7-b6 against other moves anyway. This isn’t a position where tempi
are hugely significant. In fact, that is one reason I dislike White’s position –
Black has a simple strategy (put the bishop on the long light diagonal, create
tension between the pawn chains, and land a tactical shot), whereas White
doesn’t have a clear plan. Although he controls more space, he doesn’t have
any obvious pawn breaks.
Other moves:
a) 14 Rc1 b6 15 Kh1 Bb7 16 g5 Nh5 17 Qd2 a6 18 a4 Rb8 19 Rfd1 b5 20
axb5 axb5 21 Nb4 (21 b3 with equality looks more sensible) 21 ... bxc4 22
dxc6 was played in Le Quang Liem-S.Feller, Lubbock 2011, and now I like
22 ... Nxc6 23 Nxc6 Bxc6 24 Qxd6 Qc8, when Black’s pressure on the long
diagonal and excellent pieces fully compensate for his pawn deficit.
b) 14 Kh1 b6 15 Rg1 (after 15 h3 Bb7 16 Kh2 b5 17 b3 cxd5 18 cxd5
Rc8 19 Qd2 a5 20 g5, Black has a typical positional sacrifice in these lines:
20 ... Nfxd5!? 21 Nxd5 Nxd5 22 exd5 f4 23 Bf2 e4 24 fxe4 f3 25 Bxf3 Rxf3
and the complicated sequence left Black with excellent compensation for his
pawn deficit in S.Feller-M.Vachier Lagrave, French Championship, Nimes
2009; ½-½ in 37) 15 ... Bb7 16 g5 Nh5 17 Qd2 was Le Quang Liem-Ding
Liren, Olongapo City 2010. Here Black put his queen on d7, which I don’t
like since it makes it harder to advance the queenside pawns. I’d prefer 17 ...
Rc8, followed by ... a7-a6 and ... b6-b5, with complex play in which Black’s
chances are not worse.
14 ... b6 15 Kh1 Bb7 16 Nf2 Rc8 17 Rc1 Ba8 18 Rc2 a6 19 b3 b5 20
axb5 axb5 21 dxc6 bxc4 22 Bxc4 Rxc6

Both other recaptures on c6 are also reasonable, but Bacrot plays this
game very well and I have no reason to criticize his choice.
23 Ra2
It seems more logical to target the d6-pawn with 23 Rd2, although 23 ...
Qb8 leaves the position quite unclear. The engines claim an edge for White
after 24 gxf5 gxf5 25 Bg5, but I’d prefer Black’s game based on the
possibility of a tactical accident on the long diagonal.
23 ... Rc8 24 gxf5 gxf5 25 Ra6?!
Targeting the d6-pawn like this has a significant disadvantage compared
with Rd2.
25 ... d5!
Of course. There’s nothing good about White’s position, and his weaker
king and pawn structure remain problematic.
26 exd5 Nexd5 27 Nxd5 Nxd5 28 Rxa8
The best chance, but the tactics work out fine for Black.
28 ... Nxe3 29 Qxd8 Rcxd8 30 Rxd8 Rxd8 31 Re1 Bh6

Question: How would you assess this position?

Answer: I would previously have thought that such a position (with

reduced material and opposite-coloured bishops) has strong drawing
tendencies. However, my assessment now is that White has a horribly
unpleasant defence. His structure is weak, his king remains a problem, and
the opposite-coloured bishops mean that Black will be able to avoid
exchanges. It’s instructive to see how Bacrot, whose technique is immaculate,
wins the game against experienced GM opposition in just over 10 moves.
32 Nh3?!
The knight has no prospects on g3 (which is where it is headed). 32 Bb5
was a more stubborn defence.
32 ... Kg7 33 Ng1 Rd2 34 Ne2 Kf6 35 Ng3 Bf4
It’s interesting that, on each of the past few moves, computers prefer to
take the bishop on c4. However, Bacrot doesn’t want to change the character
of the position by removing the opposite-coloured bishops (see the note to
Black’s 31st move).
36 Rg1
Here 36 Ne2 was better, though Black retains a very significant
advantage.

36 ... Bxg3!
Winning a pawn and the game.
37 hxg3
 Or 37 Rxg3 f4 38 Rg8 Nxc4 39 bxc4 Rf2 and Black wins easily, since the
connected passed e- and f-pawns will create both queening and mating
threats.
37 ... Rf2 38 Ra1 Rxf3 39 Ra7 Rxg3 40 Rxh7 Nxc4 41 bxc4 Rc3 42
Rc7 e4 0-1
This position is a simple technical exercise for someone of Bacrot’s class,
so Ftacnik resigned.

Game 12
B.Gelfand-H.Nakamura
World Team Championship, Bursa 2010

This game is one of my favourites in the book. In team events, I have often
felt a certain pressure to play more conservatively than usual. In particular,
with Black against a strong opponent, a draw is normally considered a good
result in such situations. I love the bravado with which Nakamura selects not
only the King’s Indian, but one of its most complex branches, and the flair
with which he comprehensively crushes one of the world’s best players.
1 d4 Nf6 2 c4 g6 3 Nc3 Bg7 4 e4 d6 5 Be2 0-0 6 Nf3 e5 7 0-0 Nc6 8 d5
Ne7 9 Nd2
This is an older main line, and was recommended by GM Mihail Marin
on a recent DVD.
As we will see, I am recommending one of the most ambitious and
chaotic approaches in this position, so we can hardly be accused of taking the
easy option. However, a major advantage of the system I’m proposing is that
it can also be played against the Bayonet Attack, 9 b4. Here Black’s main
lines start with 9 ... Nh5, and there are good alternatives in 9 ... a5 (which has
worked out well for me, including in the Olympiad) and 9 ... c6!?. Instead,
we will look at 9 ... Ne8, when play is likely to transpose to 9 Nd2 lines.
 a) 10 Nd2 is popular and directly transposes to the main game;
b) 10 c5 f5 11 Nd2 also transposes, while 11 a4 Nf6 12 Nd2 is the note
with 12 a4 below. 11 Ng5 is probably the most significant independent
option, when 11 ... h6 12 Ne6 Bxe6 13 dxe6 fxe4 14 cxd6 Nxd6! shows an
important advantage of having the knight on e8. White has nothing in this
position.
c) 10 a4 f5 11 a5 (Avrukh gives 11 Ba3 Nf6 12 Nd2 Bh6 as okay for
Black, while after 12 Ng5 h6 13 Ne6 Bxe6 14 dxe6 fxe4 15 b5, “White has
compensation for the pawn, but the position is very complicated”) 11 ... Nf6
12 Bg5 is another independent option. Black’s simplest response is 12 ...
Nxe4 13 Nxe4 fxe4 14 Nd2 h6 15 Bxe7 Qxe7 16 Nxe4 Bf5 17 Bd3, as in
I.Farago-F.Hujbert, Budapest 2013. Now I like 17 ... Rf7, improving co-
ordination, when Black can follow with ... Raf8 and ... h6-h5. The position is
simplified and Black doesn’t have weaknesses, so I think he has a reasonable
game.
Apart from the main lines of 9 Ne1, 9 Nd2 and 9 b4, White has tried
about 20 moves at this juncture, many of which make little or no sense, but a
couple are worth considering:
a) 9 Bg5 is a pet line of Swiss GM Yannik Pelletier. It’s also a favourite
move of Ireland’s FM Colm Daly, who has used it to achieve excellent
positions, including against Mark Hebden. White intends to take the f6-
knight after ... h7-h6, but I prefer going after the other bishop with 9 ... Nh5
10 Ne1 Nf4 11 Nd3 Nxe2+ 12 Qxe2 h6 13 Be3 f5 14 f3 f4 15 Bf2

The following is a model game for Black in this line: 15 ... g5 16 c5 h5 17

Rac1 Bd7 18 a4 a6 19 b4 Ng6 20 Nb2 g4 21 cxd6 (21 Kh1 is a useful move,
with unclear play) 21 ... cxd6 22 Nc4?! (I’m not sure if this was a sacrifice or
a blunder) 22 ... gxf3! 23 gxf3 (23 Qxf3 Bg4 is horrible for White) 23 ... Bh3
24 Kh1 Bxf1 25 Qxf1 Rc8 reaches an unclear position, but one where
White’s compensation for the exchange wasn’t obvious in P.Prohaszka-
Z.Almasi, Hungarian Championship, Szeged 2009 (0-1 in 51). I like the
directness of Almasi’s play in this example, pushing ... g6-g5 and ... h6-h5 in
thematic fashion.
b) 9 Kh1 was a favourite of Tony Miles. After 9 ... Nd7 10 g4 Kh8 11
Rg1 a5 12 Be3 Nc5 13 Rc1 Bd7 14 b3 Ng8 15 Nd2 f6, preparing ... Bh6,
Black had a decent game in Ma.Carlsen-T.Radjabov, Biel 2006 (½-½ in 30).
9 ... Ne8 10 b4 f5

11 c5
11 f3 seems a slightly less flexible move order, as after 11 ... f4 12 a4 g5
13 c5 Ng6 14 Nc4, Black can either transpose to the main game with 14 ...
Nf6 or continue with 14 ... h5.
11 ... Nf6
Black has experimented with 11 ... Kh8, but I prefer the more direct
approach, which aims to force f2-f3 so that the kingside pawns can be set in
motion. The knight also controls the g4-square, supporting the advance of the
g-pawn while preventing White from meeting ... f5-f4 with a disruptive Bg4.
12 f3
Here 12 cxd6 cxd6 13 Ba3 Bh6! is an idea worth bearing in mind,
activating the bishop on the c1-h6 diagonal. If White plays 14 exf5, as in
E.Postny-I.Nataf, Portuguese League 2006, I like 14 ... Nxf5, when the knight
establishes itself on d4 and Black is no worse.
The main alternative is 12 a4, although Black can establish a dangerous
pawn roller even if White doesn’t play f2-f3. For example, 12 ... f4 13 Nc4
g5 and now:

a) 14 Ba3 g4 15 cxd6 cxd6 16 b5 was tested in the heavyweight

encounter V.Kramnik-H.Nakamura, Khanty-Mansiysk Olympiad 2010. Now,
instead of Nakamura’s 16 ... Ne8, subsequent practice has focused on 16 ...
f3!? 17 gxf3 (17 Nxd6!? fxe2 18 Nxe2 is a piece sacrifice essayed by GMs
Melkumyan and Fier; White won both games, but neither featured the best
response: 18 ... Nh5!, when the knight is coming to f4 with very unclear play)
17 ... gxf3 (not 17 ... Ne8? 18 fxg4 and White is just winning, M.Taimanov-
R.Klovsky, USSR Team Championship 1968; 1-0 in 39) 18 Bxf3 Ne8
(despite what the computer thinks, it seems to me that Black has reasonable
compensation here) 19 Bg2 (19 Bg4 loses time, since Black no longer needs
the light-squared bishop, and after 19 ... Bxg4 20 Qxg4 Rf6 21 Qe2 Ng6,
Black is no worse; 19 Kh1 is the engine’s top line, but 19 ... Bh3 20 Rg1 Rc8
21 Be2 Ng6 sees White a long way from establishing control and all three
results are possible) 19 ... Ng6 20 Ne2 Nh4 21 f3 Bd7 22 Rf2 Rf6 and with
rooks coming to g6 and c8, Black had good compensation for the pawn in
C.Delizia-U.Beyer, correspondence 2012 (½-½ in 49).
b) 14 Nb5!? is a pet line of the very strong Hungarian GM, Ferenc
Berkes. Black has two options here and, frankly, I didn’t find clear equality
in either of them, despite several hours of work. I present my conclusions
here so you can choose which one you prefer:
b1) 14 ... Nxe4!? is the forcing option: 15 f3 (15 Bd3 Nxf2! 16 Rxf2 e4 is
fine for Black) 15 ... a6 16 Nxc7 Nc3 17 Qe1 Nxe2+ 18 Qxe2 Qxc7 19 cxd6
Qd7 20 dxe7 Qxe7.

A long, forced sequence has resulted in a position where the main

question is whether White’s d5-pawn is strong or weak.
b11) 21 Bb2 Bf5 22 Rfe1 b5 23 Na5 was played in R.Buhmann-
P.Sadilek, Vienna 2014, when Black should have grabbed the pawn: 23 ...
Qxb4! 24 Nc6 Qc5+ 25 Kh1 Rfe8 with balanced chances. Instead, 23 axb5
axb5 24 d6 is critical, though after 24 ... Qe6 25 Nxe5 Qxd6 26 Qxb5 Rab8
27 Qc4+ Be6 28 Qd4 Rfd8 Black’s bishop pair and active pieces balance the
pawn deficit, especially with such limited material remaining. I think this is
likely to end in a draw.
 b12) 21 d6! Qe6 22 Bb2 Bd7 23 Rfe1 was played in F.Berkes-I.Nyzhnyk,
Bad Wiessee 2013. Here Nyzhnyk opted for 23 ... Rae8 and went on to lose a
long game (1-0 in 91). His defence was strengthened in a subsequent
correspondence game: 23 ... Rad8 24 Qd3 (24 Bxe5 is also possible: 24 ...
Bxe5 25 Nxe5 Qxd6 26 Nxd7 Rxd7 27 Rad1 Qb6+ 28 Kh1 Rxd1 29 Rxd1
Rd8 and White retains some pressure in this endgame, but again I think Black
should hold with basic care) 24 ... Rf5 25 Re2 Be8!? (an interesting defence,
initially disregarded by the engines which are very optimistic about White’s
chances of success in the arising rook endgame; 25 ... b5 26 Na5! bxa4 27
Nb7 Rb8 28 Nc5 is slightly better for White) 26 Rae1 Bg6 27 Nxe5 (the only
try for an advantage) 27 ... Bxe5 28 Bxe5 Rxe5 29 Rxe5 Qxe5 30 Qxg6+
hxg6 31 Rxe5 Rxd6 32 Rxg5 Kg7. This was rather surprisingly agreed drawn
in M.Dard-S.Zemlyanov, correspondence 2014. Black must have some
drawing chances, despite being a pawn down, with an active rook and very
limited material, although I lost a rapid training game against Fritz rather
easily from this position. If pressed I’d say the chances are 50/50 between a
draw and a win for White.
b2) 14 ... Ne8 drops back again to cover d6 and aims for more standard
King’s Indian play:
 b21) 15 f3 Ng6 16 Bd2 Rf7 17 Be1 a6 18 Nc3 Bf8 19 Bf2 h5 20 Rb1 Rg7
with standard King’s Indian counterplay in G.Hill-W.Morrow,
correspondence 2014 (½-½ in 36).
b22) 15 Ba3 Ng6 16 Bg4 Nh4 17 Bxc8 Rxc8 gives Black excellent
hacking chances against the white king. See how quickly a strong GM lost as
White: 18 g3 a6 19 Nc3 Ng6 20 Qg4 Rf7 21 Na5 Nf6! 22 Qe2 (22 Qxg5 Qf8
is quite chaotic) 22 ... Rb8 (Black has no cause for complaint here, and White
overpresses) 23 b5 axb5 24 Nxb5?! dxc5! 25 Bxc5 f3 (a very unpleasant
move to meet) 26 Qc4?? (White had to find 26 Qxf3! Nd7 27 Qa3 Nxc5 28
Qxc5 b6 29 Nc6 bxc5 30 Nxd8 Rxd8 with another irrational position: Black
is a piece up but the a-pawn looks sufficiently dangerous to balance the
chances) 26 ... Qc8! (threatening an elementary mate with ... Qh3 and, if
necessary, ... Ng4) 27 Rfe1 Nf4! (not difficult for a player of Vallejo’s
calibre, but pretty nonetheless) 28 Re3 Qh3 29 Qf1 Ng4! 30 Qxh3 Nxh3+ 31
Kf1 Nhxf2 and White resigned in J.Werle-F.Vallejo Pons, Wijk aan Zee
2009.
b23) 15 Bg4! is the move which concerns me, as played in F.Berkes-
D.Popovic, Bosnian Team Championship 2012 (1-0 in 37). I think Black
should react with 15 ... Nf6!? (despite having just gone to e8) 16 Bxc8 Nxc8,
when the other knight takes over defensive duties, allowing Black to push
forward with ... g5-g4. To be honest, I’m concerned that the light-squared
bishops have been traded since, if Black inserts ... g4-g3 and h2-h3, he won’t
have a ... Bxh3 resource. However, it still looks like Black has reasonable
counterplay in this position.
12 ... f4 13 Nc4 g5

14 a4
Alternatively:
a) White can accelerate the b4-b5 break with 14 Ba3 Ng6 15 b5, although
it looks slightly odd to block the a-pawn in this fashion. I like Vocaturo’s
treatment: 15 ... dxc5 16 Bxc5 Rf7 17 Qb3 g4 18 d6?! (White should have
taken on g4, but Black has decent counterplay in any event) 18 ... gxf3 19
Bxf3 Be6! and ... Bf8 is about to follow, when White is struggling to hold his
position together, F.Bruno-D.Vocaturo, Italian League 2014 (½-½ in 72).
b) 14 g4 h5 15 h3 is a conservative plan known to us from other lines of
the Classical. After 15 ... Rf7 16 a4 Ng6 17 Ba3 Bf8 18 Kf2 Rh7 19 Rh1 Be7
20 Qb3 Kg7 21 Raf1 Qh8, Black was fine in S.Mazé-E.Bacrot, French
Championship, Pau 2012 (½-½ in 49).
14 ... Ng6 15 Ba3
White can play more directly with 15 cxd6 cxd6 16 Nb5 (the d6-pawn is
now under attack, so I think it’s a reasonable idea for Black to play some
prophylactic moves) 16 ... Ne8 (I tried making the untested 16 ... g4 17
Nbxd6 g3 18 h3 Bxh3 19 gxh3 work for Black, but it doesn’t seem to pass
muster, in particular because of White’s Nf5 resource)

a) 17 Bd2 aims to bring the bishop to f2 (via e1), defending the king and
attacking Black’s queenside. Then 17 ... h5 has been tested in several games,
but I recommend 17 ... Rf7 18 Be1 Bd7 and now:
a1) 19 Bf2 Qb8!? (I like this approach of refusing to create targets, while
taking advantage of the absence of the c-pawns with the ... Bg7-f6-d8
manoeuvre; 19 ... Bxb5 20 axb5 b6 was tried in M.Gurevich-A.Shirov, Wijk
aan Zee 1993, but I think White should have the better chances here) 20 Rc1
Bf6 21 Qb3 Bd8 22 Na5 Nf8 23 Bc4 Bxa5 24 bxa5 g4 25 Be2 h5 26 Qc3 was
K.Troff-F.Perez Ponsa, Cappelle la Grande 2016, and now I suggest 26 ...
Rg7 with good counterplay. Black has achieved ... g5-g4 and is holding on
the queenside. Note that ... a7-a6 and ... Bxa4 might be considered in some
positions.
a2) 19 Na5 Qb8 20 Bf2 Bf6 21 Qb3 Bd8 22 Rfc1 Bxa5 23 bxa5 h5 (the
same plan, where Black trades his potentially passive dark-squared bishop for
the active a5-knight, and then proceeds with his kingside attack) 24 Qd1 Qd8
25 a6 bxa6 26 Nxa7

Exercise: How should Black continue?

Answer: 26 ... g4! (of course! – against moves like 27 Kh1 or 27 Nc6, the
black queen comes to g5, but what happens if White takes the pawn?) 27
fxg4? f3! (this double pawn sacrifice crops up occasionally, so it’s important
to be aware of it; here, in addition to his kingside attack, Black is playing
against the a7-knight, which relies on the f2-bishop for defence) 28 Bxf3 (28
gxf3 hxg4 29 Nc6 Qg5 gives Black a fearsome attack at the price of a mere
pawn) 28 ... hxg4 29 Bxg4 Rxf2! 30 Nc6? (this simply loses, though the
position was both bad and very difficult to play; 30 Bxd7 Qxd7 31 Kxf2
Qxa7+ was the best attempt, but the pair of knights are heavy favourites
against the rook and two pawns, since they can create threats easily against
the exposed white king) 30 ... Qg5 (30 ... Rxg2+! 31 Kxg2 Qg5 was
immediately terminal) 31 Kxf2 Bxg4 32 Qe1 Nf6 (allowing the dangerous
light-squared bishop to stay on the board hasn’t improved White’s chances –
the attack is irresistible and Maiwald conducts it well) 33 Rc3 Nf4 34 g3 Ne2
35 Re3 Nxe4+! (excellent calculation) 36 Rxe4 Rf8+ 37 Kg2 Bf3+ 38 Kh3
Qf5+ 39 g4 Qh7+ and White resigned in Y.Pelletier-J.Maiwald, German
League 2000, since 40 Qh4 Bg2+ 41 Kxg2 Qxe4+! mates quickly.
b) After the slower 17 a5 Black can again consider 17 ... h5, for example,
18 Bd2 Qd7!? (attacking the knight and preparing ... g5-g4) 19 Rc1!? (not
forced, but a good move leading to a complicated position) 19 ... Qxb5 20
Nb6 Qxe2 21 Qxe2 axb6 22 Qb5!? (Black finds it easier to untangle his
position after 22 axb6 Bd7) 22 ... bxa5 23 bxa5 g4!? with unclear play, as
occurred in L.Ftacnik-R.Biolek, Czech League 2016 (½-½ in 46). The
engines heavily favour White, but that might be a combination of two
computer biases: an underestimation of Black’s attacking potential in the
King’s Indian (even without queens!) and a preference for a queen over a
combination of minor pieces. It’s true that Black is currently struggling to
activate his forces, but he’ll have his share of the chances when the position
opens up. The rest of the game is not very instructive, and Ftacnik missed a
clear win towards the end, but at this stage I think all three results are
possible in practice.
Returning to the main game:
 15 ... h5
Nakamura crushed another expert of the Classical Variation, GM
Alexander Beliavsky, after 15 ... Rf7 16 a5 h5 17 b5 dxc5 18 b6 g4 19 bxc7
Rxc7 20 Nb5 g3 21 Nxc7 Nxe4 22 Ne6? (Beliavsky prioritizes neutralizing
Black’s light-squared bishop, but the knight on e4 is also a highly dangerous
piece; the engines prefer 22 fxe4! and after 22 ... Qh4 23 h3 Bxh3! 24 gxh3
Qxh3 25 Rf2!, I haven’t found a clear continuation of Black’s attack) 22 ...
Bxe6 23 dxe6 gxh2+ 24 Kxh2 Qh4+ 25 Kg1 Ng3 26 Bxc5 e4 27 Ra4 Rc8 28
Bxa7?? (this loses to a beautiful tactic; White could have maintained equality
with precise play) 28 ... b5!! 29 Rb4 (Black isn’t bothered by en passant,
since 29 axb6 Bd4+! ends the game) 29 ... bxc4 30 Bxc4 Qh1+ (the
immediate 30 ... e3 was even better, but Nakamura has seen a clear win and
goes for it) 31 Kf2 e3+ 32 Bxe3 fxe3+ 33 Kxe3 Nxf1+ 34 Bxf1 Qg1+ and
Black won in A.Beliavsky-H.Nakamura, Amsterdam 2009, though I’m
unsure whether Beliavsky resigned or lost on time, having had to defend a
horrible position for the past dozen moves. Anyway, despite nominal material
equality, there’s no way the white king is surviving this attack.
 16 b5
Here 16 cxd6 cxd6 17 Nb5 Ne8 18 Rc1 Rf7 19 a5 Bf8 20 Rc2 Rg7 21
Qd2 was A.Ho-C.Sandipan, World U-10 Championship, Bratislava 1993.
Black had regrouped thematically and well, and now I like 21 ... Bd7 22 Nc3
Rc8, when I prefer his position.
16 ... dxc5
Changing the structure. This takes advantage of White’s refusal to insert
c5xd6 over the past several moves. Black has also played 16 ... Ne8, but this
gives White options of 17 a5 or 17 b6 which I’d prefer to avoid.
17 Bxc5 Rf7 18 a5 g4

19 b6
The most direct option, immediately creating a confrontation on the
queenside. This was Marin’s recommendation on his DVD. Instead:
a) 19 Bf2 Bf8 20 Kh1 (after 20 b6, it’s a good moment to switch kingside
attacking plans with 20 ... gxf3!? 21 Bxf3 Ng4, as in Ki.Georgiev-
I.Ivanisevic, Sofia 2011; ½-½ in 41, when Black has attacking chances in an
unclear position, just like in all the other lines!) 20 ... Rg7 21 Ra2 g3! (Black
has prepared well for this thematic sacrifice) 22 hxg3 fxg3 23 Bxg3 h4 24
Bxe5!? (trying to change the character of the play and seize the initiative – it
doesn’t work, but 24 Bf2 Nh5 would have given Black a horrible attack
based on the threat of 25 ... Ng3+) 24 ... Nxe5 25 Nxe5 Nh5 26 d6 Ng3+ 27
Kh2 (27 Kg1 h3 is also fatal) 27 ... Bxd6, when White’s pieces (including his
king) were hanging and he faced a hopeless defence despite his extra pawn in
A.Korotylev-M.Mozharov, Moscow 2010 (0-1 in 59).
b) Korotylev repeated the line two years later, but didn’t fare any better:
19 Ra2 Bf8 (I’d prefer the immediate 19 ... g3 which is unclear) 20 Bxf8
Qxf8 21 Re1 Rg7 22 Kh1 Nh4 was A.Korotylev-A.Chernobai, Moscow
2012, which Black won quickly after the players traded errors: 23 d6? (White
should have entered complications after 23 Nxe5 Nxg2 24 Rg1!!, which the
computer assesses as 0.00; of course a decision like this is almost impossible
to take over the board), and now 23 ... gxf3! was strongest: 24 gxf3 cxd6 25
Qxd6 Bh3 26 Qxf8+ Kxf8 27 Rg1 Rc8 and Black’s attack continues despite
the queen exchange.
19 ... g3
Black was successful with 19 ... cxb6 20 axb6 g3 in A.Maksimenko-
D.Vocaturo, Bastia 2016 (0-1 in 39), but I don’t like giving White the option
of 21 Rxa7.
 20 Kh1
White has tried numerous alternatives:
a) 20 Nb5? Nd7! is already winning for Black, but a couple of decent
players have stumbled into this line with White: 21 Nxc7 (or 21 Nxa7 Nxc5
22 Nxc8 Rxc8 23 bxc7 Rcxc7 and White had only avoided mate at the cost of
a piece in A.Merry-M.Hebden, Jersey 2015; 0-1 in 51) 21 ... Qh4 22 h3 Nxc5
23 Nxa8 Bxh3! led to a rapid win for Black in Sa.Nikolov-T.Bonev, Sunny
Beach 2005 (0-1 in 29). Apparently this is a forced mate in eighteen,
culminating in 41 ... a6 mate, but I’m sure you didn’t need me to tell you that.
b) 20 Qd2 gxh2+ 21 Kh1 (White hopes to use the h2-pawn as a shield,
but Black has plenty of attacking options in any event; 21 Kxh2 Ng4+! 22
fxg4 Qh4+ 23 Kg1 hxg4 is just horrible) 21 ... h4 22 bxc7 Rxc7 23 Bf2 Nh5
24 Nb5 Rd7 25 Rfc1 h3! 26 Bf1 Nh4, when White had no answer to the
unusual attacking configuration on the h-file and Black went on to win in
R.Kazantsev-E.Lobanov, correspondence 2010 (0-1 in 43).
c) 20 Qc2 gxh2+ 21 Kh1 (21 Kxh2 Ng4+! 22 fxg4 Qh4+ 23 Kg1 hxg4
was great for Black in J.Fages-A.Podgursky, correspondence 2010; ½-½ in
35) 21 ... h4 22 d6? (not a good decision; White’s position was objectively
poor and very difficult to handle – still, it’s surprising that a correspondence
player would go for this) 22 ... cxb6 23 axb6 h3! is exactly what it looks like,
namely a winning attack: A.Krapivtsev-I.Telepnev, correspondence 2009 (0-
1 in 30).
d) 20 h3

Exercise: How would you respond as Black?

Answer: While taking on b6 is okay, I hope that you instinctively

snapped off the h-pawn: 20 ... Bxh3! 21 gxh3 Qd7 (21 ... Qc8! is more
precise, when 22 Rf2 can be met by 22 ... cxb6) 22 Qd2 (22 Rf2 was the best
defence, though still quite dire) 22 ... Qxh3 23 Bd3 Nh4 24 Nxe5 (this
position can be won in several ways, but I like Black’s choice) 24 ... Ng4! 25
Nxf7 (25 Nxg4 hxg4 is absolutely terminal) 25 ... Nf2 26 Rxf2 gxf2+ 27
Qxf2 Nxf3+ 28 Qxf3 Qxf3 and White soon resigned in C.Houriez-Y.Ouerk,
Condom 2010 (0-1 in 32).
 e) 20 bxc7 Rxc7 21 Bb4? (White had to give up the bishop, albeit for
insufficient compensation, since after ... ) 21 ... Nh7! 22 Bd3 Qh4 23 h3
Bxh3! ( ... history repeated itself) 24 gxh3 Qxh3 25 Ra2 Nh4 26 Rff2 gxf2+
27 Rxf2 Bf6 28 Rh2 Nxf3+ and White resigned in O.Delisau Gil-E.Mateu
Guiu, Catalonian Team Championship 2013.
20 ... Bf8
Black also has an extremely wide choice here, with five or six reasonable
moves being tested. We’ll pick one line and stick with it! Certainly 20 ... Bf8
strikes me as one of the most logical moves in the position – the c5-bishop is
perhaps White’s best piece and worth exchanging, plus the second rank is
cleared for the black rook.
21 d6
Instead:
a) 21 bxc7? Rxc7 22 Bg1, as in B.Ivanov-F.Kumic, Senta 2011, sells
White’s queenside advantage too cheaply and he now has nothing to oppose
Black’s kingside attack. Here Black has several promising options, but I think
it’s a good moment for 22 ... h4! 23 h3 (no prizes for guessing the next move,
but White can’t allow Black to play ... h4-h3 himself) 23 ... Bxh3! 24 gxh3
Qc8 25 Kg2 Nh7, which is pretty convincing, for instance, 26 Bf2 Ng5 27
Rh1 Rxc4 28 Bxc4 Qxc4 29 Be1 Kg7 and, despite his extra exchange, White
can hardly move.
b) 21 Bg1 is very logical, shoring up the kingside. Again Black has tested
almost every logical move in the position. My recommendation is 21 ... Nh4!
and then:
 Exercise: How should Black meet 22 Re1 - ?

Answer: 22 ... Nxg2!! is crushing: 23 Kxg2 Rg7 24 Nxe5 (or 24 Be3 Qd7
and, facing imminent mate, White resigned in R.Mesias Rojas-M.Arcos
Facio, Montevideo 2013) 24 ... gxh2+ 25 Kh1 Nxe4! and White resigned in
T.Roussel Roozmon-P.Charbonneau, Montreal 2008.
In a couple of correspondence games, White tried 22 Ra2 instead, but
Black didn’t find the 22 ... Nxg2! shot, when he is doing well in all
variations. Naturally, 23 Kxg2? Rg7 creates huge problems for White.
By analogy with the main game, I checked whether 22 hxg3 fxg3 23 Be3
was working for White.
 Question: How is Black doing in this position?

Answer: Extremely well, as long as he finds 23 ... Nh7!!. The black

queen is coming to h4 (after ... Nf5, for instance) and there’s no defence.

c) Question: What about 21 Bxf8 - ?

Answer: No good. 21 ... Nxe4! 22 Bc5 Nxc5 and White resigned in

M.Nenezic-O.Taichman, European Junior Championships, Albena 2011.
21 ... axb6 22 Bg1
We are still in Marin’s analysis. After 22 dxc7 Qxd1 23 Rfxd1 Bxc5 24
Nxb6 Bxb6 25 axb6, White’s compensation for the piece was insufficient in
V.Cvetnic-Ji.Houska, correspondence 2008 (0-1 in 47), since the queenside
pawn chain is vulnerable to ... Nd7xb6; while 22 axb6 fails to the simple 22
... Rxa1 23 Qxa1 (or 23 bxc7 Rxc7!) 23 ... cxd6 and Black was better in
A.Stella-I.Gaponenko, Forni di Sopra 2011 (1-0 in 34).
 22 ... Nh4
This ambitious move works out wonderfully in the game, but I’m not sure
about its objective merits.
I’m not wild about 22 ... gxh2 either, in view of 23 Bf2! (23 Bxh2 bxa5
24 dxc7 Qxd1 25 Rfxd1 Rxc7 26 Nb6 Rb8 was level in C.Gatto-E.Miciak,
correspondence 2012; 0-1 in 55) 23 ... bxa5 24 dxc7 (24 Nb5 cxd6 25 Bb6
was also good for White in R.Cerrato-P.Cutillas Ripoll, correspondence
2013; ½-½ in 42) 24 ... Qxd1 25 Rfxd1 Rxc7, which has been tested in
several correspondence games, but White seems better after 26 Nb5!.
I think Black should prefer the logical 22 ... Be6!, completing
development (a move which is not mentioned by Ftacnik or Marin). This
position needs to be tested further but initial correspondence praxis, and my
analysis, suggest that Black is doing fine. 23 dxc7 Qxc7 24 Nxb6 (24 Nb5
Qe7 25 Nxb6 Nh4 threatens ... Bh3, with a mess in B.Fraser-I.Malichev,
correspondence 2015 (½– ½ in 32); 24 Nd5 Nxd5 25 exd5 Bh3!! 26 gxh3
Nh4 27 Rf2 gxf2 28 Bxf2 Nf5 with an unclear position in which Black’s
chances are no worse in L.Forslöf-B.Westera, correspondence 2014 (½– ½ in
41)) 24 ... Qxc3 25 Nxa8 Bh3!! 26 gxh3 Nh4 Black is a rook down but the
threat of ... g2+ levels the chances, for example 27 Bb6 Rd7 28 Rc1 Rxd1 29
Rxc3 g2+ 30 Kg1 gxf1Q+ 31 Bxf1 Bb4 and the game is likely to end in a
draw.
23 Re1?
A logical move, preparing to shore up the kingside with Bf1, but also a
big error.
GM Ftacnik recommended 23 hxg3! (a very difficult move to find over
the board) 23 ... fxg3 24 Be3 Bh3 25 Rg1 Bxg2+ 26 Rxg2 Rg7! 27 Nd5 Bxd6
28 Nxd6 (Marin gives 28 Ndxb6! Rb8 29 Nxd6 cxd6 30 Kg1, which is even
more convincing) 28 ... Qxd6 29 axb6 Rxa1 30 Qxa1 Nxg2 31 Kxg2 Nxd5 32
exd5, correctly assessing the position as good for White. At the time his notes
were made, Ftacnik assessed this as only an edge, but modern engines are
more pessimistic about Black’s chances, since the bishop pair will quickly
create decisive threats.

Exercise: How should Black continue?

 Answer:
23 ... Nxg2!!
Exposing the white king and setting off a stunning tactical sequence.
24 dxc7?
White should have tried 24 Kxg2, although his position remains
uncomfortable after 24 ... Rg7 even with computer analysis, while over the
board this position is completely unplayable. The engine’s top line is 25 dxc7
Qe7 26 Ne3!! fxe3 27 hxg3 Qe6 28 Kh1! with a chaotic position which
seems to be dynamically balanced.
24 ... Nxe1
Perhaps Gelfand overlooked this simple but beautiful idea – White is two
queens up but mated by a lowly pawn after 25 cxd8Q g2 mate. As we will
see, Nakamura hasn’t finished with the queen offers in this game.
25 Qxe1 g2+!
Again the most precise, even better than 25 ... Qxc7 (which was also very
good for Black).
26 Kxg2 Rg7+ 27 Kh1 Bh3!!
 The same theme, but in a stunning second edition.
28 Bf1 Qd3!!
And again.
29 Nxe5 Bxf1 30 Qxf1 Qxc3 31 Rc1 Qxe5 32 c8Q Rxc8 33 Rxc8 Qe6
0-1
A fabulous game by Nakamura, and a wonderful advertisement for
Black’s chances in this utterly chaotic variation.

Game 13
R.Vazquez Igarza-H.Nakamura
Gibraltar 2016

1 d4 Nf6 2 c4 g6 3 Nc3 Bg7 4 e4 d6 5 Nf3 0-0 6 Be2 e5 7 Be3

 This system is named after the Yugoslav Grandmaster Svetozar Gligoric,
who was also known for his deep expertise in the King’s Indian as Black. The
line remains a viable attempt for an advantage. White postpones castling and
fortifies his centre.
The related line 7 0-0 Nc6 8 Be3 is less threatening but has been tried by
several strong players. In My 60 Memorable Games Bobby Fischer
recommended 8 ... Re8!, when Black is just fine. For example, 9 dxe5 (9 d5
Nd4! is the tactical point) 9 ... dxe5 10 h3 b6 11 c5 Bb7 is very solid for
Black, and after 12 Bc4, as in E.Miroshnichenko-R.Polzin, European Cup,
Ohrid 2009, I like 12 ... Na5 13 Bb5 (or 13 Qa4 Nxc4 14 Qxc4 Qd7 and the
queen comes to e6 with approximate equality) 13 ... Nxe4!? 14 Bxe8 Qxe8
15 Nxe4 Bxe4 with good compensation for the exchange in the form of an
extra central pawn and a strong bishop pair on an open board.
7 ... exd4!?
A good move, and a good opportunity for us to examine a new structure.
7 ... Ng4 has traditionally been the main line, but Nakamura has recently
preferred 7 ... exd4 and I think it’s equally good.
 Question: Can’t Black just play as in the Classical Variation with 7 ...
Nc6 - ?

Answer: No. This is effectively a sophisticated opening trap. Although

some IMs and GMs have played it as Black, 7 ... Nc6 can’t be recommended,
since after 8 d5 Ne7 9 Nd2! White has established an ideal set-up. He will
continue his queenside play with c4-c5 (supported by b2-b4 if necessary),
while Black’s kingside counterplay is too slow. A few games by GM Jan
Gustafsson (admittedly, against less experienced opposition) nicely illustrate
the advantages of White’s position compared with the main lines: 9 ... Nd7 (9
... Ne8 10 b4 f5 11 f3 Nf6 12 c5 fxe4 13 fxe4 h5 14 0-0 Ng4 15 Bxg4 Bxg4
16 Qb3 Bf6 17 Nc4, as in J.Gustafsson-R.Campos Alarcon, Torrelavega
2002, is a position which the engines already assess as completely winning
for White, since Black can’t hold the queenside together; 1-0 in 37) 10 b4 f5
11 f3 f4 (11 ... a5 12 a3 axb4 13 axb4 Rxa1 14 Qxa1 Nf6 15 Qa8 Kh8 was
J.Gustafsson-J.Perez Gonzalez, Madrid 2003, and now I like 16 0-0 when
White’s queenside progress contrasts sharply with Black’s kingside “attack”,
which has yet to start) 12 Bf2 g5 13 c5 Nf6 14 Nc4 Ne8 15 a4 and White was
miles ahead in his attack in J.Gustafsson-S.Johansen, Helsingor 2011 (1-0 in
28). See how hard it is to commit to an all-out attack against the king, when it
hasn’t committed to kingside castling yet!
8 Nxd4 Re8 9 f3 c6
 A useful move, preparing ... d6-d5.
10 Bf2
White has also tried:
a) 10 Qd2 d5 11 exd5 cxd5 12 0-0 Nc6 13 c5 allows a beautiful exchange
sacrifice introduced by Kasparov: 13 ... Rxe3!? 14 Qxe3 Qf8!.
Kasparov explains his idea as follows: “Apart from the obvious
superiority in control of the centre and the powerful dark-squared bishop
(hole at e3, weak white pawn at c5), a completely new factor has appeared in
the position – the black rook is ready to come into play via b8-b4-h4. But the
sum of these pluses can hardly outweigh White’s extra exchange, and the
position is one of approximate dynamic balance.”
Dozens of games have been played here, but the ideas are well
demonstrated by Kasparov’s initial two: 15 Nxc6 (or 15 Ncb5 Qxc5 16 Rac1
Qb6 17 Qf2 Bd7 18 Rfd1 Re8 19 Bf1 Bh6 20 Rc3 Nb4 with excellent
compensation for the exchange in B.Gelfand-G.Kasparov, Linares 1992; 0-1
in 60) 15 ... bxc6 16 Kh1 Rb8 17 Na4 Rb4 18 b3 Be6 19 Nb2 Nh5 20 Nd3
Rh4 21 Qf2 Qe7 22 g4
 Exercise: White has defended excellently but Black has a forced draw.
How should he continue?

Answer: 22 ... Bd4!! 23 Qxd4 (23 Qg2 Bxa1 24 Rxa1 Qf6 and ... Nf4 is
better for Black) 23 ... Rxh2+! 24 Kxh2 Qh4+ and a draw was agreed in
A.Karpov-G.Kasparov, New York (11th matchgame) 1990. This entire game
has been repeated, but the depth of Kasparov’s ideas is stunning.
b) 10 0-0 d5 11 cxd5 Nxd5! 12 Nxd5 cxd5 is a clean equalizer:
 b1) 13 Rc1 a5!? 14 Qb3 (this was played in one of my favourite King’s
Indian games; instead, 14 Rc5 dxe4 15 Bc4 Bxd4 16 Qxd4 Qxd4 17 Bxd4
Nc6 18 fxe4 Be6 19 Bf6 Bxc4 20 Rxc4 Re6 saw the complications fizzle out
to a balanced endgame in Xu Jun-I.Cheparinov, Gibraltar 2014; ½-½ in 38)
14 ... a4! 15 Qxd5 Qxd5 16 exd5 (White has relied on the attack on the c8-
bishop to discourage Black from taking on e3, but there are other ideas in the
position) 16 ... a3! 17 b3 (on 17 bxa3 Krasenkow gives 17 ... Bd7! 18 Nc2
Rxe3 19 Nxe3 Bd4, when Black gets a second piece for the rook, though I
think White should be able to hold this position) 17 ... Nc6!! (a gorgeous
developing move) 18 Nc2 Rxe3! 19 Nxe3 Nb4! (Black has pummelled his
opponent with a series of hammer blows: he’s now an exchange and a pawn
down, but the a3-pawn is going to cause serious damage) 20 Rc4 (Krasenkow
analyses 20 Rc7 as giving better drawing chances) 20 ... Nxa2 21 Ra4?!
(trading the opponent’s remaining rook is a typical device when an exchange
up, but here the a-pawn is going all the way) 21 ... Rxa4 22 bxa4 Bd4 23 Kf2
Nb4 24 Rc1 a2! (Black continues with the theme of queening his a-pawn, and
for the umpteenth time offers a material sacrifice) 25 Rxc8+ Kg7 26 Rc1
Nxd5 (a striking picture of domination) 27 Rd1 Nxe3 and White resigned in
J.Gustafsson-V.Kramnik, Dortmund 2012.
b2) 13 Qb3 can be met in similar fashion: 13 ... a5!? (more adventurous
than 13 ... Nc6 14 Rad1 Nxd4 15 Bxd4 Bxd4+ 16 Rxd4 Qf6, which is also
fine for Black) 14 Bb5 (14 Rad1 a4 15 Qb5 Ra5 gives Black a very active
game) 14 ... Bd7 15 a4 Nc6 16 Qxd5 Nxd4 17 Qxd7 Qxd7 18 Bxd7 Red8 19
Bh3 Nc2 with a favourable endgame for Black in D.Fridman-Rau.Mamedov,
European Team Championship, Reykjavik 2015 (½-½ in 48).
10 ... d5 11 exd5 cxd5 12 0-0 Nc6 13 c5
Nakamura played another great game after 13 Nxc6 bxc6 14 Rc1?!, this
time against Bartosz Socko: 14 ... Bh6! 15 Rb1 Bf5 (Black has activated both
bishops with tempo; I’m not sure what Socko was aiming for, so perhaps he
mixed up his move order) 16 Bd3 d4 17 Bxf5 dxc3 18 Bd3 cxb2 19 Rxb2
Be3! 20 Bxe3 Rxe3

We have a classic situation where the black knight is more effective than
White’s light-squared bishop, in particular because the bishop is limited by
its own pawns and White has dark-squared weaknesses across the board. As
always in his King’s Indian games, Nakamura plays with a lot of energy to
exploit these strategic advantages: 21 Be2 Qe7! (of course Black isn’t
interested in a queen exchange yet, since the white king is in more danger
than its black counterpart) 22 Rf2 Re8 23 g3 h5! (thematic and strong: the h-
pawn advances to soften the white kingside structure and help generate
threats against the king) 24 Qd4 Rd8 25 Qf4 Kg7 26 Kg2 h4 (Black could
also proceed more deliberately with 26 ... Re5 followed by ... c6-c5 and ...
Rd4) 27 Qxh4 Qe5 28 Qf4 Qxf4 29 gxf4 Nh5 30 Bf1?! (at first glance it
looks like White should hold easily after 30 f5, but in fact Black retains a
strong initiative after 30 ... Nf4+ 31 Kg3 Rd4 32 fxg6 Re6! or 32 Bf1 gxf5;
nevertheless, perhaps this was a better chance, since in the game Nakamura
gets a superior structure at no cost) 30 ... Nxf4+ 31 Kg3 Nh5+ 32 Kg2 Rd1
and White eventually collapsed under sustained pressure against his king and
weak pawns, B.Socko-H.Nakamura, Italian League 2015 (0-1 in 54).
13 ... Bf8!?

This was the choice in 2016 of two of the greatest King’s Indian
specialists, Radjabov and Nakamura, so it seems a good idea to follow their
example. If readers need an alternative they can investigate the main line 13
... Nh5 or the speculative 13 ... Re5!?.
The bishop move is very logical, creating immediate pressure against the
c5-pawn which is at risk of being separated from the white forces. While it
might seem like White has a pleasant position with a solid blockade of the
isolated queen’s pawn, he has some structural imperfections himself. In
particular, while the f3-pawn controls important squares and limits the f6-
knight, it also creates some weaknesses on the g1-a7 diagonal. The c5-pawn
would be better placed on c2 or c3, since it is vulnerable to attack and, due to
its advance, the d5-pawn is passed (although White’s blockade is so solid that
it’s hard to imagine this pawn becoming a genuine queening risk).
14 Bb5
Other moves:
a) 14 Nxc6 bxc6 15 a3 Nh5! 16 Re1 Nf4 17 Bf1 Rxe1 18 Qxe1 Bd7 19
Qd2 Ne6 20 b4 Qf6 21 Rd1 h5 22 Ne2 a5 23 Nd4 axb4 24 axb4 Bg7 25 Nxe6
Qxe6 was level in A.Morozevich-A.Grischuk, World Blitz Championship,
Astana 2012 (0-1 in 46). Take note of the typical devices employed by
Grischuk, in particular ... Nh5-f4 and gaining space with both rook’s pawns.
b) 14 Kh1 Bd7 has also seen some high-level traffic: 15 Rc1 (after 15
Qd2 Rc8 16 Nb3 d4 17 Nxd4 Bxc5 18 Nxc6 Rxc6 19 Bxc5 Rxc5 20 Rad1
Qe7 21 Qd4 Qe5, the pieces were being hoovered off the board and a draw
was soon agreed in E.Inarkiev-T.Radjabov, Spanish League 2016, ½-½ in 38;
instead, 15 a3 Nh5 16 Qd2 Nxd4 17 Bxd4 was S.Sjugirov-P.Svidler, Russian
Championship, Moscow 2012, and now rather than 17 ... Qh4?!, which could
have created some problems with the defence of the d5-pawn, I prefer 17 ...
Bc6 18 Rad1 Qh4 19 Bf2 Qf4 and Black is no worse) 15 ... Rc8 16 Ndb5 Bf5
17 Na4 is an overambitious set-up, which was well met by 17 ... d4! 18 Nxd4
Nxd4 19 Bxd4 Rxe2 20 Bxf6 Qxd1 21 Rfxd1 Bc2 22 Rd4 Bh6 23 Ra1 Rce8,
when Black had more than enough compensation for the pawn, through his
bishop pair and more active pieces in L.Van Wely-I.Cheparinov, World
Rapid Championship, Dubai 2014 (½-½ in 56).
14 ... Bd7
 15 Nb3
The position after 15 Bxc6 bxc6 16 b4 is instructively assessed by GM
Postny: “So, White has given up his light-squared bishop, but stabilized the
structure and occupied the nice blockading spot on d4. On the other hand
Black’s development is also fine and his only weakness on c6 is difficult to
target. It’s remarkable that some engines prefer Black’s position. Of course,
White is the only side that can claim to have a slight advantage here.”
In his excellent notes to one of his own games, Postny gives a survey
from recent Olympiads where this variation was discussed. I like the
approach of Van Kampen who, as we will see throughout this book, has an
excellent knowledge and feeling for King’s Indian positions: 16 ... a5 17 a3
Nh5 18 Qd2 Qf6 19 Nce2 Ng7 20 Rae1, E.Postny-R.Van Kampen, Tromsø
Olympiad 2014, was an important game from the match between Israel and
Holland. Here Postny recommends the thematic 20 ... h5!?, saying “I would
consider this move as Black. The march of the h-pawn might provoke some
concession.”
15 ... Ne5 16 Bxd7 Qxd7 17 Qd4 Bg7 18 Rad1 Nc4 19 Nxd5 Qxd5 20
Qxd5 Nxd5 21 Rxd5 Bxb2

This position is equal – and too barren even for Nakamura to generate any
imbalances.
22 Rd7 b6 23 Re1 Rxe1+ 24 Bxe1 bxc5 25 Nxc5 Nb6 26 Rd3 Rc8 27
Bf2 Bf6 28 g4 Be7 29 Ne4 Rc1+ 30 Kg2 Rc2 ½-½

Game 14
J.Rowson-G.Jones
British Knockout Championship, London 2015

1 d4 Nf6 2 c4 g6 3 Nc3 Bg7 4 e4 0-0 5 Nf3 d6 6 Be2 e5 7 d5

 Question: What is this line?

Answer: This is known as the Petrosian Variation, named after the ninth
World Champion, Tigran Petrosian. “Iron Tigran” was known for his tough
positional style, restricting his opponent’s possibilities. He is also known for
perhaps the best insult of the King’s Indian Defence, saying that his wins
against it had fed his family for many years!
The line was also adopted by Kramnik in his early years, including two
outstanding wins against Kasparov. Indeed, Kramnik returned to it in a game
with Nakamura in the 2014 London Classic, where he posed huge problems
after 7 ... a5 8 Bg5 h6 9 Be3!? (instead of the traditional 9 Bh4).
The Petrosian Variation typically leads to complex strategic battles with
chances for both sides. My suggestion is one I picked up from a book by Joe
Gallagher several years ago, which is based on a neat tactic and, in my view,
simplifies Black’s task considerably.
7 ... Na6 8 Bg5
 This is by far the most popular move, even though it allows Black to
execute his idea.
Instead:
a) 8 Be3 transposes to a minor line of the Gligoric Variation (where it
would arise after 7 Be3 Na6 8 d5). White threatens to establish his ideal
formation with Nd2, so Black should target the e3-bishop to generate
counterplay; for example, 8 ... Ng4 9 Bg5 f6 10 Bh4 Nh6 11 Nd2 c5 12 f3
(12 a3 Qe8 13 Rb1 f5 14 f3 Nf7 15 b4 Bh6 16 Bf2 Qe7 17 0-0 b6 was similar
in Vi.Kiselev-A.Istratescu, Belgian League 2013; ½-½ in 101) 12 ... Nf7 13
a3 Bh6 14 Bf2 f5 15 h4 Bd7 16 Rb1 b6 17 b4 Qe7 and Black had a good set-
up in R.Vogel-R.Van Kampen, German League 2017 (½-½ in 55).
b) 8 Nd2 is Delchev’s suggestion in his White repertoire database for
Modern Chess, where he annotates his win against Hamdouchi (Spanish
League 2012), which continued 8 ... Ne8 9 h4 f5 10 h5 with some initiative
for White. However, he doesn’t mention my preferred move 8 ... h5!?, a
recommendation of Joe Gallagher and a recent favourite of Mark Hebden.
Gallagher’s comment is instructive: “Now there is no need to worry about
White advancing on the kingside (we got there first). And take a look at the
white position. What move do you feel like playing now? Nd2-f3, perhaps.”
Following 9 a3 (after 9 Nb3 b6 10 Bg5 Nc5 11 Nd2 a5 12 Qc2 Bd7 13 0-
0 Qe8 14 Rae1 Nh7 15 Be3 f5 16 exf5 gxf5 17 f4 e4, Black had an ideal
blockading position in M.Tratar-L.Valdes Romero, San Sebastian 2016; 0-1
in 46) 9 ... c5, Black stops b2-b4 for the moment and intends to meet an
eventual b-pawn push with ... b7-b6, holding the queenside:
 b1) 10 Nf1 Nh7 11 h4 Nc7 12 g4!? leads to chaotic play: 12 ... hxg4 13
Bxg4 f5 14 Be2 (after 14 exf5 gxf5 15 Be2 b5!? 16 cxb5 f4 and 17 ... Bf5, or
the immediate 15 ... f4, Black has good compensation in view of White’s
poor development and king in the centre) 14 ... f4 15 Bg4 Nf6 16 Bxc8 Rxc8
17 Qf3 Qe8 18 Nd2 b5 and Black had a full share of the chances in T.Hillarp
Persson-M.Hebden, Oslo 2015 (0-1 in 70).
b2) 10 Nf3 (an interesting illustration of Gallagher’s comment above) 10
... Nh7 11 h4 f6!? (not a standard move in the King’s Indian but, as we will
see from looking at the alternatives, it makes a lot of sense here: 11 ... f5 12
exf5! was played in S.Williams-M.Hebden, British Championship, Coventry
2015, when Black had to recapture with the bishop, allowing White an
enduring advantage thanks to his control of the e4-square – perhaps Hebden
had missed that the positionally desirable 12 ... gxf5? is immediately crushed
by 13 Ng5 with a winning position; previously, against Cummings in the
British League 1996, he had opted for 11 ... Nc7 but correctly avoided
repeating this move, probably in view of 12 Ng5! again, when White is
clearly better since taking the knight will be met by h4xg5, while 12 ... f6 13
Nxh7 Kxh7 14 g4 leads to an attack) 12 Be3 Nc7 13 Qc2 Bd7 14 Nd2 f5! 15
exf5 gxf5 16 f4 e4!? (17 ... Qe8 was also okay) 17 Bxh5 b5! and Black had
good play for the pawn, going on to defeat an experienced GM in
V.Ikonnikov-R.Voigt, German League 2015 (0-1 in 86).
c) 8 0-0 Nc5 9 Qc2 is David Cummings’ recommendation in his
repertoire book on the English. The critical line runs 9 ... a5 10 Bg5 h6 11
Be3 (Cummings correctly notes that Black is doing well after 11 Bh4 g5 12
Bg3 Nh5 13 Nd2 Nf4) 11 ... b6 (now Black can meet Bxc5 with ... b6xc5) 12
Nd2 Bg4 13 Qd1 (13 f3 Bd7 14 Rfe1 Nh5 15 Bf1 f5 gave Black typical
counterplay in B.Savchenko-I.Smirin, St Petersburg 2012; 0-1 in 47) 13 ...
Qc8 and now:

c1) 14 a3 Bxe2 15 Qxe2 a4 16 Bxc5 bxc5 17 Qd1 Qd7 18 Nb5 Nh7 19

Nb1 f5 20 N1c3 was T.Kriebel-M.Jurcik, Zillertal 2015, and here I like 20 ...
Ng5 21 exf5 (or 21 f3 fxe4 22 Nxe4 Nxe4 23 fxe4 Rxf1+ 24 Qxf1 h5 and the
bishop can come into play via h6) 21 ... Rxf5 22 Qe2 Raf8 and Black will
have good kingside counterplay with ... h6-h5 and ... Bh6 again. White can’t
comfortably take the a4-pawn since 23 Nxa4 e4! already gives Black a
decisive attack based on themes of ... Rf4-h4, ... Be5 and ... Nf3+.
c2) 14 Bxg4 Nxg4 15 Bxc5 bxc5 16 Nb5 is Cummings’ proposed line,
though he only examines 16 ... h5. Stockfish prefers 16 ... Nf6!, aiming to
bring the knight to f4 immediately, which seems to lead to decent play for
Black; for example, 17 Re1 Nh5 18 g3 (after 18 Nf1 Nf4, I don’t think Black
should be unhappy with his position) 18 ... Nf6, followed by ... Nh7 and ...
f7-f5, gives some kingside counterplay. Moving the pawn to g3 has slightly
weakened White’s formation and the knight might return to play via g5,
targeting the weak light squares around White’s king.
8 ... h6 9 Bh4
By contrast with the Kramnik-Nakamura game referred to above, 9 Be3
makes less sense here since ... Na6 is much more useful than ... a7-a5. After 9
... Ng4 10 Bd2 f5, Black already has an excellent game. Although this line is
almost never played, we have a nice example by Cheparinov: 11 exf5?
(White should settle for 11 0-0 fxe4 12 Nxe4 Bf5 13 Bd3 Nf6 with balanced
chances) 11 ... Bxf5 12 h3?
 Exercise: Black to play!

Answer: 12 ... Nxf2! (winning, and not a difficult tactic for someone of
Cheparinov’s level; White would have been fine had the knight retreated, but
now Black gets a huge attack) 13 Kxf2 e4 14 Kg3 (as the reader can check,
Black would win rapidly after all knight retreats) 14 ... exf3 15 gxf3 Nc5,
when White’s position was ruined and he soon lost a miniature in
P.Drenchev-I.Cheparinov, Plovdiv 2013 (0-1 in 25).
9 ... g5 10 Bg3

Exercise: See if you can spot Black’s tactical trick.

Answer:
10 ... Nxe4!?
Strong players have experimented with 10 ... Nh5, which also looks
possible, but I prefer the text.
11 Nxe4 f5
 The point. Black regains his piece, although it’s unclear whether he’ll be
taking one of the white knights or the bishop on g3.
12 Nc3
White elects to give up the bishop. 12 Nfd2 fxe4 13 Nxe4 Bf5 14 Bd3
Bxe4 15 Bxe4 Nc5 16 Qe2 Qd7 17 0-0 Nxe4 18 Qxe4 Qf5 19 f3 Qxe4 20
fxe4 g4! was comfortable for Black in T.Banusz-J.Viterbo Ferreira, European
Championship, Jerusalem 2015 (½-½ in 54).
12 ... f4 13 Nd2 Nc5 14 Nde4

Taking control of the e4-outpost before Black plays ... Bf5. 14 Nce4
Nxe4 15 Nxe4 transposes.
14 ... Nxe4 15 Nxe4 fxg3 16 hxg3 Bf5 17 Bd3 g4 18 Qd2
In an earlier game White castled straight away: 18 0-0 Qe8 19 Qb3 Qg6
20 Rae1 b6 21 Re3 h5 22 Nd2 Bh6 23 Bxf5 Rxf5 24 Re2 Bxd2 25 Rxd2 Raf8
26 Qe3
 Exercise: How can Black break through in this major piece position?

Answer: 26 ... h4! 27 gxh4 g3! puts White under severe pressure. After
28 Rd3 Gawain took with the pawn in E.Porper-G.Jones, Reykjavik 2014 (½-
½ in 71), but I prefer taking with the rook: 28 ... Rxf2 29 Rb1 R2f4! 30 Qxg3
Rg4 and White is struggling despite his extra pawn.
18 ... Qe8 19 0-0
David Howell tried much sharper 19 0-0-0!? in the next round in London.
Gawain’s resolute 19 ... b5?! 20 cxb5 a6 was well met by 21 b6!, when White
retained control and was winning after 21 ... Bxe4 22 Bxe4 Qa4 23 Qe2 Qxa2
24 b7 (1-0 in 40). Instead, I would probably try 19 ... a6 with the same idea of
pushing ... b7-b5, with unclear play.
19 ... Qg6
 Question: How would you assess this position?

Answer: At first glance it looks more comfortable for White, since his
knight is excellently placed on the e4-outpost and Black’s g7-bishop is
passive. However, I think Black is objectively at least equal and, more
importantly, his position is easier to play. He will continue with ... Rf7 and ...
Raf8, followed by ... h6-h5, after which he can activate his dark-squared
bishop by ... Bh6 or examine breakthroughs on the kingside with ... h5-h4.
The white knight on e4 is well placed but is also a tactical weakness since it
can’t easily be supported by a pawn; f2-f3 would open lines in Black’s favour
and, in many cases, drop the g3-pawn.
The assessment is informed, in part, by the present game, where White (a
very strong GM) loses largely without a fight.
20 Rfe1 Rf7! 21 Rad1 Raf8 22 Re2 h5
Inserting 22 ... b6 23 b4 doesn’t change much.
23 c5 b6 24 cxd6 cxd6 25 Qb4 Rd8 26 Rc2 Bf8 27 Rc6
 Rowson has manoeuvred his pieces to put maximum pressure on the d6-
pawn but lacks any pawn breaks to increase it. Meanwhile, Black can arrange
for ... h5-h4.
27 ... Rh7! 28 Qd2 Be7! 29 Qc2 Rf7
Not yet 29 ... h4? since 30 Nf6+ Qxf6 31 Bxf5 turns the tables, with a
clear advantage for White.
30 Rc7 Rdf8 31 Rxa7
White has won a pawn and has an active rook on the seventh rank, but
Black’s pieces are perfectly co-ordinated and his kingside attack is about to
land. Although I think the position is dynamically balanced, White needs to
be more careful.
31 ... Qh6!
Again 31 ... h4? is too early, as 32 gxh4 Bxh4 33 Rxf7 Qxf7 34 g3
reduces Black’s attacking potential.
32 Qc6?
A bad mistake, albeit a very understandable one. Instead, 32 Nc3 is the
computer’s suggestion, when Black maintains the balance with a dynamic
double pawn sacrifice: 32 ... h4! 33 gxh4 (or 33 Bxf5 Rxf5 and now 34 Ra8!
is equal, whereas 34 Rxe7?? hxg3 wins for Black) 33 ... g3! 34 fxg3 Qe3+ 35
Kh2 Bxh4 36 gxh4 Rxa7 37 Bxf5 Qf4+ 38 g3 Qxf5 39 Qxf5 Rxf5 and the
endgame is roughly level.

Exercise: How can Black gain the upper hand?

Answer:
32 ... Bxe4! 33 Bxe4 Rxf2!
Pushing the h-pawn first with 33 ... h4! is also strong.
34 Rxe7?
This just loses; 34 Qc3 was obligatory.
34 ... Qe3!
Setting up a fatal discovered check.
35 Kh2 h4??
Allowing a beautiful draw, which Rowson misses. Instead, 35 ... Qxe4
followed by ... h4 would have won easily.
 Exercise: How can White save the game?

36 gxh4??
This move reminds me of a profound comment by GM Tiger Hillarp
Persson in his notes to S.Mamedyarov-Mi.Adams, FIDE Grand Prix, Sharjah
2017 (1-0 in 43). Adams was under pressure for the whole game and missed a
draw (not dissimilar to the puzzle question) at move 30. For games in elite
events, followed by players with engines, this would typically result in online
ridicule, but Tiger was deeply sympathetic: “Chess is a very difficult game,
especially when you are short on time, and even more so when you believe
you are worse.”
Answer: 36 Rg7+!! Kxg7 (36 ... Kh8?? 37 Rxg4 even wins for White) 37
Qd7+ forces a draw by perpetual check.
36 ... Qxe4
Now White has nothing to oppose the attack on his king. His queen on c6
and rook on e7 are powerless to assist the defence.
 37 Rg1 Qf4+ 38 Kh1 Qg3 39 Qc3 Qxh4+ 40 Qh3 gxh3 41 gxh3+ Kh8
0-1

Game 15
A.Mirzoev-V.Kovalev
Turkish League 2016

1 d4 Nf6 2 Nf3 g6 3 c4 Bg7 4 Nc3 0-0 5 e4 d6 6 Be2 e5

In this game we look at one of the big worries facing an inexperienced
King’s Indian player.
7 dxe5

Question: Oh no! Isn’t this just a draw?

Answer: Black faces exchange variations in many openings, including the

Slav and the French. Of these, in my view the line where White runs the most
strategic risk (especially if he plays passively) is against the King’s Indian.
This is because Black can use the outpost on d4, whereas the d5-square can,
in the long run, be covered by ... c7-c6. If this were the only factor in the
position, White would be worse, but of course, the story is more complicated.
White gets some pressure and (often) a lead in development, so Black needs
to be precise in order to equalize. But once equality is achieved, there’s no
reason for Black to agree a draw – the positions are fully playable for both
sides, even if they are, admittedly, less exciting than the main lines of the
Classical.
We should also look at 7 0-0 Nc6 8 dxe5, which aims similarly for a risk-
free edge: 8 ... dxe5 (I prefer this to 8 ... Nxe5, which is also completely
playable) 9 Bg5 Bg4
 I’m unsure how to annotate this move. I wanted to mark 9 ... Bg4 as
“interesting”, but I’m not sure whether this is valid since it basically aims to
force a draw (even if Black has a small plus score in my database). However,
it seems to me like the best choice when compared to the alternatives.
First of all, Black can play the well-established 9 ... Qxd1 10 Rfxd1 Bg4
(10 ... Be6!? was tried recently by Dominguez Perez and could be worth
further investigation). Robin Van Kampen has pointed out that, in the
resulting positions, White has some small pressure and runs almost no risk of
losing. So this isn’t a great result.
Alternatively, there is the dynamic 9 ... Nd4!?, as played and analysed by
Van Kampen. This seems like an impractical choice to me, for different
reasons. 7 0-0 and 8 dxe5 is a line GMs bring out against each other when
they want to play without risk. Most club players aren’t even aware that this
line exists, and I think most readers of this book are unlikely to face it very
often, if at all. Learning a lot of theory for a rare (and generally non-
threatening) variation isn’t a good use of one’s preparation time, which ought
to be spent getting to grips with more serious systems like the Classical,
Sämisch etc.
So here’s the quick (and, I think, good) solution: 9 ... Bg4 10 Qxd8
(unlike the line with 9 ... Qxd1 10 Rfxd1 included, here 10 h3 isn’t just
innocuous, it’s almost dangerous for White who, after 10 ... Bxf3 11 Bxf3
Nd4 and ... c7-c6, is facing some strategic risk based on the fabulous knight
on the d4-outpost, while d5 is inaccessible to the white pieces; this type of
position can be balanced or even good for White if he can achieve c4-c5 and
b2-b4 quickly, but this version looks more than fine for Black) 10 ... Nxd8 11
Bxf6 Bxf6 (11 ... Bxf3 has also been played but is marginally less precise) 12
Nd5 Kg7! 13 Nxc7 (otherwise White is certainly no better and might risk
being worse) 13 ... Rc8 14 Nd5 Bxf3 15 gxf3 (if White wants a quick draw,
this was a fair result after 15 Bxf3 Rxc4 16 Rac1 Rxc1 in S.Matsenko-
M.Oleksienko, Ordu 2016) 15 ... Ne6 16 Kg2

16 ... Nf4+ (it’s also hard to see how White can win after 16 ... Rfd8, as in
V.Ivanov-V.Iordachescu, Moscow 1995, ½-½ in 36; however, I trust
Radjabov’s choices in such positions: sometimes they depend on deep
computer analysis which can’t be replicated over the board, but here, with
opposite-coloured bishops and one open file, I think Black’s play is quite
straightforward) 17 Nxf4 exf4 18 Rfd1 Rfd8 (Gawain Jones also used this
line to make an easy draw in a recent game: 18 ... b6 19 Rd7 Rfd8 20 Rad1
Rxd7 21 Rxd7 Bxb2 22 Rxa7 Rc5 23 Rb7 Bd4 24 Rd7 Rg5+ 25 Kf1 Bc5 26
a4 Rh5 27 Kg2 Rg5+ and the players repeated moves in N.Grandelius-
G.Jones, Reykjavik 2017) 19 Rd5 (this looks like the only realistic try, since
if all the rooks are traded down the d-file the position would be an immediate
draw) 19 ... Rxd5 20 exd5 b6 21 Rb1 a5 22 b4 (I just don’t see any ideas for
White) 22 ... axb4 23 Rxb4 Ra8 24 a4 Bd4 25 Bd1 Bc5 with an unbreakable
blockade in E.L’Ami-T.Radjabov, European Team Championship, Reykjavik
2015 (½-½ in 39). Radjabov is one of the main exponents of fortresses in his
opening preparation with Black (several games from the Schliemann Ruy
Lopez come to mind). The level of skill required to hold such positions
varies. This version seems pretty simple to me.
7 ... dxe5 8 Qxd8 Rxd8

9 Bg5
 Question: What is White’s threatening?

Answer: 10 Bxf6 Bxf6 11 Nd5, winning material. So Black has to take

evasive action.
The immediate 9 Nd5 is also possible: 9 ... Nxd5 10 cxd5 c6 11 Bc4
(after 11 Bg5 Black can transpose to the main line with 11 ... Re8, but 11 ...
f6 is probably simpler; e.g. 12 dxc6 Nxc6 13 Bc4+ Kf8 14 Be3 b6 15 Rc1
Ne7 and 16 ... Bb7 with equality) 11 ... cxd5 (Black has several options here
but I think this is simplest) 12 Bxd5 Na6, when 13 Bg5 Re8 14 0-0 Nb4 15
Bb3 Be6 16 Bxe6 Rxe6 17 Rfd1 f6 reached a level position in D.Schlecht-
A.Fedorov, Copenhagen 2004 (0-1 in 32).
Please note that 9 Nxe5 Nxe4! is less than nothing for White: 10 Nxe4
Bxe5 11 Bg5 Bxb2 12 Bxd8 Bxa1 13 Bxc7 Nc6 and Black is certainly no
worse, E.Ubilava-B.Adhiban, Benasque 2015 (½-½ in 22). This is a
completely standard tactic in the King’s Indian and we’ll see it in many
variations.
9 ... Re8
This is the main line and my recommendation. Black has alternatives if he
wants to create more imbalance, including 9 ... Rf8 and 9 ... Na6, but in doing
so he runs more risk in my view.
10 Nd5
White seizes the chance to change the structure. Instead, 10 0-0-0 Na6
(10 ... c6 11 Nxe5!? isn’t bad for Black, but we don’t need to allow this shot)
11 Ne1 Nc5 12 f3 c6 13 Be3 Ne6 14 Nc2 Bf8 was agreed drawn in T.Ravi-
S.Praneeth, Noida 2016. This is the type of position where I would begin to
become ambitious with Black, since the d4-square is weak and he has a
natural plan with ... Bc5. Although the position is objectively level, I think
White needs to be more careful.
10 ... Nxd5 11 cxd5
Now if Black doesn’t take immediate action he will be worse, since the
c7-pawn is backward.
11 ... c6! 12 Bc4 cxd5 13 Bxd5 Nd7!
The most precise. Black threatens ... Nf6 which would gain the bishop
pair.
14 Nd2
A good move, defending e4 so the bishop can retreat after ... Nf6.
14 ... Nc5
 15 0-0
After 15 0-0-0 Be6! (Black’s lack of development and the weakness of f7
means he shouldn’t get involved with ... Nd3+ and taking on f2) 16 Bxe6
Nxe6 17 Be3 Nf4 18 Bxf4 exf4 19 f3 f5, the game was agreed drawn in
J.Bartholomew-S.Nagle, Minneapolis 2014.
Instead, I faced 15 Nc4 in Mu Ke-S.Collins, Riga 2015; I responded very
poorly and was in deep trouble before salvaging a draw in a rook and pawn
endgame (½-½ in 61). I think it’s best to prevent the knight incursion to d6
with 15 ... Bf8 16 0-0 (16 Ke2 Ne6 17 Be3 Nc7 18 Rac1 Nxd5 19 exd5 b6 20
Rhd1 was another draw in N.Brunner-M.Ahn, Belgian League 2007) 16 ...
Be6 17 Rfd1 (after 17 Bxe6 Rxe6 18 f3 Rc8, Black was at least equal in
P.Certek-C.Zetocha, Slovakian League 2015; ½-½ in 47) 17 ... Rac8 (Bogner
previously faced 17 ... Bxd5, but I prefer Hunt’s approach) 18 Rac1 Bxd5 19
exd5 Ne4 20 Be3 b5 21 Na5 Rxc1 22 Rxc1 Nf6 23 Rd1 Rd8 24 Nc6 Rxd5 25
Rxd5 Nxd5 26 Nxa7 b4 and White had to play precisely to draw in S.Bogner-
A.Hunt, Hastings 2014/15 (½-½ in 40).
15 ... Be6 16 Bxe6 Nxe6 17 Be3
 Question: How would you assess this position?

Answer: To my mind, it’s a good example of how static factors can lead
to a misassessment of various King’s Indian positions. It’s tempting, for
instance, to say that the g7-bishop is “bad” (versus the “good” bishop on e3),
and that a white knight on d5 would be difficult to challenge, whereas a black
knight on d4 could be captured by White’s bishop.
However, the dynamics of the position are strongly in Black’s favour.
He’s ready to occupy the d4-square with his knight (while White is several
moves away from an invasion of d5), and taking this piece would leave Black
with a dangerous passed pawn and a dominant dark-squared bishop. The
black rooks also find work more easily than their white counterparts. Finally,
Black has the only active pawn break in the position, namely ... f7-f5
(whereas f2-f4 by White would weaken the long diagonal and the e4-pawn).
There’s little doubt but that Mirzoev (a 2525-rated GM) picked this as a
“safe” line, but he rapidly finds himself in a prospectless position. This
illustrates the strategic dangers of the Exchange Variation if White doesn’t
play energetically.
17 ... Rec8
Mirzoev agreed a draw after 17 ... Nd4 in his game with IM Ogulcan
Kanmazalp (who, incidentally, was outrated by over 250 points) in Istanbul
2010. I’m not sure what he was hoping for by repeating the variation.
18 Rfc1 Nd4 19 Kf1?!
Here 19 Rxc8+ Rxc8 20 Nf3 looks like a much sounder equalizing
attempt.
19 ... Nc2! 20 Rab1 f5

Black is playing more actively and has the initiative, even if White should
be okay with precise defence. Could we wish for a better opening outcome
than this?
21 Nb3 Nxe3+ 22 fxe3 f4!
Continuing in the same active King’s Indian style. White’s a long way
from his dream of a dominant knight on d5 against a passive dark-squared
bishop.
 23 Rxc8+ Rxc8 24 Rc1 Rf8!
It’s important to keep a pair of rooks, since rook and bishop tend to
outperform rook and knight and Black can create threats more easily with
more pieces on the board.
25 Ke2?!
The computer prefers to jettison a pawn with 25 exf4 Rxf4+, which
suggests the opening hasn’t been a success for White.
25 ... fxe3 26 Rf1

Exercise: How can Black keep playing for a win?

Answer:
26 ... Rf4!
This type of move (a partial move along an open line) can be hard to spot
but is often very effective.
27 Nc5 b6 28 Ne6
The minor piece ending after 28 Rxf4 exf4 29 Nd3 g5 30 g3 fxg3 31 hxg3
Bd4 isn’t a simple draw either.
28 ... Rxe4 29 Nxg7 Kxg7

A detailed analysis of this rook endgame is beyond our purposes. I’d

guess that the chances of a Black win or a White draw are about 50/50. In the
game, Kovalev manages to convert.
30 Rc1 Kf6 31 Rc7 Rb4 32 b3 a5 33 Rxh7 a4 34 bxa4 Rxa4 35 Kxe3
Rxa2 36 g4 Kg5 37 h3 b5 38 Ke4 b4 39 Rb7 Re2+ 40 Kd5 Kf4 41 Rxb4+
e4 42 Rb6 Rd2+ 43 Kc4 g5 44 Rf6+ Kg3 45 Rg6 Rd8 46 Re6 Kf3 47 Rf6+
Ke2 48 Rg6 e3 49 Rxg5 Kd2 50 Re5 e2 51 h4 e1Q 52 Rxe1 Kxe1 53 g5
Kf2 54 h5 Kg3 55 g6 Kg4 56 h6 Kg5 0-1
This game is one you should keep in mind the next time someone plays the
Exchange Variation against you. Active, energetic play throughout, bringing
home a fully deserved point.
Chapter Three
The Sämisch Variation
Introduction
The Sämisch is one of the most critical options against the King’s Indian, and
was a favourite of several World Champions, including Botvinnik and
Karpov. White builds a powerful pawn chain and, unlike in the Classical
main lines, retains options of castling on either side of the board.
1 d4 Nf6 2 c4 g6 3 Nc3
The immediate 3 f3 is very popular at all levels, primarily as an Anti-
Grünfeld move order. Black has tried at least half a dozen moves in this
position, but the most consistent response for King’s Indian players is 3 ...
Bg7 4 e4 d6. Now White can transpose to the Sämisch main lines with Nc3
(either now or on the next few moves), but he also has some independent
options after 5 Ne2 followed, in some cases, by Nec3 (see the notes to Game
16).
3 ... Bg7 4 e4 d6 5 f3 0-0
 This is the main starting position of our Sämisch coverage. White has
built a strong and flexible centre, which gives him a number of different
possibilities to develop his pieces.
Black, in his turn, can choose from various set-ups. Systems with ... e7-
e5, as in the Classical, are possible (and were essayed by Kasparov on several
occasions), but are not to my taste. Black can also aim for queenside play
with ... Nc6 and ... a7-a6, as favoured by English KID specialists Gawain
Jones and Mark Hebden.
My preferred systems are based on ... c7-c5, even as a pawn sacrifice.
This leads to structures characteristic of the Maróczy Bind (after ... c5xd4) or
the Benoni (after d4-d5), in which Black tends to get good counterplay on the
dark-squares. The Sämisch retards White’s development and Black ought to
use this fact to force some concessions before White brings his king to safety.
6 Be3
This is the traditional main line. 6 Nge2 and 6 Bg5 are important
alternatives which are often tried by Sämisch specialists like Alexey Dreev
(see Games 19 and 20 respectively).
 6 ... c5 7 Nge2
Keeping the tension seems to be White’s most dangerous option.
It looks like White is simply a pawn up in the endgame after 7 dxc5 dxc5
8 Qxd8 Rxd8 9 Bxc5, but in fact the picture is considerably more
complicated, as we will see in Game 16.
The immediate 7 d5 aims for a Benoni-type structure, which Black will
create with 7 ... e6 and ... e6xd5 (see Game 18).
7 ... Nc6
Taking on d4 would transpose to a rather ordinary version of the Maróczy
Bind (as might be reached through a Sicilian Accelerated Dragon) and gives
Black fewer chances for counterplay.
8 d5
Maintaining the tension with 8 Qd2 is known to be fine for Black after
the active 8 ... Qa5!, so this move is critical.
8 ... Ne5 9 Ng3
White plans Be2 and 0-0, and Black needs to decide on his response. I
have covered the most critical line:
9 ... h5!? 10 Be2 h4 11 Nf1 e6
Black aims for counterplay in a Benoni structure, and if White plays ...
12 f4
... then Black can create chaos with ...
12 ... Nxc4!! 13 Bxc4 b5!
 This is the brainchild of Alexander Grischuk (from the 2013 Candidates
tournament) and one of the most creative opening ideas I’ve ever seen.
Move orders are particularly important in the Sämisch and transpositions
are possible on almost every turn. Accordingly I’ve been entirely unfaithful
to the actual early move order of all of the featured games, replacing them
with moves which are more accurate or facilitate clearer instruction.

Game 16
M.Mchedlishvili-E.Bacrot
Tromsø Olympiad 2014

Etienne Bacrot is one of the very best King’s Indian players around, in my
opinion. As well as being extremely strong (normally rated over 2700),
Bacrot plays in quite a lot of mixed strength events (opens and team
competitions), where he needs to be able to play for a win as Black. While he
has a broad repertoire and can handle many positions well (being extremely
solid on the black side of the Queen’s Gambit Declined, for instance), for
such a player it’s extremely useful to have an opening like the King’s Indian
in his arsenal.
I was playing in the same tournament as Bacrot at the time of the present
game, though I don’t recall seeing it live. I was, and remain, deeply
impressed at Bacrot’s ability to win convincingly with Black against a strong
Georgian GM, resulting in a precious point for France in one of their most
successful Olympiads ever (France were competing for medals until the very
last rounds), under the wonderful captaincy of my friend and Gonzaga Chess
colleague (and King’s Indian expert), GM Sébastian Mazé.
1 d4 Nf6 2 c4 g6 3 Nc3
You might have noticed 3 f3 cropping up in elite tournaments.

Question: Why does White play this? And does it simply result in a
transposition to the main lines of the Sämisch?

Answer: 3 f3 is primarily directed against Grünfeld players. While Black

can (and often does) play 3 ... d5 here, after 4 cxd5 Nxd5 5 e4 the black
knight is forced to retreat since there’s no knight on c3 to exchange. This is
hardly a disaster for Black, who continues with 5 ... Nb6, but the resulting
positions are generally believed to offer some chances to White, against an
opening where top players often struggle to demonstrate anything
convincing.
The unnatural-looking 3 f3 permits a range of acceptable responses, such
as 3 ... c5 (aiming for Benoni-type structures after 4 d5) or even the odd-
looking 3 ... e6 (when play often goes into Queen’s Indian/Nimzo-
Indian/Queen’s Gambit Declined channels where, Black hopes, the f2-f3
move is more out of place than ... g7-g6). However, for King’s Indian
players, the most natural and consistent response is 3 ... Bg7. Then after 4 e4
d6, White can transpose directly to the Sämisch with 5 Nc3, but he can also
be more sophisticated and opt for 5 Ne2.
 Exercise: Why would White play this? (Hint: Try to work out how White
will develop his pieces, especially his knights – it’s not so easy!)

Answer: In the Sämisch, White struggles to find a good square for his g1-
knight. Indeed, the famous King’s Indian aficionado, GM Gufeld, suggested
that, in the Sämisch, we should ask the g1-knight what he thinks about the
move f2-f3! Clearly this move takes away the knight’s most natural
developing square, leaving two remaining candidates, each with their pros
and cons:
a) The knight could go to h3. While this looks like a poor move, the
knight is aiming for f2, which is actually not such a bad square in this
structure. The main problem with Nh3 is that White needs to wait for ... Nbd7
first, since otherwise Black will happily play ... Bxh3, spoiling White’s
structure.
b) The knight could go to e2. This is sensible, but creates another problem
– what about the bishop on f1? The bishop rarely goes to d3 in these lines
since, as we will see, Black tends to generate quick pressure on the d4-pawn,
and White doesn’t want to interfere with his queen’s defence of this pawn.
The bishop would like to go to e2, but of course both white pieces can’t
occupy the same square. And if the knight goes to e2 while the bishop is on
f1, then it will need to move again in order to allow the bishop out and White
to castle kingside.
The knight on e2 has a limited number of squares to go to, each with their
problems: on g3, it can be quickly harassed with ... h7-h5-h4, gaining space
and kingside counterplay; on f4, the knight is slightly awkward, and sets up
possibilities of ... e7-e5 (perhaps with a sacrifice on d4), forking d4 and f4;
on c1, it is rather passive, and interferes with the development of White’s
queenside. In fact, the perfect square for a knight on e2 is c3, where it is
active, has strong central influence, and doesn’t interfere with White’s other
pieces. The problem is that White’s other knight tends to be on c3 (from as
early as the third move in the Sämisch).
And now, finally, we can see White’s idea. By sending his g1-knight, via
e2, to c3, he prepares the rapid development of his kingside, avoiding the
problems outlined above. His queen’s knight, meanwhile, can develop to d2
(once the c1-bishop is brought out to e3 or g5). So White is aiming for a set-
up which, he argues, is more harmonious than he would normally obtain from
a standard Sämisch set-up.
In response we’ll proceed with the same plan of ... c7-c5 as in the main
line; i.e. 5 ... 0-0 6 Be3 c5 and now:
 a) 7 Qd2 is answered by 7 ... Nc6, and Bacrot produced a model game
with Black in this variation too: 8 d5 (taking the pawn with 8 dxc5 dxc5 9
Bxc5 is possible, but Black retains thematic compensation after 9 ... Nd7 10
Be3 Nde5 followed by ... b7-b6; we’ll see a similar continuation in the main
game) 8 ... Ne5 9 Nec3 (9 Ng3 can be met by 9 ... h5, as we will see in the
main lines) 9 ... Nh5 (not mentioned in Svetushkin’s repertoire book) 10 Be2
f5 11 Na3 f4 12 Bf2 Bd7 13 Nc2 a6 14 a4 g5 with an excellent blockade on
the dark squares in F.Caruana-E.Bacrot, Baden-Baden 2015 (½-½ in 54).
In this line Krasenkow recommends 11 exf5 gxf5 (aside from tactical
considerations such as a fork with g2-g4 which could arise on 11 ... Bxf5?,
taking with the pawn is the correct strategic choice since Black retains control
of the crucial e4-square) 12 Bh6, which he assesses as follows: “trading
Black’s strong bishop and keeping the position semi-open. In that case
Black’s king will be unsafe, which gives White good chances although the
position requires further exploration.” Here I like 12 ... f4 13 Bxg7 Nxg7,
when I think it’s hard for White to generate a kingside initiative, while Black
can prepare ... b7-b5 with ... Qe8, ... Bd7 and ... a7-a6.
 b) 7 dxc5 doesn’t make any sense here. Black has a range of decent
options, including 7 ... Nfd7, aiming to recapture with the knight, but the
simplest is 7 ... dxc5 when White can’t proceed by analogy with the main
game since 8 Qxd8 Rxd8 9 Bxc5? gives Black an excellent tactical
opportunity.

Exercise: What should Black play?

Answer: 9 ... Nxe4! is a thematic shot, taking advantage of the strength of

the dark-squared bishop on the long diagonal. After 10 Bxe7 (10 Ba3 Nc5! is
wonderful for Black) 10 ... Bxb2 11 Bxd8 Nc6 12 Bh4 Nd6, Black is doing
very well; e.g. 13 Bf6!? Bxf6 14 Nbc3 Nxc4 with more than enough
compensation for the exchange, in the form of an extra pawn, the bishop pair,
and White’s difficulties in completing development.
c) 7 d5 is a more serious move which was, for example, used successfully
by Laznicka against Radjabov in the 2015 World Blitz Championship. I think
Black should proceed in standard fashion with 7 ... e6 (I tried making the
Benko-style 7 ... b5!? work, but after 8 cxb5 a6 9 Nec3, I’m not sure that
Black has full compensation) 8 Nec3 exd5 9 cxd5 a6 10 a4 and now I like
rapidly playing for ... f7-f5 with 10 ... Nh5!?. The only game in my database
with this move went like a dream for Black, admittedly against much lower-
rated opposition: 11 Be2 (11 Qd2 Nd7 12 Na3 Ne5 13 Be2 is another
possibility, when after 13 ... Qh4+!? 14 Bf2 Qf4, White might be well
advised to repeat with 15 Be3 Qh4+ 15 Bf2 etc) 11 ... f5

c1) 12 Nd2 f4 (here 12 ... Bd4! would have been a very annoying move to
meet: 13 Bxd4? cxd4 is just winning for Black, who will follow up with ...
Qh4+ and ... fxe4 with a decisive attack; 13 Bf2 Nf4 is also very unpleasant
for White; while after 13 Nc4 Bxe3 14 Nxe3 Qh4+, the king is forced to
move since 15 g3? Nxg3 wins material) 13 Bf2 Nd7 14 Nc4 Ne5, when
Black had a perfect dark square blockade and was not worse at all in F.Lupis-
A.Cuevas, Villa Martelli 2014 (0-1 in 23).
c2) 12 0-0 is more natural, but Black has decent prospects, both after 12
... f4 (followed by 13 ... Nd7 as in Lupis-Cuevas above) and the more
enterprising 12 ... fxe4!? 13 Nxe4!? (Black is very comfortable after 13 fxe4
Rxf1+ 14 Qxf1 Nd7 with ... Ne5 to follow) 13 ... Bxb2 14 Nbd2 Bxa1 15
Qxa1, although White definitely has compensation for his large material
deficit (an exchange and a pawn), since the black king is weak and the black
queenside is completely undeveloped. In short, the position looks totally
unclear!
3 ... Bg7 4 e4 d6 5 f3 0-0 6 Be3 c5
White has a wide choice here. In this first game we’ll examine what
happens if he grabs the pawn.
7 dxc5 dxc5 8 Qxd8
After 8 Bxc5 Nc6, White has nothing better than taking on d8, with a
transposition.
8 ... Rxd8 9 Bxc5 Nc6

Question: Black sacrificed a pawn and allowed an exchange of queens!

What, if any, compensation does he have in this position?

Answer: Black’s compensation can be described in a few ways:

 1) He has a big lead in development – he has developed three minor
pieces to White’s two, castled, and brought his rook to d8. In addition, the
bishop on c5 is likely to cost White a further tempo, after ... Nd7 or ... b7-b6.
As we will see, it takes White a really long time to mobilize his kingside
pieces in this line.
2) White’s dark squares are chronically weak. The black pieces already
have excellent control of the d4-square (and the whole a1-h8 diagonal), while
additional jumping points for the knights include e5 (with ... Nd7-e5) or even
b4 (after ... a7-a5 and ... Nb4). These two factors combine to give Black full
compensation for the pawn.
Nevertheless, we are in an endgame (or, more precisely, a queenless
middlegame) and Black is a pawn down, so it’s important to have some idea
of what to do here.
10 Nge2
This seems to be the modern preference. It’s hard to think of a more
logical move: White develops a piece and fights for the d4-square. Instead:
a) 10 Ba3 was the old main line, played in classic games like V.Kramnik-
A.Shirov, German League 1992 (½-½ in 18). Black can enter forcing
variations with 10 ... a5, followed by ... Nb4, which, it’s fair to say, have been
analysed to a forced draw. However, it’s probably more practical to play 10
... b6, when a transposition to the game is possible and even likely. For
instance, 11 Rd1 Bb7 12 Nge2 Nd7 13 Nd5 Nde5 14 Nec3 e6 was the actual
move order of the Atalik game in the notes to White’s 15th move below.
Otherwise 11 Nd5 e6 12 Ne7+ Nxe7 13 Bxe7 Rd7 14 Ba3 Ba6 gives Black
excellent compensation, for example, 15 Nh3 (or 15 Rd1 Rad8 16 Rxd7
Nxd7 17 Nh3 Ne5 18 b3 Nc6 19 Be2 Bc3+ 20 Kf2 Rd2 with overwhelming
play for the pawn in O.Kroll-E.Mortensen, Danish League 1989; ½-½ in 61)
15 ... Rc8 16 Rc1 was J.Plachetka-E.Mortensen, Ostrava 1992, and now I like
16 ... Ne8!? preparing ... f7-f5 with a good game.
b) 10 Nd5 is White’s most ambitious move. The theoretically-approved
main line is to give up a second pawn with 10 ... Nd7, when Black is fine.
However, I recommend the little-played, but objectively equivalent
alternative 10 ... e6!?. After 11 Ne7+ Nxe7 12 Bxe7 Rd7, White has
demonstrated no advantage: 13 Ba3 (13 Bb4 Nh5 gives Black good play;
naturally 13 Bxf6?? Bxf6 is disastrous for White, who has traded his only
developed piece and the only piece capable of fighting for the dark squares)
13 ... b6 and now:
 b1) 14 Nh3 Ba6 15 e5 (15 Rd1 was better, with equality after 15 ... Rad8
16 Nf2 Rxd1+ 17 Nxd1 Nd7) 15 ... Ne8 16 f4 f6 17 Rd1 (17 exf6 Nxf6 18
Be2 Re8 and ... e6-e5 also gives Black good play) 17 ... Rxd1+ 18 Kxd1 fxe5
19 Kc2 Rc8 with fine play for Black in A.Karpov-H.Nakamura, Cap d’Agde
(rapid) 2008 (0-1 in 38).
b2) 14 Rd1 Ba6 15 b3 (or 15 Be2 Rad8 16 Rxd7 Nxd7 17 b3 f5 18 exf5
exf5 19 Nh3 Nc5 20 Bb4 Ne6 21 Nf2 Nf4 with full compensation in
N.Javanbakht-G.Harutjunyan, Esfahan 2011; 1-0 in 44) 15 ... Rad8 16 Rxd7
Nxd7 17 Ne2 was J.Bick-M.Cebalo, Biel 2003, and now I like 17 ... Nc5!? 18
Bxc5 (18 Nc1 Bc3+ 19 Kf2 Rd1 is unpleasant for White) 18 ... bxc5, when
the bishop pair, control of the d-file, and plan of ... a6-a5-a4 give Black more
than enough compensation for the pawn.
b3) 14 Ne2 Ba6 15 Nc3 Rc8 16 Rc1 Nh5 17 g3 Bh6 18 Rc2 (18 Rd1 is
better, when 18 ... Rxd1+ 19 Kxd1 Bxc4 20 Bxc4 Rxc4, followed by ... Bf8,
results in equality) 18 ... Bxc4 19 Nb5
 Exercise: Has Black just blundered material?

Answer: No! 19 ... Rcd8! 20 Rxc4 (20 Nd6 is better, though Black still
has good winning chances after 20 ... Rxd6 21 Bxd6 Bxf1 22 Kxf1 Rxd6) 20
... Rd1+ 21 Kf2 R8d2+ 22 Be2 (White is forced to give up the exchange to
avoid the cute 22 Kg1 Be3 mate) 22 ... Rxh1 and Black won in
Y.Ovchinnikov-E.Dyakonova, Tula 2014 (0-1 in 36).
10 ... b6
Black can also start with 10 ... Nd7, followed by ... Nde5 and ... b7-b6.
This can transpose, but if Black is aiming for a set-up with ... b7-b6, I think it
makes sense to play it straight away, developing the bishop and preparing to
bring the a8-rook into the game. 10 ... Nd7 is more popular, but 10 ... b6 has
been played by a who’s who of KID experts including Ding Liren, Radjabov
and Grischuk.
 11 Ba3
This is effectively the only move!

Question: 11 Be3 is the most popular modern response to 10 ... Nd7.

Does it work well against 10 ... b6 - ?

Answer: It does not. After 11 ... Nb4!, the threats of ... Nd3+ and ... Nc2+
cost White material: 12 Rd1 (12 Kf2 Nd3+ 13 Kg1 Nxb2 is horrible for
White) 12 ... Nd3+ 13 Rxd3 Rxd3 14 Kf2 Rd8 and Black was material up
with a superior position in R.Aghasaryan-D.Petrosian, Kajaran 2011 (½-½ in
80). 11 Bf2?? Nb4 is even worse.
11 ... Bb7 12 Rd1
The position after 12 Nd5 e6 13 Ne7+ Nxe7 14 Bxe7 has been tested in
several GM encounters. Ding Liren and Ziska chose to play 14 ... Rdc8 here,
but I prefer 14 ... Rd7 15 Bb4 Rc8 16 Nc3 (16 Rc1 Bh6 17 Rc2 Rcd8 gave
Black good play in R.Lev-Y.Kraidman, Tel Aviv 1992; 0-1 in 85) 16 ... Nh5
which was agreed drawn in I.Khenkin-I.Saric, Croatian Team Championship
2013. The game could have continued 17 Rd1 Rcd8 18 Rxd7 Rxd7 19 g3
(preventing ... Nf4) 19 ... Bd4 20 Be2 f5 with good counterplay and standard
compensation.
12 ... Nd7

13 Nd5
Other moves:
a) 13 Nf4 Nde5 14 Be2 e6 15 Nb5 Rxd1+ 16 Kxd1 Ba6 17 Ke1 was
played in M.Socko-K.Kuderinov, Bangkok (blitz) 2014. Here I think Black
should have changed the course of the game with 17 ... Bxb5!? 18 cxb5 Nd4
19 Kf2 Nxe2 20 Nxe2 a6!, when Black’s queenside pressure gives him full
compensation. For example, 21 bxa6 Nc4 22 Rb1 Rxa6 23 f4 Nxa3 24 bxa3
Bf8 and the game should end in a draw, but White is the side who needs to
prove it.
b) 13 Nc1 Nde5 14 Nb3 (14 Nd5 transposes to 14 Nc1 in the main game)
14 ... Rxd1+ 15 Kxd1 Rd8+ 16 Kc2 Ba6 17 Nb5 was Sa.Martinovic-I.Saric,
Zadar 2012, where the players repeated with 17 ... Bc8 18 Nc3 Ba6 19 Nb5
Bc8 20 Nc3 Ba6 and agreed a draw, but Black could have continued, for
instance with 17 ... f5!?.
13 ... Nde5! 14 Nec3
The e7-pawn is immune:
a) 14 Nxe7+ Nxe7 15 Bxe7 (15 Rxd8+? Rxd8 16 Bxe7 Rd7 wins
material) 15 ... Nd3+ 16 Rxd3 (16 Kd2? Nxb2+ wins) 16 ... Rxd3 and White
has two pawns for the exchange, but his poor mobilization and lack of passed
pawns or counterplay make it unlikely that he can survive here.
b) 14 Bxe7 Nxe7 15 Nxe7+ Kf8 16 Nd5 Nxc4 17 b3 Bxd5 18 Rxd5 Rxd5
19 exd5 Ne3 and Black regains his pawn with a much better position.
c) 14 Nc1 is probably White’s most prudent option at this point: 14 ... e6
15 Ne7+ Nxe7 16 Bxe7 (or 16 Rxd8+ Rxd8 17 Bxe7 Re8 18 Bd6?! Nxc4! 19
Bxc4 Rc8 20 b3 b5 and Black was better in J.Gil Capape-P.Cramling,
Benasque 1990, ½-½ in 63; 18 Ba3 improves, but Black has no problems
after 18 ... f5! with good play for the pawn) 16 ... Rxd1+ 17 Kxd1 Re8 18
Ba3 (18 Bg5 f5 19 exf5 Nxf3! was excellent for Black in E.Arlandi-
A.Khalifman, European Junior Championships, Groningen 1985; 0-1 in 37)
18 ... f5 19 exf5 Nxf3 20 fxg6 Rd8+ 21 Bd3 hxg6 22 Kc2 Ne5 23 Re1 Bxg2
and Black was better in J.Lautier-Ki.Georgiev, Biel 1992 (½-½ in 78).
14 ... e6
 15 Ne7+
Instead:
a) 15 Ne3 is assessed as dubious by Atalik in his entertaining notes to his
own game: 15 ... Rxd1+ 16 Ncxd1 Rd8 17 Be2 (after 17 Nc2 Atalik similarly
suggests 17 ... f5!? with excellent play) 17 ... Nd4 18 Kf2 Nxe2 19 Kxe2 Nd3
20 b3 Nf4+ 21 Ke1 f5! and Black had a strong initiative for the pawn in
B.Posedaru-S.Atalik, Obrenovac 2013.
b) 15 Nc7 Rxd1+ 16 Nxd1 Rd8 17 Nb5 is a typical position, where Black
enjoys long-term compensation for the pawn, in particular since White is
poorly mobilized. Many continuations are possible, but the following logical
sequence shows some of the resources available: 17 ... Na5 18 c5 Nec4 19 b4
Bc6 20 bxa5 (not 20 Nxa7? Nxa3 21 bxa5 Ba4 with a winning attack) 20 ...
Bxb5 21 Bxc4 Bxc4 22 axb6 axb6 23 cxb6 Bd4 24 Be7 (24 b7 looks safer,
for example 24 ... Rb8 25 Bb2 Bxb2 26 Nxb2 Rxb7 27 Nd1 and White
shouldn’t lose) 24 ... Rd7 25 Bb4 Bxb6 and Black’s bishop pair and more
active pieces gave him more than enough for the pawn in J.Hawkins-
M.Hebden, British Rapidplay, Halifax 2009 (0-1 in 60).
 15 ... Nxe7 16 Bxe7 Rdc8 17 c5 bxc5 18 Be2 c4

This position has been played before, but I would be very surprised if
Mchedlishvili actually aimed for it, since Black has fantastic play and has
regained his pawn.
19 Bd6
This looks pointless to me. White needs this bishop to defend his weak
dark squares, so Bxe5 isn’t a threat at all. Instead:
a) 19 f4 was played by a good GM in M.Narciso Dublan-H.Saldano
Dayer, Spanish Championship, Ceuta 2008. Black still had a good position
after meekly retreating his knight to c6, but 19 ... Nd3+! was called for, when
20 Bxd3 Bxc3+! 21 bxc3 cxd3 22 Rxd3 Bxe4 leaves White struggling to
defend his weak pawns and co-ordinate his pieces.
b) 19 Kf2 looks like the most sensible move, aiming to complete
development. Black retained some pressure after 19 ... Ba6 20 Rd6 Nc6 21
Bg5 Be5 22 Rd2 Nb4, followed by ... Nd3+, in A.Zakharchenko-H.Grötz,
Kecskemet 2005 (0-1 in 60).
19 ... Rc6
 The immediate 19 ... Ba6 is similar.
20 Ba3 Ba6 21 f4?!
This creates additional weaknesses, but it is already difficult to suggest
good moves for White.
21 ... Nd3+ 22 Bxd3 Bxc3+!
I really like this decision, for reasons I’ll discuss later. Instead, 22 ... cxd3
23 e5 tries to shut out the bishop, although Black remains on top after 23 ...
f6!.
23 bxc3 cxd3 24 Kd2

There is nothing better.

Question: How would you assess the results of the opening? And what
should Black play?

Answer: The opening has been a disaster for White. The black d3-pawn is
a powerhouse which can never be captured, since it’s supported by the bishop
on a6. White has three weaknesses (on a2, c3 and e4), whereas Black has
none. It’s also important to talk briefly about the bishops. Everyone knows
that opposite-coloured bishops can exert strong drawing tendencies, offering
drawing chances even several pawns down in “pure” opposite-coloured
bishop endgames. With other pieces on the board, however, the inferior side
can really suffer. Here, with rooks on the board, White can never neutralize
the d3-pawn, and is subject to a long-term attack on the light-squares.
The above analysis would hold true even after a lazy, automatic move
like 24 ... Rac8?. But Black has better:
24 ... Rc4!
Winning a pawn, due to the threats of ... Rxe4 and ... Ra4.
25 Bd6 Rxe4
Alternatively, 25 ... Rd8 was also strong, but Bacrot’s choice is good
enough.
26 Rhe1 Ra4

Black is now a pawn up with a better position. The rest of the game isn’t
perfect but, for our purposes, we can stop here, with Black having excellent
winning chances.
 27 a3 Rd8 28 Bb4 Bc4 29 Rb1 f5 30 Be7 Re8 31 Bd6 Ra6 32 Bb4 Rc8
33 Rb2 Kf7 34 Re5 Rb6 35 a4 h6 36 Ke3 g5 37 g3 Ba6 38 Rd2 Rbc6 39
Ra5 R8c7 40 Rf2 Kg6 41 h3 Rd7 42 Kd2 Rb7 43 Ke3 Bc4 44 Rd2 Rbc7
45 Kd4 Rd7+ 46 Ke3 a6 47 Rc5 Rxc5 48 Bxc5 gxf4+ 49 gxf4 Kh5 50 Bf8
a5 51 Rb2 Ba6 52 Kd2 Rd8 53 Rb6 Bc4 54 Bg7 Rg8 55 Rc6 Bd5 56 Rc7
Be4 57 c4 Kh4 58 Bxh6 Rg2+ 59 Kc3 Rc2+ 60 Kb3 Kxh3 61 Bg7 Kg4 62
Re7 Kf3 63 Bc3 Rh2 64 Ka3 Kxf4 65 Rxe6 Ke3 66 c5 d2 67 Rd6 Bf3 68 c6
Bxc6 69 Bxa5 Bf3 0-1

Game 17
M.Baldauf-N.Radovanovic
Novi Sad 2016

1 d4 Nf6 2 c4 g6 3 Nc3 Bg7 4 e4 d6 5 f3 0-0 6 Be3 c5 7 Nge2

White maintains the tension in the centre and would be quite happy with
7 ... cxd4 8 Nxd4, transposing to a Maróczy Bind structure in which Black is
rather passive.
 7 ... Nc6!
Continuing to offer the c5-pawn as bait.
8 d5
Instead, 8 dxc5 dxc5 9 Qxd8 Rxd8 10 Bxc5 returns to Game 16.
White can maintain the tension for another move with 8 Qd2, but this
allows Black active play with 8 ... Qa5! and then:

a) 9 d5 Ne5 10 Ng3 h5 11 Be2 h4 12 Nf1 a6 13 Nd1 Qc7 14 a4 e6 15

Bh6 exd5 16 Bxg7 Kxg7 17 cxd5 Qe7 18 Nf2 Bd7 19 Ne3 Nh5 was agreed
drawn in E.Tomashevsky-I.Khairullin, European Championship, Budva
2009.
b) 9 Rc1 a6 10 b3 b5! led to a rout in a game between two very strong
GMs: 11 dxc5 dxc5 12 Nd5 b4! (Black is already better) 13 e5?! Nd7 14 Bg5
Ndxe5 15 Bxe7 Nxe7 16 Nxe7+ Kh8 17 f4?! Ng4 (now Black is winning) 18
h3 Nf6 19 Nxc8 Raxc8 20 g3 Ne4 21 Qd3 and White resigned in E.Iturrizaga
Bonelli-L.Dominguez Perez, Istanbul Olympiad 2012.
c) 9 Rd1 Bd7 10 b3 a6 11 dxc5 dxc5 12 Nd5 Nxd5 13 exd5 Nb4 14 Kf2
was also agreed drawn in A.Dreev-J.Van Foreest, Wijk aan Zee 2016.
 d) 9 0-0-0!? is a very aggressive approach: 9 ... a6 10 Kb1 has been tested
in a few games where White has obtained some advantage in the endgame
after 10 ... b5 11 dxc5 dxc5 12 Nd5. Instead, I prefer 10 ... Rd8!, for example,
11 dxc5 (or 11 h4 b5 with counterplay) 11 ... dxc5 12 Nd5 Nxd5 13 Qxa5
Nxa5 14 cxd5 Nc4 15 Bxc5 Nxb2 16 Rd4!? e5 17 Rd2 Na4 and Black is no
worse.
8 ... Ne5

9 Ng3
Not White’s only option:
a) After 9 Nc1, I suggest proceeding in Benko Gambit style with 9 ... e6
10 Be2 exd5 11 cxd5 a6 12 a4 Bd7 13 0-0 b5!, when 14 axb5 axb5 15 Rxa8
Qxa8 16 Bxb5 Bxb5 17 Nxb5 Nc4 gave Black good compensation, and 18
Bg5 Qa6 19 Qb3 Nxb2 equalized in H.Ismail-M.Ezat, Cairo 2014 (0-1 in
101).
b) 9 Nf4 was tried in L.Ljubojevic-G.Kamsky, Monte Carlo (rapid) 1994.
Black has an effective response in 9 ... g5 10 Nh3 Bxh3 11 gxh3 h6 12 f4
gxf4 13 Bxf4 Ng6 14 Be3 Qa5 15 Qd2 Kh7 with good counterplay in
T.Niemela-A.Virtanen, Finnish League 1996 (0-1 in 29).
9 ... h5!
The standard reaction whenever a knight lands on g3.
10 Be2 h4 11 Nf1 e6

12 f4
White can play in quieter fashion with 12 Nd2 exd5 13 cxd5 h3!.
Surprisingly, this move isn’t covered in Svetushkin’s Sämisch repertoire
book, which only mentions it after 13 ... a6 14 0-0, when 14 ... h3 allows 15
g4!. As we will see, pushing the pawn on move 13 is far stronger. Then 14 g3
a6 15 a4 transposes to A.Gupta-I.Kurnosov, Nakhchivan 2013, which I’ve
annotated in detail via the 7 d5 move order (see Game 18).
Instead, 14 g4?! has been played by several GMs but allows Black a
strong initiative after 14 ... Bxg4! (a novelty unleashed by Swedish GM and
King’s Indian expert Emanuel Berg) 15 fxg4 Nfxg4 16 Bf4 (as noted by
Berg, 16 Bxg4? Qh4+ 17 Bf2? Nd3+ and wins is the main tactical
justification of the sacrifice) 16 ... Bh6! 17 Bxe5 Nxe5 (for the piece Black
has two pawns and a strong initiative, and computers slightly prefer his
position) 18 Nf3 (Berg gives 18 Qb3 Qh4+ 19 Kd1 b5) 18 ... Nxf3+ 19 Bxf3
Qh4+ 20 Ke2 b5 (Black was successful with 20 ... f5!? in A.Arias Igual-
J.Cuenca Jimenez, Roquetas de Mar 2016, 0-1 in 31, but I think Berg’s
original treatment is more accurate) 21 Rg1 (21 Qg1 b4 22 Nd1 Rae8 23 Nf2
Qf6 24 Rb1 Qd4 25 Qd1 Qe3+ 26 Kf1 f5 27 e5 was tenacious defence in
S.Brunello-D.Dochev, Greek Team Championship 2014, but Black would
still be much better after 27 ... Qxe5 28 Nxh3 g5; not 21 Kd3? b4 22 Ne2?!
Qf6, when Black collected a third pawn for the piece with a continuing attack
and a decisive advantage in C.Wallis-T.Hillarp Persson, Isle of Man 2014, 0-
1 in 37) 21 ... b4 22 Rg4 and now Black retreated his queen to f6 in
A.Istratescu-E.Berg, Greek Team Championship 2011, which gave White
some breathing room (½-½ in 88). Berg suggested the improvement 22 ...
Qe7! 23 Na4 Kg7, preparing ... f7-f5, “when the white king doesn’t have a
safe spot to go to”, and this looks clearly better for Black.
12 ... Nxc4!!
An absolutely stunning novelty, all the more impressive for having been
uncorked in a vitally important game at the highest level, namely the Svidler-
Grischuk clash from the London Candidates Tournament 2013. Black intends
to give up a whole piece for rich positional compensation. Unsurprisingly,
this move wasn’t covered in repertoire books before the Candidates (such as
Schandorff’s Playing 1 d4 series).
In J.Sriram-B.Adhiban, Indian Championship, Aurangabad 2011, Black
inserted 12 ... h3!? 13 gxh3 before launching 13 ... Nxc4!? 14 Bxc4 b5 and
went on to win (0-1 in 32). This is certainly worthy of independent
investigation, but there’s too little practical material since no one appears to
have followed Adhiban’s concept, so we’ll concentrate on the main line.

Question: Can Black reach a good position with the tactical shot 12 ...
Neg4!? 13 Bxg4 Nxg4 14 Qxg4 exd5, intending ... d5-d4 - ?
 Answer: This is tactically sound and could well be playable. However, 15
f5 d4 16 Nd5 dxe3 17 Nfxe3 reaches an odd position where White has
wonderfully placed knights and a strong initiative on the kingside. Black’s
dark-squared bishop is a monster, and indeed, can win a pawn immediately,
but that aside his position is rather passive and vulnerable. The following
high-level game is a wonderful demonstration of White’s attacking prospects
here: 17 ... Bxb2 (Black is not forced to take the pawn, but I think he needs
“something to suffer for” since, if White is allowed to consolidate, the
bishops will be fighting an uphill battle against the magnificent white
knights, not to mention Black’s vulnerable king) 18 0-0!! – a wonderful
sacrifice of the exchange (in addition to the pawn which was given up on the
last move). As noted previously, Black’s dark-squared bishop is his only
good piece. Accordingly, trading it (even for a rook!) is an attractive idea for
White. Ponomariov now takes the principled decision to accept the exchange,
distracting the rook from f1 in the process. This has been tested in several
games but I prefer White’s chances in this line. After 18 ... Bxa1 19 Rxa1
Kg7 20 Rf1, White had a very dangerous attack in E.Tomashevsky-
R.Ponomariov, European Cup, Rogaska Slatina 2011 (1-0 in 31).
Incidentally, 17 0-0-0!?, as played in K.Hovmoller-G.A.Szabo, Hallsberg
2003 (1-0 in 48), is also dangerous and has been recommended by
Mikhalchishin.
13 Bxc4 b5!

An important point, pouring more fuel on the fire.

14 Bxb5
Instead:
a) 14 Nxb5 exd5 breaks White’s co-ordination: 15 e5 dxe5 16 fxe5 Bg4
(the immediate 16 ... dxc4! is also possible, avoiding the possibility of 17
Be2, which would have given better chances to defend) 17 Qd2 dxc4! (Black
retains great compensation even in the endgame) 18 Qxd8 Rfxd8 19 exf6
Bxf6, when Black already has two pawns for the piece and b2 will drop soon
as well, J.Sarkar-D.Guthrie, London 2013 (½-½ in 55).
b) 14 e5?! bxc4 15 exf6 was L.Böttger-P.Zwahr, German League 2014,.
where 15 ... Qxf6 would give Black overwhelming compensation: he already
has two pawns and White’s co-ordination problems remain.
 c) 14 Bb3 c4 15 dxe6 (Stockfish likes 15 Bc2 b4 16 Ne2 exd5 17 Bd4
dxe4 for White, but Black has three healthy pawns for the piece and I
honestly can’t understand why he isn’t at least equal) 15 ... cxb3 16 exf7+
Rxf7 17 Qxb3 was M.Leon Hoyos-A.Shimanov, St Louis 2016. Black has
only invested a pawn for this initiative and after 17 ... Qa5!, followed by ...
b5-b4, White’s position looks dreadful.
14 ... exd5 15 e5
Amazingly, 15 exd5 Rb8! gives Black full positional compensation for
the piece. Even the engines think the game is level, and the black position is
far easier to play:

a) 16 Rc1 a6 17 Be2 Rxb2 18 Rc2 Rxc2 19 Qxc2 Re8 20 h3 Qa5 21 Bd2

Qb4 22 Kd1 Bf5 23 Bd3 Qd4 24 Bxf5 gxf5 25 Qxf5 Ne4 26 Nxe4 Rxe4 27
Qc8+ Kh7 28 Qf5+ Kg8 29 Qc8+ Kh7 30 Qf5+ Kg8 and a draw was agreed
in D.Baramidze-A.Naiditsch, Baden-Baden 2014 which, as far as I can see,
was a perfectly played game (at least by Black). You can see that bringing
more pieces onto natural attacking squares is enough to maintain equality.
Also, in the final position, White’s extra piece isn’t felt because of his poor
co-ordination (just look at his knight and rook).
b) 16 Rb1 Qe7 17 Qd2 h3 18 Bd3? Re8 19 Rg1 Ng4 20 Nd1 Bd4 21 Ke2
was I.Hrescak-K.Newrkla, Mitropa Cup, Zillertal 2015, when Black has
various ways to win, of which 21 ... Qh4! 22 g3 Qh5 23 Ke1 Bf5 is the most
elegant, with a decisive advantage since ... Rxe3+ and ... Re8 is coming.
c) 16 Bd2 was M.Parligras-I.Kurnosov, Abu Dhabi 2013, where Black
obtained a good game after developing his bishop to g4, but I prefer 16 ... h3!
with excellent play. One of the more beautiful lines here runs 17 Ne3 Ng4!!
18 Nxg4 hxg2 19 Rg1 Rxb5!! 20 Rxg2 Rxb2, when White’s position is a
mess.
d) 16 h3 Nh5 17 Be2? fails to 17 ... Rxb2 18 Rc1 Re8 and already White
can’t hold the position, P.Bersamina-A.Pichot, Baku Olympiad 2016 (0-1 in
28).
e) 16 Qd2 was A.Secheres-A.Banzea, Rumanian Championship,
Calimanesti Caciulata 2015, and now Black should regain material with 16 ...
Ng4! 17 h3 (or 17 Bc6 h3 18 g3 Qe7 19 0-0-0 Bf5 and White can’t co-
ordinate his pieces, while the black rooks will double on the b-file) 17 ...
Bxc3 18 Qxc3 Nxe3 19 Nxe3 Rxb5 20 0-0 with balanced play.
f) 16 Qe2 Bg4 17 Qd2!? is an untested but critical attempt, trying to cut
across Black’s ... Ng4 idea (17 Qc4? a6 18 Bc6 Rxb2 was much better for
Black in I.Lopez Mulet-J.Cuenca Jimenez, Spanish Rapid Championship,
Linares 2016; 0-1 in 39). I recommend 17 ... h3 18 Rg1 (18 gxh3 Bf3 makes
use of the bishop on g4) and now a fabulous shot: 18 ... Rxb5!! 19 Nxb5 Ne4,
when Black is fine despite being a whole rook down, for example, 20 Qd3
Qa5+ 21 Nd2 Re8 22 b4 Qxb4 23 Rb1 Qa5 24 Qa3 Bc3!! 25 Qxa5 Bxa5 26
gxh3 Nxd2 27 hxg4 Rxe3+ 28 Kf2 Rf3+ 29 Ke2 Rh3 30 Nxd6 Rxh2+ 31
Ke3 Rh3+ and both sides should repeat moves.
 15 ... Bg4
A similar idea of 15 ... dxe5 16 fxe5 Bg4 was played in the stem game,
but White responded well: 17 exf6!? Bxd1 18 fxg7 Kxg7 19 Bxc5 h3 20
Rxd1 hxg2 21 Rg1 gxf1Q+ 22 Kxf1with three good pieces for a queen and a
pawn in P.Svidler-A.Grischuk, London Candidates 2013 (½-½ in 41) and
later in A.Esipenko-P.Smirnov, Samara 2014 (½-½ in 41).
The immediate 15 ... Bg4 was suggested by prominent King’s Indian
analysts, Robin Van Kampen and Victor Bologan.
16 exf6
Again, White collects a bunch of pieces for the queen. 16 Be2 is a more
sober continuation: Bxe2 17 Qxe2 dxe5 18 fxe5 Nd7 19 Nxd5 (19 0-0-0?! d4
20 Bxd4? cxd4 21 Rxd4 was C.Rietze-T.Kaasen, Stockholm 2016, and now
21 ... Bxe5! 22 Rxd7 Qxd7 23 Qxe5 Rae8 is simplest, when White is doing
well from a material perspective but struggles to bring his pieces into the
game; 20 Ne4 dxe3 is also good for Black) 19 ... Nxe5 and although Black
has only one pawn for the piece, his compensation is outstanding. After 20
Rd1 Re8 21 Nf4 (21 Kf2 c4!, followed by ... Rb8, is a good suggestion by
Bologan) 21 ... Qf6, even the engines want to repeat with 22 Nd5 Qd8 23 Nf4
etc.
16 ... Bxd1 17 fxg7 Kxg7 18 Rxd1 d4
Here White shows a great resource of his own.
19 Rxd4!!
White gains three pieces for the queen after 19 Bxd4+ cxd4 20 Rxd4, but
his lack of co-ordination tells: 20 ... Qb6 21 Rd2 d5! and White can’t hold his
position together.
19 ... cxd4 20 Bxd4+ f6 21 Ne3 Rb8 22 0-0 Qa5

I’ve had this position in several blitz games. Of course the material
balance is haywire, with rook and queen for four minor pieces. The key
question is whether White is able to generate an attack against the black king.
23 Bc4 Rxb2 24 Ncd5 Rbb8!
The engines prefer Black’s position here, but this is due to an
overestimation of the queen vs. minor pieces material balance. The position
remains completely unclear and, if the white pieces manage to co-ordinate for
an attack on the black king, Black can fall into difficulties very quickly.
 Instead, 24 ... Qa3 25 f5?! Rd2 26 fxg6 was played in Sa.Martinovic-
L.Kessler, Zadar 2015, and after 26 ... Kxg6 I think Black is better, though
this is admittedly much easier to say with an analysis engine running, instead
of over the board with the clock ticking. A sample line might continue 27
Ne7+ Kh7 28 Rf5 Qc1+ 29 Rf1 Qxf1+! 30 Kxf1 Rxd4 and Black has chances
thanks to his material advantage.
25 Rd1
25 f5 g5 keeps the position unclear.
25 ... Rbc8
The engines suggest 25 ... Rbe8! with a very “human” idea: 26 f5 g5 27
Bc3 Qa3 28 Bb3 Re5! – it’s easily worth giving up an exchange to remove
White’s powerful dark-squared bishop, and protect the king. The position
remains too complex to evaluate, but Black doesn’t look worse at all.
26 Bb3 Qa3 27 Kf2 Rc1 28 Rd2 Qa6 29 Bb2 Rc5 30 g4 hxg3+ 31 hxg3

Both players have played extraordinarily well, considering the complexity

of the position. Now 31 ... Rf7 would keep things balanced. Instead, the
mistakes start coming, probably in horrible mutual time trouble.
 31 ... Qc8?? 32 g4
This gives White a clear advantage, but there was a better move: 32
Nxf6! Rxf6 33 Rxd6, which is an example of how an attack changes
everything with this material balance. There’s no reason to prefer the queen
on c8 to any of the minor pieces. White wins, since after 33 ... Rc3 34 Rxf6
Kxf6 35 Bxc3+, Black can’t recapture in view of the massive fork on d5.
32 ... Rxd5 33 Bxd5
The attack continues and White isn’t material down any more.
33 ... Re8 34 g5?
It’s worth inserting 34 Rc2 to push the queen to a worse square with a
clear advantage.
34 ... Kf8 35 gxf6 Qc5?
Here 35 ... Qh3 would give Black real chances to save the game.
36 Bd4

Now the white attack is decisive. Transferring the rook to the h-file at
once would be simplest, but White decided to finish in style by including the
king.
 36 ... Qb4 37 Rd1 Qa5 38 Kf3 Qb5 39 a4 Qa5 40 Kg4 Qa6 41 Rh1
Qe2+ 42 Kg5 1-0
The checks have run out and it’s time for mate.

Game 18
A.Gupta-I.Kurnosov
Nakhchivan 2013

1 d4 Nf6 2 c4 g6 3 Nc3 Bg7 4 e4 d6 5 f3 0-0 6 Be3 c5 7 d5

I think this move order is a good choice for White, since it cuts out the
piece sacrifice we saw in the previous game, although White has to give up
some options too.
7 ... e6!
7 ... Nbd7? 8 Nh3! lets White complete his optimal regrouping by
bringing his knight to f2.
8 Nge2
Now 8 Nh3? exd5 and 9 ... Bxh3 ruins White’s structure.
 8 ... exd5 9 cxd5 Nbd7 10 Ng3 a6 11 a4
This is almost an automatic response by White to delay Black’s ... b7-b5
break. 11 Be2 b5 12 0-0 Ne5 13 a3 Rb8 gave Black a healthy position in
P.H.Nielsen-V.Kovalev, World Blitz Championship, Dubai 2014 (½-½ in
45).
11 ... h5
Again targeting the g3-knight.
12 Be2 h4
Black can also delay this and play 12 ... Nh7!?, waiting for 13 0-0 before
pushing the pawn, for which see 11 Be2 Nh7 12 Be3 Nd7 in the notes to the
next game.
13 Nf1 Ne5 14 Nd2 h3

15 g3
Note that 15 g4?! is well met by 15 ... Bxg4! 16 fxg4 Nfxg4, a sacrifice
discovered by Emanuel Berg in a similar position (without the inclusion of
a2-a4 and ... a7-a6) and discussed in detail in the notes to Game 17. If
anything, the inclusion of the two a-pawn moves ought to help Black, since it
might lead to the opening of files on the queenside, and any opening of the
position is likely to make life more dangerous for the white king.
15 ... Bd7 16 0-0
White can force a different structure on the queenside with 16 a5. Black
has to seek counterplay with 16 ... b5 and after 17 axb6 Qxb6 18 Ra2, I like
the active 18 ... Qb4! 19 0-0 (19 Qb3, as in M.Siva-P.Konguvel, Doha 2015,
would be well met by 19 ... Rfb8 20 Qxb4 Rxb4 and ... a6-a5 with strong
play on the b-file.) 19 ... Bb5! (the bishop is very active on b5 but taking it
opens files on the queenside which would benefit Black’s major pieces) 20
g4 Rfb8 21 g5 Nfd7 (21 ... Nh5 was also good in A.Kelires-F.Perez Ponsa,
Gibraltar 2016; 0-1 in 35) 22 Qc2 Nb6 23 Nxb5 axb5 24 Rxa8 Rxa8 and
Black had an excellent position in L.Spassov-Tr.Nedev, Plovdiv 2014 (0-1 in
33).
Instead, 16 Bg5 was tried against me in an Irish weekend event. I think I
responded quite reasonably: 16 ... c4!? 17 Nxc4 Nxc4 18 Bxc4 Re8 (Black
threatens the standard trick ... Nxe4, and White needs to play carefully to
consolidate) 19 Be2 (the natural 19 0-0 is fine for Black after 19 ... Qb6+,
since 20 Rf2?! runs into 20 ... Nh7! followed by ... Bd4, collecting the
exchange, when I prefer Black’s chances) 19 ... b5 20 Qd2 Qa5 21 0-0
 Generally speaking, I think White has played pretty well to this point.
White had other possibilities, but I haven’t found a clear route to an
advantage for him, since Black retains enduring compensation in most lines,
based on his active pieces (especially the fantastic bishop on g7), strong
queenside counterplay, and the favourable kingside structure (the pawn on h3
creates enduring threats, since it could easily promote if Black lands a
sacrifice on h2 or g3). K.Moynihan-S.Collins, Bunratty 2016, continued 21 ...
bxa4 22 Bd1 (this might be an inaccuracy, after which Black’s initiative is
hard to neutralize – a consolidating move like 22 Bd3 might have been better,
when there are often Nxa4 tricks in the position, which looks rather unclear)
22 ... Bb5 23 Re1 Rab8 24 Bxa4 (White correctly avoids 24 Nxa4 Qxd2 25
Bxd2 Nxd5!, when taking the piece leads to an amusing mate after 26 exd5?
Bd4+ 27 Kh1 Rxe1+ 28 Bxe1 Bf1! with an unavoidable ... Bg2 mate, and
blocking on e3 only holds out for a couple of moves longer; I was quite
disappointed by my opponent’s choice, since I thought there was a decent
chance of him falling for this, especially since the game was played early on
a Saturday morning!) 24 ... Bxa4 25 Nxa4 Qb5 (25 ... Rb4! was better, with
full positional compensation for the pawn) 26 Bxf6? (White prevents the
threatened regrouping ... Nd7-e5, or maybe to c5, but at a huge positional
price: Black will now dominate the dark squares; 26 Bf4 would have left me
struggling to prove my compensation) 26 ... Bxf6 27 Re2, and now I should
have played 27 ... Rec8, when the extra pawn isn’t felt at all and Black is in
full control. Note that White’s king is weak, in addition to his other positional
woes.
Going back, 16 ... b5 was also possible, allowing White to execute his
threat, since after 17 f4 Neg4 (the typical 17 ... Nc4?? simply loses to 18
Nxc4 bxc4 19 e5 etc) 18 Bxg4 Bxg4 19 Bxf6, Black has the intermediate
move 19 ... Qd7!, when 20 Qc2? Bxf6 is just superb for Black, who has a
lead in development, a fabulous pair of bishops, and has executed the
standard Benoni breaks on both sides of the board. In fact White may well be
objectively lost here. This isn’t quite the end of the story, since White can
sacrifice his queen with 20 Bxg7! Bxd1 21 Bxf8 Bxa4 22 Bh6 Bc2 with an
odd material balance (rook and two knights for queen and pawn). Strong
analysis engines often favour the queen in such positions, but if White can
co-ordinate his pieces he could generate a very dangerous attack. I would be
quite reluctant to go into this position as Black.
 Exercise: How can Black generate counterplay?

Answer:
16 ... b5!
A typical sacrifice to open lines on the queenside.
17 axb5
Instead, 17 Qc2 was tried in Wang Puchen-W.Sean, Budapest 2014.
Black continued 17 ... bxa4 18 Nxa4 Bb5!, which is a typical attempt to
generate play in a Benoni structure (as we saw in the note to White’s 16th
move) but leaves Black with weak pawns. I prefer the flexible 17 ... Rb8
when Black has good counterplay.
17 ... axb5 18 Rxa8?!
After this White certainly has no advantage and, from a practical
perspective, has the more difficult position to handle. He should probably opt
for 18 Nxb5 Qb6 19 Rxa8 Rxa8 20 Na3 Qxb2 21 Nac4 Nxc4 22 Nxc4 Qb8
23 Qc2, as in A.Salem-M.Al Modiahki, Dubai 2011, when White has some
strategic trumps and Black needs to play accurately to create counterplay. I
like 23 ... Ba4! 24 Qd2 Nd7, after which Black has various active options ( ...
Ne5, ... Bb3, or ... Bb5 followed by ... Ra1) and doesn’t stand worse.
18 ... Qxa8 19 Bxb5
White has a choice of captures on b5, but each has its own disadvantages.
19 Nxb5?! was played against me by a Romanian FM in a recent game. The
position after 19 ... Bxb5 20 Bxb5 Rb8 21 Be2 Rxb2 is obviously superb for
Black, and showcases some typical tactical shots. For instance:
a) 22 Qa1 Ra2 23 Qc1 Qa4 24 Kh1 Qb4 25 Rd1

Exercise: Black to play and win!

Answer: 25 ... Nxe4! and since taking with the pawn leads quickly to
mate after ... Rxd2 and ... Qxe4+, White soon resigned in M.Feige-P.Zwahr,
German League 2013 (0-1 in 27).
b) 22 f4 Qa3 23 Bf2
 Exercise: How does Black win here?

Answer: 23 ... Rxd2 24 Qxd2 Qf3! and White resigned in T.Ringoir-

I.Cheparinov, World Rapid Championship, Dubai 2014. I actually calculated
this variation during my game:
c) 22 Rf2 Qa3 23 Nf1 Qa2 24 Qe1 Rb1 25 Qd2 Qb3 26 Bf4 Nh5 27 Qa5
Nxf4 28 gxf4
 Exercise: How can Black end the game?

Answer: 28 ... Nxf3+! 29 Bxf3 (if 29 Rxf3 then 29 ... Bd4+ leads to the
quickest mate, but I had calculated 29 ... Qxf3 when White gets mated or
goes into a losing endgame after 30 Qd8+ Kh7 31 Qh4+ Qh5 32 Qxh5+
gxh5) 29 ... Bd4 30 Be2 Qe3 and White resigned in C.Gheorghiu-S.Collins,
Gonzaga Classic, Dublin 2017.
It’s possible that White should refuse the pawn and try to consolidate
with something like 19 Qc2, but Black is clearly very comfortable.
19 ... Bxb5 20 Nxb5 Qa6 21 Nc3
Here 21 Na3 Rb8 22 Bf4 Nd3 was outstanding for Black in M.Muha-
U.Uurtsaikh, Junior Olympiad, Gyor 2014 (0-1 in 31), and 22 ... Nfd7 would
have been even better.
21 ... Rb8 22 Qc2 Nfd7
 Black has excellent compensation for the pawn. His pieces are more
active, the b2-pawn is weak, and the h3-pawn is a huge asset. I haven’t found
any way for White to equalize, and his position goes downhill without any
obvious errors.
23 Rb1 Nc4 24 Qd3 Nde5 25 Qf1 Ra8!
Or 25 ... Qa5 26 Nxc4 Nxc4 27 Qxc4 Bxc3 and Black will regain the
pawn on b2 while maintaining all the advantages of his position.
26 Nxc4 Nxc4 27 Bf4 Bd4+ 28 Kh1
 Exercise: How should Black continue?

Answer:
28 ... Nxb2!
A simple tactical shot.
29 Nb5
White can’t relieve the pressure with exchanges: 29 Qxa6 Rxa6 30 Rxb2
Bxc3 wins for Black. The white king will be boxed in, the entire f3-d5 pawn
chain is weak, while the passed c-pawn is unstoppable.
29 ... Qa2 30 g4? Nd3 31 Bg3 Rb8 32 g5 Qc2 0-1
White can’t move anything.

Game 19
A.Giri-E.Bacrot
European Cup, Bilbao 2014
1 d4 Nf6 2 c4 g6 3 Nc3 Bg7 4 e4 d6 5 f3 0-0 6 Nge2

Probably the most flexible move, since White retains options of

developing the bishop to e3 or g5, or first completing his kingside
development.
6 ... c5 7 d5 e6 8 Ng3 exd5 9 cxd5 a6 10 a4 h5
This is almost the same position as in Game 18, but the fact that the c1-
bishop is still at home (rather than on e3) alters the play somewhat.
11 Bg5
If 11 Be2 then 11 ... Nh7! is a nice move order subtlety pointed out by
GM Van Kampen (aiming to avoid 11 ... h4 12 Nf1, after which the knight
regroups to e3). White generally replies 12 Be3, after which recent games
have gone 12 ... h4 13 Nf1 f5, but I prefer 12 ... Nd7, aiming to drive the
knight back to h1 (this seems more consistent to me, since if White plays 12
0-0 we are likely to transpose after 12 ... h4 13 Nh1 Nd7 14 Be3 anyway).
Rather surprisingly, h1 isn’t a bad square for a knight in the Sämisch since it
can come to f2 which is one of its best positions, out of the way of White’s
other pieces and supporting an eventual f3-f4. Then 13 0-0 h4 14 Nh1 f5 is an
important position which has been tested by Sämisch experts like Dreev and
Korchnoi against generations of KID players.

Some examples:
a) 15 Qd2 is answered by 15 ... Qf6 and then:
a1) 16 exf5 gxf5 17 Nf2 (or 17 f4 Qg6 18 Nf2 Ndf6 and Black had good
counterplay in E.Postny-E.Can, Bursa 2010; ½-½ in 72) 17 ... f4! (directed
against Nh3-f4) 18 Ng4 Qg5 19 Bf2 Ne5 20 Ne4 Qe7 21 Nxe5 Bxe5 22 Rae1
Bf5 and Black had excellent play on the kingside in A.Dreev-T.Radjabov,
European Championship, Warsaw 2005 (0-1 in 39).
a2) 16 f4 is a very committal move. White avoids the dark square bind
which might arise once Black plays ... f5-f4 himself, but severely weakens
his light squares, including his entire pawn centre. After 16 ... fxe4 17 Bg4?!
(White should retain his bishop to cover his light squares, though after 17
Nf2 Qe7 18 Ncxe4 Ndf6 19 Nxf6+ Qxf6, Black had comfortably equalized in
A.Dreev-A.Karpov, Reykjavik rapid 2004, ½-½ in 71 – the only example
I’ve ever seen of Karpov taking the black pieces in a King’s Indian!) 17 ...
Nb6 18 Bxc8 (admitting his mistake with 18 Be2 was preferable, with
equality, but such moves are very tough to make) 18 ... Raxc8 19 Qe2? (but if
19 Nxe4 Qxb2 and White is likely to lose a pawn) 19 ... Qf7, White was
unable to defend his centre and his position collapsed in P.Sowray-
M.Hebden, Coventry 2013 (0-1 in 37).
b) 15 exf5 gxf5 16 Nf2 was tested in a game between Korchnoi and
Cheparinov: 16 ... f4!? (otherwise White would bring his knight via h3 to f4,
where it is well placed) 17 Bd2 Ne5 18 Kh1 Nf6!? (an enterprising pawn
sacrifice, as we have come to expect from the Bulgarian Super-GM) 19 Bxf4
Nh5 20 Bd2 Ng3+! 21 hxg3 hxg3 22 Nh3 Qh4 23 Bg5! Qh7 24 Qe1!
(Korchnoi was one of the best defenders in the world and coolly dealt with
Black’s kingside pressure, but White ought to have no advantage in this
position) 24 ... Bxh3 25 Qxg3 Bf5+ 26 Qh2.

Here Cheparinov understandably kept the queens on with 26 ... Qg6 27 f4

Ng4 28 Bxg4 Bxg4 29 Rae1?! Rf7 30 Qg3? Bd4, when Black’s attack should
have been decisive in V.Korchnoi-I.Cheparinov, European Team
Championship, Porto Carras 2011 (½-½ in 49), but 29 Qh4! Bd7 30 g4!
would leave White better in a very murky position.
 Instead, Black can improve his move order with 26 ... Rae8! 27 Rae1 (27
Qxh7+ Kxh7 gives full positional compensation for the pawn in the
endgame) 27 ... Qg6 28 f4 Ng4 29 Bxg4 Bxg4 and if White plays as above
with 30 Qh4 Bd7 31 g4, Black obtains fully sufficient counterplay with 31 ...
Rxe1 32 Rxe1 Qd3, when both kings are exposed and a draw by perpetual is
likely.
11 ... Nbd7 12 Be2 Qc7 13 Qd2
Or 13 0-0 c4, when 14 Qd2 Rb8 will transpose, while Van Kampen gives
14 Rf2 Rb8 15 a5 b5 16 axb6 Nxb6, intending ... Nfd7-c5, with counterplay.
13 ... Rb8 14 0-0 c4

Freeing the c5-square for the knight and preparing ... b7-b5.
15 Rab1
Instead:
a) 15 Be3 h4 16 Nh1 h3 17 Nf2 hxg2 18 Kxg2 Nh5 19 f4 Nhf6 was quite
unclear in V.Anand-H.Nakamura, Moscow 2011 (½-½ in 33). White has
some strategic trumps but his weakened king position will always give Black
potential counterplay.
 b) 15 Rac1 b5 16 axb5 axb5 17 b4 cxb3 18 Nxb5 Qd8 19 Nd4 was
J.Ehlvest-V.Kovalev, Viljandi 2014, and now I like 19 ... Qb6 20 Be3 Nc5 21
Nc6 b2 22 Rb1 Rb7 with counterplay.
c) 15 Rfc1 b5 16 axb5 axb5 17 b4 cxb3 18 Nxb5 was I.Khenkin-
D.Vocaturo, Frascati 2005, and here Black might play 18 ... Qb6+ 19 Be3
Nc5 at once, followed by ... Nfd7, when I already prefer Black.
15 ... Re8
The thematic 15 ... b5 looks like a decent move too. Perhaps Bacrot
rejected it because, after 16 axb5 axb5 17 b4, he can’t take en passant since
the b5-pawn would be weak. However, 17 ... Ne5 18 h3 Bd7 19 Rfc1 (or 19
f4 Nd3 20 Bxd3 cxd3 21 Qxd3 Nh7 and Black exchanges the dark-squared
bishop, with excellent compensation) 19 ... Nh7 20 Be3 h4 21 Nf1 g5!,
followed by ... f7-f5, gives him a good position. Note that 22 Bxg5? Nxg5 23
Qxg5 Ng6 leaves Black with overwhelming compensation on the dark
squares.
16 b4 cxb3 17 Rxb3 Nc5 18 Rb4 Bd7 19 Rfb1 Rec8

This is a different structure than we’ve encountered previously, and a

very complex position.
20 Be3
Here 20 e5!? was a strong idea, when 20 ... dxe5?! 21 Rb6! wins material,
while after 20 ... Ne8 21 e6 fxe6 22 dxe6 Bxe6 23 Nd5 Bxd5 24 Qxd5+ Qf7
25 Qxf7+ Kxf7 26 Bc4+, White has good compensation for the pawn.
20 ... h4 21 Nh1 Nh5 22 Nf2 Qd8
Now 22 ... Nxa4! 23 Nxa4 b5 24 Nb2 Bc3 was a good tactical
opportunity: 25 Rc4 Bxd2 26 Rxc7 Bxe3 27 Rxd7 Rc2 and Black regains the
piece with equality.
23 Nfd1 f5 24 exf5 Bxf5 25 R1b2 Rc7 26 Bg5 Bf6 27 f4 Ng7 28 Nf2
Rbc8 29 Ng4 Bxg4 30 Bxg4 Nf5

The position is dynamically balanced. White has the bishop pair but
Black has excellent squares for all his pieces.
31 Rb1 a5 32 R4b2 Qf8 33 Bxf6 Qxf6 34 Re1 Re7 35 Rxe7 Qxe7 36
Rb1 Re8 37 Nb5?
Most likely a time trouble error. 37 Bxf5 gxf5 38 h3 is level.
37 ... Nxa4 38 Bxf5 gxf5 39 Nd4 Nc5 40 Ne6 Nxe6 41 dxe6 Qxe6 42
Qxa5

Black now has an extra pawn but White has reasonable drawing chances
due to the reduced material, isolated black pawns and exposed black king.
42 ... Qe3+
Here 42 ... h3! was stronger, opening up the white king with the better
chances.
43 Kh1 Qxf4 44 Qd5+ Kf8 45 Qxb7 h3 46 gxh3 Qe4+ 47 Kg1!
After 47 Qxe4 fxe4, Black has excellent winning chances.
47 ... Qe3+ 48 Kh1 Qe4+ 49 Kg1 d5 50 Rf1 Re6 51 Qh7 Qd4+ 52 Kh1
Qe4+ 53 Kg1 Rf6 54 Qh8+ Kf7 55 Qh7+ Kf8 56 Qh8+ Kf7 57 Qh7+ Kf8
½-½

Game 20
V.Neverov-M.Nenezic
Obrenovac 2011

1 d4 Nf6 2 c4 g6 3 Nc3 Bg7 4 e4 d6 5 f3 0-0 6 Bg5

 This time White develops the bishop more actively, at least for the
moment.
6 ... c5 7 d5 e6

8 Qd2
Here 8 Nge2 is a tricky move order, which gives White the option of
meeting 8 ... exd5 with 9 Nxd5!? Be6 10 Nec3. I’d rather avoid this, and so I
suggest following Gallagher’s recommendation 8 ... h6 9 Be3 exd5 10 cxd5
a6 11 a4 Nbd7, when we will transpose to Game 18 after 12 Ng3 h5 (albeit
with each side having spent an extra tempo) or else to the main game after 12
Qd2 h5 or 12 Nc1 Ne5 13 Qd2 h5. In the last line, White can play 13 Be2
instead, when 13 ... h5 loses some of its point since there will be no knight on
g3 to harass and the pawn isn’t under attack yet. Accordingly, I recommend
changing tack with 13 ... Nh7!? 14 0-0 (or 14 Qd2 f5, since the h6-pawn is
indirectly defended) 14 ... f5, when Black has good kingside counterplay, for
example, 15 f4 (15 Qd2 g5 16 exf5 Bxf5 17 a5 Qe7 18 Ra3 Rae8 was
excellent for Black in L.Spassov-D.Bojkov, Bulgarian Team Championship
2010; 0-1 in 25) 15 ... Ng4 16 Bxg4 fxg4 17 Nd3 Re8 18 Qd2 Bxc3 19 bxc3
Rxe4 20 Nf2 Re7 21 f5 Bxf5 22 Bxh6 Qd7 and White was struggling to
demonstrate full compensation in M.Sadler-I.Glek, Dutch League 2001 (½-½
in 33).
8 ... exd5 9 cxd5
White can change the structure, at least for the moment, with 9 Nxd5.
Black’s light-squared bishop often struggles to find a role in Benoni
formations, which explains the radical solution: 9 ... Be6 10 Ne2 Bxd5!? 11
cxd5 b5 and Black has good counterplay, for example, 12 Nc3 a6 13 Rc1
Nbd7 14 b3 Qb6 15 Be2 Rfc8 16 0-0 Ne8 17 Kh1 Qa5 18 Rc2 Nb6 was
unclear in A.Riazantsev-A.Morozevich, Moscow (blitz) 2014 (0-1 in 36).
9 ... a6 10 a4

The active bishop on g5 requires a different approach from Black.

10 ... h6! 11 Be3
After 11 Bxh6, Black plays a standard tactical trick: 11 ... Nxe4! 12 Nxe4
Qh4+ 13 g3 Qxh6 14 Qxh6 Bxh6 15 Nxd6 Nd7, when Black remains a pawn
down but has compensation in the form of some initiative, a lead in
development, the bishop pair, and an unsafe white king. In practice I think he
has the easier game and GMs have steered well clear of this with White in
recent years:

a) 16 Rd1 Rd8 17 b3 Nf6 18 Nxc8 Raxc8 19 Bc4 b5 20 axb5 axb5 21

Bxb5 Nxd5 22 Ne2 Ne3 23 Rxd8+ Rxd8 24 Kf2 Nd1+ 25 Kg2 Rd2 26 Bc4
Ne3+ 27 Kf2 Nd1+ 28 Kg2 Ne3+ 29 Kf2 Nd1+ and a draw was agreed in
Z.Varga-V.Loginov, Budapest 1993.
b) 16 a5 Rd8 17 Bh3 Nb6 18 Bxc8?! (18 Nxc8 Nxc8 19 f4 Bg7 is “only”
full compensation for the pawn) 18 ... Rxd6! 19 Bxb7 Re8+ 20 Kf2 Nc4 21
Bc6 Be3+ 22 Kg2 Rb8 23 Ra2 Bd4 and White was suffering despite her
material advantage in N.Zhukova-V.Kotronias, Kavala 2002 (0-1 in 57).
c) After 16 Bh3, several games have proceeded with 16 ... f5, but 16 ...
Nb6! is again more accurate, as your engine will tell you, for example, 17
Bxc8 Nxc8 18 Nxc8 Raxc8 19 Ne2 Rfd8 20 Rd1 b5 21 axb5 axb5 22 Kf2 c4,
when the d5- and b4-pawns are weak and White (if anyone) needs to be
careful.
11 ... h5
11 ... Nbd7?! 12 Nh3! permits White to obtain his ideal formation with a
knight on f2. Remember to keep the c8-h3 diagonal clear (so that Nh3 can be
met by ... Bxh3) until this knight has committed to e2.
12 Nge2 Nbd7

Now putting the knight on g3 is nonsensical in view of 13 ... h4, so White

should try an alternative formation.
13 Nc1
Instead:
a) 13 Nd1 Nh7 14 Nec3 f5 15 exf5 gxf5 16 Be2 Qf6 gave Black good
counterplay in C.Ionescu-A.Kovacevic, Bucharest 1997 (0-1 in 31).
b) 13 Nf4 has been tried a couple of times by Natalia Zhukova, though 13
... Ne5 14 Be2 Bd7 15 0-0 Qa5 16 Kh1 c4 17 Bd1?! b5 18 b4 cxb3 19 Bxb3
Rfc8 was fine for Black at this stage in N.Zhukova-O.Girya, Moscow 2008
(1-0 in 39).
13 ... Ne5 14 Be2 Nh7 15 0-0 f5 16 f4
Here 16 exf5 Bxf5 17 f4 Ng4 18 Bxg4 hxg4 was excellent for Black in
D.Singh-S.Arun Prasad, Mumbai 2008 (½-½ in 23). Alternatively 16 N1a2
Bd7 17 exf5 gxf5 18 Kh1 was R.Akesson-J.Dworakowska, Tromsø 2006,
and now I like 18 ... Qf6 followed by play on the kingside.
16 ... Nf7 17 e5

Achieving the e4-e5 break is one of White’s main objectives in Benoni

structures, but he isn’t very well placed to follow it up here (if his c1-knight
were on f3 and his a1-rook on e1, this push could lead to a serious
advantage).
17 ... dxe5 18 Bxc5 Re8 19 Nd3
Instead, 19 fxe5 Nxe5 20 a5 Nf6 was fine for Black in J.Houska-
Y.Dembo, European Women’s Team Championship, Porto Carras 2011 (0-1
in 55), but perhaps this was the preferable course since, as the game
demonstrates, Black’s e-pawn can become quite a strong asset.
19 ... e4 20 Ne1 Bd7 21 a5 Rc8 22 Bb6 Qe7 23 Nc2 Nd6
 Already I prefer Black’s position. He has no weaknesses, a perfect
blockading knight on d6, a full share of central space, and prospects of
advancing on the kingside. Meanwhile, White struggles to find an active
plan. I’m very impressed by this game, where Nenezic disregards a 200
rating point deficit to convincingly beat a very strong and experienced GM
with Black.
24 Ne3 h4 25 Bd4 Rc7!
An outstanding reaction to White’s last move, preparing to lift the rook to
the kingside.
26 Ra3 g5!
Pushing on the side where Black is better. Soon all of his pieces will join
the attack.
27 Bxg7 Qxg7 28 Rb3 gxf4 29 Rxf4 Qg5 30 Ned1 Bc8!
An elegant regrouping (almost a mirror image of the ... Rf7, ... Bf8 and ...
Rg7 plan we often see in the Classical Variation). Black frees the second rank
for his rook to join the attack and overdefends his b7-pawn in the process. It
looks like the d6-knight is loose (and if it moves, White could gain
counterplay by advancing d5-d6 himself), but as we will see, White has no
time to attack it.
31 Rb6?
31 Qe3 was preferable, but then White would be much worse objectively
and without a decent plan.

Exercise: How should Black build up his attack?

Answer:
31 ... Rg7!
The correct way to start. Bringing all your pieces to the party is the first
rule of attacking play. Black threatens mate in one and White struggles to
defend against it. Note that the f4-rook is pinned, and Ne3 would leave it
undefended.
32 Bf1

Exercise: What now?

 Answer:
32 ... e3! 33 Qd4 e2!
The triumph of the e-pawn, and a damning indictment of White’s
decision on move 19. I should note that changing the move order earlier with
31 ... e3 32 Qd4 Rg7 would have given White more defensive chances after
33 g4 hxg3 34 h4, though Black is still clearly better.
34 Nxe2 Rxe2 35 Ne3

Exercise: Your experienced GM opponent is on the ropes, but accuracy is

required until the end of the game. What is Black’s best move?

Answer:
35 ... Rd2!!
By far the strongest. Now 36 Qxd2 Qxf4 would leave White pinned and a
piece down without counterplay, while other moves give Black more time to
prosecute his attack.
 36 Qe5 Nf7
Either 36 ... Re7 or 36 ... Ne4 was even stronger, but the text is more than
good enough.
37 Qe8+?! Nf8 38 Rf3 f4! 39 Qxc8 fxe3 0-1
For the second time in the game, a black pawn arrives on e3 and threatens to
advance to e2. White resigned.
Chapter Four
Systems with h2-h3
Introduction
The King’s Indian has been around a long time, and it’s difficult to find a
reasonable line which hasn’t been tried against it at some point. This doesn’t
mean that the KID is immune to fashion, quite the opposite. One of the major
changes in this respect over the past several years has been the surge in
popularity of the systems based on Nf3 and h2-h3.
1 d4 Nf6 2 c4 g6 3 Nc3 Bg7 4 e4 d6 5 Nf3
5 h3 0-0 transposes after 6 Nf3, but White can also play 6 Bg5 as
developed by Aronian (see Game 24).
5 ... 0-0 6 h3 e5 7 d5

This variation is referred to as the “Makogonov System” in several

sources. However, its modern resurgence must surely be primarily associated
with Russian Super-GM Evgeny Tomashevsky, who has relied on this system
in high-profile encounters at top events against such King’s Indian experts as
Ding Liren and Grischuk.

Question: What’s the point of 6 h3 - ?

Answer: There are several and I’m sure I’ll miss some in this explanation,
but it seems to me that h2-h3 is primarily a prophylactic measure against
Black’s counterplay with ... f7-f5. There are two aspects to this:
1. White has the option of playing g2-g4. While this doesn’t strictly
prevent ... f7-f5, which would be supported three times (by the g6-pawn, the
c8-bishop, and the f8-rook), Black would need to take into account two new
positional features: (i) a possible attack on the g-file (if Black responds to
g4xf5 by capturing with the g-pawn); or (ii) White gaining control of the e4-
square (if Black recaptures on f5 with a piece).
2. White can also play more subtly to render ... f7-f5 less desirable or
more difficult to achieve. One idea is to meet ... Nh5 with Nh2, when after ...
f7-f5 and e4xf5 Black can’t comfortably recapture, since ... g6xf5 would
hang the h5-knight to the queen on d1, while ... Bxf5 or ... Rxf5 would run
into g2-g4, again winning a piece.

Question: Sounds good. What are the downsides to 6 h3 - ?

Answer: Firstly, White’s king safety tends to be an issue in these lines.

He is unlikely to find a safe haven on the kingside, especially if he plays g2-
g4. Keeping his king in the centre or castling queenside will also give Black
his fair share of attacking chances.
Secondly, White is spending time which Black can use to accelerate his
own counterplay. If Black achieves ... f7-f5 under favourable conditions,
White’s set-up will start looking shaky.
We should also note that play by White on the queenside is more
problematic in these lines than in the Classical systems, due either to king
safety or a lack of development (the two details discussed above). Black can
often get a knight to c5 and either support it with ... a7-a5, or exploit the
loosening of White’s queenside structure after b2-b4.
Generally speaking, the positions tend to be fresh and very double edged,
which means that King’s Indian players should welcome h2-h3, in my
opinion.
There is considerable debate over the move order here, so I’ll spend some
time describing why I chose the recommended path.
7 ... Nh5
Playing for an immediate ... f7-f5 looks like the most logical plan.
Kotronias recommends 7 ... a5 in his King’s Indian repertoire series. It’s a
reasonable move, preparing ... Na6-c5. However, Robin Van Kampen makes
a compelling case for omitting or delaying ... a7-a5, since meeting ... Na6-c5
with b2-b4 can be very double edged, especially if the position opens. More
concretely, I don’t like giving White the option of 8 g4, preventing ... Nh5.
8 g3
White’s most popular move, controlling the f4-square. 8 Nh2 is a serious
alternative (see the notes to Game 23).
Of course 8 g4 would be happily met by 8 ... Nf4. Even if White wins a
pawn after 9 Bxf4 exf4, Black would get fabulous compensation with the
liberation of his dark-squared bishop (which now has no opponent) down the
long diagonal.
 8 ... Na6
I think this is the most practical choice. 8 ... a5 has also been played but,
as mentioned above, I prefer to delay or omit ... a7-a5.
Instead, 8 ... f5 is very logical and often transposes to the lines under
discussion. This move was endorsed by one of the top KID experts, elite
Chinese GM Ding Liren: “I have had this position with both sides and come
to the conclusion that maybe the plan of 7 ... Nh5 followed by ... f5 suited the
spirit of the position most.”
The big downside of this move is that White gets a couple of very
dangerous options after 9 exf5 gxf5, in particular 10 Ng5 and 10 Nh4, which
have been tried in top-level games, including some by Tomashevsky. These
are huge lines to prepare for Black and it makes more sense to simply avoid
them. White gains some additional options (compared to 8 ... f5) after 8 ...
Na6 too, but these seem less threatening to me, even in the simple sense of
not involving an enemy knight moving towards the weakened black kingside.
With regard to Ding’s comment, I should note that Black is going ... f7-f5 on
the next move against most of White’s options.
Game 21
Ni Hua-Ding Liren
Chinese Team Championship 2015

1 d4 Nf6 2 c4 g6 3 Nc3 Bg7 4 e4 d6 5 Nf3 0-0 6 h3 e5 7 d5 Nh5 8 g3 Na6

I’ve varied the move order of this game, which went 8 ... f5 9 exf5 gxf5
10 Ng5 Qe8 11 Be2 Nf6 12 Be3 Na6.
9 Be3

The most popular move in this position. White develops his bishop to its
most natural square and prepares both queenside castling and a standard
regrouping (9 ... Nc5 10 Nd2!) which we have seen in other lines, e.g. in the
Gligoric Variation.
We will examine the alternatives 9 Be2 and 9 Nh2 in Games 22 and 23
respectively.
9 ... f5
The most active option, and clearly the best in my view. I tried making
other lines work but found nothing satisfactory. Although this invites a
transposition to variations which can be reached after 8 ... f5 (as stated in the
note to Black’s 8th move), we have avoided White’s most dangerous
possibilities.
10 exf5 gxf5 11 Ng5
Instead:
a) 11 Nxe5?! is never a problem in the h2-h3 lines. After 11 ... Nxg3 12
fxg3 Bxe5, Black was already better in J.Tisdall-T.R.Hansen, Norwegian
League 2011 (0-1 in 33).
b) 11 Nh4 Nf6! is an example of a favourable transposition to 8 ... f5
lines, since after 8 ... f5 9 exf5 gxf5 10 Nh4 Nf6 (10 ... Nf4!? is a sharp
alternative), Tomashevsky has employed 11 Bg5 with success, while 11
Bd3!? e4 12 Be2 is also reasonable. In the present position White has
committed to the less active Be3, so Black has comfortable play. For
example, 12 Rg1 Kh8 13 Bd3 e4 14 Be2 Nd7 15 Ng2 Ne5 16 Nf4 Nc5 gave
Black good counterplay in O.Ladva-An.Kovalev, Pardubice 2016 (0-1 in 34).
11 ... Qe8 12 Be2 Nf6 13 Qd2
 13 ... Nc5!
Ding Liren’s improvement on his own play: 13 ... Bd7, as in the
wonderful game T.Radjabov-Ding Liren, Wijk aan Zee 2015 (0-1 in 40). In
Wijk 2015 Ding beat two of the world’s elite, Aronian and Radjabov, with
Black in the h2-h3 KID. While this was an awesome achievement, Ding
criticized his 13th move against Radjabov as based on a miscalculation, so
we’ll follow his new recipe.
14 0-0-0
Ding analysed the following lines after 14 b4 Na4 15 Nb5 (15 Nxa4 Qxa4
16 0-0 Qe8! or 16 b5 a6 is fine for Black) 15 ... Qe7!, concluding “I had
missed this simple and strong move. Or should I say I don’t want to leave my
knight on a4, but he can’t find a safe place for his king either.”
14 ... h6
The most direct, clarifying the position of the knight. The cost is some
weakening of the light squares on the kingside.
15 Nf3
Ding Liren assessed 15 Bxc5 hxg5 (or 15 ... dxc5 16 Nf3 e4 17 Nh4 Qf7)
16 Ba3 Qg6 17 Kb1 b6 as fine for Black.
15 ... Nce4 16 Nxe4 Nxe4 17 Qc2 c6!
 Generating counterplay in the centre and on the kingside, for example,
after 18 g4 cxd5 19 Rxd5 Be6, as given by Ding.
18 dxc6
An earlier game went 18 Nh4 cxd5 19 Rxd5 Be6 20 Rhd1 Bxd5?? (a very
understandable move, winning an exchange, but also a near-fatal error) 21
Rxd5 Qe6 22 Nxf5 Qxf5 23 g4 Qg6 24 Bd3 with overwhelming play on the
light squares in M.Simantsev-J.Orzechowski, Ceske Budejovice 2015 (1-0 in
40). As we see throughout these lines with h2-h3, Black needs to carefully
guard his light squares, and his light-squared bishop is his most important
piece; hence 20 ... Qd7 was better, or 20 ... Qc6 as in D.Otero Acosta-
P.Jimenez Fraga, Santa Clara 2017 (½-½ in 35), with unclear prospects in
both cases.
18 ... bxc6 19 g4
White can play more conservatively with 19 Nd2 Nf6 20 Nb3, when 20 ...
d5 (20 ... Qe7 21 f4 Re8 22 Rhe1 a5 23 fxe5 dxe5 24 Nc5 Rb8 25 Qd2 a4 26
Qd6 Qxd6 27 Rxd6 a3 28 b3 f4 was agreed drawn in V.Korchmar-H.Ziska,
Riga 2015, but perhaps White was motivated by his opponent’s strength
rather than the position, since Black’s game looks dubious here) 21 f4 d4 22
Bg1 Ne4 23 fxe5 c5 led to unbalanced play where Black was not worse in
F.Levin-U.Tuncer, German League 2015 (1-0 in 36).
19 ... d5 20 Nh4
Instead, 20 gxf5 Bxf5 21 Nh4 Bh7 22 Rhg1 Ng5 23 Qd2 Ne4 was
analysed by Ding Liren as a possible repetition.

20 ... d4
In his notes for Chess Publishing, Vigorito suggests 20 ... f4 as even
stronger, and I agree. After 21 cxd5 (or 21 f3 Ng3 22 Bc5 e4! with an
extremely strong central pawn chain and fabulous play for Black) 21 ... cxd5
(even more incisive than Vigorito’s 21 ... fxe3) 22 Rxd5 fxe3 23 Qxe4 Be6,
White’s pieces (including his king) are hanging.
21 gxf5 Nc5 22 Rhg1 dxe3 23 Rxg7+!
This should lead to a draw with best play.
23 ... Kxg7 24 Rg1+ Kf6
Arguably the riskier of Black’s alternatives, but Ding Liren has never
lacked courage or calculating ability. 24 ... Kh8 is safer, with a repetition
after 25 Ng6+ Kh7 26 f6 e4 27 Nxf8+ Qxf8 28 Qc3 Nd7 29 Rg7+ Kh8 30
Rg6 Kh7 31 Rg7+ Kh8.
25 Rg6+
Vigorito’s suggestion 25 f4! seems to create more problems for Black.
25 ... Ke7 26 Qc3 Bxf5

27 Qxe5+?!
This look like the wrong check. After 27 Rg7+ Kf6 28 Nxf5 exf2 29 Qf3
Ne6 30 Rh7 Nf4, the game should end in a draw, for example, 31 Nxh6 (or
31 Rxh6+ Kxf5 32 Qg4+ Ke4 33 Qf3+ Kf5 etc) 31 ... Ke6 32 Bf1 Rd8 33
Qg4+ Kd6 34 Qd1+ Ke6 35 Qg4+ with a perpetual, since 34 ... Kc5? 35 b4+
is too dangerous.
27 ... Kd8 28 Qxc5 Bxg6 29 Nxg6?
A further mistake, but 29 Qd6+ Qd7 30 Qxf8+ Kc7, threatening mate on
d2, is an important point. After 31 Qf4+ Kb7 32 Qxe3 Re8, White is okay
materially (with two pawns for the exchange), but his poor co-ordination is
likely to cost at least a pawn.
29 ... Rxf2 30 Ne5 Rxe2 31 Nxc6+
 I can see why Ni Hua might have been tempted by this – a queen and a
knight, coupled with a safe square on d5, would often create sufficient threats
for at least a draw against a completely exposed king. Unfortunately, here it’s
not enough.
31 ... Kd7 32 Ne5+ Ke6 33 Qd5+ Ke7 34 Nc6+ Kf8 35 Qd6+ Kg7 36
Qd4+ Kh7 37 Qd3+ Qg6 0-1

Game 22
P.Barbot-J.Van Foreest
World U-20 Championship, Khanty-Mansiysk 2015

1 d4 Nf6 2 c4 g6 3 Nc3 Bg7 4 e4 d6 5 Nf3 0-0 6 h3 e5 7 d5 Nh5 8 g3 Na6 9

Be2
Developing and targeting the h5-knight, while staying flexible with the
c1-bishop.
9 ... f5 10 exf5 gxf5 11 Bg5
11 Nxe5?! is still not a problem: 11 ... Nxg3 12 fxg3 Bxe5 and Black is
better.
11 ... Nf6
Slightly less popular than 11 ... Qe8, but a healthy move in my opinion.
GM Van Kampen recommended 11 ... Nf6 after losing a game to Delchev
with 11 ... Qe8.
 12 Qc2
Preparing queenside castling.
a) 12 a3 can be met by 12 ... c6, undermining the white centre, as noted
by Van Kampen.
b) 12 g4 Nc5 13 gxf5 Bxf5 14 Nh4 Qd7 15 Rg1 was P.Vishnu-
S.Guramishvili, London 2015. Up to this point the game has proceeded
logically, but here Black ought to have exchanged a pair of knights: 15 ...
Nfe4 16 Nxe4 Bxe4 with unclear play, for example, 17 Rg4 Bf5 18 Nxf5
Qxf5 19 Be3 a5.
c) The text was Axel Bachmann’s improvement over a game he lost to
Anish Giri, which proceeded 12 Nd2 Qe8 13 Nb3 Qg6 14 Qd2 Nb4 15 f4 h6
16 Bh4 exf4 17 gxf4 Qg2 (Black could play 17 ... Ne4! immediately, since
after 18 Nxe4 fxe4 19 Qxb4 Rxf4, White can’t hold on to his bishop on h4;
e.g. 20 Bf2? Qg2 21 Rf1 Bxh3 and White’s position collapses) 18 0-0-0 Ne4
19 Nxe4 Qxe4 20 Rdg1 (or 20 Qxb4 Qxf4+ 21 Qd2 Qxh4 with an extra pawn
and a superior position) 20 ... Nxa2+ 21 Kd1 Qb1+ 22 Nc1 Bd7 23 Bd3 Ba4+
24 b3 Qxc1+ 25 Qxc1 Bxb3+ 26 Qc2 Bxc2+ 27 Bxc2 Nc3+ and Black won
easily in A.Bachmann-A.Giri, Tromsø Olympiad 2014 (0-1 in 34).
12 ... h6 13 Be3 c6!
The exclamation mark is Van Kampen’s. Indeed, this seems to be a
strong move. Instead, 13 ... Nb4 14 Qd2 f4 15 gxf4 Bf5 16 fxe5! was much
better for White in A.Bachmann-M.Lokander, Stockholm 2016 (1-0 in 27). It
would be interesting to learn what Bachmann had prepared against the
stronger 13 ... c6.

14 0-0-0
Van Kampen gives 14 dxc6 bxc6 15 0-0-0 (or 15 Rd1 Ne8) 15 ... Qc7 16
Qd2 Ne4 with counterplay.
14 ... cxd5 15 cxd5 Bd7 16 Nh4 Rc8
A natural alternative to Van Kampen’s suggested 16 ... Rf7!?.
17 Bxa6 bxa6 18 Nxf5 Bxf5 19 Qxf5 Qa5 20 Kd2
Black has full compensation after 20 Qc2 Rc4.
20 ... Qb4 21 Rc1 Rc4 22 Ke2 Qxb2+ 23 Rc2 Qb7 24 Rhc1 Rfc8 25
Qe6+ Kh8 26 Kf1 Qc7
 Black is at least equal here. White can barely move due to the horrible pin
on the c-file, but it’s not easy for Black to improve his position either. The
rest of the game isn’t very accurate, so we’ll go through it quickly.
27 Kg2 Rf8 28 g4 Kh7?
A serious mistake. 28 ... Ng8! followed by ... e5-e4 was fine for Black.
29 Qf5+ Kh8 30 Qd3?
Missing 30 Ne4! Rxc2 31 Nxf6 Rxf6 32 Qxc2, when Black doesn’t have
enough activity to compensate for his weaknesses and White seems to be
winning.
30 ... Rc8 31 Qg6 Nh7?
Again 31 ... Ng8 was the correct way to defend the kingside, with
equality.
32 Bxh6 Bxh6 33 Qxh6 Qe7 34 Qd2 Qg5?
The queen should return to c7.
35 Qd3 Rd4 36 Ne4??
Here 36 Qxa6 should win for White.
36 ... Rxc2 37 Qxc2 Rxe4 38 Qc8+ Qg8
 39 Qf5?
Having already missed the win, White now loses the initiative as well,
and just before the time control.
39 ... Rf4 40 Rc8??
40 Qe6 kept some defensive chances.
40 ... Rxf5 0-1

Game 23
J.M.Lopez Martinez-J.Cuenca Jimenez
Roquetas de Mar 2015

1 d4 Nf6 2 c4 g6 3 Nc3 Bg7 4 e4 d6 5 Nf3 0-0 6 h3 e5 7 d5 Nh5 8 g3

White can also play 8 Nh2 on this turn.
 Now Black has the option of entering forcing lines after 8 ... Qe8 9 Be2
Nf4 10 Bf3 f5 11 g3 Nxh3 12 Bg2 fxe4 13 Be3 Na6. Many strong KID
players have gone for these successfully, but it strikes me as an impractical
choice against such an unimpressive sequence as h2-h3 and Nh2. Moreover,
such lines are highly sensitive to computer analysis. One important
continuation runs 14 Nxe4 Nf4 (this piece sacrifice isn’t forced, but it’s a
typical resource in these lines; Black has racked up a huge score from here in
over-the-board encounters, but in correspondence chess he’s been routed) 15
gxf4! (without engine assistance players have mostly opted for the compliant
15 Bf3, when Black has excellent play) 15 ... exf4 16 Bd4 f3 17 Bxf3 Rxf3
18 Qxf3 Bf5 19 Bxg7! (improving on 19 Ng4, as in M.Yilmaz-Y.Zherebukh,
Moscow 2012; 0-1 in 40) 19 ... Bxe4 20 Qc3 Bxh1+ 21 Kf1 Qd7 22 Bh6. It
takes computers a while to understand how dangerous Black’s position is
here. His extra pawn is no comfort against the permanent mate threat on g7,
and this looks quite unplayable for Black.
Instead, I recommend 8 ... Na6, as in the main game. Then 9 g3 is by far
the most popular move, returning to 8 g3 lines, which already shows some of
the practical value of playing in this fashion.
Alternatively: 9 Be3 (Van Kampen gives 9 a3 Nc5 10 b4 Nd7 11 Be3 a5
with counterplay; while 9 Be2 Nf4 10 Bf3 Nc5 11 0-0 Nfd3 was better for
Black in S.Sethuraman-P.Harikrishna, Olongapo City 2010, 1-0 in 68) 9 ...
Qe8 (preparing ... f7-f5 so that after an exchange of pawns on f5, the h5-
knight will be defended by the queen) 10 Be2 (attacking the knight again and
creating the positional threat of Bxh5) 10 ... f5! (the critical move, creating
maximum complications; Black has rich counterplay in this position)

Some examples:
a) 11 Bxh5 gxh5 12 g4 was tried by Grigoryan a couple of times, with
poor results: 12 ... hxg4 13 hxg4 fxe4 (13 ... f4 was also fine for Black in
A.Grigoryan-K.Piorun, Warsaw 2008; 0-1 in 45) 14 Qb1 (White has other
moves but nothing to frighten Black; note that 14 Nxe4 is very well met by
14 ... b5! with excellent counterplay) 14 ... Qg6 15 Qxe4 Bd7 and Black was
already better in A.Grigoryan-Rau.Mamedov, European Championship,
Rijeka 2010 (½-½ in 28).
b) 11 exf5 Nf4! (showing another point to the queen on e8 – she exerts
pressure down the e-file and is prepared to recapture if White plays f5xg6) 12
Bxf4 (12 0-0 was played in Cu.Hansen-G.Kasparov, 2nd matchgame,
Svendborg 1990, and here the strongest KID player of all time took on f5
with the bishop, but I’d prefer 12 ... Nxe2+ 13 Qxe2 gxf5, when Black has a
dream position) 12 ... exf4 13 0-0 (White can get “something to suffer for”
with 13 fxg6, but 13 ... Qxg6 14 Kf1 Nc5 leaves Black with fabulous
compensation; then 15 Nf3 Bd7 16 Qd2 was agreed drawn in P.Spacek-
R.Chytilek, Czech League 2010, but I’d play this position all day with Black
after 16 ... Rae8) 13 ... Bxf5 14 Nf3 Nc5 15 Rc1 a5 gave Black a very
comfortable position in D.Flores-D.Vocaturo, San Cristobal 2012 (½-½ in
69).
8 ... Na6 9 Nh2

Now ... f7-f5 is out for the moment, but Black can improve his game on
the queenside.
9 ... Nc5 10 Be3
This was Aronian’s latest choice (admittedly in a blitz game, and one
played only three days after a game where he essayed 10 b4), so I have taken
it as my main line.
Naturally, 10 b4 is possible. Then 10 ... Nd7! is better than 10 ... Na6
(which has also been tried by strong players), since Black wants to attack the
queenside pawn chain with ... a7-a5. The following variations are, I think, a
good indication of Black’s counterplay against an early b2-b4 and, more
generally, support the decision to delay ... a7-a5.

a) 11 Be2 a5 12 b5 f5 13 Bxh5 gxh5 14 Qxh5 Nf6 15 Qe2 Nxe4 16 Nxe4

fxe4 17 Ng4 Qe8 18 Bh6 Bxh6 19 Nxh6+ Kh8 20 Ng4 Bxg4 21 Qxg4 e3 was
played in D.Jakovenko-B.Amin, FIDE World Cup, Baku 2015, which ended
in a draw shortly after 22 0-0 (½-½ in 31). Black’s idea was 22 fxe3 e4!,
followed by ... Qe5, with full positional compensation for the pawn, for
instance, 23 Qd1 Qe5 24 Qd4 Qxd4 25 exd4 Rf3 26 g4 Raf8 and Black can’t
be worse.
b) 11 Bd2 a5 12 a3 f5 13 Bg2? (13 exf5 e4 14 Rc1 Rxf5! is good for
Black; note that 15 g4? Rxf2! gives Black a decisive attack in view of 16
Kxf2 Qh4+ which, my engine informs me, is mate in fifteen!) 13 ... Nb6! and
a 2600-rated GM was already clearly worse with White. After 14 exf5 Bxf5
15 c5 Nc4 16 Ne4 Nxd2 17 Qxd2 axb4 18 Qxb4 Nf6, Black went on to win
in A.Pashikian-T.Nedev, European Championship, Plovdiv 2012 (0-1 in 67).
c) 11 Rb1 (this prophylactic move was Aronian’s choice) 11 ... a5 12 a3
axb4 13 axb4 f5! (more incisive than 13 ... Nhf6, even if it led to a quick
Black victory in B.Morchiashvili-A.Fier, Georgian Team Championship
2014; 0-1 in 29) 14 exf5 (14 Bg2 was played in S.Milanovic-M.Pavlovic,
Serbian League 2013, where Black essayed 14 ... Nf4 and won from an
objectively bad position when White misplayed the defence, 0-1 in 21;
nevertheless, 14 ... Ndf6 is a sounder move, offering good counterplay, e.g.
15 0-0 Nxe4 16 Nxe4 fxe4 17 Ng4 Bd7 18 Qe2 Nf6 and Black is very
comfortable) 14 ... e4 15 Nb5 Ne5 16 g4 (16 fxg6 hxg6! doesn’t help White
at all) 16 ... Qh4 17 Qc2 was L.Aronian-Rau.Mamedov, World Rapid
Championship, Berlin 2015. Here Black took on f5 with the bishop and made
a draw (½-½ in 35), but he had much stronger alternatives, including the
spectacular 17 ... Rxf5!! (the quiet 17 ... Bd7!! is also horribly strong) 18
gxf5 Nd3+ 19 Bxd3 exd3 20 Qxd3 Bxf5 21 Qe2 Bxb1 22 0-0 Bf5, when
White’s position is in tatters.
10 ... a5 11 Bg2 b6!
 This was another recommendation of GM Van Kampen, preparing to
recapture on c5 with the b-pawn if needed.
12 0-0 Qe8
Black can also start with 12 ... Bd7 and has done so in some
correspondence games: 13 Qe2 Qe8 (Black is fully developed and ready for
... f7-f5) 14 Nf3 (14 Nb5 Rc8 15 a4 f5 was similar in A.Müller-V.Dolganiuc,
correspondence 2013; ½-½ in 33) 14 ... f5 15 exf5 gxf5 and Black had good
counterplay in K.Schmitzer-H.Koch, correspondence 2011 (½-½ in 32).
13 Nb5
The knight turns out to be somewhat exposed on this square. 13 Qe2 Bd7
transposes to the previous note.
13 ... Na6 14 Nf3 Bd7 15 Re1?!
This doesn’t look like a great sacrifice to me, but 15 Qe2 f5 is very
comfortable for Black in any event.
15 ... Bxb5 16 cxb5 Qxb5 17 b3
 White doesn’t appear to have full compensation here, and Black plays
without risk.
17 ... Nf6
Cuenca tries to play solidly and consolidate his extra pawn, but this
doesn’t bring him much. I would prefer the more thematic 17 ... Nc5 18 Nd2
Qe8, when Black again prepares ... f7-f5. White can only delay this
temporarily: 19 Bf3 Nf6! 20 Qc2 Nfd7 and again ... f7-f5 is coming. White’s
queenside chances do not seem to have improved with the loss of his c-pawn,
while Black’s kingside play still looks promising, especially against the
slightly advanced h- and g-pawns. The rest of the game isn’t so important for
our purposes and is presented with only light annotations.
18 Nd2 Nc5 19 a3 Rfc8 20 Rb1 Qd7 21 Qe2 c6
I’d prefer to prepare this move with 21 ... Rab8.
22 Nc4 Rcb8?! 23 Bxc5?!
Here 23 dxc6! Qxc6 24 Rec1 would leave Black under unpleasant
pressure.
23 ... bxc5 24 dxc6 Qxc6 25 a4 Rb4 26 Red1 Ne8 27 h4!
 White now has decent compensation. After mutual mistakes Black
managed to get a winning position, but ultimately the game ended in a draw.
27 ... Qa6 28 h5 Rab8 29 hxg6 hxg6 30 Qc2 Bf8 31 Bf1 Qa8 32 Be2
Nf6 33 Bf3 Nh7 34 Ne3 Ng5 35 Bg2 Ne6 36 Rd3 Nd4 37 Qd1 Qb7 38 Nd5
Rxa4 39 Qg4 Bg7 40 Rdd1 Ra3 41 Bf1 Rxb3 42 Ra1 Kf8 43 Rxa5 Qc8 44
Qxc8+ Rxc8 45 Ra6 Rb7 46 Bc4 Rd8 47 Rda1 Nf3+ 48 Kg2 Nd2 49 Rxd6
Rdb8 50 Bd3 c4 51 Bc2 Bh6 52 Rda6 Nb3 53 R1a2 Nc1 54 R2a4 Rc8 55
Ra8 Rxa8 56 Rxa8+ Kg7 57 Rc8 Nb3 ½-½

Game 24
S.Movsesian-Ding Liren
World Team Championship, Antalya 2013

1 c4 g6 2 d4 Nf6 3 Nc3 Bg7 4 e4 d6 5 h3 0-0 6 Bg5

A pet line of Levon Aronian. 6 Be3 has also been played in a few
hundred games. White is better placed to meet ... c7-c5 here than in the
Sämisch since h2-h3 is less weakening to his light squares, so I suggest 6 ...
e5 7 d5 Na6 and then:

a) 8 Bd3 Nh5 9 g3 Nc5 10 Be2 Nf6 11 Qc2 a5, when White hadn’t
achieved much with his manoeuvres and Black had a decent game in
I.Kovalenko-H.Ziska, Riga 2013 (½-½ in 73).
b) 8 g4 Nc5 9 Qc2 a5 10 Nge2 c6 11 Ng3 cxd5 12 cxd5 Bd7 13 Be2 Rc8
14 Rc1 a4 15 Qd2 b5 with good counterplay in N.Grandelius-S.Shyam, Isle
of Man 2016 (1-0 in 45).
c) 8 Nf3 Nc5 9 Nd2 succeeds in getting the knight to d2, but Black can
generate counterplay with 9 ... a5 10 g4 c6 11 Be2 a4!? 12 a3 (or 12 Bxc5
dxc5 13 Nxa4 Bh6!? 14 Nxc5 cxd5 15 cxd5 Bxd2+ 16 Qxd2 b6 and Black
will regain the pawn with a good position) 12 ... Bd7 13 h4 Ne8 14 g5 cxd5
15 exd5 Bf5 and Black had a healthy position in A.Riazantsev-
A.Demchenko, Vladivostok 2014 (0-1 in 46).
6 ... c5 7 d5 e6 8 Bd3
8 Nf3, followed by Bd3, will transpose to the lines we examine below but
loses some flexibility since the knight might prefer to go to e2 instead.
8 ... exd5
 Now White faces a choice which structure to adopt.
9 exd5
Instead:
a) 9 Nxd5 was tried in F.Mittendorf-R.Grüttner, Bad Zwesten 2006. Here
I like 9 ... Nbd7 10 Nc3 (10 Nf3 h6 is fine for Black in view of 11 Bh4?! g5!
or 11 Be3 b5!) 10 ... h6 11 Bh4 Re8 followed by ... Ne5 with good
counterplay.
b) 9 cxd5 Re8 threatens ... Nxe4 tricks and leaves White with a choice of
squares for his king’s knight:

b1) 10 Nge2 Nbd7 11 0-0 h6 12 Bh4 (or 12 Be3 Ne5 13 a4 Nxd3 14

Qxd3 b6 and having traded a minor piece and gained the bishop pair, Black
had no complaints in this Benoni structure in R.Ponomariov-V.Topalov, Wijk
aan Zee 2007; 0-1 in 58) 12 ... Ne5 13 Bb5 Bd7 14 f4 Bxb5 15 Nxb5 Nc4 16
Qc1 a6! 17 Nbc3 b5 with excellent counterplay in Vl.Georgiev-I.Cheparinov,
Albena 2014 (½-½ in 35).
b2) 10 Nf3 c4 11 Bc2 b5 12 a3 Nbd7 13 0-0 a6 14 Qd2 Bb7 15 Rfe1 Rc8
16 Rad1 Qc7. Mihail Marin assessed a similar position as slightly better for
White, writing “Despite his advantage of space on the queenside, Black has
not equalized completely. His queen’s bishop is passive, while the d4-square
will serve White quite well to increase his pressure.” However, I think Black
has reasonable chances here. for example, 17 Nd4 (17 Bh6 Bh8 18 Nd4 Nc5
19 Re3 Nfd7 20 Nf5 was N.S.Mohammad-A.Vovk, Gallipoli 2016, and now
I like 20 ... Be5 21 Nd4 Qb6 22 Bb1 Bh8, when White has nothing better
than repeating moves with 23 Rg3 Be5 24 Re3) 17 ... Nc5 18 Nc6?! (18 Bh6
Bh8 is level) 18 ... Bxc6 19 dxc6 Re6!, when Black collected the pawn on c6
and went on to win in V.Kramnik-A.Grischuk, Monte Carlo (rapid) 2011 (0-1
in 33).
9 ... Nbd7

10 Nf3
After 10 f4 Qa5 11 Kf2 (if 11 Qd2 then 11 ... a6 and ... b7-b5 gives
excellent counterplay) 11 ... Nh5!? 12 Nge2 f6 13 Bh4 Bh6 14 g3 a6 15 Qc2
f5 16 a4 Ndf6 17 Rhe1 Bd7 18 Ng1 Bg7 19 Nf3 h6 20 Re7, Black had a fine,
active position in M.Kazhgaleyev-I.Ivanisevic, Wijk aan Zee 2011. The most
straightforward continuation is to trade all rooks with 20 ... Rae8 21 Rae1
Rxe7 22 Rxe7 Re8, when Black is at least equal.
10 ... Re8+ 11 Kf1
On 11 Be3 Black’s simplest response is 11 ... Nh5 (11 ... Bh6 12 0-0
Bxe3 13 fxe3 Rxe3 14 Qd2 gives White some initiative) 12 0-0 Ne5 13 Nxe5
Bxe5 with comfortable play in R.Ignatescu-V.Nevednichy, Rumanian Team
Championship 2002 (0-1 in 34).
11 ... Ne5 12 Nxe5 Rxe5 13 Qd2

Exercise: Here Ding found an inventive way to active Black’s game. Can
you see it?

Answer:
13 ... b5!
An enterprising pawn sacrifice which works well for Black.
14 Nxb5
14 cxb5 a6 15 Kg1 axb5 16 Bxb5 Bd7 gives Black excellent Benko-style
compensation.
 14 ... Ne4 15 Bxe4
Ze.Stankovic-M.Dann, Basel 2015, saw 15 Bxd8? Nxd2+ 16 Kg1, when
16 ... Bf5! 17 Bxf5 a6!! leads to a clear advantage for Black, for example, 18
Bc7 axb5 19 Bd7 Re7 20 Bc6 Rxc7! 21 Bxa8 Bxb2 22 Re1 bxc4 and the
passed c-pawn(s) more than compensate the exchange deficit.
15 ... Rxg5 16 f4?!
The more conservative 16 Bd3 looks better, though Black retains good
long-term compensation for the pawn after 16 ... Re5 17 g3 Qb6 18 Kg2 Bd7
19 Nc3 Rae8.
16 ... Rg3 17 Qf2
Perhaps White intended 17 Nxd6, aiming to trap the rook on g3, only to
notice that 17 ... Rb8! 18 Nb5 a6 19 Nc3 Qh4 gives Black an overwhelming
attack on the dark squares.
17 ... Qh4 18 Rd1 Bh6 19 Nc3 Bxf4 20 Ne2 Bg4!

Black has regained his pawn while retaining the trumps of his position.
21 Bf3 Bf5?!
Here 21 ... Be5! was stronger, when the tactics work out in Black’s
favour. Note that 22 Nxg3 Bxg3 doesn’t help since the h1-rook will never
enter the game.
22 Nxf4 Qxf4 23 Be4 Qg5 24 h4?!
After 24 Bxf5 gxf5, Black would only be slightly better.
24 ... Qg4 25 Bf3 Qxc4+ 26 Kg1 Qf4 27 Be4 Qxf2+ 28 Kxf2 Rg4 29
Bxf5 Rf4+ 30 Kg3 Rxf5
Movsesian did well to survive to a rook endgame, but Black still has
excellent winning chances since White’s king remains exposed and there are
weaknesses in his structure.
31 Rhe1 Rb8 32 Rd2 Kf8 33 Re3 a5 34 Ra3 Rb5 35 Ra4 h5 36 b3 Kg7 37
Kh3 Kf6 38 g3 Ke5 39 Kg2 f6 40 Ra3 Ke4 41 Rd1 g5 42 Ra4+ Ke5 43
Ra3 g4 44 b4 Rxb4 0-1
Chapter Five
Systems with Be2
Introduction
1 d4 Nf6 2 c4 g6 3 Nc3 Bg7 4 e4 d6 5 Be2 0-0

As well as transposing to the Classical with 6 Nf3, White has some

important independent options in this position:
6 Bg5 is the Averbakh Variation, named after the famous Russian
opening and endgame theorist and grandmaster. I have picked quite a forcing
response which was advocated by Robin Van Kampen in his video lectures.
Black needs to know what he’s doing in this line, as illustrated when Loek
Van Wely mixed up his move order against Magnus Carlsen and lost without
a fight.
6 Be3 is a move I hadn’t seen until a couple of years ago, but which
makes some sense. White adopts a similar set-up to the Sämisch, but claims
that his developing move Be2 will leave him better placed to deal with the ...
c7-c5 sacrifice, and that ... e7-e5 can be met by an early g2-g4. This line is
double edged with chances for both sides, so King’s Indian players should be
happy to see it, even if some precision is required.

Game 25
A.Fier-J.Vakhidov
Hastings 2014/15

1 d4 Nf6 2 c4 g6 3 Nc3 Bg7 4 e4 d6 5 Be2 0-0 6 Bg5

The Averbakh System, which King’s Indian players can expect to face
from time to time. Its popularity might increase, having been recommended
by Alexei Kornev in his White repertoire series.
6 ... c5
As with most of the systems where White develops his dark-squared
bishop early, I favour this Benoni-type response. The main alternative is 6 ...
Na6, preparing ... e7-e5 (the immediate 6 ... e5?? drops material, as I’m sure
readers can work out for themselves).
7 d5
7 dxc5 is another important option. Black normally responds thematically
with 7 ... Qa5 (threatening ... Nxe4; some strong players have experimented
with 7 ... dxc5!?, which also looks fine for Black) 8 Bd2 and now hundreds of
games have proceeded with 8 ... Qxc5, but this version of the Maróczy Bind
looks quite reasonable for White, who has enjoyed excellent results after 9
Nf3 Bg4 10 Be3 Qa5 11 0-0 Nc6 12 a3. I really fail to see why Black should
be happy with this version.
Instead I would like to recommend the less popular move 8 ... dxc5!?.

The structure after 8 ... dxc5 is potentially risky for Black, in particular if
White can shut out the dark-squared bishop with e4-e5 and maintain this
pawn (often with f2-f4). However, in this particular position I think White is
too passively placed to keep the bind:
a) 9 Nf3 Nc6 10 0-0 Bg4 is more than comfortable for Black. Having the
pawn on c5 (instead of d6) makes a huge difference, since White no longer
has a space advantage and it’s harder for him to expand on the queenside,
while Black has a straightforward plan to occupy the d4-square. An
encounter between two GMs (both King’s Indian experts) resulted in a rout:
11 Re1 Nd7! (a thematic move, increasing Black’s control over the central
dark squares, when ... Bxf3, ... Nd4, and ... Nde5 are on the cards; White can
insert Nd5 but Black will respond ... Qd8, with a possible ... e7-e6 to chase
away the knight) 12 h3 Bxf3 13 Bxf3 Nde5 (this knight can be a target so
perhaps 13 ... e6 was better, simply improving the position: White’s central
dark squares will remain weak) 14 Be2 Qd8 15 Be3 Nd4 (Black could also
have exchanged queens first, when he is at least not worse in the endgame)
16 f4 Nec6 17 Bf3? (a natural-looking move, but a big error; 17 e5 would
have kept the position double edged)

Exercise: How can White’s last move be punished?

Answer: 17 ... Nb4! (an extraordinarily strong move: the threats of either
knight moving to c2, or ... Nd3, are too much to meet; the engines
recommend simply giving up the exchange, but White’s chances for survival
would have been limited in any event) 18 Bxd4 Qxd4+ 19 Kh2 Nd3 20 Qb3
Nxe1 21 Rxe1 Qd2 and the f4-pawn falls, so White resigned in W.Uhlmann-
I.Glek, Bad Zwesten 2000.
b) As mentioned above, 9 e5 is critical. Now the less popular retreat 9 ...
Ne8! is the best (and best-scoring) move for Black.

b1) 10 Nd5 Qd8 11 f4 (11 Bg5 just encourages the ... f7-f6 break: 11 ... f6
12 Nxf6+ Nxf6 13 exf6 was B.Szczechowicz-M.Nurkiewicz, Krakow 1996,
and now Black should have recaptured with the pawn: 13 ... exf6! 14 Qxd8
Rxd8 15 Be3 Nc6! 16 Bxc5 f5, when he has excellent compensation in the
endgame, not dissimilar to a good version of the Sämisch gambit) 11 ... f6 12
Nf3 Nc6 13 0-0 was tested in S.Savchenko-E.Ragozin, Riga 1988. Here,
rather than taking on e5, which gives White promising compensation based
on his strong knight on d5, I prefer playing for activity: 13 ... e6 14 Nxf6+
Nxf6 15 exf6 Bxf6 with unclear play after, for example, 16 Bc3!? Qxd1 17
Raxd1 Bxc3 18 bxc3 Rxf4.
b2) 10 f4 has been met by several moves in practice. GMs Sutovsky and
Ragozin were successful with 10 ... f6, immediately striking at the white
centre: 11 Nf3 (11 exf6 was essayed in B.Alterman-E.Sutovsky, Israeli Team
Championship 2010, when Sutovsky recaptured with the pawn and won, but
this time playing for maximum activity with 11 ... Nxf6!? 12 Nf3 Nc6 13 0-0
Qc7 looks better to me – Black will develop his light-squared bishop to f5,
probably followed by ... Rad8; the position is balanced, but Black must be
satisfied with the complete destruction of White’s centre) 11 ... Nc6 12 Qb3
Qb4 13 Qxb4 Nxb4 14 0-0-0 was K.Bischoff-E.Ragozin, Passau 1996. Now I
recommend 14 ... fxe5 15 Nxe5 (after 15 fxe5 Nc7, Black can target the e5-
pawn; this interesting position looks roughly balanced) 15 ... Nd6 16 a3
Nc6!?. It’s hard for White to establish control in this position, since the black
pieces are quite active and the pawn pushes to f4 and c4 have weakened some
important squares; of course, Black’s structure is far from perfect either. All
three results are possible.
7 ... e6

8 Qd2
In such structures 8 dxe6 is rarely a problem for Black and gives him a
choice of ways to recapture. One good treatment was demonstrated by a
(then) future World Champion, against a (then) future GM: 8 ... Bxe6 9 h3
(this is quite slow, but Black’s plan can be played even after 9 Nf3) 9 ... Nc6
10 Nf3 Qb6 11 b3 Nd4 12 0-0 Rfe8 13 Be3 Nxe2+ 14 Qxe2 Qc6 and Black
had comfortably equalized. E.Grivas-V.Anand, World Junior Championships,
Sharjah 1985, continued 15 Qd3?! Nxe4! 16 Qxe4? Bxc3 17 Qxc6 bxc6 18
Rad1, when Black had won a pawn and would have had good winning
chances after 18 ... Rad8.
8 ... exd5 9 exd5
By far the main line. While this is a less ambitious recapture than c4xd5,
White keeps more control of the position in this symmetrical structure.
Instead:
a) 9 Nxd5 isn’t a serious move. The knight is vulnerable to exchange by
... Be6xd5, when Black gets rid of a piece which often struggles to find a role
in such structures. After 9 ... Be6 10 Bd3 Nc6 11 Ne2 Re8 12 0-0 Bxd5 13
exd5 Ne5, Black was very comfortable in B.Hibner-R.Loncar, Velika Gorica
2002 (0-1 in 20).
b) 9 cxd5 has been played in 60 games in my database and deserves
respect, especially since Mamedyarov defeated Topalov with this line. After
9 ... Re8 10 f3 a6 11 a4 (of course, White doesn’t allow ... b7-b5), Black has
tried several moves, but I like the active 11 ... Qa5! which GM Sandipan has
faced a couple of times: 12 Ra3 (Black was threatening ... b7-b5 again,
exploiting the pin on the a-file, and the rook can be well placed on a3,
especially if it manages to exert influence along the third rank; there’s a
downside to this square, however, as we will see, so perhaps 12 Ra2 ought to
have been preferred) 12 ... Qb4
 The queen is extremely active on the b4-square, it can’t be attacked by
any white pawns, and harassment by the king’s knight is also unlikely since
the route to d3 (via e2) is blocked by the bishop. Then:
b1) 13 Bf4 has been met by 13 ... Bf8 in practice, but I’d like to propose a
more active solution: 13 ... Nbd7! (13 ... Nh5 14 Bxd6 f5 is another thematic
variation, but I haven’t quite managed to make it work here) 14 Bxd6 (14 a5
b5 15 axb6 Nh5 16 Bxd6 Qxb6 17 Bg3 Nxg3 18 hxg3 Rb8 19 Na4 Qb4!
keeps excellent compensation, even in the endgame, and 19 Ra2 f5 is a
position a King’s Indian player should be delighted to play: at the cost of a
pawn, Black dominates the dark squares, while the white king struggles to
find safety) 14 ... Ne5 15 Nd1 (after 15 Bxe5 Rxe5, Black has rich
compensation in the form of unopposed dark square control) 15 ... Nxd5!! (an
incredible machine resource) 16 Bxe5 (or 16 exd5 Nc4 17 Qxb4 cxb4 18
Bxb4 Nxa3 19 Bxa3 Bd7 and White is likely to lose two or three pawns while
getting his kingside out) 16 ... Bxe5 17 exd5 Bxb2 18 Re3 Rxe3 19 Qxb4
cxb4 20 Nxe3 Bd7 and this position is evaluated by the engines as “equal”,
but for a human it is chaos. Black only has one pawn for the piece, but his
bishops are hyperactive and the queenside pawns are dangerous.
b2) 13 Be3 Nbd7 14 Nh3 Ne5 15 Nf2 Nfd7 (I prefer this to 15 ... Bd7, as
played in C.Sandipan-H.Nezad, Dubai 2016; 1-0 in 48) 16 0-0?? (amazingly,
this natural move is a blunder; White could have retained equality with 16
Bf4)

Exercise: How can Black punish White’s mistake?

Answer: 16 ... Nb6! (the knight is landing on c4, where it will cause a lot
of damage) 17 Ra2 Nbc4 18 Qc1 Nxe3 19 Qxe3 Nc4 20 Bxc4 Qxc4 and
White managed to retain nominal material equality, but there’s no way his
knights can be said to match Black’s bishops (especially the monster on g7).
Black went on to overcome an almost 200 rating point deficit and win in
C.Sandipan-R.Kevlishvili, Leiden 2016 (0-1 in 47).
9 ... Qb6 10 Nf3 Bf5 11 Nh4
This is the critical move, and the one demanding exact knowledge from
Black.
 The quieter 11 0-0 Ne4 12 Nxe4 Bxe4 is very comfortable for Black, who
has traded a pair of knights, thus reducing the effect of his slightly cramped
position, while the light-squared bishop (often a problem piece) is well
placed and likely to be traded for the remaining white knight. After 13 Rae1
Nd7 14 b3 Rfe8 15 Qc1 Bxf3 16 Bxf3 Qa5 17 a4 Qb4 18 Qd1 Bc3 19 Re7
Rxe7 20 Bxe7 Be5, Black had no problems holding the balance in
I.Kovalenko-T.Radjabov, Baku Olympiad 2016 (½-½ in 37).
11 ... Ne4 12 Nxe4 Bxe4 13 f3

This is a critical position where Black has to know what he’s doing.
13 ... h6!
An excellent move.
As Krasenkow notes after 13 ... Qxb2?! 14 Rc1!, “This line of the
Averbakh System was popular a couple of decades ago. It is quite risky for
Black, who has won the b2-pawn but is considerably behind in
development.” There’s theory here but we shouldn’t get involved, since 13 ...
h6 is a much cleaner solution. Loek Van Wely fell into this line against
Magnus Carlsen in Wijk aan Zee 2013, and lost in crushing fashion: 14 ...
Bf5?! 15 Nxf5 gxf5 16 Qxb2 Bxb2 17 Rb1 Bc3+ 18 Kd1 Re8 (Csaba Balogh
gives 18 ... b6 as a tougher defence, though still much better for White after
19 Bd3 Re8 20 Bxf5) 19 Rxb7 Na6 20 a3! (coolly preventing ... Nb4) 20 ...
Rab8 21 Rxb8 Nxb8 22 Bd3 (I enjoyed Krasenkow’s note here: “White has
the bishop pair, and Black has weaknesses on f5 and d6. Such positions are
peanuts for GM Magnus Carlsen.”) 22 ... Nd7 23 Kc2 Bd4 24 Rb1! Nb6
(now Black doesn’t have ... Ne5) 25 Bf4 Be5 26 Re1 Kg7 27 Bg3 Re7 28 f4
Bf6 29 Rxe7 Bxe7 30 Be1 h5 31 g3 Bf6 32 Kb3 Kg6 33 h3! – a beautifully
quiet finish, threatening g3-g4 to win the f5-pawn. In a simplified (but
completely hopeless!) position with material equality, Van Wely resigned.
14 Bxh6
The only try for an advantage. 14 Be3?! Qxb2 15 Rc1 Qf6 16 g3 Bf5
shows an important difference from the previous note. Black was very
comfortable in N.Nodirjanova-T.Nguyen An, Asian Women’s
Championship, Mashhad 2011 (0-1 in 36).
14 ... Bxh6 15 Qxh6 Qxb2 16 0-0!
16 Rc1 is well met by the cool 16 ... Re8! and then:
 a) 17 fxe4?! (taking the piece brings nothing but trouble) 17 ... Rxe4 18 0-
0 Qd4+ (the tactical justification behind Black’s 16th move; 18 ... Rxe2?? 19
Rxf7! is crushing) 19 Kh1 Rxh4 20 Qg5 Nd7 21 Qe7 Ne5 22 Qxd6 was
A.Demianjuk-M.Vladimirov, Pavlodar 2015, and now Black should have
brought in the last piece with 22 ... Re8, when White’s position is very
unpleasant.
b) 17 Qd2 Qf6 18 g3 Bf5 19 Nxf5 Qxf5 20 0-0 Nd7 was fine for Black in
I.Jamieson-V.Bolenkov, correspondence 2012 (0-1 in 28). He has no real
weaknesses and I think the knight has more potential than the bishop in this
structure.
c) 17 Qe3? was recently essayed by a top Norwegian junior, but after 17
... Nd7! 18 Qc3 (or 18 fxe4 Rxe4 19 Qxe4 Qxc1+ 20 Bd1 Nf6, followed by ...
Re8+ with a huge initiative which more than compensates for the piece – just
look at the sorry state of White’s development) 18 ... Qxa2 19 Kf2, Black
went badly wrong with 19 ... Bxf3?, leading to an eventual draw in
J.Salomon-M.Al Sayed, Gibraltar 2016 (½-½ in 41). Instead, 19 ... Bb1! 20
Rhe1 b5!? 21 cxb5 Rab8 leaves White struggling to stay in the game, since
Rxb1 will always be met by ... Rxe2(+), regaining the piece with a very
strong initiative.
16 ... Bc2
 This admittedly looks a little artificial, in particular when Black’s
queenside development is incomplete, but White also struggles to co-ordinate
his pieces, with the knight on h4 and the bishop on e2 presenting tactical
targets in some lines.
Not 16 ... Qxe2?? 17 fxe4, which is already game over, since Black has
no good way to protect his weaknesses (including the d6-pawn) against
White’s activity: 17 ... Qxe4 (preventing Nxg6) 18 Nf5! gxf5 19 Qg5+ Kh8
20 Rf4 is a nice point.
17 Kh1
A good prophylactic move, stepping out of ... Qd4+ ideas, though White
has several other possibilities:
a) 17 Qd2 Qd4+ (17 ... Qf6 18 g3 Ba4 19 f4 gives White some pressure)
18 Qxd4 cxd4 19 Rac1 d3 20 Bd1 Nd7 21 Bxc2 dxc2 22 Rxc2
 Question: How would you assess this position? Has White not simply
won a pawn?

Answer: White is, of course, a pawn up, but Black’s compensation is

quite compelling. First, there are some weaknesses in White’s structure, in
particular the c4-pawn and the c5-square (which will be an excellent outpost
for the black knight). The knight on h4 is badly misplaced and will take
several moves to bring back into the game – note that the most natural
method (f3-f4 and Nf3) would create another potential outpost on e4.
GM Van Kampen assessed this position as “unclear/with compensation”
after 22 ... Rfe8, and subsequently a practical test confirmed this assessment:
22 ... Rac8 23 f4 Rfe8 24 Nf3 Re4 25 Nd2 Re2 26 Rf2 Re1+ 27 Rf1 Re2 28
Rfc1 (White is in a fighting mood and declines the repetition, but Black is no
worse in this position) 28 ... Nc5 29 Kf1 (Black also seems to have full
compensation in the rook endgame after 29 Nb3 Re4 30 Nxc5 Rxc5) 29 ...
Rce8 (29 ... Re3 first might have been more precise, preventing Nb3 and with
some ideas of ... Nd3) 30 Nb3 Nxb3 31 axb3 R2e4 32 g3 b6 33 b4 (otherwise
... a7-a5 seems to create a solid blockade on the queenside) 33 ... Rd4 (with
possible ideas of ... b6-b5) 34 Re1?! (easing Black’s task; the engines prefer
the counter-intuitive 34 b5!?, though it looks to me that Black retains good
drawing chances) 34 ... Rc8! 35 Re7 (35 Rec1 b5! liquidates the queenside;
for example, 36 cxb5 Rxc2 37 Rxc2 Rxb4 38 Rc6 Rxb5 39 Rxd6 a5 and the
d5- and a5-pawns will be traded, reaching a position which should be drawn
with some care) 35 ... b5 36 c5 dxc5 37 bxc5 Rxd5 38 c6 Rd6 39 Rxa7
Rdxc6 40 Rxc6 Rxc6 and a draw was agreed in B.Lalith-Lin Chen, Kolkata
2015.
b) 17 Rac1 Re8 and then:

b1) 18 Rf2 Ba4 has been tested a couple of times. Both games featured 19
Bf1 Qg7 (19 Bd3 Qg7 is similar), when Black is fine if the queens are
exchanged (K.Ozturk-Wang Jue, Turkish League 2014; 0-1 in 93) or if
they’re not (F.Novoa-S.Gimenez, Buenos Aires 2011; ½-½ in 59). The
critical line is 19 Nxg6 fxg6 20 Qxg6+ Qg7 (the e8-rook is protected!) 21
Qxd6 Nd7 22 Qf4, when Van Kampen gives 22 ... Ne5 as unclear, whereas I
like 22 ... Re7, followed by 23 ... Rae8. I prefer the piece to the three pawns
in this position.
b2) 18 Qd2 Qf6 19 g3 (Black was very comfortable after 19 Rxc2 Qxh4
20 Rb2 b6 21 f4 Nd7 in P.Dolinski-W.Niewiadomski, correspondence 2007;
0-1 in 36) 19 ... Ba4 20 f4 is given by Kornev as slightly better for White,
stating “His prospects in the middle game will be preferable, because his king
is much safer (3 pawns against 2 for Black’s king). After the exchange of the
queens (for example after ... Qf6-e7-e3), White can exploit the vulnerability
of the enemy d6-pawn, after Nh4-f3-g5, Bf3 and Ne4.”
I think Kornev does a good job of describing the strategic dangers in the
position, but I also believe that Black is no worse. G.Flear-R.Thogersen,
Hastings 2016/17, continued 20 ... Nd7 21 Nf3 Qe7 22 Kf2 b5 23 Rfe1,
where Black blundered horribly and resigned after 23 ... bxc4?? 24 Bd1!.
Instead, 23 ... Nf6 would give him fully acceptable play, for instance, 24 Ng5
bxc4 25 Bxc4 Qd8 26 Rxe8+ Bxe8 27 Rb1 Rb8, when the white king is no
safer than its black counterpart. In fact I’d argue it’s less safe because of the
f3-f4 move. Black has excellent play.
c) White can also play directly with 17 f4 Nd7 18 f5 Qg7 and now:
 c1) 19 Qd2 Qd4+ leads to similar play to line ‘a’: 20 Qxd4 cxd4 21 Rac1
d3 22 Bd1 Bxd1 23 Rcxd1 (23 Rfxd1 Nc5 24 fxg6 fxg6 25 Nxg6 Rf6 26 Nh4
Re8 is risky for White, since the d3-pawn is very strong) 23 ... Rac8 and
Black has no cause for complaint.
c2) 19 Qg5 was M.Kvetny-Max.Hess, Willingen 2015, when 19 ... Rae8!
would have left White in serious trouble; for instance, 20 Rae1 Qd4+ 21 Kh1
Qe3 22 Rf4 (after 22 Qxe3 Rxe3, Black has a serious advantage in the
endgame) 22 ... Ne5 and White’s attack has run out of steam, while his
exposed position contains several weaknesses.
17 ... Re8 18 Qd2 Qd4! 19 Qxc2
Not 19 Qxd4?? cxd4 and Black would even win in view of the threats of
... Rxe2 and ... g6-g5.
19 ... Qxh4
 In principle, Black should be fine in this position. Once his queenside is
mobilized, his knight will be at least as effective as White’s bishop (which is
limited by the pawns on c4 and d5).
20 Rab1 b6 21 f4 Nd7 22 f5 Kg7
Here 22 ... g5!? followed by ... f7-f6 is more ambitious – the knight will
defend the whole kingside from e5, and the king will be quite secure on g7.
23 Rb3 Qe7
And here 23 ... Rh8 looks logical, for instance, 24 Rh3 Qe7 25 fxg6 fxg6
and White has nothing better than beautifully forcing a draw with 26 Qxg6+!!
Kxg6 27 Bd3+ etc.
24 Bd1 Qe4 25 Qf2 Qd4 26 Qf3 Qf6 27 fxg6 fxg6 28 Qxf6+ Nxf6 29
Rbf3 Rf8 30 Kg1 Rae8 31 Bc2 Nd7 32 Rxf8 Nxf8
 A level endgame has arisen. Black went on to lose but the result of the
game had nothing to do with the opening.
33 h4 Re3 34 Rf3 Re2 35 Rf2 Re1+ 36 Rf1 Re5 37 Rf3 Rh5 38 g3 Nd7
39 Ra3 a5 40 Rb3 Rh8 41 Rb1 Rb8 42 Bd3 Ne5 43 Be2 g5 44 Kg2 gxh4 45
gxh4 Ng6 46 Kg3 Re8 47 Bh5 Rb8 48 Bxg6 Kxg6 49 Kg4 Kf6 50 Re1 b5
51 Re6+ Kf7 52 cxb5 Rxb5 53 Rxd6 Rb2 54 a3 Rb3 55 h5 Rxa3 56 h6
Ra1 57 h7 Rg1+ 58 Kf5 Rf1+ 59 Ke4 Re1+ 60 Kd3 Kg7 61 Rd7+ Kh8 62
d6 Rc1 63 Ke4 Rd1 64 Ke5 Re1+ 65 Kd5 Rd1+ 66 Kxc5 Rc1+ 67 Kd5
Rd1+ 68 Ke6 Re1+ 69 Kf6 Rf1+ 70 Ke7 Re1+ 71 Kf8 Rd1 72 Rd8 Kxh7
73 Ke7 Re1+ 74 Kd7 Kg7 75 Re8 Rf1 76 Ra8 Ra1 77 Ra7 Kf6 78 Kc8
Rc1+ 79 Rc7 1-0

Game 26
A.Kelires-H.Ziska
European Small Nations Championship, Luxembourg 2016

1 d4 Nf6 2 c4 g6 3 Nc3 Bg7 4 e4 d6 5 Be2 0-0 6 Be3

 This system, which is an odd hybrid of the Sämisch, Gligoric, and h2-h3
lines, was branded the “No Name System” by IM David Vigorito, and I’m
happy to adopt his terminology! To be honest, I didn’t learn of its existence
until recently, when 6 Be3 began popping up in the games of several strong
GMs (Alexander Riazantsev, for example) who are looking for a complex
fight without the huge theory of the main lines.
I was very impressed by the text game, won by Helgi Dam Ziska on his
way to winning the 2016 Small Nations Championship and a valuable place
in the World Cup. His opponent needed a win for first place, whereas a draw
was enough for Ziska to claim the top spot, which explains some of the
decisions taken in the game.
6 ... e5
This is the principled move. A couple of months later Ziska varied with 6
... Na6. This is a tricky move order, since 7 g4 is more doubled edged when
the centre isn’t closed – for example, Black can meet it with 7 ... c5!?, and
interesting counterplay. In the game White went for 7 Qc2, when Black
changed plan with 7 ... Ng4, earning a quick draw after 8 Bxg4 Bxg4 9 Nge2
c5 10 0-0 cxd4 11 Bxd4 Bxe2 12 Nxe2 Bxd4 13 Nxd4 Rc8 in P.Vezzosi-
H.Ziska, Forni di Sopra 2016.
7 d5 Na6 8 g4!?
Gaining space and trying to discourage an eventual ... f7-f5. Naturally,
White takes on some additional risk playing in this fashion, since he weakens
his position and delays his development.
After 8 Nf3, White is threatening to establish his ideal set-up with Nd2.
Black ought to disrupt this with the typical 8 ... Ng4! 9 Bg5 f6 10 Bh4 Nh6
11 Nd2 c5 12 f3 Nf7 13 a3 Bh6 14 Bf2 f5 15 h4 Bd7 16 Rb1 b6 17 b4 Qe7
and Black was doing fine in R.Vogel-R.Van Kampen, German League 2017
(½-½ in 55).
8 ... Nc5

9 Bf3
The alternative is 9 f3. As Ziska observes, “The downside of that move is
that it allows Black to meet h4 with ... h5. On the other hand, the bishop on
e2 keeps the c4-pawn defended. In the game, the undefended c4-pawn turns
out to be an important detail.” Here 9 ... a5 10 h4 c6 (Black can proceed with
10 ... h5 as flagged by Ziska, but I don’t see anything wrong with the text) 11
h5 looks like trouble for Black, but he can establish queenside counterplay
with the following thematic sequence: 11 ... cxd5 12 cxd5 a4 13 Nh3 Bd7 14
Qd2 Qa5 15 Nf2 b5 16 Kf1 b4 17 Ncd1 Bb5 18 Kg2 Rfc8. Black had played
in model fashion and had decent counterplay in M.Rivas Pastor-M.Narciso
Dublan, Spanish Championship, Leon 2006 (½-½ in 54).
9 ... a5 10 h4 c6!?
Varying from his previous game, which continued 10 ... Nfd7 11 h5 Nb6
and ended in another impressive win for Black in B.Chatalbashev-H.Ziska,
Reykjavik 2011 (0-1 in 41). Ziska’s new move involves a pawn sacrifice,
which seems quite justified to me considering White’s lack of development
and potentially unsafe king.
11 g5
Declining the pawn with 11 h5 cxd5 12 cxd5 Bd7 13 g5 Ne8 14 Nge2 b5
gave Black excellent counterplay in A.Korbal-W.Maes, Police 2013 (1-0 in
31).
11 ... Nfd7
11 ... Nh5?! is a highly speculative pawn sacrifice, known from some
lines of the Modern Benoni, which was essayed in V.Petkov-D.Stets, Nantes
2011 (½-½ in 49). Despite the result of the game, this version seems quite
dubious to me.
12 dxc6
The critical try, grabbing the pawn. Instead:
a) 12 Nge2 Nb6 13 b3 cxd5 14 cxd5 f5 15 Bg2 Bd7 16 f3 Rc8 17 Qd2
Na8!? 18 Nc1 Nc7 19 Nd3 N7a6 and Black was more than fine in
Vi.Kiselev-E.Can, Voronezh 2007 (0-1 in 26).
b) 12 h5 cxd5 13 cxd5 a4 (13 ... b5!? 14 Nxb5 Ba6 could also be
considered) 14 Bg4 was T.Nabaty-G.Akash, Chennai 2011. Now I like 14 ...
f5!, when the opening of the position heavily favours Black, who is far ahead
in development.
12 ... bxc6 13 Qxd6 Qb6
 Question: What does Black have for the pawn?

Answer: Black frequently sacrifices his d6-pawn in the King’s Indian,

and the compensation tends to assume a similar form. His play on the dark
squares is excellent – the d4-square is weak and the knight stands very well
on c5. Black’s play down the b-file is also very promising, especially since
the white king can hardly castle short. To improve your understanding of
these sacrifices, I highly recommend you analyse the game between Karpov
and Kasparov from Linares 1993, which is one of the most crushing losses in
Karpov’s career.
In P.Korobkov-V.Spasov, Bansko 2010, Black tried 13 ... Ne6 14 Nge2
c5 and ... Nd4. He retains some compensation playing in this fashion, but I
prefer Ziska’s treatment.
14 0-0-0
Ziska notes that 14 Qd2? fails to the cute 14 ... Qxb2!. After 14 Nge2, the
queen should avoid taking on b2 (as Rb1 would leave it short of squares).
Instead, Ziska had planned 14 ... Re8 15 0-0 Bf8 16 Qd2 Qb4 with reasonable
compensation.
14 ... Rb8 15 Rd2
Ziska gives 15 Qd2? Nd3+ 16 Kc2 N7c5 with a strong attack, in view of
17 Na4? Qb4! 18 Bxc5 Qxa4+ 19 Kxd3 Rd8+ and wins.
15 ... Re8
Now Black threatens to catch the white queen with 16 ... Bf8.
16 Na4 Qb4 17 Nxc5
17 Bxc5 Qxc4+ 18 Rc2 Qxa4 is also nice for Black.
17 ... Bf8 18 Qxc6
Here 18 Qxd7!? Bxd7 19 Nxd7 is an attempt to keep the game going, but
Ziska notes that Black can still force a draw with 19 ... Red8! 20 Nxb8
Qxc4+ 21 Kb1 Qf1+ 22 Kc2 Qc4+ etc.
18 ... Nxc5 19 Qxe8 Qxc4+

20 Kb1??
Trying to avoid the draw, Kelires blunders. 20 Kd1?? was no better, as
Black wins with 20 ... Nd3!. White had to acquiesce to 20 Rc2 Qf1+ 21 Kd2
(not 21 Bd1?? Nd3+ winning in view of 22 Kd2 Qe1+ 23 Kxd3 Ba6+) 21 ...
Qd3+, which was the forced draw Ziska had calculated at move 15.
20 ... Nxe4!
Ironically, even here Black could force a draw with 20 ... Be6 21 Qxb8
Qxa2+, but Ziska’s choice is far stronger.
21 Rc2
Or 21 Bxe4 Qxe4+ 22 Ka1 Bf5 and Black wins.
21 ... Nc3+ 22 Kc1 Qf1+ 0-1
White had missed that 23 Kd2 Nb1 mate is an attractive checkmate.
Chapter Six
Other Lines
Introduction
1 d4 Nf6 2 c4 g6 3 Nc3 Bg7

The King’s Indian is an enormous opening. Because Black does not

immediately contest the centre, White has a very wide choice of how to set
up his pieces. The lines we have covered in the preceding chapters will
feature in over 95% of your games, but White has some other systems
available which it is necessary to study.
The most important line in this section is the Four Pawns Attack (4 e4 d6
5 f4). This is a very ambitious attempt to seize and hold a huge central space
advantage, and if White is permitted time to develop his pieces in comfort,
Black’s game will be almost unplayable. Black has a range of responses here,
but I’ve covered one I used to win the Irish Championship a couple of years
ago, which generates quick counterplay against the centre (see Game 27).
After 4 e4 d6, White has tried some minor alternatives of which the most
significant, in my view, are 5 Bd3 and 5 Nge2. Black can get a good game in
several ways, as we’ll see in Game 28.
Finally, White can try systems with an early Bf4 or Bg5, in which he
normally puts his e-pawn on e3 rather than e4. These are examined in Game
29.

Game 27
D.Komarov-A.Motylev
Niksic 2000

1 d4 Nf6 2 c4 g6 3 Nc3 Bg7 4 e4 d6 5 f4

The Four Pawns Attack has always been viewed as dubious or, at a
minimum, not a significant challenge to the King’s Indian from a theoretical
perspective. However, this system contains real poison, not least because in
several of the main lines Black has to remember a lot of theory and the price
of a single misstep is high. Of course, the same could be said of the whole
King’s Indian, but I still don’t recommend going into the traditional main
lines after 5 ... c5 6 d5 0-0 7 Nf3 e6 8 Be2 (8 dxe6 is also a decent move,
which I used to win a game in an Irish Championship) 8 ... exd5 9 cxd5, as
while Black has several playable moves here (9 ... Bg4, 9 ... Re8, and 9 ...
Nbd7) the position is sharp and has more theory than I like for a line you’re
only likely to encounter once every couple of years.
Instead, I’d like to suggest an alternative approach which I used to win
my second Irish Championship.
5 ... 0-0 6 Nf3

Question: What do you think of the immediate 6 e5 - ?

Answer: This is not a good idea and leaves White’s centre over-extended
after 6 ... Nfd7 7 Nf3 c5!. If you haven’t seen it already, I’d encourage you to
have a look at the game R.Letelier Martner-R.J.Fischer, Leipzig Olympiad
1960 (0-1 in 23), which remains the model game for dealing with such play
(albeit with ... Ne8 rather than ... Nfd7).
6 ... Na6
 In my view, this is a good alternative to the main line 6 ... c5. Black
develops a piece and prepares ... e7-e5 in many positions since, as we will
see, f4xe5 ... d6xe5 d4-d5 is often White’s best response, when the knight
will find a good square on c5.
7 Bd3
The routine 7 Be2 gives less protection to e4 and allows the immediate 7
... e5! 8 fxe5 dxe5, when 9 Nxe5 (9 d5 is much better, but Black was okay
after 9 ... Nc5 10 Bg5 h6 11 Bxf6 Qxf6 12 b4 Na6 13 a3 c5 in Ba.Jobava-
Wan Yunguo, Vlissingen 2015; ½-½ in 79) 9 ... c5 10 Be3 looks like
excellent compensation for Black, though he needs to be precise to prove it:
10 ... Nb4! (the best move, when all the tactics work in Black’s favour) 11
Rc1 (or 11 0-0 cxd4 12 Bxd4 Nc2! and White is struggling to hold his
position together) 11 ... Ng4!! (never played over the board in my database,
but known from correspondence games) 12 Bxg4 Bxg4 13 Nxg4 cxd4 and
Black was at least equal in A.Herrera-G.Netolitzky, correspondence 2008 (0-
1 in 42).
7 ... Bg4 8 0-0 Nd7
 Black creates a tactical threat of taking on d4, and prepares to hit the
centre with ... c7-c5 or ... e7-e5.
9 Be3
This move is the most common by far, but it allows Black to execute his
idea.
Instead 9 Be2! is a good move, losing a tempo but placing his pieces
correctly and reinforcing d4. The best I’ve found here is 9 ... e5 10 fxe5 Bxf3
11 gxf3 (after either recapture with a piece on f3, 11 ... c5! is effective) 11 ...
dxe5 12 d5 Qh4 followed by ... Nf6-h5 with good play on the dark squares to
compensate White’s massive centre and bishop pair.
9 ... e5! 10 fxe5 c5!

A typical counterstroke in the centre. Black will gain good control of the
central dark squares, although White has his own trumps.
11 d5
Other moves:
a) 11 Nd5 cxd4 12 Bxd4 Nac5 13 Bc2 was J.O’Connor-S.Collins, Irish
Championship, Dublin 2014, and now 13 ... Nxe5 would have been very
comfortable for Black.
b) 11 Be2 Bxf3 12 gxf3 cxd4 13 Qxd4 (13 Bxd4 dxe5 14 Be3 Nac5 15
Qd2 Ne6 16 Rad1 Nd4 gave Black decent counterplay in C.Jepson-
J.Bejtovic, Swedish League 2014; 1-0 in 75) 13 ... dxe5 14 Qd6 Ndc5 15
Qxd8 Rfxd8 16 Nd5 Ne6 was a balanced endgame in S.Agdestein-
E.Gullaksen, Norwegian League 2005 (½-½ in 33).
c) 11 dxc5 dxc5 12 Be2 Bxf3 13 gxf3 Nxe5 14 f4 Nc6 15 e5 was
T.Taylor-D.Bojkov, Los Angeles 2011, and now 15 ... f6! would give good
counterplay against the white centre.
11 ... Nxe5 12 Be2 Nxf3+
The safest option. Black has had mixed results after the more dynamic 12
... Bxf3 13 gxf3 f5!?, but I think White is better after 14 f4!.
13 Bxf3 Bd7!

I prefer this move to the more common 13 ... Bxf3, since White struggles
to find a role for his light-squared bishop.
14 Qd2 Qe7 ½-½
The players agreed a draw here.
 A subsequent blitz game, between a great expert on the Four Pawns
Attack and a great expert on the King’s Indian, also suggests that the position
is balanced: 14 ... Qe7 15 Bg5 Qe5 16 Bf4 Qe7 17 Ne2 Rae8 18 Rae1 Bc8 19
a3 Nb8 20 b4 b6 21 bxc5 bxc5 22 Rb1 Nd7 23 Bg5 Bf6 24 Bh6 Bg7 25 Bg5
Bf6 26 Bh6 Bg7 27 Bg5 and a draw was agreed in A.Vaisser-M.Golubev,
playchess.com (blitz) 2004.

Game 28
A.Mastrovasilis-A.Naiditsch
Baku Olympiad 2016

1 d4 Nf6 2 c4 g6 3 Nc3 Bg7 4 e4 d6

Here I’d like to deal with some lesser played options which contain some
venom.
5 Bd3
I associate this line with GM Ivan Sokolov, and it has been developed
recently by GM Alexander Moiseenko.
5 Nge2 0-0 6 Ng3 is another sideline with some decent ideas. I haven’t
found a good way to avoid ending up in a Sämisch more or less a tempo
down after 6 ... c5 7 d5, so I think this is a good place to adopt an alternative
set-up with 6 ... e5 7 d5 a5, followed by ... Na6.
 For example: 8 Be2 (or 8 Bd3 Na6 9 0-0 h5 10 Bg5 Qe8 11 h3 Nh7 12
Be3 Qe7 13 a3 Bd7 14 Rb1 Bf6 15 Nge2 Bg5 16 f4 exf4 17 Nxf4 Nc5 and
Black was okay in M.Lagarde-J.M.Degraeve, French League 2015; 0-1 in 51)
8 ... Na6 9 h4 (9 0-0 h5 10 Bg5 Qe8 11 Qd2 Nh7 12 Be3 h4 13 Nh1 Bd7 also
looked fine for Black in V.Bologan-V.Kovalev, World Rapid Championship,
Dubai 2014; 0-1 in 42) 9 ... h5 10 Bg5 Qe8 (Radjabov notes that 10 ... Nc5?!
11 Bxh5! gxh5 12 Nxh5 is very dangerous for Black) 11 Qd2 was
Ma.Carlsen-T.Radjabov, Shamkir 2014. Radjabov went on to win a fine
game after 11 ... Nc5 12 0-0-0 Ng4 (0-1 in 51), but later suggested 11 ...
Nh7!? 12 Be3 Qe7 13 Nf1 f5 14 exf5 gxf5 15 Bg5 Nxg5 16 Qxg5 Qxg5 17
hxg5 a4 “and I am sure Black has strong compensation for the pawn on the h-
file.”
5 ... 0-0 6 Nge2 c5
Black has good options in 6 ... e5 or even 6 ... Nc6, but I prefer inviting a
Benoni formation here.
7 d5 e6
A familiar idea from the Sämisch and Averbakh variations.
 8 Bg5
This is the modern trend, pioneered by Moiseenko.
Black has few problems after 8 0-0 exd5 9 cxd5 a6 10 a4 Re8 11 Ng3
Qc7 12 h3 Nbd7 and then:

a) 13 Be3 h5 (a typical reaction against a knight on g3, as we saw in the

chapter on the Sämisch) 14 Re1 Rb8 15 Qe2 (after 15 a5 b5 16 axb6 Rxb6 17
Na4 Rb8, Black was no worse in M.Romanko-M.Simantsev, Serpukhov
2009, 0-1 in 62; while 15 Rc1 c4 16 Bc2 Nc5 17 Bd4 b5 gave Black great
counterplay in B.Taborov-M.Simantsev, Kiev 2004, 1-0 in 36) 15 ... h4 16
Nf1 Nh5 17 a5 Be5 18 Qg4 Bf6 19 Qe2 Be5 20 Qg4 Bf6 21 Qe2 Be5 22 Qg4
and a draw was agreed in R.Torma-M.Colpe, Budapest 2017.
b) 13 f4 (a more aggressive set-up) 13 ... Rb8 14 Be3 c4 15 Bc2 b5 16
axb5 axb5 17 Ra7 Qd8 18 Qf3 (after 18 e5?! dxe5 19 f5 b4 20 Nce4 Rb5,
Black was at least equal in J.W.De Jong-I.Cheparinov, Andorra 2002; 0-1 in
31) 18 ... b4 19 Na4 Bb7 20 Rc1 h5 21 Bf2 was V.Erdos-R.Lötscher, Szeged
2007, and now I like 21 ... Qe7 22 Re1 b3 23 Bb1 Nxd5!? 24 exd5 Qxe1+ 25
Bxe1 Rxe1+ 26 Kh2 Bd4, when Black can force a draw or play for more.
 8 ... h6 9 Bh4 exd5

10 exd5
Alternatively:
a) 10 cxd5 Nbd7 11 0-0 Ne5 12 f4 was tested a few times by Vladimir
Georgiev, but Black has reasonable chances after 12 ... Nxd3 13 Qxd3 Re8.
b) 10 Nxd5 has also been tried by Moiseenko; for example, 10 ... Nc6 11
0-0 (11 Nec3 Be6 12 0-0 Bxd5 13 Nxd5 g5 14 Bg3 Nxd5 15 cxd5 Ne5 was
okay for Black in A.Moiseenko-E.Bacrot, FIDE World Cup, Tromsø 2013;
½-½ in 23) 11 ... g5 12 Bg3 Nxd5 13 exd5 Nb4 14 Nc3 f5 15 f4 Nxd3 16
Qxd3 a6 17 a4 Qf6 and Black was fine in A.Moiseenko-R.Van Kampen,
German League 2016 (½-½ in 38).
10 ... Nbd7 11 f4 g5!?
 A very dynamic and concrete choice.
12 fxg5
Instead, 12 Bg3 Ng4 13 0-0 Ne3 14 Qd2 Nxf1 15 Rxf1 was
A.Moiseenko-E.Bacrot, FIDE World Cup (rapid play-off), Tromsø 2013 (1-0
in 36). Here Szabo recommends 15 ... Qe7 16 fxg5 (or 16 Re1 Re8 17 Nb5
Qe3+ 18 Qxe3 Rxe3 “and Black hasn’t got any problem”) 16 ... hxg5 17 Rf5
and now I like 17 ... Bxc3 18 Nxc3 f6, followed by ... Ne5, when the best I
can see for White is a draw: 19 Ne4 Ne5 20 Nxf6+ Rxf6 21 Qxg5+ Kf7 22
Qh5+ Kf8 23 Qh6+ Kf7 24 Qh5+ with a repetition.
12 ... Ng4 13 Qd2 Nde5 14 Ne4
On 14 Bc2 Szabo gives 14 ... f6! 15 Bg3 (after 15 g6?! Nxc4, or 15
gxf6?! Bxf6 16 Bxf6 Qxf6, “Black has a huge initiative for the pawn”) 15 ...
fxg5 and Black has good play.
14 ... hxg5?!
A mistake in a very sharp position. 14 ... Qe7 15 0-0 Nxc4 16 Bxc4 Qxe4
is fine for Black.
15 Bxg5 f6 16 Bf4 f5 17 N4g3
 Here 17 Ng5 looks clearly better for White.
17 ... b5!?
A typical counterattacking move, in the style of the Benko Gambit. 18
cxb5 a6 19 0-0 Nxd3 20 Qxd3 axb5 gives Black reasonable compensation for
the pawn.
18 0-0 Nxd3 19 Qxd3 bxc4

Black has equalized but goes on to lose his way in the complications.
20 Qf3 Rb8
20 ... Bxb2! 21 Rab1 Qh4 22 h3 Ne5 23 Qe3 Rb8 would leave Black with
his full share of the chances in an unclear position.
21 Nh5! Bh8 22 h3 Rxb2 23 Rae1! Qh4?
Now 23 ... Qe8 was best, but the position is very difficult for Black to
play.
24 Bxd6
All of White’s pieces participate in a winning attack. The execution (from
both sides) is not the most precise, but White gets there in the end.
24 ... Ne5 25 Qe3 Re8 26 Qh6 Rb7 27 Bxe5 Rxe5 28 Qg6+
 28 Qc6! wins.
28 ... Kf8 29 Neg3 Rf7 30 Qh6+? Bg7 31 Qc6 Bb7 32 Qd6+? Kg8?
32 ... Qe7 would keep the game going.
33 Nxg7 Rxe1 34 Rxe1 Kxg7 35 Re8?!
35 Re7! was a nice winning shot.
35 ... Qd4+ 36 Kh2 Qxd5 37 Nh5+ Kh7 38 Qg3 c3? 39 Nf4 1-0

Game 29
S.B.Hansen-R.Van Kampen
German League 2016

Here I want to examine some of White’s attempts with an early Bg5 or Bf4,
often in combination with e2-e3. These systems are not viewed as critical by
theory, but it’s important to have some idea of Black’s best set-ups and to
remain alert in the opening phase. I recall losing a painful game to IM
Vladimir Hamitevici in these lines a few years ago.
1 d4 Nf6 2 c4 g6 3 Nc3 Bg7 4 Nf3
This is the normal move order for players who like an early Bg5 or Bf4.
One point is to avoid a direct transposition into the Grünfeld with 4 ... d5, but
I also think White doesn’t gain anything by delaying or omitting Nf3. For
example, after 4 Bf4 0-0 5 e3 d6, White doesn’t have a good way to avoid
transposing to the early Nf3 lines, since 6 Nf3 and 6 h3 Nbd7 7 Nf3 are his
best options here.
The immediate 4 Bg5 0-0 gives White an interesting, independent option
in 5 e4 d6 6 Qd2, but Black obtains excellent play in Benoni/Benko fashion
with 6 ... c5 7 d5 e6 8 Nge2 (8 Bd3 exd5 9 cxd5 Re8 10 Nge2 Nbd7 11 0-0
a6 12 a4 Ne5 13 Bc2 Rb8 14 b3 b5 15 axb5 axb5 16 f4 Ned7 was excellent
for Black in A.Morozevich-V.Topalov, World Blitz Championship, Astana
2012; ½-½ in 56) 8 ... exd5 9 cxd5 b5! (yet again, we see Ding Liren
sacrificing his b-pawn for active counterplay) 10 Nxb5 h6 11 Bxf6 (if the
bishop retreats Black will win the e4-pawn with an overwhelming position)
11 ... Bxf6 12 Nec3 a6 13 Na3
 Question: What does Black have for his pawn?

Answer: We have seen dozens of positional pawn sacrifices in this book.

The present version shows, in a pure and uncomplicated form, an idea which,
I think, Ding Liren believes about the King’s Indian, namely that having an
unopposed dark-squared bishop on an open diagonal is worth a pawn. Many
of his games feature sacrifices to open up this bishop, including pawn and
exchange sacrifices. What is worth noting is how calmly he continues to
develop, being confident that he has full positional compensation for the
pawn.
R.Ponomariov-Ding Liren, Danzhou 2014, continued 13 ... Nd7 14 Bd3
(not 14 Qxh6?? Bg5 15 Qh3 Ne5 and Black’s initiative is decisive) 14 ... Ne5
15 0-0 Rb8 16 Nc4 Re8 17 Rfe1 Nxc4 18 Bxc4 Bd4 (an amazing post for the
bishop) 19 Rac1 Qg5! (happily playing the endgame, by analogy with the
Benko Gambit) 20 Qxg5 hxg5 21 b3 g4! (fixing targets in the white kingside
structure) 22 g3 Ra8 23 Na4 Rb8 24 Nc3 Ra8 25 Na4 Rb8 26 Nc3 and a draw
was agreed. I find it quite inspiring how Ding managed to completely
neutralize a sideline, against a very strong opponent, just using typical ideas.
4 ... 0-0
I think this is the most flexible move order, though some players prefer 4
... d6.
5 Bg5
After 5 Bf4 it becomes more attractive for Black to aim for ... e7-e5,
which will gain a tempo on the bishop and drive it to a passive position: 5 ...
d6 6 h3 (after 6 e3 Nbd7, White can transpose with 7 h3 or play 7 Be2 Nh5 8
Bg5 h6 9 Bh4, when the common 9 ... g5 is well met by 10 g4!, but 9 ... c5
offers Black good counterplay) 6 ... Nbd7 7 e3 (7 e4 allows the immediate 7
... e5! with equality) 7 ... b6 8 Be2 Bb7 and Black has a good position.

A few examples by KID specialists: 9 0-0 (after 9 Qc2 e5! White can’t
win a pawn on e5 since g2 would hang, and 10 Bh2 exd4 11 exd4 Re8 12 0-0
Ne4 gave Black an excellent position in V.Hamitevici-G.Jones, British
League 2013; 0-1 in 20) 9 ... Ne4 10 Qc2 (or 10 Nxe4 Bxe4 11 Nd2 Bb7 12
Bf3 Qc8 and Black was fine in R.Gajek-M.Bosiocic, St Veit 2016; ½-½ in
43) 10 ... Nxc3 11 Qxc3 e5! 12 Bg5 (not 12 dxe5? dxe5 and Black wins
material since 13 Nxe5 Nxe5 14 Bxe5 drops a piece to 14 ... Qg5) 12 ... f6 13
Bh4 g5 14 Bg3 f5 (White’s position is already unpleasant, but it went
downhill quickly) 15 Qd2? f4 16 exf4 gxf4 17 Bh4 Qe8 18 d5 Qg6 19 Rfe1
e4 20 Ng5 Ne5 21 Kh1 Bh6 and White resigned in M.Massoni-E.Bacrot,
Calvi (rapid) 2014.
5 ... c5 6 d5 d6 7 e3
Now if 7 e4, Black replies 7 ... h6 and then:

a) 8 Bf4 allows Black to sacrifice the d6-pawn with 8 ... e5! 9 dxe6 Bxe6
10 Bxd6 (10 Qxd6 Qa5 is dangerous for White) 10 ... Re8 11 Be2 (instead,
11 Bxc5 Na6 12 Qxd8 Raxd8 13 Be3 was L.Lengyel-L.Kavalek, Tel Aviv
Olympiad 1964, and here I suggest 13 ... Ng4 14 Bxa7 Nb4, when Black is
three (!) pawns down but has sufficient counterplay; the main line runs 15
Rc1 Bf5 16 Nd2 Bxc3 17 Rxc3 Bxe4 18 Nxe4 Rxe4+ 19 Be2 Nxa2 20 Rc2
Ne5 21 Be3 Nb4 22 Rc3 Nbd3+ 23 Kf1 Nxb2 24 Bxh6 f6 and Black will
collect the c4-pawn with a likely draw) 11 ... Qb6 gives Black good
compensation. Some examples: 12 Bxb8 (12 0-0 Rd8 13 e5 Ne8 14 Nd5
Bxd5 15 Qxd5 Nxd6 16 Rad1 Nc6 17 exd6 Nd4 was fine for Black in
A.Aleksandrov-R.Wojtaszek, Warsaw 2009; ½-½ in 29) 12 ... Raxb8 13 Qb3
(or 13 Qc2 Nh5 14 g3 Bh3 with good play for the pawn in J.Piket-L.Van
Wely, Wijk aan Zee 1997; 1-0 in 32) 13 ... Qa5 14 0-0 b5! with a strong
initiative in A.Konaplev-N.Matinian, Moscow 2016 (0-1 in 52).
b) 8 Bh4 is the most common move, but then 8 ... Qa5 seizes the
initiative, and 9 Nd2 g5 10 Bg3 Nxe4 11 Ndxe4 f5 12 Nd2 f4 13 Be2 Bxc3
14 bxc3 Qxc3 was excellent for Black at this stage in L.Bruzon Batista-
W.Arencibia Rodriguez, Cuban Championship, Santa Clara 2005 (1-0 in 39).
7 ... a6 8 Nd2

8 ... h6
The bishop is well placed on g5 and an unpleasant pin will be created
after Black plays ... e7-e6, so I think it’s worth kicking it back. As we will see
in the game, one of Black’s ideas is to break the pin with ... g6-g5, which
weakens his kingside but limits White’s dark-squared bishop.
9 Bh4 e6 10 a4
Otherwise Black could execute ... b7-b5 after ... e6xd5 and c4xd5.
 10 ... exd5 11 cxd5 Nbd7 12 Be2

From this position there are a number of different set-ups and move
orders which Black can utilize. I recommend using both the set-up and move
order adopted by Van Kampen, which involves putting the queen on e7
(rather than c7, where it is more passive), the knight on e5, and breaking the
pin with ... g6-g5. Thereafter Black has a couple of active formations, which
include: (i) ... Bf5 and ... Rae8 (as in the main game), and (ii) ... Ng6, ...
Nfd7-e5 and ... Nf4 (as in the Jobava game given in the notes).
12 ... Qe7
The following shows what can happen when White pushes f2-f4 with his
pawn still on e3: 12 ... Re8 13 0-0 Ne5 (I prefer this to 13 ... Rb8 14 a5 Qc7
15 e4 g5 16 Bg3 Ne5 17 Qc2 Nfd7 as in I.Sokolov-S.Drazic, Serbian Team
Championship 2010, since with the queen on c7 White is under less risk of a
kingside attack) 14 h3 g5 15 Bg3 Bf5 16 Bh2 Qe7 17 Kh1 Rad8 18 a5 Bh7
19 f4 gxf4 20 exf4 and now in N.Sulava-M.Apicella, European Cup, Kemer
2007, Black went for unclear complications with 20 ... Nd3 21 Nc4 Nxb2!?
(0-1 in 36), but 20 ... Ned7 is also good, for example, 21 Bc4 Ne4!? 22
Ncxe4 Bxe4 23 Re1 Nf6 24 Bd3 Bxd3 25 Rxe7 Rxe7 with a highly unclear
position in which Black has reasonable compensation for the queen.
13 0-0 Ne5
Here 13 ... Re8 led to similar play in H.Herraiz Hidalgo-Ba.Jobava, Dos
Hermanas (online blitz) 2006. After 14 h3 Rb8 15 Re1 Ne5 16 Bg3 g5 17 e4
Nfd7 18 Nf1 Ng6 19 Bd3 Nde5 20 Ne3 Nf4 21 Bf1 Qf6 22 Qc2 Bd7, Black
established a good set-up with counterplay and went on to win (0-1 in 34).
14 h3 g5 15 Bg3 Bf5 16 Kh1 Rae8

All of Black’s pieces are actively placed and he has decent chances in a
very complicated position.
17 Ra3 Qd7 18 e4
Not 18 f4 gxf4 19 exf4? in view of 19 ... Nd3.
18 ... Bg6 19 f4 gxf4 20 Bxf4 Nh7?!
This knight will be poorly placed for several moves to come. 20 ... Qe7,
followed by 21 ... Nfd7, looks like a more harmonious arrangement. With
such strong control over the e5-outpost I think Black is fine.
21 Qe1 Qe7 22 Rb3 Rc8 23 Qg3 Rc7 24 Rb6 Rd7 25 Nf3 Re8 26 Nxe5
Bxe5 27 Bxe5 Qxe5 28 Qxe5 Rxe5 29 Bg4

Exploiting Black’s lack of co-ordination and winning a pawn. However,

one of the characteristics of the King’s Indian is that even when Black goes
wrong and his position is poor, he typically retains some resources for
counterplay.
29 ... Rd8 30 Rxb7 Ng5 31 Bf5 Nxe4 32 Bxe4 Bxe4 33 Nxe4 Rxe4 34
Rfxf7
Or 34 a5 Ra4 35 Rfxf7 Rxa5 with a similar position.
34 ... Rxa4
A tough endgame to assess. Of course White’s construction on the
seventh is threatening and assures him of a draw in the worst case, but there
are no targets on this rank and White has some weaknesses too. Despite the
engine’s strong enthusiasm for the white side, I think Black has excellent
drawing chances.
35 Rg7+ Kh8 36 Rh7+ Kg8 37 Rbg7+ Kf8 38 Ra7 Kg8 39 Rxh6 Rd4
40 Rh5
After 40 Rxa6 Rxd5, Black will generate good counterplay with his
passed d-pawn.
40 ... Rd2 41 Rxa6 Kg7 42 Ra7+
Or 42 Rb6 Kg6 43 Re5 Kf6 44 Re6+ Kf5 45 Rbxd6 Rxd6 46 Rxd6 Rxb2
and despite the two pawn deficit, the extremely active black pieces in
combination with the passed c-pawn seem sufficient to hold the balance.
42 ... Kf6 43 Rah7 Rg8!
A precise move, preparing to meet Rh6+ with ... Rg6.
44 g4 Ra8 45 R7h6+ Kg7 ½-½
Index of Complete Games
Baldauf.M-Radovanovic.N, Novi Sad 2016
Barbot.P-Van Foreest.J, World U-20 Championship, Khanty-Mansiysk
2015
Efimov.I-Mazé.S, Francophonie Championship, Menton 2016
Fier.A-Vakhidov.J, Hastings 2014/15
Ftacnik.L-Bacrot.E, German League 2015
Gelfand.B-Nakamura.H, World Team Championship, Bursa 2010
Giri.A-Bacrot.E, European Cup, Bilbao 2014
Gupta.A-Kurnosov.I, Nakhchivan 2013
Hansen.S-Van Kampen.R, German League 2016
Jakovenko.D-Fedoseev.D, Taganrog 2011
Kasimdzhanov.R-Nakamura.H, FIDE Grand Prix, London 2012
Kelires.A-Ziska.H, European Small Nations Championship,
Luxembourg 2016
Komarov.D-Motylev.A, Niksic 2000
Krassowizkij.J-Van Kampen.R, German League 2016
Lopez Martinez.J-Cuenca Jimenez.J, Roquetas de Mar 2015
Maghsoodloo.P-Tabatabaei.M, World U-16 Championship, Khanty-
Mansiysk 2016
Mastrovasilis.A-Naiditsch.A, Baku Olympiad 2016
Matnadze.A-Ushenina.A, European Women's Championship, Mamaia
2016
Mchedlishvili.M-Bacrot.E, Tromsø Olympiad 2014
Mirzoev.A-Kovalev.V, Turkish League 2016
Movsesian.S-Ding Liren, World Team Championship, Antalya 2013
Naumann.A-Naiditsch.A, German League 2016
Neverov.V-Nenezic.M, Obrenovac 2011
Ni Hua-Ding Liren, Chinese Team Championship 2015
Ragger.M-Grischuk.A, European Cup, Skopje 2015
Rowson.J-Jones.G, British Knockout Championship, London 2015
So.W-Nakamura.H, Sinquefield Cup, St Louis 2015
Vazquez Igarza.R-Nakamura.H, Gibraltar 2016
Zviaginsev.V-Akopian.VFIDE World Rapid Championship, Khanty-
Mansiysk 2013