John Cox The Berlin Wall QUALITY CHESS g £Games/Chess The Berlin Wall The Berlin variation of the Spanish is one of the most popular openings among world class players. After 1.64 e5 2.4 f3 Aic6 3.2b5 Afé Black can play a completely sound opening based mainly on understanding rather than memorizing theory. Many opening books make this claim, but the scarcity of forcing lines in the Berlin means that in this case it is true. The trick is to gain the requisite understanding, and this is where John Cox’s eloquent prose comes into its own. After reading his explanations it will be clear why this robust opening has gained the nickname The Berlin Wall. “You learn so much about chess from reading this book it is amazing! It is ‘incredibly well-written; careful attention is given to all relevant details; everything that deserves explanation is explained, and everything that needs analysis is analyzed. Even if you don't have any interest in this opening, you should buy this book. If you are interested in becoming a better chess player and want to understand the pawn structures in the Exchange Ruy Lopez or Berlin Wall, this book will carry you far. It is truly fantastic and should be a candidate for book of the year. 2 kk &" - FM Carsten Hansen, ChessCafe.com + Expert guidance on one of the soundest openings + Play the opening that Kramnik used to dethrone Kasparov + Atrustworthy opening that requires little memorization of theory John Cox is a lawyer and International Master from London. He is an experienced author whose previous books have received high praise. €23.99 £16.99 $29.95 ISBN 978-91-857790-2-4 95 529 & QUALITY CHESS Mil < = 9°789185 779024 www. qualitychess.co.ukBibliography Introduction Part 1 ‘Understanding the Berlin Wall Chapter 1 Positional Introduction Chapter 2 Typical Berlin Endings Pawn Endings fhe Endings Knight vs. dark-squared Bishop Knight vs. light-squared Bishop Bishop vs. Knight Bishops of the same Colour Opposite-coloured Bishops Rook Endings Dark-squared Bishops & Knights Opposite-coloured Bishops & Knights ‘Two Knights vs. Bishop & Knight The Two Bishops Rooks & dark-squared Bishops Rook & Knight vs. Rook & Queen's Bishop Rook & Knight vs. Rook 8& King’s Bishop Double-rook Endings Rooks & Bishops of opposite Colours Endings with three Minor Pieces each Chapter 3 Positional Themes ‘The e6 Break Sactificing e5 Actacks with the Rook Pawns The QdG/Gt Trick ‘The Exchange Sacrifice Capturing on d5 13, 21 21 23 26 27 Bos 41 47 51 53 55 60 61 63 66 67, 69 74 79 79 84 93 94 975 vacseill Fe (Chapter 3 continued) The g4 Trick The Classical Blockade Set-up Black's ...d4 Correct and incorrect handling of Black’s Queenside Pawns Black's ...<2ec6! Restraining g4 ~ forestalling it with ...h5-h4 Striking back with ...h5 after g4 Black's 4 Break: a Controversial Undoubling 5 Break . Developing Black’s Rook with ...a5 Black's ...2xe5 Trick Blacks ...Lbxe3 Black's early King Activation Part 2 ‘The Theory of the Berlin Wall Chapter 4 --De7 Systems without h3 Chapter 5 «-e7 Systems with h3 Chapter 6 .«.De7 Systems without an immediate ... Ag6 Chapter 7 : Chapter 8 whe? Systems Chapter 9 Berlin Endgame: White Alternatives and Miscellaneous Black Systems Chapter 10 White Plays 4.3, Chapter 11 Other White Tries Index of Variations Index of Theoretical Games 99 100 107 108 109 110 114 116 123 127 129 131 132 oo. 134 136 os 141 159 181 205 233 oo. 281 295 323 332Books: Bareev/Levitov: From London to Elista, New in Chess 2008 (comments on the Kasparov-Kramnik games that are rather different from Kramnik’s) Bologan: Victor Bologan: Selected Games 1985-2004, Russell 2007 Bronstein: 200 Open Games, Dover 1992 Comas Fabrego: True Lies in Chess, Quality Chess 2007 Davies: Play I e4 €5!, Everyman 2005 Gershon & Nor: San Luis 2005, Quality Chess 2007 Greet: Play the Ruy Lopez, Everyman 2007 Hansen, L.B.: Secrets of Chess Endgame Strategy, Gambit 2006 (some interesting general reflections on the Berlin and a chapter annotating six games, plus some examples in the general text) Kaufman: The Chess Advantage in Black and White, McKay 2004 (repertoire coverage for Black of ...£1€7 systems, assisted by Sherzer) Khalifman: Opening for White According to Anand: Volume 1, Chess Stars 2003 (biased but authoritative, wide-ranging and interesting repertoire coverage for White) Shamkovich: The Chess Terrorists Handbook, American Chess Promotions 1995 (some little- known ideas in lines with 5.Ze1 and 2d3) Seohl: Garry Kasparov's Greatest Chess Games: Volumes 1 and 2, Gambit 2005 and 2006 Periodicals: Almasi: New in Chess Yearbooks (various) Chess Informant 1-98 Electronic Resources: ChessPublishing.com Kramnik: My Path to the Top, ChessBase DVD 2007 (great comments on Kasparov-Kramnik Berlins from 2000) Ree: Dutch Treat column, ChessCafe.com Mega Database 2008 (various annotators including Wedberg, Stohl, Postny, Muller and Hecht) ‘The Week In Chess Of the above Almasi, Khalifman and Kaufman are the main theoretical sources, and Kramnik by far the most relevant of the others.To anyone who has followed top-flight chess in the 21st century, the Berlin will need little introduction. Since Vladimir Kramnik brought it out of retirement in his 2000 world title clash with Kasparov, where the opening played a huge part in his victory, a clear majority of the world’s top 100 players have incorporated it into their repertoires to a greater or lesser extent, Its most high-level patron today is Kramnik’s great rival and stylistic antipode Veselin Topalov, but Levon Aronian, Ruslan Ponomariov, Magnus Carlsen, Sergei Karjakin and others of the most diverse styles have used it. ‘The main reason is simple, First and foremost, it’s a very strong and solid opening which gives White huge difficulty in showing any advantage. But the Berlin also has a,couple of more practical merits which make it a useful addition to anyone's repertoire, First, it dictates the nature of the play. Once 3...f6 hits the board, White really doesn’t have too much alternative to going in for the so-called ‘Berlin endgame’ if he wants to play for an edge. So clearly if White is an attacker who hates the idea of a queen exchange and plays better in tactical positions, the Berlin is a handy option. Second, i's very easy to prepare, and very hard to prepare against. There isn't any theory which Black absolutely has to know, and once Black has a good general idea of what he’s trying to achieve in the opening and what he’s trying to avoid, he won't do too badly if he knows no theory at all. White, on the other hand, cannot simply learn a line and play it. For one thing Black has a great flexibility in move orders and plans, and for another learning a line isn't enough; you have to know what you're doing, and the Berlin is deceptive; White can be in trouble before he knows it. You only have to look at the many games solid grandmasters lost against it in the 1990's to see how difficult it can be to orientate yourself in the resulting positions if they're not familiar to you. Nowadays they are much more familiar to professionals, but at a lower level it is still common to see White players flailing hopelessly. So even if you don't want to play the Berlin all the time, it’s a very useful shot to have in the locker. ‘There are grandmasters who despite its drawish reputation and the early exchange of queens are happy to play it even against weaker opponents in open events, but even they usually choose their victims with care (juniors are a favourite). Which brings me on to the main objection players tend to express to the opening; isnt it dull as ditchwater? Drawish as anything?8 The Berlin Wall Well, I hope to convince you that it isn’t dull. Dull is in the eye of the beholder. When I told one British GM and Berlin expert chat I was contemplating taking the opening up and writing this book he pursed his lips and looked dubious. ‘Irs very sharp, you know’, he said. Not altogether a compliment, I felt, but a wise observation. Its true that if White wants to get something against the Berlin he often has to do something quickly and surprisingly critical positions can arise quite early. I've certainly enjoyed playing the opening, but it’s true that you'd better like playing endgames and be prepared if necessary for a long positional struggle. Irs more drawish than the Dragon Sicilian, no doubt about it, but if you want a sound, solid game with Black, chen you have to accept that the value of each move is going to be less than in sharper openings, and hence the margin of the draw will be wider. I's always handy to havea hero in an opening, a player whose games you can look at for inspiration and education. Its obviously hard to look past two towering figures such as Topalov and Kramnik, but I would also suggest Alexei Aleksandrov of Belarus and Zoltan Almasi of Hungary, who have both been playing the opening since the early nineties and have produced many model victories with it, and for the ...8e7 systems specifically, the young German Leonid Kritz, who is a virtuoso in these. ‘The opening is called the ‘Berlin’ defence after its advocacy by the Pleiades, a collection of scars from that city in the middle of the nineteenth century, including von der Lasa and Bilguer. It was popular at the curn of the century, when Tatrasch and Lasker played some surprisingly modern games from the Berlin Endgame position, but then fell into disuse, althqugh occasionally played (notably by Bisguier), until che 1990's, when Alexandrov, Almasi and to a lesser extent Tony Miles began to demonstrate the possibilities of the opening. The 2000 London match, however, produced an absolute explosion of interest which took it from the position of an occasional and slightly inferior try to one of the most respected defences Black has to 1 e4: many more grandmaster games, over ten times as many I suspect, have been played in the last ten years with the opening as in the preceding hundred and fife. A word about computers and the Berlin: Kramnik says that one of the reasons he chose the opening against Kasparov was to neutralise the latter’s legendary expertise with the machines. In 2000 I gather the beasts still really steuggled with the opening; these days they are better. In writing this book I used Rybka 2.2 and Fritz 9 and 10. My experiences have been thus. In Berlin blockade-type positions Rybka is almost always about half a pawn more optimistic for Black than Fritz. With Fritz, once Black gets the evaluation down to only about +0.75, generally he is fine in such positions. With Rybka, on the other hand, you need to be down to 0.25 or so. Both machines have a much greater tendency than human players to make the move @h2 for White as preparation for g4 (and here I suspect they are right and this is often a more useful semi-waiting move than humans realise). And both cend to overestimate the sort of play Black can generate with moves like ...2d5, ...2e5+ or ..h4 after a g4 hg; hg sequence. If this doesn't do something quickly chen quite often it's scary but White can push Black back; often the evaluation improves steadily for White as one goes on in such a position. In the endgame Rybka has a horrible tendency to overrare Black’s defensive chances because it doesnt realise fast enough that a pawn endgame is lost; Fritz is much better like this.Introduction 9 The structure of the book is as follows. Chapter 1 isa brief positional introduction and explanation of what I consider the basic principles of the opening. Chapter 2 deals with typical endings arising from it. To a greater extent than most openings, you need to know where you're trying to arrive before you work out how co get there. Chapter 3 deals with typical positional themes. These two chapters are in some ways the most important of the book; if you want to take up the opening thorough study of them and a glance at the theory is probably more useful than equivalent study of the theory and a glance at chapters 2 and 3. Chapters 4-6, 7 and 8 deal with Black’s three main set-ups in the so-called Berlin Endgame (the position which arises after 1.e4 €5 2.563 Dc6 3.865 Df 4.0-0 Dxe4 5.d4 Dd6 6.8xc6 dxc6 7.dxe5 D5 8.Wxd8t Hxd8). These are respectively ...e7 plans, ...d7 plans and ...8e7 plans, and I attempt so far as possible to give helpful coverage of both sides’ ideas, Chapter 9 deals with lesser options for Black in the endgame and also with irregular White plans at move 9. Chapter 10 deals with White's most dangerous alternative to the endgame, 4 d3, and is repertoire coverage only from Black's point of view (advocating what one might call the pure Berlin move 4...8c5), and Chapter 11 deals with White's other tries, again strictly on a Black repertoire basis. Iconly remains for me to thank Jacob Aagaard and John Shaw for their faith that | would eventually produce this book in spite of the disgraceful stream of broken promises and missed deadlines I subjected them to, Andrew Greet and Jonathan Parker for their very helpful comments on their own games and in particular The Berlin Bishop Endgame (see pages 38 and 35), the on-line community at chesspub.com (especially ‘Bibs’ and ‘micawber’ for some helpful suggestions about sources and saving me from analytical error respectively), Richard Palliser for his usual generous assistance with computers and databases and so on, and Mark Ferguson and Jonathan Parker again for various inspiration and insights into the opening,Positional Introduction ‘The heart of this book is the position which is reached after the moves 1.e4 €5 2.063 @c6 3.2b5 DFG 4.0-0 Axed 5.d4 Dd6 6.8xc6 dxc6 7.dxe5 DFS 8.Wxd8t doxd8, the so-called Berlin Wall. Chapters 4-9 attempt to pro- vide comprehensive analysis of this position from both sides, while Chapter 2 deals with typical endings arising, and Chapter 3 with typical middlegame themes. After 3...0f6 this sequence is usually considered White's only serious try for advantage, and Chapters 10 and 11 provide repertoire coverage only from Black’s point of view of the various de- viations White has between moves four and eight: with the exception of 4.d3 these are more common at club level than international, level. Let us go through the initial eight moves and see why this might be so. L.e4 €5 2.63 De6 3.2b5 AFG Iewas Morphy who first proposed that Black ought to insert 3...a6 in order to give himself the option to break the pin quickly. Basically the great man was right: the only variations of the Ruy in which Black does not benefit from having ...a6 &a4 thrown in are those in which the bishop is at some moment attacked on b5, the Bird (3.,.@d4), the Schliemann (3...f5), the Classical (3...8c5) and the present debut, and those in which Black does nor intend to fight against the formation of the c3/d4 pawn centre and does not want to push the bishop towards its ideal spot on c2, the Cozio (3...Dge7) and the variously-named 3...g6 lines (Pillsbury? Smyslov?). 4.d3 and 4.We2 are respectable ways to avoid Black’s main idea and are dealt with in Chapter 10 and Game 55 respectively. The former en- visages either the old Steinitz plan with d3/c3 Dbd2-F-g3 before castling or else a build-up with c3 and d4, the latter perhaps 0-0/Ed1/ ©3/d4 along the lines of the Worrall Atcack in the normal Closed Ruy. However from a14 ‘The Berlin Wall logical standpoint 4.43 should not be the most critical test: if the game had gone 3.2c4 Df6 4.d3 then most people would not think that White was opting to press Black particularly in the opening, while if he isn't going co exploit the pressure created on the e-pawn by 3.2b5 to force the concessions (queenside weaknesses or surtender of central space, basically), which are typical of the main lines of the Ruy, then it’s not clear why White put his bishop on b5 instead of c4 at all. 4.2c3 is the Spanish Four Knights, which could of course have arisen by 3.8c3 @f6 4.8b5, and is not covered in this work: read- ets are referred to grandmaster Mihail Marin’s recent Beating the Open Games for (excellent) coverage. 4.d4 (game 56) is the Central Attack, and is not so effective before Black is committed to ...d6 and can still go ...d5 in one, as the traditional reply shows: 4...exd4 5.0-0 a6 6.824 Be7 7.05 (after 7.2e1 b5 8.2b3 d6 White’s tragedy is that 9.@xd42? falls into the Noah’s Ark trap with 9...Qxd4 10.fxd4 c5 and ...c4, so he has either to gambit a pawn for vague com- pensation only with 9.c3, or else give up the bishop with 9.2d5) 7...e4 8.@xd4 0-0 9.F5 d5. 4.8xc6 (game 57), like in the Exchange Vari- ation (3...a6 4.8xc6) is not so bad, but obvi- ously Black would rather have played 3... than 3...06. 4...Dxe4 Were Black to play 4...2e7 now, analogous to the normal Chigorin defence with ...a6/2a4 added, he would quickly find out the wis- dom of Morphy’s advice: White continues with 5.#el defending his own e-pawn and so threatening to win a pawn by &xc6 and ®xe5, thus forcing 5...d6 6.d4 renews the threat and forces 6...8d7 if Black wants to maintain a pawn on e5, and now after 7.2c3 Black finds thar 7...0-0 loses material after 8.8xc6 Bxc6 9.dxe5 dxe5 10.Bxd8 Haxd8 11.8xe5, and if 11..Bxe4? 12.Dxe4 Ded 13.0d3 65 14.63 Qh4 15.g3, the famous Tarrasch Trap, and so he is forced to cede central space to White with 7...exd4, transposing to the old Steinicz defence. 4,..2c5, the Classical Berlin, is another reason- able line which is not covered in this book, but by omitting ...6 Black usually telegraphs his intention to play the text. Black plays ina way akin to the Open Defence (5...xe4 with the inclusion of ...a6/2a4). 5.d4 5.Hel (games 58-59) is possible and is in some ways the most natural move. In the nor- mal Open Defence this move is rubbish be- cause 6...2c5 attacks the bishop on a4 and simply trades it off with a slight edge for Black. Here Black has to go 5...d6 to gain the same tempo, which of course blocks his develop- ment and gives White possibilities, but even so it turns out that Black’ difficulties can be fairly easily surmounted. 5.€e2 is also possible and is dealt with in Game 60. \d6 ‘This move, the trademark of the Berlin Wall, was the whole point of leaving out ...aG. Both here and in the Open Defence proper 5...exd4 is frowned upon because of the hair-raising se- quence 6.8e1 d5 7.Oxd4 £d6 8.2xc6 &xh2+ 9.2h1 Wh4 10.Bxe4> ded 11.Wd8t Bxds 12.Qxd8+ txd8 13.cexh2, so in the Open proper Black normally plays 6...b5 to enable .d5 (in fact he can try to reach the same posi- tion here by 5...a6 6.84 b5 7.£b3 d5). The text move however hits b5 and threatens toChapter 1 - Positional Introduction 15 consolidate Black’s gains with ...e4, so White is forced to concede the bishop. Black can also try the strange 5...2e7 6.8el 2d6 7.8xc6 bxc6 8.dxe5 ®b7, known in some circles as the Rio de Janeiro varia- tion (although properly this refers to a Black plan later on). This book does not cover this option. 6.Lxc6 Since 6.224 allows Black easy equality after 6...exd4 or 6...e4 White doesn't seem to have much choice, but in fact 6.dxe5 Dxb5 7.24 (or 7.c4) is possible since the knight is trapped. In either case Black has the choice between returning the piece with equality or accepting a risky pawn sacrifice: see Games 61 and 62. White actually has still another try in 6.85 (game 63), which again gives Black a choice between steady play with 6....2e7 or accept- ing the challenge and the piece sacrifice with 6...f6, when White probably does not have enough compensation. 6...dxc6 7.dxe5 7.Bxe5 (game 64) is utterly feeble; White should obviously translate his d-pawn to the kingside to obtain a working majority there compared to Black's crippled one, as in the Exchange Variation, not to mention displac- ing Black’s king by the forthcoming queen ex- change. 7 DBS 7..4e4 is a dubious alternative virtually re- futed by 8.Me2 2f5 9.8d1 Wc8 10.Dd4 Bes 11.b4 &b6 12.f3 Wd7 13.223 0-0-0 14.24 (Khalifman) and is not covered. 8.Wxd8t Nothing else makes a lot of sense. 8.¥e2 is often played by White players with Oedipus complexes, but the whole point of 7...Df5 (as opposed to 7...Ae4) was to meet that with 8...Dd4 9.Dxd4 Wxd4, when Black can trade the queens anyway if he wants to after 10.8d1 &g4, and obtain comfortable play (game 65). And here we are. I hope the above preamble has convinced you that this position is critical for 3...2f6, so it makes sense to take a long look at it, especially since in my opinion most texts fundamentally mis-state where Black’s advantages lie. White's assets are fairly clear. One, he is ahead in development (and also in space). Two, if all the pieces but the kings were magically removed from the board Black would have to resign. And three, Black's king is stuck in the centre of the board and will almost always block at least one of his rooks from en- tering the game along the back rank for some time to come. ‘These considerations suggest that White will win games in this opening in two ways: first by obtaining the initiative, perhaps by opening the centre with a pawn sacrifice, and exploiting his active pieces to force decisive gains, second- ly by slowly and carefully exchanging pieces, advancing his majority, creating a passed pawn and winning the ending. This impression is more or less correct and,16 ‘The Berlin Wall of the two, the former is the more common. It is a great mistake as White in this opening to imagine that exchanges will automatically lead the game towards victory. Just as fre- quently they serve the opposite purpose of dis- sipating White's initiative and converting the centralised Black king from a weakness into a strength. Moreover, when we consider Black’s strengths we will see that the most fundamen- tal of these is the possibility of blockade. More often than not, once Black manages to erect his fortifications White's chances of victory are negligible. White must strive if at all possible to strike before then. ‘And what then of Black? Optically his most obvious asset is the ewo bishops, and this is what many manuals focus upon. However, if this is where his compensation for his other difficulties cruly lies, why is it that Black is so frequently anxious ro exchange his dark- squared bishop for an enemy knight? In fact classical wo-bishop compensation in an open position is quite rare in the Berlin, although there are examples (especially in chapter 8). The matter can best be understood by com- paring the position with the structure arising in the Exchange Ruy proper, which has the same pieces and the same basic structure save that White’ e-pawn has remained on e4. The differences are wholly in Black’s favour. ‘They are these. First, the cl-bishop is obstructed. Try and find a good place for it. You won't suc- ceed. There isnt one. On d2 or €3 it blocks a file; on £4 or b2 itis blocked by its own pawn, and on g5 it Moats vaguely and merely encour- ages the useful ...h6. Second, White's pawn majority is tendered annoyingly inflexible. In the Exchange Ruy proper he has a number of plans; for example he sometimes sets up the Lasker — Capablanca structure with e4 and 5, playing for space. Or he can support the e-pawn with and advance with g4. Here, assuming that Black plays ...h5 and when called for ...g6, White has basically only one option (especially since in a simplified end- ing he normally finds himself forced into £% to protect e5); first he has to arrange g4, then 5, then e6. ‘That gives him a number of problems. First of all, each time he wants to advance a pawn, say (after f4 and ...h5) he wants to play g4, then so long as Black still has his light-squared bishop, White has to tie up two pieces, one to defend h3, the other g4, Second, Black can pur a rook in ambush on the h-file (or more frequently just leave the one he already has there where it is), and the moment White plays g4, Black can exchange the h-pawns and have the active against the passive rook. Third, White may find that the advanced e-pawn causes still other annoyances. Very often White finds himself constrained in the near future to play h3, partly to prevent ...g4, which might otherwise threaten co take the knight on {3, which itself cannot move because it has to defend e5. Frequently Black then plays ...h5-h4, and now White has no pawn move on the kingside at all which doesn't cre- ate more pawn islands and threaten to trans- form his majority into a mere collection of weaknesses This last paragraph (and the name of the open- ing) give one the clue to how Black’ assets can most simply be expressed: his possibilities of blockading and then playing against the white pawn majority. Fundamentally intertwined with this point is the fact that Black’s light- squared bishop is the best minor piece on the board. From e6 it assists tremendously in lam- ing White's majority, and it also has possibili ties on the queenside. Those white queenside pawns will need defending, Black's ideal ar- rangement of his own pawns is a5/b6/c5/c7/ £7lg7(g6)/h5(h4). If one imagines that Black has set up thus and that White's a2- and c2- pawns have at some point been attacked byChapter | - Positional Introduction 17 the bishop, then generally White can do one of two things. Either he can advance them to a4 and c4 and defend them by the b3-pawn, or (if he has not developed his bishop on b2) he can place the a and c-pawns more mod- estly on a3 and c3, protected by the pawn on b2. The former is usually fine in the middle- game, but in a simplified endgame the need to defend the pawns constantly against a bishop sortie to ¢2, or perhaps a knight coming to d4 which cannot be taken for fear of the resulting passed pawn and dark-square weakness, fur- ther ties White's hands when he is trying to advance his own pawns. On che other hand, the a3/b2/c3 option weakens b3 horribly and may allow annoying moves like ...b3 trying to take the d-file in a middlegame, or more seriously give a pathway to the black king in an endgame. ‘When ic goes well, Black’s play in the Ber- lin gives perfect expression to Philidor’s rule of two hundred years ago: the pawns should go on the same colour as the bishop on the flank where you are defending, but on the op- posite colour where you are attacking. Black must always have it in mind co use his own majority actively. The only ching wrong with it is thar it cannot create a passed pawn. The good thing about it is thar it can gain space without leaving weaknesses behind it, and by threatening to undouble itself the enemy's actions are constrained (for example White can hardly ever leave his pawns on a2/b3/c2 because of the threats of ..a5-a4 and ...c4. We shall see examples of all these ideas in chap- ter 3 on positional themes, For the moment though, my poinc is that the Berlin can best be chought of as a batdle between competing ma- jorities. Black has made an obvious concession with regard co his own majority, but he has acquired a number of subtler but real advan- cages, which assist him in blockading Whites majority and erecting the so-called Wall. Besides, Black’s majority is not so helpless as it appears. In chat context the answer to my ear- lier rhetorical question becomes apparent: we can now see why Black’s dark-squared bishop is the piece he would most like to exchange, and preferably for a knight. White's knights can help in advancing his majority over g4 and f5, his bishop cannot. Meanwhile Black’s bishop can find little role in constructing the light-squared blockade, and offensively too it cannot help much in advancing Black's own majority; against an a4/b3/c4 structure it is confined by its own pawns on a5 and c5, and against a3/b2/c3 it is restrained by the enemy pawns and can seldom find a way past White's queen's bishop to b2. From the previous diagram White usually plays 9.c3. Black then has a multitude of plans which I will try co explain in the indi- vidual chapters as we go along, but broadly chapters 4-6 deal with the immediate 9...2e7, chapter 7 with the move 9...2d7, intending co hide the king on the queenside (this was the plan with which Kramnik revived the Berlin in 2000, although he was not the first to play id, chapter 8 with various ideas with ....2€7, whose most obvious idea is to trade off the f5-knight on h4, and chapter 9 with three slighty more offbeat systems; 9...22€6, whose Unique Selling Point is the inten- tion (o recapture on e6 with the king should White harass the bishop with a knight and 9...t8e8 followed either by 10...5 or 6, the former largely a waiting strategy albeit with some subtle positive points, and the latter intended to force some sort of con- cessions by the idea of ...2a6. Let me conclude this section by showing a short game in which White doesn't play too well and things donc ger as far as Black need- ing to bother restraining White’s majority, but which illustrates perfectly the offensive op- portunities the opening offers the unopposed 8-bishop, the real hero of the Berlin.18 ‘The Berlin Wall Pruijssers ~ L’Ami Amsterdam 2006 leh 5 2.063 Dc6 3.4b5 Df 4.0-0 Dxed 5.d4 Dd6 6.2xc6 duc6 7.dxe5 DFS 8.Yxd8t thxd8 Irs noteworthy that LAmi, a young Dutch GM with a rating of 2585 at the time of the game, was perfectly happy to play the Berlin even in an open tournament against an oppo- nent whom he must have been keen to beat (Pruijssers is a Dutch junior with a rating of 2355). There's nothing necessarily drawish about an exchange of queens. 9.Dc3 De7 A strange-looking move, but on g6 the knight ties down the f3-knight and so plays its parc in restraining the majority, since f4 is kept impossible, whereas on {5 the piece was merely a target for g4 at some moment. 10.h3 Dg6 11.De4 a bede One of many moves: probably not the most pressing, but the knight looks pretty enough here. 11...h6 Black's first priority in the Berlin is often to get in this move and his next, in order to pro- rect the bishop on 6 from being attacked by a knight (in conjunction with the knii 6 covering £4). Once this set-up is achieved Black is already virtually fireproof. 12.b3 White often develops the bishop like this, but he isn’t pressing Black quickly enough. 12...c5 One of the drawbacks of the early ed was that White allowed this move too easily: with the knight on c3 it can often be well met with @b5 or Ads. 13.04 White doesn't want to allow ...c4 undou- bling the pawns (which Black would surely not have played for some time to come any- way, since the price of undoubling is losing the important control over d4 the pawn gives him, and the consequent security for his e6-bishop), but this move doesn’t do anything positive for White's game. 13...8.€6 14.862 sed7! If you saw that coming then either you've been around the Berlin block a bit or you ought to give ita try. In fact c6 is often a dream square for Black’s king in the Berlin. It’s enor- mously safe there (if you gave White a hun- dred free non-capturing moves with his pieces he still couldn't even check it sensibly); it con- nets the rooks quickly, and it is well placed to assist with either offensive action on the queenside or defensive play from 6 in the case of simplification. More often, though, it has to trek round via c8-b7. The direct route saves wo tempi (since ...b6 isn’t necessary), but nor- mally White has stirred himself enough to put a rook on dl before now (which in fact means that ...2d7 and later ...2€6 is necessary, soChapter 1 - Positional Introduction 19 perhaps Black is now three tempi up on where he should be). 15.Bad1t Too late. 15...che6 16.03 White notices the d5-square and directs his knight there, but it all cakes time, and any- way what exactly does the knight do on that square? 16...2¢7 Black doesn’t often manage to connect his rooks so quickly and easily. 17.0d5 a5 fg h abe Black waited for the knight to leave c3 before playing this. White seldom wants to allow ...a4 giving Black the option to open the file when he chooses, but a4 has its own problems. 18.Dh2 White readies himself for 4-5, firing up his own majority, but his timing is off. Black is already more than fine, but White needed to play Bd? so as at east to keep his knight on d5. 18...2hd8 19.2xe7¢ White perceives this as necessary, since d5 is en prise, but in fact he might as well have pressed ahead with 19.4 &xd5 20.cxd5t ixd5 21.8xd5 thxd5 22.f5 2B 23.16 gxf6 24.exf6 2d6 25.2g4, when he has at least good chanc- ¢€s to cause some confusion. The exchange in the text is meat and drink for Black. 19...)xe7 20.04 This really is a dreadful move: White still needed to give 20.F4 a go, although it isn't very pleasant, for example 20...h5 21.g4 g6 22.cbf2 (22.24 hxg4 23.hxg4 Bh8 24.8d3 Bhd 25.8g3 Bd8 is also vile) 22...04 23.263 axb3 24.axb3 bS. See how effective Black's bishop is on both flanks, and also how good the king is on c6. 20...£165 And that’s it. White can’t avoid ...{2c2 fol- lowed by the removal of most of his queen- side. 21.6 21.Bcl Qg6 followed by ...Df4 and ...8d3 only delays the inevitable, although all the text achieves is to ensure that there will be another pawn on the menu when Black gets around to taking it. 21.6 22.Bdel ‘A blunder, but it no longer makes any dif- ference. 22.8xd8 Bxd8 23.8e1 S&c2 was hardly any better. 22...8d3 0-1 A slight game, but it illustrates as well as any other exactly for what it was that Black al- lowed his pawns to be crippled and his king inconyenienced.Typical Berlin Endings In this chapter I want to start by looking at the opening backwards. You can't learn any opening without considering the typical endings it gives rise to, and this is more true of the Berlin than most. Lets start at the end with pawn endings with the typical Berlin Wall pawn structure. . Pawn Endings ‘There would be many good reasons to call David Bronstein back to life, but one small question | would have for him would be why he wrote, in 200 Open Games, “IF you have time, check whether the pawn ending is won. ‘That's a very difficule problem, bur there is a solution.” I have a feeling I must be missing the great man's point. In any case according to me the pawn ending is generally hopeless for Black; White creates a kingside passer, decoys the black king with it and wins on the queenside in classical Ruy Exchange style. The only thing he has to be a little careful about is to ensure there isn't a kingside pawn left after the decoying process, but this isn't hard. Here's Kasparov showing that even having his king well placed on d5 doesn't help Black. Kasparov - Bazan Simultaneous, Germany 1992 Truly a horrible move. Black was worse of course but with the bishop he still had some chances to defend. I think probably 30...b4 was best: the bishop is in some danger of being trapped. 31.dxd3 Simplest although I think 31.cxd3 does win as well, and makes quite a nice finish: 31...a5 32.b3 ad 33.bxa4 bxad 34.23 c6 35.263 shd422 ‘The Berlin Wall 36.cRf2! c4 (36...h2xd3 37.f5! 04 38.hel c3 39.shd1) 37.dxc4 ted (37...tbxe4 38.25 thd5 39.5 doxe5 40.16 gxf6 41.gxhé is the point, a typical trick with this kingside structure.) 38.sbg3 c5 39.g5 hxg5 40.fxg5 texe5 41.cbg4, and White wins, eg. 41...Re6 42.h6 gxho 43.gxh6 bf6 44.h5 31...c4F 32.the3 06 33.c3 a5 34.03 ab 35.063 10 With this queenside structure Black can play 25 and ...b4 if he likes, but White just keeps tempoing his king and in the end Black has to lec it in to e4, and the rest is easy. “The importance of the e4-square in these pawn endings is great. IF White establishes his king there, even having dissolved his doubled pawn. may not be enough to save Black. Janev - Marcelin Bois Colombes 2003 1.248 ‘The game actually continued with 1...h5 2.the3 Hh6 3.65 He6 4.2c2 b5 5.thed hb6 6.2d5 Q6 7.Fig6 Exg6 8.262 Bob 9.Exf7 4 10.e6 c3 11.e7 Bc5t 12,4d6 Bc6t 13.hd7 c2 14.21 1-0. But couldn’ Black have just opposed rooks? The answer is no. White wins quite simply after: 2.Sixd8 shxd8 3.8e3 bd7 4.be4 heb 5.g4 bS 6.h4 a5 7.h5 For example: 7.196 8.46 04 Lerting the king to d5 is hopeless. 9.bxc4 bred 10.shd4 3 11.2xc3 shd5 12.chd3 g5 13.she3 a4 14.23 heb 15.04 gxf4 16.2ix64 £6 17.exf6 chxf6 18.951 Le6 19.chg4 Sf7 20.265 shg8 21.she6 ‘And so forth; the white king crosses at once to the queenside. Bur there is one important pawn ending which is a draw and which has turned up quite often in practice, Korneev - Fontaine Cap d'Agde 2002 gh ‘White has just exchanged Black's active rook on £3, but now he finds that he cannot win, abede fChapter 2 - Typical Berlin Endings 2B 1...g6t ‘The only move. 1...g5 loses after 2.fxg5 thg6 3.chf4 the7 4g3 shg6 5.tth4 c6 6.03 she7 7.26, and 1...s%e7 2.f5 the8 3.cbf4 she7 4.chg5 is also hopeless. But now White cannot break through. Korneev tries some aimless tacking but so long as Black sticks to king moves until £5 is played he cannot lose. 2.she4 he7 3.he3 hf7 4.263 hfs 5.sbg2 shgs 6.th3 bh7 7.2h4 ho Black has to be able to meet ®h4 with this move of course; the same position can occur in the same-coloured bishop ending after 2h4 Sixh4; shxhd, 8.2h3 Hh7 9.chg2 gs 10.bg1 bg7 11h, BAB 12.chel sh7 13.chf1 ohfB 14.chf2. of7 15.cf3 Sf 16.65 g5! This is the point: Black locks the position and the extra protected.passed pawn is not enough. 17.8e4 he7 18.f6F Mh 18.fxe6?? thxe6, of course would be terrible, actually losing for White. ‘This demonstrates another point about this ending, though: if White had not played b3 then Black would be lost. All White need do is play (using the move numbering above) 17.fxe6 the7 18.the3! dxe6 19.cbe4 c6 and now the triumphant 20.b3, and White wins. Luckily for Black though the pawn usually ison 63, since Black usually doesn’t play ...a5 unless there is a hook on b3, and if he does White wouldn't normally bother with a4. It follows from considering this ending that if the same position arises with pawns still on g3/h3 and Black with a pawn on h5, White should not gaily press ahead with g4 hxg4; hxg4??. Instead he should recapture with the king. Knight Endings A knight endgame is, as a great man once said, really a pawn endgame, and White’ chances are better in a pure knight ending than any other but the pawn ending. Knights are not much affected by blockades ~ perhaps this is why Black usually wants to trade White’ knights if he can. Here is one typical example. Carlsen - Kharlov European Championship, Warsaw 2005 oh gh Black is fairly active and would not be worse atall if White had his bishop remaining instead of the knight, but as it is Black is probably lost. L.wa5 2.c3 White must make a pawn move on the queenside to prevent the well-known ...a5-a4, .=-b3 axb3 ...c3 breakthrough being a factor later on, and as usual in the Ruy Exchange structures he wants the c- and a-pawns advanced rather than the b-pawn so that Black has no chance to trade his advanced c-pawn. On the other hand this introduces a complication later on... 2...b3 3.a3 Dg8 4.A\d1 Ae7 5.De3t hes24 ‘The Berlin Wall 6.204 a4 7.f1 White needs to drive back the black king before he goes for it. The immediate 7.f5 loses if White isn’t careful: 7...gxf5t 8.gxf5 exfSt 9.DxfS DdS 10.c6 Dxc3t 11.cheS (11.63 @dst+ 12.hd2 &c6) 11..Ad5 and now White has to find 12.3 (12.43? loses to 12...€3 13.Dxd5 c2 14.\F4 hed 15.Ae2 hd3) 7 bb5 8.Dd2 he5 Black cannot afford to play 8...c5: White would just bring the knight to 3. 9D e6 10,d44 kd7 11.De2 Le6 12.Ad4} &2d7 White has driven the black king back as far as he can, and now he plays his trump card. 13.65 exfSt 14.gxf5 c5? Tes hard to say whether this was a blunder or a desperate try, This being an official FIDE event the players were presumably suffering under one of FIDE’s ludicrous anti-endgame time controls, judging from the play. The logical continuation was 14...gxf5t 15.2xf5 Dg 16.23 ke6 17.Axc4, and gradually White wins. 15.e642? Now the position is a draw. 15.6 was immediately decisive since after 15...cxd4 16.7 White queens. 15.8 16.82 Oxf 17.04 De7 18.he5 g5 19.Dhb5 bd8 20.hd6 DBT 21.2e5 De7 22.hf6 Ac? Georgios Makropoulos must have been a proud man at this moment. 22...Ad5+ draws because Black keeps the g-pawn and can thwart White's triangulation attempts: 23.067 (23.coxg5? Dxc3! even loses) 23...g4 24.Dg3 De7 25.De4 Dd5 26.268 De7 27.2g3 Agot! 28.g7 De7 29.f7 Ac6! 30.06 Dest 31.8f6 Dc 32.€7t sed7 and so on. 23.cheugs the7 24.064 hd6 ‘The game now resumes its natural course: 24...2Da7 25.85 DbS 26.Ad5t dhd6 27.3 Be7 28.Axc4 Ac7 29.he5 Dxe6 30.Db6 was no better. 25.26 De7 26.f7 Ac6 27.266 De7 27...De5 28.Dd5 Dgdt 29.285 De5 30.c7 hd7 31.66 Dgst 32.07 Dest 33.cbg7! (cugzwang) 33...2\c6 34.92f8 (Mueller) is a typical technique in knight endings. 28.Dg2 Dest 29.2g7 De7 After 29...c8xe6 the knight arrives just in time: 30.¢2xg8 eS 31.2e3 hes 32.Dxc4 hd3 33.Db6 tec? 34.4)xa4 30.247 DG 31.066 Dds Black cannot hold the blockade: 31...De7 32.De3 DgSt 33.che7 De7 34.b7 DcB 35.8286 32.07 hd7 33.27 10 ‘Another example of the knight endgame is Game 1, Shirov ~ Almasi, on page 141. Knight vs. dark-squared Bishop ‘A pure ending of white knight against dark- squared bishop with the usual Berlin structure does not often arise, for the excellent reason that in order for it to happen Black must have permitted ewo trades, one of his light-squared bishop for a knight, the other of his knight for White's bishop, both of which he should in principle be avoiding. If we look at a paradigmChapter 2 - Typical Berlin Endings 25 position where an exchange has occurred on e6, we may see why Black should be avoiding such a position. Example position: KH Nw RUA» Ow abedefgh 1.c2£3 she8 2.94 hxgat 2...@£7 3.Dg3 leads to the same thing since after 3...h4 4.2e4 White proceeds with g5 and shgdxh4 and wins easily. 3.coxg4 £48 3... BET? 4.Dg5t 4.25 2e7 4... 2d7? 5.D7 Be7 6.Dh8 5.h4 06, Black has no other move. 6.h5 gxhSt 7.soxh5 27 8.0g6 2d8 9.57 2c7 10.e4 £d8 11.066 b5 11.,.Gc7 12.Dg8+ hes 13.chF6 ded7 14.67 and 6+ wins the e-pawn. 12.e4 bxc4 13.bxc4 2b6 14.296 227 15.Dd6 &b8 16.Dc8t A flashy finish: although White could also start winning pawns with 16.b7 the text is much quicker. 16...2d7 17.27 doxc8 18,2xe6 White queens a pawn, for example: 18...2d8 19.97 More likely perhaps is a scenario where Black has chosen to play ...2g4xf3, which could lead toan endgame like the next diagram. Normally perhaps Black would have kept his blockading knight on 5 and simply allowed ®xf5 gxf5 if necessary, but if he for some reason chooses to exchange the knight for White's bishop an ending like the next diagram might arise. This is equal, for example: Mi “a y, |b ee. 1.£4 g6 2.shg2 hd7 3.cbf3 che6 4.23 £5 5S.exf6 5.De2 Bc5 is no better, for example: 6.c4 bS 7.b3 Ba7 8.Dcl 2d4 9.Ad3 c5 10.24 06 11.cbg3 che7 12.h4 h5 13.8f3 cheb 5..chxf6 6.hed 5 7.6 0g] 8.h3 bs And since White cannot dissolve his doubled pawns with 9.65 in view of oan fBE2 he has no sensible way to try and make progress.26 ‘The Berlin Wall In this structure Black should certainly beware, however, of taking a bishop on ¢3 with his knight. Witness the following ending (which incidentally computers tend to underestimate gravely for White). HE NW RUA AO abcde f gh L.hf2 bd7 2.64 he6 3.2f3 b5 4.Dg3 g6 5.2e4 Bc A main point is that 5...f5t 6.exf6 2xf6 now, while still possible, obviously leaves Black much worse with e4-e5 and f5 on the cards. 6.De2 Le7 7.Dd4t hd7 8.65 gxfSt After 8...c5 9.fixg6 fixg6 10.e2 Black cannot prevent the king invading: 10...c4 (10...%%e6 LL.@64t hf7 12.sed5 and 10...06 11.f4 gS 12.g6 are no better) 11.A\F4 g5 12.De2 Bb4 13.c3 Be7 14.8f5 c5 15.Dg3 b4 16.c6¢ eB 17.%e5 and so on. 9.sbx65 5 10.03 2d8 11.b3 2e7 12.04 6 13.Del 2g5 14.Dg2 2d8 15.064 25 16.%2¢4 hd8 16.,.02c7? 17.€6 Sxf4 18.exf4 fre6 19.c2h5 wins. 17.2h5 shd7 18.h4 Se7 19.e6¢ fre6 20.Dg6 2f6 21.2xh6 we8 22.c4 wE7 23.e5 2d8 24.0h7 Lc7 25.h5 a5 26.04 bxa4 27.bxa4 2d8 28.h6 £c7 29.Db8t PB 30.c6g6 Sxe5 31.07 2c3 32.b7 White wins. This line has very litte value as analysis - I generated it by playing against Rybka’s top line after giving it abour ten seconds to think — but it does demonstrate that Black has real problems to solve. For practical purposes it isn’t worth delving too deeply inco this ending, as long as I’ve convinced you that Black shouldn't be playing ...2)xe3 in such positions. Knight vs. light-squared Bishop ‘An ending with knight against light-squared bishop and the structure unchanged also arises quite rarely. If the queenside is 24/b3/c4 against a5/b6/c7/c5 then Black will surely be able to hold; one of White's pieces will be tied forever to d2 or thereabouts. If the queenside is otherwise it may still be quite difficult for White to avoid counterplay arising from 3 and a3 being forced and the black king threatening to come to c4 and b3. All in all Black shouldn't have much to fear, but I want to show a more common ending, which arises after an exchange of bishop for knight on d4. Kruppa — Aleksandrov Minsk 1997 HN we UDAChapter 2 - Typical Berlin Endings 7 29.063? This is already perhaps a mistake. White is underestimating the danger of his position: something which is still common today for newcomers to the Berlin, and at that time was possible even for experienced grandmasters. It is actually not clear to me that White can hold at all, though. Another idea is to hasten to d2 with the king to defend c2, bur even then the difficulties persist. For example: 29.cbF1 Add 30.h4 2c6 31.hel Bed 32.ced2 she6 33.44 hd7 Now let us suppose thar White resists the temptation to eliminate the knight and sits tight with: 34.2al 34.8xd4 axd4 35.2c4 he6 36.3 cbd5 37.Qa3 dxc3+ 38.cbxc3 Bf5 39.0b5 shed 40.Dxc7 BAB 41, Dd5 thugs 42.Dxb6 dxh4 loses very much as in the game, the bishop is about to demonstrate in very clear fashion its traditional strengths when play is on both wings. 34...He6 35.203 Black’s best way forward now is to take play into the pure opposite-coloured bishop ending. 35...D65! 36.Dxf5 Soxf5 ‘Without attempting an exhaustive analysis, one can say that White has a difficult task to hold this. A sample line: 37.01 Sop4 38.21 Sef3 39.cbd2 BFS 40.chd1 Be6 41 thd? Bp4 ‘With this neat manoeuvre Black ensures he can get in on the kingside, and now White has little choice but to bust out with: 42.cbd1 she3t 43.cl He2 44.6 Oxe6 45.803 BE 46.265 6 47.2c7 bS 48.2xa5 sbxg3 49.8d8 But it looks to me as though Black should win: 49...bxa4 50.bxa4 soxf4 51.2¢7 cbf 52.a5 Bcd 53.Bxc5 £6 54.chd2 g5 And so on. There are other similar lines, but I don’t see salvation for White. 29...0d4t 30.fxd4 After 30.the4 Qxc? 31.Axc2 BFSt 32.cbd5 Gxc2 33.chc4 Bd1 34.8d2 Bert 35.chc3 cA! (a typical method) 36.bxc4 &dI Black gradually wins. 30...cxd4 31.Dc4 he6 32.the4 the5 33.2d3 bd5 34.c3 White is lost in any case: 34.4 &f5+ 35.cbd2 the4 36.the? 2g4t 34.065 35.02 duc3t 36.bxc3 Bet And the black king simply sails in and takes White's kingside; the co-ordination of black pawns and light-squared bishop in restricting the knight is striking. 37.f4 BAB 38.%2d4 Se6 39.0d2t cg? 40.4 bxh2 o1 Iesallitde surprising how effortlessly Black won here, and ithas a lot to do with the white knight being poorly placed at the beginning of the ending ~ if it could have got to g5 the outcome might have been different. But this ending does quite often arise and in general Black will be the one with the winning chances. Bishop vs. Knight The ending with White's bishop against Black’s knight is one of the most favourable for Black. Especially if the queenside pawns have become fixed on a4/b3/c4, so that Black is always threatening ...Ad4, it is difficult for White to make any progress. The sort of breakthrough we saw in Carlsen - Kharlov simply won't work, since Black can just trade pawns down28 ‘The Berlin Wall and arrange to blockade the e-pawn with «eG after f5 and exchanges on £5. ab cde if gh Indeed this position was agreed drawn without further ado in Ponomariov — Kramnik, USSR-ROW 2002, while in J. Polgar — Milov, FIDE KO 2005, White did no more than play h3 and p4 before giving up. Black cannot play completely passively: he mustn't allow the king into g5/h6 and the bishop co the h4-d8 diagonal, but unless his pieces are very badly placed at the outset of the ending it is easy enough to stop this by keeping the knight on 6 within striking range of d4. However, the position isn't necessarily a draw if White hasn't gone £4, The next position is reached by Khalifman in his analysis, and as the ex-FIDE champion says, White has much better chances in this version. Rw RUA Ao He has basically three ideas: getting the king to g5 (he would have to be willing to lose b3 to do this, but that may not matter; unlike the similar situation in the opposite-bishop ending where losing b3 generally is the end of the world); penetrating with the bishop to 8 (a pipe dream just now but may become important); and, most significantly, making a passed pawn on the kingside (one significance of not having played £4 is that White can play g4 and present Black with a dilemmas cither he takes and allows a potential passed h-pawn, or he doesn’t and ensures that there will be a pawn left on the kingside after gxh5 ...gxh5 and a later f4-f5, which greatly increases White's winning chances). Indeed I suspect White may be winning here. I won't attempt a detailed analysis, but here is a sample continuation which shows Black's difficulties. 29...B€7 30.he4 White should go straight for it; if he dallies with 30.2c3 then Black can put great difficulties in his way with 30...c8 31.be4 D6 32.chf4 Dd4 33.chg5 AxBt 34.coxh5 @xh2 35.%g6 chd7, for example 36.hxg7 she7 37.chg6 c6 38.0h5 chf7 39.24 chg7 40.g5 DEB 41.g6 hfB 42.chg4 Dd4 43.82 Axb3 44.2h6+ hes 45.chg5 Dd4 46.16 bS 47.cxb5 cxbS 48.axb5 Axb5 49.hoxc6 a4 50.thd7 Add 51.c6 DFS 52.c7 Dxe7 53.doxe7 a3 30.006 31.bf4 Dd4 32.cbg5 Dxb3 33.dxh5 White doesn’t need to worry about losing his c4-pawn; although it seems that 33.2c3 2d4 34.f4 wins also. 33...0d2 34.bg6 Oxc4 35.2c3 c6 35...c0c6 36.toxg7 De 37.4 LST 38.c2f6 ded5 39.h5 and 35...2a3 36.f4! (and not 36.2%xg7 b5 37.axb5 a4 when BlackChapter 2 - Typical Berlin Endings 29 is suddenly better) 36...b5 37.axb5 &xb5 (37...a4 38.f5 exf5 39.6 Dxb5 40.2e5 queens) 38.8xa5 are no better. 36.S2xg7 b5 37.h4 b4 38.2a1 b3 39.hS b2 40.xb2 Dxb2 41.h6 c4 42.h7 c3 43.h8Y4 2 44.h6 Dd3 45.86 cl 46.8h7t Hb6 47.Bixd3 I suppose that White should win this queen ‘endgame. It does have to be said though that Black's difficulties stem to a large extent from. the fact that his king is in the next parish ac the start of the endgame, and Khalifman has Black playing some pretty bad moves shortly before the diagram to ensure that should be so. Let's see a more cheerful example for Black. Berescu - Mastrovasilis Kavala 2005 e abcd cn Here we see the same situation without the queenside pawns frozen, which you would think ought to favour the bishop. On the other hand Black has already made good progress with his king, and if anything it is White who already needs to be careful. In the game i turns out that the decisive factor is the dominating black king on d5, which is possible precisely because White doesn't have his pawn on c4. 25...2e6 26.2263?! White doesn't appreciate the danger. He ought to have taken control of d5 with 26.c4, and if Black challenges that control with 26...b5 then he should stick to his guns with 27.b3. If Black could immediately attack the c4-pawn this would be a disaster, but White is in time to meet 27...bxc4 28.bxe4 e7 with 29.825. 26...ed5, 26..Ad4¢ 27.84 Dxc2 28.654 hd? 29.s2d3. was impossible, of course, but now White needs to guard against this threat. 27.03 ‘An unpleasant move to have to make, but since the pawn ending after 27.2¢3 c5 28.g3 Dd4¢ 29.8xd4 exd4 is lost White has to play it sooner or later. 27... Dh4t 28.293 AFSt 29.063 5 30.2e1 4 31.32! A slightly curious move, Black could hardly have hoped to win after 31.h3 h4: of course 32.chp4 hed 33.xh4 Oxh4 34.chxh4 exts is bad, but White could just play 32.2/2-e1 instead and Black could never move his knight and thus couldn't make any progress, So Black shouldn't play 31...h4, bur in that case there was no need to lose a tempo with 31.g3. 31..b5 32.h3 a5 33.g4 hxgét 34.hxg4 De7 35.2h4 Deb 36.032! White can’t go on with his kingside play and is therefore already worse, but there wasn't any need to make this concession. I may be wrong, but T don’t see how Black could have won if White had waited passively. He can go ...b4- b3, but that takes away the b3-square. The significance of that is apparent in the note to White's 38ch. 36...a4 37.2e3 Db8 38.24830 ‘The Berlin Wall When you see the game you wonder whether White could have held by keeping this bishop inside his own lines, but he does need to prevent, .Da5, for example 38.81 c5 39.2d2 Dc6 40.23 a5 41.cl Db3 42.3 (White has to keep on this diagonal to prevent ...!cl-d3) 42...96 43.g5 b4 and White is in zugewang and has to let the black king in to e4 or the knight to cl and d3. 38...c5 39.27 Ac6 40.0013 b4 41.65 b3! 41.,.Dxe54? 42.8xe5 dxe5 43.c%e3 is only a draw, but Black has a better idea. 42.02f4 Dbél 43.bg5 Dd3 44.06 fxe6 45.96 gxf6t 46.coxf6 DAZ 47.95 Dest 48.296 Dxgs 49.coxg5 hed A final insult to Whites bishop, which is comprehensively outgunned by Black’s king. 1 R. Perez - Narciso Dublan Capablanca Memorial, Havana 1999 ‘The next example comes from 2 game quoted by Khalifman as an example of a model exploitation of White’ possibilities in the Berlin endgame, so it’s a pity that he fails to mention what an utter hash Black makes of holding what should have been a fairly simply drawn ending, Vw AUDA» oo 29.04 Te’s not totally clear this was necessary, but in practical terms ics 2 mean little teaser. 29...De7 Obviously 29.,.2d4? would be terrible: the pawn ending after 30.8xd4 cxd4 is simply lost (White also wins, perhaps slightly deceptively at first sight, with the pawn back on a7). But the question is what is happening in the ‘other’ pawn ending after: 29...Db4 30.8xb4 cxb4 Always the right recapture, and also correct and drawing if White's pawns are on c4 and b3 (in Shaw — Neubauer, European Team ‘Champs 2005, Black played ...axb4? in such a position and should have lost). This ending is tricky and all three results would be well possible in practice, but I think it’s a draw. 31.hed White's problem is that following ...cxb4 Black’s pawns can now make a passed pawn on their own, and so he has either to make £5 work somehow without neutralising the pawns first (which it turns out he can’t do) or else organise c3 bxc3; bxc3 and then c4, which will win ifhe can organise it peacefully, but it turns out he can't do that either. 31.c3? at the moment loses to 31...bxc3 32.bxc3 b5! so White has to get nearer. 6 Black has to stop the threat of 3 immediately. 31...S8%e7? loses to 32.c3! only. Others don't work: a) 32.1522 actually loses to 32...ex 33.gxf5 gxfSt 34.xf5 b5 35.b3 bxa4 36.bxa4 c5 37.she4 4 38.thd4 c3. b) 32.s8d4 6 33.cbe4 (33.b32! g5 introduces a new unpleasantness for White, although he can still draw after 34.fxg5 sbf7 35.23 dog 36.0f4 b5 37.c2e3 SoxgS 38.cbd4 brad 39.bxa4 Pxgd 40.c8c5 BFS 41.hb6! hxes 42.sbxa5 5 43.%b5 dé and so on)Chapter 2 - Typical Berlin Endings 31 33...b5 34.b3 bxa4 35,bxa4.c5 36.9243 ded7 37.t2c4 the6 holds easily. ©) 32.c4 doesnt quite do the trick: 32...06 33.f5 (33.thd4 is subtler and forces Black to play 33...c2d8! [33...che8 loses to the exquisite line 34.c5 b5 35.axb5 cxb5 36.c6 a4 37.65 gxfS 38.gxf5 a3 39.fre6! shd8 40.bxa3 bxa3 41.c2c3; 33...c2d7 34.c5 bS 35.axb5 exbS 36.f5 exf5 37.ced5 is also winning,] 34.b3 [34.c5 b5 is now dangerous only for White] 34...%2d7 35.c5 b5 36.axb5 exbS 37.f5 gxf5 is now fine for Black since al is check after b3 has been played.) 33...exf5t 34.gxf5 extt 35.tbxf5 b3! draws (iE White had played b3 earlier to avoid this then in this position ...b5! would draw; Black queens on b1 with checl ) So the winning lines go: 32.c3! bxc3 33.bxe3 6 34.04 "bdy 35.5 gS 36,he3 (only not 36.c5?? exfSt!=) 36...ed8. (or 36...c5 37 fre6t the7 38.3 doxe6 39.che4) 37.5; this line illustrates White's basic idea. 32.b3 32.c3 now is no good because of 32...bxc3 33.bxc3 b5 again, although at least White can still draw with 34.5! 32.s2d42! too comes close to losing although White can also still draw here with 32...g5 33.fxg5 tef7 34.c4 bxc3 35.bxc3 b5 36.axb5 cxb5 37.<8c5 a4 38.shb4 chg6 39.c4 bxc4 40.chxad thes 41. bb4 xg 4 42.coxc4. 32...the7 33.65 ‘The only try: we already saw 33.c0d4 g5, although Black can also wait. 33...exf5t! 34.gx65 5! This is the whole point; the natural 34...gxf5+ 35.cbxf5 27 (or 35...b5 36.64) loses after 36.c6¢ she7 37.c8e5 bS 38.cbd4 shd6 39.07 cSt 40.cbd3! (not 40.she4 shxe7 41.axb5 [41.cbd52? c4—+] 41...bd6=) 40...Bxe7 41.axb5, White is now close to losing, but ifhe collects himself sufficiently he can still draw with 35.2£3 b5 36.c2g4 bxa4 37.bxad c5 38.f6+ SofB 39.c2xg5 4 40.€6 b3 41.cxb3 cxb3 42.chg6 b2 43.e7¢ the8 44.chg7 bIW 45.67t thxe7 46.£8Wt So all in all my conclusion is that Black could have held with 29...2b4, although by no means trivially. Probably he would have had to have seen this position-type before to be confident over the board. 30.the4 2£7 ‘An interesting point. In the similar position with Black having a bishop, the king needs to be on d7, but here it needs to be on £7. The reason is this. Black has two problems. He has to prevent £d8, and he also has to prevent White's king penetrating on the kingside. With the bishop, Black needs to escape with the bishop on to the far side of the pawn chain once White takes the h4-d8 diagonal, and thereafter the bishop can hold the king back from entering via g5, as we shall see in some of the next few examples, but it can't defend the queenside pawns, so the king has to guard 8. The knight by contrast can protect the queenside pawns, but it cant stop the king entering once it is dominated by the bishop on. hd, so the king needs to be on £7. 31.2e1 De6 32.2432 ‘The Berlin Wall 32...0d42 Ies hard to say whether Black missed White's next move completely or thought he was losing if he held still and just panicked. In fact Black has a fortress and could have held on by doing nothing. 32...8e8 33.c3 BE7 34.65 pxfSt 35.gxf5 DB As usual in these positions Black cannot let the king to £5. Instead he has to allow f6 if White wants it, but after that White can’t do anything atall; the king defends the kingside and the knight the queenside by ...4 and White can’t make any progress; Black’s next move is to kick the bishop out with ...%e8, and White doesn’t have time to get in on the queenside after: 37.2h4 DbB 38.66 Deb 39.293 c4. Black can also play 32...c4 at once, followed by tempoing with his king unless pressed. I don't see how White can try and break this fortress either, for example: a) 33.f5? gxft 34.exf5 exfSt 35.chxf5 dat b) 33.3 Db4 34.82d4 (34.c3 Ad3; 34.chd2 @d5) 34...Axc2t 35.%oxe4 the8 c) 33.c3 eB 34.65 pxft 35.exf5 we 36.225 b8, and the pawn being on c4 makes no difference; White can’t win c4 without playing fxe6, and then he loses ¢5. 33.2d8 b5 34.axb5 Dxb5 35.c4 1-0 Bishops of the same Colour If White is left with the dark-squared bishop the best balance of forces for him is for Black also to have a dark-squared bishop, and in that event, other things being equal, White will have atleast good winning chances. The generic ending normally arises after an exchange of knight for bishop on e6 at some point, as in the following position Deep Fritz - Kramnik Bahrain (1) 2002 HEN whUDA OE abede fgh This endgame is very close and, as we shall see in a moment, strong grandmasters have lost it as Black and others have published analysis which (in my opinion, naturally) is mistaken, so it’s worth knowing something about it. As a general observation, Black’s chances to defend are based on blockade, and White's main weapon is zugzwang, It is possible to see quite easily how the play will go; at some moment White will play g4 and £5, with his king on e4. At this point Black cannor gaily trade all the pawns and leave himself facing a lone pawn on €5 with the king on 5; this position is lost. Instead he has to wait and hope that neither 6 nor fxe6 wins. He also has to bear in mind that he can rarely permit White to seize the h4-d8 diagonal, since the obligation ro prevent 248 as well as hold the kingside will usually be too much. Black’s best approach in this endgame is the following.Chapter 2 - Typical Berlin Endings 33 1. The best place for the king is d7. From there it defends both d8 and e6. Black has to manage on the kingside without it. The king should come towards f7 only in case of need: either to meet 85 by ...Q2£7, or to meet Sxh4 ...fixh4 sbxh4 with ...2h6. 2. The move ...c6 should not be played. This has several disadvantages: first, as we saw in Korneev — Fontaine it means that after g4 wohxg4, hxg4 and later 2h4 ...Bxh4, bxh4 .-€2h6 the pawn endgame is lost; secondly it means that fre6t ...2xe6 Bd8 cannot usefully be met by ...22d7, and thirdly it weakens d6. 3. £5 should be met by ...gxf5t, gxf ...d8h4 or -2g5. The h-pawns can either be exchanged or not, Generally it will be simpler to do so, but especially if White’s b-pawn is still on b2, then Black has to be careful about playing «-shxg4 before £5 is played. More about this in the next example, Howell — Parker. If this programme is followed, Black should not lose. 24.94 White might «ry 24.g3 with the same idea as Parker's suggested 28.h4 in the next game, but Black can deal with it easily with 24.,.%e8 25.h4 shf7, and if 26.8g5 &xg5 27.hxg5 bg6 28.f4 FS 29.bF2 cog4 30.hg2 h4 31.gxh4 x4 32.2h3 g6 White even loses. 24.96 This move caused much speculation chat Kramnik had analysed this endgame in his preparation, and perhaps he had. Another try to draw, which is here both unnecessary and unsuccessful, but is important in other positions, is 24...hxg4 25.hxg4 g5 26.%g2 eB 27.4 exf4 28.8xf4 f7, but White can prevail here with 29.@h3 (there is no time to dally; if Black had another tempo, for example after 29..d2? hg6, then White can do nothing at all since 30.%h3, with the idea of &h4, can be met with 30...28 31.261 27) sig6 30.83 Sg5 (30...8f8 31.8h4 Bg7 32.8f6 2xf6 33.exf6 Sxf6 is a winning pawn ending) 31.8h4 Bf4 32.816 shf7 33.2g2 hee 34.263 2d2 35.248 c6 36.2xb6 2b4; at first sight this looks like a draw but White can break Black’s fortress with 37.8d8 &f7 38.2h4 g6 39.22 chgS 40.Lc3+ dg6 (40...2h4 41.2 f4 decides at once) 41.2! (zugewang; the black king has co retreat) &g7 42.8d2 (this is possible now that the black king is two moves from g5), and White wins. 25.h4?! Not the most compelling ty. 25.gxh5 exh5 would be hopeless, of course, allowing the black king to come to g6, but Stohl gave analysis to prove that Black holds even after the natural: 25.64 the8 26.2 hxg4 This move is not obligatory: Black also draws after 26...2d7 27.23 2d8 28.04 Be7 29.gxh5 gxh5 30.65 h4 31.2h6 2g3 32.2g5 Bel when White does not have any way to make progress: this is important if the pawn is on b2, as we shall see in the next game, since this is then Black’s only efficient way of making a draw. 27 bxg4 f7 28.213 28.2h3 the7 29.83 bh7 30.8f2 &e7 31.8h4 Sxh4 32.drxh4 h6 is drawing, as we saw on p. 22-23, 28...2d8 It is also possible to reach the correct set- up with 28...8h4? 29.%e4 Bel 30.65 gxfSt 3l.gxfS be7 32.he5+ hd7 33.hf4 Bd2+ 34.4g4 Bel 35.8f4 che7 36.cbg5 £7, when again White cannot go any further. 29.he434 ‘The Berlin Wall HR wRUaAre 29...g5? A desperate attempt, but I don't believe Black needs to panic like this. Better was: 29.867 30.65 Sh4 (30.848 31.£h6 followed by sbf4 and 2g5 wins: Black has to try and prevent this) 31.2h6 (31.cbf4 pxfS 32.gxf BF2 33.c2g4 Bd4 also keeps White out) 31...gxf5t (31...g5 32.3 wins since Black has to let the bishop out of the cage) 32.gxf5 the7! (always to d7: 32...8g3? 33.25, but even 32...c6 might draw since after 33.04 Bel 34.85 Bd2t 35.che4 2c3 36.216 24 37.04 Black has 37...b5, and I don’t see how White can win since if 38.axb5? cxbS 39.cxb5 c4 he loses.) 33.4 Bel! (33...2d7? 34.027 he7 35.sbg4 Bel 36.866+ shd7 37.cb¢5 and Black has to give way) 34.09p4 (34.cg5 wf7 forces retreat, and certainly not 34...2d2+?? 35.96 &xh6 36.6t) 34...2d7 35.264 che7! (else &g5) 36.8g5+ hed7 37.26 8d2 HE Reka! White can't progress. This timely control of the cl-h6 diagonal is Black’s fundamental defensive idea. 29...s8e8 is che right idea but is badly timed since Black cantt fight against the f4-p4-g5 plan. 30.65 gxfSt 31.gxf5 shd7 32.hf4 Sh4 33.00g4 S£2 34.chg5 and wins. 30.65 c6 31.263 bg7 ee eh Stohl though Black holds here but, as Jacob Aagaard pointed out to me, White wins with: 32.fre6 Bp6 33.2f2 Se7 34.893 dB 35.¢7 Bxe7 36.06 66 Or 36...8d8 37.2d6 f6 38.7. 37.Lc7 hoxeb 38.2xb6 25..hxgd Kramnik could even have fallen into the computer's trap ~ after 25...8xh4 26.g5 White can win the black bishop, but he can’t win the game — Black's position is unbreachable, as the reader can check for himself. 26.285 Sixgs 27.hxg5 the8 28.c2g2 ‘The pawn endgame is an easy draw. The variation with 29...Qe7 is very important, since it proves that so long as his pawn is on c7 Black can even hold by keeping the king on £7 and playing ...hxg4. ory fy oto ogChapter 2 - Typical Berlin Endings 35 In the next diagram Black is very significantly worse placed than in Frice - Kramnik because White's pawn is still on b2. Howell - Parker HN weEUDAA e& 7 cd soa gon This means that che pawn endgame is lost after £4, g4 ...hxg4, hxg4, and consequently that Black has to be very careful about ...hxg4, hxgé since if this is played White can take his king to h3 and go 2h4, winning the crucial diagonal. This is only immediately facal if Black allows his bishop co be trapped in his ‘own camp; otherwise | believe he can survive even this, bur White can be cunning and cake his bishop to e] before breaking with e4. 27.04 ‘White needs to play this at some point, since otherwise Black does have the possibility of counterplay after ...c4. 27...e8 Wuugd tries to avoid White's h4 idea, but loses in the same way as shown in the note to Black's 24th move in Fritz — Kramnik above. 28.f4? White can win at this point with a different idea aleogether, which has never been shown as faras | know in any of the various commentaries on this type of ending: 28.h4! (this move was pointed out by Jonathan Parker) 28...2e7 29.82h3 chf7 30.p4 Se8 (30...hxg4t 31 .coxgd cb£7 32.2895 followed by h5 wins comfortably) 31.gxh5 (not 31.225? Sxg5 32 hxgs hxg4t drawing as in Fria — Kramnik) 31...gxh5 32.825 Of8 33.64 Sg7 34.cg3 hd7 (34...c6 35.063 thd7 36.the4 shc8 37.£5 ted7 38.b3! Again! 38.28 39.fxe6t Bxe6 40.2d8 Bho 41.8xb6 &cl 42.8xc5, and White wins wich the wo extra pawns) 35.02f3 Bf8 36.he4 297 37.5 (ugewang; I think White could always force this position with Black to move by tempoing shf3-e3 if needs be, but even if it were somehow White's move he would have b3!) 37...818 38.2.6 &h6 39.fxe6t @xe6 40.2d8 ded7 41.25! (the point) and White wins. This idea works because of some quite specific circumstances; White hasnt yet played £4 and Black has played ...g6, so that his king cannot get to that square. 28...2£7 29.283 62 Tr seems to me that this move ought to have been the fatal error. Black needed to hold his position as it was, avoid ...hxg4, and draw in the way hidden in the comments on Black's 26th (26...42d7) in the hypothetical line given in the note to White's 25th in the Deep Fritz — Kramnik example, starting with 25.f4 (page 33). 30.g42! In my opinion this is premature, although as it happens I think White is still winning. Before playing g4 it would be good to take control of h4 in order to prevent the bishop escaping from its own camp and gaining the c1-h6 diagonal. To achieve this no good is 30.2e1 g5 gaining the g6-square for Black’s king and drawing easily, bur White can win with first 30.4 2e7 31.Bel Qd8 32.24.36 ‘The Berlin Wali This places Black in an unpleasant dilemma; either he keeps the h-pawns and loses as in the game, or he exchanges, but then White can win without f5, thus: 32...hxg4 (32...2e7 33.0g3 2d8 34.gxh5 gxh5 35.65 2p5 36.264 Bh4 37.26 is like the game) 33.hxg4 2e7 34,003 Qd8 35.c%g3 PB 36.2h3, and now Black either goes in for 36...g5 37.65, when he loses as shown in the notes to White's 25th in the last game, or he allows &h4 with his bishop still trapped in his own camp, which is fatal. 30...0h4! ‘Ihe only move, to prevent the lines given in the last note: 30...hxg4¥? 31.doxg4 27 32.He1 eS 33.8h4 B68 34.895 thd7 35.h4 and h5 wins easily. abede White isn't yer ready for 31.gxh5 in view of 31...gxh5 32.che4 dogo. 31..the7 Ie. big decision whether or not to exchange h-pawns, but as it happens Black is lost either way, albeit only by some clever tempo play after: 31..-hxgd 32.hxg4 eB 33.15 33.83 hd7 34.263 Sel 35.82 Bd2 36.8h4 Sc} 37.b3 2d? 38.de4 2cl 39.26 2d2 is less effective; by his messing about White has given Black the chance to force b3, and now White cannot do any better than 40.f5 gxf5t 41.gxf5 &cl (41...2c3 also draws) 42.fre6+ txe6 43.248 Bd? 44.2xb6 b4, and Black holds. 33.n.gxb5t 34.exf5 Bd7 35.26 the7 Necessary to prevent 2f8-d6, for example 35...g3 36,218 the8 37.2d6 ced7 38.(6 eB 39.2c7. 36.24 Bel Not 36....Rf2? 37.2g5t Sd7 38.2f6 and ches. 37.Bg5t 37.025 ©f7 38.82¢4 tke7 doesn't improve White's position. 37...8d7 38.fxe6t 38.26 is now met by 38...2d2t. 38... Hxe6 39.8d8 2d2¢ 40.che4 Bel 40...2e1 41.2xb6 Bb4 42.b3 41.Bxb6 Bxb2 42.dxa5 Bxe5 43.8b6 Bd4 ‘Any bishop move leads to the same finale. 44.a5 d7 45.2f5 BB 46.06 Bg7 47.26 218 48.2f6! Bd6 49.267 Bp3 By his last manoeuvre White has driven the bishop to this diagonal, and now he wins d6 for his king and with it the game. 50.dhe6 84 51.8xc5 Bp3 51...08¢7 52.2d6+! is the point. 52.2d4 Bf4 53.05 Bd2 54.2d6 And White picks up 06 and wins. 32.gxh5 gxh5 33.65 £7 34.3?! It was possible to play 34.2h6 e7 35.4 d7 36.8f8 at once, but Howell prefers to mark time a little until the time control. Something like 34.0e3 was a much better way of doing this, though. b2-b3 is a terrible move for at least two reasons: it cakes away a reserve tempo which is useful in many lines, and it gives Black a square on c3 which White didn't have to give him.Chapter 2 - Typical Berlin Endings 37 34.tbe7 35.2h6 £7 36.2F4 Bel 37.4 Bh4 38.23 Le7 39.291? “This really is taking the policy of reaching the time control too far. White could still have won cleanly with: 39.86 27 39..e1 40.2g5+ d7 41.fxe6t sexe6 42.8d8 2c3 43.2xb6 2b4 is the same as 41...Be1. 40.84 Bel 41,hg5 Lg3 42.fre6t xed 43.Sxh5 BxeS 43...Lixe5 44,2296 44.cag4 Sel 45.2F4+ ted4 46.263 2d2 47.h4 Bho 47...8c3 48.2f4 Bel 49.h5 Sexb3 50.h6 2c3 51.0265 Cxa4 52.h7 Hb3 53.2.5 48.24 2p7 49.5865 39. nhhg3?! Black could have tried 39....2e1, and now I don’t see that White can do better than: 40.23 (40.84 Bd2+ 41.83 Bcd [41...Bxe34?? 42.Sxe3 xf 43.0f4 eG 44.h4 wins] 42.95 Bxc5 43.06 Bxe6 44.cxh5 BFS draws easily; Black wins the queenside pawns after something like 45 Bg5 Wed.) 40...2c3 4).2g5t Bd7 42.fxe6t Lxe6 43.8d8 Sxe5 44.8xb6 2c3 45.2xc5 “This endgame looks as though White ought to be able to win somehow, but I’m far from sure how he can do it. He is handicapped by the fact that often an exchange of bishops and then ...c5 is a draw. 40.2¢3 Gh4 41.81 &f7 41,01 42.fg5t d7 43.fxe6t kxe6 44.88 definitely wins because Black doesn't have time to pick up e5 in return for c5 without also allowing b4, as can be seen here: 44,..d2¢3 45.2xb6 (or 45.c7) 45...2b4 46.227 B23 (46...8c3 47.2xc5 Bxe5 48.b4) 47.2b8 2b2 48. £d6 &b4 49.cbF4 41...ed7 42.8h6 the7 43.2F4 is the game. 42,2h6 he7 43.sbF4 om 8 oO oO abedefenh Now White has what he wants. Either king move allows a fatal bishop entry, and if the bishop moves then a fatal king entry at g5 follows. 43...8d7 43..001 44.eg5 (44.095¢ Shd7 45.fxe6+ Lxe6 46.248 £c3 47.27 Rd2t 48.he4 Bc3 49.2xb6 also wins) 44...2267 45.fxe6t Lxe6 46.7 h4 47.4 Bg3 48.216, and &xh4 leaves White with a winning ending. 44268 And the added threat of £2d6 decides: if ..c6 had not been played ...;8e8 would now defend comfortably. 44...$€1 45.295 £293 46.2266 exf5 47.2x5 h4 48.2f6 eB 49.2d6 hd7 50.0f7 Ld8 51.2b8 28 52.06 dexb8 53.7 e7 54.088 10 ‘This game demonstrates that Black’s problems are greater with the pawn on b2. His difficulty is that if White is careful and takes the king to e4 and the bishop to el before going g4 then ...hxg4, hxg4 simply loses because of the38 ‘The Berlin Wall possibility of coming round to h3 with the king and going £h4. So Black has to permit gxh5 ...gxh5. He can then still draw as in the line with 26...22d7 in the notes to White's 25ch in Deep Fritz - Kramnik, but only so. Ie is obvious that with any worse set-up Black is probably going to be lost. Another intriguing example is the following. At first sight the pawn structure looks worse for Black, and on the kingside there is no doubt this is true, but pethaps the possibility of ...b5, leaving only 5 as a weakness on the queenside and more importandly still allowing the bishop an easy route to the critical areas via a5, will assist Blacle Greet - Howell Gausdal 2005 HN we DAA OW abede aoe ee 31BA2 c6 32.Hh4 2c7 33.04 You can see why Greet wanted to fix the queenside, but this move isn’t immediately necessary and in fact in the game continuation it may be that it was harmful and threw away the win, See the note to Black’s 48th move. 33.688 34.63, Here this move does no harm because Black is anyway going to be forced to play ..b5. 34...°0£7 35.0263 the8 36.295 whf7 37.chg3 she8 38.shg4 thf7 39.h4 eB 40.h5 White commits himself. The alternative plan of g4 doesn't work so well when Black's bishop is still free, for example 40.chf3 chf7 41.94 a6 42.thed b5 43.65 gxf5 44.gxf5 bxad 45.bxad a5, and the active bishop can thwart any efforts to bring the king to g5. 40...gxb5t ‘This might look a bit on the compliant side, but Black cannot avoid it forever, for example 40.,.2E7 41.2h4 a6 42.225 b5 43.2h4 Bas 44.202 Bb4 (44...2b6 45.2c3 bxa4 46.bxa4 £27 47.25 is worse) 45.2g5 gxhS 46.cbxh5 23 47.24 and now 47...g6¢ is forced, with similar play to the game, since if 47...2b4 48.5 White is winning: 48....2c3 49.2g3 &d4 50.fxe6+ Sexe6 51.22g6 bxc4 52.bxc4 Bxe5 53.dixeS SexeS 54.cixg7 hf 55.c2f6 coxe4 56.he6 hfs 57.2d6 thed 58.thxc5 edd 59.25 2c3 60.2b6! 41.coxh5 s6f7 42.24 a6 Black has no choice but to go in for this, whether it is weakening or not, since 42...2b8 43.2d8 the8 44.tog6 thxd8 45.c2xg7 wins easily for White. 43.g4 b5 44.042 g6t 45.coh42 Now this one I don't understand: 45,c2h6 is like the game a couple of moves later. 45..bxe4 46.bxc4 2b6?! Black could have driven the king back and in that case I don’t see how White could have made even the small progress represented by winning the front c-pawn, for example 46...Ld8t 47.253 Bb6 48.0263 the7 49.che4 ET 50.24 (50.F5 gxf5t 51.gxf5 he7 52.64 WET 53.03 he7 54.chg5 ET is the same) 50...8a5 51.65 gxfSt 52.gxf5 &c3 53.262 2b4 54.F4 Bd2t 55.hog4 &b4 56.2c3 £3, and so on,Chapter 2 - Typical Berlin Endings 39 47 Shg5 2d82? Its inexplicable that Black should miss the chance for 47.27. 48.2h6 Le7 ‘Wherever the bishop goes Black falls into zugewang and has to let the c-pawn go, for example 48,..2b6 49.%c3 £27 50.a5. What I don't understand however is why Black didn’t insist in return for his pawn on driving the white king back, thus: 48...2a5 49.8xc5 If White goes for a breakthrough like the game but without winning the front c-pawn then Black draws handily, for example 49.$03 &c3 50.g5 Bd4 51.8d2 2b2 52.65 cxf 53.c6¢ (53.8f4 Bd4 54.c6¢ thxe6 55.exg6 &e5 56.23 is more prudent and at least draws for White!) 53...2xe6 54.2xp6 eS 55.8c3 f4 and already White has to think about saving himself. 49...0d2 50.82g5 &cl 51.862 Bd2 52.803 2c3 53.cbh4 2d2 54.c¢h3 White can insinuate his bishop into position on the kingside, but not his king: 54.@h2 Be3 55.hog3 Bd2 56.013 Bcl 57.223 Ld2 58.ih4 Sct 59.hg5 2d2 60.s%g3 Le3 61.S2h4 2£2+ (or even worse the amusing 61.2h6 te8 62.coh4 chf7 63.8g597 Bf, and Black wins). 54,..e3 55.c2g2 Bd4 56.c0f3 2b6 57.22 B05 58.c8e4 8d2 59.65 gxfSt 60.gxP5 2c3 I dont see how White can make any progress; against proper defence he can’t get his king in through either the d4-c5 gap or via b4. However, it is perhaps at this poine that the mischief of White's 33rd move can be seen. If the pawn were still back on a2, then White could win this position; he plays {6 and takes the king over to a4, and, while he has to give up his ¢5- and f6-pawns in the process, he can then force ta5xa6, and as far as I can see if he does it right (he should ensure that the tempi are such that the black king is on £7 and not 8, for example), he can then prevail in the race between a- and c-pawns on the one hand and the e-pawn on the other. Of course a lot of water has flowed under the bridge since 33.24, and one would hesitate to say that it was a decisive error, but I think it might have been. 49.2h7 2d8 49...2£8 50.2h4 is worse; Black loses in a few moves. 50.0xc5 &h4 51.23 293 52.0h6 2h4 53.2d2 £248 54.21 2b6 55.2g3 &a5 55...2c3 would have transposed into the note to Black's 48th move; of course White didnt need to allow Black this second chance. 56.2£2 248 As would 56...2d2 now, but this was the last opportunity. Clearly Howell was following the alternative line of defence by trying to keep the king ‘cut off on h6, but Greet proves that the king is quite active there too... 57.21 2b6 58.2d2 2d8 59.25 Howell may have thought he had matters in hand here, but che text is the prelude to a fine breakthrough 59...2b6 60.f5! exfS 61.e6t White can also win with 61.84 &d4 62.e6+ hexe6 63.chxg6 Be5 64.8c3 £4 65.Rg1 66.2h7 FS 67.g6 hed 68.c5 hd3 69.g7 2xg7 70.chxg7 the2 (70...tc4 is like the note to Black’s 65th) 71.92f6. G1...exe6 62.cexg6 2d4 63.2h7 he5 64.96 hed 65.81 65.c5 was winning also, but the text leads to a pretty finish. 65.105 After 65...f4 White wins by one tempo:40. ‘The Berlin Wall 66.5 3 67.67 Sxg7 68.doxg7 ded4 69.2624 shed 70.2 f6 bb4 71.5 bxad 72.he4 a5 73.2xf3 Lbs 74.Le4 a4 75.he5 23 76.246 3277.4 66.2£2 Bal 67.2xc5 shd3 68.2441! 4 The point is 68...2xd4 69.c5 F4 70.c6 3 71.c7 £2. 72.c8Y FLY 73. YxaGt. 69.97 £3 70.g8H Sxd4 71.Wg6t bd2 72.8e4 10 Another one in which the half point was handed to and fro, but in the end a strong player went down with Black. One final example. Here White has succeeded in fixing the kingside pawns on dark squares, ue as compensation the queenside structure is much more favourable for Black, offering chances of genuine counterplay in some lines. Anand - Almasi Bundesliga 2003 Uk "@ HE Nw BU A> ow gh 29.263 A difficult decision, Wedberg. offers cwo alternatives. “The first is 29.g5 hxgs 30.2g4 g6 31.fxg5 28, of which he asks rhetorically ‘Can White win this?’ co which I would say the answer is a definite ‘no’. ‘The second is 29.f4. Black cannot then react passively because after g5xh6, f3-e4, and f5, he will fall into zugzwang, but after 29...bxc4 he should be fine: White can never win after 23, bxo3 ...c4, when he has to defend a3 all the time and Black’s bishop is free, while after 30.22 2d8 he can never actually approach the pawns with his king anyway in view of t2d2 2a54, so his bishop is tied down in a similar fashion on the queenside just the same. 29.4195 With the queenside structure from Deep Fritz — Kramnik White wins after this move with hxg6, followed by bringing the king to h3 and the bishop to h4, allowing ...h5xg4 if needs be, and augzwang, as in Dominguez = Dzhakacv, Linares 2002, and Dominguer’s notes in Informant 84, but here there can be no zugewang, and Anand has to be more cunning, 30.81! 30.hxg6t chxg6 31.%g3 (or 31.tee4 h5 32.gxh5t fxh5) 31...h5 32.f4 hxgd 33.doxg4 bxc4 both draw easily for Black. 30...b422 Falling for the trap. Almasi reckons that after .-b3- next he will be safe, although if he had asked himself why his opponent had chosen his last move he might have come up with the right answer. 30...bxc4 31.2d2 (otherwise 31...3 32.bxc3 cf) 31...8d8 32.he4 thg7 33.F4 gxf4 34.Bxf4 e7 draws. As before White can't get at the queenside pawns. Black can also ask White for his point with 30...2g7. 31.b3!! axb3,Chapter 2 - Typical Berlin Endings 41 31...bxa3 32.bxa4 a2 33.2b2 is hopeless. 32.a4 the8 33.che2 Qd8 34.chd1 &c7 35.8e3 Bxe5 36.2xc5 c3 This looks as if it loses a crucial tempo, but im fact after 36...2g7 37.2xb4 thd7 38.203 Le7 39.s8e1 Ebb 40.8b2 BIB 41.23 he5 42.a5 Six 43.26 Be5 44.87 White wins just the same. 37.hel 2g7 38.Exb4 2d4 38...2d8 39.243 fc7 40.2b2 Bf8 41.265t kb6 42.chb2 collects b3 in time to defend a4 and wins trivially. ‘This is the line Black would have been hoping to stop with 36...£g7. 39.246 Sxf2 40.bb2 2d4t 41.c2xb3 2g7 42.a5 bd7 43.06 he8 44.82 10 Another one gets away from Black’s point of view. x To sum up same-coloured bishop endings, they are always difficult to defend although Black should usually be able to draw. ‘Ihe main problem from Black's point of view is having no squares at all for his bishop and being forced ro make concessions as a result. IF his queenside structure is anything other than the dreaded a5/b6/c5/c6 then he should be able to defend, and almost always his defence involves activating his bishop when possible to attack ¢5, % and anything it can find on the queenside, and to prevent king penetration via g5 in particular. Opposite-coloured Bishops Opposite-colouted bishop endings, of course, are at the other end of the spectrum. White cant really hope to succeed simply with his extra kingside pawn. To win he will generally need to do one of three things: win a kingside pawn (in return for a queenside one of his own if needs be) so as to set up two connected passed pawns, have an enormously more active king (usually on the Kingside dark squares), or generate a passed pawn on the queenside also, often in connection with some sacrificial breakthrough. Naturally that commonplace of opposite-coloured bishop _endgames, zugewang, will play an active role. In our first example White’s 32.88 has just presented Black with a sharp choice between two of those alternatives Wang Pin - Gokhale Asian Women's Championship 2003 32...$8¢7? Unnatural and wrong. After 32...h5 33.gxh5 gxh5 34.25 Gokhale must have been afraid of White taking the king to ¢7 and breaking through with £4-F5, 06 ...£e6, £6, chen winning the bishop for the fpawn and winning the ending with the bishop and h-pawn against c-, e- and h-pawns. This would indeed have occurred after: 34...2e642 ‘The Berlin Wall Itwas possible however to draw by preventing this set-up with 34...2¢2 35.02f6 2c4 36.64 cB 37.65 (37.Le7 &e6 prevents f5 and holds since any attempt to put Black in zugzwang fails: if nothing else Black can simply jettison the c-pawn, for example 38.h4 &2c7 39.82e8 5 40.8xc5 kG 41.203 Bc7 42.f5 Axfs 43.doxt7 chd7.) 37...82d7 38.0hg7 tee8 39.25 Ba2, and White can do nothing. This set-up draws even without a c-pawn. 35,h4 Sp4 36.2f6 2e6 37.4 2b3 38.65 thc7 38...2b6 39.the7 cS 40.6 fre6 41.f6 5 42.67 4 43.ce8+ Skb5 44.2h6 Sxf7t 45.axf7 Bed 46.8496 ed3 47.8exh5 c5 48.294 is no better since after 48...e3 the unkind 49.8xe3 ¢2xe3 50.h5 wins with che spear check From hé6 at the end. 39.8¢7 5 40.8h6 Bc6 41.6 fre6 42.f6 5 43.7 White wins. 33.Sxh6 hd7 34.225 the6 35.64 Sf 36.h4 <5 37.268 c4 38.2b4 2d3 39.2c3 Le7 40.2b4t Leb 41.83 he7 42.h5! Correct. Amateurish would be 42.f5 gxf5 43.gxf5 o2f8 when White cannot win: 44.5 gS 45.h6 Sc2 46.2f6 Bh7 47.cox67 Bxf5 48.2d2 c3 49.Sc1 LeGt 50.82F6 2.51.06 2d3 52.e7 Bg6 53.te6 LeB 54.2d6 Bb5 55.87 hg6 56.hd8 BE7 57.hb2 hg6 and Black is saved by the c-pawn: without this White could indeed win by this method by taking her bishop to £8 here (she needs to cover both g7 and h6). 42...gxh5 43.2xh5 Of course not 43.gxh5 which fails as in the last note. 43.018 44.hb4¢ chg8 45.cbg5 Be2 46.295 2g7 47.203 Hf 48.25 he7 Black had no way to prevent White achieving this position, and now she cannot prevent White bringing her king to {6 followed by a decisive breakthrough. 48...2297 49.06} LEB 50.Lb4t he8 S1.exf7t Shxf7 52.g6t also loses for much the same reasons as the text. For the reason given in the note to White's 53rd, you might think that Black needed to try and get the bishop to d5 to prevent 6 ..fxe6, SExe6, but the bishop check from b4 followed by an appropriate king invasion is still fatal. For example 48...263 49.%2f6 sed 50.65 Bd5 51.Bb4t thes 52.82e7 Bed 53.96 fxg6 54.66 £2d5 55.c6, and White just queens. ‘The white bishop on b4 and the black pawn on 7 shutting the black king out from the king- or queenside, as the case may be, is a recurrent theme of Berlin opposite-bishop endings. 49.2b4t he8 50.016 2d3 51.65 B04 52.06 fre6 53.dixe6 For anyone such as myself who can use an occasional refresher on basic endgame theory, connected pawns on the fifth rank with opposite-coloured bishops win unless the defender is able to establish a set-up such as White Shed Ae5 £5 (any), Black &c8 7. Black plays &c8-d7-c8, and since White can never play cither S2d5 or e6+, he can't progress. With pawns on f5 and g5, the equivalentChapter 2 - Typical Berlin Endings 43 set-upwouldbet#f7 2h7.Sincei8isnotavailable to the bishop, though, White then passes with his bishop, and Black has to allow the fatal g6 (for the same reason, connected pawns on the sixth win almost always). It follows that White has already reached a theoretically winning position (the c-pawn makes no difference). 53..0d3 54.26 c3 55.2xc3 268 56.066 shg8 57.chg5 Bc2 57...2b5 58.96 28 59.2b4 illustrates the zugzwang point I was making before, 58.16 &b3 59.2b4 2c4 60.f4 Bb3 61.2e5 Bch 62.chd6 1-0 Blockade of course is a fundamental feature not only of opposite-bishop endings but of the Berlin in general. Here it looks as though Black has a fireproof construction, but even at two seconds per move (to be fair, a long period of aimless tacking led up to this position) Balogh proves this wrong. Balogh - Chen De Internet blicz 2003 Hye RU DAA & 57.4! Decisive. There's really no more to be said. 57..2xc4 58.04 27 59.25 Le6 60.26 2cB 61.a7 &b7 62.hd6 b4 63.c6 c5 64.e7 1-0 Patterson - Pitre Canadian Championships 2001 abcde f gh Another blocked-looking situation which White can resolve with a breakthrough. 55.e6! fre6 56.¢5 Black’s problem is that the kingside position freezes both of his pieces, so he can obtain no counterplay. 56...2F7 57.2d6 Bed 58.2e5 b5 59.xc5 a4 60.bxa4 bxa4 61.0b4 &c2 62.825 Le4 63.86 03 64.c5 a2 65.c2d7 Putting Black into zugewang, although in fact 65.h7 &xh7 66.%2xc6 was winning very simply. The position well illustrates the theme of the ‘single diagonal’ in opposite-bishop endings; White's bishop renders both Black’s pawns irrelevant from its position, and indeed defends £6 as well, so that White just wins the bishop for the c-pawn in peace and Black can't do anything at all. If the e-pawn were on say 6, so that the bishop couldn't do this, the outcome would be different.44 ‘The Berlin Wall 65.8452? 65.818 was more resilient — Black can spare the e6-pawn casily — and would have forced White to find the plan mentioned in the last note. 66.h7 1-0 Berescu - Madan Romanian Championships 2005 8 1 6 5 4 3 2 1 Material is equal but White's active king and, most important of all, his ability to cut the black king off from the h-pawn, enable him to win. 49.207 In fact White could also win with 49.246 at once since after 49...2d7 50.tbe5 a4 it is crue that he cannot retreat the bishop so as to go ‘hf6-g7 without leaving it en prise, but on the other hand he doesn't need to: 51.926 daxd6 52.cexf7 wins easily. Still, lengthening the d6- £8 diagonal for the bishop as in the next note is thematic. 49...a4 Ie’s tempting co hold two with one, but this move loses because White's capture attacks another pawn and Black either has to let his king be shut out or go two pawns down. No better however was 49...b5, not because of 50.2xa5? FB 51.%8e5 c4, when Black's bishop can defend the queenside single-handed from d7 and his king holds the kingside from g8, but 50.2d6!, shutting the king off decisively and following with te5-f6-g7 and wins. 50...82d7 51.he5 c4 52.25 tec6 53.818 even preserves White's bishop as well. 50.2xb6 50.246 Hd8 51.c%2e5 was still winning casily, bur the text is also adequate. 50.228 50...c4 51.Bc5 loses just like 49...b5, but Black must have placed his hopes on this move; it looks as though his king can hold one flank and the bishop the other, and indeed this would be so had he just one more tempo... 51.2xc5t tg8 52.b4! Just so. “Box”, as the Informant symbols have it. If Black could play ...2b3 he would be safe, since in order to remove the bishop from that square White would need to simplify the kingside in order to prevent the black king simply playing from g8 to h7. And then Black can hang on, thus: 52.244 2b3 53.g6 fxg6 54.hxg6 WF8 55.c2g5 tgs 56.82h6 AB! (56...8c2 57.b4 axb3 58.g7 is like che game: if 58...b2 59.2xb2 £7 60.24) 57.2h7 (57.816 &g8 58.227 shifts the bishop but to no avail: 58...8c2 59.b4 axb3 60.2b2 &xg6=) 57...&c2 (a typical device for preventing pawns advancing in opposite-bishop endings), and White can't progress, Now b4 is useless since the king heads for a8 and the bishop controls, the g-pawn, so there follows, for example, 58.8c5t theB 59.c2g7 (59.2h6 Bb3 60.g7 &g8 61.c2g6 wd7 gives Black a blockade) 59...204 60.%8h7 2c2 61.h6 2b3 with a draw.Chapter 2 - Typical Berlin Endings 45 52...axb3 53.244 ‘And it’s over. White wins the bishop for the a-pawn, returns and snacks on the b-pawn and wins on the kingside at his leisure. Black can't obtain any counterplay because once again the bishop multi-tasks on the single diagonal, preventing ...b2 but also preventing any assault by ...f6 on the white kingside pawns. 53..02h7 53...c2f8 54,h6 the7 loses on the kingside instead: 55.a4 thd6 56.a5 thd5 57.2b2 2e6 58.26 tec 59.ce5 chb6 60.c2f6-g7 and so on, 54.2b2 2d1 55.h6 2 g6 56.a4 Be2 57.206 2d1 58.a5 Le2 59.the3 a6 60.2d2 1.0 Tseitlin - Postny Dov Porath Memorial, Israel 2000 fo gh My final example of White victory again shows the vital importance of establishing widely-spaced passed pawns in these endings, but also introduces a theme which more commonly works in Black's favour, where one side in a fixed pawn structure is able simply to remove his opponent's entire pawn skeleton on abcede cone wing, since these are fixed on the colour of the opponent's bishop, the defending king is far away and the bishop can't help them. We join the game as White has been struggling for some moves to show that his rook really is more effective than the mighty bishop on e4, which is both athletic and stable. 49...b4 ‘The young Evgeny Postny (a Berlin expert by the way) must have been feeling good about this move, 50.8d2 4 51.2Qxf4 bxc3+ isn’t particularly clear at all. 50.2h6t!! Maybe the rook wasn't better than the bishop after all? 50...2xh6 50...S2f7 51.2d2 £4 can now be met by the deflating 52.8h4, so Black doesn't have a lot of choice. 51.gah6 bxc3t ‘After any other move White wins with ¢4, hcl-d2, and then &c7xa5xb4, for example: 51...2d3 52.04 Be4 53.chel BF3 54.cbd2 Be4 55.87 Soxh6 56.Lxa5 hp5 (56...f4 57.2xb4 25 58.25 Bc8 59.chd3 thes 60.chd4 shfs 61.c8c5 £3 62.21 changes nothing) 57.2xb4 Shf6 58.a5 the 59.26 thd7 60.a7 c5 61.2xc5 52.chxc3 2d5 Nothing makes any difference: after 52...c5 53.the4 Qc2 54.thxc5 &xb3 55.00b5 it’s the single diagonal again; the bishop holds f4 and h6 and White's king wins the bishop. 53.b4 axb4t 54.chxb4 Even if the c-pawn were on 2 now it wouldn't help Black; the f4 bishop holds everything. 54...8:g8 55.05 h5 56.26 tig4 57.07 1-046 ‘The Berlin Wall Frequently however Black actually has the advantage in these endings. For the reasons T discussed in chapter 1, his bishop can be a mighty piece which defends the whole kingside from say 6, while also helping on the queenside. As we will see in the theory section; very often the structure a4/b3/c4 vs. a5/b6/ €71c5 arises, and you won't have to play very many internet blitz games with che opening before you win an opposite-bishops ending for Black which starts with the move ...fic2, and continues with the removal of White's entire queenside. This fate can overtake even the world’s elite as White. Here Naiditsch is actually a pawn down, but still the threat of ...&d1 gives him the advantage. In the normal way Black's pawn would be on c5 rather than d4, and if that were the case White would have no defence at all against this procedure. Here at least he can trade some of his queenside pawns rather than just losing them. Bologan - Naiditsch European Cup, St Vincent 2005 35.2d2 b6 36.b4 36.h4 amounts to resignation after 36...8d1 37.d2g2 Qxb3 38.h5 Lxc4 39.h6 £d3, so the only alternative is 36.c5 bxc5 37.xa5 dl, when Postny (and presumably Bologan, since he can’t have been under many illusions about the text) thinks thar Black is clearly winning, although I'm not entirely sure. After the long variation 38.8c1 &xb3 39.a5 c4 40.h4 c3 (40.,.8d1? 41.06 Bc6 42.82 forces 42...d3 with a draw) 41.h5 eG 42.6 Bc? 43.812 Lc5 44.26 Bed 45.06! fre6 46.7 2 47.284 Bxa8 48.2xd4t Bxd4 49.h7 cl 50.h8Yt 5 51.!xa8 it seems probable that Black is winning, but accidents can occur in this sort of line. 36...axb4 37.2xb4 c5 38.2d2 Re2 39.24 Qxc4 40.g3 243 41.264 04 42.2b4 3. 43.h4 be6 44.h5 hd5 45.25 bxa5 46.2xa5 heh 47.06 Black was getting ready for ...&h7 and 23 followed by ...c2, so it was time ro get desperate. 47 x06 48.2¢5°BH7 49.95 d3 50.0xc3 chxc3 51.g6 d2 0-1 52.gxh7 d1¥ 53.h84 Wd4t settles the matter. Kilian - Sielecki Germany 2005 By yy mw HNO RUA ADO eeChapter 2 - Typical Berlin Endings a7 Black's pieces have occupied the critical posts, and now he transposes into the opposite- bishops ending. 27.63 28.Dxf3 28.cbf2 @xg5 29.fxg5 would have prevented the immediate disaster, but Black still has every chance to win in the long run. 28...Exf3 29.g5 Ed1f 30.2xd1 &xd1 White's pawns on the kingside are now crippled, while on considering the queenside it is hard to refrain from the expression ‘ripe apples. 31.b4 axb4 32.Sxb4t c5 33.243 Sxc2 34.05 ‘The only way to keep from losing two pawns on the queenside. It is Black's next move which gives this example its value. 34.8d3t ‘ After 34...bxa5 J don't see how Black can win at all, and if he can its certainly a lot harder than the text. For example: 35.faxc5t eG 36.82 hd5 37.8f8 96 38.47! ‘A typical method of defence: pawns don't count compared to obtaining a single diagonal from which the bishop can do everything. 38...265 38...c8e4 39.06 fre 40.865 5 41.82 Bd3 42.9%c3 a4 43.8d6 thd5 44.865 Led 45.8h8 23 46.hd2 dac4 47.25 Hb3 48.26 4 49.05 da? 50.hcl &d3 51.2d4 is the same: as long as White doesn't let the king to d3 I don't see what Black can do. 39.c6 Bxe6 40.265 5 41.hd3 a4 42.02 hed 43.8d6 hd4 44.he5¢ ed 45.266 a3 46.205 B£5+ 47.822 Black can’t go further. After the text however Black retains two connected passed pawns and wins easily. 35.2 heb 36.03 a6 37.24 2b7t 38.%0e3 g6 39.c2d3 ed5, 0-1 Black is ready for ...b5 and thereafter with any kind of care the connected pawns win. Rook Endin; Is hard to generalise about rook endings since they don't often arise without some changes in the structure, and activity of the pieces is everything. What can be said is that the characteristic ending which will be familiar to Ruy Exchange practitioners is usually a draw. ‘A yypical example is the following: Gdanski - Grabarczyk Polish Championships, Warsaw 2001 HN WRU AA @ 42.8h5 247 In view of 42...2e6 43.85 He7 44.8 £6 Black cannot simply defend his pawns, so he decides to go active at once. 43.2 White’ plan to defend the b3-pawn with his king does not bring success, and generally will48 ‘The Berlin Wall not do so in such positions. If White wants to try and win, and doesnit mind taking a chance con overstepping the mark and losing, then he has to risk abandoning his queenside with 43.Bxh6, After that the position is still within the bounds of the draw, but Black definitely has to play more carefully, and right away has a decision to make, 1) Greediest is 43...2d2+ (computers tend to favour this move) when Black has at least to play very accurately: 44,323 (44.chg3 Bd3+ 45.8h4 gets nowhere: 45...82d7) 44...Hb2 45.Bf6 Exb3+ 46.the4 Exh3 47.2xf7 and now Black has a choice: a) 47..S2d8 48.g5 HeB (48...8g3 49.hf5 EE3t 50.chg6 Ba3 51.0h7 Bxad 52.g6 Excd 53.87 Sigs 54.g8¥t Exes 55.chxg8 is lost, for example 55...a4 56.227 b5 57.2266 b4 58.d26 shc8 59.chd5 b3 60.2F1 a3 61.%8c4. If Black wanted to do this he shouldnt have played ~-Xd8.) 49.96 Bc3 50.8F5 EBT (50...Bxc4? 51.Bf6 Bel 52.g7) 51.che6 Exf7 S2.gxf7t sf 53.chd7 (and not 53.cf622 expecting 53...c6 55 S86 and wins but in fact running into 53...b5 0-1) 53...b5 54.axbS a4 55.bxc7 a3 56.b6 thxf7 57.b7 a2 58.b8H al 59.8b7 Black has a fairly horrid defence ahead. b) 47...8c3 48.g5 Exedt 49.00f5 Bxa4 50.g6 Bal 51.g7 gl 52.868+ hb7 53.g8Y Exgs 54.2xg8 with a position I hesitate to assess; it is clear that White is not losing, and it is interesting to note that if Black plausibly heads for the well-known draw with a3/b4/c5 and is so incautious as also to play ...c6, he is likely to lose because W:a4 the4 B: 2b6, a3, b4, c5 and 6G is zugewang, whereas without the offending ©6-pawn ...2c6 is a theoretical draw. c) 47...8h1 48.g5 8g] 49.5865 b5 50.cxb5 4 51.Ef6 hd7 51...€3 52.86 Hcl 53.b6 52.86 Bft I think this is the only move: 52...fcl 53.g6 3 54.g7 BAt 55.hg6 Bgl t 56.07 EFT 57.thp8 BF3 58.Bcd! Bd3 59.Bc5! Zugawang! 59...2d6 (59... ¢3 60.2d5¢ eG [60...22c8 61.8h5 c2 62.8h1) 61.Bh5 c2 62.85 Bg2 63.2xc7 ed6 64.b6) 60.cf7 BBt 61.che8 Be3+ 62.c2d8 Bg3 63.Bxc7 c2 64.8xc2 Fixg7 65.2c6t and White wins. 53.c2e5 Bgl 54.0266 SFI t 55.8297 Bel Only once White has been forced to block his own pawn. 56.g6 c3 57.deg8 c2 58.g7 dB 59.85 ‘Threatening 2d5}-d2-h2. 59...€28! 60.b6 c6! 61.8xc6+ 2b7 And Black holds. 2) 43...8d3 “This is easier. 44.86 Bxh3 44...Lxb3 45.8292 2d7 should also draw. 45x17 dB 46.g5he8 47. 2B! Or 47.26 Bh6 48.876 te7 49.806 we8 (49.872? 50.g7 wins) 50.43 Bhi 51.994 Bb] 52.8xc7 Exb3 53.265 28+ 54.c0—5 Hc3 55.0hf6 BGT S6.che6 Bd3 57.27 8d4, 47..8h5! Much worse is the immediate 47..2h2t 48.93 Hh 49.che4. 48.96 Bh2t! Certainly not 48...2g5? 49 2g3, nor 48... £h8? 49.23 Bg 50.8g3 PB 51.che4 hg7 52.shd5 Be8 53.8c6 Be7, when White breaks the fortress with 54.82b7-c8-d8) 49.93 BhG 50.286 the7 51.26 bf 51...2d7?? 52.g7 we know already. 52.chg4 chg8 53.00f5 Bh3 54.Bxc7 Bxb3 55.06 Bd3 And Black draws. 43..8e7t 44.cbd2 Bd7t 45.cbc2 Be7 46.8xh6 He3 47.2h7 8g3 48.2xf7 Exh3 Ya-Va