|
|
|
The Instructor Mark Dvoretsky
|
There's Always a Defense! Eternal vigilance is the price
of supremacy. I'll start by thanking Charles Sullivan for his instructive addition to the analysis of the ending of Lev-Shirov, London 1991 (from my January column).
1...? In the game there followed 33...Rg4? 34.Nc5 b6 35.d7 Rd4 36.Rh8+ Kf7 37.d8Q Rxd8 38.Rxd8 bc 39.Rd5, and White capitalized on winning the exchange. Sullivan pointed out the correct plan of defense: play to exchange off all the remaining white pawns on the board, preparing to give up the bishop for the most dangerous of them: the passed d6-pawn. 33...a4!! 34.Nc5 (34.ba Ra2 35.d7 Rxa3+; 34.d7 Bxd7 35.Rxd7 Ra2) 34...ab 35.d7 Bxd7 36.Nxd7+ Ke8 37.Nf6+ Kd8 38.Rxb7 Rg3+ 39.Ke4 Kc8 40.Rb4 b2 41.a4 Ra3= with the subsequent 42...b1Q. I'll use this case to reply to a comment from one of my readers about the question marks under the diagrams. I put them there in all cases when the position is suitable as an exercise for you to solve on your own. Moreover, it doesn't have to be connected with the subject being discussed at all. When I'm working with strong players I don't limit myself to merely describing the material I've prepared, but I stop at every appropriate case and suggest that my student solves the problem that's facing the players in the game we're studying. Then we discuss his choice and evaluate the right ideas and the mistakes in the decision he made. I won't insist (although it probably is the case) that this approach is effective in lessons with weaker players, but I've already been convinced many times of its expediency for my students. It's useful for practical players to constantly test their mastery in the most varied spheres, including, of course, those that don't directly fit into the lesson program. On the same principle I also stick with articles that basically imitate my lessons to a great extent. The reader doesn't have to solve all the exercises, of course, but I think I'm right in pointing them out in the text for those who want to train themselves. Furthermore, I would also like to recommend that other authors use the same approach in their articles and books. Another of Sullivan's suggestions concerning one of the exercises from the December column served as inspiration for writing this column. R. Liberzon-Belov
1.? White is a piece down and 1...Rxa2 is threatened. He needs to outrace his opponent by immediately starting an attack against the enemy king. But how to do this? He can't achieve his aim with 1.Nxf6? Bxf6 2.Rgh1 h6 3.Qxf4 Kg8! 4.Rxh6 gh 5.Rg1+ (5.Qxh6 Bg7 6.Rg1 f6-+) 5...Kh7! (a draw after 5...Bg5 6.Rxg5+ hg 7.Qxg5+ Kf8! 8.Qd8+ Kg7 offers little for Black) 6.Rh1 Bg7 7.f6 Qb4! 8.Qf5+ Kg8 9.Rg1 Qd4 10.Rxg7+ Kf8-+, and White doesn't succeed in continuing his attack because of the impending threat of 11...Qd1+ 12.Kb2 Rxa2+! 13.Kxa2 Ra7+. An impressive combination will help, starting with a deflecting queen sacrifice. 1.Qe1!! Bxe1 (1...Rxa2 is refuted by 2.Rxh7+! Nxh7 3.Qe8+ or 2...Kxh7 3.Nxf6+) 2.Nxf6 gf (2...h6 3.Rxh6+!) 3.Rgh1 Kg8 4.Rxh7 (threatening 5.Rh8+ Kg7 6.R1h7) 4...Kf8 (4...Bh4 5.R7xh4!) 5.Rh8+ Ke7 6.Rxe1+ Kd7 7.Rhe8! with unavoidable mate. White's first move is strictly an only move (all others lose), and we've checked the main variations associated with accepting the sacrifice, and also with an immediate counterattack on the queenside – so the exercise has been solved! In a practical game you wouldn't have to search for your opponent's best defense: it's his problem, as they say. There's no point doing this in training, either, so as not to develop the bad habit of calculating unnecessary variations, expending time and energy in vain. If a new problem is posed it's another matter, though. After 1.Qe1!! is Black really doomed to defeat? What is his best and most stubborn defense? To answer these questions we must again check and conclusively refute the most natural continuations that we previously analyzed, and then look for new candidate moves. Above all, 2.Nxf6 is threatened with a subsequent mate on e8. Let's try to cover this square without being troubled by the loss of the bishop (it's extraneous in any case). Let's calculate 1...Qb5!? 2.Qxc3 (2.Nxf6 Bxf6 isn't dangerous) 2...Nxg4.
1.? A nice discovery: 3.Rxg4? is impossible because of 3...Qf1+ 4.Kb2 Rxa2+! 5.Kxa2 Ra7+, and Black mates. On 3.Rhg2? the reply 3...Qb4! is strong. However, White has a very strong retort: 3.Rhh1!!. The first rank has been protected, White wants to take the knight, and it can't retreat: 3...Nf6? 4.Rxg7!. No good now is 3...Qb4? because of 4.Qh3 Nh6 (or 4...Nf6) 5.Rxg7!. He has to play 3...f6, and after 4.Rxg4 White wins back the piece, obtaining a clear advantage. Then again, this advantage isn't a win yet, and it's clear that Black's situation here is better than in the variations we looked at earlier. And that's besides the question of whether his opponent will find the subtle move 3.Rhh1!!. However, before making our choice, let's see if Black has anything else. The e8-square can also be protected by a rook, and it isn't difficult to work out which rook. Let's study the position after 1...Ra8! 2.Qxc3.
1...? Here 2...Nxg4? 3.Rxg4 no longer works, although 2...Ne4 looks tempting. Calculation shows that the rook sacrifice 3.Rxh7+ Kxh7 4.Qh3+ Kg8 5.Qh5 is enough, at most, to maintain equality (Black can, for example, play 5...Kf8 6.Qh8+ Ke7 7.Qxa8 Ra7). Considerably more dangerous is 3.Qh3! Ng5 (the only move) 4.Qh5 Qa7 5.a4. In the event of 5...f6 6.Qg6 Black has no satisfactory defense to the multiple threats. Also clearly in White's favor is 5...c4!? 6.Rgh1! f6 (defending from 7.Qxh7+!) 7.Qg6 h6 8.Nxh6 gh 9.Qxf6+ Kg8 10.Rxh6 Nh7 11.Qe6+ Kh8 12.Qxd6 cb 13.cb. It's necessary to immediately switch to a counterattack by means of 2...c4!. Precisely this move in combination with the previous one was suggested by Charles Sullivan. When I checked his recommendation I came to the conclusion that a drawn rook endgame arises from best play from both sides. 3.Rgh1 (the rook moves away from the queen's attack) 3...cb 4.Qxb3 Qa7! 5.Nxf6 h6 6.Ng4! (he has to act as energetically as possible: in the event of 6.Kc1? Rb8 Black will attack) 6...Rb7 7.Nxh6
In the variation 7...Rxb3+?! 8.cb f6 9.Nf7+ Kg8 10.Nxd6 Kf8 11.Rh8+ Ke7 12.Nc8+ Rxc8 13.Rxc8 the two white rooks are evidently stronger than the queen (although the computer evaluates the position that has arisen as equal). 7...gh 8.f6! (8.Rxh6+? Kg8 loses – the rook is needed not only for the attack, but also to defend the second rank) 8...Qxa2+ After 8...Kg8 9.Rg2+ Kf8 10.Rxh6 Qxa2+ 11.Kc1 Qa1+ 12.Kd2 Qxf6! 13.Rxf6 Rxb3 14.cb Ra2+ 15.Kc3 Rxg2 16.Rxf4 Black still has some problems, although objectively the position is probably drawn. 9.Kc1 Qa1+ 10.Kd2 Qxh1 11.Rxh1 Rxb3 12.Rxh6+ Kg8 13.cb Ra2+ 14.Kd3 f3 with a certain draw. I emphasize: here, as in all the subsequent examples, we're not talking about refuting your opponent's idea. The problem posed for discussion and training is how to avoid losing heart under powerful attacks; stay calm and find a relatively good way out of the difficult situation that has been created. Those of you who carefully studied my article about critical moments in the battle that was published in the July, August, and September 2008 columns will most likely have noticed one of the most important conclusions drawn in it: In many critical positions the successful solution found by one of the players doesn't objectively upset the balance of power very much, but simply poses a problem for the opponent. The latter often can't cope with it, and only then starts to endure serious difficulties. In other words, positions often turn out to be critical for both sides at once. And sometimes a critical position for the opponent arises slightly later - after a series of more or less forced moves. In literally all the examples analyzed in that article the weaker side could defend successfully, but didn't solve the problem posed by the opponent and lost. I've observed a similar scenario when I offer my students (strong players) the positions given below. They confidently determined the correct path for the attacking side, but when I then asked them to find the best defense against the idea they'd suggested, the grandmasters couldn't cope with the new task. This hardly means that attacking is easier than defending. More likely, it's because of their lack of sufficient practice at solving complex defensive problems. In collections of exercises the overwhelming majority of examples are usually associated not with defending your own positions, but with breaking through the enemy defenses. Well, psychology also plays a significant role, of course. On the receiving end of an unexpected and strong attack from his opponent, a player often becomes flustered, his belief in his own success is weakened, and he's no longer capable of calmly and precisely analyzing the variations that arise. In my article dedicated to critical positions the weaker side mainly had to face strategic threats, so here the problems will be purely tactical. Let's take an example that has appeared in many books, although it's hardly ever analyzed adequately. I found the solution to the position many years ago, and computer analysis later allowed me to discover or clarify a number of new interesting details. Fischer-Sherwin
1.? Useless is 30.Bxf7+?! Kh8. It's possible to fight for the advantage only by delivering the straightforward combinational blow 30.Rxf7!. The rook is unassailable, of course, but White had to consider his opponent's more aggressive retorts that could have refuted his idea. 30...Qxd5? doesn't work because of 31.Rxf8+ Kxf8 32.Qf1+!. On 30...Qc1+? White doesn't reply 31.Rf1+? Kh8, but 31.Qf1! with an immediate win. On the move that followed in the game, 30...Rc1+? it was necessary to foresee the same powerful reply 31.Qf1!! (31.Rf1+? Kh8 even loses). After 31...Rxf1+ 32.Rxf1+ Qxd5 33.Rxf8+ Kxf8 32.ed White is left the exchange and a pawn up. No help either is 31...h5 32.Qxc1! (with 32.Rxf8+? Kh7 Black is fine). That's everything White had to calculate in undertaking the combination. The game ended like this: 32...Qh4 33.Rxf8+ Kh7 34.h3 Qg3 35.hg h4 36.Be6 1-0 And now let's try and answer a more difficult question: how should Black have defended? Before trying to find the strongest defense let's look at some other, less successful tries. The counterattack 30...Ne3? is most convincingly refuted by means of 31.Qxe3!. Almost as strong too is 31.Qf1!. The move 30...Rfc8? doesn't pose any serious problems for White either. The most precise retort is: 31.Rc4! Rc1+ 32.Qf1, but also possible is 31.Qf1!, not fearing the combination 31...Qxd5!? 32.ed Ne3 because of the impending sacrifice: 33.Rxg7+! Kxg7 34.Ra7+. It's more difficult to find something wrong with a retreat to с8 by the other rook: 30...Rcc8?!.
1.? Black intends to move the king away to the corner, retaining a defensible position. His idea fails only because of the effective deflection 31.Ra8!! Rxa8 32.Rf5+ Kh8 33.Rxg5, and if 33...Rac8, then 34.Bc4!. For Black this variation is preferable to the previous ones from a practical point of view, as it poses concrete problems for his opponent, leaving open the possibility that he'll make a mistake. But still, he has a much better solution. 30...h5!!
1.? Black has defended from the move 31.Qf1 that was threatening him, on which now follows 31...Kh7! (but not 31...Rxf7? 32.Qxf7+ Kh7 33.Qg8+ Kh6 34.Qh8+ Kg6 35.Bf7+! Kxf7 36.Ra7+ Kf6 37.Qd8+ Kg6 38.Rxg7+ +-) 32.Rxf8 Rc1, and White already has to play accurately to avoid getting into difficulties. For example, losing is 33.Be6? Ne3! 34.Bf5+ Nxf5 35.Rxf5 Rxf1+ 36.Rxf1 Qb5! (an unexpected double attack!) and the rook on a6 is lost. Nor does 31.Rxf8+ Kxf8 32.Qf1+ Qf6! achieve the aim (weaker is 32...Nf6 33.Rc4+/-). For example, 33.Qb1 Nf2+ 34.Kg1 Nh3+! 35.Kh1 (bad is 35.gh? Qg5+) 35...Nf2+ with perpetual check, or 33.Ra8+ Ke7 34.Ra7+ Kd8 35.Rf7 Qxf1+ (also possible is 35...Nf2+ 36.Kg1 Qh6!) 36.Rxf1 Ne3 with a drawn endgame, or 33.Qf6+ Nxf6 34.h3 Nxd5 35.ed Rd3 36.Ra5 Kf7 – White's extra pawn has no significance, as his pieces are positioned worse than the enemy pieces. White's advantage is only preserved with 31.Rc4! Rxc4 (mistaken is 31...Rxf7? 32.Bxf7+ Kxf7 33.Rxc3 Qe5 34.Rf3+ Ke7 35.g3+- or 32...Kh7!? 33.Rxc3 Qe5 34.g3 Qxc3 35.Bxh5+-) 32.Qxc4! Significantly weaker is 32.Rxf8+?! Kxf8 33.Qxc4 – Black plays either 33...Nf2+!? 34.Kg1 Nh3+ 35.Kf1 Qf6+ with real chances to save himself, or 33...Qe5!? (but not 33...Qf4? 34.Qc8+ Ke7 35.Qc7+ Ke8 36.Bc6+ Kf8 37.Qd8+ Kf7 38.Qe8+ Kf6 39.Qf8+) 34.Qc8+ (34.Qf1+ Ke7 35.Qg1 Qf4! 36.Qa7+ Kd8! 37.Qb8+ Kd7!=) 34...Ke7, and the exchange of queens on f5, which White can force, doesn't guarantee him a win. 32...Rxf7 33.Bxf7+ Kh7 (worse is 33...Kf8 34.Be6+/-)
1.? The double attack 34.Qc8?! (threatening 35.Qg8+ and 35.Qf5+) is parried by 34...Nh6! 35.Bd5 Qd2 36.h3 Qe1+ 37.Kh2 Ng4+! with perpetual check. He has to play 34.Qf1! g6 35.h3!?, repelling his opponent's direct threats and retaining the extra pawn. But the entire battle is still ahead, of course, and its outcome remains undetermined. If we recall the solutions to the positions we've just studied, we can draw the conclusion that training in this area also enables us to develop one of the most important elements of the technique of calculating variations: the ability to determine candidate moves. Spielmann-Tartakower
Rudolf Spielmann sacrificed the exchange for the sake of an attack on the king, and his strategy justified itself after a serious blunder by his opponent: 15...Ne5?. The game ended like this: 16.Nxg7! Qd8 17.Ngf5 Ng6 18.Qh6 Ne8 19.Nf3! Bxe3+ 20.fe Qf6 21.Ng5 Qh8 22.Ne7+ 1-0 It was possible to defend using different methods. For example, 15...Rfe8!? or 15...g6!? 16.Qh6 Ne8 – in both cases with reciprocal chances. But we'll concentrate on discussing the move mentioned by Spielmann, 15...Kh8!?. The first problem: how should White continue the attack? This recommendation by Spielmann doesn't withstand criticism: 16.g4? g6 17.g5 Ng8 (probably stronger was 17...Nd5!) 18.Rg1 with the subsequent Rg1-g3. Black easily repels the threats by playing 18...Rfe8 with the subsequent 19...Nf8, or, even better, 18...Ne5!. The correct answer to the question is that he has to sacrifice a piece, 16.Nxg7!.
1...? Of course, the answer is incomplete – we also have to bring in the accompanying variations. In contrast to the previous examples, I won't do this right away, so as not to make the solution to the next problem, the most important one, any easier: searching for the best defense for Black. In practice when we receive this kind of blow, as a rule we first have to figure out our opponent's idea and check whether it's correct, and once we've convinced ourselves that "straightforward" play won't help us, look for additional, "side" resources. It might seem that on 16...Kxg7 White was preparing 17.Nf5+ Kh8 18.Bg5, intending after 18...Ng8? to continue the attack by means of either 19.Be4! Bxf2 20.g3+-, or 19.Ne7! f5 20.Bxf5+-. However, the knight retreats to another square: 18...Nd5!. For example, 19.Ne7 (19.c4!? Bxf2 unclear) 19...f5 20.Bxf5 Rxf5 21.Nxf5 Bxf2 22.Qe4 Rg8 with reciprocal chances. The variation isn't completely forced, but, in any case, no direct threats for Black are visible, and he's right to go into this kind of play. In actual fact, as grandmaster Igor Platonov has shown, taking the knight can be effectively refuted. 16...Kxg7? 17.Bxh7!! Nxh7 18.Ne6+! fe (no help either is 18...Kh8 19.Bxc5+-) 19.Rxd7+ Rf7 20.Qh6+ Kg8 21.Qg6+ Kh8 22.Rxf7 (or 22.Qxf7) 22...Bxe3+ 23.Kb1! with unavoidable mate. Grandmaster Vadim Zviagintsev has established that by declining the sacrifice Black preserves equality. 16...Rg8! 17.Ngf5 Rg4 Dangerous is 17...Rxg2 at least because of 18.Ng3 (even stronger, apparently, is 18.Nh6 Rf8 19.Bf5). 18.Qh6 Rg6 Doubtful is 18...Rag8?! 19.Nf3!, but 18...Rxg2!? already makes more sense than it did a move earlier: as White doesn't have a single one of the retorts mentioned in the previous note. A rook retreat still looks more solid. On 19.Qh3 follows 19...Ne5, and in the event of 19.Qh4 or 19.Qf4 it's possible to at least repeat moves: 19...Rg4. With that we'll take a break for now. You'll see some new examples on the same theme in a month's time. © 2010 ChessCafe.com. All Rights Reserved. Comment on this month's column via our Contact Page! Pertinent responses will be posted below daily. |
Purchases from our shop help keep ChessCafe.com freely accessible:
|
|
|
|
[ChessCafe
Home Page] [Book Review]
[Columnists] © 2010 BrainGamz, Inc. All Rights Reserved. |
|
