(1) Radjabov,T (2735) - Van Wely,L (2681) [D43]
Corus A Wijk aan Zee NED (3), 14.01.2008
Black has been pressing for a long time, but now he allows his opponent an unexpected transposition to a theoretical endgame.
[Garry Kasparov pointed out that van Wely missed an "elementary" win with
. John Nunn, who had all the necessary tablebases installed, confirmed that it was "elementary".]
This is an important theoretical position. The evaluation and the recommended methods of play have varied along the decades. Of course, things have been settled from theoretic point of view after the creation of the 6 men tablebases, but in many cases one does understand very little about the endgame by learning that a cople of moves draw, while all the others lose in 38 moves or less.
In the subsequent analysis, I have mainly relied on the lines provided by Averbakh in his 5-tomes endgame manual (the best ever published, in my humble opinion). It is remarkable that his analysis does not contain "real" mistakes. Although the tablebases sometimes recommend lines that win with... 1 or 2 moves earlier (36 instead of 38, for example), I have stuck to the human lines, because they are easier to explain as a part of the same general plan. (In fact, there is a lot to be explained, because Averbakh's lines have huge instructive value, but only if you draw the rules yourself, because the verbal explanations are rather restrained.) Computer lines sometimes rely on casual tactics, which are nice, but are more difficult to remember if you get this endgame in an over-the-board game.
Among other things, chess is supposed to be an exciting, educative and entertaining game, or, to put it briefly, just fun. By watching the "cold" recommendations of the tablebases, one risks getting a feeling of false strength, similar to that of driving a BMW at 200 kms per hour, which has nothing to do with one's physical condition. It tends to generate exaggerated criticism of top GMs, who sometimes overlook "such an obvious win". The notion of "fun" is basically related with the pure human effort in difficult and complex situations; the current position is complex enough, as we shall convince ourselves soon.
Here is a small episode that illustrates quite well what I mean. When this endgame was reached, none other than Garry Kasparov became enthusistiacally involved in the evaluation of the position, mentioning the crucial game Salwe-Rubinstein (see below) and also two of his own games (against Pinter and Jussupow, both played in 1993; more about these games later). The great champion tried to remember how or if at all it could be won, what were the best chances, etc. rather than simply connecting to the tablebases. A human effort from a great player who loves chess passionately...
Let us return to our analysis now. To start with, the diagrammed position is drawn.
Black's only winning chance consists of sacrificing the exchange for the a4-pawn, but he cannot achieve this under favourable circumstances. Things would be different if the pair of pawns would be placed on any other file, but it is known that rook pawns offer additional saving chances in pawn endings. However, in order to reach the desired result, White has to avoid 3 (!) main situations. 1) His king has to avoid being pushed onto the b-file or further. 2) His king has to avoid being pushed onto the 5th rank or further. 3) His king has to avoid being trapped in the immediate neighbourhood of the pawn! The first two points are easy to explain. In case of an exchange sacrifice, the king has to be in time to reach the b1-square. However, the latter aspect is far from obvious and, for many decades, remained unnoticed by theoreticians. Let us move on now and take contact with the mentioned situations as they arise in some of the sidelines and with the methods of achieving (or, on the contrary, avoiding) certain situations.
[The only good move! 48.Kb2?
would lead to the situation corresponding to point 3), which was met (with reversed wings) in a famous game Salwe-Rubinstein, Prague 1908, quoted in most ending books. Rubinstein managed to win the game, but only by a... mistaken method! He chased the king away from the dangerous zone and eventually succeeded in pushing it until the 5th rank, after which the win was relatively easy. In 1954, Baranov discovered that Black can avoid being pushed too far and that the logical result would have been a draw. His conclusions are valid up to this moment, but only for the case when the defending king escapes the zone surrounding the pawns. Later, the renown endgame expert Maizelis (who was one of Averbakh's collaborators for the first edition of his books, published in the late '50s), discovered the dangers facing the king in the neighbourhood of the pawns. Lysytzin's famous endgame manual, first published in 1956, only mentions Baranov's conclusions, without showing how Rubinstein could have won... without the help of his opponent. Anyway, after 48...Kd2
we reach a position from Maizelis' analysis.Black's pieces are not optimally placed yet. In order to win, he has to carry out a major regrouping. His king should reach b4, in order to restrict the bishop's mobility, but the enemy king should not be allowed to escape from the dangerous zone at the same time! 51...Rh3
Black has strengthened his domination, but the b4-square is not available yet. 55.Bf7
(White stubbornly keeps his king on a3, keeping b4 defended, but his bishop will not be able to return to the b5-e8 diagonal under favourable circumstances. The voluntary retreat 56.Ka2
leads to a crucial position after 56...Rc1
Black has completed his regrouping and White will get in successive zugzwang positions. It is typical for such endings that the rook restricts both enemy pieces at the same time. The next step is to push the enemy king on the back rank. 59.Bb5
Apart from king retreats, which is precisely what Black aims to provoke, White has only one move: 60.Bd7
, but this places the bishop on a vulnerable square, allowing the rook to switch to lateral attacks, without letting the king escape from the corner! This will be an important element in several phases of the line actually played in the game Radjabov-Van Wely. 60...Rd8!
Black has made further progress and now the same method as on the previous step ensures him the win. 63.Kb1
White is in zugzwang again. The only move that maintains the pawn defended and avoids mate in one is 65.Bc6
but this loses the bishop to 65...Rc8
) Let us go back to the position after 56.Ba2. 56...Rc1
(Avoiding the trap 56...Ra1??
with complete domination and... draw by stalemate!) 57.Bb3
The king retreats in order to enable lateral attacks of the rook. With the a4-pawn vulnerable now, White cannot save the game. 59.Kb2
(A desperate attempt to return with the bishop on the b5-e8 diagonal. 60.Bb1
leaves the c4-square undefended, allowing a relatively easy win after 60...Kc4
and so on, like in the variation starting with 56.Ka2 above.; 60.Bb3
keeps the c4-square defended, but deprives the own king of the b3-square and places the bishop on a vulnerable square. 60...Kc5!
With the bishop on a favourable square, the white king could escape now to the right wing, but here this is impossible. 62.Kb2
and it is all over.) 60...Re7
with a familiar position.) 61...Kc4
Finally, the bishop has returned on the "good" diagonal, but Black has carried out his essential regrouping in the meanwhile. 63...Rc7
and we have reached a familiar zugzwang position.
I will return now to the aforementioned games of Kasparov. In both cases, the situation was slightly different: the defending part had an advantage of space (here, this would mean that pawns are on a5 and a6). If the king is free, the attacker has to push it one file further than in our current game, namely untill the opposite edge of the board. In Kasparov's game against Pinter, the Hungarian GM managed to avoid this and achieved a draw. Of more interest is the Jussupov-game. Just like in the lines above, the king was trapped in the surroundings of the pawns. Apparently, this should be less dangerous, because the king has more space available, but on the other hand the bishop's "good" diagonal is shorter! Curiously, the evaluation remains the same: the side with a rook should win by successive zugzwang positions. For a while, Kasparov conducted his attack in the best way, but then let the enemy king escape, with inevitable draw...]
[We can notice an important element. Being a long-ranged piece, the bishop, too, can act on both wings. Here, the control of the d3-square is as important as the defence of the pawn. For instance, the careless 52.Bc6?
would lose to 52...Rd2+!
with a double (and deadly) threat.) 53...Kd3
followed by ...Kc3.; However, 52.Be8
is playable, too. After 52...Rd2+
the bishop is safe on e8, because the interposition of the black king on the e-file prevents the double attack ...Re2+.]
Finally, Black switches plans. He will try to push the enemy king as far as possible.
[One important defensive method is the diagonal opposition. In case of the careless 62.Kf4?
Black achieves the situation from point 1) with 62...Re6
followed by the exchange sacrifice on a4 ensures an elementary win.]
[The king has to approach the dangerous 5th rank in order to maintain the diagonal opposition. 77.Kf2?
loses to 77...Rd3
with zugzwang. If the bishop moves, it will be attacked by the rook, followed by a check on the c-file, while 79.Kf1
immediately drives the king into the forbidden zone.]
[Again, the control of the d3-square is essential. After a careless move such as 78.Bc6?
Black establishes a frontal opposition with 78...Kd3
, forcing the king to go upper than he would have wished. 80.Kf5
is met by the familiar manoeuvre 80...Re7
and now Black starts his decisive attack on the opposite wing with 83...Kc5
[Radjabov deals correctly with the last dangerous moment for White in this game. 82.Bd7?
loses to 82...Re3+
and so on.]
and, having convinced himself that Radjabov knows "everything" about this endgame, Van Wely resigned himself to the inevitable. 1/2-1/2