|Feb-29-08|| ||Knight13: I bet this game's in the book "Basic Chess Endings" by Reuben Fine.|
|Jul-17-08|| ||keypusher: Part I
Played in Munich on September 28-29
<Translated from Tarrasch's book on the match (http://www.google.com/books?id=0CgC...) with additional commentary from Shredder and me. Tarrasch's comments are in plain text; all other comments are in brackets. German that I was particularly unsure about (or particularly fond of) is also in brackets, right after the translation. As always, corrections are welcome.>
1. d4 d5 2. Nf3 c5
This move, with which Black plays the queen's gambit a move behind, has long been known to be the best response to the queen's pawn game.
The gambit, which can be advantageously offered by the first player, is obviously even better for the second player, as the maximum advantage of the first move may be transferred to Black. <Das Gambit anzunehmen, kann mann sich im Anzuge naturlich weit eher erlauben als im Nachzuge, da man damit hochstens den Vorteil des Anzuges an Schwarz abtritt.>
Even in the ordinary queen's gambit it is best to play Nf3, so as to stop the double advance of the king's pawn; in the gambit a move behind ...Nf6 is almost forced, because, as is soon evident, after ...e6 there arises after e2-e4 a variation of the French Defense, namely 1. e4 e6 2. d4 d5 3. exd5 exd5 4. Nf3 c5? 5. dxc5 Bxc5, in which ...c7-c5 instead of ...Nf6 is at the least premature. Therefore Black should have steered clear of this variation of the French Defense, and for this purpose he should have played ...Nf6 rather than ...e6.
4. e4 Bxc5 5. exd5 exd5 6. Bb5+ Nc6 7. 0-0 Nf6!
Much better than ...Ne7, when the knight would stand badly; now on Re1+, ...Ne4 would follow.
8. Nc3 0-0 9. Bg5 Be7
Instead of this, ...Be6 must first be played; Black underestimates the consequences of the following combination and so loses a pawn.
Lasker says that it perhaps would have been better to forego the win of a pawn, and rather strengthen his position with Re1. I think, however, that the win of a pawn must bring him the advantage. It makes an enormous difference whether the pawn was sacrificed or lost! <Es macht doch einen gewaltigen Unterschied, ob der Gegner einen Bauern opfert oder verliert!>
10....Bxf6 11. Qxd5
If Black now plays 11....Qxd5 12. Nxd5 Bxb2, as I originally intended, he loses a piece to the surprising combination 13. Rb1 Ba3 14. Nc7 Rb8 15. Bxc6! A terrible disaster! <Ein boser Reinfall!>
click for larger view
Now the best chance for Black; there also came into consideration ...Be6 or ...Qb6, with the idea of compensating for the lost pawn with the bishop-pair, but this would not have proven satisfactory.
12. Qxd8 Rxd8 13. bxc3 Na5
Of course the knight stands badly here initially <zunachst>, but Black prevents White from taking the knight, closing the c-file and making a frontal attack on the doubled pawns difficult to carry out. <Interestingly, 13....Na5 is Shredder's immediate preference.>
14. Rfe1 Bd7
If Black prevents the rook penetrating to e7 with ...Kf8, Re5 with the subsequent doubling of White's rooks could follow. With the text move Black wins a tempo.
15. Bd3 <Here Shredder likes the simple 15. Bxd7 Rxd7 16. Rad1 Rxd1 (16....Rad8?? 17. Rxd7) 17. Rxd1 Re8, but I am not sure it is any better than the game.>
Black must at first stay on the defensive and oppose the strength of the white rook on the e-file. Nothing good comes of the tempting attacking move ...Rac8. White can play Re3 (...Rxc3?, Bxh7+) or Rad1 (...Rxc3?, Bf5!), and Black cannot take the pawn on c3. <Shredder nevertheless prefers ...Rac8, as it thinks the game continuation is bad for Black. Apparently, though, Tarrasch assesses the position arising after Black's 18th move more accurately than the machine does.>
|Jul-17-08|| ||keypusher: Part II
The double exchange of rooks (Rxe8+ and Re1) would hardly improve White's chances. <Mit doppeltem Turmtausch (Txe8+ nebst Te1) wurde Weiss seine Chancen schwerlich verbessern.>
Black still cannot attack on the c-file, because after 16....Rac8, then 17. Bf5 Bxf5 18. Nxf5 could follow, and the Black rooks are chained to the back rank. <Schwarz kann immer noch nicht auf der c-Linie angreifen, denn auf 16....Tac8 kann 17. Lf5 Lxf5 18. Sxf5 folgen und die schwarzen Turme sind an die letzte Reihe gefesselt.>
Here there is no easy response for Black. The retreat of the knight to c6 would be very weak, because White would then with 18. Nc5 Bc8 19. Be4 completely paralyze the black game. To take the knight on b3 hardly came into consideration, as it would give white a real pawn advantage, which would barely count at first, but in all likelihood its aggressive strength would unfold <aber aller Warscheinlichkeit nach keine Aggressivkraft entfalten wird>. So only the move ...b6 was left, which at first looked awful, but on further consideration turned out to be quite satisfactory.
click for larger view
17....b6(!) <A gutty move by Tarrasch, who would lose the match on one more defeat, and must have been having unpleasant memories of game 7 around here.>
This seems natural and seductive enough, but it seems to be not very strong. Lasker thought that he should have played c3-c4. In fact, this move seems most like to lead to a secure advantage for White, i.e. 18. c4 Rac8 19. Rxe8+ B or Kxe8 20. c5 Nxb3 21. axb3 bxc5 22. Rxa7; or 18. c4 Nb7 19. Be4 <Shredder finds a useful improvement here: 18. c4 Rxe1+ 19. Rxe1 Nb7, and Black should be able to hold>; or 18. c4 Rec8 19. Nxa5 bxa5 20. Re5 as in the game, but with an extra move.
18....bxa5 19. Kf1 Rac8
Finally there comes the long-planned attack on the c-file, which in a short time brings about a drawn game.
20. c4 Be6 21. Re5
Apart from this move, the defense of the pawn with Re4 also came into question, but Black would have responded with ...Rc5 and the doubling of rooks on the c-file, and been at least no worse off than in the game continuation.
21....Bxc4 22. Rxa5 Bxd3+ 23. cxd3
Now Black has succeeded, via multiple exchanges, in forcing a rook ending, where an extra pawn is often not sufficient to win.
Admittedly, it depends on the pawn; a middle pawn, for example, plays a very different role than a rook pawn! Therefore, Black immediately attacks the really dangerous pawn.
click for larger view
A mistake was ...Rc2 with the intention of doubling rooks on the second rank; White would take the a-pawn and on ...Ree2 would by the enchanting twist <fesselnde Wendung> Ra8+, ...Ke7, Re1 thwart Black's plan.
After this White remains with only an a-pawn and it is no longer difficult for Black to achieve the draw. But attempts to defend the important d-pawn lead to no better result, e.g. Rd1, ...Re7 with the threat to achieve "horizontal doubling" <Querverdopplung> with ...Rc2; or even Rd5, ...Rc2 with the same idea; or finally d3-d4, ...Rd3, d4-d5, ...Rd8.
24....Rxd3 25. Re1
The rook exchange is the last thing White needs <Den Turmtausch brauchte Weiss nicht gerade anzustreben>, for now the game is a sure draw. But after g2-g3 (to secure the king and so make the queen rook mobile), there follows ...Rd2 and horizontal doubling.
|Jul-17-08|| ||keypusher: Part III
Black may not exchange rooks himself, as then his other rook would not reach the second rank, which is so important for him.
26. Rxe8+ Kxe8
click for larger view
Now a typical position has arisen; White has on one flank an extra passed pawn, the attacking rook stands badly (much better would be a1) and the defending rook stands very well. In such positions the rook pawn wins only in exceptional cases, as I will show in the appendix <the book included a 23-page appendix on endings of this type>.
27. Ke1 Rb2 28. a4 g6 29. a5 Ra2 30. a6 Kf8 31. Ra8+
With this move the game was adjourned. The sealed move is not good, as the rook would have been better remaining on the 7th rank, so that, if the black rook took one of the three kingside pawns, with Rc7, ...Ra2, a6-a7 White could achieve a change of front that would be very dangerous for Black, since the a-pawn has arrived at the 7th rank and the white rook is freed from the guard post <Zwangstellung> on the a-file. Now that the white rook has abandoned the 7th rank, in the analogous case he can only "change his front" only by playing Rb8 and Rb6, so the passed pawn remains on the 6th rank. <Nachdem der Weisse Turme aber jetzt die siebente Reihe verlassen hat, kann er im analogen Falle seine Frontveranderung nur derart vollziehen, dass er also der Freibauer noch auf der sechsten Reihe verbleibt.>
31....Kg7 32. g4
Or 32. h4 with a similar result.
White threatened to paralyze the pawns with g4-g5 and thus to win, as will be shown in the Appendix. Although it's true that a pawn on g5, or rather, its guard cannot be held, Black must play, not to win a pawn, but rather to obtain a passed pawn, either to tie down the White king, or (with the support of its own king) to make it so dangerous that the black rook can, in an emergency, be sacrificed for the a-pawn. <Zwar ware der Bauer g5 bzw. seine Deckung nicht zu halten, allein Schwarz muss in dieser Situation nicht darauf spielen, einen Bauern zu erobern, sondern vielmehr darauf, sich einem Freibauern zu machen, den der weissen Konig fesselt oder, vom schwarzen Konig unterstutzt, so gefahrlich wird, dass der schwarzen Turm sich im Notfalle gegen den a-Bauern opfern kann.> <I was baffled by the first clause of that sentence -- whiteshark, nescio, can you help?>
33. h4 h5 34. gxh5
If White defends with f2-f3, the best continuation is 34....Ra4, and Black gets a passed pawn.
Now Black can win the h-pawn and make his own h-pawn passed. The draw is already clear.
In general it is better not to advance the pawn, leaving it at a6 so that the a7 square may serve as a landing place for the White king to escape from rook checks. But in this position it would not give White any greater winning chances.
35....Kg7 36. Kd1 Kh7 37. Kc1 Kg7 38. Kb1 Ra4 39. Kc2 Ra2+ 40. Kc3 Ra3+ 41. Kc4 Ra4+ 42. Kb5 Ra1 43. f4
If the white pawn was still on a6, the white king could get to b6 and a7, whereupon Rb8, Rb6, Kb7, a6-a7-a8/Q would win.
43....Rb1+ 44. Kc5 Ra1
Black offered a draw here, which White rejected.
click for larger view
45. Kd6 Kf6 46. Kd7
White cannot play Rh8 because of ...Ra6+; if the king then goes to the seventh rank, then Black takes the a-pawn with check; if instead the king goes to the fifth rank, then ...Rxa7, and the white rook cannot take the h-pawn, as it is then lost to a rook check.
With this move, with which the game was adjourned for a second time, Black plays for the win, as far as is possible. But with 47. Ke7 f6 48. Kf7 Ra6 49. Kg7 Kxf4 50. Kh6 Ra5 51. Rg8 Rxa7 52. Kxh5 Rh7+ 53. Kg7 Rxh4, White can still achieve the draw.
47. Kc6 Kxf4 48. Kb6 Rb1+ 49. Kc6 Rc1+ 50. Kb6 Rb1+ 51. Kc6 Ra1 52. Kb6 Drawn.
This game is of importance for opening study <Eroffnungslehre>, has an interesting middlegame, and is the latest word in the theory of rook endings. As such, it suggested to me the analytical study that follows in the appendix.
|Jan-27-11|| ||Llawdogg: By move 26, Tarrasch had forced the kind of rook endgame that he knew he could draw.|
|Dec-14-18|| ||RookFile: I guess, but the opening play was risky by Tarrasch. He should have felt fortunate to get a draw in this one.|