|
< Earlier Kibitzing · PAGE 11 OF 11 ·
Later Kibitzing> |
| Jan-11-12 | | erniecohen: <njchess> See the lines above. Each of those lines ends in what looks like a clearcut draw. Where do you suggest improvement for white? In the alternative, I challenge anyone to play the White side of the position after 36...a6 against me (via correspondence), for any stake up to $100. I take Black and win the bet if I can draw. Any takers? |
 |
Jan-11-12
 | | visayanbraindoctor: The wonder of this game can never be overestimated. Capablanca from a middlegame steers the game into an endgame that to most human and probably to computer (given tournament time constraints) evaluation would initially look downright disadvantageous.
click for larger viewIn this position, what would most white players do? Black being a bit cramped is planning d5, liquidating pieces. Retreat the white Queen? Capa does not. Instead he pushes 23. h4
click for larger viewThis allows 23...d5
click for larger viewThis leads to a series of exchanges, the end position of which Capa has already foreseen as allowing him to push the his h-pawn one more square forward to h5. |
|
Jan-11-12
| | visayanbraindoctor: 24. cxd5 exd5 25. Qxe8+ Qxe8 26. Rxe8+ Kxe8
click for larger viewThere is absolutely no doubt Capa had seen this position before he allowed the liquidating exchanges above, and had accurately seen that 27. h5!! (and also Black's logically defense Rf6) would give him an endgame advantage. 
click for larger viewTartakover replies with the natural 27.. Rf6
click for larger viewwhich not only defends the g6 pawn but also threatens the rook swing Rc6 to the other side of the board and attack White's Queenside pawns. Capa allows him! Just all part of the plan. |
|
Jan-11-12
| | visayanbraindoctor: 28. hxg6 hxg6
click for larger viewThis is a position that Capa must have already seen and evaluated before allowing the above middlegame liquidation. Yet how would most chess players regard it? White has three pawn islands, glaring weaknesses in his c4 and b3 squares, and a weak c3 pawn. If White does nothing, black would simply shift his rook Rc6 attacking the c3 pawn and his knight could then begin jumping all over white's queenside. Most human chess players (and I suspect even computers), would not risk such an endgame if they were given the position on Black's 22nd move and just before 23. h4, the move that initiated all these.
click for larger viewFrom the above to this
click for larger viewis something that Capa must have seen accurately in his mind's eye and evaluated as being advantageous for white. |
|
Jan-11-12
| | visayanbraindoctor: Capa (still all part of the plan) ignores Black Rc6 threat and continues his attack on the kingside, portentously foreseen with his earlier h4 and h5 pawn push. I suspect that Capa had already seen the following ensuing play even in the middlegame in his chess eye, or he would not have goner into this risky looking variation at all. 29. Rh1!
click for larger viewTartakover also plays logically 29.. Kf8 30. Rh7 Rc6
click for larger viewGiven their tendency for materialism, computers given tournament time controls might even evaluate this position as advantageous for black, as black is now threatening to wipe out white's queenside pawns. Capa is unfazed. He proceeds according to his Kingside attack plan. Note that he adheres to the following principles: Capa never lets go of the initiative and at all times maximizes piece activity over pawn structure. He plays this endgame as though he were still in the middlegame. 31. g4! Nc4
click for larger viewTartakover has activated his knight and threatens havoc on the queenside with Ne3+ and Rxc3. Capa does not allow himself to get tied down on a thankless defensive task. Instead instead he goes for the initiative and piece activity, with an end view to directly attack black's King! 32. g5 Ne3+ 33. Kf3 Nf5 34. Bxf5 gxf5 |
|
Jan-11-12
| | visayanbraindoctor: 
click for larger viewNow Capablanca sacs his queenside pawns in going for the logical conclusion that he must have visualized a long time ago, perhaps when he brilliantly allowed the game to be steered into this endgame with 23. h4! He uses his king, rook, and passed pawn to create a direct attack on the black King; and never mind the sacked pawns. Now comes the King march attack. 35. Kg3! Rxc3 36. Kh4 Rf3 37. g6 Rxf4+ 38. Kg5 Re4 39. Kf6 
click for larger viewWhite has three isolated separated pawns; Black has connected pawns in the Queenside and is winning two pawns. Yet white's piece activity is has now created a direct and unstoppable attack on the black king. What did we learn from this game?
Endgames that are unsymmetrical should be played like middlegames- with priority on the initiative and piece activity. Pawn structure is of secondary importance. (Naturally Capablanca also had to do tons of accurate far-seeing calculations before even deciding to go into this endgame in the first place. I think he had already seen the outlines of his final winning King march attack way back in the middlegame; or else a world champion as experienced and strong as he was would never have gone into such a risky line that entails sacrificing loads of pawns in the first place.) |
|
| Jan-11-12 | | erniecohen: <visayanbraindoctor> It's probably best to avoid including so many diagrams from the game - that's what the board applet is for. It makes it pretty hard to follow conversations that were in progress. |
|
Jan-12-12
 | | andrewjsacks: Thank you, King Death; I agree with you wholeheartedly. |
|
| Jan-14-12 | | Tigranny: Why is this Game of the Day twice? |
|
| Jan-14-12 | | AlphaMale: It was adjourned overnight. |
|
| Jan-18-12 | | LIFE Master AJ: http://www.lifemasteraj.com/old_af-... |
|
| Feb-16-12 | | Penguincw: 1-0
 ENDGAMES STATISTICS
  vs.
The superior side (in this case white) wins 46.7% of the time (as in this game). A draw happens 51.3% of the time.
The inferior side (in this case black) 2% of the time. |
|
| Feb-25-12 | | Tigranny: Why not 29...Kd8? |
|
| Feb-25-12 | | King Death: <Tigranny> Tartakower was probably afraid of something like 30.Rh6 Rc6 (30...Nc4 31.Bc4 dc 32.Kf3 looks good for White) 31.Bg6 Rc3 32.f5 when without deep analysis my judgment is that the passed pawns and bishop against knight in this position should be much better. Tartakower's judgment in my opinion was right. |
|
Mar-10-12
 | | kingscrusher: Very instructive game for Rook on the 7th - I did this video about part of this classic game: http://www.youtube.com/watch?v=FaCC... |
|
| Mar-19-12 | | nolanryan: after move 39 when it becomes obvious that black is lost, I suspect Tartakower was repeating to himself, "Any stalemate tricks? Any stalemate tricks? Any stalemate tricks?" Kinda like curtains, aka, Greg Shahade, when totally lost in blitz. |
|
| Apr-28-12 | | Rook e2: 27. h5!
I know at sight what a position contains. What could happen? What is going to happen? You figure it out; I know it! --- Capablanca |
|
May-06-12
 | | Peligroso Patzer: This game is analyzed as Game #28 in the recently-published book, <The Greatest Ever chess endgames> (<sic>), by Giddins, Steve, Everyman Chess ©2012, at pp. 131-136. IM Giddins considers at length two possible improvements for Black, the first of which is <35. … Kg8><!?> (in lieu of Tartakower’s <35. … Rxc3><+>), which idea he attributes to the Russian Master, Vladimir Goldin, as originally analyzed in issue 12/2003 of “64”. One of the lines analyzed by Goldin goes <36. Rd7 Rxc3+ 37. Kh4 Rf3 38. Kh5! Rxf4 39. Kg6 Kf8 40. Kf6 Re4 41. Rf7><+> <Kg8 42. Rxc7 Re8 43. Kxf5 Re4 44. Kf6 Ke5 45. Ke5 Rg4 46. Rxa7 Rxg5><+> <47. Kd6 Rg6><+>, bringing about this position:
click for larger view
Here, Giddins says that Goldin in his analysis in 64, 12/2003 claimed that after the further <48. Kxd5>, the defense with <48. … Kf8> would lead to a draw (op. cit., p. 132). Giddins credits Igor Arkadievich Zaitsev with an analytical breakthrough in finding the “much stronger <48. Kc7><!> which appears to win”, and which was analyzed by Zaitsev in issue 5/2004 of “64”, as cited by Giddins (op. cit., p. 132). The Zaitsev analysis is presented at length by Giddins, but he does not elaborate on the statement that <48. Kxd5 Kf8> would lead to a draw, and it seems to me that Black’s defense in this line would remain extremely difficult after <49. Rc7>, producing this position: 
click for larger view … with the idea of next playing <50. Rc6> to eliminate the last Black pawn. Although Black can capture the White a-pawn (for example, after <49. … Ke8 50. Rc6 Rg5+ 51. Kd6 Ra5 52. Rc8><+> <Kf7 53. Rc4 b5 54. axb5 Rxb5>, the resulting position: 
click for larger view is a tablebase win for White.
Can anyone explain why <48. Kxd5 Kf8> from the position given in the first diagram, above, should lead to a draw? |
|
May-06-12
| | Peligroso Patzer: The second possible improvement for Black considered at length by IM Stephen Giddins is <36. … a6><!>, again originally suggested by Vladimir Goldin. (Op. cit., pp. 132-135.) To summarize briefly this analysis (which runs to three pages in Giddins’s book), if Tartakower had played <36. … a6><!>, correct now for White would have been to continue <37. Kh5> (not <37. g6>, by [flawed] analogy to the actual game continuation) and then after <37. … b5>, White wins with the brilliant <38. Kg6><!!> (instead of the obvious, but weaker, <38. axb5>). (According to IM Giddins, the idea of <38. Kg6><!!> also originated with Igor Arkadievich Zaitsev in the article in “64” cited in my earlier post.) The main continuation after <38. Kg6><!!> goes: <38. … Kg8 39. Rg7><+><Kf8 40. Rf7><+><Kg8 41. Rf6><!>, leading to this position: 
click for larger view
Of this position, IM Giddins writes, “Now we see the point of refraining from the exchange on b5. White will capture the a6-pawn, at the same time threatening mate, and his combined force of king, rook and passed g-pawn will again be too strong, just as in the game continuation.” (Op. cit., p. 135) |
|
| May-06-12 | | Rook e2: What was the top 3 in the greatest ever endgames? |
|
| May-20-12 | | blazerdoodle: In the 3rd paragraph, after 44, the analysis is mixed unreadable. Even if I leave out the two Ke5's, something is missing. |
|
| May-27-12 | | Tigranny: How about 29...Ke7? |
|
| Dec-11-12 | | CharlesSullivan: The posting by <Peligroso Patzer> of May-06-12 is a good summary of why 36...a6 does not save the game for Black. And <erniecohen> is right to insist on Jan-10-12 that 33...Nd1 would have drawn. Letting Houdini 1.5a run for 42 hours to a depth of 40 shows White has nothing: 33...Nd1 34.Rh6 Kg7 35.f5 Nb2 36.Bb1 Nxa4 37.Kf4 b5! For example: 38.Ke5 Nb6 39.Rh2 Nd7+ 40.Kxd5 Rd6+ 41.Ke4 gxf5+ 42.Kxf5 Nf8 43.Ra2 a6 44.Be4 Ng6 45.Ra5 Ne7+ 46.Kg4 Nd5 is very drawish:
click for larger view
Now 47.Bxd5 Rxd5 48.Rxa6 wins a pawn but not the game: both 48...b4 and 48...c5 lead to book draws. |
|
Dec-11-12
 | | Fusilli: The final comment from Reti is priceless:
<Capablanca's management of the endgame gives the impression of being so natural that one easily forgets the difficulty of such precise play. The difficulty is chiefly psychological. In chess, as in life, one is so accustomed to place value on the material factors that it is not easy to conceive the idea of indulging in pawn sacrifices when there is so little available material.> |
|
|
|
< Earlier Kibitzing · PAGE 11 OF 11 ·
Later Kibitzing> |
Annotations by 