< Earlier Kibitzing · PAGE 2 OF 2 ·
Later Kibitzing> 
Jan2104   Whitehat1963: Where? 

Jan2104   ughaibu: Move 34. "white to play and win". 

Jan2104   Whitehat1963: 34. Bxf7, right?? 

Jan2104   Whitehat1963: Whoops, didn't see the pin on the queen. So what's the answer? 

Jan2104
  technical draw: If 34.Bxf7+ Bxf7, 35.Re5! 

Jan2104   Whitehat1963: Great! I didn't see 35. Re5! Thanks. 

May0804
  acirce: "Rubinstein is a rook ending of a chess game that was started by God a thousand years ago."  Tartakower, quoted i Dvoretsky's Endgame Manual. 

May2504   Hinchliffe: <ughaibu> Did you spot the missed win yourself? 

May2504   Everett: Technical Draw
What if 35.... g6? What's white's next? 

May2604   Everett: Ahhh, now I see it, completely didn't see the hanging rook on e2. Nice find ughaibu. 

May2604   Hinchliffe: < Technical draw> As always a pleasure to read your comments. But what if black refutes and plays K  h8 ? I know his position looks dire but I was interested to see what the continuation could be and I know you'll have done it. Once again always enjoy your comments and suggestions. 

May2604   ughaibu: Hinchliffe: The first time I saw this game was in Lasker's book of the tournament, he pointed out 34.Bf7 

May2604   drukenknight: are you sure there's a win there? Here's one try w/ the crappy computer, have no idea if I started black out right or not.... 34. Bxf7+ Kh8
35. Bxg6 hxg6
36. b4 Rxf2+
37. Rg2 Rxg2+
38. Kxg2 Qxd4
39. Qxg6 Qd2+
40. Kg3 Qe3+
41. Kh2 Rb6
42. Qg3 Qd2+
43. Kh1 Qd1+
44. Kh2 

Oct2705
  sisyphus: <"Rubinstein is a rook ending of a chess game that was started by God a thousand years ago."> This is the rook ending analyzed by Danny Kopec on http://www.chess.fm, starting with 41.... Ra8. 

Dec2206   who: <drukenknight> 40.Kh1 is more accurate, but in your line 42.Rh5+ Kg8 43.Qc2 and the checks are almost over. 

Jan1007   Karpova: This game is simply stunning!
First, Spielmann plays very forcefully and Rubinstein commits an error with 27...b4 (Reinfeld suggests 27...Qd6) but Spielmann misses 34.Bxf7+ as Lasker pointed out.
And then, after the 40th move Rubinstein plays the rook ending perfectly. He attacks the weak white pawns first and then moves his king forward. He rejects to win the d4pawn at once but calmly improves his kingside position.
47...Kc4 (and if 48.Rb7 Rxa3+ 49.Ke4 d5+ 50.Kf5 Rxh3).
Regrouping of the rook to the efile is also nice and the d4pawn is taken afterwards. Too many pawn islands and a passive rook  there's no better example showing the disadvantages. 

Jun2410
  GrahamClayton: Analysis of this endgame after White's 56th move can be found at: http://www.chessending.com/psitn039... 

Feb0712   Chessmaster9001: One of the most instructive endgames ever. Essential Rubinstein! 

May0212
  Peligroso Patzer: In the position after <44. … Kd5>: click for larger view
Spielmann set a clever trap with <45. Ke2!?> (instead of the obvious <45. Ke3>, routinely defending the dpawn). Although there was little or no chance that Rubinstein would fall for the trap, if Black now goes for the pawn grab (<45. … Rxd4?>), White’s reply would be <46. Ke3!>, forcing the exchange of Rooks. In the resulting pawn ending, although White has one pawn less, his outside passer on the afile means it is White that is playing for a win. Spielmann apparently learned well the painful lesson from this game: Spielmann vs Duras, 1907, in which a similar trick forced a pawn ending that (in the earlier game) was losing for Spielmann. 

May0412
  Peligroso Patzer: <Jan2104 ughaibu: Move 34. "[W]hite to play and win".> <Jan2104 technical draw: If 34.Bxf7+ Bxf7, 35.Re5!> <May2504 Hinchliffe: <ughaibu> Did you spot the missed win yourself? > <May2604 ughaibu: Hinchliffe: The first time I saw this game was in Lasker's book of the tournament, he pointed out 34.Bf7 > <May2604 drukenknight: are you sure there's a win there? *** > Lasker’s comment in the tournament book reads as follows: “With this move [<34. Qxd6>] White loses the fruit of the efforts he has made. The right move was <34. Bxf7><+>. If then <34. … Bxf7 35. Re5> makes the queen mobile and White mates or wins the rook. And if <34. … Kxf7 35. Rxc7<+> Ke6 36. d5<+> Qxd5 37. Qg4<+>> winning easily.” <The International Chess Congress: St. Petersburg 1909>, by LASKER, Emanuel, Russell Enterprises, Inc. ©2008, at pp. 159160.> [<Source note>: The title page and colophon of this edition omit to identify the translation; the Foreword (ibid., at p. 5) identifies it as one primarily done by Richard Teichmann originally published in New York in 1910, with the moves converted from descriptive to algebraic notation.] To return to the World Champion’s analysis, it does not consider Black’s best defense (after <34. Bxf7><+>), which would have been <34. … Kh8> (as previously proposed by other kibitzers). White then would obtain a clear advantage and excellent winning chances with the simple <35. Bxg6>, creating an additional weakness in the already difficult Black position; but it seems to me it could not be conclusively stated that the position resulting from this continuation is outright winning for White. 

Aug1212   Karpova: Another version of Dr. Emanuel Lasker's annotation on move 43: <The manner in which Rubinstein treats the following endgame is beyond all praise.> From the St. Petersburg 1909 tournament book, translated by Richard Teichmann. 

Aug1712
  Naniwazu: If instead of 43. Kg3 White plays 43. d5 as Lasker suggests Black can play 43...g5 as mentioned by Dvoretsky in his 'Endgame Manual', 44. Kg2 Kf6 45. Rf3+ Kg6 (Δ Rd4) 46. Rd3 f6 Δ ...Kf5. 

Oct1012   vinidivici: After the match the stunnedSpielmann shouted "Akiba, if u lived in the Middle Ages you would have been burned at the stake, what u do in rook endgames can only be called witchcraft!" 

Jun1816
  AylerKupp: Spielmann vs Rubinstein, 1909 (part 1 of 2) <<Peligroso Patzer> Spielmann set a clever trap with <45. Ke2!?> (instead of the obvious <45. Ke3>, routinely defending the dpawn). Although there was little or no chance that Rubinstein would fall for the trap, if Black now goes for the pawn grab (<45. … Rxd4?>), White’s reply would be <46. Ke3!>, forcing the exchange of Rooks. In the resulting pawn ending, although White has one pawn less, his outside passer on the afile means it is White that is playing for a win. From the position after 45.Ke2:
click for larger viewIf now 45...Rxd4 46.Ke3 Rxd3 47.Kxd3 we reach this position:
click for larger viewThe FinalGen tablebase generator came up with the following results after each of Black's possible moves: 47...g5 Black wins or Draw
47...f6 White wins or Draw
47...Kc5 White wins or Draw
47...Kc6 White wins in 30
47...Ke5 White wins in 30
47...Ke6 White wins in 28
47...f5 White wins in 18
So, in a practical sense, you are correct; White has the better winning chances since Black has more opportunities to go wrong. But, in the theoretical sense, after 47...g5 it is Black that has the winning chances since, with best play, White can only draw. And, because the results after 47...g5, 47...f6, and 47...Kc5 are ambiguous, FinalGen does not provide explicit lines so I can't tell you in any of these lines the possibilities that White wins, Black wins, or the game is drawn. And here, for example, are what FinalGen considers the best line for both sides following 47...Kc6: 48.Kc4 g5 49.Kd4 f6 50.Kc4 f5 51.f3 d5+ 52.Kd4 Kd6 53.a4 Kc6 54.a5 f4 55.a6 Kb6 56.Kxd5 Kxa6 57.Ke4 Kb5 58.Kf5 Kc6 59.Kxg5 Kd6 60.Kxf4 Ke7 61.Kg5 Ke8 62.Kg6 Kf8 63.h4 Kg8 64.h5 Kf8 65.h6 Kg8 66.Kh5 Kh7 67.hxg7 Kxg7 68.Kg5 Kf7 69.Kf5 Ke7 70.Kg6 Kd8 71.Kf7 Kc7 72.Ke7 Kb8 73.f4 Kc7 74.f5 Kb8 75.f6 Kc7 76.f7 Kb877.f8Q+ And here FinalGen ends its analysis since it considers the position won for White. 

Jun1816
  AylerKupp: Spielmann vs Rubinstein, 1909 (part 2 of 2) In contrast, here is how two of today's best engines assess the position after 47.Kxd3 using 5piece Syzygy tablebases: Komodo 10, d=49:
1. [0.00]: 47...g5 48.a4 Kc5 49.a5 f6 50.a6 Kb6 51.Kd4 Kxa6 52.Kd5 Kb5 53.Kxd6 Kc4 54.Ke6 Kd3 55.Kf7 f5 56.Kxg7 g4 57.hxg4 2. [0.00]: 47...f6 48.f4 Kc5 49.h4 Kd5 50.a4 Kc5 51.Kc3 f5 52.Kb3 Kd4 53.Kb4 d5 54.a5 Ke3 55.a6 d4 56.a7 d3 57.a8Q d2 58.Qd5 Ke2 59.Qb5+ Ke1 60.Qe5+ Kf1 61.Qa1+ Ke2 62.Qa6+ Ke1 63.Qxg6 d1Q 64.Qxf5 Qd4+ 65.Kb5 Kf2 66.Qc5 Qxc5+ 67.Kxc5 3. [0.00]: 47...Kc5 48.f4 Kd5 49.a4 Kc5 50.Kc3 f6 51.Kb3 f5 52.h4 Kd4 53.Kb4 d5 54.a5 Ke3 55.a6 d4 56.a7 d3 57.a8Q d2 58.Qd5 Ke2 59.Qb5+ Ke1 60.Qe5+ Kf1 61.Qd5 Ke1 62.Qh1+ Ke2 63.Qg2+ Ke1 64.Qxg6 d1Q 65.Qxf5 Qd4+ 66.Kb5 Kf2 67.Qe5 Qd2 68.Ka6 Kf3 69.Qh5+ Kf2 70.Qg5 Kf3 71.Kb6 Qxf4 72.Qxg7 Stockfish 7, d=55:
1. [0.00]: 47...g5 48.a4 f5 49.a5 f4 50.a6 Kc6 51.Kc4 Kb6 52.Kd5 Kxa6 53.Kxd6 Kb5 54.Ke5 Kc4 55.Kf5 Kd3 56.Kxg5 Ke2 57.Kxf4 2. [0.00]: 47...f6 48.f4 f5 49.h4 Kc5 50.Kc3 Kd5 51.a4 Kc5 52.Kb3 Kd4 53.a5 Kc5 54.Ka4 g5 55.hxg5 g6 56.a6 Kb6 57.Kb4 Kxa6 58.Kc4 Kb6 59.Kd5 Kc7 60.Ke6 Kc6 61.Kf6 d5 62.Kxg6 d4 63.Kxf5 3. [0.00]: 47...Kc5 48.f4 f6 49.Kc3 Kd5 50.h4 Kc5 51.Kd3 Kd5 52.a4 Kc5 53.Kc3 f5 54.Kb3 Kd4 55.a5 Kc5 56.Ka4 g5 57.hxg5 g6 58.a6 Kb6 59.Kb4 Kxa6 60.Kc4 Kb6 61.Kd5 Kc7 62.Ke6 Kc6 63.Kf6 d5 64.Kxg6 d4 65.Kxf5 So FinalGen, Komodo 10, and Stockfish 7 all agree; 47...g5, 47...f6, and 47...Kc5 are Black's 3 top moves but, after that, neither Komodo nor Stockfish could find winning lines for either side. Earlier, at d=26 for Komodo and d=35 for Stockfish, the two engines evaluated the position as wins for White after both 47...f6 and 47...Kc5 (evaluations higher than [+4.00]) but they reverted back to [0.00] evaluations in a few more search plies. Of course, the chances that either player would have been able to find the best moves throughout is unlikely, hence the possible results remain theoretical. But it was an interesting exercise. 

< Earlier Kibitzing · PAGE 2 OF 2 ·
Later Kibitzing> 