Ehrhardt Post vs Nimzowitsch 1905, draw or not ? Crazy rook

[Vinvin]

Here Ehrhardt Post played Rxf3 !? (or ?!), shallow analyze computer show a very high score but no win because zillions of checks ... what's the result with an accurate play ?
[d]8/1R6/8/6K1/p3P2r/k4P2/3R4/5r2 b - - 3 60


[pgn]
[Event "Barmen Meisterturnier B"]
[Site "Barmen DEU"]
[Date "1905.08.22"]
[EventDate "1905.08.14"]
[Round "8"]
[Result "1/2-1/2"]
[White "Alfred Ehrhardt Post"]
[Black "Aron Nimzowitsch"]
[ECO "D07"]
[WhiteElo "?"]
[BlackElo "?"]
[PlyCount "196"]

1. d4 d5 2. Nf3 Bg4 3. c4 Nc6 4. cxd5 Bxf3 5. gxf3 Qxd5 6. e3
e5 7. Nc3 Bb4 8. Bd2 Qd7 9. d5 Nce7 10. Qb3 Bxc3 11. Bxc3 Qxd5
12. Qxd5 Nxd5 13. Bxe5 f6 14. Bd4 Nge7 15. Rg1 O-O 16. Bc4 Kh8
17. Bc5 Rfe8 18. O-O-O Nb6 19. Bf7 Rf8 20. Bh5 g6 21. Bxe7 Rf7
22. Bxg6 hxg6 23. Bb4 Kg7 24. Rd4 c5 25. Bxc5 Rc8 26. b4 Nd7
27. Kb1 Nxc5 28. bxc5 Rxc5 29. Rdg4 Rb5+ 30. Ka1 g5 31. h4 Rc7
32. hxg5 Rc2 33. gxf6+ Kf7 34. Rb1 Ra5 35. Rxb7+ Kxf6 36. Rf4+
Ke6 37. a4 Rxf2 38. Rb2 Rf1+ 39. Ka2 Re5 40. Re2 Kd5 41. Kb2
Kc5 42. Kc3 Rc1+ 43. Kd2 Ra1 44. Rg4 Rf5 45. Rf2 Re5 46. e4
Kd4 47. Rg7 Ra2+ 48. Ke1 Rxa4 49. Re2 Rh5 50. Rd7+ Kc4
51. Rc2+ Kb3 52. Rcd2 Ra1+ 53. Ke2 a5 54. Rb7+ Kc3 55. Rc7+
Kb3 56. Kd3 Rf1 57. Rb7+ Ka3 58. Ke3 a4 59. Kf4 Rh4+ 60. Kg5
Rxf3 61. Kxh4 Rf4+ 62. Kg3 Rf3+ 63. Kg4 Rf4+ 64. Kg5 Rf5+
65. Kg6 Rf6+ 66. Kg7 Rg6+ 67. Kh8 Rg8+ 68. Kh7 Rh8+ 69. Kg6
Rh6+ 70. Kf5 Rf6+ 71. Kg5 Rf5+ 72. Kg6 Rf6+ 73. Kg7 Rg6+
74. Kh8 Rg8+ 75. Kh7 Rh8+ 76. Kg6 Rh6+ 77. Kg5 Rg6+ 78. Kf4
Rg4+ 79. Kf3 Rf4+ 80. Ke2 Rxe4+ 81. Kd1 Re1+ 82. Kc2 Rc1+
83. Kd3 Rc3+ 84. Kd4 Rc4+ 85. Kd5 Rc5+ 86. Kd6 Rc6+ 87. Kd7
Rb6 88. Rc7 Rb2 89. Rd4 Rb8 90. Rcc4 Ra8 91. Rb4 Ra6 92. Ke7
Ra7+ 93. Ke6 Ra6+ 94. Ke5 Ra5+ 95. Ke4 Ra6 96. Kd3 Ra8 97. Kc2
Rc8+ 98. Kb1 Rb8 1/2-1/2[/pgn]

[Vinvin]

Ok, after 61.Kxh4 there's 7 pieces and white seems to be forced to take the e4 pawn, and after that the 6 pieces is winning ...

[kgburcham]

[D] 8/1R6/8/6K1/p3P2r/k4P2/3R4/5r2 b - -




6 processor(s) found, POPCNT available
NUMA configuration with 1 node(s), offset 0
4096 MB Large Page Hash
Nalimov 6 men EGTB available - 999 MB cache
Engine: Houdini 4 Pro x64 (4096 MB)
by Robert Houdart
23/62 0:01 0.00 60...Rxf3 61.Kxh4 Rf4+ 62.Kg5 Rf5+ 63.Kg6 Rf6+ 64.Kg7 Rg6+ 65.Kh8 Rg8+ 66.Kh7 Rh8+ 67.Kg6 Rh6+ 68.Kg5 Rg6+ (18.716.646) 15430 TB:23.925
24/62 0:01 0.00 60...Rxf3 61.Kxh4 Rf4+ 62.Kg5 Rf5+ 63.Kg6 Rf6+ 64.Kg7 Rg6+ 65.Kh8 Rg8+ 66.Kh7 Rh8+ 67.Kg6 Rh6+ 68.Kg5 Rg6+ (22.565.064) 15768 TB:30.298
25/64 0:01 0.00 60...Rxf3 61.Kxh4 Rf4+ 62.Kg5 Rf5+ 63.Kg6 Rf6+ 64.Kg7 Rg6+ 65.Kh8 Rg8+ 66.Kh7 Rh8+ 67.Kg6 Rh6+ 68.Kg5 Rg6+ (28.482.135) 15797 TB:43.120
26/64 0:02 0.00 60...Rxf3 61.Kxh4 Rf4+ 62.Kg5 Rf5+ 63.Kg6 Rf6+ 64.Kg7 Rg6+ 65.Kh8 Rg8+ 66.Kh7 Rh8+ 67.Kg6 Rh6+ 68.Kg5 Rg6+ (37.030.886) 16313 TB:57.100
27/65 0:03 0.00 60...Rxf3 61.Kxh4 Rf4+ 62.Kg5 Rf5+ 63.Kg6 Rf6+ 64.Kg7 Rg6+ 65.Kh8 Rg8+ 66.Kh7 Rh8+ 67.Kg6 Rh6+ 68.Kg5 Rg6+ (52.867.109) 17026 TB:81.248
28/66 0:04 0.00 60...Rxf3 61.Kxh4 Rf4+ 62.Kg5 Rf5+ 63.Kg6 Rf6+ 64.Kg7 Rg6+ 65.Kh8 Rg8+ 66.Kh7 Rh8+ 67.Kg6 Rh6+ 68.Kg5 Rg6+ (81.551.083) 17840 TB:123.471
29/66 0:05 0.00 60...Rxf3 61.Kxh4 Rf4+ 62.Kg5 Rf5+ 63.Kg6 Rf6+ 64.Kg7 Rg6+ 65.Kh8 Rg8+ 66.Kh7 Rh8+ 67.Kg6 Rh6+ 68.Kg5 Rg6+ (106.540.153) 18100 TB:171.620
30/66 0:08 0.00 60...Rxf3 61.Kxh4 Rf4+ 62.Kg5 Rf5+ 63.Kg6 Rf6+ 64.Kg7 Rg6+ 65.Kh8 Rg8+ 66.Kh7 Rh8+ 67.Kg6 Rh6+ 68.Kg5 Rg6+ (151.807.317) 18583 TB:266.903
31/68 0:11 0.00 60...Rxf3 61.Kxh4 Rf4+ 62.Kg5 Rf5+ 63.Kg6 Rf6+ 64.Kg7 Rg6+ 65.Kh8 Rg8+ 66.Kh7 Rh8+ 67.Kg6 Rh6+ 68.Kg5 Rg6+ (209.394.418) 18893 TB:369.265
32/68 0:14 0.00 60...Rxf3 61.Kxh4 Rf4+ 62.Kg5 Rf5+ 63.Kg6 Rf6+ (285.575.147) 19154 TB:528.861
33/68 0:21 0.00 60...Rxf3 61.Kxh4 Rf4+ 62.Kg5 Rf5+ 63.Kg6 Rf6+ 64.Kg7 Rg6+ 65.Kh8 Rg8+ 66.Kh7 Rh8+ 67.Kg6 Rh6+ 68.Kg5 Rg6+ (399.312.189) 18931 TB:740.046
34/68 0:28 0.00 60...Rxf3 61.Kxh4 Rf4+ 62.Kg5 Rf5+ 63.Kg6 Rf6+ 64.Kg7 Rg6+ 65.Kh8 Rg8+ 66.Kh7 Rh8+ 67.Kg6 Rh6+ 68.Kg5 Rg6+ (547.531.566) 19267 TB:1.040.767
35/68 0:39 0.00 60...Rxf3 61.Kxh4 Rf4+ 62.Kg5 Rf5+ 63.Kg6 Rf6+ 64.Kg7 Rg6+ 65.Kh8 Rg8+ 66.Kh7 Rh8+ 67.Kg6 Rh6+ 68.Kg5 Rg6+ (773.457.564) 19418 TB:1.518.220
36/70 1:01 0.00 60...Rxf3 61.Kxh4 Rf4+ 62.Kg5 Rf5+ 63.Kg6 Rf6+ 64.Kg7 Rg6+ 65.Kh8 Rg8+ 66.Kh7 Rh8+ 67.Kg6 Rh6+ 68.Kg5 Rg6+ (1.222.053.501) 19936 TB:3.087.038
37/70 1:20 0.00 60...Rxf3 61.Kxh4 Rf4+ 62.Kg5 Rf5+ 63.Kg6 Rf6+ 64.Kg7 Rg6+ 65.Kh8 Rg8+ 66.Kh7 Rh8+ 67.Kg6 Rh6+ 68.Kg5 Rg6+ (1.627.005.320) 20318 TB:4.780.377
38/70 1:40 0.00 60...Rxf3 61.Kxh4 Rf4+ 62.Kg5 Rf5+ 63.Kg6 Rf6+ 64.Kg7 Rg6+ 65.Kh8 Rg8+ 66.Kh7 Rh8+ 67.Kg6 Rh6+ 68.Kg5 Rg6+ (2.034.269.143) 20179 TB:5.609.714
39/70 2:25 0.00 60...Rxf3 61.Kxh4 Rf4+ 62.Kg5 Rf5+ 63.Kg6 Rf6+ 64.Kg7 Rg6+ 65.Kh8 Rg8+ 66.Kh7 Rh8+ 67.Kg6 Rh6+ 68.Kg5 Rg6+ (3.002.860.044) 20598 TB:7.519.540
40/76 3:02 0.00 60...Rxf3 61.Kxh4 Rf4+ 62.Kg5 Rf5+ 63.Kg6 Rf6+ 64.Kg7 Rg6+ 65.Kh8 Rg8+ 66.Kh7 Rh8+ 67.Kg6 Rh6+ 68.Kg5 Rg6+ (3.770.358.571) 20637 TB:9.500.385
41/78 5:01 0.00 60...Rxf3 61.Kxh4 Rf4+ 62.Kg5 Rf5+ 63.Kg6 Rf6+ 64.Kg7 Rg6+ 65.Kh8 Rg8+ 66.Kh7 Rh8+ 67.Kg6 Rh6+ 68.Kg5 Rg6+ (6.321.562.070) 20994 TB:14.114.672
42/78 5:49 0.00 60...Rxf3 61.Kxh4 Rf4+ 62.Kg5 Rf5+ 63.Kg6 Rf6+ 64.Kg7 Rg6+ 65.Kh8 Rg8+ 66.Kh7 Rh8+ 67.Kg6 Rh6+ 68.Kg5 Rg6+ (7.339.209.801) 20993 TB:16.367.535


[Event "Barmen Meisterturnier B"]
[Site "Barmen DEU"]
[Date "1905.08.22"]
[Round "8"]
[White "Alfred Ehrhardt Post"]
[Black "Aron Nimzowitsch"]
[ECO "D07"]
[Result "1/2-1/2"]

1. d4 d5 2. Nf3 Bg4 3. c4 Nc6 4. cxd5 Bxf3 5. gxf3 Qxd5
6. e3 e5 7. Nc3 Bb4 8. Bd2 Qd7 9. d5 Nce7 10. Qb3 Bxc3
11. Bxc3 Qxd5 12. Qxd5 Nxd5 13. Bxe5 f6 14. Bd4 Nge7
15. Rg1 O-O 16. Bc4 Kh8 17. Bc5 Rfe8 18. O-O-O Nb6 19. Bf7
Rf8 20. Bh5 g6 21. Bxe7 Rf7 22. Bxg6 hxg6 23. Bb4 Kg7
24. Rd4 c5 25. Bxc5 Rc8 26. b4 Nd7 27. Kb1 Nxc5 28. bxc5
Rxc5 29. Rdg4 Rb5+ 30. Ka1 g5 31. h4 Rc7 32. hxg5 Rc2
33. gxf6+ Kf7 34. Rb1 Ra5 35. Rxb7+ Kxf6 36. Rf4+ Ke6
37. a4 Rxf2 38. Rb2 Rf1+ 39. Ka2 Re5 40. Re2 Kd5 41. Kb2
Kc5 42. Kc3 Rc1+ 43. Kd2 Ra1 44. Rg4 Rf5 45. Rf2 Re5 46. e4
Kd4 47. Rg7 Ra2+ 48. Ke1 Rxa4 49. Re2 Rh5 50. Rd7+ Kc4
51. Rc2+ Kb3 52. Rcd2 Ra1+ 53. Ke2 a5 54. Rb7+ Kc3 55. Rc7+
Kb3 56. Kd3 Rf1 57. Rb7+ Ka3 58. Ke3 a4 59. Kf4 Rh4+
60. Kg5 Rxf3 61. Kxh4 Rf4+ 62. Kg3 Rf3+ 63. Kg4 Rf4+
64. Kg5 Rf5+ 65. Kg6 Rf6+ 66. Kg7 Rg6+ 67. Kh8 Rg8+ 68. Kh7
Rh8+ 69. Kg6 Rh6+ 70. Kf5 Rf6+ 71. Kg5 Rf5+ 72. Kg6 Rf6+
73. Kg7 Rg6+ 74. Kh8 Rg8+ 75. Kh7 Rh8+ 76. Kg6 Rh6+ 77. Kg5
Rg6+ 78. Kf4 Rg4+ 79. Kf3 Rf4+ 80. Ke2 Rxe4+ 81. Kd1 Re1+
82. Kc2 Rc1+ 83. Kd3 Rc3+ 84. Kd4 Rc4+ 85. Kd5 Rc5+ 86. Kd6
Rc6+ 87. Kd7 Rb6 88. Rc7 Rb2 89. Rd4 Rb8 90. Rcc4 Ra8
91. Rb4 Ra6 92. Ke7 Ra7+ 93. Ke6 Ra6+ 94. Ke5 Ra5+ 95. Ke4
Ra6 96. Kd3 Ra8 97. Kc2 Rc8+ 98. Kb1 Rb8 1/2-1/2

[Uri Blass]

Houdini's analysis seems to be wrong and white should win.

With no tablebases I do not get a mate score fast but I guess that with tablebases stockfish should be able to get a mate score.

[Isaac]

Houdini's analysis seems to be wrong and white should win.

With no tablebases I do not get a mate score fast but I guess that with tablebases stockfish should be able to get a mate score.

I also think the same about Houdini. I have a slow pc and only the 5 men syzygy.
Stockfish (recent dev version) increases its eval at every increase of depth.
+3.52 evaluation at depth 35. 1437 M nodes searched.
I expect someone with the 6 men syzygy to reach higher scores at lower depths.

[Lyudmil Tsvetkov]

Here Ehrhardt Post played Rxf3 !? (or ?!), shallow analyze computer show a very high score but no win because zillions of checks ... what's the result with an accurate play ?
[d]8/1R6/8/6K1/p3P2r/k4P2/3R4/5r2 b - - 3 60


[pgn]
[Event "Barmen Meisterturnier B"]
[Site "Barmen DEU"]
[Date "1905.08.22"]
[EventDate "1905.08.14"]
[Round "8"]
[Result "1/2-1/2"]
[White "Alfred Ehrhardt Post"]
[Black "Aron Nimzowitsch"]
[ECO "D07"]
[WhiteElo "?"]
[BlackElo "?"]
[PlyCount "196"]

1. d4 d5 2. Nf3 Bg4 3. c4 Nc6 4. cxd5 Bxf3 5. gxf3 Qxd5 6. e3
e5 7. Nc3 Bb4 8. Bd2 Qd7 9. d5 Nce7 10. Qb3 Bxc3 11. Bxc3 Qxd5
12. Qxd5 Nxd5 13. Bxe5 f6 14. Bd4 Nge7 15. Rg1 O-O 16. Bc4 Kh8
17. Bc5 Rfe8 18. O-O-O Nb6 19. Bf7 Rf8 20. Bh5 g6 21. Bxe7 Rf7
22. Bxg6 hxg6 23. Bb4 Kg7 24. Rd4 c5 25. Bxc5 Rc8 26. b4 Nd7
27. Kb1 Nxc5 28. bxc5 Rxc5 29. Rdg4 Rb5+ 30. Ka1 g5 31. h4 Rc7
32. hxg5 Rc2 33. gxf6+ Kf7 34. Rb1 Ra5 35. Rxb7+ Kxf6 36. Rf4+
Ke6 37. a4 Rxf2 38. Rb2 Rf1+ 39. Ka2 Re5 40. Re2 Kd5 41. Kb2
Kc5 42. Kc3 Rc1+ 43. Kd2 Ra1 44. Rg4 Rf5 45. Rf2 Re5 46. e4
Kd4 47. Rg7 Ra2+ 48. Ke1 Rxa4 49. Re2 Rh5 50. Rd7+ Kc4
51. Rc2+ Kb3 52. Rcd2 Ra1+ 53. Ke2 a5 54. Rb7+ Kc3 55. Rc7+
Kb3 56. Kd3 Rf1 57. Rb7+ Ka3 58. Ke3 a4 59. Kf4 Rh4+ 60. Kg5
Rxf3 61. Kxh4 Rf4+ 62. Kg3 Rf3+ 63. Kg4 Rf4+ 64. Kg5 Rf5+
65. Kg6 Rf6+ 66. Kg7 Rg6+ 67. Kh8 Rg8+ 68. Kh7 Rh8+ 69. Kg6
Rh6+ 70. Kf5 Rf6+ 71. Kg5 Rf5+ 72. Kg6 Rf6+ 73. Kg7 Rg6+
74. Kh8 Rg8+ 75. Kh7 Rh8+ 76. Kg6 Rh6+ 77. Kg5 Rg6+ 78. Kf4
Rg4+ 79. Kf3 Rf4+ 80. Ke2 Rxe4+ 81. Kd1 Re1+ 82. Kc2 Rc1+
83. Kd3 Rc3+ 84. Kd4 Rc4+ 85. Kd5 Rc5+ 86. Kd6 Rc6+ 87. Kd7
Rb6 88. Rc7 Rb2 89. Rd4 Rb8 90. Rcc4 Ra8 91. Rb4 Ra6 92. Ke7
Ra7+ 93. Ke6 Ra6+ 94. Ke5 Ra5+ 95. Ke4 Ra6 96. Kd3 Ra8 97. Kc2
Rc8+ 98. Kb1 Rb8 1/2-1/2[/pgn]

I am not certain if it is a draw or not, but this is a good position to generalize some stalemate rules, as I see that recently even SF tried to implement some improved stalemate detection in the eval. Whether stalemate belongs to the search or not, that is another matter.

As you easily see from the diagram, in the generalized case stalemate occurs when:

- king of the weaker side can not move, has zero mobility, like Ka3
- none of the pawns of the weaker side can move, all of them have zero mobility, either because they are blocked by enemy pawns, or pawns, own king or other pieces; on the diagram, the black pawn on a4 is blocked by the own king
- weaker side has 1 piece at most; indeed, after white captures Kh4, black is left with Rf3 as its only piece
- weaker side is side to move

Similar stalemating motives occur also when weaker side has only one queen, bishop or knight that would give an unavoidable check to the enemy king before stalemate arises. So that, although zero king mobility is a bad thing in the general case, especially in the endgame, what concerns stalemating motives zero king mobility is an exception. As Don used to say, there are always exceptions.

The first 2 conditions are also obligatory when the weaker side has no other pieces besides the king, like in king + edge pawn + wrong-squared bishop vs king. I think the latter simplistic stalemate motive would occur in only about 10% of all stalemating cases, but it catches the attention because of the surprising material imbalance leading to the draw.

[zullil]

Houdini's analysis seems to be wrong and white should win.

With no tablebases I do not get a mate score fast but I guess that with tablebases stockfish should be able to get a mate score.

I also think the same about Houdini. I have a slow pc and only the 5 men syzygy.
Stockfish (recent dev version) increases its eval at every increase of depth.
+3.52 evaluation at depth 35. 1437 M nodes searched.
I expect someone with the 6 men syzygy to reach higher scores at lower depths.

Stockfish (recent dev version) no tablebases:

Code: Select all

info depth 71 seldepth 112 score cp 9970 lowerbound nodes 9505207143 nps 3331875 time 2852810 multipv 1 pv g5h4 f3f4 h4h5 f4f5 h5h6 f5f6 h6g7 f6g6 g7h8 g6g8 h8h7 g8h8 h7g6 h8g8 g6f5 g8f8 f5e5 f8e8 e5d4 e8e4 d4d5 e4e5 d5c6 e5c5 c6d7 c5b5 b7c7 b5h5 d7d6 a3b3 d2d3 b3b4 d3d4 b4b3 c7b7 b3c3 d4a4 h5h6 d6d5 h6h5 d5e6 h5h6 e6f5 h6d6 b7b4 d6d5 f5e6 d5d8 b4c4 c3b3 a4b4 b3a3 b4b7 d8e8 e6d5 e8h8 c4c6 h8h5 d5d4 h5h4 d4d3 h4h3 d3d2 h3h2 d2e1 h2h1 e1f2 h1h2 f2g3 h2g2 g3g2

[AdminX]

8/1R6/8/6K1/p3P2r/k4P2/3R4/5r2 b - - 0 1

Analysis by Komodo TCECr 64-bit :

60...Rxf3 61.Kxh4 Rf4+ 62.Kg5 Rg4+ 63.Kf6 Rf4+ 64.Ke6 Rxe4+ 65.Kd5 Re5+ 66.Kc6 Rc5+ 67.Kd7 Rb5 68.Rc7 Rh5 69.Kd6 Kb3 70.Rd3+ Kb4 71.Rd4+ Kb3 72.Rb7+ Kc2 73.Rxa4 Rh6+ 74.Ke5 Rh5+ 75.Kf6 Rh6+ 76.Kg5 Rc6 77.Ra3 Rc5+ 78.Kf6 Rc6+ 79.Ke5 Rc8 80.Ke6 Rh8 81.Kd6 Rh6+ 82.Kc5 Rh4 83.Rd7 Rh5+ 84.Kc6 +- (2.42) Depth: 30 00:05:15 3584MN

(06.05.2014)

[Lyudmil Tsvetkov]

[d]7k/7p/7P/2BQ4/4P3/P7/1q6/6K1 b - - 0 1
Black king and pawn can not move, black has only one piece apart from the king and Qg2 stalemates the black king

[d]7k/5R1p/7P/3Q4/4P1n1/P7/6PP/6NK b - - 0 1
Nf2 stalemates the black king, but here the black king can move on g8, so it is not clear if it is not better to leave stalemate to search after all

[d]8/6kP/8/6N1/5K2/8/8/8 w - - 0 1
Another surprising endgame position leading to a draw because of stalemate

Vincent, sorry, you posted a nice position, I am not going to hijack your thread by posting any more comments. It is just that other people think in terms of scores and me in terms of rules.

[yanquis1972]

i analyzed this w/ SF 6man for about an hour. the score was 123.xx after Rxf3 & never altered. so either i didnt give it enough time to solve for mate or houdini has some nice bit of code that allowed it to recognize draw.