Drawn position from 2011 Australian Open

[stevenaaus]

How many engines eval this as zero ?
Black erroneously resigned this position... costing him the tournament!

[D] 2n5/2P5/1Pk5/2B5/8/p7/8/1K6 b - - 2 86

[Event "Australian Open"]
[Site "NSLCC, Cammeray, NSW"]
[Date "2011.01.07"]
[Round "6.4"]
[White "Illingworth, Max"]
[Black "Ly, Moulthun"]
[Result "1-0"]
[WhiteElo "2311"]
[BlackElo "2321"]
[ECO "B80"]
[EventDate "2011.01.02"]
[PlyCount "171"]
[EventType "swiss"]
[EventRounds "11"]
[EventCountry "AUS"]
[SourceDate "2011.01.02"]

1. e4 c5 2. Nf3 d6 3. d4 cxd4 4. Nxd4 Nf6 5. Nc3 a6 6. Be3 e6 7. Qd2 b5 8.
f3 Nbd7 9. g4 Nb6 10. O-O-O Bb7 11. Bd3 Nfd7 12. Kb1 Rc8 13. g5 Ne5 14.
Qe1 Be7 15. h4 Nec4 16. Bc1 d5 17. exd5 Nxd5 18. Nxd5 Bxd5 19. Be4 O-O 20.
g6 Bxe4 21. gxf7+ Rxf7 22. Qxe4 Qd5 23. Nxe6 Qxe4 24. fxe4 Rc6 25. Nd4
Rc8 26. Nf5 Bf8 27. Rhg1 Rfc7 28. Bf4 Ra7 29. h5 Kf7 30. Rgf1 Ke8 31. c3
g6 32. hxg6 hxg6 33. Nd6+ Bxd6 34. Bxd6 Rf7 35. Rh1 Rc6 36. Rh8+ Kd7 37.
Bc5+ Kc7 38. Ra8 Rf4 39. Ra7+ Kb8 40. Rd8+ Rc8 41. Rxc8+ Kxc8 42. Rxa6
Rxe4 43. b3 Ne5 44. Re6 g5 45. Bd4 Nf3 46. Rxe4 Nd2+ 47. Kc2 Nxe4 48. Kd3
Nd6 49. Be3 Nf7 50. Ke4 Kb7 51. Kd5 g4 52. Ke6 Nd8+ 53. Kf5 g3 54. Kg4 g2
55. Kf3 Kc6 56. Kxg2 Kd5 57. Kf3 Nc6 58. Bd2 Ne5+ 59. Kf4 Nd3+ 60. Ke3 Nc5
61. Bc1 Ne4 62. Bb2 Nd6 63. Ba3 Nf7 64. Bf8 Ne5 65. Bg7 Nf7 66. Kd3 Nd6
67. Kc2 Nf5 68. Bf6 Ke6 69. Bh8 Kd5 70. Kb2 Ne7 71. Kc2 Nc8 72. Kd3 Nd6
73. Bg7 Nb7 74. Bd4 Nd6 75. Be3 Nb7 76. a4 bxa4 77. c4+ Kc6 78. b4 a3 79.
b5+ Kd6 80. Kc2 Na5 81. c5+ Kd5 82. Kb1 Nb7 83. c6 Nd6 84. c7 Nc8 85. b6
Kc6 86. Bc5 {And Black resigned, although Stephen Solomon pointed out that
Black can just put his king on a8, and White can make no further progress
because of stalemate defences.} 1-0

[Vinvin]

Very interesting !!

Which engine evaluate this position as roughly equal ?
[d]k1n5/2P5/1P6/2B5/1K6/8/8/8 w - - 0 5

[Tom Barrister]

None that I tried had any idea. This isn't surprising, since engines are primiarly brute force anyway. The existence of tablebases means that most engines contain about as much native endgame information as they do information about piloting a helicopter. If the engine can't find it through brute force, it won't find it at all. Take this position:


[D] 8/8/1q3k2/7p/8/7R/6P1/7K w - - 0 1

This is a known draw. White shuttles the Rook from h3 to f3 and the King around g1/h1/h2, and Black can't make progress. Of course the tablebases know that. Try explaining it to an engine. The engine will send Black's King and Queen on a tour of the entire board before it gives up trying.

[stevenaaus]

This isn't surprising, since engines are primiarly brute force anyway.

Yes, our new overlords do have their limits ;>

The fact that it's a draw isn't obvious really. There's quite a few continuations, but none that don't end in stalemate.