![[gb]](/mwf/v12/flags/gb.png "United Kingdom")
Here is a curious game from a Noomen2012 opening set 8 engine tournament I am running. Deep Rybka 4.1MP (2 threads) misses winning moves at move 41. and goes on to not only throw away almost a 2 pawn advantage but ends up blundering at move 54 and not seeing it until move 58 giving a 6.00+ adverse evaluation chenge.
I ran this with full 4 threads and 2Gb LP hash but still took over 10 minutes to see it was already lost at move 57.g4 move before evaluation jump.
Thought this multiple iteration issue had ben addressed from Rybka 3 but still shown to be present here with no updates in engine output window and very little information being shown in current move window.
PeterG
[Event "Noomen2012 8 eng"]
[Site "Newport, South Wales"]
[Date "2012.01.30"]
[Round "18.2"]
[White "Deep Rybka 4.1 x64 2T"]
[Black "Zappa Mexico II 2T"]
[Result "0-1"]
[ECO "B12"]
[WhiteElo "2990"]
[BlackElo "2875"]
[Annotator "0.20;0.32"]
[PlyCount "156"]
[EventDate "2012.01.23"]
[Source "Grayson"]
{Intel(R) Core(TM)2 Quad CPU Q9550 @ 2.83GHz 3536 MHz W=17.7 plies; 222kN/ s;
9,792 TBAs B=15.7 plies; 1,536kN/s; 9,744 TBAs} 1. e4 c6 2. d4 d5 3. e5 Bf5 4.
Nd2 e6 5. Nb3 Nd7 6. Nf3 Qc7 {0.32/13 22} 7. Be2 {0.20/14 9 (Bd3)} f6 {0.20/14
13 (c5)} 8. O-O {0.22/15 11 (Bf4)} O-O-O {0.19/14 21 (Ne7)} 9. Bd3 {0.27/14 22
(Bf4)} Bg4 {-0.12/15 13 (Ne7)} 10. Bf4 {0.22/15 12 (exf6)} fxe5 {-0.27/14 17
(g5)} 11. dxe5 {0.31/15 11} c5 {-0.28/14 11 (Ngf6)} 12. Nbd2 {0.21/16 24} h6 {
-0.32/15 0 (Ne7)} 13. c4 {0.32/15 14 (Bg3)} Ne7 {-0.16/14 25} 14. h3 {0.30/18
0 (cxd5)} Bh5 {-0.23/13 8} 15. cxd5 {0.35/17 0 (Bh2)} Nxd5 {-0.14/14 10} 16.
Bh2 {0.23/17 2} N7b6 {0.02/14 17 (Kb8)} 17. Qb1 {0.35/15 11 (Be4)} g5 {0.03/13
7} 18. a4 {0.29/17 0} Nb4 {0.09/13 7} 19. Bg6 {0.29/18 0} Bxf3 {0.08/13 6} 20.
Nxf3 {0.29/18 0} Rg8 {0.11/13 7} 21. Qe4 {0.29/17 0 (Be4)} Rg7 {0.05/13 11} 22.
a5 {0.27/16 0} Nc4 {0.00/14 15} 23. Qxc4 {0.30/17 0 (Bf5)} Rxg6 {-0.11/7 0} 24.
Rfd1 {0.30/17 15 (Qe4)} Qxa5 {-0.02/13 5 (Kb8)} 25. Nd4 {0.35/17 10} Qb6 {0.00/
15 0} 26. Nb5 {0.35/17 14} Kb8 {0.03/15 0} 27. Rxd8+ {0.47/15 11} Qxd8 {0.02/
16 0} 28. Nd6 {0.47/17 14 (Nxa7)} b6 {0.00/14 5 (Nc6)} 29. Ra3 {0.48/16 11} Nd5
{0.19/15 1} 30. Qg4 {0.63/17 11 (Rf3)} Qd7 {0.50/15 18} 31. Rd3 {0.74/18 0} Nb4
{0.83/15 11 (Rg8)} 32. Rf3 {0.92/18 17} Be7 {0.93/17 0} 33. Rf7 {1.04/18 14
(Qe4)} Rg8 {0.96/13 5 (Nd5)} 34. Rh7 {1.00/18 19} Nc6 {0.99/15 0} 35. Rxh6 {1.
12/17 11} Nd4 {1.02/15 0} 36. Qe4 {1.18/16 5} Rf8 {1.03/14 2 (Rd8)} 37. Bg3 {
1.46/14 8} Qc7 {1.15/14 1 (Qc6)} 38. Rh7 {1.45/17 5 (Kh2)} Nf5 {1.35/15 8} 39.
Qc4 {1.68/17 1} Qd7 {1.41/15 6 (Nxg3)} 40. Nxf5 {1.67/17 5} Rxf5 {1.51/16 0}
41. Rh6 {1.67/18 14} (41. Qe4 $1) (41. h4 $1) 41... b5 {1.59/16 0} 42. Qxe6 {
1.67/19 43} Qxe6 {1.48/18 0} 43. Rxe6 {1.67/19 6} Rf7 {1.48/18 0} 44. Rg6 {1.
39/18 10 (Kf1)} Kb7 {1.50/16 15} 45. f3 {1.39/18 0 (Kf1)} c4 {1.31/15 9 (b4)}
46. Kf1 {1.30/17 12} b4 {1.28/16 3} 47. Ke2 {1.30/18 2} a5 {1.32/15 12} 48. Bf2
{1.30/18 0} b3 {1.06/15 23 (a4)} 49. Bd4 {0.75/14 10 (e6)} Bb4 {0.64/14 10} 50.
Rxg5 {0.75/15 1 (e6)} Rd7 {0.39/12 6} 51. Rg4 {0.48/14 0} c3 {0.39/12 0} 52. e6
{0.00/16 9} Rxd4 {-0.01/14 0} 53. Rxd4 {0.00/20 9} cxb2 {-0.01/16 0} 54. Rd7+
$2 $17 {0.00/20 12} (54. Rd1 a4 55. Kd3 a3 56. e7 a2 (56... Bxe7 57. Kc3 a2 58.
Kxb2) 57. e8=Q b1=Q+ 58. Rxb1 axb1=Q+ 59. Kc4 Ba5) 54... Kc6 $19 {-0.01/16 0}
55. Rd1 {0.00/21 10} a4 {-0.01/17 0} 56. Kd3 {0.00/21 10} Kb5 {-0.01/17 31} 57.
g4 {0.00/20 0 (f4)} Be7 {-6.59/14 10} 58. g5 {-6.11/13 0 (Kc3)} Kb4 {-7.44/13 0
} 59. g6 {-8.91/10 4} a3 {-7.47/14 0} 60. g7 {-10.52/14 5} a2 {-8.83/14 0} 61.
g8=Q {-11.34/15 8 (Ke2)} b1=Q+ {-8.92/8 0} 62. Rxb1 {-10.51/13 4} axb1=Q+ {-10.
39/14 11} 63. Kd4 {-#16/17 0} Qd1+ {-10.45/14 23 (Bc5+)} 64. Ke3 {-#14/16 10}
Qc1+ {-11.74/14 21} 65. Ke2 {-#14/17 0 (Ke4)} Qc2+ {-#16/14 19} 66. Ke3 {-#11/
19 0} Bc5+ {-#15/14 0} 67. Kf4 {-#79/8 0} Bd6+ {-#14/14 0 (b2)} 68. Ke3 {-#11/
18 17} b2 {-#13/15 9} 69. Qa8 {-#10/18 0 (Qg4+)} b1=Q {-#10/13 9 (Bc5+)} 70.
Qb7+ {-#10/14 11} Kc3 {-#9/15 3} 71. Qc6+ {-#8/15 3} Kb3 {-#8/15 0} 72. Qb7+ {
-#7/17 6} Bb4 {-#7/14 0} 73. Qd5+ {-#6/19 7} Ka3 {-#6/13 0} 74. Qa8+ {-#5/21 8}
Kb2 {-#5/14 0} 75. Qh8+ {-#4/24 8} Kb3 {-#4/22 0} 76. Qb2+ {-#3/24 8} Qbxb2 {
-#3/63 0} 77. Kf4 {-#2/43 9} Qf6+ {-#2/63 0 (Qg6)} 78. Kg4 {-#1/55 12 (Kg3)}
Qcg6# {-#1/63 0} 0-1
I ran this with full 4 threads and 2Gb LP hash but still took over 10 minutes to see it was already lost at move 57.g4 move before evaluation jump.
Thought this multiple iteration issue had ben addressed from Rybka 3 but still shown to be present here with no updates in engine output window and very little information being shown in current move window.
PeterG
[Event "Noomen2012 8 eng"]
[Site "Newport, South Wales"]
[Date "2012.01.30"]
[Round "18.2"]
[White "Deep Rybka 4.1 x64 2T"]
[Black "Zappa Mexico II 2T"]
[Result "0-1"]
[ECO "B12"]
[WhiteElo "2990"]
[BlackElo "2875"]
[Annotator "0.20;0.32"]
[PlyCount "156"]
[EventDate "2012.01.23"]
[Source "Grayson"]
{Intel(R) Core(TM)2 Quad CPU Q9550 @ 2.83GHz 3536 MHz W=17.7 plies; 222kN/ s;
9,792 TBAs B=15.7 plies; 1,536kN/s; 9,744 TBAs} 1. e4 c6 2. d4 d5 3. e5 Bf5 4.
Nd2 e6 5. Nb3 Nd7 6. Nf3 Qc7 {0.32/13 22} 7. Be2 {0.20/14 9 (Bd3)} f6 {0.20/14
13 (c5)} 8. O-O {0.22/15 11 (Bf4)} O-O-O {0.19/14 21 (Ne7)} 9. Bd3 {0.27/14 22
(Bf4)} Bg4 {-0.12/15 13 (Ne7)} 10. Bf4 {0.22/15 12 (exf6)} fxe5 {-0.27/14 17
(g5)} 11. dxe5 {0.31/15 11} c5 {-0.28/14 11 (Ngf6)} 12. Nbd2 {0.21/16 24} h6 {
-0.32/15 0 (Ne7)} 13. c4 {0.32/15 14 (Bg3)} Ne7 {-0.16/14 25} 14. h3 {0.30/18
0 (cxd5)} Bh5 {-0.23/13 8} 15. cxd5 {0.35/17 0 (Bh2)} Nxd5 {-0.14/14 10} 16.
Bh2 {0.23/17 2} N7b6 {0.02/14 17 (Kb8)} 17. Qb1 {0.35/15 11 (Be4)} g5 {0.03/13
7} 18. a4 {0.29/17 0} Nb4 {0.09/13 7} 19. Bg6 {0.29/18 0} Bxf3 {0.08/13 6} 20.
Nxf3 {0.29/18 0} Rg8 {0.11/13 7} 21. Qe4 {0.29/17 0 (Be4)} Rg7 {0.05/13 11} 22.
a5 {0.27/16 0} Nc4 {0.00/14 15} 23. Qxc4 {0.30/17 0 (Bf5)} Rxg6 {-0.11/7 0} 24.
Rfd1 {0.30/17 15 (Qe4)} Qxa5 {-0.02/13 5 (Kb8)} 25. Nd4 {0.35/17 10} Qb6 {0.00/
15 0} 26. Nb5 {0.35/17 14} Kb8 {0.03/15 0} 27. Rxd8+ {0.47/15 11} Qxd8 {0.02/
16 0} 28. Nd6 {0.47/17 14 (Nxa7)} b6 {0.00/14 5 (Nc6)} 29. Ra3 {0.48/16 11} Nd5
{0.19/15 1} 30. Qg4 {0.63/17 11 (Rf3)} Qd7 {0.50/15 18} 31. Rd3 {0.74/18 0} Nb4
{0.83/15 11 (Rg8)} 32. Rf3 {0.92/18 17} Be7 {0.93/17 0} 33. Rf7 {1.04/18 14
(Qe4)} Rg8 {0.96/13 5 (Nd5)} 34. Rh7 {1.00/18 19} Nc6 {0.99/15 0} 35. Rxh6 {1.
12/17 11} Nd4 {1.02/15 0} 36. Qe4 {1.18/16 5} Rf8 {1.03/14 2 (Rd8)} 37. Bg3 {
1.46/14 8} Qc7 {1.15/14 1 (Qc6)} 38. Rh7 {1.45/17 5 (Kh2)} Nf5 {1.35/15 8} 39.
Qc4 {1.68/17 1} Qd7 {1.41/15 6 (Nxg3)} 40. Nxf5 {1.67/17 5} Rxf5 {1.51/16 0}
41. Rh6 {1.67/18 14} (41. Qe4 $1) (41. h4 $1) 41... b5 {1.59/16 0} 42. Qxe6 {
1.67/19 43} Qxe6 {1.48/18 0} 43. Rxe6 {1.67/19 6} Rf7 {1.48/18 0} 44. Rg6 {1.
39/18 10 (Kf1)} Kb7 {1.50/16 15} 45. f3 {1.39/18 0 (Kf1)} c4 {1.31/15 9 (b4)}
46. Kf1 {1.30/17 12} b4 {1.28/16 3} 47. Ke2 {1.30/18 2} a5 {1.32/15 12} 48. Bf2
{1.30/18 0} b3 {1.06/15 23 (a4)} 49. Bd4 {0.75/14 10 (e6)} Bb4 {0.64/14 10} 50.
Rxg5 {0.75/15 1 (e6)} Rd7 {0.39/12 6} 51. Rg4 {0.48/14 0} c3 {0.39/12 0} 52. e6
{0.00/16 9} Rxd4 {-0.01/14 0} 53. Rxd4 {0.00/20 9} cxb2 {-0.01/16 0} 54. Rd7+
$2 $17 {0.00/20 12} (54. Rd1 a4 55. Kd3 a3 56. e7 a2 (56... Bxe7 57. Kc3 a2 58.
Kxb2) 57. e8=Q b1=Q+ 58. Rxb1 axb1=Q+ 59. Kc4 Ba5) 54... Kc6 $19 {-0.01/16 0}
55. Rd1 {0.00/21 10} a4 {-0.01/17 0} 56. Kd3 {0.00/21 10} Kb5 {-0.01/17 31} 57.
g4 {0.00/20 0 (f4)} Be7 {-6.59/14 10} 58. g5 {-6.11/13 0 (Kc3)} Kb4 {-7.44/13 0
} 59. g6 {-8.91/10 4} a3 {-7.47/14 0} 60. g7 {-10.52/14 5} a2 {-8.83/14 0} 61.
g8=Q {-11.34/15 8 (Ke2)} b1=Q+ {-8.92/8 0} 62. Rxb1 {-10.51/13 4} axb1=Q+ {-10.
39/14 11} 63. Kd4 {-#16/17 0} Qd1+ {-10.45/14 23 (Bc5+)} 64. Ke3 {-#14/16 10}
Qc1+ {-11.74/14 21} 65. Ke2 {-#14/17 0 (Ke4)} Qc2+ {-#16/14 19} 66. Ke3 {-#11/
19 0} Bc5+ {-#15/14 0} 67. Kf4 {-#79/8 0} Bd6+ {-#14/14 0 (b2)} 68. Ke3 {-#11/
18 17} b2 {-#13/15 9} 69. Qa8 {-#10/18 0 (Qg4+)} b1=Q {-#10/13 9 (Bc5+)} 70.
Qb7+ {-#10/14 11} Kc3 {-#9/15 3} 71. Qc6+ {-#8/15 3} Kb3 {-#8/15 0} 72. Qb7+ {
-#7/17 6} Bb4 {-#7/14 0} 73. Qd5+ {-#6/19 7} Ka3 {-#6/13 0} 74. Qa8+ {-#5/21 8}
Kb2 {-#5/14 0} 75. Qh8+ {-#4/24 8} Kb3 {-#4/22 0} 76. Qb2+ {-#3/24 8} Qbxb2 {
-#3/63 0} 77. Kf4 {-#2/43 9} Qf6+ {-#2/63 0 (Qg6)} 78. Kg4 {-#1/55 12 (Kg3)}
Qcg6# {-#1/63 0} 0-1
i thought rybkas elo is over 3000? it even states it on either rybka 3 dvd plastic cover or rybka 4 dvd plastic cover, did they lie?
ELO should NOT be used as some absolute strength measurement in chess and besides the example Peter shows, there are many more where a club chess player would find stronger moves than R4.1 in certain positions.
Rybka is undoubtedly a strong, super GM+ engine but from time to time these oddities arise that show there may be some underlying knowledge gap, over-aggressive pruning, hashing issue or just some error in the evaluation. If these can be addressed and improved on, it should make the engine stronger overall. Rybka is not alone here and if you test this with some other engines they also struggle.
PeterG
PeterG
![[us]](/mwf/v12/flags/us.png "United States")
Other engines play the same game here.
Even Houdini is confident of a win, confident of a draw at 55(like rybka), and doesn't see the lose coming.
I don't see this as a fault by rybka's programming, just the role of time in the game :)
Given more time the win would have been found.
[Event "?"]
[Site "?"]
[Date "????.??.??"]
[Round "?"]
[White "?"]
[Black "?"]
[Result "0-1"]
1. e4 c6 2. d4 d5 3. e5 Bf5 4. Nd2 e6 5. Nb3 Nd7 6. Nf3 Qc7 7. Be2 f6 8. O-O
O-O-O 9. Bd3 Bg4 10. Bf4 fxe5 11. dxe5 c5 12. Nbd2 h6 13. c4 Ne7 14. h3 Bh5
15. cxd5 Nxd5 16. Bh2 N7b6 17. Qb1 g5 18. a4 Nb4 19. Bg6 Bxf3 20. Nxf3 Rg8 21.
Qe4 Rg7 22. a5 Nc4 23. Qxc4 Rxg6 24. Rfd1 Qxa5 25. Nd4 Qb6 26. Nb5 Kb8 27.
Rxd8+ Qxd8 28. Nd6 b6 29. Ra3 Nd5 30. Qg4 Qd7 31. Rd3 Nb4 32. Rf3 Be7 33. Rf7
Rg8 34. Rh7 Nc6 35. Rxh6 Nd4 36. Qe4 Rf8 37. Bg3 Qc7 38. Rh7 Nf5 39. Qc4 Qd7
40. Nxf5 Rxf5 41. Rh6 (41. h4 gxh4 42. Bxh4 Qd1+ 43. Kh2 Bxh4 44. Qxe6 Qd3 45.
Rxh4 Qd8 46. Qxf5 {+Mate in 38} ) 41... b5 42. Qxe6 Qxe6 43. Rxe6 Rf7 44. Rg6
Kb7 45. f3 c4 46. Kf1 b4 47. Ke2 a5 48. Bf2 b3 49. Bd4 Bb4 50. Rxg5 Rd7 51.
Rg4 c3 52. e6 Rxd4 53. Rxd4 cxb2 54. Rd7+ Kc6 55. Rd1 a4 56. Kd3 Kb5 57. g4
Be7 58. g5 Kb4 59. g6 a3 60. g7 a2 61. g8=Q b1=Q+ 62. Rxb1 axb1=Q+ 63. Kd4
Qd1+ 64. Ke3 Qc1+ 65. Ke2 Qc2+ 66. Ke3 Bc5+ 67. Kf4 Bd6+ 68. Ke3 b2 69. Qa8
b1=Q 70. Qb7+ Kc3 71. Qc6+ Kb3 72. Qb7+ Bb4 73. Qd5+ Ka3 74. Qa8+ Kb2 75. Qh8+
Kb3 76. Qb2+ Qbxb2 77. Kf4 Qf6+ 78. Kg4 Qcg6# 0-1
Even Houdini is confident of a win, confident of a draw at 55(like rybka), and doesn't see the lose coming.
I don't see this as a fault by rybka's programming, just the role of time in the game :)
Given more time the win would have been found.
[Event "?"]
[Site "?"]
[Date "????.??.??"]
[Round "?"]
[White "?"]
[Black "?"]
[Result "0-1"]
1. e4 c6 2. d4 d5 3. e5 Bf5 4. Nd2 e6 5. Nb3 Nd7 6. Nf3 Qc7 7. Be2 f6 8. O-O
O-O-O 9. Bd3 Bg4 10. Bf4 fxe5 11. dxe5 c5 12. Nbd2 h6 13. c4 Ne7 14. h3 Bh5
15. cxd5 Nxd5 16. Bh2 N7b6 17. Qb1 g5 18. a4 Nb4 19. Bg6 Bxf3 20. Nxf3 Rg8 21.
Qe4 Rg7 22. a5 Nc4 23. Qxc4 Rxg6 24. Rfd1 Qxa5 25. Nd4 Qb6 26. Nb5 Kb8 27.
Rxd8+ Qxd8 28. Nd6 b6 29. Ra3 Nd5 30. Qg4 Qd7 31. Rd3 Nb4 32. Rf3 Be7 33. Rf7
Rg8 34. Rh7 Nc6 35. Rxh6 Nd4 36. Qe4 Rf8 37. Bg3 Qc7 38. Rh7 Nf5 39. Qc4 Qd7
40. Nxf5 Rxf5 41. Rh6 (41. h4 gxh4 42. Bxh4 Qd1+ 43. Kh2 Bxh4 44. Qxe6 Qd3 45.
Rxh4 Qd8 46. Qxf5 {+Mate in 38} ) 41... b5 42. Qxe6 Qxe6 43. Rxe6 Rf7 44. Rg6
Kb7 45. f3 c4 46. Kf1 b4 47. Ke2 a5 48. Bf2 b3 49. Bd4 Bb4 50. Rxg5 Rd7 51.
Rg4 c3 52. e6 Rxd4 53. Rxd4 cxb2 54. Rd7+ Kc6 55. Rd1 a4 56. Kd3 Kb5 57. g4
Be7 58. g5 Kb4 59. g6 a3 60. g7 a2 61. g8=Q b1=Q+ 62. Rxb1 axb1=Q+ 63. Kd4
Qd1+ 64. Ke3 Qc1+ 65. Ke2 Qc2+ 66. Ke3 Bc5+ 67. Kf4 Bd6+ 68. Ke3 b2 69. Qa8
b1=Q 70. Qb7+ Kc3 71. Qc6+ Kb3 72. Qb7+ Bb4 73. Qd5+ Ka3 74. Qa8+ Kb2 75. Qh8+
Kb3 76. Qb2+ Qbxb2 77. Kf4 Qf6+ 78. Kg4 Qcg6# 0-1
It is the massive jump in evaluation after 57.g4 that highlights a blindspot that may be useful to address. However other than missing at least 2 better, probably winning moves, at 41. h4 or 41.Qe6 the critical point to avoid the loss is move 54. For example, Critter 1.4 sees 54.Rd1 in a few seconds and evaluates 57.g4 or f4 as already lost very quickly. Interestingly DR4.1 found 54.Rd1 very quickly when given the full 4 cores and 2 Gb hash. So probably a combination of different factors.
The large eval jump suggests to me something missing as opposed to the progressively deteriorating/improving score. Just as a human would a blunder for example. In this case engines need to see 54.Rd7+ as losing.
PeterG
The large eval jump suggests to me something missing as opposed to the progressively deteriorating/improving score. Just as a human would a blunder for example. In this case engines need to see 54.Rd7+ as losing.
PeterG
>It is the massive jump in evaluation after 57.g4 that highlights a blindspot that may be useful to address
Hi,
Turning off TBs helps to find the loss much faster, Rybkas node count was halved on my machine with the default TB probe of 36
Regards
> Turning off TBs helps to find the loss much faster, Rybkas node count was halved on my machine with the default TB probe of 36
Thank you for suggestion. I tried without TB's but found that DR4.1 still has similar problem at 57.g4 - see clipped analysis. However, what is clear is that when using full 4 threads and 2Gb hash it would likely have played a different game, playing different game moves finding 41.h4 within typical game move time window but also, importantly, it did not even register 54.Rd7+ immediately playing instead the draw saving 54.Rd1 instead. However, if it had played the same game moves then after 54.Rd7 it still showed a zero score finding the adverse evaluation at 57.g4 in 18 ply but still taking well over the typical game move time window. So with or without TB's, on my set up, it still has the problem.
Interestingly, as black, it sees 54..Kc6 as better for black fairly quickly and evaluates a significant adverse score for white at moves 55 and 56 so it suggests there is significant difference in the evaluation criteria for White and Black where in this case, the criteria for black is more accurate.
See clipped analysis as White after 56. Kd3
To simulate actual game play I clicked on the black moves for white analysis or white moves for back analysis at the appropriate time interval. So in this case there is information in hash as there would be in the actual game.
Regards,
PeterG
White analysis at move 57.
1: Deep Rybka 4.1 x64 2T - Zappa Mexico II 2T, Noomen2012 8 eng 2012
8/8/4P3/1k6/pb6/1p1K1P1P/1p4P1/3R4 w - - 0 1
Analysis by Deep Rybka 4.1 x64 NoTB:
57.g4 Be7
= (0.00) Depth: 6 00:00:00 0kN
57.g4 Be7
= (0.00) Depth: 7 00:00:00 0kN
57.g4 Be7
= (0.00) Depth: 8 00:00:00 0kN
57.g4 Be7
= (0.00) Depth: 9 00:00:00 0kN
57.g4 Be7
= (0.00) Depth: 10 00:00:00 0kN
57.g4 Be7
= (0.00) Depth: 11 00:00:00 0kN
57.g4 Be7
= (0.00) Depth: 12 00:00:00 0kN
57.g4 Be7
= (0.00) Depth: 13 00:00:00 0kN
57.g4 Be7
= (0.00) Depth: 14 00:00:00 0kN
57.g4 Be7
= (0.00) Depth: 15 00:00:00 0kN
57.g4 Be7
= (0.00) Depth: 16 00:00:00 0kN
57.g4 Be7
= (0.00) Depth: 17 00:00:00 0kN
57.g4 Be7
= (0.00) Depth: 18 00:00:00 0kN
57.h4 a3 58.e7
-+ (-8.50) Depth: 19 00:06:05 180mN
57.h4 Be7[] 58.h5 Kb4[] 59.h6 a3 60.h7 a2 61.h8Q b1Q+[] 62.Ke2 Qc2+[] 63.Rd2 Qc4+[] 64.Rd3 Qxe6+ 65.Kf2 Qc4[] 66.Qb8+ Ka3[] 67.Qa7+ Kb2 68.Rd2+ Kc1[] 69.Qd7 Bc5+ 70.Kg3 Qf1 71.Rd5 Bf2+ 72.Kg4
-+ (-10.61) Depth: 20 00:07:33 224mN
(Grayson, Newport, South Wales 20.02.2012)
Black analysis at move 56..
1: Deep Rybka 4.1 x64 2T - Zappa Mexico II 2T, Noomen2012 8 eng 2012
8/8/2k1P3/8/pb6/1p1K1P1P/1p4P1/3R4 b - - 0 1
Analysis by Deep Rybka 4.1 x64 NoTB:
56...Kb5 57.g4 Bd6 58.Kc3 Be5+ 59.Kd2 Bf6 60.f4 Kc4 61.g5 Be7 62.g6 a3 63.g7 a2 64.g8Q b1Q 65.Qc8+ Bc5 66.Qa6+ Kb4 67.Qb7+ Ka5 68.Qa8+ Kb5 69.Qb7+ Bb6 70.Qd5+ Kb4 71.Ke2 Qc2+
= (0.00) Depth: 6 00:00:00 0kN
56...Kb5 57.g4 Bd6 58.Kc3 Be5+ 59.Kd2 Bf6 60.f4 Kc4 61.g5 Be7 62.g6 a3 63.g7 a2 64.g8Q b1Q 65.Qc8+ Bc5 66.Qa6+ Kb4 67.Qb7+ Ka5 68.Qa8+ Kb5 69.Qb7+ Bb6 70.Qd5+ Kb4 71.Ke2 Qc2+
= (0.00) Depth: 7 00:00:00 0kN
56...Kb5 57.g4 Bd6 58.Kc3 Be5+ 59.Kd2 Bf6 60.f4 Kc4 61.g5 Be7 62.g6 a3 63.g7 a2 64.g8Q b1Q[] 65.Qc8+ Bc5[] 66.Qa6+ Kb4 67.Qb7+ Ka5 68.Qa8+ Kb5 69.Qb7+ Bb6[] 70.Qd5+ Kb4 71.Ke2 Qc2+
= (0.00) Depth: 8 00:00:00 0kN
56...Kb5 57.g4 Bd6 58.Kc3 Be5+ 59.Kd2 Bf6 60.f4 Kc4 61.g5 Be7 62.g6 a3 63.g7 a2 64.g8Q b1Q[] 65.Qc8+ Bc5[] 66.Qa6+ Kb4 67.Qb7+ Ka5 68.Qa8+ Kb5 69.Qb7+ Bb6[] 70.Qd5+ Kb4 71.Ke2 Qc2+
= (0.00) Depth: 9 00:00:00 0kN
56...Kb5 57.g4 Bd6 58.Kc3 Be5+ 59.Kd2 Bf6 60.f4 Kc4 61.g5 Be7 62.g6 a3 63.g7 a2 64.g8Q b1Q[] 65.Qc8+ Bc5[] 66.Qa6+ Kb4 67.Qb7+ Ka5 68.Qa8+ Kb5 69.Qb7+ Bb6[] 70.Qd5+ Kb4 71.Ke2 Qc2+
= (0.00) Depth: 10 00:00:00 0kN
56...Kb5 57.g4 Bd6 58.Kc3 Be5+ 59.Kd2 Bf6 60.f4 Kc4 61.g5 Be7 62.g6 a3 63.g7 a2 64.g8Q b1Q[] 65.Qc8+ Bc5[] 66.Qa6+ Kb4 67.Qb7+ Ka5 68.Qa8+ Kb5 69.Qb7+ Bb6[] 70.Qd5+ Kb4 71.Ke2 Qc2+
= (0.00) Depth: 11 00:00:00 0kN
56...Kb5 57.g4 Bd6 58.Kc3 Be5+ 59.Kd2 Bf6 60.f4 Kc4 61.g5 Be7 62.g6 a3 63.g7 a2 64.g8Q b1Q[] 65.Qc8+ Bc5[] 66.Qa6+ Kb4 67.Qb7+ Ka5 68.Qa8+ Kb5 69.Qb7+ Bb6[] 70.Qd5+ Kb4 71.Ke2 Qc2+
= (0.00) Depth: 12 00:00:00 0kN
56...Kb5 57.g4 Bd6 58.Kc3 Be5+ 59.Kd2 Bf6 60.f4 Kc4 61.g5 Be7 62.g6 a3 63.g7 a2 64.g8Q b1Q[] 65.Qc8+ Bc5[] 66.Qa6+ Kb4 67.Qb7+ Ka5 68.Qa8+ Kb5 69.Qb7+ Bb6[] 70.Qd5+ Kb4 71.Ke2 Qc2+
= (0.00) Depth: 13 00:00:00 0kN
56...Kb5 57.g4 Be7
= (0.00) Depth: 14 00:00:00 0kN
56...Kb5 57.g4 Be7
= (0.00) Depth: 15 00:00:00 0kN
56...Kb5 57.g4 Be7
= (0.00) Depth: 16 00:00:00 0kN
56...Kb5 57.g4 Be7
= (0.00) Depth: 17 00:00:00 0kN
56...Kb5 57.g4 Be7
= (0.00) Depth: 18 00:00:00 0kN
56...Kb5 57.g4 Be7
= (0.00) Depth: 19 00:00:00 0kN
56...Kb5 57.g4 Be7
= (-0.12 !) Depth: 20 00:00:00 0kN
56...Kb5 57.g4 Be7
=/+ (-0.27 !) Depth: 20 00:00:00 0kN
56...Kb5 57.g4 Be7[]
=/+ (-0.67 !) Depth: 20 00:00:00 153kN
56...Kb5 57.g4 Be7[]
-+ (-1.47 !) Depth: 20 00:00:02 1321kN
56...Kb5 57.g4 Be7[]
-+ (-3.07 !) Depth: 20 00:00:16 10124kN
56...Kb5 57.g4 Be7[]
-+ (-6.27 !) Depth: 20 00:00:50 29200kN
(Grayson, Newport, South Wales 20.02.2012)
![[kw]](/mwf/v12/flags/kw.png "Kuwait")
This is nothing new, and many engines have problems with horizon effect. For a human it is quite obvious that connected pawns are extremely dangerous and that the black king better watch out and either push his own pawns forward and try to create counterplay, or send his king to the other side and try to stop the pawn roller. To be fair to Rybka, I think with more time and bigger hardware it would have not lost the position. Pawn storms are extremely hard for engines (and humans) to calculate as they depend on super accurate deep calculations and one loss of tempi means that one side queens first and wins. So in the initial position black sees a pawn win and goes for it as that outweighs the static evaluation of the pawns (they are still not far advanced). I still think black is not losing after taking the pawn, but again horizon effect will not show danger on the engines search and black managed to do a series of poor moves due to not seeing the danger way down in the horizon. I would not worry too much about this as this problem will disappear as hardware strength improves. A static evaluation change that would change the engine behavior would probably backfire in other positions and cause a net loss in ELO. These type of weaknesses were very common with early engines, and strong GM's used them to beat engines regularly. This is not the case anymore and such losses are extremely rare.
One thing to note is that Rybka and all Rybka 3 based engines (Ippolit, Ivanhoe, Houdini etc..) are strong because they prune heavily and choose a few good moves to search through. So you can say that they don't have a deep comprehensive search of all moves, this makes them focus on stronger moves and in general that makes them incredibly strong ... of course you will find some positions where it backfires (like here), but over all the ELO strength is dramatically better.
I think even at this point the game is drawn
Interesting position here because it is a race to see who queens first. One missed tempo and it is game over so deep accurate calculation is necessary. I think Rd1! draws and anything else loses and here Rybka did not see it, or at least not quickly and neither does Houdini. I am pretty sure with more time they would see it though.
Actually now that I look at it, Rybka 4.1 does see it quite quickly on my Octa at 4 Ghz. Here is a clip analysis at 1 minute
Deep Rybka 4.1 x64 2T - Zappa Mexico II 2T, Noomen2012 8 eng 2012
Analysis by Deep Rybka 4.1 x64: (8 core)
54.Rd1 a4 55.e7 Bxe7 56.Kd3 a3 57.Kc3 a2 58.Kxb2 Bf6+ 59.Kxb3 a1Q 60.Rxa1 Bxa1 61.f4 Bf6 62.Kc4
+/= (0.49) Depth: 6 00:00:00 2kN
54.Rd7+ Kc6 55.Rd1 a4 56.e7 Bxe7 57.Kd3 Kc5 58.g3 Kb4 59.Re1 a3 60.Re4+ Kb5 61.Re5+
+/= (0.28) Depth: 7 00:00:00 6kN
54.Rd7+ Kc6 55.Rd1 a4 56.e7 Bxe7 57.Kd3 Kc5 58.g3 Kb4 59.Re1 a3 60.Re4+ Kb5 61.Re5+ Ka4 62.Re1
+/= (0.40 !) Depth: 8 00:00:00 7kN
54.Rd7+ Kc6 55.Rd1 a4 56.e7 Bxe7 57.Kd3 Kc5 58.Re1 Kb4 59.h4 a3 60.Re4+ Kc5 61.Re5+ Kc6 62.Re1 a2 63.Re6+
+/= (0.55 !) Depth: 8 00:00:00 7kN
54.Rd7+ Kc6 55.Rd1 a4 56.Kd3 a3 57.e7 Bxe7 58.Kc3 a2 59.Kxb2 Bf6+ 60.Kxb3 a1Q 61.Rxa1 Bxa1 62.Kc4 Bf6 63.f4 Kd6 64.g4 Ke6 65.h4 Bxh4 66.Kd4 Bg3 67.Ke4
+/= (0.66) Depth: 8 00:00:00 7kN
54.Rd7+ Kb6[] 55.Rd1 a4 56.Kd3 a3 57.e7 a2 58.e8Q b1Q+ 59.Kc4 Qxd1 60.Qb8+ Kc6 61.Qc8+ Kb6 62.Qb8+ Kc6 63.Qc8+ Kb6 64.Qb8+ Kc6 65.Qc8+ Kb6 66.Qb8+ Kc6 67.Qc8+ Kb6 68.Qb8+ Kc6 69.Qc8+
= (0.00) Depth: 9 00:00:00 62kN
54.Rd7+ Kb6[] 55.Rd1 a4 56.Kd3 a3 57.e7 a2 58.e8Q b1Q+ 59.Kc4 Qxd1 60.Qb8+ Kc6 61.Qc8+ Kb6 62.Qb8+ Kc6 63.Qc8+ Kb6 64.Qb8+ Kc6 65.Qc8+ Kb6 66.Qb8+ Kc6 67.Qc8+ Kb6 68.Qb8+ Kc6 69.Qc8+
= (0.00) Depth: 10 00:00:00 64kN
54.Rd7+ Kb6 55.Rd1 a4 56.Kd3 a3 57.e7 a2 58.e8Q b1Q+ 59.Kc4 Qxd1 60.Qb8+ Kc6 61.Qc8+ Kb6 62.Qb8+ Kc6 63.Qc8+ Kb6 64.Qb8+ Kc6 65.Qc8+ Kb6 66.Qb8+ Kc6 67.Qc8+ Kb6 68.Qb8+ Kc6 69.Qc8+
= (0.00) Depth: 11 00:00:00 87kN
54.Rd7+ Kb6 55.Rd1 a4 56.Kd3 Kb5 57.f4 a3 58.e7 Bxe7 59.Kc3 Ka4 60.Rd4+ Ka5 61.Rd1 b1Q 62.Rxb1 Ka4 63.Rxb3 a2 64.Kb2 a1Q+ 65.Kxa1 Kxb3 66.g4 Kc3 67.f5 Bg5 68.Kb1 Kd3 69.Kb2
= (-0.06) Depth: 12 00:00:00 184kN
54.Rd1 a4 55.Kd3 a3 56.e7 a2 57.e8Q[] b1Q+ 58.Rxb1 axb1Q+ 59.Kc4 Ba5 60.Qb5+ Bb6[] 61.Qd7+ Bc7 62.Qb5+ Bb6 63.Qd7+ Bc7 64.Qb5+ Bb6 65.Qd7+ Bc7 66.Qb5+ Bb6 67.Qd7+ Bc7 68.Qb5+ Bb6 69.Qd7+
= (0.00) Depth: 13 00:00:00 303kN, tb=2
54.Rd1 a4 55.Kd3 a3 56.e7 a2 57.e8Q[] b1Q+ 58.Rxb1 axb1Q+ 59.Kc4[] Ba5 60.Qb5+ Bb6[] 61.Qd7+ Bc7 62.Qb5+ Bb6 63.Qd7+ Bc7 64.Qb5+ Bb6 65.Qd7+ Bc7 66.Qb5+ Bb6 67.Qd7+ Bc7 68.Qb5+ Bb6 69.Qd7+
= (0.00) Depth: 14 00:00:00 351kN, tb=2
54.Rd1 a4 55.Kd3 a3 56.e7 a2 57.e8Q[] b1Q+ 58.Rxb1 axb1Q+ 59.Kc4[] Ba5 60.Qb5+ Bb6[] 61.Qd7+ Bc7 62.Qb5+ Bb6 63.Qd7+ Bc7 64.Qb5+ Bb6 65.Qd7+ Bc7 66.Qb5+ Bb6 67.Qd7+ Bc7 68.Qb5+ Bb6 69.Qd7+
= (0.00) Depth: 15 00:00:00 471kN, tb=4
54.Rd1 a4[] 55.Kd3 a3 56.e7 a2 57.e8Q[] b1Q+[] 58.Rxb1 axb1Q+ 59.Kc4[] Ba5 60.Qb5+ Bb6[] 61.Qd7+ Bc7 62.Qb5+ Bb6 63.Qd7+ Bc7 64.Qb5+ Bb6 65.Qd7+ Bc7 66.Qb5+ Bb6 67.Qd7+ Bc7 68.Qb5+ Bb6 69.Qd7+
= (0.00) Depth: 16 00:00:01 690kN, tb=10
54.Rd1 a4[] 55.Kd3 a3 56.e7 a2 57.e8Q[] b1Q+[] 58.Rxb1 axb1Q+ 59.Kc4[] Ba5 60.Qb5+ Bb6[] 61.Qd7+ Bc7 62.Qb5+ Bb6 63.Qd7+ Bc7 64.Qb5+ Bb6 65.Qd7+ Bc7 66.Qb5+ Bb6 67.Qd7+ Bc7 68.Qb5+ Bb6 69.Qd7+
= (0.00) Depth: 17 00:00:02 1175kN, tb=18
54.Rd1 a4[] 55.Kd3 a3 56.e7 a2 57.e8Q[] b1Q+[] 58.Rxb1 axb1Q+ 59.Kc4[] Ba5 60.Qb5+ Bb6[] 61.Qd7+ Bc7 62.Qb5+ Bb6 63.Qd7+ Bc7 64.Qb5+ Bb6 65.Qd7+ Bc7 66.Qb5+ Bb6 67.Qd7+ Bc7 68.Qb5+ Bb6 69.Qd7+
= (0.00) Depth: 18 00:00:03 1910kN, tb=44
54.Rd1 a4[] 55.Kd3 a3 56.e7 a2 57.e8Q[] b1Q+[] 58.Rxb1 axb1Q+ 59.Kc4[] Ba5 60.Qb5+ Bb6[] 61.Qd7+ Bc7 62.Qb5+ Bb6 63.Qd7+ Bc7 64.Qb5+ Bb6 65.Qd7+ Bc7 66.Qb5+ Bb6 67.Qd7+ Bc7 68.Qb5+ Bb6 69.Qd7+
= (0.00) Depth: 19 00:00:04 3150kN, tb=100
54.Rd1 a4[] 55.Kd3 a3 56.e7 a2 57.e8Q[] b1Q+[] 58.Rxb1 axb1Q+ 59.Kc4[] Ba5 60.Qb5+ Bb6[] 61.Qd7+ Bc7 62.Qb5+ Bb6 63.Qd7+ Bc7 64.Qb5+ Bb6 65.Qd7+ Bc7 66.Qb5+ Bb6 67.Qd7+ Bc7 68.Qb5+ Bb6 69.Qd7+
= (0.00) Depth: 20 00:00:08 5408kN, tb=185
54.Rd1 a4[] 55.Kd3[] a3 56.e7 a2 57.e8Q[] b1Q+[] 58.Rxb1 axb1Q+ 59.Kc4[] Ba5 60.Qb5+ Bb6[] 61.Qd7+ Bc7 62.Qb5+ Bb6 63.Qd7+ Bc7 64.Qb5+ Bb6 65.Qd7+ Bc7 66.Qb5+ Bb6 67.Qd7+ Bc7 68.Qb5+ Bb6 69.Qd7+
= (0.00) Depth: 21 00:00:14 9151kN, tb=334
54.Rd1 a4[] 55.Kd3[] a3 56.e7 a2 57.e8Q[] b1Q+[] 58.Rxb1 axb1Q+ 59.Kc4[] Ba5 60.Qb5+ Bb6[] 61.Qd7+ Bc7 62.Qb5+ Bb6 63.Qd7+ Bc7 64.Qb5+ Bb6 65.Qd7+ Bc7 66.Qb5+ Bb6 67.Qd7+ Bc7 68.Qb5+ Bb6 69.Qd7+
= (0.00) Depth: 22 00:00:23 15844kN, tb=599
54.Rd1 a4[] 55.Kd3[] a3 56.e7 a2 57.e8Q[] b1Q+[] 58.Rxb1 axb1Q+ 59.Kc4[] Ba5 60.Qb5+ Bb6[] 61.Qd7+ Bc7 62.Qb5+ Bb6 63.Qd7+ Bc7 64.Qb5+ Bb6 65.Qd7+ Bc7 66.Qb5+ Bb6 67.Qd7+ Bc7 68.Qb5+ Bb6 69.Qd7+
= (0.00) Depth: 23 00:00:39 25945kN, tb=1115
54.Rd1 a4[] 55.Kd3[] a3 56.e7 a2 57.e8Q[] b1Q+[] 58.Rxb1 axb1Q+ 59.Kc4[] Ba5 60.Qb5+ Bb6[] 61.Qd7+ Bc7 62.Qb5+ Bb6 63.Qd7+ Bc7 64.Qb5+ Bb6 65.Qd7+ Bc7 66.Qb5+ Bb6 67.Qd7+ Bc7 68.Qb5+ Bb6 69.Qd7+
= (0.00) Depth: 24 00:01:07 44076kN, tb=2052
(, Microsoft 17.02.2012)
And here is a clip analysis on 1 core after clearing hash to avoid the undetermenistic behavior of MP search
Deep Rybka 4.1 x64 2T - Zappa Mexico II 2T, Noomen2012 8 eng 2012
Analysis by Deep Rybka 4.1 x64: (1 core)
54.Rd1 a4 55.e7 Bxe7 56.Kd3 a3 57.Kc3 a2 58.Kxb2 Bf6+ 59.Kxb3 a1Q 60.Rxa1 Bxa1 61.Kc4 Kc6 62.g3
+/= (0.46) Depth: 6 00:00:00 2kN
54.Rd7+ Kc6 55.Rd1 a4 56.Kd3 a3 57.e7 Bxe7 58.Kc3 a2 59.Kxb2 Bf6+ 60.Kxb3 a1Q 61.Rxa1 Bxa1 62.Kc4 Be5 63.g4
+/= (0.58 !) Depth: 7 00:00:00 10kN
54.Rd7+ Kc6 55.Rd1 a4 56.Kd3 a3 57.e7 Bxe7 58.Kc3 a2 59.Kxb2 Bf6+ 60.Kxb3 a1Q 61.Rxa1 Bxa1 62.Kc4 Be5 63.g4
= (0.23) Depth: 7 00:00:00 10kN
54.Rd7+ Kc6 55.Rd1 a4 56.Kd3 a3 57.e7 Bxe7 58.Kc3 a2 59.Kxb2 Bf6+ 60.Kxb3 a1Q 61.Rxa1 Bxa1 62.Kc4 Be5 63.g4
+/= (0.35 !) Depth: 8 00:00:00 11kN
54.Rd7+ Kc6 55.Rd1 a4 56.Kd3 a3 57.e7 Bxe7 58.Kc3 a2 59.Kxb2 Bf6+ 60.Kxb3 a1Q 61.Rxa1 Bxa1 62.Kc4 Kd6 63.f4 Ke6 64.g4
+/= (0.50 !) Depth: 8 00:00:00 11kN
54.Rd7+ Kc6 55.Rd1 a4 56.Kd3 a3 57.e7 Bxe7 58.Kc3 a2 59.Kxb2 Bf6+ 60.Kxb3 a1Q 61.Rxa1 Bxa1 62.h4 Kd6 63.g4 Be5 64.g5 Ke6 65.h5 Kf7 66.g6+ Kg7 67.Kc4 Bf4 68.Kd5
+/= (0.66) Depth: 8 00:00:00 14kN
54.Rd7+ Kb6[] 55.Rd1 a4 56.Kd3 a3 57.e7 a2 58.e8Q b1Q+ 59.Kc4 Qxd1 60.Qb8+ Kc6 61.Qc8+ Kb6 62.Qb8+ Kc6 63.Qc8+ Kb6 64.Qb8+ Kc6 65.Qc8+ Kb6 66.Qb8+ Kc6 67.Qc8+ Kb6 68.Qb8+ Kc6 69.Qc8+
= (0.00) Depth: 9 00:00:00 83kN, tb=1
54.Rd7+ Kb6 55.Rd1 a4 56.Kd3 a3 57.e7 a2 58.e8Q b1Q+ 59.Kc4 Qxd1 60.Qb8+ Kc6 61.Qc8+ Kb6 62.Qb8+ Kc6 63.Qc8+ Kb6 64.Qb8+ Kc6 65.Qc8+ Kb6 66.Qb8+ Kc6 67.Qc8+ Kb6 68.Qb8+ Kc6 69.Qc8+
= (0.00) Depth: 10 00:00:00 88kN, tb=1
54.Rd7+ Kb6 55.Rd1 a4 56.Kd3 a3 57.e7 a2 58.e8Q b1Q+ 59.Kc4 Qxd1 60.Qb8+ Kc6 61.Qc8+ Kb6[] 62.Qb8+ Kc6 63.Qc8+ Kb6[] 64.Qb8+ Kc6 65.Qc8+ Kb6[] 66.Qb8+ Kc6 67.Qc8+ Kb6[] 68.Qb8+ Kc6 69.Qc8+
= (0.00) Depth: 11 00:00:00 103kN, tb=1
54.Rd7+ Kb6 55.Rd1 a4 56.Kd3 Kb5 57.f4 a3 58.e7 Bxe7 59.Kc3 Ka4 60.Rd4+ Ka5 61.Rd1 b1Q 62.Rxb1 Ka4 63.Rxb3 a2 64.Kb2 a1Q+ 65.Kxa1 Kxb3 66.g4 Kc3 67.g5 Kd3 68.g6 Bf6+ 69.Ka2
= (-0.10) Depth: 12 00:00:01 175kN, tb=1
54.Rd7+ Kb6 55.Rd1[] a4[] 56.Kd3 Kb5 57.f4 a3 58.e7 Bxe7 59.Kc3 Ka4 60.Rd4+ Ka5 61.Rd1 b1Q 62.Rxb1 Ka4 63.Rxb3 a2 64.Kb2 a1Q+ 65.Kxa1[] Kxb3 66.h4 Bxh4 67.g4 Kc4 68.g5 Kd5 69.Ka2
= (0.00) Depth: 13 00:00:02 259kN, tb=2
54.Rd7+ Kb6 55.Rd1[] a4[] 56.Kd3 Kb5 57.f4 a3 58.e7 Bxe7 59.Kc3 Ka4 60.Rd4+ Ka5 61.Rd1 b1Q 62.Rxb1 Ka4 63.Rxb3 a2 64.Kb2 a1Q+ 65.Kxa1[] Kxb3 66.h4 Bxh4 67.g4 Kc4 68.g5 Kd5 69.Ka2
= (0.00) Depth: 14 00:00:02 344kN, tb=3
54.Rd7+ Kb6 55.Rd1[] a4[] 56.Kd3 Kb5 57.f4 a3 58.e7 Bxe7 59.Kc3 Ka4 60.Rd4+ Ka5 61.Rd1 b1Q 62.Rxb1 Ka4 63.Rxb3 a2 64.Kb2 a1Q+ 65.Kxa1[] Kxb3 66.h4 Bxh4 67.g4 Kc4 68.g5 Kd5 69.Ka2
= (0.00) Depth: 15 00:00:02 410kN, tb=4
54.Rd7+ Kb6 55.Rd1[] a4[] 56.Kd3 Kb5 57.f4 a3 58.e7 Bxe7[] 59.Kc3[] Ka4 60.Rd4+[] Ka5 61.Rd1 b1Q 62.Rxb1 Ka4 63.Rxb3 a2 64.Kb2 a1Q+ 65.Kxa1[] Kxb3 66.h4 Bxh4 67.g4 Kc4 68.g5 Kd5 69.Ka2
= (0.00) Depth: 16 00:00:04 684kN, tb=9
54.Rd1 a4 55.Kd3 a3 56.e7 a2 57.e8Q b1Q+ 58.Rxb1 axb1Q+ 59.Kc4 Bd6 60.Qd7+ Bc7 61.Qb5+ Kc8 62.Qe8+ Kb7 63.Qb5+[] Kc8 64.Qe8+ Kb7 65.Qb5+[] Kc8 66.Qe8+ Kb7 67.Qb5+[] Kc8 68.Qe8+ Kb7 69.Qb5+[]
= (0.00) Depth: 17 00:00:11 1842kN, tb=102
54.Rd1 a4 55.Kd3 a3 56.e7 a2 57.e8Q b1Q+ 58.Rxb1 axb1Q+ 59.Kc4 Bd6 60.Qd7+ Bc7 61.Qb5+ Kc8 62.Qe8+ Kb7 63.Qb5+[] Kc8 64.Qe8+ Kb7 65.Qb5+[] Kc8 66.Qe8+ Kb7 67.Qb5+[] Kc8 68.Qe8+ Kb7 69.Qb5+[]
= (0.00) Depth: 18 00:00:17 2730kN, tb=214
54.Rd1 a4 55.Kd3 a3 56.e7 a2 57.e8Q b1Q+ 58.Rxb1 axb1Q+ 59.Kc4 Bd6 60.Qd7+ Bc7 61.Qb5+ Kc8 62.Qe8+ Kb7 63.Qb5+[] Kc8 64.Qe8+ Kb7 65.Qb5+[] Kc8 66.Qe8+ Kb7 67.Qb5+[] Kc8 68.Qe8+ Kb7 69.Qb5+[]
= (0.00) Depth: 19 00:00:25 4140kN, tb=394
54.Rd1 a4[] 55.Kd3 a3 56.e7 a2 57.e8Q[] b1Q+[] 58.Rxb1 axb1Q+ 59.Kc4[] Bd6 60.Qd7+ Bc7 61.Qb5+ Kc8 62.Qe8+ Kb7 63.Qb5+[] Kc8 64.Qe8+ Kb7 65.Qb5+[] Kc8 66.Qe8+ Kb7 67.Qb5+[] Kc8 68.Qe8+ Kb7 69.Qb5+[]
= (0.00) Depth: 20 00:00:40 6531kN, tb=774
54.Rd1 a4[] 55.Kd3 a3 56.e7 a2 57.e8Q[] b1Q+[] 58.Rxb1 axb1Q+ 59.Kc4[] Bd6 60.Qd7+ Bc7 61.Qb5+ Kc8 62.Qe8+ Kb7 63.Qb5+[] Kc8 64.Qe8+ Kb7 65.Qb5+[] Kc8 66.Qe8+ Kb7 67.Qb5+[] Kc8 68.Qe8+ Kb7 69.Qb5+[]
= (0.00) Depth: 21 00:01:04 10565kN, tb=1491
(, Microsoft 17.02.2012)
Houdini also seems to find it after a little more time on 8 cores, but seems to take a little long on single core
Deep Rybka 4.1 x64 2T - Zappa Mexico II 2T, Noomen2012 8 eng 2012
Analysis by Houdini 2.0c Pro x64: (8 cores)
54.Rd1 a4 55.Kd3 Kc6 56.Kc4 Be7 57.f4
+- (1.61) Depth: 6/16 00:00:00 2kN
54.Rd1 Kc6 55.Rb1 a4 56.Rxb2 Kd5 57.Kd3 Kxe6 58.g4
+/- (1.24) Depth: 7/16 00:00:00 3kN
54.Rd1 a4 55.e7 Bxe7 56.Kd3 a3 57.Kc3 a2 58.Kxb2 Bf6+ 59.Kxb3 a1Q 60.Rxa1 Bxa1
= (0.02) Depth: 8/24 00:00:00 34kN
54.Rd1
= (0.07 !) Depth: 9/24 00:00:00 36kN
54.Rd1 a4 55.e7 Bxe7 56.Kd3 a3 57.Kc3 a2 58.Kxb2 Bf6+ 59.Kxb3 a1Q 60.Rxa1 Bxa1 61.h4 Kc6 62.h5 Bg7 63.g4
= (0.13) Depth: 9/24 00:00:00 40kN
54.Rd1
= (0.17 !) Depth: 10/26 00:00:00 53kN
54.Rd1
+/= (0.26 !) Depth: 10/29 00:00:00 57kN
54.Rd1 a4
= (0.08 --) Depth: 10/29 00:00:00 98kN
54.Rd1 a4
= (0.00 --) Depth: 10/29 00:00:00 124kN
54.Rd1 a4
= (-0.21 --) Depth: 10/29 00:00:00 156kN
54.Rd7+ Kc6 55.Rd1 a4 56.Kd3 a3 57.e7 Bxe7 58.Kc3 a2 59.Kxb2 Bf6+ 60.Kxb3 a1Q 61.Rxa1 Bxa1 62.g4 Bf6 63.h4 Bxh4 64.g5 Bxg5 65.Kc4 Bf6
= (0.02) Depth: 10/30 00:00:00 172kN
54.Rd7+ Kc6 55.Rd1 a4 56.Kd3 a3 57.e7 Bxe7 58.Kc3 a2 59.Kxb2 Bf6+ 60.Kxb3 a1Q 61.Rxa1 Bxa1 62.g4 Bf6 63.h4 Bxh4 64.g5 Bxg5 65.Kc4 Bf6
= (0.02) Depth: 11/30 00:00:00 174kN
54.Rd7+ Kc6 55.Rd1 a4 56.Kd3 a3 57.e7 Bxe7 58.Kc3 a2 59.Kxb2 Bf6+ 60.Kxb3 a1Q 61.Rxa1 Bxa1 62.g4 Kd6 63.h4 Ke6 64.g5 Kf5 65.f4 Kxf4 66.Kc4 Kf5
= (0.02) Depth: 12/30 00:00:00 248kN
54.Rd7+ Kb6 55.Rd1 a4 56.Kd3 a3 57.e7 a2 58.e8Q b1Q+ 59.Kc4 Qxd1 60.Qb8+ Kc6 61.Qc8+ Kb6 62.Qb8+ Kc6
= (0.00) Depth: 13/34 00:00:00 481kN
54.Rd7+ Kb6 55.Rd1 a4 56.Kd3 a3 57.e7 a2 58.e8Q b1Q+ 59.Kc4 Qxd1 60.Qb8+ Kc6 61.Qc8+ Kb6 62.Qb8+ Kc6
= (0.00) Depth: 14/34 00:00:00 665kN
54.Rd7+ Kb6 55.Rd1 a4 56.Kd3 a3 57.e7 a2 58.e8Q b1Q+ 59.Kc4 Qxd1 60.Qb8+ Kc6 61.Qc8+ Kb6 62.Qb8+ Kc6
= (0.00) Depth: 15/38 00:00:00 992kN
54.Rd7+ Kb6 55.Rd1 a4 56.Kd3 a3 57.e7 a2 58.e8Q b1Q+ 59.Kc4 Qxd1 60.Qb8+ Kc6 61.Qc8+ Kb6 62.Qb8+ Kc6
= (0.00) Depth: 16/42 00:00:00 1811kN
54.Rd7+ Kb6 55.Rd1 a4 56.Kd3 a3 57.e7 a2 58.e8Q b1Q+ 59.Kc4 Qxd1 60.Qb8+ Kc6 61.Qc8+ Kb6 62.Qb8+ Kc6
= (0.00) Depth: 17/42 00:00:00 3071kN
54.Rd7+ Kb6 55.Rd1 a4 56.Kd3 a3 57.e7 a2 58.e8Q b1Q+ 59.Kc4 Qxd1 60.Qb8+ Kc6 61.Qc8+ Kb6 62.Qb8+ Kc6
= (0.00) Depth: 18/45 00:00:00 5264kN
54.Rd7+ Kb6 55.Rd1 a4 56.Kd3 a3 57.e7 a2 58.e8Q b1Q+ 59.Kc4 Qxd1 60.Qb8+ Kc6 61.Qc8+ Kb6 62.Qb8+ Kc6
= (0.00) Depth: 19/49 00:00:00 9721kN
54.Rd7+ Kb6 55.Rd1 a4 56.Kd3 a3 57.e7 a2 58.e8Q b1Q+ 59.Kc4 Qxd1 60.Qb8+ Kc6 61.Qc8+ Kb6 62.Qb8+ Kc6
= (0.00) Depth: 20/51 00:00:01 15466kN
54.Rd7+ Kb6 55.Rd1 a4 56.Kd3 a3 57.e7 a2 58.e8Q b1Q+ 59.Kc4 Qxd1 60.Qb8+ Kc6 61.Qc8+ Kb6 62.Qb8+ Kc6
= (0.00) Depth: 21/54 00:00:01 25161kN, tb=2
54.Rd7+ Kb6 55.Rd1 a4 56.Kd3 a3 57.e7 a2 58.e8Q b1Q+ 59.Kc4 Qxd1 60.Qb8+ Kc6 61.Qc8+ Kb6 62.Qb8+ Kc6
= (0.00) Depth: 22/55 00:00:03 43049kN, tb=13
54.Rd7+ Kb6
= (-0.04 --) Depth: 23/57 00:00:08 124mN, tb=45
54.Rd1 a4 55.Kd3 a3 56.e7 a2 57.e8Q b1Q+ 58.Rxb1 axb1Q+ 59.Kc4 Ba5 60.Qb5+ Bb6 61.Qd5+ Kb8 62.Qd6+ Kb7 63.Qd5+ Kb8
= (0.00) Depth: 23/57 00:00:10 166mN, tb=174
54.Rd1 a4 55.Kd3 a3 56.e7 a2 57.e8Q b1Q+ 58.Rxb1 axb1Q+ 59.Kc4 Ba5 60.Qb5+ Bb6 61.Qd5+ Kb8 62.Qd6+ Kb7 63.Qd5+ Kb8
= (0.00) Depth: 24/57 00:00:13 209mN, tb=281
54.Rd1 a4 55.Kd3 a3 56.e7 a2 57.e8Q b1Q+ 58.Rxb1 axb1Q+ 59.Kc4 Ba5 60.Qb5+ Bb6 61.Qd5+ Kb8 62.Qd6+ Kb7 63.Qd5+ Kb8
= (0.00) Depth: 25/59 00:00:17 282mN, tb=474
54.Rd1 a4 55.Kd3 a3 56.e7 a2 57.e8Q b1Q+ 58.Rxb1 axb1Q+ 59.Kc4 Ba5 60.Qb5+ Bb6 61.Qd5+ Kb8 62.Qd6+ Kb7 63.Qd5+ Kb8
= (0.00) Depth: 26/60 00:00:24 410mN, tb=799
54.Rd1 a4 55.Kd3 a3 56.e7 a2 57.e8Q b1Q+ 58.Rxb1 axb1Q+ 59.Kc4 Ba5 60.Qb5+ Bb6 61.Qd5+ Kb8 62.Qd6+ Kb7 63.Qd5+ Kb8
= (0.00) Depth: 27/65 00:00:49 861mN, tb=2477
(, Microsoft 17.02.2012)
Deep Rybka 4.1 x64 2T - Zappa Mexico II 2T, Noomen2012 8 eng 2012
Analysis by Houdini 2.0c Pro x64: (1 core)
54.Rd1 a4 55.Kd3 Kc6 56.Kc4 Be7 57.f4
+- (1.61) Depth: 6/16 00:00:00 2kN
54.Rd1 Kc6 55.Rb1 a4 56.Rxb2 Kd5 57.Kd3 Kxe6 58.g4
+/- (1.24) Depth: 7/19 00:00:00 4kN
54.Rd1 a4 55.e7 Bxe7 56.Kd3 a3 57.Kc3 a2 58.Kxb2 Bf6+ 59.Kxb3 a1Q 60.Rxa1 Bxa1
= (0.02) Depth: 8/24 00:00:00 31kN
54.Rd1
= (0.07 !) Depth: 9/24 00:00:00 39kN
54.Rd1 a4 55.h4 Kc6 56.Kd3 a3 57.e7 Bxe7 58.Kc3 a2 59.Kxb2 Bf6+ 60.Kxb3 a1Q 61.Rxa1 Bxa1 62.h5
= (0.09) Depth: 9/24 00:00:00 42kN
54.Rd1 a4 55.h4 Kc6 56.Kd3 a3 57.e7 Bxe7 58.Kc3 a2 59.Kxb2 Bf6+ 60.Kxb3 a1Q 61.Rxa1 Bxa1 62.h5 Bg7 63.g4
= (0.13) Depth: 10/25 00:00:00 58kN
54.Rd1 a4
= (0.08 --) Depth: 11/27 00:00:00 88kN
54.Rd1 a4
= (0.00 --) Depth: 11/27 00:00:00 119kN
54.Rd1 a4
= (-0.21 --) Depth: 11/27 00:00:00 155kN
54.Rd7+ Kc6 55.Rd1 a4 56.Kd3 a3 57.e7 Bxe7 58.Kc3 a2 59.Kxb2 Bf6+ 60.Kxb3 a1Q 61.Rxa1 Bxa1 62.Kc4 Bg7 63.h4 Kd6 64.f4 Ke6 65.g4 Kf6 66.h5 Bf8 67.g5+ Kf5
= (0.03) Depth: 11/32 00:00:00 199kN
54.Rd7+ Kc6 55.Rd1 a4 56.Kd3 a3 57.e7 Bxe7 58.Kc3 a2 59.Kxb2 Bf6+ 60.Kxb3 a1Q 61.Rxa1 Bxa1 62.Kc4 Bg7 63.h4 Kd6 64.f4 Ke6 65.g4 Kf6 66.h5 Bf8 67.Kd5
= (0.08) Depth: 12/33 00:00:00 268kN
54.Rd7+ Kc6
= (0.03 --) Depth: 13/37 00:00:00 521kN
54.Rd7+ Kb6 55.Rd1 a4 56.Kd3 a3 57.e7 a2 58.e8Q b1Q+ 59.Kc4 Qxd1 60.Qb8+ Ka6 61.Qa8+ Kb6 62.Qb8+ Ka6
= (0.00) Depth: 13/37 00:00:00 631kN
54.Rd7+ Kb6 55.Rd1 a4 56.Kd3 a3 57.e7 a2 58.e8Q b1Q+ 59.Kc4 Qxd1 60.Qb8+ Ka6 61.Qa8+ Kb6 62.Qb8+ Ka6
= (0.00) Depth: 14/40 00:00:00 865kN
54.Rd7+ Kb6 55.Rd1 a4 56.Kd3 a3 57.e7 a2 58.e8Q b1Q+ 59.Kc4 Qxd1 60.Qb8+ Ka6 61.Qa8+ Kb6 62.Qb8+ Ka6
= (0.00) Depth: 15/41 00:00:00 1250kN
54.Rd7+ Kb6 55.Rd1 a4 56.Kd3 a3 57.e7 a2 58.e8Q b1Q+ 59.Kc4 Qxd1 60.Qb8+ Ka6 61.Qa8+ Kb6 62.Qb8+ Ka6
= (0.00) Depth: 16/44 00:00:00 2000kN
54.Rd7+ Kb6 55.Rd1 a4 56.Kd3 a3 57.e7 a2 58.e8Q b1Q+ 59.Kc4 Qxd1 60.Qb8+ Ka6 61.Qa8+ Kb6 62.Qb8+ Ka6
= (0.00) Depth: 17/46 00:00:01 3332kN
54.Rd7+ Kb6 55.Rd1 a4 56.Kd3 a3 57.e7 a2 58.e8Q b1Q+ 59.Kc4 Qxd1 60.Qb8+ Ka6 61.Qa8+ Kb6 62.Qb8+ Ka6
= (0.00) Depth: 18/49 00:00:01 5639kN
54.Rd7+ Kb6 55.Rd1 a4 56.Kd3 a3 57.e7 a2 58.e8Q b1Q+ 59.Kc4 Qxd1 60.Qb8+ Ka6 61.Qa8+ Kb6 62.Qb8+ Ka6
= (0.00) Depth: 19/51 00:00:02 8303kN
54.Rd7+ Kb6 55.Rd1 a4 56.Kd3 a3 57.e7 a2 58.e8Q b1Q+ 59.Kc4 Qxd1 60.Qb8+ Ka6 61.Qa8+ Kb6 62.Qb8+ Ka6
= (0.00) Depth: 20/53 00:00:04 13578kN
54.Rd7+ Kb6 55.Rd1 a4 56.Kd3 a3 57.e7 a2 58.e8Q b1Q+ 59.Kc4 Qxd1 60.Qb8+ Ka6 61.Qa8+ Kb6 62.Qb8+ Ka6
= (0.00) Depth: 21/55 00:00:07 21832kN, tb=1
54.Rd7+ Kb6 55.Rd1 a4 56.Kd3 a3 57.e7 a2 58.e8Q b1Q+ 59.Kc4 Qxd1 60.Qb8+ Ka6 61.Qa8+ Kb6 62.Qb8+ Ka6
= (0.00) Depth: 22/57 00:00:11 35663kN, tb=7
54.Rd7+ Kb6 55.Rd1 a4 56.Kd3 a3 57.e7 a2 58.e8Q b1Q+ 59.Kc4 Qxd1 60.Qb8+ Ka6 61.Qa8+ Kb6 62.Qb8+ Ka6
= (0.00) Depth: 23/64 00:00:24 75579kN, tb=20
54.Rd7+ Kb6 55.Rd1 a4 56.Kd3 a3 57.e7 a2 58.e8Q b1Q+ 59.Kc4 Qxd1 60.Qb8+ Ka6 61.Qa8+ Kb6 62.Qb8+ Ka6
= (0.00) Depth: 24/64 00:00:38 118mN, tb=46
54.Rd7+ Kb6 55.Rd1 a4 56.Kd3 a3 57.e7 a2 58.e8Q b1Q+ 59.Kc4 Qxd1 60.Qb8+ Ka6 61.Qa8+ Kb6 62.Qb8+ Ka6
= (0.00) Depth: 25/64 00:01:09 206mN, tb=100
54.Rd7+ Kb6
= (-0.04 --) Depth: 26/72 00:02:25 434mN, tb=152
54.Rd1 a4 55.Kd3 a3 56.e7 a2 57.e8Q b1Q+ 58.Rxb1 axb1Q+ 59.Kc4 Ba5 60.Qb5+ Bb6 61.Qd7+ Kb8 62.Qe8+ Ka7 63.Qd7+ Ka6 64.Qa4+ Kb7 65.Qd7+ Kb8
= (0.00) Depth: 26/72 00:04:10 752mN, tb=814
(, Microsoft 17.02.2012)
So this might be more to do with you not having your EGTB's installed, or maybe some hash corruption due to LP. If you are using LP make sure that during testing you use engine vs engine match rather than a tourney. Loading and unloading of engine with LP creates huge memory fragmentation under Win OS. You should always have at least the 5 pieces loaded as they take minimal space on your HDD and Rybka 4.1 is tested with them on.
One thing to note is that Rybka and all Rybka 3 based engines (Ippolit, Ivanhoe, Houdini etc..) are strong because they prune heavily and choose a few good moves to search through. So you can say that they don't have a deep comprehensive search of all moves, this makes them focus on stronger moves and in general that makes them incredibly strong ... of course you will find some positions where it backfires (like here), but over all the ELO strength is dramatically better.
I think even at this point the game is drawn
8/1k6/4P3/p7/1b1R4/1p3P1P/1p2K1P1/8 w - - 0 54
Interesting position here because it is a race to see who queens first. One missed tempo and it is game over so deep accurate calculation is necessary. I think Rd1! draws and anything else loses and here Rybka did not see it, or at least not quickly and neither does Houdini. I am pretty sure with more time they would see it though.
Actually now that I look at it, Rybka 4.1 does see it quite quickly on my Octa at 4 Ghz. Here is a clip analysis at 1 minute
Deep Rybka 4.1 x64 2T - Zappa Mexico II 2T, Noomen2012 8 eng 2012
8/1k6/4P3/p7/1b1R4/1p3P1P/1p2K1P1/8 w - - 0 1
Analysis by Deep Rybka 4.1 x64: (8 core)
54.Rd1 a4 55.e7 Bxe7 56.Kd3 a3 57.Kc3 a2 58.Kxb2 Bf6+ 59.Kxb3 a1Q 60.Rxa1 Bxa1 61.f4 Bf6 62.Kc4
+/= (0.49) Depth: 6 00:00:00 2kN
54.Rd7+ Kc6 55.Rd1 a4 56.e7 Bxe7 57.Kd3 Kc5 58.g3 Kb4 59.Re1 a3 60.Re4+ Kb5 61.Re5+
+/= (0.28) Depth: 7 00:00:00 6kN
54.Rd7+ Kc6 55.Rd1 a4 56.e7 Bxe7 57.Kd3 Kc5 58.g3 Kb4 59.Re1 a3 60.Re4+ Kb5 61.Re5+ Ka4 62.Re1
+/= (0.40 !) Depth: 8 00:00:00 7kN
54.Rd7+ Kc6 55.Rd1 a4 56.e7 Bxe7 57.Kd3 Kc5 58.Re1 Kb4 59.h4 a3 60.Re4+ Kc5 61.Re5+ Kc6 62.Re1 a2 63.Re6+
+/= (0.55 !) Depth: 8 00:00:00 7kN
54.Rd7+ Kc6 55.Rd1 a4 56.Kd3 a3 57.e7 Bxe7 58.Kc3 a2 59.Kxb2 Bf6+ 60.Kxb3 a1Q 61.Rxa1 Bxa1 62.Kc4 Bf6 63.f4 Kd6 64.g4 Ke6 65.h4 Bxh4 66.Kd4 Bg3 67.Ke4
+/= (0.66) Depth: 8 00:00:00 7kN
54.Rd7+ Kb6[] 55.Rd1 a4 56.Kd3 a3 57.e7 a2 58.e8Q b1Q+ 59.Kc4 Qxd1 60.Qb8+ Kc6 61.Qc8+ Kb6 62.Qb8+ Kc6 63.Qc8+ Kb6 64.Qb8+ Kc6 65.Qc8+ Kb6 66.Qb8+ Kc6 67.Qc8+ Kb6 68.Qb8+ Kc6 69.Qc8+
= (0.00) Depth: 9 00:00:00 62kN
54.Rd7+ Kb6[] 55.Rd1 a4 56.Kd3 a3 57.e7 a2 58.e8Q b1Q+ 59.Kc4 Qxd1 60.Qb8+ Kc6 61.Qc8+ Kb6 62.Qb8+ Kc6 63.Qc8+ Kb6 64.Qb8+ Kc6 65.Qc8+ Kb6 66.Qb8+ Kc6 67.Qc8+ Kb6 68.Qb8+ Kc6 69.Qc8+
= (0.00) Depth: 10 00:00:00 64kN
54.Rd7+ Kb6 55.Rd1 a4 56.Kd3 a3 57.e7 a2 58.e8Q b1Q+ 59.Kc4 Qxd1 60.Qb8+ Kc6 61.Qc8+ Kb6 62.Qb8+ Kc6 63.Qc8+ Kb6 64.Qb8+ Kc6 65.Qc8+ Kb6 66.Qb8+ Kc6 67.Qc8+ Kb6 68.Qb8+ Kc6 69.Qc8+
= (0.00) Depth: 11 00:00:00 87kN
54.Rd7+ Kb6 55.Rd1 a4 56.Kd3 Kb5 57.f4 a3 58.e7 Bxe7 59.Kc3 Ka4 60.Rd4+ Ka5 61.Rd1 b1Q 62.Rxb1 Ka4 63.Rxb3 a2 64.Kb2 a1Q+ 65.Kxa1 Kxb3 66.g4 Kc3 67.f5 Bg5 68.Kb1 Kd3 69.Kb2
= (-0.06) Depth: 12 00:00:00 184kN
54.Rd1 a4 55.Kd3 a3 56.e7 a2 57.e8Q[] b1Q+ 58.Rxb1 axb1Q+ 59.Kc4 Ba5 60.Qb5+ Bb6[] 61.Qd7+ Bc7 62.Qb5+ Bb6 63.Qd7+ Bc7 64.Qb5+ Bb6 65.Qd7+ Bc7 66.Qb5+ Bb6 67.Qd7+ Bc7 68.Qb5+ Bb6 69.Qd7+
= (0.00) Depth: 13 00:00:00 303kN, tb=2
54.Rd1 a4 55.Kd3 a3 56.e7 a2 57.e8Q[] b1Q+ 58.Rxb1 axb1Q+ 59.Kc4[] Ba5 60.Qb5+ Bb6[] 61.Qd7+ Bc7 62.Qb5+ Bb6 63.Qd7+ Bc7 64.Qb5+ Bb6 65.Qd7+ Bc7 66.Qb5+ Bb6 67.Qd7+ Bc7 68.Qb5+ Bb6 69.Qd7+
= (0.00) Depth: 14 00:00:00 351kN, tb=2
54.Rd1 a4 55.Kd3 a3 56.e7 a2 57.e8Q[] b1Q+ 58.Rxb1 axb1Q+ 59.Kc4[] Ba5 60.Qb5+ Bb6[] 61.Qd7+ Bc7 62.Qb5+ Bb6 63.Qd7+ Bc7 64.Qb5+ Bb6 65.Qd7+ Bc7 66.Qb5+ Bb6 67.Qd7+ Bc7 68.Qb5+ Bb6 69.Qd7+
= (0.00) Depth: 15 00:00:00 471kN, tb=4
54.Rd1 a4[] 55.Kd3 a3 56.e7 a2 57.e8Q[] b1Q+[] 58.Rxb1 axb1Q+ 59.Kc4[] Ba5 60.Qb5+ Bb6[] 61.Qd7+ Bc7 62.Qb5+ Bb6 63.Qd7+ Bc7 64.Qb5+ Bb6 65.Qd7+ Bc7 66.Qb5+ Bb6 67.Qd7+ Bc7 68.Qb5+ Bb6 69.Qd7+
= (0.00) Depth: 16 00:00:01 690kN, tb=10
54.Rd1 a4[] 55.Kd3 a3 56.e7 a2 57.e8Q[] b1Q+[] 58.Rxb1 axb1Q+ 59.Kc4[] Ba5 60.Qb5+ Bb6[] 61.Qd7+ Bc7 62.Qb5+ Bb6 63.Qd7+ Bc7 64.Qb5+ Bb6 65.Qd7+ Bc7 66.Qb5+ Bb6 67.Qd7+ Bc7 68.Qb5+ Bb6 69.Qd7+
= (0.00) Depth: 17 00:00:02 1175kN, tb=18
54.Rd1 a4[] 55.Kd3 a3 56.e7 a2 57.e8Q[] b1Q+[] 58.Rxb1 axb1Q+ 59.Kc4[] Ba5 60.Qb5+ Bb6[] 61.Qd7+ Bc7 62.Qb5+ Bb6 63.Qd7+ Bc7 64.Qb5+ Bb6 65.Qd7+ Bc7 66.Qb5+ Bb6 67.Qd7+ Bc7 68.Qb5+ Bb6 69.Qd7+
= (0.00) Depth: 18 00:00:03 1910kN, tb=44
54.Rd1 a4[] 55.Kd3 a3 56.e7 a2 57.e8Q[] b1Q+[] 58.Rxb1 axb1Q+ 59.Kc4[] Ba5 60.Qb5+ Bb6[] 61.Qd7+ Bc7 62.Qb5+ Bb6 63.Qd7+ Bc7 64.Qb5+ Bb6 65.Qd7+ Bc7 66.Qb5+ Bb6 67.Qd7+ Bc7 68.Qb5+ Bb6 69.Qd7+
= (0.00) Depth: 19 00:00:04 3150kN, tb=100
54.Rd1 a4[] 55.Kd3 a3 56.e7 a2 57.e8Q[] b1Q+[] 58.Rxb1 axb1Q+ 59.Kc4[] Ba5 60.Qb5+ Bb6[] 61.Qd7+ Bc7 62.Qb5+ Bb6 63.Qd7+ Bc7 64.Qb5+ Bb6 65.Qd7+ Bc7 66.Qb5+ Bb6 67.Qd7+ Bc7 68.Qb5+ Bb6 69.Qd7+
= (0.00) Depth: 20 00:00:08 5408kN, tb=185
54.Rd1 a4[] 55.Kd3[] a3 56.e7 a2 57.e8Q[] b1Q+[] 58.Rxb1 axb1Q+ 59.Kc4[] Ba5 60.Qb5+ Bb6[] 61.Qd7+ Bc7 62.Qb5+ Bb6 63.Qd7+ Bc7 64.Qb5+ Bb6 65.Qd7+ Bc7 66.Qb5+ Bb6 67.Qd7+ Bc7 68.Qb5+ Bb6 69.Qd7+
= (0.00) Depth: 21 00:00:14 9151kN, tb=334
54.Rd1 a4[] 55.Kd3[] a3 56.e7 a2 57.e8Q[] b1Q+[] 58.Rxb1 axb1Q+ 59.Kc4[] Ba5 60.Qb5+ Bb6[] 61.Qd7+ Bc7 62.Qb5+ Bb6 63.Qd7+ Bc7 64.Qb5+ Bb6 65.Qd7+ Bc7 66.Qb5+ Bb6 67.Qd7+ Bc7 68.Qb5+ Bb6 69.Qd7+
= (0.00) Depth: 22 00:00:23 15844kN, tb=599
54.Rd1 a4[] 55.Kd3[] a3 56.e7 a2 57.e8Q[] b1Q+[] 58.Rxb1 axb1Q+ 59.Kc4[] Ba5 60.Qb5+ Bb6[] 61.Qd7+ Bc7 62.Qb5+ Bb6 63.Qd7+ Bc7 64.Qb5+ Bb6 65.Qd7+ Bc7 66.Qb5+ Bb6 67.Qd7+ Bc7 68.Qb5+ Bb6 69.Qd7+
= (0.00) Depth: 23 00:00:39 25945kN, tb=1115
54.Rd1 a4[] 55.Kd3[] a3 56.e7 a2 57.e8Q[] b1Q+[] 58.Rxb1 axb1Q+ 59.Kc4[] Ba5 60.Qb5+ Bb6[] 61.Qd7+ Bc7 62.Qb5+ Bb6 63.Qd7+ Bc7 64.Qb5+ Bb6 65.Qd7+ Bc7 66.Qb5+ Bb6 67.Qd7+ Bc7 68.Qb5+ Bb6 69.Qd7+
= (0.00) Depth: 24 00:01:07 44076kN, tb=2052
(, Microsoft 17.02.2012)
And here is a clip analysis on 1 core after clearing hash to avoid the undetermenistic behavior of MP search
Deep Rybka 4.1 x64 2T - Zappa Mexico II 2T, Noomen2012 8 eng 2012
8/1k6/4P3/p7/1b1R4/1p3P1P/1p2K1P1/8 w - - 0 1
Analysis by Deep Rybka 4.1 x64: (1 core)
54.Rd1 a4 55.e7 Bxe7 56.Kd3 a3 57.Kc3 a2 58.Kxb2 Bf6+ 59.Kxb3 a1Q 60.Rxa1 Bxa1 61.Kc4 Kc6 62.g3
+/= (0.46) Depth: 6 00:00:00 2kN
54.Rd7+ Kc6 55.Rd1 a4 56.Kd3 a3 57.e7 Bxe7 58.Kc3 a2 59.Kxb2 Bf6+ 60.Kxb3 a1Q 61.Rxa1 Bxa1 62.Kc4 Be5 63.g4
+/= (0.58 !) Depth: 7 00:00:00 10kN
54.Rd7+ Kc6 55.Rd1 a4 56.Kd3 a3 57.e7 Bxe7 58.Kc3 a2 59.Kxb2 Bf6+ 60.Kxb3 a1Q 61.Rxa1 Bxa1 62.Kc4 Be5 63.g4
= (0.23) Depth: 7 00:00:00 10kN
54.Rd7+ Kc6 55.Rd1 a4 56.Kd3 a3 57.e7 Bxe7 58.Kc3 a2 59.Kxb2 Bf6+ 60.Kxb3 a1Q 61.Rxa1 Bxa1 62.Kc4 Be5 63.g4
+/= (0.35 !) Depth: 8 00:00:00 11kN
54.Rd7+ Kc6 55.Rd1 a4 56.Kd3 a3 57.e7 Bxe7 58.Kc3 a2 59.Kxb2 Bf6+ 60.Kxb3 a1Q 61.Rxa1 Bxa1 62.Kc4 Kd6 63.f4 Ke6 64.g4
+/= (0.50 !) Depth: 8 00:00:00 11kN
54.Rd7+ Kc6 55.Rd1 a4 56.Kd3 a3 57.e7 Bxe7 58.Kc3 a2 59.Kxb2 Bf6+ 60.Kxb3 a1Q 61.Rxa1 Bxa1 62.h4 Kd6 63.g4 Be5 64.g5 Ke6 65.h5 Kf7 66.g6+ Kg7 67.Kc4 Bf4 68.Kd5
+/= (0.66) Depth: 8 00:00:00 14kN
54.Rd7+ Kb6[] 55.Rd1 a4 56.Kd3 a3 57.e7 a2 58.e8Q b1Q+ 59.Kc4 Qxd1 60.Qb8+ Kc6 61.Qc8+ Kb6 62.Qb8+ Kc6 63.Qc8+ Kb6 64.Qb8+ Kc6 65.Qc8+ Kb6 66.Qb8+ Kc6 67.Qc8+ Kb6 68.Qb8+ Kc6 69.Qc8+
= (0.00) Depth: 9 00:00:00 83kN, tb=1
54.Rd7+ Kb6 55.Rd1 a4 56.Kd3 a3 57.e7 a2 58.e8Q b1Q+ 59.Kc4 Qxd1 60.Qb8+ Kc6 61.Qc8+ Kb6 62.Qb8+ Kc6 63.Qc8+ Kb6 64.Qb8+ Kc6 65.Qc8+ Kb6 66.Qb8+ Kc6 67.Qc8+ Kb6 68.Qb8+ Kc6 69.Qc8+
= (0.00) Depth: 10 00:00:00 88kN, tb=1
54.Rd7+ Kb6 55.Rd1 a4 56.Kd3 a3 57.e7 a2 58.e8Q b1Q+ 59.Kc4 Qxd1 60.Qb8+ Kc6 61.Qc8+ Kb6[] 62.Qb8+ Kc6 63.Qc8+ Kb6[] 64.Qb8+ Kc6 65.Qc8+ Kb6[] 66.Qb8+ Kc6 67.Qc8+ Kb6[] 68.Qb8+ Kc6 69.Qc8+
= (0.00) Depth: 11 00:00:00 103kN, tb=1
54.Rd7+ Kb6 55.Rd1 a4 56.Kd3 Kb5 57.f4 a3 58.e7 Bxe7 59.Kc3 Ka4 60.Rd4+ Ka5 61.Rd1 b1Q 62.Rxb1 Ka4 63.Rxb3 a2 64.Kb2 a1Q+ 65.Kxa1 Kxb3 66.g4 Kc3 67.g5 Kd3 68.g6 Bf6+ 69.Ka2
= (-0.10) Depth: 12 00:00:01 175kN, tb=1
54.Rd7+ Kb6 55.Rd1[] a4[] 56.Kd3 Kb5 57.f4 a3 58.e7 Bxe7 59.Kc3 Ka4 60.Rd4+ Ka5 61.Rd1 b1Q 62.Rxb1 Ka4 63.Rxb3 a2 64.Kb2 a1Q+ 65.Kxa1[] Kxb3 66.h4 Bxh4 67.g4 Kc4 68.g5 Kd5 69.Ka2
= (0.00) Depth: 13 00:00:02 259kN, tb=2
54.Rd7+ Kb6 55.Rd1[] a4[] 56.Kd3 Kb5 57.f4 a3 58.e7 Bxe7 59.Kc3 Ka4 60.Rd4+ Ka5 61.Rd1 b1Q 62.Rxb1 Ka4 63.Rxb3 a2 64.Kb2 a1Q+ 65.Kxa1[] Kxb3 66.h4 Bxh4 67.g4 Kc4 68.g5 Kd5 69.Ka2
= (0.00) Depth: 14 00:00:02 344kN, tb=3
54.Rd7+ Kb6 55.Rd1[] a4[] 56.Kd3 Kb5 57.f4 a3 58.e7 Bxe7 59.Kc3 Ka4 60.Rd4+ Ka5 61.Rd1 b1Q 62.Rxb1 Ka4 63.Rxb3 a2 64.Kb2 a1Q+ 65.Kxa1[] Kxb3 66.h4 Bxh4 67.g4 Kc4 68.g5 Kd5 69.Ka2
= (0.00) Depth: 15 00:00:02 410kN, tb=4
54.Rd7+ Kb6 55.Rd1[] a4[] 56.Kd3 Kb5 57.f4 a3 58.e7 Bxe7[] 59.Kc3[] Ka4 60.Rd4+[] Ka5 61.Rd1 b1Q 62.Rxb1 Ka4 63.Rxb3 a2 64.Kb2 a1Q+ 65.Kxa1[] Kxb3 66.h4 Bxh4 67.g4 Kc4 68.g5 Kd5 69.Ka2
= (0.00) Depth: 16 00:00:04 684kN, tb=9
54.Rd1 a4 55.Kd3 a3 56.e7 a2 57.e8Q b1Q+ 58.Rxb1 axb1Q+ 59.Kc4 Bd6 60.Qd7+ Bc7 61.Qb5+ Kc8 62.Qe8+ Kb7 63.Qb5+[] Kc8 64.Qe8+ Kb7 65.Qb5+[] Kc8 66.Qe8+ Kb7 67.Qb5+[] Kc8 68.Qe8+ Kb7 69.Qb5+[]
= (0.00) Depth: 17 00:00:11 1842kN, tb=102
54.Rd1 a4 55.Kd3 a3 56.e7 a2 57.e8Q b1Q+ 58.Rxb1 axb1Q+ 59.Kc4 Bd6 60.Qd7+ Bc7 61.Qb5+ Kc8 62.Qe8+ Kb7 63.Qb5+[] Kc8 64.Qe8+ Kb7 65.Qb5+[] Kc8 66.Qe8+ Kb7 67.Qb5+[] Kc8 68.Qe8+ Kb7 69.Qb5+[]
= (0.00) Depth: 18 00:00:17 2730kN, tb=214
54.Rd1 a4 55.Kd3 a3 56.e7 a2 57.e8Q b1Q+ 58.Rxb1 axb1Q+ 59.Kc4 Bd6 60.Qd7+ Bc7 61.Qb5+ Kc8 62.Qe8+ Kb7 63.Qb5+[] Kc8 64.Qe8+ Kb7 65.Qb5+[] Kc8 66.Qe8+ Kb7 67.Qb5+[] Kc8 68.Qe8+ Kb7 69.Qb5+[]
= (0.00) Depth: 19 00:00:25 4140kN, tb=394
54.Rd1 a4[] 55.Kd3 a3 56.e7 a2 57.e8Q[] b1Q+[] 58.Rxb1 axb1Q+ 59.Kc4[] Bd6 60.Qd7+ Bc7 61.Qb5+ Kc8 62.Qe8+ Kb7 63.Qb5+[] Kc8 64.Qe8+ Kb7 65.Qb5+[] Kc8 66.Qe8+ Kb7 67.Qb5+[] Kc8 68.Qe8+ Kb7 69.Qb5+[]
= (0.00) Depth: 20 00:00:40 6531kN, tb=774
54.Rd1 a4[] 55.Kd3 a3 56.e7 a2 57.e8Q[] b1Q+[] 58.Rxb1 axb1Q+ 59.Kc4[] Bd6 60.Qd7+ Bc7 61.Qb5+ Kc8 62.Qe8+ Kb7 63.Qb5+[] Kc8 64.Qe8+ Kb7 65.Qb5+[] Kc8 66.Qe8+ Kb7 67.Qb5+[] Kc8 68.Qe8+ Kb7 69.Qb5+[]
= (0.00) Depth: 21 00:01:04 10565kN, tb=1491
(, Microsoft 17.02.2012)
Houdini also seems to find it after a little more time on 8 cores, but seems to take a little long on single core
Deep Rybka 4.1 x64 2T - Zappa Mexico II 2T, Noomen2012 8 eng 2012
8/1k6/4P3/p7/1b1R4/1p3P1P/1p2K1P1/8 w - - 0 1
Analysis by Houdini 2.0c Pro x64: (8 cores)
54.Rd1 a4 55.Kd3 Kc6 56.Kc4 Be7 57.f4
+- (1.61) Depth: 6/16 00:00:00 2kN
54.Rd1 Kc6 55.Rb1 a4 56.Rxb2 Kd5 57.Kd3 Kxe6 58.g4
+/- (1.24) Depth: 7/16 00:00:00 3kN
54.Rd1 a4 55.e7 Bxe7 56.Kd3 a3 57.Kc3 a2 58.Kxb2 Bf6+ 59.Kxb3 a1Q 60.Rxa1 Bxa1
= (0.02) Depth: 8/24 00:00:00 34kN
54.Rd1
= (0.07 !) Depth: 9/24 00:00:00 36kN
54.Rd1 a4 55.e7 Bxe7 56.Kd3 a3 57.Kc3 a2 58.Kxb2 Bf6+ 59.Kxb3 a1Q 60.Rxa1 Bxa1 61.h4 Kc6 62.h5 Bg7 63.g4
= (0.13) Depth: 9/24 00:00:00 40kN
54.Rd1
= (0.17 !) Depth: 10/26 00:00:00 53kN
54.Rd1
+/= (0.26 !) Depth: 10/29 00:00:00 57kN
54.Rd1 a4
= (0.08 --) Depth: 10/29 00:00:00 98kN
54.Rd1 a4
= (0.00 --) Depth: 10/29 00:00:00 124kN
54.Rd1 a4
= (-0.21 --) Depth: 10/29 00:00:00 156kN
54.Rd7+ Kc6 55.Rd1 a4 56.Kd3 a3 57.e7 Bxe7 58.Kc3 a2 59.Kxb2 Bf6+ 60.Kxb3 a1Q 61.Rxa1 Bxa1 62.g4 Bf6 63.h4 Bxh4 64.g5 Bxg5 65.Kc4 Bf6
= (0.02) Depth: 10/30 00:00:00 172kN
54.Rd7+ Kc6 55.Rd1 a4 56.Kd3 a3 57.e7 Bxe7 58.Kc3 a2 59.Kxb2 Bf6+ 60.Kxb3 a1Q 61.Rxa1 Bxa1 62.g4 Bf6 63.h4 Bxh4 64.g5 Bxg5 65.Kc4 Bf6
= (0.02) Depth: 11/30 00:00:00 174kN
54.Rd7+ Kc6 55.Rd1 a4 56.Kd3 a3 57.e7 Bxe7 58.Kc3 a2 59.Kxb2 Bf6+ 60.Kxb3 a1Q 61.Rxa1 Bxa1 62.g4 Kd6 63.h4 Ke6 64.g5 Kf5 65.f4 Kxf4 66.Kc4 Kf5
= (0.02) Depth: 12/30 00:00:00 248kN
54.Rd7+ Kb6 55.Rd1 a4 56.Kd3 a3 57.e7 a2 58.e8Q b1Q+ 59.Kc4 Qxd1 60.Qb8+ Kc6 61.Qc8+ Kb6 62.Qb8+ Kc6
= (0.00) Depth: 13/34 00:00:00 481kN
54.Rd7+ Kb6 55.Rd1 a4 56.Kd3 a3 57.e7 a2 58.e8Q b1Q+ 59.Kc4 Qxd1 60.Qb8+ Kc6 61.Qc8+ Kb6 62.Qb8+ Kc6
= (0.00) Depth: 14/34 00:00:00 665kN
54.Rd7+ Kb6 55.Rd1 a4 56.Kd3 a3 57.e7 a2 58.e8Q b1Q+ 59.Kc4 Qxd1 60.Qb8+ Kc6 61.Qc8+ Kb6 62.Qb8+ Kc6
= (0.00) Depth: 15/38 00:00:00 992kN
54.Rd7+ Kb6 55.Rd1 a4 56.Kd3 a3 57.e7 a2 58.e8Q b1Q+ 59.Kc4 Qxd1 60.Qb8+ Kc6 61.Qc8+ Kb6 62.Qb8+ Kc6
= (0.00) Depth: 16/42 00:00:00 1811kN
54.Rd7+ Kb6 55.Rd1 a4 56.Kd3 a3 57.e7 a2 58.e8Q b1Q+ 59.Kc4 Qxd1 60.Qb8+ Kc6 61.Qc8+ Kb6 62.Qb8+ Kc6
= (0.00) Depth: 17/42 00:00:00 3071kN
54.Rd7+ Kb6 55.Rd1 a4 56.Kd3 a3 57.e7 a2 58.e8Q b1Q+ 59.Kc4 Qxd1 60.Qb8+ Kc6 61.Qc8+ Kb6 62.Qb8+ Kc6
= (0.00) Depth: 18/45 00:00:00 5264kN
54.Rd7+ Kb6 55.Rd1 a4 56.Kd3 a3 57.e7 a2 58.e8Q b1Q+ 59.Kc4 Qxd1 60.Qb8+ Kc6 61.Qc8+ Kb6 62.Qb8+ Kc6
= (0.00) Depth: 19/49 00:00:00 9721kN
54.Rd7+ Kb6 55.Rd1 a4 56.Kd3 a3 57.e7 a2 58.e8Q b1Q+ 59.Kc4 Qxd1 60.Qb8+ Kc6 61.Qc8+ Kb6 62.Qb8+ Kc6
= (0.00) Depth: 20/51 00:00:01 15466kN
54.Rd7+ Kb6 55.Rd1 a4 56.Kd3 a3 57.e7 a2 58.e8Q b1Q+ 59.Kc4 Qxd1 60.Qb8+ Kc6 61.Qc8+ Kb6 62.Qb8+ Kc6
= (0.00) Depth: 21/54 00:00:01 25161kN, tb=2
54.Rd7+ Kb6 55.Rd1 a4 56.Kd3 a3 57.e7 a2 58.e8Q b1Q+ 59.Kc4 Qxd1 60.Qb8+ Kc6 61.Qc8+ Kb6 62.Qb8+ Kc6
= (0.00) Depth: 22/55 00:00:03 43049kN, tb=13
54.Rd7+ Kb6
= (-0.04 --) Depth: 23/57 00:00:08 124mN, tb=45
54.Rd1 a4 55.Kd3 a3 56.e7 a2 57.e8Q b1Q+ 58.Rxb1 axb1Q+ 59.Kc4 Ba5 60.Qb5+ Bb6 61.Qd5+ Kb8 62.Qd6+ Kb7 63.Qd5+ Kb8
= (0.00) Depth: 23/57 00:00:10 166mN, tb=174
54.Rd1 a4 55.Kd3 a3 56.e7 a2 57.e8Q b1Q+ 58.Rxb1 axb1Q+ 59.Kc4 Ba5 60.Qb5+ Bb6 61.Qd5+ Kb8 62.Qd6+ Kb7 63.Qd5+ Kb8
= (0.00) Depth: 24/57 00:00:13 209mN, tb=281
54.Rd1 a4 55.Kd3 a3 56.e7 a2 57.e8Q b1Q+ 58.Rxb1 axb1Q+ 59.Kc4 Ba5 60.Qb5+ Bb6 61.Qd5+ Kb8 62.Qd6+ Kb7 63.Qd5+ Kb8
= (0.00) Depth: 25/59 00:00:17 282mN, tb=474
54.Rd1 a4 55.Kd3 a3 56.e7 a2 57.e8Q b1Q+ 58.Rxb1 axb1Q+ 59.Kc4 Ba5 60.Qb5+ Bb6 61.Qd5+ Kb8 62.Qd6+ Kb7 63.Qd5+ Kb8
= (0.00) Depth: 26/60 00:00:24 410mN, tb=799
54.Rd1 a4 55.Kd3 a3 56.e7 a2 57.e8Q b1Q+ 58.Rxb1 axb1Q+ 59.Kc4 Ba5 60.Qb5+ Bb6 61.Qd5+ Kb8 62.Qd6+ Kb7 63.Qd5+ Kb8
= (0.00) Depth: 27/65 00:00:49 861mN, tb=2477
(, Microsoft 17.02.2012)
Deep Rybka 4.1 x64 2T - Zappa Mexico II 2T, Noomen2012 8 eng 2012
8/1k6/4P3/p7/1b1R4/1p3P1P/1p2K1P1/8 w - - 0 1
Analysis by Houdini 2.0c Pro x64: (1 core)
54.Rd1 a4 55.Kd3 Kc6 56.Kc4 Be7 57.f4
+- (1.61) Depth: 6/16 00:00:00 2kN
54.Rd1 Kc6 55.Rb1 a4 56.Rxb2 Kd5 57.Kd3 Kxe6 58.g4
+/- (1.24) Depth: 7/19 00:00:00 4kN
54.Rd1 a4 55.e7 Bxe7 56.Kd3 a3 57.Kc3 a2 58.Kxb2 Bf6+ 59.Kxb3 a1Q 60.Rxa1 Bxa1
= (0.02) Depth: 8/24 00:00:00 31kN
54.Rd1
= (0.07 !) Depth: 9/24 00:00:00 39kN
54.Rd1 a4 55.h4 Kc6 56.Kd3 a3 57.e7 Bxe7 58.Kc3 a2 59.Kxb2 Bf6+ 60.Kxb3 a1Q 61.Rxa1 Bxa1 62.h5
= (0.09) Depth: 9/24 00:00:00 42kN
54.Rd1 a4 55.h4 Kc6 56.Kd3 a3 57.e7 Bxe7 58.Kc3 a2 59.Kxb2 Bf6+ 60.Kxb3 a1Q 61.Rxa1 Bxa1 62.h5 Bg7 63.g4
= (0.13) Depth: 10/25 00:00:00 58kN
54.Rd1 a4
= (0.08 --) Depth: 11/27 00:00:00 88kN
54.Rd1 a4
= (0.00 --) Depth: 11/27 00:00:00 119kN
54.Rd1 a4
= (-0.21 --) Depth: 11/27 00:00:00 155kN
54.Rd7+ Kc6 55.Rd1 a4 56.Kd3 a3 57.e7 Bxe7 58.Kc3 a2 59.Kxb2 Bf6+ 60.Kxb3 a1Q 61.Rxa1 Bxa1 62.Kc4 Bg7 63.h4 Kd6 64.f4 Ke6 65.g4 Kf6 66.h5 Bf8 67.g5+ Kf5
= (0.03) Depth: 11/32 00:00:00 199kN
54.Rd7+ Kc6 55.Rd1 a4 56.Kd3 a3 57.e7 Bxe7 58.Kc3 a2 59.Kxb2 Bf6+ 60.Kxb3 a1Q 61.Rxa1 Bxa1 62.Kc4 Bg7 63.h4 Kd6 64.f4 Ke6 65.g4 Kf6 66.h5 Bf8 67.Kd5
= (0.08) Depth: 12/33 00:00:00 268kN
54.Rd7+ Kc6
= (0.03 --) Depth: 13/37 00:00:00 521kN
54.Rd7+ Kb6 55.Rd1 a4 56.Kd3 a3 57.e7 a2 58.e8Q b1Q+ 59.Kc4 Qxd1 60.Qb8+ Ka6 61.Qa8+ Kb6 62.Qb8+ Ka6
= (0.00) Depth: 13/37 00:00:00 631kN
54.Rd7+ Kb6 55.Rd1 a4 56.Kd3 a3 57.e7 a2 58.e8Q b1Q+ 59.Kc4 Qxd1 60.Qb8+ Ka6 61.Qa8+ Kb6 62.Qb8+ Ka6
= (0.00) Depth: 14/40 00:00:00 865kN
54.Rd7+ Kb6 55.Rd1 a4 56.Kd3 a3 57.e7 a2 58.e8Q b1Q+ 59.Kc4 Qxd1 60.Qb8+ Ka6 61.Qa8+ Kb6 62.Qb8+ Ka6
= (0.00) Depth: 15/41 00:00:00 1250kN
54.Rd7+ Kb6 55.Rd1 a4 56.Kd3 a3 57.e7 a2 58.e8Q b1Q+ 59.Kc4 Qxd1 60.Qb8+ Ka6 61.Qa8+ Kb6 62.Qb8+ Ka6
= (0.00) Depth: 16/44 00:00:00 2000kN
54.Rd7+ Kb6 55.Rd1 a4 56.Kd3 a3 57.e7 a2 58.e8Q b1Q+ 59.Kc4 Qxd1 60.Qb8+ Ka6 61.Qa8+ Kb6 62.Qb8+ Ka6
= (0.00) Depth: 17/46 00:00:01 3332kN
54.Rd7+ Kb6 55.Rd1 a4 56.Kd3 a3 57.e7 a2 58.e8Q b1Q+ 59.Kc4 Qxd1 60.Qb8+ Ka6 61.Qa8+ Kb6 62.Qb8+ Ka6
= (0.00) Depth: 18/49 00:00:01 5639kN
54.Rd7+ Kb6 55.Rd1 a4 56.Kd3 a3 57.e7 a2 58.e8Q b1Q+ 59.Kc4 Qxd1 60.Qb8+ Ka6 61.Qa8+ Kb6 62.Qb8+ Ka6
= (0.00) Depth: 19/51 00:00:02 8303kN
54.Rd7+ Kb6 55.Rd1 a4 56.Kd3 a3 57.e7 a2 58.e8Q b1Q+ 59.Kc4 Qxd1 60.Qb8+ Ka6 61.Qa8+ Kb6 62.Qb8+ Ka6
= (0.00) Depth: 20/53 00:00:04 13578kN
54.Rd7+ Kb6 55.Rd1 a4 56.Kd3 a3 57.e7 a2 58.e8Q b1Q+ 59.Kc4 Qxd1 60.Qb8+ Ka6 61.Qa8+ Kb6 62.Qb8+ Ka6
= (0.00) Depth: 21/55 00:00:07 21832kN, tb=1
54.Rd7+ Kb6 55.Rd1 a4 56.Kd3 a3 57.e7 a2 58.e8Q b1Q+ 59.Kc4 Qxd1 60.Qb8+ Ka6 61.Qa8+ Kb6 62.Qb8+ Ka6
= (0.00) Depth: 22/57 00:00:11 35663kN, tb=7
54.Rd7+ Kb6 55.Rd1 a4 56.Kd3 a3 57.e7 a2 58.e8Q b1Q+ 59.Kc4 Qxd1 60.Qb8+ Ka6 61.Qa8+ Kb6 62.Qb8+ Ka6
= (0.00) Depth: 23/64 00:00:24 75579kN, tb=20
54.Rd7+ Kb6 55.Rd1 a4 56.Kd3 a3 57.e7 a2 58.e8Q b1Q+ 59.Kc4 Qxd1 60.Qb8+ Ka6 61.Qa8+ Kb6 62.Qb8+ Ka6
= (0.00) Depth: 24/64 00:00:38 118mN, tb=46
54.Rd7+ Kb6 55.Rd1 a4 56.Kd3 a3 57.e7 a2 58.e8Q b1Q+ 59.Kc4 Qxd1 60.Qb8+ Ka6 61.Qa8+ Kb6 62.Qb8+ Ka6
= (0.00) Depth: 25/64 00:01:09 206mN, tb=100
54.Rd7+ Kb6
= (-0.04 --) Depth: 26/72 00:02:25 434mN, tb=152
54.Rd1 a4 55.Kd3 a3 56.e7 a2 57.e8Q b1Q+ 58.Rxb1 axb1Q+ 59.Kc4 Ba5 60.Qb5+ Bb6 61.Qd7+ Kb8 62.Qe8+ Ka7 63.Qd7+ Ka6 64.Qa4+ Kb7 65.Qd7+ Kb8
= (0.00) Depth: 26/72 00:04:10 752mN, tb=814
(, Microsoft 17.02.2012)
So this might be more to do with you not having your EGTB's installed, or maybe some hash corruption due to LP. If you are using LP make sure that during testing you use engine vs engine match rather than a tourney. Loading and unloading of engine with LP creates huge memory fragmentation under Win OS. You should always have at least the 5 pieces loaded as they take minimal space on your HDD and Rybka 4.1 is tested with them on.
Thank you for the deeper analysis confirming 54.Rd1 draws. Generally I play engine matches using 2 PC's, one PC for each engine but because here I am running an 8 engine tournament, 1680 games, even at 40 moves in 5 minutes repeating, this is taking 3.5 to 4 weeks of continuous running. Therefore I am running on one PC as I cannot afford for both PC's to be tied up that long. Each engine has 2 threads and 512 Mb of hash so 1Gb hash being used out of 4 Gb available. I am stopping and resetting the PC every 24hours to make sure there is no resource issue and that seems to be working OK. I have been checking just before reset and engines that can use LP seem to be OK up to that time.
Started late January, on schedule to finish early next week. All 5 man EGTB's and some 6-man EGTB's. Gaviota table bases available to those engines that can use them. GUI EGTB enabled to save some time.
Only 1 engine having a problem that is Zappa Mexico II. Has big time loss issue when playing Stockfish 2.2.2. Seems as if Stockfish is interfering somehow with time control. No problem with other engines, only when these two meet. Will probably re-run their games on 2 PC's on completion to redress the time control issue.
PeterG
Started late January, on schedule to finish early next week. All 5 man EGTB's and some 6-man EGTB's. Gaviota table bases available to those engines that can use them. GUI EGTB enabled to save some time.
Only 1 engine having a problem that is Zappa Mexico II. Has big time loss issue when playing Stockfish 2.2.2. Seems as if Stockfish is interfering somehow with time control. No problem with other engines, only when these two meet. Will probably re-run their games on 2 PC's on completion to redress the time control issue.
PeterG
Hi Peter, would be great if you can share the games, especially interesting ones like this game! With ZM II I noticed that it is one of the engines that uses the CPU to the very max (heats it up the most), so maybe somehow SF is interfering and causing TL. As for LP, I experimented a lot with it and came up with the conclusion that the performance gain is not worth all the hassle. I can see where it can be good if you want every last ounce of hardware strength possible in a certain deep analysis of a position, but in most cases I can live without it. You sort of have to check every once in a while to see if LP is really on or the memory fragmentation has somehow made it not work. Maybe by now there are better memory fragmenters that work on the fly or maybe a better way of using it is there, but I have been out of the testing loop for a while. Nelson or Nick might be good to PM to find out about running LP on Win OS without problems, I would think that they would have figured out a way to do it without the memory fragmentation problems by now.
![[mx]](/mwf/v12/flags/mx.png "Mexico")
> As for LP, I experimented a lot with it and came up with the conclusion that the performance gain is not worth all the hassle.
+1
> Hi Peter, would be great if you can share the games, especially interesting ones like this game!
There are some interesting games in the batch to date. There is also value in looking at the games to identify any mistakes the engines make such as here but it also identifies which openings are best suited for a particular engine. It also tells me that although there is not much difference in the Houdini 2.0c to 1.5a head to head matches with 2.0c marginally stronger so far, 2.0c has about 30 Elo better performance against the engines compared to 1.5a. The results pattern against other engines also suggests a significant difference between 2.0c and 1.5a
On the theme of engine mistakes, the following recent game in the tournament highlighted Stockfish 2.2.2 unable to see stalemate and allowed Depp Rybka 4.1 to draw from what was a mate in 25 position at move 85. DR4.1 latched onto that and got the draw by 3-fold repetition.
I like to check through the games first before I post. I'd consider private mailing. Once this is completed I have to rerun the Zappa MexicoII vs Stockfish 2.2.2 games on 2 PC's to take out Zappa's time control issue in those matches.
PeterG
Stockfish 2.2.2 played 85.Rd7 allowing Deep Rybka to take advantage of the stalemate to get 3-fold repetition.
1r2r2k/R3PKRP/8/8/8/8/6p1/8 w - - 0 85
[Event "Noomen2012 8 eng"]
[Site "Newport, South Wales"]
[Date "2012.02.18"]
[Round "53.1"]
[White "Stockfish 2.2.2 JA 64bit 2T"]
[Black "Deep Rybka 4.1 x64 2T"]
[Result "1/2-1/2"]
[ECO "D80"]
[WhiteElo "3010"]
[BlackElo "2990"]
[Annotator "0.24;0.00"]
[PlyCount "195"]
[EventDate "2012.01.23"]
[EventType "tourn"]
[Source "Grayson"]
{Intel(R) Core(TM)2 Quad CPU Q9550 @ 2.83GHz 3536 MHz W=29.0 plies; 2,
744kN/s B=20.0 plies; 177kN/s; 2,462,248 TBAs} 1. d4 Nf6 2. c4 g6 3. Nc3 d5 4.
e3 Bg7 5. Qb3 e6 6. Qa3 Bf8 {0.00/16 11} 7. Qa4+ {0.24/20 9 (Qb3)} Bd7 {0.06/
16 21} 8. Qd1 {0.32/22 5 (Qc2)} Be7 {0.00/16 14} 9. Nf3 {0.28/22 0} O-O {0.00/
15 15} 10. cxd5 {0.24/22 3} exd5 {0.01/17 7} 11. Bd3 {0.24/22 5} c5 {0.01/16 0}
12. O-O {0.20/21 18 (dxc5)} c4 {-0.16/17 15 (Bg4)} 13. Bc2 {0.04/20 8 (Be2)}
Nc6 {-0.06/14 14} 14. Bd2 {0.08/20 0 (Ne5)} Be6 {-0.24/14 23} 15. b3 {0.00/24
0 (a3)} cxb3 {-0.26/15 10} 16. axb3 {0.00/21 0} a6 {-0.26/15 7 (Rc8)} 17. Qc1 {
-0.12/20 14 (h3)} Rc8 {-0.41/15 14} 18. Qb2 {0.00/21 2} Bd6 {-0.41/14 2} 19.
Rfc1 {-0.20/22 19 (h3)} Nb4 {-0.43/14 7 (Re8)} 20. Bb1 {-0.20/22 15} Qe7 {-0.
43/16 0} 21. Ne1 {-0.24/22 12 (h3)} Rc6 {-0.43/15 11 (Rfe8)} 22. f4 {-0.36/21
14 (Nd3)} Rfc8 {-0.48/15 9 (Rc7)} 23. Nf3 {-0.32/24 30} R6c7 {-0.52/17 0} 24.
Ne5 {-0.24/23 20} Ng4 {-0.50/15 0 (Qd8)} 25. h3 {-0.60/20 14} Nf6 {-0.50/16 0
(Nh6)} 26. Be1 {-0.40/19 18} Ne8 {-0.53/16 0} 27. Nf3 {-0.44/20 19 (Qd2)} Ng7 {
-0.63/14 8 (f6)} 28. Qd2 {-0.36/20 9} f6 {-0.63/15 0} 29. g4 {-0.28/21 14} Bd7
{-0.61/15 4} 30. Qf2 {-0.32/21 2} b6 {-0.61/14 7} 31. Bd2 {-0.40/21 8} a5 {-0.
61/15 0 (Qe6)} 32. Ne1 {-0.24/20 5 (Qe2)} f5 {-0.61/14 42 (Bc6)} 33. g5 {-0.40/
21 5} Ne6 {-0.41/14 0 (Bc6)} 34. Bd3 {0.00/21 4} Nxd3 {-0.34/13 3} 35. Nxd3 {
0.00/19 0} Bc6 {-0.34/16 9} 36. Qg2 {0.00/22 0} Ba3 {-0.22/15 21 (Rd8)} 37.
Rcb1 {0.00/21 8} Qd6 {-0.14/16 0} 38. b4 {0.00/24 6} Bxb4 {-0.11/16 0} 39. Nxb4
{0.00/24 11} axb4 {-0.11/17 0} 40. Na2 {0.00/24 5 (Na4)} Bb5 {-0.20/16 9} 41.
Bxb4 {0.00/26 7} Qc6 {-0.20/17 1 (Qd7)} 42. Be1 {0.12/26 11 (Nc3)} Bc4 {-0.05/
18 33} 43. Nb4 {0.16/27 0} Qb7 {-0.05/19 16 (Qd6)} 44. Qd2 {0.20/23 12 (Qc2)}
Nd8 {-0.01/18 15} 45. Nd3 {0.08/24 10} Nf7 {-0.01/17 0} 46. Ne5 {0.36/24 17}
Nxe5 {0.00/19 0 (b5)} 47. fxe5 {0.28/20 14} Rc6 {0.00/19 0 (Rf7)} 48. Qd1 {0.
16/21 20 (Rb2)} h6 {0.08/17 23 (b5)} 49. gxh6 {0.60/21 9} g5 {0.08/17 0} 50.
Qf3 {0.52/22 9 (h4)} Qh7 {0.00/16 9} 51. Rb2 {0.80/23 20} Rxh6 {0.06/19 0} 52.
h4 {0.72/23 13} g4 {0.06/19 0} 53. Qf4 {0.84/24 25} Qg6 {0.21/19 0 (b5)} 54.
Ra7 {0.80/20 9} Rh5 {0.21/17 0 (Qh5)} 55. Rb1 {1.13/22 9 (Re7)} Bd3 {0.39/17
28 (Qc6)} 56. Rba1 {1.01/21 7} Qe6 {0.32/16 0 (Bb5)} 57. Rb7 {0.96/19 7} Be4 {
0.32/15 0 (Qc6)} 58. Raa7 {1.13/20 6 (Ra6)} Qh6 {1.38/15 27 (Qc6)} 59. Qxh6 {
2.18/22 5} Rxh6 {1.26/16 0} 60. Kh2 {2.02/22 16} f4 {1.24/16 0} 61. exf4 {1.81/
18 4} Rhc6 {1.63/15 1} 62. f5 {2.10/21 5} Bxf5 {1.63/15 0} 63. Kg3 {1.81/21 13}
Rc4 {1.68/15 1 (Rc1)} 64. Rg7+ {2.34/17 3} Kh8 {1.68/15 2} 65. Rgf7 {2.30/19 1}
Be4 {1.91/15 8 (Bc2)} 66. e6 {2.82/16 4} Rxd4 {1.91/15 2 (Re8)} 67. h5 {3.03/
18 3 (Bf2)} Rd1 {2.35/14 11} 68. e7 {3.71/22 0} Kg8 {2.57/15 15} 69. Rg7+ {5.
33/22 0} Kh8 {3.02/17 6} 70. Bf2 {6.02/22 1} Re8 {3.02/17 5} 71. h6 {5.93/23 0}
Bf5 {3.02/16 17} 72. Bxb6 {6.66/26 0 (Kf4)} Rd3+ {3.35/15 14 (d4)} 73. Kf4 {9.
49/25 12} Bc8 {3.60/17 0} 74. Bc5 {10.64/25 10 (Ke5)} d4 {6.72/14 12} 75. Ke4 {
12.26/26 0} Re3+ {8.00/15 14} 76. Kxd4 {18.03/26 0} Rf3 {8.02/16 6} 77. Ke4 {
23.03/24 3} Bf5+ {8.02/16 13} 78. Kd5 {16.30/26 0} Rf4 {9.88/14 11} 79. Bd4 {
82.82/32 0} g3 {9.88/16 10} 80. Ke5 {82.67/34 0 (Be5)} Rh4 {7.52/14 5 (Re4+)}
81. Kxf5 {82.72/35 18} Rxd4 {7.63/19 0} 82. Ke6 {82.72/32 0} Rb4 {80.00/18 18
(g2)} 83. Kf7 {82.97/28 21} Rbb8 {#28/21 0} 84. h7 {82.97/28 16 (Rxg3)} g2 {8.
05/21 31} 85. Rd7 {82.72/31 0 (Rxg2)} (85. Rg8+ {#25}) 85... g1=Q {8.05/21 11}
86. Rxg1 {0.00/38 0} Rxe7+ {8.05/26 7} 87. Rxe7 {0.00/56 5 (Kxe7)} Rf8+ {1.15/
40 10} 88. Ke6 {0.00/60 3} Rf6+ {1.15/42 4} 89. Kd5 {0.00/64 7 (Ke5)} Rd6+ {1.
15/44 10} 90. Kc4 {0.00/66 5 (Ke4)} Rd4+ {0.00/43 10 (Rc6+)} 91. Kb3 {0.00/69
10 (Kc5)} Rd3+ {0.00/47 16} 92. Ka2 {0.00/71 0 (Ka4)} Ra3+ {0.00/50 8 (Rd2+)}
93. Kb2 {0.00/74 12 (Kxa3)} Rb3+ {0.00/47 9} 94. Ka1 {0.00/75 2 (Kc2)} Ra3+ {
0.00/52 8} 95. Kb1 {0.00/77 7 (Kb2)} Rb3+ {0.00/49 11} 96. Ka1 {0.00/77 3 (Ka2)
} Ra3+ {0.00/59 8} 97. Kb1 {0.00/100 0 (Kb2)} Rb3+ {0.00/58 8} 98. Ka1 {0.00/
78 9} 1/2-1/2
Date 2012-02-21 03:28
Powered by mwForum 2.27.4 © 1999-2012 Markus Wichitill