Hi Hi Hi Endgame

[Henk]

With about 17 seconds left on the clock white wasn't able to move.

[pgn]
[Event "Computer Chess Game"]
[Site "HP"]
[Date "2017.02.22"]
[Round "-"]
[White "Fairy-Max 4.8S"]
[Black "SkipperWinb"]
[Result "0-1"]
[TimeControl "120"]
[Annotator "1. +0.05 1... +0.24"]

1. e4 {+0.05/8} e6 {+0.24/63 2.1} 2. Nc3 {+0.15/8 8} Bb4 {+0.00/57 2.0} 3.
d3 {+0.13/8 5} Ne7 {+0.65/66 2.0} 4. a3 {+0.06/9 4} Bxc3+ {+0.59/74 2.0} 5.
bxc3 {+0.12/9 1.5} Nbc6 {+0.35/67 1.9} 6. Nf3 {+0.25/8 1.7} O-O
{+0.37/64 1.9} 7. Be2 {+0.47/9 1.3} h6 {+0.20/68 1.9} 8. O-O {+0.50/10 2.3}
d5 {+0.12/65 1.8} 9. c4 {+0.45/9 1.3} dxc4 {+0.50/69 1.8} 10. dxc4
{+0.32/10 1.6} e5 {+0.51/64 1.8} 11. Qxd8 {+0.40/10 2.2} Rxd8
{+0.28/70 1.8} 12. Be3 {+0.41/9 2.2} Nd4 {+0.40/57 1.7} 13. Nxd4
{+0.38/11 1.9} exd4 {+0.29/70 1.7} 14. Rad1 {+0.29/10 1.2} c5
{+0.40/68 1.7} 15. c3 {+0.41/11 1.4} Nc6 {+0.52/71 1.7} 16. cxd4
{+0.40/12 2.6} Nxd4 {+0.46/67 1.6} 17. Bd3 {+0.34/11 1.8} Rd6
{+0.35/62 1.6} 18. h4 {+0.37/9 1.5} Bg4 {+0.53/65 1.6} 19. f3 {+0.36/9 1.1}
Be6 {+0.41/60 1.5} 20. f4 {+0.35/9 1.2} Ra6 {+0.96/63 1.5} 21. f5
{+0.42/11 1.6} Bc8 {+0.81/62 1.5} 22. Ra1 {+0.29/10 1.1} Rb6 {+0.75/65 1.5}
23. Bxd4 {+0.49/9 1.9} cxd4 {+0.83/75 1.5} 24. e5 {+0.49/10 1.2} Rb3
{+0.66/70 1.4} 25. Rf3 {+0.43/11 2.2} Rb8 {+0.63/66 1.4} 26. Be4
{+0.52/9 1.1} Rxf3 {+0.25/78 1.4} 27. gxf3 {+0.55/12 1.2} b5 {+0.08/81 1.4}
28. c5 {+1.09/11 1.5} Bb7 {-0.24/76 1.4} 29. c6 {+1.10/12 1.1} Ba8
{-0.36/78 1.3} 30. e6 {+1.03/11 2.0} Kf8 {-0.08/71 1.3} 31. a4
{+1.28/10 1.3} b4 {-0.54/72 1.3} 32. Rd1 {+1.27/10 1.0} Rc8 {-0.60/74 1.3}
33. Rxd4 {+1.12/11 1.0} fxe6 {-1.47/74 1.2} 34. fxe6 {+0.86/11 1.1} Bxc6
{-1.33/80 1.2} 35. Rc4 {+0.72/12 0.9} Bb7 {-1.05/83 1.2} 36. Rxb4
{+0.60/13 0.7} Bxe4 {-1.27/78 1.2} 37. Rxe4 {+0.67/13 0.7} Ke7
{-0.78/85 1.2} 38. Rg4 {+0.66/11 1.8} Kf6 {-0.70/82 1.2} 39. h5
{+0.75/11 1.0} Rc7 {-0.83/84 1.2} 40. Rg6+ {+0.80/12 0.7} Kf5
{-0.61/95 1.2} 41. Kf2 {+0.81/13 1.4} Re7 {-0.91/93 1.2} 42. a5
{+0.69/13 1.0} Kf4 {-0.75/94 1.1} 43. a6 {+0.74/14 1.2} Kf5 {+0.10/108 1.1}
44. Ke3 {+0.73/14 1.0} Rxe6+ {+0.45/115 1.1} 45. Rxe6 {+0.67/15 1.1} Kxe6
{+0.54/143 1.1} 46. Ke4 {+0.60/18 0.6} Kf6 {+0.17/142 1.1} 47. f4
{+0.64/18 0.6} Ke6 {+0.14/140 1.1} 48. f5+ {+0.68/21 0.6} Kf7
{+0.23/129 1.1} 49. Ke5 {+0.63/22 0.7} Ke7 {+0.00/131 1.0} 50. Kd5
{+0.62/21 0.8} Kd7 {+0.00/147 1.0} 51. Kd4 {+0.63/22 0.6} Kd6
{+0.00/139 1.0} 52. Ke4 {+0.67/24 1.3} Ke7 {+0.00/156 1.0} 53. Ke5
{+0.64/24 0.6} Kf7 {-1.40/143 1.0} 54. Kd6 {+0.78/23 1.6} Kf6
{+0.00/133 1.0} 55. Kc6 {+0.75/23 0.6} Ke5 {+0.00/148 1.0} 56. Kc7
{+2.43/21 0.5} Kd5 {+0.00/156 0.9} 57. Kd7 {+2.40/22 0.5} Ke4
{-2.09/131 1.0} 58. Ke6 {+2.14/22 0.5} Kf4 {-1.13/137 0.9} 59. f6
{+2.09/24 0.8} gxf6 {-2.52/142 0.9} 60. Kxf6 {+2.10/27 0.6} Ke4
{-1.98/159 0.9}
{Black wins on time} 0-1
[/pgn]

[Henk]

Second case: more than 30 seconds left

[pgn]

[Event "Computer Chess Game"]
[Site "HP"]
[Date "2017.02.22"]
[Round "-"]
[White "Fairy-Max 4.8S"]
[Black "SkipperWinb"]
[Result "0-1"]
[TimeControl "120"]
[Annotator "1. +0.14 1... -0.08"]

1. Nc3 {+0.14/8} Nf6 {-0.08/63 2.1} 2. d4 {+0.10/8 2.0} h6 {+0.15/60 2.0}
3. e3 {+0.03/8 5} e6 {+0.42/64 2.0} 4. Nf3 {-0.02/8 3} Bb4 {+0.46/68 2.0}
5. a3 {-0.03/9 2.2} Bxc3+ {+0.40/71 1.9} 6. bxc3 {+0.28/9 1.5} Ne4
{+0.32/69 1.9} 7. c4 {+0.30/9 1.8} Nc3 {+0.00/67 1.9} 8. Qd3 {+0.37/10 2.2}
Na4 {+0.18/71 1.8} 9. Bd2 {+0.26/9 2.1} O-O {+0.29/62 1.8} 10. O-O-O
{+0.35/8 1.5} Nc6 {+0.85/62 1.8} 11. c5 {+0.33/10 2.3} a6 {+0.73/64 1.8}
12. c4 {+0.41/10 2.7} Na7 {+0.70/64 1.7} 13. Qb3 {+0.41/10 2.5} b5
{+0.72/65 1.7} 14. cxb5 {+0.51/11 1.8} axb5 {+0.29/69 1.7} 15. Bxb5
{+0.44/11 11} Nxb5 {-0.12/73 1.7} 16. Qxb5 {+0.72/10 1.0} d6 {-0.11/68 1.6}
17. Be1 {+0.72/9 4} dxc5 {+0.01/63 1.6} 18. dxc5 {+0.66/10 1.5} Qe7
{+0.00/68 1.6} 19. Bb4 {+0.78/9 1.8} Rd8 {+0.07/68 1.5} 20. h4
{+1.33/9 2.0} Rxd1+ {+0.09/70 1.5} 21. Rxd1 {+1.28/10 2.0} Qf6
{-0.38/65 1.5} 22. Rd4 {+1.32/9 1.8} Qe7 {-1.37/70 1.5} 23. Ne5
{+2.04/10 1.8} Qf8 {-1.01/69 1.4} 24. Nc6 {+2.47/10 1.3} Kh7 {-3.79/71 1.4}
25. Rd8 {+3.16/10 1.9} Ba6 {-3.63/74 1.4} 26. Rxa8 {+3.23/11 0.9} Bxb5
{-4.10/77 1.4} 27. Rxf8 {+4.23/14 1.6} Bxc6 {-4.89/90 1.4} 28. Rxf7
{+3.94/13 1.8} Bxg2 {-5.05/85 1.4} 29. Rxc7 {+4.01/13 1.5} Kg8
{-5.00/81 1.3} 30. c6 {+5.09/11 0.7} g5 {-6.08/80 1.3} 31. Rc8+
{+7.04/13 0.7} Kf7 {-7.96/80 1.3} 32. c7 {+7.19/13 0.9} Bb7 {-9.24/74 1.3}
33. Rb8 {+7.24/14 0.8} Ba6 {-9.01/70 1.3} 34. Ra8 {+7.32/14 2.2} Bb7
{-9.53/71 1.2} 35. Rxa4 {+7.38/14 5} gxh4 {-8.87/86 1.2} 36. Bd6
{+7.33/13 0.9} h3 {-9.31/77 1.2} 37. Rh4 {+7.85/13 0.7} Kg6 {-9.53/76 1.2}
38. Rxh3 {+7.89/13 0.5} h5 {-10.32/73 1.1} 39. Rg3+ {+7.92/12 0.6} Kf5
{-10.39/72 1.1} 40. Rg8 {+8.10/11 0.5} h4 {-10.41/70 1.1} 41. c8=Q
{+8.80/14 0.7} Bxc8 {-10.70/86 1.1}
{Black wins on time} 0-1
[/pgn]

[hgm]

The problem is usually that the opponent played the expected move at the point where null-move is sqwitched off. The root gets a high-depth exact hash hit, and tries to improve on the depth of that without null move.

Fairy-Max 5.0 should be less sensitive to this problem.

[Sven]

The problem is usually that the opponent played the expected move at the point where null-move is sqwitched off. The root gets a high-depth exact hash hit, and tries to improve on the depth of that without null move.

Fairy-Max 5.0 should be less sensitive to this problem.

In general I think that probing TT at the root should not be used for a "TT cutoff" (if I understood you right that you believe that might have happened here) but only to get the best move for move ordering. So I see no relation to null-move at this point.

[Henk]

Skipper's time management:

[pgn]
[Event "Computer Chess Game"]
[Site "HP"]
[Date "2017.02.23"]
[Round "-"]
[White "Fairy-Max 4.8S"]
[Black "SkipperWinb"]
[Result "1-0"]
[TimeControl "120"]
[Annotator "1. +0.06 1... -0.04"]

1. d3 {+0.06/8} Nc6 {-0.04/90 2.2} 2. c4 {+0.16/8 2.5} e6 {+0.20/80 2.0} 3.
Nf3 {+0.16/7 2.0} h6 {+0.24/80 2.0} 4. Nc3 {-0.02/8 2.7} Nf6
{+0.37/100 1.9} 5. e4 {-0.10/8 4} Bb4 {+0.47/80 1.9} 6. Be2 {+0.00/8 1.8}
d5 {+0.47/90 1.9} 7. exd5 {+0.08/8 2.6} exd5 {+0.28/90 1.8} 8. O-O
{+0.06/8 2.5} Bxc3 {+0.75/100 1.8} 9. bxc3 {+0.12/10 3} dxc4
{+0.31/100 1.8} 10. dxc4 {+0.03/10 1.4} O-O {+0.37/100 1.8} 11. Qxd8
{+0.45/10 2.2} Rxd8 {+0.37/90 1.7} 12. Bf4 {+0.49/10 2.1} Rd7
{+0.42/100 1.7} 13. h4 {+0.34/10 1.9} Re7 {+0.65/90 1.7} 14. Bd3
{+0.18/10 2.2} Nh5 {+0.55/90 1.7} 15. Be3 {+0.16/10 2.8} Bg4 {+0.45/80 1.6}
16. Rfb1 {+0.26/9 6} Bxf3 {+0.31/80 1.6} 17. gxf3 {+0.60/10 1.0} b6
{+0.15/90 1.7} 18. Be4 {+0.69/10 1.3} Re6 {+0.13/100 1.6} 19. Bf5
{+0.58/9 1.4} Rf6 {+0.43/100 1.6} 20. Be4 {+0.01/11 1.7} Re8
{+0.67/110 1.5} 21. Rb5 {+0.24/9 1.1} Ne5 {+1.09/100 1.5} 22. Bd4
{+0.26/10 1.2} c5 {+1.15/110 1.5} 23. Bxe5 {+0.26/11 1.2} Rxe5
{+1.26/110 1.5} 24. Rd1 {+0.25/10 1.2} Re8 {+1.33/110 1.5} 25. Rd7
{+0.13/10 1.6} Nf4 {+1.03/110 1.4} 26. Kh2 {+0.07/9 1.2} a6 {+1.13/110 1.4}
27. Rb2 {-0.07/10 1.1} Ng6 {+1.23/110 1.4} 28. Ra7 {-0.04/10 1.4} a5
{+1.01/90 1.5} 29. Kg3 {-0.03/10 0.9} Ne7 {+0.67/100 1.4} 30. Ra6
{+0.46/10 1.1} Nc8 {+0.58/110 1.3} 31. a3 {+0.42/10 0.9} Rf8 {+0.57/80 1.3}
32. a4 {+0.42/11 2.9} h5 {+0.71/80 1.3} 33. Ra8 {+0.38/11 1.0} Re8
{+1.05/100 1.3} 34. Rd2 {+0.40/12 1.0} Kf8 {+0.91/110 1.3} 35. Rd5
{+0.33/11 0.7} g6 {+0.96/100 1.3} 36. Rb8 {+0.33/12 0.9} Nd6
{+1.13/110 1.3} 37. Rxe8+ {+0.46/13 0.7} Kxe8 {+1.06/120 1.2} 38. Re5+
{+0.38/13 1.2} Kd7 {+1.03/110 1.2} 39. Bd3 {+0.04/13 0.9} Kc8
{+1.18/120 1.2} 40. Rg5 {+0.19/12 1.5} Re6 {+1.28/110 1.2} 41. Rd5
{-0.27/12 1.0} Kb7 {+1.19/100 1.2} 42. Bf1 {+0.01/10 1.1} Kb8
{+1.23/90 1.2} 43. Rd1 {+0.20/11 0.7} Nf5+ {+1.22/110 1.2} 44. Kh3
{+0.21/11 0.8} Ng7 {+1.17/110 1.1} 45. Rd8+ {+0.36/11 0.9} Kc7
{+1.18/120 1.1} 46. Ra8 {+0.36/13 2.9} Rf6 {+1.08/110 1.1} 47. Be2
{+0.26/13 2.0} Kb7 {+1.29/120 1.1} 48. Rh8 {+0.23/12 0.8} Ne6
{+1.19/110 1.1} 49. Kg3 {+0.09/12 0.5} Kc6 {+1.19/100 1.1} 50. Rc8+
{+0.42/11 1.0} Kd7 {+1.11/110 1.1} 51. Rg8 {+0.41/12 1.1} Ke7
{+1.09/100 1.0} 52. Rb8 {+0.43/11 1.2} Nf4 {+1.09/120 1.2} 53. Rb7+
{+0.75/11 0.5} Ke8 {+1.09/130 1.0} 54. Bf1 {+0.68/14 0.7} Kf8
{+1.02/130 1.0} 55. Rd7 {+0.78/12 0.5} Kg7 {+0.88/120 1.0} 56. Rd2
{+0.61/12 0.8} Ne6 {+1.02/110 1.0} 57. Be2 {+0.42/12 1.1} Nc7
{+1.12/100 1.0} 58. Rd1 {+0.13/11 0.7} Re6 {+1.12/100 0.9} 59. Bd3
{+0.14/12 0.5} Kh6 {+1.14/80 0.9} 60. f4 {+0.30/11 0.5} Rd6 {+1.19/100 0.9}
61. Bc2 {+0.08/12 0.4} Rxd1 {+1.22/120 0.9} 62. Bxd1 {+0.29/16 0.4} Ne8
{+1.20/140 1.0} 63. Be2 {+0.30/12 0.7} Nd6 {+1.39/140 0.9} 64. Bd3
{+0.29/12 0.6} Kg7 {+1.45/130 0.9} 65. Kg2 {+0.26/12 0.7} f5
{+1.15/130 0.9} 66. f3 {+0.40/12 0.5} Kf7 {+1.39/130 0.9} 67. Kf2
{+0.39/12 0.4} Ke6 {+1.67/120 1.0} 68. Ke3 {+0.40/13 0.5} Nb7
{+1.31/140 0.8} 69. Be2 {+0.22/14 0.6} Kd7 {+1.36/140 0.8} 70. Bf1
{+0.23/13 0.4} Nd6 {+1.31/150 0.8} 71. Bd3 {+0.34/12 0.4} Kc6
{+1.31/140 0.8} 72. Kd2 {+0.32/13 0.4} Kc7 {+1.31/150 0.8} 73. Be2
{+0.32/13 0.6} Kd7 {+1.31/150 0.8} 74. Kd3 {+0.36/14 0.7} Nf7
{+1.31/150 0.8} 75. Ke3 {+0.33/14 0.7} Ke6 {+1.31/160 0.8} 76. Bd3
{+0.35/14 0.5} Nd6 {+1.31/160 0.8} 77. Ke2 {+0.33/13 0.6} Ke7
{+1.31/130 0.8} 78. Ke1 {+0.32/13 0.5} Ne8 {+1.35/130 0.8} 79. Kf2
{+0.32/13 0.5} Kd7 {+1.35/140 0.8} 80. Kg2 {+0.25/13 0.8} Nd6
{+1.53/130 0.7} 81. Kg3 {+0.34/12 0.4} Ke6 {+1.56/140 0.8} 82. Be2
{+0.34/13 0.4} Nc8 {+1.47/130 0.7} 83. Kg2 {+0.30/13 0.3} Kd7
{+1.35/150 0.9} 84. Kf2 {+0.28/13 0.3} Nd6 {+1.33/160 0.7} 85. Kg3
{+0.34/13 0.5} Nb7 {+1.31/130 0.7} 86. Bd3 {+0.27/13 0.3} Nd8
{+1.35/130 0.7} 87. Kf2 {+0.27/12 0.2} Ne6 {+1.38/150 0.7} 88. Ke3
{+0.35/12 0.3} Nc7 {+1.35/150 0.7} 89. Bf1 {+0.25/12 0.2} Ke6
{+1.35/140 0.7} 90. Kf2 {+0.32/12 0.2} Kf6 {+1.35/140 0.7} 91. Bd3
{+0.36/13 0.4} Kg7 {+1.35/130 0.7} 92. Ke3 {+0.31/12 0.3} Kh6
{+1.35/140 0.6} 93. Bc2 {+0.27/13 0.3} Ne6 {+1.35/150 0.6} 94. Bb3
{+0.31/13 0.3} Kg7 {+1.38/160 0.7} 95. Bd1 {+0.32/13 0.4} Kf7
{+1.38/170 0.6} 96. Bc2 {+0.36/13 0.3} Kf6 {+1.38/160 0.6} 97. Bd3
{+0.37/13 0.3} Ke7 {+1.38/160 0.6} 98. Be2 {+0.32/13 0.4} Kd8
{+1.38/170 0.6} 99. Bf1 {+0.27/13 0.6} Nf8 {+1.38/170 0.6} 100. Bd3
{+0.28/13 0.2} Nd7 {+1.38/150 0.6} 101. Be2 {+0.24/12 0.3} Kc7
{+1.35/140 0.6} 102. Kd2 {+0.25/12 0.2} Nf8 {+1.38/160 0.6} 103. Bd3
{+0.27/12 0.2} Kd7 {+1.35/140 0.6} 104. Ke3 {+0.27/13 0.3} Kd8
{+1.35/150 0.6} 105. Bf1 {+0.22/12 0.1} Ke7 {+1.35/150 0.6} 106. Bg2
{+0.27/12 0.4} Ne6 {+1.38/170 0.6} 107. Bh3 {+0.29/11 0.2} Kd7
{+1.38/140 0.6} 108. Bf1 {+0.28/12 0.2} Kd8 {+1.35/150 0.7} 109. Be2
{+0.27/13 0.3} Ke8 {+1.35/160 0.5} 110. Bf1 {+0.01/13 0.2} Nc7
{+1.35/160 0.5} 111. Kf2 {+0.25/12 0.2}
{White wins on time} 1-0
[/pgn]

[hgm]

In general I think that probing TT at the root should not be used for a "TT cutoff" (if I understood you right that you believe that might have happened here) but only to get the best move for move ordering. So I see no relation to null-move at this point.

Well, actually it doesn't really cutoff in the root. (Because there is no such thing as 'sufficient depth' in the root.) But if the previous move was searched to depth=N the first N-2 iterations all get sufficient-depth hash hits from the ply=1 level, and thus take zero time. Then it happily starts on the N-1 iteration. But it now has to complete that without the benefit of null move, while the previous depth N was reached using null move. This will take very long, not only because of the higher EBF, but also because it has to find completely new refutation trees for moves that before were refuted by a null move.

[Uri Blass]

In general I think that probing TT at the root should not be used for a "TT cutoff" (if I understood you right that you believe that might have happened here) but only to get the best move for move ordering. So I see no relation to null-move at this point.

Well, actually it doesn't really cutoff in the root. (Because there is no such thing as 'sufficient depth' in the root.) But if the previous move was searched to depth=N the first N-2 iterations all get sufficient-depth hash hits from the ply=1 level, and thus take zero time. Then it happily starts on the N-1 iteration. But it now has to complete that without the benefit of null move, while the previous depth N was reached using null move. This will take very long, not only because of the higher EBF, but also because it has to find completely new refutation trees for moves that before were refuted by a null move.

so better may be never to switch off null move because I guess it save lines in the code and also prevent losing on time.

[hgm]

But it cannot win KRK and KPK (or draw KKP)with null move on. And the time forfeits mostly occur in lost positions.