Dev Builds » 20131107-1846

Use this dev build

NCM plays each Stockfish dev build 20,000 times against Stockfish 7. This yields an approximate Elo difference and establishes confidence in the strength of the dev builds.

Summary

Host Duration Avg Base NPS Games Wins Losses Draws Elo
ncm-et-3 06:02:57 1985658 2420 123 1183 1114 -163.22 +/- 10.11
ncm-et-4 06:02:36 1986501 2432 122 1202 1108 -165.83 +/- 10.14
ncm-et-5 01:53:40 2019162 782 41 415 326 -180.92 +/- 18.83
ncm-et-6 01:53:34 2024265 764 28 400 336 -184.84 +/- 18.39
ncm-et-7 01:53:03 1986658 750 38 354 358 -156.11 +/- 17.78
ncm-et-8 01:53:20 1971842 753 27 379 347 -176.09 +/- 18.01
ncm-et-9 06:03:09 1989979 2438 119 1191 1128 -163.94 +/- 10.03
ncm-et-10 06:02:47 1983470 2440 102 1185 1153 -165.73 +/- 9.87
ncm-et-12 01:53:34 2017442 774 30 378 366 -168.23 +/- 17.51
ncm-et-13 06:02:55 1990302 2450 131 1199 1120 -162.32 +/- 10.09
ncm-et-14 01:53:39 2019454 783 38 415 330 -182.37 +/- 18.68
ncm-et-15 06:03:06 1987634 2439 104 1220 1115 -171.71 +/- 10.08
ncm-et-16 01:53:37 2002079 775 47 408 320 -175.35 +/- 19.04
20000 950 9929 9121 -167.94 +/- 3.53

Test Detail

ID Host Started (UTC) Duration Base NPS Games Wins Losses Draws Elo CLI PGN
99898 ncm-et-15 2019-10-15 12:01 00:23:12 1974831 163 6 71 86 -146.68 +/- 35.69
99897 ncm-et-3 2019-10-15 12:00 00:23:39 1968018 159 3 89 67 -210.34 +/- 41.19
99896 ncm-et-4 2019-10-15 12:00 00:24:27 1966936 158 9 68 81 -136.33 +/- 37.33
99895 ncm-et-9 2019-10-15 11:59 00:25:16 1975775 168 9 79 80 -154.14 +/- 37.85
99894 ncm-et-13 2019-10-15 11:59 00:25:32 1977157 181 10 79 92 -139.49 +/- 35.02
99893 ncm-et-10 2019-10-15 11:58 00:25:57 1965096 171 8 78 85 -151.08 +/- 36.43
99892 ncm-et-15 2019-10-15 10:45 01:14:55 1964169 500 23 239 238 -160.64 +/- 21.79
99891 ncm-et-4 2019-10-15 10:44 01:14:20 1987708 500 23 236 241 -158.08 +/- 21.63
99890 ncm-et-3 2019-10-15 10:44 01:15:17 1963994 500 31 232 237 -148.02 +/- 21.97
99889 ncm-et-9 2019-10-15 10:43 01:14:42 1968467 500 24 235 241 -156.39 +/- 21.64
99888 ncm-et-13 2019-10-15 10:42 01:15:24 1973603 500 19 238 243 -163.21 +/- 21.43
99887 ncm-et-10 2019-10-15 10:41 01:15:42 1965260 500 21 246 233 -168.4 +/- 22.03
99886 ncm-et-9 2019-10-15 09:28 01:14:00 1969093 500 19 267 214 -189.0 +/- 23.12
99885 ncm-et-4 2019-10-15 09:28 01:15:14 1974986 500 31 230 239 -146.36 +/- 21.86
99884 ncm-et-15 2019-10-15 09:27 01:16:45 1982671 500 21 264 215 -184.42 +/- 23.09
99883 ncm-et-3 2019-10-15 09:27 01:15:57 1971413 500 29 232 239 -149.68 +/- 21.84
99882 ncm-et-13 2019-10-15 09:27 01:14:05 1972658 500 28 252 220 -167.53 +/- 22.89
99881 ncm-et-10 2019-10-15 09:26 01:14:05 1972065 500 23 248 229 -168.4 +/- 22.3
99880 ncm-et-4 2019-10-15 08:12 01:15:14 1967849 500 27 266 207 -180.8 +/- 23.66
99879 ncm-et-10 2019-10-15 08:12 01:13:40 1960472 500 21 240 239 -163.21 +/- 21.7
99878 ncm-et-13 2019-10-15 08:12 01:14:19 1972808 500 31 250 219 -163.21 +/- 22.98
99877 ncm-et-15 2019-10-15 08:12 01:14:35 1965394 500 20 261 219 -182.61 +/- 22.83
99876 ncm-et-9 2019-10-15 08:12 01:15:33 1972449 500 26 234 240 -153.86 +/- 21.73
99875 ncm-et-3 2019-10-15 08:12 01:14:25 1959546 500 25 245 230 -164.07 +/- 22.27
29448 ncm-et-9 2018-05-13 11:25 00:38:35 2026719 270 13 141 116 -179.04 +/- 31.57
29447 ncm-et-7 2018-05-13 11:25 00:38:38 2002155 250 13 107 130 -137.37 +/- 29.29
29446 ncm-et-8 2018-05-13 11:25 00:38:39 1932972 253 8 138 107 -197.32 +/- 32.72
29445 ncm-et-6 2018-05-13 11:25 00:38:47 2023610 264 10 138 116 -183.9 +/- 31.41
29444 ncm-et-3 2018-05-13 11:25 00:38:58 2024917 261 10 129 122 -171.0 +/- 30.45
29443 ncm-et-4 2018-05-13 11:24 00:39:07 2025245 274 10 136 128 -172.72 +/- 29.69
29442 ncm-et-13 2018-05-13 11:24 00:39:30 2022140 269 13 132 124 -165.1 +/- 30.37
29441 ncm-et-12 2018-05-13 11:24 00:39:29 2021977 274 6 136 132 -179.21 +/- 28.85
29440 ncm-et-10 2018-05-13 11:24 00:39:40 2020023 269 10 118 141 -147.81 +/- 27.85
29439 ncm-et-15 2018-05-13 11:24 00:39:48 2017908 276 8 141 127 -182.55 +/- 29.73
29438 ncm-et-16 2018-05-13 11:23 00:41:20 1982508 275 19 135 121 -156.31 +/- 31.0
29437 ncm-et-5 2018-05-13 11:22 00:41:24 2023773 282 16 145 121 -171.66 +/- 30.97
29436 ncm-et-14 2018-05-13 11:22 00:41:28 2020836 283 16 146 121 -172.5 +/- 30.98
29435 ncm-et-7 2018-05-13 10:10 01:14:25 1971161 500 25 247 228 -165.8 +/- 22.39
29434 ncm-et-10 2018-05-13 10:09 01:13:43 2017908 500 19 255 226 -178.11 +/- 22.4
29433 ncm-et-8 2018-05-13 10:09 01:14:41 2010713 500 19 241 240 -165.8 +/- 21.6
29432 ncm-et-13 2018-05-13 10:09 01:14:05 2023446 500 30 248 222 -162.35 +/- 22.8
29431 ncm-et-12 2018-05-13 10:09 01:14:05 2012907 500 24 242 234 -162.35 +/- 22.03
29430 ncm-et-16 2018-05-13 10:09 01:12:17 2021651 500 28 273 199 -186.25 +/- 24.18
29429 ncm-et-9 2018-05-13 10:09 01:15:03 2027374 500 28 235 237 -153.02 +/- 21.93
29428 ncm-et-4 2018-05-13 10:09 01:14:14 1996285 500 22 266 212 -185.33 +/- 23.29
29427 ncm-et-15 2018-05-13 10:09 01:13:51 2020836 500 26 244 230 -162.35 +/- 22.29
29426 ncm-et-14 2018-05-13 10:09 01:12:11 2018072 500 22 269 209 -188.08 +/- 23.48
29425 ncm-et-3 2018-05-13 10:09 01:14:41 2026063 500 25 256 219 -173.67 +/- 22.91
29424 ncm-et-6 2018-05-13 10:09 01:14:47 2024920 500 18 262 220 -185.33 +/- 22.73
29423 ncm-et-5 2018-05-13 10:09 01:12:16 2014552 500 25 270 205 -186.25 +/- 23.77

Commit

Commit ID ecd07e51d0f03ccd3e41e5634518b299989254dd
Author Lucas Braesch
Date 2013-11-07 18:46:51 UTC
Retire eval margin and gains 1/ eval margin and gains removed: - gains removed by Value(128): search() and qsearch() now behave consistently! 2/ futility_margin() - testing showed that there is no added value in this weird (log(depth), movecount) formula, and a much simpler linear formula is just as good. In fact, it is most likely better, as it is not yet optimally tuned. - the new simplified formula also means we get rid of FutilityMargins[], its initialization code, and more importantly ss->futilityMoveCount, and the hacky code that updates it throughout the search(). - the current formula gives negative futility margins, and there is a hidden interaction between the move coutn pruning formula and the futility margin one: what happens is that MCP is supposed to be triggered before we use the non-sensical negative futility margins. 3/ unify pre & post futility pruning - pre futility pruning (what SF calls value based pruning) used depth < 7 plies, while post futility pruning (what SF calls static null move pruning) used depth < 4 plies. - also the condition depth < 7 in pre futility pruning was not obvious, and it seemd to be depth < 16 (futility_margin() returns an infinite value when depth >= 7). Tested with fixed number of games both at short TC: ELO: 0.82 +-2.1 (95%) LOS: 77.3% Total: 40000 W: 7939 L: 7845 D: 24216 And long TC ELO: 0.59 +-2.0 (95%) LOS: 71.9% Total: 40000 W: 6876 L: 6808 D: 26316 bench: 10206576
Copyright 2011–2024 Next Chess Move LLC