Dev Builds » 20191007-2030

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 08:26:52 1975430 3343 2058 48 1237 +241.51 +/- 9.56
ncm-et-4 08:29:12 1963856 3334 2042 68 1224 +236.56 +/- 9.65
ncm-et-9 08:26:37 1973055 3334 2028 77 1229 +232.89 +/- 9.65
ncm-et-10 08:29:11 1967925 3328 1986 65 1277 +228.71 +/- 9.42
ncm-et-13 08:24:36 1975700 3331 2043 53 1235 +239.43 +/- 9.58
ncm-et-15 08:26:22 1971620 3330 2018 59 1253 +234.53 +/- 9.51
20000 12175 370 7455 +235.58 +/- 3.9

Test Detail

ID Host Started (UTC) Duration Base NPS Games Wins Losses Draws Elo CLI PGN
99070 ncm-et-10 2019-10-09 00:16 00:49:52 1968312 328 201 5 122 +239.5 +/- 30.61
99069 ncm-et-9 2019-10-09 00:15 00:50:12 1978596 334 210 5 119 +248.4 +/- 31.07
99068 ncm-et-15 2019-10-09 00:15 00:50:28 1980500 330 195 8 127 +223.26 +/- 30.09
99067 ncm-et-13 2019-10-09 00:15 00:50:29 1976071 331 190 11 130 +210.29 +/- 29.82
99066 ncm-et-4 2019-10-09 00:14 00:51:27 1970817 334 208 7 119 +241.8 +/- 31.17
99065 ncm-et-3 2019-10-09 00:13 00:52:08 1978885 343 211 3 129 +244.33 +/- 29.58
99064 ncm-et-15 2019-10-08 22:59 01:15:03 1972044 500 308 8 184 +240.82 +/- 24.91
99063 ncm-et-13 2019-10-08 22:59 01:15:32 1967089 500 314 9 177 +246.3 +/- 25.48
99062 ncm-et-9 2019-10-08 22:58 01:16:36 1974212 500 303 8 189 +235.45 +/- 24.53
99061 ncm-et-4 2019-10-08 22:58 01:15:29 1968942 500 314 12 174 +243.0 +/- 25.79
99060 ncm-et-10 2019-10-08 22:58 01:17:07 1964007 500 295 14 191 +220.88 +/- 24.55
99059 ncm-et-3 2019-10-08 22:57 01:15:31 1966771 500 332 4 164 +273.0 +/- 26.4
99058 ncm-et-13 2019-10-08 21:44 01:14:01 1967703 500 300 6 194 +234.38 +/- 24.1
99057 ncm-et-15 2019-10-08 21:42 01:15:37 1971722 500 288 13 199 +214.85 +/- 23.97
99056 ncm-et-10 2019-10-08 21:41 01:15:20 1967094 500 286 7 207 +218.86 +/- 23.24
99055 ncm-et-9 2019-10-08 21:40 01:16:55 1963998 500 303 15 182 +228.08 +/- 25.23
99054 ncm-et-4 2019-10-08 21:40 01:16:47 1975312 500 302 9 189 +233.32 +/- 24.56
99053 ncm-et-3 2019-10-08 21:39 01:16:19 1963535 500 301 6 193 +235.45 +/- 24.17
99052 ncm-et-13 2019-10-08 20:26 01:16:26 1970336 500 320 5 175 +257.59 +/- 25.5
99051 ncm-et-15 2019-10-08 20:24 01:17:34 1959389 500 322 9 169 +255.3 +/- 26.13
99050 ncm-et-9 2019-10-08 20:23 01:15:39 1966781 500 315 10 175 +246.3 +/- 25.67
99049 ncm-et-10 2019-10-08 20:23 01:17:21 1968778 500 277 8 215 +208.93 +/- 22.76
99048 ncm-et-3 2019-10-08 20:23 01:15:49 1968488 500 307 7 186 +240.82 +/- 24.72
99047 ncm-et-4 2019-10-08 20:22 01:17:13 1948615 500 306 11 183 +235.45 +/- 25.07
99046 ncm-et-13 2019-10-08 19:09 01:16:16 1975603 500 301 9 190 +232.26 +/- 24.49
99045 ncm-et-15 2019-10-08 19:07 01:15:21 1970480 500 320 10 170 +251.89 +/- 26.08
99044 ncm-et-10 2019-10-08 19:06 01:15:47 1968262 500 313 10 177 +244.1 +/- 25.51
99043 ncm-et-3 2019-10-08 19:06 01:15:33 1986502 500 299 15 186 +223.94 +/- 24.93
99042 ncm-et-9 2019-10-08 19:06 01:16:14 1979563 500 283 10 207 +212.86 +/- 23.35
99041 ncm-et-4 2019-10-08 19:02 01:18:46 1933912 500 294 7 199 +227.04 +/- 23.78
99040 ncm-et-13 2019-10-08 17:51 01:16:52 1975454 500 315 5 180 +251.89 +/- 25.1
99039 ncm-et-9 2019-10-08 17:50 01:15:02 1976523 500 297 10 193 +227.04 +/- 24.3
99038 ncm-et-15 2019-10-08 17:50 01:16:51 1974363 500 291 7 202 +223.94 +/- 23.58
99037 ncm-et-10 2019-10-08 17:49 01:16:06 1971481 500 309 9 182 +240.82 +/- 25.09
99036 ncm-et-3 2019-10-08 17:49 01:16:05 1979594 500 305 8 187 +237.58 +/- 24.68
99035 ncm-et-4 2019-10-08 17:46 01:14:24 1981851 500 311 10 179 +241.91 +/- 25.35
99034 ncm-et-13 2019-10-08 16:35 01:15:00 1997645 500 303 8 189 +235.45 +/- 24.53
99033 ncm-et-15 2019-10-08 16:33 01:15:28 1972842 500 294 4 202 +230.16 +/- 23.45
99032 ncm-et-9 2019-10-08 16:33 01:15:59 1971718 500 317 19 164 +238.66 +/- 26.69
99031 ncm-et-3 2019-10-08 16:32 01:15:27 1984241 500 303 5 192 +238.66 +/- 24.2
99030 ncm-et-10 2019-10-08 16:30 01:17:38 1967541 500 305 12 183 +233.32 +/- 25.09
99029 ncm-et-4 2019-10-08 16:30 01:15:06 1967547 500 307 12 181 +235.45 +/- 25.24

Commit

Commit ID 0150da5c2bd4661996b05dec4a1eca473515e9d7
Author Alayan
Date 2019-10-07 20:30:04 UTC
Adjust aspiration window with eval This patch changes the base aspiration window size depending on the absolute value of the previous iteration score, increasing it away from zero. This stems from the observation that the further away from zero, the more likely the evaluation is to change significantly with more depth. Conversely, a tighter aspiration window is more efficient when close to zero. A beneficial side-effect is that analysis of won positions without a quick mate is less prone to waste nodes in repeated fail-high that change the eval by tiny steps. STC: LLR: 2.96 (-2.94,2.94) [0.50,4.50] Total: 60102 W: 13327 L: 12868 D: 33907 http://tests.stockfishchess.org/tests/view/5d9a70d40ebc5902b6cf39ba LTC: LLR: 2.95 (-2.94,2.94) [0.00,3.50] Total: 155553 W: 25745 L: 25141 D: 104667 http://tests.stockfishchess.org/tests/view/5d9a7ca30ebc5902b6cf4028 Future work : the values used in this patch were only a reasonable guess. Further testing should unveil more optimal values. However, the aspiration window is rather tight with a minimum of 21 internal units, so discrete integers put a practical limitation to such tweaking. More exotic experiments around the aspiration window parameters could also be tried, but efficient conditions to adjust the base aspiration window size or allow it to not be centered on the current evaluation are not obvious. The aspiration window increases after a fail-high or a fail-low is another avenue to explore for potential enhancements. Bench: 4043748
Copyright 2011–2024 Next Chess Move LLC