Dev Builds » 20181014-1833

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 08:48:37 1990643 3397 1825 93 1479 +195.45 +/- 8.72
ncm-et-4 08:48:26 1990011 3372 1812 102 1458 +194.16 +/- 8.8
ncm-et-9 08:48:03 1931310 3255 1769 184 1302 +184.85 +/- 9.43
ncm-et-10 08:48:27 1965277 3343 1758 84 1501 +191.2 +/- 8.62
ncm-et-13 08:48:23 1921539 3236 1744 84 1408 +196.91 +/- 8.93
ncm-et-15 08:48:26 1982167 3397 1911 83 1403 +208.99 +/- 8.97
20000 10819 630 8551 +195.25 +/- 3.64

Test Detail

ID Host Started (UTC) Duration Base NPS Games Wins Losses Draws Elo CLI PGN
56943 ncm-et-13 2018-10-18 07:49 00:37:42 1985622 236 124 5 107 +192.82 +/- 32.3
56942 ncm-et-9 2018-10-18 07:46 00:41:16 1950698 255 131 11 113 +177.48 +/- 31.87
56941 ncm-et-10 2018-10-18 07:33 00:54:02 1978890 343 183 3 157 +202.53 +/- 26.31
56940 ncm-et-4 2018-10-18 07:29 00:58:11 1989718 372 209 10 153 +207.44 +/- 27.31
56939 ncm-et-15 2018-10-18 07:26 01:00:32 1982164 397 225 15 157 +204.54 +/- 27.14
56938 ncm-et-3 2018-10-18 07:24 01:03:10 1990349 397 202 13 182 +179.93 +/- 24.87
56937 ncm-et-9 2018-10-18 06:24 01:20:45 1916660 500 280 18 202 +202.15 +/- 23.87
56936 ncm-et-13 2018-10-18 06:21 01:27:24 1776602 500 285 11 204 +213.85 +/- 23.58
56935 ncm-et-10 2018-10-18 06:14 01:17:41 1983105 500 285 15 200 +209.91 +/- 23.94
56934 ncm-et-4 2018-10-18 06:08 01:19:42 1989558 500 245 15 240 +172.78 +/- 21.5
56933 ncm-et-15 2018-10-18 06:07 01:18:21 1979506 500 288 10 202 +217.85 +/- 23.68
56932 ncm-et-3 2018-10-18 06:04 01:18:13 1991298 500 270 10 220 +200.24 +/- 22.52
56931 ncm-et-9 2018-10-18 05:01 01:21:17 1925698 500 289 16 195 +212.86 +/- 24.31
56930 ncm-et-13 2018-10-18 05:01 01:19:01 1960200 500 265 15 220 +190.85 +/- 22.66
56929 ncm-et-10 2018-10-18 04:55 01:18:11 1981224 500 258 14 228 +185.33 +/- 22.16
56928 ncm-et-4 2018-10-18 04:49 01:18:13 1989243 500 260 23 217 +179.0 +/- 23.0
56927 ncm-et-15 2018-10-18 04:47 01:18:24 1984048 500 276 14 210 +202.15 +/- 23.26
56926 ncm-et-3 2018-10-18 04:46 01:16:55 1988927 500 266 8 226 +198.34 +/- 22.08
56925 ncm-et-13 2018-10-18 03:39 01:20:40 1930476 500 271 20 209 +191.78 +/- 23.45
56924 ncm-et-9 2018-10-18 03:37 01:22:45 1916572 500 272 31 197 +182.61 +/- 24.32
56923 ncm-et-10 2018-10-18 03:33 01:21:13 1948043 500 266 17 217 +189.92 +/- 22.89
56922 ncm-et-4 2018-10-18 03:29 01:18:10 1987983 500 285 12 203 +212.86 +/- 23.67
56921 ncm-et-15 2018-10-18 03:29 01:17:27 1983262 500 275 11 214 +204.07 +/- 22.93
56920 ncm-et-3 2018-10-18 03:28 01:17:27 1991140 500 271 19 210 +192.71 +/- 23.37
56919 ncm-et-13 2018-10-18 02:18 01:19:20 1966019 500 271 11 218 +200.24 +/- 22.67
56918 ncm-et-9 2018-10-18 02:16 01:20:39 1906922 500 263 47 190 +160.64 +/- 24.71
56917 ncm-et-10 2018-10-18 02:15 01:16:48 1983105 500 249 12 239 +179.0 +/- 21.46
56916 ncm-et-4 2018-10-18 02:11 01:17:04 1993197 500 282 14 204 +207.95 +/- 23.65
56915 ncm-et-15 2018-10-18 02:10 01:17:24 1979660 500 291 10 199 +220.88 +/- 23.88
56914 ncm-et-3 2018-10-18 02:10 01:16:55 1991141 500 276 15 209 +201.19 +/- 23.35
56913 ncm-et-13 2018-10-18 00:56 01:21:42 1895472 500 277 11 212 +206.01 +/- 23.05
56912 ncm-et-10 2018-10-18 00:54 01:20:02 1941378 500 263 11 226 +192.71 +/- 22.19
56911 ncm-et-9 2018-10-18 00:53 01:21:38 1949946 500 263 34 203 +171.9 +/- 23.95
56910 ncm-et-4 2018-10-18 00:52 01:18:36 1990981 500 266 15 219 +191.78 +/- 22.72
56909 ncm-et-15 2018-10-18 00:51 01:18:23 1985304 500 284 14 202 +209.91 +/- 23.79
56908 ncm-et-3 2018-10-18 00:50 01:18:44 1990981 500 273 8 219 +205.04 +/- 22.51
56907 ncm-et-4 2018-10-17 23:32 01:18:30 1989401 500 265 13 222 +192.71 +/- 22.49
56906 ncm-et-10 2018-10-17 23:32 01:20:30 1941199 500 254 12 234 +183.51 +/- 21.75
56905 ncm-et-9 2018-10-17 23:32 01:19:43 1952674 500 271 27 202 +185.33 +/- 23.98
56904 ncm-et-15 2018-10-17 23:32 01:17:55 1981225 500 272 9 219 +203.11 +/- 22.55
56903 ncm-et-13 2018-10-17 23:32 01:22:34 1936385 500 251 11 238 +181.7 +/- 21.49
56902 ncm-et-3 2018-10-17 23:32 01:17:13 1990665 500 267 20 213 +188.08 +/- 23.2

Commit

Commit ID 97d2cc9a9c1c4b6ff1b470676fa18c7fc6509886
Author Joost VandeVondele
Date 2018-10-14 18:33:52 UTC
Randomize draw eval The patch adds a small random component (+-1) to VALUE_DRAW for the evaluation of draw positions (mostly 3folds). This random component is not static, but potentially different for each visit of the node (hence derived from the node counter). The effect is that in positions with many 3fold draw lines, different lines are followed at each iteration. This keeps the search much more dynamic, as opposed to being locked to one particular 3fold. An example of a position where master suffers from 3fold-blindness and this patch solves quickly is the famous TCEC game 53: FEN: 3r2k1/pr6/1p3q1p/5R2/3P3p/8/5RP1/3Q2K1 b - - 0 51 master doesn't see that this is a lost position (draw eval up to depth 50) as Qf6-e6 d4-d5 (found by patch at depth 23) leads to a loss. The 3fold-blindness is more important at longer TC, the patch was yellow STC and LTC, but passed VLTC: STC LLR: -2.95 (-2.94,2.94) [0.00,5.00] Total: 46328 W: 10048 L: 9953 D: 26327 http://tests.stockfishchess.org/tests/view/5b9c0ca20ebc592cf275f7c7 LTC LLR: -2.95 (-2.94,2.94) [0.00,5.00] Total: 54663 W: 8938 L: 8846 D: 36879 http://tests.stockfishchess.org/tests/view/5b9ca1610ebc592cf27601d3 VLTC LLR: 2.95 (-2.94,2.94) [0.00,5.00] Total: 31789 W: 4512 L: 4284 D: 22993 http://tests.stockfishchess.org/tests/view/5b9d1a670ebc592cf276076d Credit to @crossbr for pointing to this problem repeatedly, and giving the hint that many draw lines are typical in those situations. Bench: 4756639
Copyright 2011–2024 Next Chess Move LLC