Dev Builds » 20140714-2314

You are viewing an old NCM Stockfish dev build test. You may find the most recent dev build tests using Stockfish 15 as the baseline here.

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 06:28:22 1986563 2388 148 880 1360 -110.04 ± 8.92
ncm-et-4 06:29:13 1944898 2331 166 886 1279 -110.94 ± 9.3
ncm-et-5 01:55:37 2022793 714 38 265 411 -114.42 ± 16.13
ncm-et-6 01:55:40 2023365 726 52 260 414 -102.41 ± 16.25
ncm-et-7 01:55:35 1945888 665 38 265 362 -123.56 ± 17.43
ncm-et-8 01:55:44 2020837 712 43 276 393 -118.04 ± 16.71
ncm-et-9 06:29:24 1992246 2391 164 899 1328 -110.37 ± 9.1
ncm-et-10 06:29:17 1982524 2395 140 892 1363 -112.9 ± 8.9
ncm-et-11 01:55:42 2016285 716 50 256 410 -102.86 ± 16.3
ncm-et-12 01:55:36 2015098 724 39 301 384 -131.69 ± 16.97
ncm-et-13 06:29:01 1988531 2401 167 860 1374 -103.21 ± 8.9
ncm-et-14 01:55:46 2021733 728 61 258 409 -96.42 ± 16.47
ncm-et-15 06:29:19 1985819 2391 172 862 1357 -103.19 ± 8.98
ncm-et-16 01:55:23 2022303 718 50 303 365 -127.91 ± 17.61
20000 1328 7463 11209 -110.12 ± 3.12

Test Detail

ID Host Started (UTC) Duration Base NPS Games Wins Losses Draws Elo CLI PGN
87920 ncm-et-4 2019-07-06 05:04 00:20:19 1875267 115 8 44 63 -112.54 ± 42.16
87919 ncm-et-15 2019-07-06 04:57 00:27:47 1970337 169 11 61 97 -105.96 ± 33.53
87918 ncm-et-10 2019-07-06 04:56 00:28:25 1957704 172 12 75 85 -133.46 ± 36.75
87917 ncm-et-3 2019-07-06 04:55 00:29:47 1966767 178 8 63 107 -110.98 ± 31.19
87916 ncm-et-13 2019-07-06 04:55 00:29:56 1964455 179 12 64 103 -103.92 ± 32.54
87915 ncm-et-9 2019-07-06 04:54 00:30:48 1971784 187 16 59 112 -81.35 ± 31.12
87914 ncm-et-4 2019-07-06 03:35 01:28:30 1834805 500 34 201 265 -120.67 ± 20.55
87913 ncm-et-15 2019-07-06 03:34 01:21:32 1966614 500 30 173 297 -102.22 ± 18.91
87912 ncm-et-10 2019-07-06 03:33 01:22:03 1965379 500 24 184 292 -115.23 ± 19.02
87911 ncm-et-13 2019-07-06 03:33 01:21:03 1963994 500 33 183 284 -107.54 ± 19.6
87910 ncm-et-3 2019-07-06 03:32 01:21:29 1974099 500 32 181 287 -106.78 ± 19.43
87909 ncm-et-9 2019-07-06 03:32 01:20:54 1979077 500 37 192 271 -111.37 ± 20.3
87908 ncm-et-15 2019-07-06 02:12 01:21:03 1969864 500 38 181 281 -102.22 ± 19.82
87907 ncm-et-13 2019-07-06 02:11 01:20:39 1985181 500 35 181 284 -104.49 ± 19.63
87906 ncm-et-10 2019-07-06 02:11 01:21:28 1969086 500 29 179 292 -107.54 ± 19.13
87905 ncm-et-4 2019-07-06 02:10 01:22:57 1945963 500 35 200 265 -119.11 ± 20.57
87904 ncm-et-3 2019-07-06 02:10 01:20:49 1964613 500 41 188 271 -105.25 ± 20.35
87903 ncm-et-9 2019-07-06 02:10 01:20:54 1964624 500 24 200 276 -127.76 ± 19.82
87902 ncm-et-3 2019-07-06 00:48 01:21:12 1962918 500 20 174 306 -110.6 ± 18.24
87901 ncm-et-13 2019-07-06 00:48 01:21:49 1970015 500 35 181 284 -104.49 ± 19.63
87900 ncm-et-15 2019-07-06 00:48 01:23:14 1967081 500 47 188 265 -100.71 ± 20.7
87899 ncm-et-10 2019-07-06 00:48 01:21:37 1965535 500 32 181 287 -106.78 ± 19.43
87898 ncm-et-9 2019-07-06 00:48 01:21:03 1988649 500 40 184 276 -102.97 ± 20.09
87897 ncm-et-4 2019-07-06 00:48 01:21:35 1963841 500 38 170 292 -93.95 ± 19.29
21928 ncm-et-7 2018-04-22 20:47 00:27:01 2026063 165 12 61 92 -106.38 ± 34.75
21927 ncm-et-9 2018-04-22 20:40 00:33:35 2025408 204 13 69 122 -97.88 ± 29.55
21926 ncm-et-8 2018-04-22 20:40 00:33:53 2020023 212 11 97 104 -149.54 ± 33.02
21925 ncm-et-3 2018-04-22 20:39 00:34:35 2023282 210 15 69 126 -91.39 ± 29.16
21924 ncm-et-5 2018-04-22 20:39 00:34:49 2021978 214 11 75 128 -107.18 ± 28.66
21923 ncm-et-15 2018-04-22 20:39 00:35:00 2019859 222 15 78 129 -101.38 ± 29.0
21922 ncm-et-4 2018-04-22 20:39 00:35:05 2025410 216 11 75 130 -106.13 ± 28.38
21921 ncm-et-16 2018-04-22 20:38 00:35:18 2022630 218 16 101 101 -143.04 ± 33.89
21920 ncm-et-11 2018-04-22 20:38 00:35:40 2017907 216 18 71 127 -87.03 ± 29.33
21919 ncm-et-10 2018-04-22 20:38 00:36:02 2018724 223 6 90 127 -137.65 ± 28.64
21918 ncm-et-13 2018-04-22 20:38 00:35:58 2024590 222 17 78 127 -97.98 ± 29.43
21917 ncm-et-12 2018-04-22 20:38 00:36:08 2017745 224 13 102 109 -146.08 ± 32.35
21916 ncm-et-6 2018-04-22 20:37 00:36:41 2022140 226 18 81 127 -99.48 ± 29.57
21915 ncm-et-14 2018-04-22 20:37 00:36:44 2021978 228 20 85 123 -101.87 ± 30.33
21914 ncm-et-3 2018-04-22 19:18 01:20:30 2027702 500 32 205 263 -125.39 ± 20.62
21913 ncm-et-7 2018-04-22 19:17 01:28:34 1865713 500 26 204 270 -129.35 ± 20.16
21912 ncm-et-16 2018-04-22 19:17 01:20:05 2021977 500 34 202 264 -121.46 ± 20.6
21911 ncm-et-6 2018-04-22 19:17 01:18:59 2024590 500 34 179 287 -103.73 ± 19.47
21910 ncm-et-12 2018-04-22 19:17 01:19:28 2012451 500 26 199 275 -125.39 ± 19.91
21909 ncm-et-8 2018-04-22 19:17 01:21:51 2021651 500 32 179 289 -105.25 ± 19.34
21908 ncm-et-15 2018-04-22 19:17 01:20:43 2021162 500 31 181 288 -107.54 ± 19.37
21907 ncm-et-5 2018-04-22 19:17 01:20:48 2023609 500 27 190 283 -117.55 ± 19.53
21906 ncm-et-14 2018-04-22 19:17 01:19:02 2021488 500 41 173 286 -93.95 ± 19.62
21905 ncm-et-10 2018-04-22 19:17 01:19:42 2018720 500 37 183 280 -104.49 ± 19.86
21904 ncm-et-4 2018-04-22 19:17 01:20:47 2024102 500 40 196 264 -112.14 ± 20.68
21903 ncm-et-13 2018-04-22 19:17 01:19:36 2022956 500 35 173 292 -98.44 ± 19.24
21902 ncm-et-11 2018-04-22 19:17 01:20:02 2014664 500 32 185 283 -109.83 ± 19.63
21901 ncm-et-9 2018-04-22 19:17 01:22:10 2023936 500 34 195 271 -116.0 ± 20.25

Commit

Commit ID 67a5e1ecf97eae3c74f5c84ebdbb7e3719bf90bb
Author lucasart
Date 2014-07-14 23:14:58 UTC
Contempt = 20 Also raise the admissible bounds to (-100,100), as there is no reason to prevent users from using high values if they want to. Does not regress in self play: ELO: 0.10 +-2.0 (95%) LOS: 53.7% Total: 40000 W: 7084 L: 7073 D: 25843 master vs SF 3 ELO: 182.86 +-2.7 (95%) LOS: 100.0% Total: 40000 W: 21843 L: 2541 D: 15616 Contempt = 20 vs SF 3 ELO: 189.25 +-2.8 (95%) LOS: 100.0% Total: 40000 W: 22721 L: 2859 D: 14420 Diff is therefore 6.4 +/- 3.9 elo against a 180-190 elo weaker engine, which is significantly positive, as expected. This elo difference is likely understated, because of FishTest aggressive draw adjudication though. We could push Contempt further, but after 20cp, it would get in the way of FishTest draw adjudication rule, and is likely to reduce the testing throughput as a result. bench 8198667
Copyright 2011–2024 Next Chess Move LLC