Dev Builds » 20230223-1227

Use this dev build

NCM plays each Stockfish dev build 20,000 times against Stockfish 15. This yields an approximate Elo difference and establishes confidence in the strength of the dev builds.

Summary

Host Duration Avg Base NPS Games WLD Standard Elo Ptnml(0-2) Gamepair Elo
ncm-dbt-01 07:01:06 583788 3998 1217 807 1974 +35.76 ± 5.18 3 275 1036 679 6 +71.74 ± 10.55
ncm-dbt-02 06:59:13 586077 4000 1235 838 1927 +34.6 ± 5.15 3 275 1050 666 6 +69.35 ± 10.47
ncm-dbt-03 07:00:45 586220 3996 1209 820 1967 +33.93 ± 5.03 1 263 1084 646 4 +67.98 ± 10.26
ncm-dbt-04 07:01:05 570051 4016 1246 821 1949 +36.91 ± 5.12 5 258 1058 681 6 +74.48 ± 10.42
ncm-dbt-05 07:01:15 583541 3990 1198 788 2004 +35.83 ± 5.09 2 263 1056 671 3 +72.25 ± 10.42
20000 6105 4074 9821 +35.4 ± 2.29 14 1334 5284 3343 25 +71.16 ± 4.66

Test Detail

ID Host Base NPS Games WLD Standard Elo Ptnml(0-2) Gamepair Elo CLI PGN
411920 ncm-dbt-04 569669 16 4 7 5 -65.87 ± 110.97 1 2 4 1 0 -88.74 ± 190.27
411919 ncm-dbt-05 583824 490 146 101 243 +32.0 ± 15.02 1 34 131 77 2 +63.08 ± 29.66
411918 ncm-dbt-03 585084 496 146 103 247 +30.19 ± 14.88 0 39 128 80 1 +59.41 ± 30.12
411917 ncm-dbt-01 584076 498 157 105 236 +36.41 ± 14.58 1 31 133 83 1 +73.64 ± 29.4
411916 ncm-dbt-02 585169 500 158 104 238 +37.67 ± 15.25 2 33 125 89 1 +77.71 ± 30.5
411915 ncm-dbt-04 570669 500 149 101 250 +33.46 ± 14.21 0 34 134 82 0 +67.55 ± 29.3
411914 ncm-dbt-05 584369 500 156 104 240 +36.26 ± 14.26 0 33 132 85 0 +73.34 ± 29.56
411913 ncm-dbt-03 586646 500 162 113 225 +34.16 ± 14.26 0 33 136 80 1 +67.55 ± 29.02
411912 ncm-dbt-01 582318 500 151 102 247 +34.16 ± 14.13 0 32 138 79 1 +67.55 ± 28.74
411911 ncm-dbt-02 587113 500 172 102 226 +48.96 ± 15.25 0 34 113 102 1 +98.44 ± 32.13
411910 ncm-dbt-04 569111 500 158 94 248 +44.72 ± 15.65 1 35 115 97 2 +89.48 ± 31.85
411909 ncm-dbt-05 582987 500 152 102 246 +34.86 ± 14.44 0 35 130 85 0 +70.44 ± 29.84
411908 ncm-dbt-01 584369 500 144 96 260 +33.46 ± 14.86 0 39 124 87 0 +67.55 ± 30.65
411907 ncm-dbt-03 585928 500 151 97 252 +37.67 ± 13.95 0 30 136 84 0 +76.25 ± 28.98
411906 ncm-dbt-02 586012 500 141 108 251 +22.96 ± 14.53 0 40 139 69 2 +43.3 ± 28.7
411905 ncm-dbt-04 569669 500 149 111 240 +26.46 ± 14.39 1 35 140 73 1 +53.22 ± 28.53
411904 ncm-dbt-01 583237 500 155 116 229 +27.15 ± 15.08 0 44 123 83 0 +54.65 ± 30.79
411903 ncm-dbt-03 586054 500 163 105 232 +40.48 ± 14.52 0 33 126 91 0 +82.1 ± 30.36
411902 ncm-dbt-02 588132 500 159 109 232 +34.86 ± 14.17 0 33 134 83 0 +70.44 ± 29.29
411901 ncm-dbt-05 582903 500 138 94 268 +30.65 ± 14.02 0 34 138 78 0 +61.79 ± 28.77
411900 ncm-dbt-04 570388 500 161 100 239 +42.6 ± 14.65 0 33 123 94 0 +86.52 ± 30.76
411899 ncm-dbt-02 586223 500 146 103 251 +29.95 ± 14.37 0 36 136 77 1 +58.93 ± 29.06
411898 ncm-dbt-03 585928 500 152 92 256 +41.89 ± 13.66 0 26 138 86 0 +85.04 ± 28.63
411897 ncm-dbt-01 583489 500 160 100 240 +41.89 ± 14.99 0 35 121 93 1 +83.57 ± 31.04
411896 ncm-dbt-05 583614 500 144 98 258 +32.05 ± 14.25 0 35 134 81 0 +64.66 ± 29.31
411895 ncm-dbt-04 570589 500 146 93 261 +36.97 ± 13.63 0 28 141 81 0 +74.79 ± 28.27
411894 ncm-dbt-02 585169 500 165 105 230 +41.89 ± 14.48 0 32 126 92 0 +85.04 ± 30.35
411893 ncm-dbt-03 585674 500 156 106 238 +34.86 ± 13.9 0 31 138 81 0 +70.44 ± 28.73
411892 ncm-dbt-01 585337 500 154 94 252 +41.89 ± 14.61 1 29 130 89 1 +85.04 ± 29.79
411891 ncm-dbt-04 569350 500 153 102 245 +35.56 ± 15.0 1 34 130 83 2 +70.44 ± 29.84
411890 ncm-dbt-05 583112 500 151 108 241 +29.95 ± 14.89 0 40 128 81 1 +58.93 ± 30.13
411889 ncm-dbt-02 584537 500 145 111 244 +23.66 ± 13.92 1 34 145 70 0 +48.96 ± 27.86
411888 ncm-dbt-03 585211 500 135 109 256 +18.08 ± 14.02 1 37 148 63 1 +36.26 ± 27.5
411887 ncm-dbt-05 582861 500 156 97 247 +41.19 ± 13.76 0 27 137 86 0 +83.57 ± 28.79
411886 ncm-dbt-04 570549 500 163 101 236 +43.3 ± 14.42 1 27 132 89 1 +88.0 ± 29.49
411885 ncm-dbt-01 583656 500 155 95 250 +41.89 ± 14.87 0 34 123 92 1 +83.57 ± 30.77
411884 ncm-dbt-02 586266 500 149 96 255 +36.97 ± 14.44 0 33 132 84 1 +73.34 ± 29.56
411883 ncm-dbt-05 584664 500 155 84 261 +49.67 ± 14.38 1 25 126 98 0 +102.97 ± 30.28
411882 ncm-dbt-03 589240 500 144 95 261 +34.16 ± 14.39 0 34 134 81 1 +67.55 ± 29.3
411881 ncm-dbt-04 570469 500 163 112 225 +35.56 ± 13.81 0 30 139 81 0 +71.89 ± 28.58
411880 ncm-dbt-01 583824 500 141 99 260 +29.25 ± 14.06 1 31 144 73 1 +58.93 ± 27.95

Commit

Commit ID 69639d764bde566e524b8c2566119bf677cb2622
Author Linmiao Xu
Date 2023-02-23 12:27:57 UTC
Reintroduce nnue pawn scaling with lower lazy thresholds Params found with the nevergrad TBPSA optimizer via nevergrad4sf modified to: * use SPRT LLR with fishtest STC elo gainer bounds [0, 2] as the objective function * increase the game batch size after each new optimal point is found The params were the optimal point after TBPSA iteration 7 and 160 nevergrad evaluations with: * initial batch size of 96 games per evaluation * batch size increase of 64 games after each iteration * a budget of 512 evaluations * TC: fixed 1.5 million nodes per move, no time limit nevergrad4sf enables optimizing stockfish params with TBPSA: https://github.com/vondele/nevergrad4sf Using pentanomial game results with smaller game batch sizes was inspired by: Use of SPRT LLR calculated from pentanomial game results as the objective function was an experiment at maximizing the information from game batches to reduce the computational cost for TBPSA to converge on good parameters. For the exact code used to find the params: https://github.com/linrock/tuning-fork Passed STC: https://tests.stockfishchess.org/tests/view/63f4ef5ee74a12625bcd114a LLR: 2.94 (-2.94,2.94) <0.00,2.00> Total: 66552 W: 17736 L: 17390 D: 31426 Ptnml(0-2): 164, 7229, 18166, 7531, 186 Passed LTC: https://tests.stockfishchess.org/tests/view/63f56028e74a12625bcd2550 LLR: 2.94 (-2.94,2.94) <0.50,2.50> Total: 71264 W: 19150 L: 18787 D: 33327 Ptnml(0-2): 23, 6728, 21771, 7083, 27 closes https://github.com/official-stockfish/Stockfish/pull/4401 bench 3687580
Copyright 2011–2025 Next Chess Move LLC