Dev Builds » 20210525-1940

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 06:46:12 583682 4010 351 1789 1870 -130.39 ± 4.98 88 1297 585 35 0 -283.78 ± 14.06
ncm-dbt-02 06:45:31 585770 4000 353 1777 1870 -129.35 ± 5.2 98 1281 568 53 0 -277.3 ± 14.3
ncm-dbt-03 06:46:03 585877 4000 345 1754 1901 -127.86 ± 5.09 82 1297 570 50 1 -277.93 ± 14.28
ncm-dbt-04 06:46:25 566679 3996 364 1795 1837 -130.19 ± 5.17 102 1272 579 45 0 -278.65 ± 14.16
ncm-dbt-05 06:45:27 582163 3994 346 1767 1881 -129.26 ± 4.95 73 1319 561 44 0 -284.86 ± 14.39
20000 1759 8882 9359 -129.41 ± 2.27 443 6466 2863 227 1 -280.49 ± 6.36

Test Detail

ID Host Base NPS Games WLD Standard Elo Ptnml(0-2) Gamepair Elo CLI PGN
455370 ncm-dbt-01 584285 10 3 5 2 -70.35 ± 177.62 1 1 2 1 0 -70.44 ± 307.49
455369 ncm-dbt-05 579083 494 47 215 232 -123.05 ± 13.75 5 165 70 7 0 -275.4 ± 41.13
455368 ncm-dbt-04 565194 496 42 214 240 -125.69 ± 14.88 10 161 68 9 0 -271.31 ± 41.7
455367 ncm-dbt-03 586816 500 41 216 243 -126.97 ± 13.58 8 163 75 4 0 -280.42 ± 39.57
455366 ncm-dbt-02 585464 500 49 221 230 -124.6 ± 13.97 7 165 71 7 0 -275.45 ± 40.84
455365 ncm-dbt-01 584537 500 43 217 240 -126.17 ± 13.42 8 161 78 3 0 -277.93 ± 38.66
455364 ncm-dbt-04 568196 500 49 227 224 -129.35 ± 14.91 14 156 74 6 0 -273.0 ± 39.95
455363 ncm-dbt-05 586604 500 37 218 245 -131.74 ± 12.93 5 175 66 4 0 -304.07 ± 42.38
455362 ncm-dbt-03 585295 500 50 223 227 -125.38 ± 14.63 10 161 71 8 0 -270.57 ± 40.83
455361 ncm-dbt-01 586350 500 44 223 233 -130.14 ± 14.23 10 165 69 6 0 -285.49 ± 41.44
455360 ncm-dbt-02 584916 500 48 226 226 -129.35 ± 15.86 17 153 71 9 0 -265.78 ± 40.81
455359 ncm-dbt-04 566691 500 46 219 235 -125.38 ± 14.63 11 158 74 7 0 -268.17 ± 39.97
455358 ncm-dbt-02 585632 500 34 223 243 -138.18 ± 14.76 13 170 60 7 0 -304.07 ± 44.5
455357 ncm-dbt-01 583614 500 46 221 233 -126.97 ± 14.11 10 160 75 5 0 -275.45 ± 39.63
455356 ncm-dbt-05 583028 500 42 222 236 -130.94 ± 13.51 10 162 76 2 0 -288.06 ± 39.11
455355 ncm-dbt-03 584076 500 39 211 250 -124.6 ± 14.8 12 155 76 7 0 -263.42 ± 39.41
455354 ncm-dbt-01 582402 500 45 228 227 -133.34 ± 13.81 12 161 75 2 0 -290.66 ± 39.4
455353 ncm-dbt-02 584369 500 45 206 249 -116.0 ± 15.0 11 148 82 9 0 -240.82 ± 37.9
455352 ncm-dbt-03 586985 500 42 231 227 -138.18 ± 14.06 13 166 68 3 0 -304.07 ± 41.65
455351 ncm-dbt-04 564762 500 49 224 227 -126.97 ± 13.76 8 164 73 5 0 -280.42 ± 40.21
455350 ncm-dbt-05 580738 500 35 230 235 -143.07 ± 14.81 14 174 55 7 0 -318.25 ± 46.48
455349 ncm-dbt-01 581693 500 45 214 241 -122.24 ± 14.66 11 154 78 7 0 -258.75 ± 38.88
455348 ncm-dbt-02 586774 500 45 231 224 -135.76 ± 15.14 14 166 62 8 0 -293.29 ± 43.72
455347 ncm-dbt-03 587028 500 49 216 235 -120.67 ± 14.01 7 160 76 7 0 -263.42 ± 39.41
455346 ncm-dbt-04 566533 500 33 219 248 -135.76 ± 14.29 15 158 75 2 0 -290.66 ± 39.4
455345 ncm-dbt-05 580199 500 48 218 234 -123.02 ± 14.33 6 168 66 10 0 -273.0 ± 42.27
455344 ncm-dbt-05 583572 500 47 223 230 -127.76 ± 14.93 13 157 73 7 0 -270.57 ± 40.25
455343 ncm-dbt-01 585169 500 40 220 240 -130.94 ± 14.22 10 166 68 6 0 -288.06 ± 41.76
455342 ncm-dbt-03 585169 500 35 223 242 -137.37 ± 14.08 12 168 66 4 0 -304.07 ± 42.38
455341 ncm-dbt-04 566533 500 53 235 212 -132.54 ± 15.52 17 155 71 7 0 -275.45 ± 40.84
455340 ncm-dbt-02 587240 500 47 209 244 -116.77 ± 14.53 11 146 87 6 0 -243.0 ± 36.62
455339 ncm-dbt-05 579166 500 54 223 223 -122.24 ± 13.66 9 154 84 3 0 -263.42 ± 37.11
455338 ncm-dbt-01 582235 500 49 229 222 -130.94 ± 13.87 12 158 78 2 0 -282.94 ± 38.55
455337 ncm-dbt-02 586139 500 36 240 224 -150.51 ± 13.24 11 185 51 3 0 -356.21 ± 48.54
455336 ncm-dbt-04 566020 500 52 237 211 -134.95 ± 15.48 19 152 74 5 0 -277.93 ± 39.92
455335 ncm-dbt-03 584243 500 43 218 239 -126.97 ± 14.45 11 159 74 6 0 -273.0 ± 39.95
455334 ncm-dbt-05 584916 500 36 218 246 -132.54 ± 14.01 11 164 71 4 0 -290.66 ± 40.76
455333 ncm-dbt-04 569509 500 40 220 240 -130.94 ± 13.51 8 168 70 4 0 -293.29 ± 41.07
455332 ncm-dbt-02 585632 500 49 221 230 -124.6 ± 14.64 14 148 84 4 0 -258.75 ± 37.2
455331 ncm-dbt-01 582861 500 36 232 232 -143.89 ± 14.07 14 171 62 3 0 -321.19 ± 43.77
455330 ncm-dbt-03 587409 500 46 216 238 -123.02 ± 15.44 9 165 64 11 1 -268.17 ± 42.75

Commit

Commit ID 9d53129075177cb11b63b43236556051ba60f7dd
Author Tomasz Sobczyk
Date 2021-05-25 19:40:51 UTC
Expose the lazy threshold for the feature transformer PSQT as a parameter. Definition of the lazy threshold moved to evaluate.cpp where all others are. Lazy threshold only used for real searches, not used for the "eval" call. This preserves the purity of NNUE evaluation, which is useful to verify consistency between the engine and the NNUE trainer. closes https://github.com/official-stockfish/Stockfish/pull/3499 No functional change
Copyright 2011–2026 Next Chess Move LLC