Dev Builds » 20200927-0624

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 06:42:54 581742 3976 227 1872 1877 -152.91 ± 4.57 88 1484 401 15 0 -366.06 ± 17.03
ncm-dbt-02 06:42:30 586497 4030 232 1933 1865 -156.42 ± 4.73 115 1488 395 17 0 -369.59 ± 17.16
ncm-dbt-03 06:41:01 586763 4000 232 1896 1872 -153.86 ± 4.82 110 1468 398 24 0 -360.55 ± 17.11
ncm-dbt-04 06:41:32 566335 3998 237 1903 1858 -154.16 ± 4.72 108 1465 411 15 0 -362.66 ± 16.81
ncm-dbt-05 06:42:58 580685 3996 225 1906 1865 -155.83 ± 4.56 94 1508 381 15 0 -376.26 ± 17.48
20000 1153 9510 9337 -154.64 ± 2.09 515 7413 1986 86 0 -366.96 ± 7.64

Test Detail

ID Host Base NPS Games WLD Standard Elo Ptnml(0-2) Gamepair Elo CLI PGN
447981 ncm-dbt-02 587707 30 0 15 15 -190.77 ± 76.29 3 9 3 0 0 -381.58 ± 480.22
447980 ncm-dbt-01 580199 476 38 220 218 -139.95 ± 13.29 7 171 57 3 0 -326.64 ± 45.74
447979 ncm-dbt-05 582819 496 27 236 233 -156.12 ± 12.6 11 188 48 1 0 -380.15 ± 50.02
447978 ncm-dbt-04 564762 498 29 230 239 -148.69 ± 13.94 15 173 59 2 0 -335.66 ± 44.87
447977 ncm-dbt-03 587409 500 32 246 222 -158.93 ± 14.04 18 180 50 2 0 -366.78 ± 49.02
447976 ncm-dbt-02 586181 500 26 240 234 -158.93 ± 13.64 15 187 45 3 0 -377.87 ± 51.84
447975 ncm-dbt-01 581444 500 27 235 238 -153.86 ± 13.69 14 183 50 3 0 -359.68 ± 49.05
447974 ncm-dbt-05 578753 500 32 234 234 -148.85 ± 13.7 11 185 49 5 0 -349.43 ± 49.46
447973 ncm-dbt-04 562495 500 25 246 229 -164.93 ± 11.93 12 197 41 0 0 -419.6 ± 54.3
447972 ncm-dbt-03 586901 500 20 234 246 -158.93 ± 12.8 11 195 41 3 0 -393.6 ± 54.43
447971 ncm-dbt-02 587919 500 24 238 238 -158.93 ± 13.01 14 187 48 1 0 -381.7 ± 50.02
447970 ncm-dbt-01 580116 500 32 233 235 -148.02 ± 12.94 10 183 55 2 0 -349.43 ± 46.59
447969 ncm-dbt-04 568037 500 30 229 241 -146.36 ± 13.8 13 176 58 3 0 -333.32 ± 45.35
447968 ncm-dbt-05 578588 500 29 239 232 -155.54 ± 12.34 9 194 45 2 0 -385.59 ± 51.85
447967 ncm-dbt-03 586266 500 34 229 237 -143.07 ± 13.15 8 183 55 4 0 -336.46 ± 46.65
447966 ncm-dbt-02 585379 500 28 240 232 -157.23 ± 13.32 15 183 51 1 0 -370.41 ± 48.42
447965 ncm-dbt-01 583279 500 27 245 228 -162.35 ± 12.36 13 192 45 0 0 -401.91 ± 51.63
447964 ncm-dbt-04 566730 500 38 242 220 -150.51 ± 13.44 12 183 52 3 0 -352.8 ± 48.05
447963 ncm-dbt-05 580240 500 30 242 228 -157.23 ± 12.68 12 189 48 1 0 -381.7 ± 50.02
447962 ncm-dbt-03 586985 500 29 229 242 -147.19 ± 13.96 12 181 52 5 0 -339.63 ± 47.98
447961 ncm-dbt-02 585253 500 29 238 233 -154.7 ± 12.61 10 191 47 2 0 -377.87 ± 50.67
447960 ncm-dbt-01 583531 500 34 243 223 -154.7 ± 13.45 15 180 54 1 0 -359.68 ± 46.95
447959 ncm-dbt-04 566928 500 27 233 240 -152.18 ± 13.16 12 184 52 2 0 -359.68 ± 48.01
447958 ncm-dbt-05 582569 500 24 236 240 -157.23 ± 13.32 15 183 51 1 0 -370.41 ± 48.42
447957 ncm-dbt-03 586731 500 31 238 231 -153.02 ± 12.26 9 190 50 1 0 -374.11 ± 48.94
447956 ncm-dbt-02 585717 500 31 238 231 -153.02 ± 13.53 14 181 53 2 0 -356.21 ± 47.52
447955 ncm-dbt-04 566849 500 27 239 234 -157.23 ± 13.52 14 187 46 3 0 -374.11 ± 51.25
447954 ncm-dbt-01 583405 500 27 231 242 -150.51 ± 12.4 9 187 53 1 0 -363.2 ± 47.43
447953 ncm-dbt-03 586097 500 33 228 239 -143.07 ± 13.15 9 180 58 3 0 -333.32 ± 45.35
447952 ncm-dbt-05 577848 500 24 231 245 -153.02 ± 12.48 9 191 48 2 0 -374.11 ± 50.1
447951 ncm-dbt-02 587155 500 35 239 226 -150.51 ± 14.02 15 177 55 3 0 -342.85 ± 46.64
447950 ncm-dbt-04 567799 500 31 240 229 -154.7 ± 13.45 15 180 54 1 0 -359.68 ± 46.95
447949 ncm-dbt-03 587325 500 31 245 224 -158.93 ± 13.64 17 181 51 1 0 -370.41 ± 48.42
447948 ncm-dbt-01 584285 500 17 232 251 -159.78 ± 13.6 14 191 41 4 0 -385.59 ± 54.3
447947 ncm-dbt-05 579413 500 28 247 225 -163.21 ± 12.53 12 197 39 2 0 -410.58 ± 55.96
447946 ncm-dbt-02 586477 500 29 241 230 -157.23 ± 13.32 14 186 48 2 0 -374.11 ± 50.1
447945 ncm-dbt-04 567086 500 30 244 226 -158.93 ± 13.23 15 185 49 1 0 -377.87 ± 49.47
447944 ncm-dbt-03 586393 500 22 247 231 -168.4 ± 15.72 26 178 41 5 0 -377.87 ± 54.12
447943 ncm-dbt-02 586689 500 30 244 226 -158.93 ± 13.64 15 187 45 3 0 -377.87 ± 51.84
447942 ncm-dbt-01 577684 500 25 233 242 -153.86 ± 11.53 6 197 46 1 0 -389.56 ± 51.18
447941 ncm-dbt-05 585253 500 31 241 228 -155.54 ± 13.4 15 181 53 1 0 -363.2 ± 47.43

Commit

Commit ID 1dbd2a1ad548b3ca676f7da949e1a998c64b836b
Author Stéphane Nicolet
Date 2020-09-27 06:24:50 UTC
Tweak nnue scaling to keep more material Current master uses a constant scale factor of 5/4 = 1.25 for the output of the NNUE network, for compatibility with search and classical evaluation. We modify this scale factor to make it dependent on the phase of the game, going from about 1.5 in the opening to 1.0 for pure pawn endgames. This helps Stockfish to avoid exchanges of pieces (heavy pieces in particular) when she has the advantage, keeping more material on the board when attacking. Passed STC: LLR: 2.95 (-2.94,2.94) {-0.25,1.25} Total: 14744 W: 1771 L: 1599 D: 11374 Ptnml(0-2): 87, 1184, 4664, 1344, 93 https://tests.stockfishchess.org/tests/view/5f6fb0a63b22d6afa506904f Passed LTC: LLR: 2.95 (-2.94,2.94) {0.25,1.25} Total: 8912 W: 512 L: 393 D: 8007 Ptnml(0-2): 7, 344, 3637, 459, 9 https://tests.stockfishchess.org/tests/view/5f6fcf533b22d6afa5069066 closes https://github.com/official-stockfish/Stockfish/pull/3154 Bench: 3943952
Copyright 2011–2026 Next Chess Move LLC