Dev Builds » 20210618-1929

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 06:39:35 583939 4000 487 1595 1918 -98.82 ± 5.1 51 1091 773 85 0 -204.31 ± 12.22
ncm-dbt-02 06:40:18 586145 4000 450 1599 1951 -102.69 ± 5.11 55 1118 749 77 1 -213.6 ± 12.42
ncm-dbt-03 06:38:13 585565 4000 501 1578 1921 -95.91 ± 5.2 45 1094 754 107 0 -198.34 ± 12.4
ncm-dbt-04 06:39:11 566661 4000 490 1538 1972 -93.2 ± 5.16 41 1073 780 105 1 -192.71 ± 12.19
ncm-dbt-05 06:39:31 582231 4000 460 1609 1931 -102.69 ± 5.09 56 1112 758 73 1 -213.36 ± 12.34
20000 2388 7919 9693 -98.65 ± 2.3 248 5488 3814 447 3 -204.36 ± 5.5

Test Detail

ID Host Base NPS Games WLD Standard Elo Ptnml(0-2) Gamepair Elo CLI PGN
454033 ncm-dbt-01 586393 500 57 203 240 -104.49 ± 14.66 7 143 89 11 0 -217.85 ± 36.33
454032 ncm-dbt-02 585126 500 58 207 235 -106.77 ± 13.89 7 141 96 6 0 -223.94 ± 34.66
454031 ncm-dbt-04 567918 500 65 185 250 -85.04 ± 14.92 4 130 98 18 0 -174.55 ± 34.58
454030 ncm-dbt-05 579166 500 65 214 221 -106.77 ± 15.42 13 133 94 10 0 -211.87 ± 35.26
454029 ncm-dbt-03 587792 500 66 199 235 -94.69 ± 14.69 6 134 97 13 0 -194.57 ± 34.74
454028 ncm-dbt-01 584327 500 57 201 242 -102.97 ± 14.02 6 140 96 8 0 -215.85 ± 34.78
454027 ncm-dbt-02 585000 500 53 202 245 -106.77 ± 14.67 10 137 95 8 0 -217.85 ± 34.98
454026 ncm-dbt-04 568713 500 66 180 254 -80.63 ± 13.79 2 123 112 13 0 -167.53 ± 32.08
454025 ncm-dbt-05 578177 500 54 198 248 -102.97 ± 15.23 9 138 92 10 1 -211.87 ± 35.7
454024 ncm-dbt-03 582402 500 64 212 224 -106.01 ± 15.55 9 145 81 15 0 -217.85 ± 38.05
454023 ncm-dbt-01 582485 500 66 200 234 -95.44 ± 14.25 4 138 96 12 0 -200.24 ± 34.91
454022 ncm-dbt-02 586985 500 46 205 249 -114.45 ± 14.05 6 155 81 8 0 -247.41 ± 38.13
454021 ncm-dbt-03 585717 500 60 187 253 -90.22 ± 14.46 3 136 96 15 0 -189.0 ± 34.95
454020 ncm-dbt-04 565666 500 55 187 258 -93.95 ± 14.82 6 133 99 11 1 -194.57 ± 34.34
454019 ncm-dbt-05 583656 500 54 199 247 -103.73 ± 13.55 6 138 101 5 0 -217.85 ± 33.59
453462 ncm-dbt-01 584411 500 60 190 250 -92.45 ± 14.65 7 128 103 12 0 -187.16 ± 33.6
453461 ncm-dbt-02 587240 500 58 194 248 -96.94 ± 13.96 4 138 98 10 0 -204.07 ± 34.47
453460 ncm-dbt-04 566928 500 60 198 242 -98.44 ± 13.98 4 140 96 10 0 -207.95 ± 34.86
453459 ncm-dbt-05 584243 500 41 205 254 -118.33 ± 13.35 5 159 81 5 0 -261.07 ± 38.02
453458 ncm-dbt-03 585632 500 64 183 253 -84.3 ± 14.33 4 125 107 14 0 -172.78 ± 32.96
453457 ncm-dbt-04 566454 500 67 204 229 -97.69 ± 15.17 6 141 87 16 0 -202.15 ± 36.72
453456 ncm-dbt-01 584159 500 56 201 243 -103.73 ± 15.1 9 139 90 12 0 -211.87 ± 36.13
453455 ncm-dbt-05 579372 500 55 204 241 -106.77 ± 14.67 6 149 83 12 0 -226.0 ± 37.66
453454 ncm-dbt-02 585717 500 63 194 243 -93.2 ± 14.66 6 132 99 13 0 -190.85 ± 34.36
453453 ncm-dbt-03 588345 500 66 199 235 -94.69 ± 14.69 4 140 91 15 0 -198.34 ± 35.92
453452 ncm-dbt-04 565666 500 65 193 242 -90.96 ± 14.47 4 134 98 14 0 -189.0 ± 34.56
453451 ncm-dbt-05 583363 500 52 196 252 -102.97 ± 13.71 4 144 94 8 0 -219.87 ± 35.19
453450 ncm-dbt-01 580821 500 72 192 236 -85.04 ± 13.9 2 129 106 13 0 -178.11 ± 33.1
453449 ncm-dbt-02 586520 500 59 204 237 -103.73 ± 15.24 9 139 91 10 1 -213.85 ± 35.9
453448 ncm-dbt-03 580199 500 62 193 245 -93.2 ± 14.95 6 134 95 15 0 -190.85 ± 35.14
453447 ncm-dbt-01 582569 500 56 201 243 -103.73 ± 14.34 7 140 94 9 0 -215.85 ± 35.23
453446 ncm-dbt-04 568514 500 50 193 257 -102.22 ± 15.23 11 132 96 11 0 -204.07 ± 34.89
453445 ncm-dbt-05 584034 500 61 199 240 -98.44 ± 14.44 7 134 99 10 0 -202.15 ± 34.28
453444 ncm-dbt-02 585843 500 53 192 255 -99.2 ± 15.04 8 136 93 13 0 -202.15 ± 35.52
453443 ncm-dbt-03 586350 500 68 201 231 -94.69 ± 14.39 5 135 98 12 0 -196.45 ± 34.53
453442 ncm-dbt-04 563431 500 62 198 240 -96.94 ± 14.27 4 140 94 12 0 -204.07 ± 35.31
453441 ncm-dbt-01 586350 500 63 207 230 -102.97 ± 14.49 9 134 99 8 0 -209.91 ± 34.19
453440 ncm-dbt-02 586731 500 60 201 239 -100.7 ± 14.01 5 140 96 9 0 -211.87 ± 34.83
453439 ncm-dbt-05 585843 500 78 194 228 -82.1 ± 14.42 6 117 114 13 0 -164.07 ± 31.75
453438 ncm-dbt-03 588089 500 51 204 245 -109.83 ± 14.37 8 145 89 8 0 -230.16 ± 36.26

Commit

Commit ID 86afb6a7cf33640a4d7bb7dbb2a8203450e34938
Author SFisGOD
Date 2021-06-18 19:29:14 UTC
Update default net to nn-aa9d7eeb397e.nnue Optimization of vondele's nn-33c9d39e5eb6.nnue using SPSA https://tests.stockfishchess.org/tests/view/60ca68be457376eb8bcab28b Setting: ck values are default based on how large the parameters are The new values for this net are the raw values at the end of the tuning (80k games) The significant changes are in buckets 1 and 2 (5-12 pieces) so the main difference is in playing endgames if we compare it to nn-33c9. There is also change in bucket 7 (29-32 pieces) but not as substantial as the changes in buckets 1 and 2. If we interpret the changes based on an experiment a few months ago, this new net plays more optimistically during endgames and less optimistically during openings. STC: LLR: 2.93 (-2.94,2.94) <-0.50,2.50> Total: 49504 W: 4246 L: 4053 D: 41205 Ptnml(0-2): 140, 3282, 17749, 3407, 174 https://tests.stockfishchess.org/tests/view/60cbd752457376eb8bcab478 LTC: LLR: 2.95 (-2.94,2.94) <0.50,3.50> Total: 88720 W: 4926 L: 4651 D: 79143 Ptnml(0-2): 105, 4048, 35793, 4295, 119 https://tests.stockfishchess.org/tests/view/60cc7828457376eb8bcab4fa closes https://github.com/official-stockfish/Stockfish/pull/3566 Bench: 4758885
Copyright 2011–2026 Next Chess Move LLC