Dev Builds » 20210519-1834

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 06:49:06 581801 4010 341 1805 1864 -132.98 ± 5.09 100 1304 561 40 0 -288.26 ± 14.38
ncm-dbt-02 06:48:40 585371 4000 293 1781 1926 -135.76 ± 4.92 79 1371 509 41 0 -304.42 ± 15.11
ncm-dbt-03 06:48:01 586966 4000 335 1814 1851 -134.85 ± 5.27 106 1325 511 58 0 -292.3 ± 15.08
ncm-dbt-04 06:47:26 566557 4000 282 1800 1918 -138.79 ± 4.94 94 1362 512 32 0 -309.64 ± 15.06
ncm-dbt-05 06:49:32 580495 3990 325 1798 1867 -134.62 ± 5.06 91 1334 528 41 1 -296.76 ± 14.84
20000 1576 8998 9426 -135.4 ± 2.26 470 6696 2621 212 1 -298.15 ± 6.65

Test Detail

ID Host Base NPS Games WLD Standard Elo Ptnml(0-2) Gamepair Elo CLI PGN
437477 ncm-dbt-01 581194 10 0 5 5 -190.62 ± 11.11 0 5 0 0 0 -1199.83 ± 231.31
437476 ncm-dbt-05 583824 490 48 220 222 -127.37 ± 14.99 12 155 71 7 0 -271.21 ± 40.83
437475 ncm-dbt-02 584495 500 45 217 238 -124.6 ± 13.97 8 162 74 6 0 -273.0 ± 39.95
437474 ncm-dbt-03 587452 500 48 227 225 -130.14 ± 14.57 12 161 71 6 0 -280.42 ± 40.83
437473 ncm-dbt-04 563509 500 43 226 231 -133.34 ± 14.68 13 163 68 6 0 -288.06 ± 41.76
437472 ncm-dbt-01 582444 500 45 226 229 -131.74 ± 14.71 14 158 73 5 0 -280.42 ± 40.21
437471 ncm-dbt-05 578424 500 43 217 240 -126.17 ± 14.62 13 153 79 5 0 -265.78 ± 38.54
437470 ncm-dbt-02 587028 500 34 210 256 -127.76 ± 13.92 8 166 70 6 0 -282.94 ± 41.13
437469 ncm-dbt-03 588089 500 42 228 230 -135.76 ± 14.98 16 159 70 5 0 -288.06 ± 41.11
437468 ncm-dbt-04 564801 500 39 219 242 -130.94 ± 14.39 11 164 69 6 0 -285.49 ± 41.44
437467 ncm-dbt-01 582193 500 41 231 228 -138.99 ± 13.85 10 175 60 5 0 -315.35 ± 44.57
437466 ncm-dbt-05 580240 500 38 232 230 -142.25 ± 14.66 14 172 58 6 0 -315.35 ± 45.33
437465 ncm-dbt-02 585169 500 43 221 236 -129.35 ± 14.07 6 175 60 9 0 -293.29 ± 44.36
437464 ncm-dbt-03 584706 500 42 223 235 -131.74 ± 15.53 14 163 63 10 0 -280.42 ± 43.24
437463 ncm-dbt-04 566493 500 33 226 241 -141.44 ± 13.59 10 177 59 4 0 -324.17 ± 44.97
437462 ncm-dbt-01 582736 500 46 214 240 -121.45 ± 14.17 11 150 85 4 0 -256.44 ± 36.95
437461 ncm-dbt-02 586181 500 36 235 229 -146.36 ± 13.61 12 178 57 3 0 -336.46 ± 45.77
437460 ncm-dbt-05 575924 500 45 225 230 -130.94 ± 14.22 10 166 68 6 0 -288.06 ± 41.76
437459 ncm-dbt-04 568554 500 29 229 242 -147.19 ± 13.58 12 179 56 3 0 -339.63 ± 46.2
437458 ncm-dbt-03 587664 500 39 233 228 -142.25 ± 14.3 14 170 62 4 0 -315.35 ± 43.81
437457 ncm-dbt-01 581735 500 42 231 227 -138.18 ± 14.23 13 167 66 4 0 -304.07 ± 42.38
437456 ncm-dbt-02 586393 500 24 218 258 -142.25 ± 13.18 10 176 62 2 0 -327.18 ± 43.69
437455 ncm-dbt-05 580613 500 44 224 232 -130.94 ± 15.05 10 170 61 8 1 -290.66 ± 44.0
437454 ncm-dbt-04 568395 500 32 219 249 -136.56 ± 13.73 10 171 65 4 0 -306.84 ± 42.73
437453 ncm-dbt-03 585801 500 46 225 229 -130.14 ± 15.06 13 161 68 8 0 -277.93 ± 41.74
437452 ncm-dbt-01 579620 500 38 222 240 -134.15 ± 13.8 13 159 77 1 0 -290.66 ± 38.7
437451 ncm-dbt-04 566849 500 33 233 234 -147.19 ± 14.33 15 174 57 4 0 -330.23 ± 45.79
437450 ncm-dbt-02 582736 500 44 225 231 -131.74 ± 14.2 11 164 70 5 0 -288.06 ± 41.11
437449 ncm-dbt-05 583070 500 37 231 232 -142.25 ± 13.56 12 172 64 2 0 -321.19 ± 42.95
437448 ncm-dbt-03 586774 500 37 221 242 -134.15 ± 13.61 8 173 64 5 0 -304.07 ± 43.1
437447 ncm-dbt-01 583154 500 50 218 232 -121.45 ± 15.3 10 160 68 12 0 -258.75 ± 41.54
437446 ncm-dbt-04 565744 500 44 224 232 -130.94 ± 13.51 8 168 70 4 0 -293.29 ± 41.07
437445 ncm-dbt-02 584916 500 33 230 237 -144.71 ± 14.41 13 177 54 6 0 -327.18 ± 47.0
437444 ncm-dbt-05 578712 500 38 225 237 -136.56 ± 13.17 8 174 65 3 0 -312.48 ± 42.67
437443 ncm-dbt-03 586477 500 46 228 226 -132.54 ± 15.83 15 163 61 11 0 -280.42 ± 43.82
437442 ncm-dbt-01 580613 500 41 226 233 -134.95 ± 14.65 14 162 69 5 0 -290.66 ± 41.43
437441 ncm-dbt-04 568116 500 29 224 247 -143.07 ± 13.91 15 166 68 1 0 -315.35 ± 41.44
437440 ncm-dbt-02 586054 500 34 225 241 -139.81 ± 13.83 11 173 62 4 0 -315.35 ± 43.81
437439 ncm-dbt-03 588771 500 35 229 236 -142.25 ± 15.18 14 175 52 9 0 -315.35 ± 47.52
437438 ncm-dbt-05 583154 500 32 224 244 -140.62 ± 13.99 12 172 62 4 0 -315.35 ± 43.81
437437 ncm-dbt-01 582527 500 38 232 230 -142.25 ± 14.48 15 168 63 4 0 -312.48 ± 43.44

Commit

Commit ID 754fc8a8b5ca7466926d54465eeb1df4d4a481ac
Author Stéphane Nicolet
Date 2021-05-19 18:34:37 UTC
Remove Tempo The Tempo variable was introduced 10 years ago in our search because the classical evaluation function was antisymmetrical in White and Black by design to gain speed: Eval(White to play) = -Eval(Black to play) Nowadays our neural networks know which side is to play in a position when they evaluate a position and are trained on real games, so the neural network encodes the advantage of moving as an output of search. This patch shows that the Tempo variable is not necessary anymore. STC: LLR: 2.94 (-2.94,2.94) <-2.50,0.50> Total: 33512 W: 2805 L: 2709 D: 27998 Ptnml(0-2): 80, 2209, 12095, 2279, 93 https://tests.stockfishchess.org/tests/view/60a44ceace8ea25a3ef03d30 LTC: LLR: 2.95 (-2.94,2.94) <-2.50,0.50> Total: 53920 W: 1807 L: 1760 D: 50353 Ptnml(0-2): 16, 1617, 23650, 1658, 19 https://tests.stockfishchess.org/tests/view/60a477f0ce8ea25a3ef03d49 We also tried a match (20000 games) at STC using purely classical, result was neutral: https://tests.stockfishchess.org/tests/view/60a4eebcce8ea25a3ef03db5 Note: there are two locations left in search.cpp where we assume antisymmetry of evaluation (in relation with a speed optimization for null moves in lines 770 and 1439), but as the values are just used for heuristic pruning this approximation should not hurt too much because the order of magnitude is still true most of the time. closes https://github.com/official-stockfish/Stockfish/pull/3481 Bench: 4015864
Copyright 2011–2025 Next Chess Move LLC