Dev Builds » 20210519-1834

Use this dev build

NCM plays each Stockfish dev build 20,000 times against Stockfish 7. This yields an approximate Elo difference and establishes confidence in the strength of the dev builds.

Summary

Host Duration Avg Base NPS Games Wins Losses Draws Elo
ncm-et-3 08:12:37 1965906 3337 2757 8 572 +405.98 +/- 14.23
ncm-et-4 08:12:33 1962160 3352 2694 14 644 +381.24 +/- 13.41
ncm-et-9 08:12:10 1962361 3325 2675 21 629 +379.96 +/- 13.58
ncm-et-10 08:12:16 1954779 3310 2660 15 635 +380.82 +/- 13.51
ncm-et-13 08:12:31 1961121 3348 2722 12 614 +391.0 +/- 13.73
ncm-et-15 08:12:23 1960953 3328 2683 17 628 +382.74 +/- 13.59
20000 16191 87 3722 +386.77 +/- 5.57

Test Detail

ID Host Started (UTC) Duration Base NPS Games Wins Losses Draws Elo CLI PGN
147123 ncm-et-10 2021-05-21 02:35 00:46:52 1953308 310 255 2 53 +397.85 +/- 47.63
147122 ncm-et-9 2021-05-21 02:34 00:48:13 1961076 325 259 3 63 +370.13 +/- 43.53
147121 ncm-et-15 2021-05-21 02:33 00:48:58 1959083 328 259 4 65 +360.94 +/- 42.83
147120 ncm-et-3 2021-05-21 02:32 00:50:18 1965689 337 282 1 54 +417.12 +/- 47.15
147119 ncm-et-13 2021-05-21 02:30 00:51:31 1960780 348 280 1 67 +383.37 +/- 42.06
147118 ncm-et-4 2021-05-21 02:29 00:52:51 1957401 352 280 2 70 +372.04 +/- 41.18
147117 ncm-et-10 2021-05-21 01:21 01:13:21 1962150 500 410 2 88 +397.72 +/- 36.65
147116 ncm-et-15 2021-05-21 01:19 01:12:58 1970018 500 401 4 95 +375.98 +/- 35.27
147115 ncm-et-9 2021-05-21 01:19 01:13:40 1962465 500 398 3 99 +372.25 +/- 34.49
147114 ncm-et-3 2021-05-21 01:18 01:12:59 1960039 500 413 0 87 +408.38 +/- 36.75
147113 ncm-et-13 2021-05-21 01:16 01:13:20 1955625 500 408 2 90 +393.6 +/- 36.22
147112 ncm-et-4 2021-05-21 01:15 01:13:14 1961074 500 409 1 90 +397.72 +/- 36.17
147111 ncm-et-15 2021-05-21 00:06 01:12:44 1964558 500 405 2 93 +387.57 +/- 35.6
147110 ncm-et-10 2021-05-21 00:05 01:14:52 1954200 500 389 2 109 +357.94 +/- 32.76
147109 ncm-et-9 2021-05-21 00:05 01:13:43 1961552 500 401 2 97 +379.78 +/- 34.83
147108 ncm-et-3 2021-05-21 00:04 01:14:00 1972033 500 409 2 89 +395.65 +/- 36.43
147107 ncm-et-4 2021-05-21 00:02 01:12:14 1970336 500 408 1 91 +395.65 +/- 35.96
147106 ncm-et-13 2021-05-21 00:02 01:13:45 1962461 500 393 1 106 +366.78 +/- 33.19
147105 ncm-et-15 2021-05-20 22:52 01:13:46 1976537 500 402 1 97 +383.64 +/- 34.77
147104 ncm-et-10 2021-05-20 22:50 01:14:23 1947065 500 395 4 101 +364.98 +/- 34.16
147103 ncm-et-3 2021-05-20 22:49 01:13:45 1973585 500 411 1 88 +401.92 +/- 36.6
147102 ncm-et-9 2021-05-20 22:48 01:16:08 1942441 500 408 5 87 +387.57 +/- 36.91
147101 ncm-et-13 2021-05-20 22:48 01:13:51 1956964 500 414 1 85 +408.38 +/- 37.27
147100 ncm-et-4 2021-05-20 22:47 01:14:45 1946826 500 395 1 104 +370.41 +/- 33.52
147099 ncm-et-15 2021-05-20 21:36 01:14:58 1947303 500 400 0 100 +381.69 +/- 34.16
147098 ncm-et-10 2021-05-20 21:36 01:14:14 1955437 500 391 1 108 +363.2 +/- 32.86
147097 ncm-et-4 2021-05-20 21:35 01:11:34 1961269 500 393 3 104 +363.2 +/- 33.62
147096 ncm-et-9 2021-05-20 21:34 01:13:36 1967853 500 411 0 89 +404.05 +/- 36.31
147095 ncm-et-13 2021-05-20 21:34 01:13:20 1960622 500 419 1 80 +419.61 +/- 38.47
147094 ncm-et-3 2021-05-20 21:34 01:14:58 1955671 500 396 1 103 +372.25 +/- 33.69
147093 ncm-et-10 2021-05-20 20:22 01:13:03 1953891 500 409 3 88 +393.6 +/- 36.68
147092 ncm-et-13 2021-05-20 20:20 01:12:47 1967553 500 397 4 99 +368.59 +/- 34.52
147091 ncm-et-9 2021-05-20 20:20 01:13:35 1971748 500 398 3 99 +372.25 +/- 34.49
147090 ncm-et-4 2021-05-20 20:20 01:14:31 1965867 500 408 3 89 +391.57 +/- 36.46
147089 ncm-et-15 2021-05-20 20:20 01:15:46 1939926 500 393 3 104 +363.2 +/- 33.62
147088 ncm-et-3 2021-05-20 20:19 01:13:30 1971873 500 426 2 72 +433.94 +/- 40.69
147087 ncm-et-9 2021-05-20 19:06 01:13:15 1969398 500 400 5 95 +372.25 +/- 35.28
147086 ncm-et-10 2021-05-20 19:06 01:15:31 1957405 500 411 1 88 +401.92 +/- 36.6
147085 ncm-et-15 2021-05-20 19:06 01:13:13 1969247 500 423 3 74 +424.28 +/- 40.13
147084 ncm-et-4 2021-05-20 19:06 01:13:24 1972349 500 401 3 96 +377.87 +/- 35.05
147083 ncm-et-3 2021-05-20 19:06 01:13:07 1962457 500 420 1 79 +421.93 +/- 38.73
147082 ncm-et-13 2021-05-20 19:06 01:13:57 1963845 500 411 2 87 +399.81 +/- 36.87

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 2011–2024 Next Chess Move LLC