Dev Builds » 20200816-0620

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 08:39:01 1961712 3342 2409 53 880 +304.74 +/- 11.48
ncm-et-4 08:40:09 1966399 3309 2462 38 809 +324.58 +/- 11.96
ncm-et-9 08:41:41 1966956 3338 2437 46 855 +312.69 +/- 11.64
ncm-et-10 08:39:34 1959811 3329 2420 52 857 +309.17 +/- 11.63
ncm-et-13 08:40:24 1960945 3327 2441 42 844 +316.12 +/- 11.71
ncm-et-15 08:40:38 1965994 3355 2452 33 870 +316.08 +/- 11.52
20000 14621 264 5115 +313.8 +/- 4.75

Test Detail

ID Host Started (UTC) Duration Base NPS Games Wins Losses Draws Elo CLI PGN
135061 ncm-et-4 2020-08-16 19:41 00:46:42 1980820 309 222 4 83 +305.11 +/- 37.65
135060 ncm-et-13 2020-08-16 19:36 00:51:19 1962149 327 229 3 95 +295.36 +/- 35.0
135059 ncm-et-10 2020-08-16 19:36 00:51:15 1959402 329 242 6 81 +313.43 +/- 38.22
135058 ncm-et-3 2020-08-16 19:34 00:53:42 1964455 342 251 10 81 +304.54 +/- 38.14
135057 ncm-et-9 2020-08-16 19:34 00:53:17 1967849 338 249 7 82 +312.46 +/- 37.98
135056 ncm-et-15 2020-08-16 19:33 00:54:46 1968169 355 259 3 93 +316.16 +/- 35.47
135055 ncm-et-4 2020-08-16 18:22 01:18:05 1973115 500 369 4 127 +322.67 +/- 30.3
135054 ncm-et-10 2020-08-16 18:17 01:18:30 1954958 500 357 5 138 +304.07 +/- 29.02
135053 ncm-et-13 2020-08-16 18:16 01:18:54 1957552 500 353 4 143 +299.97 +/- 28.44
135052 ncm-et-3 2020-08-16 18:16 01:17:14 1961535 500 367 6 127 +316.8 +/- 30.35
135051 ncm-et-9 2020-08-16 18:16 01:17:19 1971721 500 364 6 130 +312.48 +/- 29.98
135050 ncm-et-15 2020-08-16 18:14 01:17:36 1964633 500 357 5 138 +304.07 +/- 29.02
135049 ncm-et-4 2020-08-16 17:00 01:21:05 1928439 500 382 10 108 +333.32 +/- 33.01
135048 ncm-et-10 2020-08-16 16:59 01:17:02 1969922 500 361 10 129 +302.7 +/- 30.16
135047 ncm-et-3 2020-08-16 16:58 01:17:03 1962003 500 347 13 140 +280.42 +/- 28.93
135046 ncm-et-9 2020-08-16 16:58 01:17:02 1962475 500 369 5 126 +321.19 +/- 30.45
135045 ncm-et-13 2020-08-16 16:56 01:19:01 1963686 500 361 7 132 +306.84 +/- 29.77
135044 ncm-et-15 2020-08-16 16:56 01:17:21 1964472 500 361 6 133 +308.24 +/- 29.62
135043 ncm-et-4 2020-08-16 15:40 01:18:15 1973587 500 366 8 126 +312.48 +/- 30.51
135042 ncm-et-10 2020-08-16 15:40 01:18:02 1959406 500 370 10 120 +315.35 +/- 31.3
135041 ncm-et-9 2020-08-16 15:38 01:17:58 1967857 500 369 13 118 +309.64 +/- 31.53
135040 ncm-et-13 2020-08-16 15:37 01:17:58 1960316 500 376 10 114 +324.17 +/- 32.12
135039 ncm-et-3 2020-08-16 15:37 01:19:24 1959081 500 371 9 120 +318.25 +/- 31.3
135038 ncm-et-15 2020-08-16 15:37 01:18:06 1965557 500 363 4 133 +313.91 +/- 29.56
135037 ncm-et-4 2020-08-16 14:21 01:18:40 1971273 500 371 5 124 +324.17 +/- 30.71
135036 ncm-et-10 2020-08-16 14:20 01:18:55 1957860 500 369 8 123 +316.8 +/- 30.9
135035 ncm-et-9 2020-08-16 14:20 01:17:54 1961537 500 357 3 140 +306.84 +/- 28.72
135034 ncm-et-15 2020-08-16 14:18 01:17:20 1965074 500 360 4 136 +309.64 +/- 29.21
135033 ncm-et-13 2020-08-16 14:18 01:18:03 1962455 500 384 3 113 +347.77 +/- 32.18
135032 ncm-et-3 2020-08-16 14:18 01:17:56 1961535 500 356 6 138 +301.33 +/- 29.05
135031 ncm-et-4 2020-08-16 13:01 01:18:20 1968313 500 376 5 119 +331.77 +/- 31.38
135030 ncm-et-3 2020-08-16 13:01 01:16:07 1962149 500 352 6 142 +295.94 +/- 28.61
135029 ncm-et-10 2020-08-16 13:00 01:18:22 1957099 500 367 9 124 +312.48 +/- 30.77
135028 ncm-et-13 2020-08-16 13:00 01:17:12 1960309 500 364 10 126 +306.84 +/- 30.53
135027 ncm-et-9 2020-08-16 13:00 01:18:43 1968629 500 360 9 131 +302.7 +/- 29.91
135026 ncm-et-15 2020-08-16 13:00 01:17:29 1973585 500 375 4 121 +331.77 +/- 31.08
135025 ncm-et-3 2020-08-16 11:42 01:17:35 1961228 500 365 3 132 +318.25 +/- 29.64
135024 ncm-et-10 2020-08-16 11:42 01:17:28 1960031 500 354 4 142 +301.33 +/- 28.54
135023 ncm-et-4 2020-08-16 11:41 01:19:02 1969252 500 376 2 122 +336.46 +/- 30.86
135022 ncm-et-13 2020-08-16 11:41 01:17:57 1960154 500 374 5 121 +328.7 +/- 31.11
135021 ncm-et-15 2020-08-16 11:41 01:18:00 1960470 500 377 7 116 +330.23 +/- 31.84
135020 ncm-et-9 2020-08-16 11:39 01:19:28 1968624 500 369 3 128 +324.17 +/- 30.13

Commit

Commit ID cd0b8b4cf28208fffef931322749205a0ddc6066
Author Joost VandeVondele
Date 2020-08-16 06:20:32 UTC
Use NNUE more for fortresses Increases the use of NNUE evaluation in positions without captures/pawn moves, by increasing the NNUEThreshold threshold with rule50_count. This patch will force Stockfish to use NNUE eval more and more in materially unbalanced positions, when it seems that the classical eval is struggling to win and only manages to shuffle. This will ask the (slower) NNUE eval to double-check the potential fortress branches of the search tree, but only when necessary. passed STC: https://tests.stockfishchess.org/tests/view/5f36f1bf11a9b1a1dbf192d8 LLR: 2.93 (-2.94,2.94) {-0.50,1.50} Total: 51824 W: 5836 L: 5653 D: 40335 Ptnml(0-2): 264, 4356, 16512, 4493, 287 passed LTC: https://tests.stockfishchess.org/tests/view/5f37836111a9b1a1dbf1936d LLR: 2.93 (-2.94,2.94) {0.25,1.75} Total: 29768 W: 1747 L: 1590 D: 26431 Ptnml(0-2): 33, 1347, 11977, 1484, 43 closes https://github.com/official-stockfish/Stockfish/pull/3011 Bench: 4173967
Copyright 2011–2024 Next Chess Move LLC