Dev Builds » 20200806-1437

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 06:32:23 584355 4008 62 2584 1362 -257.16 ± 6.65 604 1315 84 1 0 -663.6 ± 37.69
ncm-dbt-02 06:31:10 586221 4000 59 2529 1412 -250.48 ± 6.43 562 1347 90 1 0 -651.27 ± 36.38
ncm-dbt-03 06:31:04 584377 4000 64 2528 1408 -249.64 ± 6.44 559 1348 91 2 0 -645.56 ± 36.16
ncm-dbt-04 06:30:42 566943 4000 57 2540 1403 -252.32 ± 6.44 567 1352 78 3 0 -667.43 ± 39.11
ncm-dbt-05 06:32:41 581842 3992 78 2492 1422 -243.39 ± 6.25 523 1368 105 0 0 -627.37 ± 33.6
20000 320 12673 7007 -250.57 ± 2.88 2815 6730 448 7 0 -650.5 ± 16.13

Test Detail

ID Host Base NPS Games WLD Standard Elo Ptnml(0-2) Gamepair Elo CLI PGN
462309 ncm-dbt-01 585674 8 1 6 1 -254.34 ± 561.59 2 1 1 0 0 -337.69 ± 541.44
462308 ncm-dbt-05 580075 492 9 299 184 -235.14 ± 16.47 56 178 12 0 0 -640.8 ± 110.27
462307 ncm-dbt-03 584411 500 12 323 165 -253.02 ± 19.42 76 159 15 0 0 -603.84 ± 95.91
462306 ncm-dbt-04 565941 500 11 311 178 -240.82 ± 18.34 66 169 14 1 0 -592.29 ± 99.42
462305 ncm-dbt-02 585126 500 8 328 164 -263.41 ± 19.4 80 160 10 0 0 -676.04 ± 124.55
462304 ncm-dbt-01 582277 500 12 317 171 -246.3 ± 19.31 72 162 15 1 0 -581.4 ± 95.42
462303 ncm-dbt-03 585421 500 9 315 176 -247.41 ± 17.38 66 174 10 0 0 -676.04 ± 124.55
462302 ncm-dbt-05 584495 500 7 320 173 -255.29 ± 17.54 70 173 7 0 0 -739.05 ± 163.27
462301 ncm-dbt-01 583866 500 7 322 171 -257.59 ± 18.68 75 165 10 0 0 -676.04 ± 124.55
462300 ncm-dbt-02 588302 500 10 319 171 -250.76 ± 19.39 75 159 16 0 0 -592.27 ± 92.24
462299 ncm-dbt-04 568156 500 6 327 167 -264.59 ± 19.27 80 161 9 0 0 -694.7 ± 134.18
462298 ncm-dbt-03 583614 500 7 313 180 -247.41 ± 17.1 65 176 9 0 0 -694.7 ± 134.18
462297 ncm-dbt-05 584327 500 6 319 175 -255.29 ± 18.66 74 165 11 0 0 -659.13 ± 116.77
462296 ncm-dbt-01 582027 500 8 332 160 -268.16 ± 19.71 83 158 9 0 0 -694.7 ± 134.18
462295 ncm-dbt-02 586477 500 8 307 185 -239.73 ± 16.91 61 177 12 0 0 -643.67 ± 110.29
462294 ncm-dbt-04 566612 500 9 330 161 -264.59 ± 20.89 84 155 9 2 0 -629.43 ± 125.74
462293 ncm-dbt-03 584832 500 10 323 167 -255.29 ± 19.97 79 155 16 0 0 -592.27 ± 92.24
462292 ncm-dbt-01 586012 500 9 322 169 -255.29 ± 19.19 76 161 13 0 0 -629.41 ± 104.8
462291 ncm-dbt-05 582235 500 14 312 174 -238.65 ± 18.05 65 168 17 0 0 -581.38 ± 88.96
462290 ncm-dbt-02 586689 500 9 329 162 -263.41 ± 20.48 83 155 11 1 0 -629.43 ± 115.29
462289 ncm-dbt-04 563431 500 6 312 182 -247.41 ± 16.82 64 178 8 0 0 -715.51 ± 146.53
462288 ncm-dbt-03 584369 500 9 318 173 -250.76 ± 19.39 74 162 13 1 0 -603.86 ± 103.96
462287 ncm-dbt-02 584832 500 8 316 176 -249.64 ± 17.94 69 170 11 0 0 -659.13 ± 116.77
462286 ncm-dbt-01 587197 500 11 313 176 -243.0 ± 18.13 67 168 15 0 0 -603.84 ± 95.91
462285 ncm-dbt-05 581985 500 12 314 174 -243.0 ± 18.13 67 168 15 0 0 -603.84 ± 95.91
462284 ncm-dbt-04 566967 500 13 315 172 -243.0 ± 18.38 68 166 16 0 0 -592.27 ± 92.24
462283 ncm-dbt-03 581444 500 3 303 194 -240.82 ± 16.23 58 185 6 1 0 -715.55 ± 177.2
462282 ncm-dbt-02 587919 500 7 310 183 -244.09 ± 17.22 64 175 11 0 0 -659.13 ± 116.77
462281 ncm-dbt-05 583866 500 8 304 188 -236.51 ± 17.26 61 174 15 0 0 -603.84 ± 95.91
462280 ncm-dbt-01 584622 500 2 322 176 -263.41 ± 17.36 73 174 3 0 0 -887.58 ± 222.18
462279 ncm-dbt-04 569829 500 2 318 180 -258.74 ± 16.78 69 178 3 0 0 -887.58 ± 222.18
462278 ncm-dbt-03 585506 500 9 314 177 -246.3 ± 18.04 68 169 13 0 0 -629.41 ± 104.8
462277 ncm-dbt-05 579413 500 11 306 183 -235.44 ± 16.86 59 177 14 0 0 -616.18 ± 100.06
462276 ncm-dbt-02 584285 500 5 313 182 -249.64 ± 17.67 68 172 10 0 0 -676.04 ± 124.55
462275 ncm-dbt-01 583405 500 6 327 167 -264.59 ± 19.55 81 159 10 0 0 -676.04 ± 124.55
462274 ncm-dbt-04 568434 500 3 309 188 -247.41 ± 17.38 66 174 10 0 0 -676.04 ± 124.55
462273 ncm-dbt-03 585421 500 5 319 176 -256.44 ± 18.25 73 168 9 0 0 -694.7 ± 134.18
462272 ncm-dbt-02 586139 500 4 307 189 -244.09 ± 16.67 62 179 9 0 0 -694.7 ± 134.18
462271 ncm-dbt-04 566178 500 7 318 175 -253.02 ± 17.82 70 171 9 0 0 -694.7 ± 134.18
462270 ncm-dbt-05 578342 500 11 318 171 -248.52 ± 18.59 71 165 14 0 0 -616.18 ± 100.06
462269 ncm-dbt-01 584117 500 6 323 171 -259.9 ± 18.41 75 167 8 0 0 -715.51 ± 146.53

Commit

Commit ID 84f3e867903f62480c33243dd0ecbffd342796fc
Author nodchip
Date 2020-08-06 14:37:45 UTC
Add NNUE evaluation This patch ports the efficiently updatable neural network (NNUE) evaluation to Stockfish. Both the NNUE and the classical evaluations are available, and can be used to assign a value to a position that is later used in alpha-beta (PVS) search to find the best move. The classical evaluation computes this value as a function of various chess concepts, handcrafted by experts, tested and tuned using fishtest. The NNUE evaluation computes this value with a neural network based on basic inputs. The network is optimized and trained on the evalutions of millions of positions at moderate search depth. The NNUE evaluation was first introduced in shogi, and ported to Stockfish afterward. It can be evaluated efficiently on CPUs, and exploits the fact that only parts of the neural network need to be updated after a typical chess move. [The nodchip repository](https://github.com/nodchip/Stockfish) provides additional tools to train and develop the NNUE networks. This patch is the result of contributions of various authors, from various communities, including: nodchip, ynasu87, yaneurao (initial port and NNUE authors), domschl, FireFather, rqs, xXH4CKST3RXx, tttak, zz4032, joergoster, mstembera, nguyenpham, erbsenzaehler, dorzechowski, and vondele. This new evaluation needed various changes to fishtest and the corresponding infrastructure, for which tomtor, ppigazzini, noobpwnftw, daylen, and vondele are gratefully acknowledged. The first networks have been provided by gekkehenker and sergiovieri, with the latter net (nn-97f742aaefcd.nnue) being the current default. The evaluation function can be selected at run time with the `Use NNUE` (true/false) UCI option, provided the `EvalFile` option points the the network file (depending on the GUI, with full path). The performance of the NNUE evaluation relative to the classical evaluation depends somewhat on the hardware, and is expected to improve quickly, but is currently on > 80 Elo on fishtest: 60000 @ 10+0.1 th 1 https://tests.stockfishchess.org/tests/view/5f28fe6ea5abc164f05e4c4c ELO: 92.77 +-2.1 (95%) LOS: 100.0% Total: 60000 W: 24193 L: 8543 D: 27264 Ptnml(0-2): 609, 3850, 9708, 10948, 4885 40000 @ 20+0.2 th 8 https://tests.stockfishchess.org/tests/view/5f290229a5abc164f05e4c58 ELO: 89.47 +-2.0 (95%) LOS: 100.0% Total: 40000 W: 12756 L: 2677 D: 24567 Ptnml(0-2): 74, 1583, 8550, 7776, 2017 At the same time, the impact on the classical evaluation remains minimal, causing no significant regression: sprt @ 10+0.1 th 1 https://tests.stockfishchess.org/tests/view/5f2906a2a5abc164f05e4c5b LLR: 2.94 (-2.94,2.94) {-6.00,-4.00} Total: 34936 W: 6502 L: 6825 D: 21609 Ptnml(0-2): 571, 4082, 8434, 3861, 520 sprt @ 60+0.6 th 1 https://tests.stockfishchess.org/tests/view/5f2906cfa5abc164f05e4c5d LLR: 2.93 (-2.94,2.94) {-6.00,-4.00} Total: 10088 W: 1232 L: 1265 D: 7591 Ptnml(0-2): 49, 914, 3170, 843, 68 The needed networks can be found at https://tests.stockfishchess.org/nns It is recommended to use the default one as indicated by the `EvalFile` UCI option. Guidelines for testing new nets can be found at https://github.com/glinscott/fishtest/wiki/Creating-my-first-test#nnue-net-tests Integration has been discussed in various issues: https://github.com/official-stockfish/Stockfish/issues/2823 https://github.com/official-stockfish/Stockfish/issues/2728 The integration branch will be closed after the merge: https://github.com/official-stockfish/Stockfish/pull/2825 https://github.com/official-stockfish/Stockfish/tree/nnue-player-wip closes https://github.com/official-stockfish/Stockfish/pull/2912 This will be an exciting time for computer chess, looking forward to seeing the evolution of this approach. Bench: 4746616
Copyright 2011–2026 Next Chess Move LLC