Dev Builds » 20210619-2124

You are viewing an old NCM Stockfish dev build test. You may find the most recent dev build tests using Stockfish 15 as the baseline here.

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 08:20:31 1968899 3363 2818 8 537 +419.11 ± 14.7
ncm-et-4 08:20:38 1964434 3344 2775 10 559 +409.32 ± 14.4
ncm-et-9 08:20:40 1961737 3331 2749 15 567 +402.74 ± 14.31
ncm-et-10 08:20:32 1957589 3311 2753 8 550 +411.75 ± 14.52
ncm-et-13 08:20:40 1959317 3340 2785 8 547 +414.41 ± 14.56
ncm-et-15 08:20:39 1956330 3311 2720 13 578 +399.37 ± 14.17
20000 16600 62 3338 +409.37 ± 5.89

Test Detail

ID Host Started (UTC) Duration Base NPS Games Wins Losses Draws Elo CLI PGN
149083 ncm-et-10 2021-06-20 06:46 00:46:37 1962304 311 252 0 59 +391.86 ± 44.81
149082 ncm-et-15 2021-06-20 06:45 00:47:23 1954987 311 257 1 53 +405.29 ± 47.58
149081 ncm-et-9 2021-06-20 06:43 00:49:27 1971416 331 274 3 54 +400.58 ± 47.19
149080 ncm-et-13 2021-06-20 06:42 00:50:24 1960016 340 286 1 53 +422.21 ± 47.62
149079 ncm-et-4 2021-06-20 06:41 00:51:36 1964301 344 273 0 71 +375.61 ± 40.67
149078 ncm-et-3 2021-06-20 06:39 00:53:20 1972339 363 308 3 52 +424.54 ± 48.14
149077 ncm-et-10 2021-06-20 05:31 01:14:40 1963226 500 415 2 83 +408.38 ± 37.78
149076 ncm-et-15 2021-06-20 05:29 01:15:47 1951190 500 422 2 76 +424.28 ± 39.56
149075 ncm-et-9 2021-06-20 05:28 01:14:10 1968010 500 412 2 86 +401.92 ± 37.09
149074 ncm-et-13 2021-06-20 05:27 01:14:38 1957399 500 430 0 70 +449.35 ± 41.19
149073 ncm-et-4 2021-06-20 05:24 01:15:56 1962322 500 425 0 75 +436.43 ± 39.73
149072 ncm-et-3 2021-06-20 05:24 01:15:01 1972804 500 428 0 72 +444.08 ± 40.59
149071 ncm-et-10 2021-06-20 04:15 01:15:26 1961841 500 417 1 82 +415.05 ± 37.98
149070 ncm-et-9 2021-06-20 04:14 01:13:10 1972339 500 406 2 92 +389.56 ± 35.8
149069 ncm-et-13 2021-06-20 04:11 01:14:58 1967567 500 421 1 78 +424.28 ± 38.99
149068 ncm-et-15 2021-06-20 04:11 01:17:12 1958017 500 414 2 84 +406.2 ± 37.55
149067 ncm-et-4 2021-06-20 04:10 01:13:50 1974360 500 419 1 80 +419.61 ± 38.47
149066 ncm-et-3 2021-06-20 04:10 01:13:26 1972137 500 410 1 89 +399.81 ± 36.38
149065 ncm-et-10 2021-06-20 02:58 01:15:28 1957555 500 414 2 84 +406.2 ± 37.55
149064 ncm-et-9 2021-06-20 02:58 01:15:44 1956211 500 421 2 77 +421.93 ± 39.29
149063 ncm-et-15 2021-06-20 02:55 01:14:53 1958623 500 409 4 87 +391.57 ± 36.91
149062 ncm-et-4 2021-06-20 02:55 01:14:17 1969245 500 410 3 87 +395.65 ± 36.9
149061 ncm-et-13 2021-06-20 02:55 01:15:52 1953892 500 415 0 85 +412.8 ± 37.2
149060 ncm-et-3 2021-06-20 02:54 01:15:00 1970211 500 411 2 87 +399.81 ± 36.87
149059 ncm-et-10 2021-06-20 01:42 01:16:06 1949648 500 424 1 75 +431.48 ± 39.8
149058 ncm-et-9 2021-06-20 01:41 01:16:23 1954546 500 416 1 83 +412.8 ± 37.74
149057 ncm-et-13 2021-06-20 01:40 01:14:33 1958472 500 410 3 87 +395.65 ± 36.9
149056 ncm-et-4 2021-06-20 01:40 01:14:47 1958000 500 414 1 85 +408.38 ± 37.27
149055 ncm-et-15 2021-06-20 01:39 01:15:21 1962153 500 403 1 96 +385.59 ± 34.96
149054 ncm-et-3 2021-06-20 01:39 01:14:04 1962172 500 424 0 76 +433.94 ± 39.45
149053 ncm-et-4 2021-06-20 00:24 01:14:27 1963728 500 416 3 81 +408.38 ± 38.29
149052 ncm-et-10 2021-06-20 00:24 01:16:46 1953437 500 410 2 88 +397.72 ± 36.65
149051 ncm-et-3 2021-06-20 00:24 01:14:42 1968322 500 411 1 88 +401.92 ± 36.6
149050 ncm-et-9 2021-06-20 00:24 01:16:51 1956943 500 417 1 82 +415.05 ± 37.98
149049 ncm-et-15 2021-06-20 00:23 01:15:28 1951935 500 412 2 86 +401.92 ± 37.09
149048 ncm-et-13 2021-06-20 00:23 01:15:51 1957719 500 414 2 84 +406.2 ± 37.55
149047 ncm-et-3 2021-06-19 23:08 01:14:58 1964312 500 426 1 73 +436.43 ± 40.36
149046 ncm-et-10 2021-06-19 23:08 01:15:29 1955114 500 421 0 79 +426.65 ± 38.66
149045 ncm-et-4 2021-06-19 23:08 01:15:45 1959084 500 418 2 80 +415.05 ± 38.52
149044 ncm-et-15 2021-06-19 23:08 01:14:35 1957405 500 403 1 96 +385.59 ± 34.96
149043 ncm-et-13 2021-06-19 23:08 01:14:24 1960156 500 409 1 90 +397.72 ± 36.17
149042 ncm-et-9 2021-06-19 23:08 01:14:55 1952699 500 403 4 93 +379.78 ± 35.66

Commit

Commit ID ba01f4b95448bcb324755f4dd2a632a57c6e67bc
Author MichaelB7
Date 2021-06-19 21:24:35 UTC
Make net nn-75980ca503c6.nnue the default. trained with the Python command c:\nnue>python train.py i:/bin/all.binpack i:/bin/all.binpack --gpus 1 --threads 4 --num-workers 30 --batch-size 16384 --progress_bar_refresh_rate 300 --smart-fen-skipping --random-fen-skipping 3 --features=HalfKAv2^ --lambda=1.0 --max_epochs=440 --seed %random%%random% --default_root_dir exp/run_10 --resume-from-model ./pt/nn-3b20abec10c1.pt ` all.binpack equaled 4 parts Wrong_NNUE_2.binpack https://drive.google.com/file/d/1seGNOqcVdvK_vPNq98j-zV3XPE5zWAeq/view?usp=sharing plus two parts of Training_Data.binpack https://drive.google.com/file/d/1RFkQES3DpsiJqsOtUshENtzPfFgUmEff/view?usp=sharing Each set was concatenated together - making one large Wrong_NNUE 2 binpack and one large Training so the were approximately equal in size. They were then interleaved together. The idea was to give Wrong_NNUE.binpack closer to equal weighting with the Training_Data binpack . Net nn-3b20abec10c1.nnue was chosen as the --resume-from-model with the idea that through learning, the manually hex edited values will be learned and will not need to be manually adjusted going forward. They would also be fine tuned by the learning process. passed STC: https://tests.stockfishchess.org/tests/view/60cdf91e457376eb8bcab66f LLR: 2.95 (-2.94,2.94) <-0.50,2.50> Total: 18256 W: 1639 L: 1479 D: 15138 Ptnml(0-2): 59, 1179, 6505, 1313, 72 passed LTC: https://tests.stockfishchess.org/tests/view/60ce2166457376eb8bcab6e1 LLR: 2.94 (-2.94,2.94) <0.50,3.50> Total: 18792 W: 654 L: 542 D: 17596 Ptnml(0-2): 9, 490, 8291, 592, 14 closes https://github.com/official-stockfish/Stockfish/pull/3570 Bench: 5020972
Copyright 2011–2024 Next Chess Move LLC