Dev Builds » 20230223-1227

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 09:37:57 1948969 3330 2869 1 460 +451.04 +/- 15.87
ncm-et-4 09:38:16 1954029 3351 2885 3 463 +449.41 +/- 15.83
ncm-et-9 09:38:07 1939070 3304 2847 5 452 +449.58 +/- 16.03
ncm-et-10 09:38:05 1947220 3322 2854 5 463 +446.2 +/- 15.84
ncm-et-13 09:38:14 1960369 3354 2889 1 464 +450.78 +/- 15.8
ncm-et-15 09:37:45 1953889 3339 2874 2 463 +449.54 +/- 15.82
20000 17218 17 2765 +449.42 +/- 6.46

Test Detail

ID Host Started (UTC) Duration Base NPS Games Wins Losses Draws Elo CLI PGN
176438 ncm-et-9 2023-03-10 11:09 00:53:25 1926294 304 257 0 47 +430.74 +/- 50.59
176436 ncm-et-3 2023-03-10 11:06 00:56:07 1948998 330 288 0 42 +467.09 +/- 53.79
176435 ncm-et-10 2023-03-10 11:05 00:57:01 1949641 322 288 1 33 +496.22 +/- 61.35
176434 ncm-et-15 2023-03-10 11:03 00:58:20 1953299 339 291 0 48 +447.23 +/- 50.09
176433 ncm-et-4 2023-03-10 11:02 01:00:02 1958839 351 297 0 54 +431.67 +/- 47.06
176432 ncm-et-13 2023-03-10 11:01 01:01:18 1958027 354 303 0 51 +443.99 +/- 48.52
176415 ncm-et-9 2023-03-10 09:40 01:27:32 1947373 500 443 0 57 +487.45 +/- 45.89
176413 ncm-et-3 2023-03-10 09:38 01:26:38 1955727 500 420 1 79 +421.93 +/- 38.73
176412 ncm-et-10 2023-03-10 09:36 01:28:15 1931114 500 428 2 70 +438.95 +/- 41.29
176411 ncm-et-15 2023-03-10 09:36 01:26:55 1958011 500 437 0 63 +468.95 +/- 43.54
176410 ncm-et-4 2023-03-10 09:35 01:26:41 1948344 500 428 0 72 +444.08 +/- 40.59
176409 ncm-et-13 2023-03-10 09:34 01:25:56 1959094 500 434 1 65 +457.52 +/- 42.89
176391 ncm-et-3 2023-03-10 08:12 01:25:36 1951071 500 442 0 58 +484.24 +/- 45.47
176390 ncm-et-9 2023-03-10 08:12 01:27:39 1942702 500 430 1 69 +446.7 +/- 41.57
176389 ncm-et-10 2023-03-10 08:09 01:26:24 1959284 500 436 0 64 +466.03 +/- 43.18
176388 ncm-et-15 2023-03-10 08:08 01:26:48 1958317 500 431 2 67 +446.7 +/- 42.25
176387 ncm-et-13 2023-03-10 08:08 01:25:40 1976382 500 438 0 62 +471.92 +/- 43.9
176386 ncm-et-4 2023-03-10 08:08 01:26:26 1958335 500 429 0 71 +446.7 +/- 40.89
176369 ncm-et-3 2023-03-10 06:46 01:25:36 1963383 500 428 0 72 +444.08 +/- 40.59
176368 ncm-et-9 2023-03-10 06:44 01:27:15 1945259 500 439 1 60 +471.92 +/- 44.73
176366 ncm-et-15 2023-03-10 06:42 01:25:35 1964349 500 416 0 84 +415.04 +/- 37.43
176365 ncm-et-10 2023-03-10 06:42 01:26:27 1952070 500 431 1 68 +449.35 +/- 41.89
176364 ncm-et-4 2023-03-10 06:41 01:26:03 1958286 500 422 1 77 +426.65 +/- 39.25
176363 ncm-et-13 2023-03-10 06:40 01:27:13 1954974 500 426 0 74 +438.95 +/- 40.01
176346 ncm-et-3 2023-03-10 05:17 01:27:58 1951914 500 414 0 86 +410.58 +/- 36.97
176344 ncm-et-9 2023-03-10 05:16 01:27:24 1937713 500 430 1 69 +446.7 +/- 41.57
176343 ncm-et-15 2023-03-10 05:16 01:25:44 1957399 500 429 0 71 +446.7 +/- 40.89
176342 ncm-et-4 2023-03-10 05:16 01:24:50 1963703 500 438 0 62 +471.92 +/- 43.9
176341 ncm-et-10 2023-03-10 05:14 01:27:04 1940455 500 409 1 90 +397.72 +/- 36.17
176340 ncm-et-13 2023-03-10 05:14 01:25:44 1960018 500 421 0 79 +426.65 +/- 38.66
176323 ncm-et-3 2023-03-10 03:49 01:27:47 1924279 500 445 0 55 +494.02 +/- 46.77
176322 ncm-et-9 2023-03-10 03:48 01:27:04 1922997 500 427 1 72 +438.95 +/- 40.66
176321 ncm-et-15 2023-03-10 03:48 01:27:03 1945401 500 430 0 70 +449.35 +/- 41.19
176320 ncm-et-10 2023-03-10 03:48 01:25:53 1951778 500 419 0 81 +421.93 +/- 38.15
176319 ncm-et-13 2023-03-10 03:47 01:25:43 1956201 500 436 0 64 +466.03 +/- 43.18
176318 ncm-et-4 2023-03-10 03:47 01:28:02 1932435 500 442 1 57 +481.1 +/- 45.96
176300 ncm-et-4 2023-03-10 02:20 01:26:12 1958263 500 429 1 70 +444.09 +/- 41.26
176299 ncm-et-10 2023-03-10 02:20 01:27:01 1946203 500 443 0 57 +487.45 +/- 45.89
176298 ncm-et-13 2023-03-10 02:20 01:26:40 1957891 500 431 0 69 +452.04 +/- 41.5
176297 ncm-et-3 2023-03-10 02:20 01:28:15 1947413 500 432 0 68 +454.76 +/- 41.82
176296 ncm-et-15 2023-03-10 02:20 01:27:20 1940448 500 440 0 60 +477.98 +/- 44.67
176295 ncm-et-9 2023-03-10 02:20 01:27:48 1951153 500 421 1 78 +424.28 +/- 38.99

Commit

Commit ID 69639d764bde566e524b8c2566119bf677cb2622
Author Linmiao Xu
Date 2023-02-23 12:27:57 UTC
Reintroduce nnue pawn scaling with lower lazy thresholds Params found with the nevergrad TBPSA optimizer via nevergrad4sf modified to: * use SPRT LLR with fishtest STC elo gainer bounds [0, 2] as the objective function * increase the game batch size after each new optimal point is found The params were the optimal point after TBPSA iteration 7 and 160 nevergrad evaluations with: * initial batch size of 96 games per evaluation * batch size increase of 64 games after each iteration * a budget of 512 evaluations * TC: fixed 1.5 million nodes per move, no time limit nevergrad4sf enables optimizing stockfish params with TBPSA: https://github.com/vondele/nevergrad4sf Using pentanomial game results with smaller game batch sizes was inspired by: Use of SPRT LLR calculated from pentanomial game results as the objective function was an experiment at maximizing the information from game batches to reduce the computational cost for TBPSA to converge on good parameters. For the exact code used to find the params: https://github.com/linrock/tuning-fork Passed STC: https://tests.stockfishchess.org/tests/view/63f4ef5ee74a12625bcd114a LLR: 2.94 (-2.94,2.94) <0.00,2.00> Total: 66552 W: 17736 L: 17390 D: 31426 Ptnml(0-2): 164, 7229, 18166, 7531, 186 Passed LTC: https://tests.stockfishchess.org/tests/view/63f56028e74a12625bcd2550 LLR: 2.94 (-2.94,2.94) <0.50,2.50> Total: 71264 W: 19150 L: 18787 D: 33327 Ptnml(0-2): 23, 6728, 21771, 7083, 27 closes https://github.com/official-stockfish/Stockfish/pull/4401 bench 3687580
Copyright 2011–2024 Next Chess Move LLC