Dev Builds » 20230806-2016

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 14. 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 09:56:28 1195206 3330 1418 289 1623 +122.65 +/- 5.31 1 43 483 1102 36 +273.65 +/- 15.51
ncm-dbt-02 09:55:59 1215429 3314 1404 272 1638 +123.65 +/- 5.25 0 30 510 1072 45 +273.0 +/- 15.05
ncm-dbt-03 09:56:35 1241089 3338 1423 343 1572 +116.6 +/- 5.42 0 47 540 1037 45 +251.97 +/- 14.65
ncm-dbt-04 09:56:30 1224466 3340 1438 315 1587 +121.54 +/- 5.31 0 42 504 1083 41 +268.11 +/- 15.18
ncm-dbt-05 09:56:05 1229400 3334 1364 296 1674 +115.36 +/- 5.32 2 55 503 1087 20 +257.52 +/- 15.2
ncm-dbt-06 09:53:45 1234963 3344 1421 314 1609 +119.52 +/- 5.42 2 46 509 1073 42 +262.33 +/- 15.11
20000 8468 1829 9703 +119.87 +/- 2.18 5 263 3049 6454 229 +264.3 +/- 6.16

Test Detail

ID Host Base NPS Games WLD Standard Elo Ptnml(0-2) Gamepair Elo CLI PGN
199023 ncm-dbt-02 1236038 314 134 26 154 +124.57 +/- 15.9 0 2 47 106 2 +284.95 +/- 50.23
199022 ncm-dbt-01 1198549 330 145 28 157 +128.77 +/- 15.32 0 2 46 115 2 +299.28 +/- 50.9
199021 ncm-dbt-05 1222658 334 138 34 162 +111.9 +/- 17.4 0 9 46 111 1 +250.07 +/- 50.65
199020 ncm-dbt-03 1237773 338 142 37 159 +111.62 +/- 16.06 0 2 64 99 4 +239.54 +/- 42.31
199019 ncm-dbt-04 1247969 340 143 33 164 +116.59 +/- 16.88 0 5 54 107 4 +253.89 +/- 46.91
199018 ncm-dbt-06 1237450 344 139 34 171 +109.54 +/- 17.33 0 5 63 98 6 +227.86 +/- 43.14
199017 ncm-dbt-02 1192572 500 211 51 238 +115.22 +/- 15.0 0 9 83 147 11 +238.66 +/- 37.66
199016 ncm-dbt-01 1176294 500 213 53 234 +115.22 +/- 14.37 0 9 79 155 7 +247.41 +/- 38.64
199015 ncm-dbt-03 1260978 500 214 46 240 +121.45 +/- 14.01 0 7 75 161 7 +265.78 +/- 39.69
199014 ncm-dbt-05 1237005 500 213 44 243 +122.24 +/- 13.48 0 5 77 162 6 +270.57 +/- 39.07
199013 ncm-dbt-06 1245284 500 211 56 233 +111.37 +/- 14.06 1 7 82 156 4 +245.2 +/- 37.88
199012 ncm-dbt-04 1206912 500 213 48 239 +119.11 +/- 13.69 0 7 76 162 5 +263.42 +/- 39.41
199011 ncm-dbt-06 1217822 500 219 44 237 +126.97 +/- 14.11 0 9 63 172 6 +285.49 +/- 43.31
199010 ncm-dbt-02 1231754 500 211 43 246 +121.45 +/- 12.58 0 4 76 168 2 +277.93 +/- 39.29
199009 ncm-dbt-03 1235863 500 217 46 237 +123.81 +/- 13.98 0 9 66 170 5 +277.93 +/- 42.33
199008 ncm-dbt-05 1231041 500 200 46 254 +110.6 +/- 13.56 0 10 77 162 1 +247.41 +/- 39.15
199007 ncm-dbt-01 1201273 500 208 32 260 +127.76 +/- 12.84 0 2 76 166 6 +288.06 +/- 39.11
199006 ncm-dbt-04 1235243 500 216 45 239 +123.81 +/- 13.63 0 5 76 162 7 +273.0 +/- 39.35
198963 ncm-dbt-02 1227620 500 223 41 236 +132.54 +/- 13.65 0 3 72 165 10 +293.29 +/- 40.39
198962 ncm-dbt-05 1244314 500 198 43 259 +111.37 +/- 13.22 0 7 83 158 2 +247.41 +/- 37.61
198961 ncm-dbt-01 1200689 500 216 54 230 +116.78 +/- 13.87 1 6 77 162 4 +261.07 +/- 39.14
198960 ncm-dbt-03 1233422 500 211 48 241 +117.55 +/- 13.87 0 6 82 155 7 +254.16 +/- 37.82
198959 ncm-dbt-04 1248570 500 229 50 221 +130.14 +/- 13.88 0 4 73 163 10 +285.49 +/- 40.16
198958 ncm-dbt-06 1237810 500 216 48 236 +121.45 +/- 12.95 0 4 78 164 4 +273.0 +/- 38.74
198957 ncm-dbt-02 1205797 500 217 37 246 +130.94 +/- 13.87 0 4 72 164 10 +288.06 +/- 40.46
198956 ncm-dbt-05 1243189 500 205 42 253 +117.55 +/- 13.36 0 7 76 164 3 +263.42 +/- 39.41
198955 ncm-dbt-01 1206713 500 208 43 249 +119.11 +/- 14.68 0 10 73 159 8 +256.44 +/- 40.21
198954 ncm-dbt-04 1213418 500 215 49 236 +119.89 +/- 12.98 0 5 77 165 3 +270.57 +/- 39.07
198953 ncm-dbt-06 1245552 500 209 48 243 +116.0 +/- 14.05 0 7 82 154 7 +249.64 +/- 37.86
198952 ncm-dbt-03 1229682 500 203 53 244 +107.54 +/- 14.05 0 10 84 152 4 +232.26 +/- 37.43
198951 ncm-dbt-05 1201898 500 210 48 242 +116.78 +/- 14.37 1 10 68 168 3 +263.42 +/- 41.6
198950 ncm-dbt-02 1197795 500 208 37 255 +123.81 +/- 12.73 0 3 77 166 4 +280.42 +/- 38.93
198949 ncm-dbt-01 1181594 500 212 36 252 +127.76 +/- 13.21 0 5 69 171 5 +290.66 +/- 41.43
198948 ncm-dbt-06 1226142 500 221 36 243 +134.95 +/- 14.13 0 6 63 171 10 +301.33 +/- 43.45
198947 ncm-dbt-04 1209325 500 213 42 245 +123.81 +/- 14.48 0 11 63 170 6 +275.45 +/- 43.16
198946 ncm-dbt-03 1257381 500 216 49 235 +120.67 +/- 14.18 0 8 74 161 7 +263.42 +/- 39.97
198945 ncm-dbt-05 1225696 500 200 39 261 +116.0 +/- 14.04 1 7 76 162 4 +258.75 +/- 39.42
198944 ncm-dbt-02 1216430 500 200 37 263 +117.55 +/- 13.53 0 5 83 156 6 +256.44 +/- 37.51
198943 ncm-dbt-06 1234687 500 206 48 246 +113.68 +/- 14.38 1 8 78 158 5 +249.64 +/- 38.9
198942 ncm-dbt-01 1201333 500 216 43 241 +125.38 +/- 13.79 0 9 63 174 4 +285.49 +/- 43.31
198941 ncm-dbt-03 1232526 500 220 64 216 +112.14 +/- 14.38 0 5 95 139 11 +230.16 +/- 34.79
198940 ncm-dbt-04 1209829 500 209 48 243 +116.0 +/- 13.54 0 5 85 154 6 +251.89 +/- 37.03

Commit

Commit ID 0ad9b51deaaa1f2a8273ed064fbf6425cfbbe4f2
Author Linmiao Xu
Date 2023-08-06 20:16:52 UTC
Remove classical psqt Based on vondele's deletepsqt branch: https://github.com/vondele/Stockfish/commit/369f5b051 This huge simplification uses a weighted material differences instead of the positional piece square tables (psqt) in the semi-classical complexity calculation. Tuned weights using spsa at 45+0.45 with: int pawnMult = 100; int knightMult = 325; int bishopMult = 350; int rookMult = 500; int queenMult = 900; TUNE(SetRange(0, 200), pawnMult); TUNE(SetRange(0, 650), knightMult); TUNE(SetRange(0, 700), bishopMult); TUNE(SetRange(200, 800), rookMult); TUNE(SetRange(600, 1200), queenMult); The values obtained via this tuning session were for a model where the psqt replacement formula was always from the point of view of White, even if the side to move was Black. We re-used the same values for an implementation with a psqt replacement from the point of view of the side to move, testing the result both on our standard book on positions with a strong White bias, and an alternate book with positions with a strong Black bias. We note that with the patch the last use of the venerable "Score" type disappears in Stockfish codebase (the Score type was used in classical evaluation to get a tampered eval interpolating values smoothly from the early midgame stage to the endgame stage). We leave it to another commit to clean all occurrences of Score in the code and the comments. ------- Passed non-regression LTC: LLR: 2.94 (-2.94,2.94) <-1.75,0.25> Total: 142542 W: 36264 L: 36168 D: 70110 Ptnml(0-2): 76, 15578, 39856, 15696, 65 https://tests.stockfishchess.org/tests/view/64c8cb495b17f7c21c0cf9f8 Passed non-regression LTC (with a book with Black bias): https://tests.stockfishchess.org/tests/view/64c8f9295b17f7c21c0cfdaf LLR: 2.94 (-2.94,2.94) <-1.75,0.25> Total: 494814 W: 125565 L: 125827 D: 243422 Ptnml(0-2): 244, 53926, 139346, 53630, 261 ------ closes https://github.com/official-stockfish/Stockfish/pull/4713 Bench: 1655985
Copyright 2011–2024 Next Chess Move LLC