Dev Builds » 20240121-1145

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 14:41:30 1182110 5010 2193 359 2458 +133.37 +/- 4.31 2 54 629 1748 72 +304.2 +/- 13.59
ncm-dbt-03 14:44:32 1199168 4992 2183 386 2423 +130.93 +/- 4.31 0 56 657 1713 70 +295.88 +/- 13.3
ncm-dbt-04 14:44:22 1210517 4998 2202 377 2419 +133.0 +/- 4.24 0 37 680 1702 80 +300.16 +/- 13.05
ncm-dbt-06 14:43:33 1194126 5000 2191 374 2435 +132.3 +/- 4.29 1 50 653 1723 73 +299.97 +/- 13.34
20000 8769 1496 9735 +132.4 +/- 2.14 3 197 2619 6886 295 +300.04 +/- 6.65

Test Detail

ID Host Base NPS Games WLD Standard Elo Ptnml(0-2) Gamepair Elo CLI PGN
326173 ncm-dbt-01 1186783 10 5 0 5 +190.62 +/- 11.11 0 0 0 5 0 +1199.83 +/- 231.31
326172 ncm-dbt-03 1196773 492 213 42 237 +126.0 +/- 13.38 0 4 73 163 6 +282.14 +/- 40.14
326171 ncm-dbt-06 1216562 500 210 29 261 +131.74 +/- 13.3 0 4 68 171 7 +298.62 +/- 41.71
326170 ncm-dbt-04 1243801 498 220 40 238 +131.52 +/- 13.53 0 4 69 168 8 +295.12 +/- 41.39
326169 ncm-dbt-01 1177156 500 214 41 245 +125.38 +/- 13.79 0 8 66 171 5 +282.94 +/- 42.37
326168 ncm-dbt-03 1202268 500 227 40 233 +136.56 +/- 14.27 0 6 62 171 11 +304.07 +/- 43.81
326167 ncm-dbt-04 1214927 500 217 35 248 +132.54 +/- 12.91 0 3 68 173 6 +304.07 +/- 41.65
326166 ncm-dbt-06 1188529 500 220 38 242 +132.54 +/- 13.1 0 5 63 177 5 +306.84 +/- 43.45
326165 ncm-dbt-01 1179842 500 238 30 232 +153.86 +/- 12.22 0 2 46 194 8 +381.7 +/- 51.25
326164 ncm-dbt-03 1202947 500 224 34 242 +138.99 +/- 13.48 0 5 58 179 8 +321.19 +/- 45.36
326163 ncm-dbt-06 1187484 500 218 33 249 +134.95 +/- 13.22 0 4 64 175 7 +309.64 +/- 43.08
326162 ncm-dbt-01 1180291 500 214 40 246 +126.17 +/- 13.95 0 6 72 164 8 +277.93 +/- 40.53
326161 ncm-dbt-04 1224115 500 212 41 247 +123.81 +/- 14.32 0 5 80 154 11 +263.42 +/- 38.27
326160 ncm-dbt-03 1194969 500 213 51 236 +116.77 +/- 13.71 0 8 76 162 4 +258.75 +/- 39.42
326159 ncm-dbt-04 1201782 500 225 30 245 +143.07 +/- 13.34 0 2 62 175 11 +327.18 +/- 43.69
326158 ncm-dbt-01 1183650 500 216 25 259 +139.81 +/- 13.45 0 5 57 180 8 +324.17 +/- 45.77
326157 ncm-dbt-06 1197051 500 224 32 244 +140.62 +/- 12.44 0 0 66 176 8 +327.17 +/- 41.95
326156 ncm-dbt-03 1202856 500 223 34 243 +138.18 +/- 13.12 0 3 63 176 8 +318.25 +/- 43.39
326155 ncm-dbt-01 1178951 500 213 39 248 +126.18 +/- 14.12 1 6 67 170 6 +285.49 +/- 42.08
326154 ncm-dbt-04 1211194 500 228 34 238 +142.25 +/- 13.75 0 2 65 170 13 +318.25 +/- 42.59
326153 ncm-dbt-06 1192136 500 212 42 246 +123.02 +/- 14.16 0 9 68 167 6 +273.0 +/- 41.7
326152 ncm-dbt-03 1195254 500 219 29 252 +138.99 +/- 13.85 0 7 54 181 8 +321.19 +/- 46.91
326151 ncm-dbt-06 1189915 500 224 44 232 +130.94 +/- 13.87 0 6 66 170 8 +293.29 +/- 42.41
326150 ncm-dbt-04 1234023 500 216 41 243 +126.97 +/- 13.58 0 6 69 169 6 +285.49 +/- 41.44
326149 ncm-dbt-01 1169435 500 222 27 251 +143.07 +/- 11.91 0 2 55 189 4 +349.43 +/- 46.59
326148 ncm-dbt-06 1192826 500 214 42 244 +124.6 +/- 13.8 0 8 67 170 5 +280.42 +/- 42.05
326147 ncm-dbt-03 1194861 500 221 31 248 +138.99 +/- 13.1 0 4 59 180 7 +324.17 +/- 44.97
326146 ncm-dbt-04 1193232 500 212 42 246 +123.02 +/- 13.47 0 6 73 166 5 +275.45 +/- 40.24
326145 ncm-dbt-01 1181668 500 222 40 238 +132.54 +/- 14.7 0 6 69 162 13 +285.49 +/- 41.44
326144 ncm-dbt-03 1201933 500 212 45 243 +120.67 +/- 13.5 0 7 73 166 4 +270.57 +/- 40.25
326143 ncm-dbt-04 1196088 500 225 42 233 +133.34 +/- 14.51 0 6 67 165 12 +290.66 +/- 42.08
326142 ncm-dbt-06 1193244 500 223 37 240 +135.76 +/- 13.76 0 5 63 173 9 +306.84 +/- 43.45
326141 ncm-dbt-01 1184200 500 215 44 241 +123.81 +/- 15.13 1 9 67 164 9 +270.57 +/- 41.96
326140 ncm-dbt-03 1199133 500 216 30 254 +135.76 +/- 13.76 0 4 66 170 10 +304.07 +/- 42.38
326139 ncm-dbt-01 1191694 500 205 36 259 +122.24 +/- 13.3 0 7 70 170 3 +277.93 +/- 41.14
326138 ncm-dbt-06 1196163 500 223 44 233 +130.15 +/- 14.4 1 5 67 168 9 +290.66 +/- 42.08
326137 ncm-dbt-04 1189675 500 219 35 246 +134.15 +/- 11.86 0 1 67 179 3 +318.25 +/- 41.77
326136 ncm-dbt-01 1189547 500 229 37 234 +140.62 +/- 13.04 0 3 60 179 8 +327.18 +/- 44.54
326135 ncm-dbt-06 1187352 500 223 33 244 +138.99 +/- 13.48 0 4 61 176 9 +318.25 +/- 44.18
326134 ncm-dbt-04 1196340 500 228 37 235 +139.81 +/- 12.26 0 2 60 183 5 +333.32 +/- 44.47
326133 ncm-dbt-03 1200691 500 215 50 235 +119.11 +/- 13.69 0 8 73 165 4 +265.78 +/- 40.25

Commit

Commit ID a901474bf9579ba259179eb09618d8401a156f64
Author Robert Nurnberg @ elitebook
Date 2024-01-21 11:45:03 UTC
Update the WDL model Update the internal WDL model. After the dual net merge, the internal evaluations have drifted upwards a bit. With this PR `NormalizeToPawnValue` changes from `328` to `345`. The new model was fitted based on about 200M positions extracted from 3.4M fishtest LTC games from the last two weeks, involving SF versions from 6deb88728fb141e853243c2873ad0cda4dd19320 to current master. Apart from the WDL model parameter update, this PR implements the following changes: WDL Model: - an incorrect 8-move shift in master's WDL model has been fixed - the polynomials `p_a` and `p_b` are fitted over the move range [8, 120] - the coefficients for `p_a` and `p_b` are optimized by maximizing the probability of predicting the observed outcome (credits to @vondele) SF code: - for wdl values, move will be clamped to `max(8, min(120, move))` - no longer clamp the internal eval to [-4000,4000] - compute `NormalizeToPawnValue` with `round`, not `trunc` The PR only affects displayed `cp` and `wdl` values. closes https://github.com/official-stockfish/Stockfish/pull/5002 No functional change
Copyright 2011–2024 Next Chess Move LLC