Dev Builds » 20240121-1145

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 06:54:57 584056 4000 1442 625 1933 +71.98 ± 4.87 1 123 942 926 8 +149.26 ± 11.01
ncm-dbt-02 06:55:39 586542 3996 1459 656 1881 +70.78 ± 4.95 1 138 924 927 8 +146.53 ± 11.15
ncm-dbt-03 06:56:05 586009 4012 1519 614 1879 +79.74 ± 5.03 0 136 840 1019 11 +166.52 ± 11.74
ncm-dbt-04 06:56:31 571179 3994 1432 640 1922 +69.82 ± 5.09 1 155 906 921 14 +143.11 ± 11.29
ncm-dbt-05 06:56:36 583154 3998 1438 647 1913 +69.66 ± 5.0 2 142 928 917 10 +143.77 ± 11.13
20000 7290 3182 9528 +72.39 ± 2.23 5 694 4540 4710 51 +149.76 ± 5.03

Test Detail

ID Host Base NPS Games WLD Standard Elo Ptnml(0-2) Gamepair Elo CLI PGN
391784 ncm-dbt-03 585000 12 4 4 4 -0.0 ± 81.87 0 1 4 1 0 -0.0 ± 173.65
391783 ncm-dbt-04 574459 494 182 74 238 +77.2 ± 13.49 0 11 119 115 2 +159.42 ± 30.82
391782 ncm-dbt-05 587494 498 183 84 231 +70.0 ± 14.5 1 17 115 114 2 +144.54 ± 31.74
391781 ncm-dbt-02 584916 496 177 79 240 +69.56 ± 13.35 0 14 122 112 0 +145.19 ± 30.51
391780 ncm-dbt-01 583572 500 194 97 209 +68.27 ± 14.12 0 18 119 111 2 +138.99 ± 31.13
391779 ncm-dbt-03 586181 500 186 74 240 +79.17 ± 14.9 0 20 101 126 3 +162.35 ± 34.05
391778 ncm-dbt-02 584285 500 189 86 225 +72.61 ± 13.54 0 15 117 118 0 +152.18 ± 31.35
391777 ncm-dbt-05 582527 500 174 82 244 +64.66 ± 15.07 1 23 111 113 2 +132.54 ± 32.44
391776 ncm-dbt-04 571070 500 177 82 241 +66.82 ± 14.62 0 23 110 116 1 +137.37 ± 32.59
391775 ncm-dbt-01 583321 500 180 77 243 +72.61 ± 13.39 0 14 119 117 0 +152.18 ± 31.0
391774 ncm-dbt-03 585000 500 186 79 235 +75.52 ± 14.65 0 20 105 123 2 +155.54 ± 33.38
391773 ncm-dbt-05 585590 500 185 77 238 +76.25 ± 13.51 0 14 114 122 0 +160.64 ± 31.79
391772 ncm-dbt-02 584453 500 180 89 231 +63.94 ± 15.04 0 25 112 110 3 +127.76 ± 32.29
391771 ncm-dbt-04 570148 500 188 80 232 +76.25 ± 14.25 0 16 113 118 3 +155.54 ± 32.01
391770 ncm-dbt-01 582987 500 186 72 242 +80.63 ± 13.18 0 11 114 125 0 +171.02 ± 31.66
391769 ncm-dbt-03 586139 500 195 73 232 +86.52 ± 14.38 0 17 95 137 1 +183.51 ± 35.13
391768 ncm-dbt-05 583363 500 175 91 234 +58.93 ± 13.58 0 19 128 103 0 +121.46 ± 29.83
391767 ncm-dbt-02 585717 500 188 72 240 +82.1 ± 13.98 0 14 108 126 2 +171.02 ± 32.78
391766 ncm-dbt-01 583279 500 171 77 252 +66.1 ± 14.03 0 19 119 111 1 +135.76 ± 31.16
391765 ncm-dbt-04 572235 500 174 80 246 +66.1 ± 14.73 0 23 112 113 2 +134.15 ± 32.28
391764 ncm-dbt-03 585970 500 191 82 227 +76.97 ± 13.53 0 11 122 114 3 +157.24 ± 30.38
391763 ncm-dbt-02 585717 500 181 82 237 +69.71 ± 14.31 0 19 115 114 2 +142.26 ± 31.77
391762 ncm-dbt-05 583572 500 187 81 232 +74.79 ± 14.34 0 18 110 120 2 +153.86 ± 32.54
391761 ncm-dbt-04 572034 500 185 78 237 +75.52 ± 13.64 0 13 119 116 2 +155.54 ± 30.96
391760 ncm-dbt-01 585801 500 169 81 250 +61.79 ± 14.55 0 24 115 110 1 +126.18 ± 31.84
391759 ncm-dbt-03 584537 500 192 72 236 +85.04 ± 14.35 0 17 97 135 1 +179.9 ± 34.76
391758 ncm-dbt-02 589027 500 183 82 235 +71.16 ± 13.04 0 11 128 110 1 +147.19 ± 29.47
391757 ncm-dbt-01 583237 500 181 73 246 +76.25 ± 13.05 0 9 126 113 2 +157.24 ± 29.64
391756 ncm-dbt-03 585843 500 201 66 233 +96.19 ± 13.95 0 14 87 149 0 +209.91 ± 36.75
391755 ncm-dbt-05 578588 500 190 65 245 +88.74 ± 13.68 0 10 108 129 3 +185.33 ± 32.63
391754 ncm-dbt-04 568673 500 169 90 241 +55.36 ± 14.79 0 27 119 102 2 +110.6 ± 31.28
391753 ncm-dbt-02 588728 500 180 79 241 +71.16 ± 13.64 0 16 117 117 0 +148.85 ± 31.38
391752 ncm-dbt-01 585548 500 184 71 245 +79.9 ± 14.21 0 17 104 128 1 +167.53 ± 33.52
391751 ncm-dbt-05 580613 500 169 79 252 +63.23 ± 14.2 0 21 119 109 1 +129.35 ± 31.2
391750 ncm-dbt-03 586985 500 192 76 232 +82.1 ± 14.13 0 16 103 130 1 +172.78 ± 33.69
391749 ncm-dbt-04 570549 500 180 79 241 +71.16 ± 15.58 1 26 95 127 1 +148.85 ± 34.96
391748 ncm-dbt-02 589497 500 181 87 232 +66.11 ± 15.0 1 24 105 120 0 +138.99 ± 33.37
391747 ncm-dbt-01 584706 500 177 77 246 +70.44 ± 13.47 1 11 126 111 1 +147.19 ± 29.83
391746 ncm-dbt-04 570268 500 177 77 246 +70.43 ± 13.76 0 16 119 114 1 +145.54 ± 31.07
391745 ncm-dbt-05 583489 500 175 88 237 +61.07 ± 13.82 0 20 123 107 0 +126.18 ± 30.59
391744 ncm-dbt-03 588430 500 172 88 240 +58.93 ± 13.72 0 20 126 104 0 +121.46 ± 30.15

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 2011–2025 Next Chess Move LLC