Dev Builds » 20220328-2043

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 06:53:54 576865 4002 973 1000 2029 -2.34 ± 5.19 2 473 1076 450 0 -4.34 ± 10.35
ncm-dbt-02 06:52:33 585844 4000 963 1027 2010 -5.56 ± 5.1 6 462 1124 406 2 -10.43 ± 10.08
ncm-dbt-03 06:54:31 584446 3992 981 994 2017 -1.13 ± 4.95 1 423 1161 410 1 -2.26 ± 9.86
ncm-dbt-04 06:54:03 568825 4008 934 1062 2012 -11.1 ± 4.95 3 482 1160 358 1 -21.87 ± 9.87
ncm-dbt-05 06:53:20 581982 3998 977 1012 2009 -3.04 ± 4.93 5 419 1183 390 2 -5.56 ± 9.73
20000 4828 5095 10077 -4.64 ± 2.25 17 2259 5704 2014 6 -8.9 ± 4.46

Test Detail

ID Host Base NPS Games WLD Standard Elo Ptnml(0-2) Gamepair Elo CLI PGN
417032 ncm-dbt-01 572356 2 1 1 0 0.0 ± 34.02 0 0 1 0 0 0.0 ± 30.47
417031 ncm-dbt-04 569151 8 3 3 2 0.0 ± 8.51 0 0 4 0 0 0.0 ± 7.61
417030 ncm-dbt-03 584411 492 120 115 257 +3.53 ± 14.19 0 50 141 55 0 +7.06 ± 28.43
417029 ncm-dbt-05 579248 498 120 117 261 +2.09 ± 14.54 0 55 136 58 0 +4.19 ± 29.14
417028 ncm-dbt-02 585632 500 117 125 258 -5.56 ± 14.02 0 56 147 46 1 -12.51 ± 27.7
417027 ncm-dbt-01 576128 500 122 137 241 -10.43 ± 14.7 0 66 133 51 0 -20.87 ± 29.53
417026 ncm-dbt-04 569270 500 116 137 247 -14.6 ± 14.0 0 64 143 43 0 -29.25 ± 28.2
417025 ncm-dbt-03 585801 500 121 120 259 +0.69 ± 13.96 0 52 145 53 0 +1.39 ± 27.97
417024 ncm-dbt-05 585000 500 124 118 258 +4.17 ± 13.75 0 48 148 54 0 +8.34 ± 27.56
417023 ncm-dbt-02 586477 500 117 128 255 -7.64 ± 14.84 1 61 137 50 1 -15.3 ± 29.02
417022 ncm-dbt-01 578218 500 119 110 271 +6.25 ± 13.95 0 48 145 57 0 +12.51 ± 27.96
417021 ncm-dbt-04 567720 500 112 128 260 -11.12 ± 13.57 0 58 150 42 0 -22.27 ± 27.27
417020 ncm-dbt-05 584285 500 133 141 226 -5.56 ± 14.54 0 60 139 50 1 -12.51 ± 28.76
417019 ncm-dbt-03 582736 500 124 128 248 -2.78 ± 13.49 0 51 152 47 0 -5.56 ± 27.02
417018 ncm-dbt-02 586774 500 118 134 248 -11.12 ± 14.24 0 63 140 47 0 -22.27 ± 28.62
417017 ncm-dbt-01 576088 500 121 126 253 -3.47 ± 14.99 1 60 132 57 0 -5.56 ± 29.66
417016 ncm-dbt-04 566849 500 122 133 245 -7.64 ± 13.94 1 55 148 46 0 -13.9 ± 27.56
417015 ncm-dbt-05 579702 500 121 124 255 -2.08 ± 13.96 0 54 145 51 0 -4.17 ± 27.97
417014 ncm-dbt-03 588558 500 128 113 259 +10.43 ± 14.82 0 52 131 67 0 +20.87 ± 29.78
417013 ncm-dbt-02 585253 500 112 135 253 -15.99 ± 14.25 2 61 145 42 0 -29.25 ± 27.93
417012 ncm-dbt-01 577028 500 127 127 246 -0.0 ± 14.8 0 59 132 59 0 -0.0 ± 29.66
417011 ncm-dbt-04 569629 500 120 137 243 -11.82 ± 15.31 1 69 126 54 0 -22.27 ± 30.41
417010 ncm-dbt-05 579951 500 108 120 272 -8.34 ± 13.59 1 53 153 43 0 -15.3 ± 26.87
417009 ncm-dbt-03 586435 500 128 127 245 +0.69 ± 13.42 0 48 153 49 0 +1.39 ± 26.88
417008 ncm-dbt-02 584537 500 124 121 255 +2.09 ± 14.23 1 50 144 55 0 +5.56 ± 28.1
417007 ncm-dbt-04 569629 500 118 141 241 -15.99 ± 14.25 1 64 142 43 0 -30.65 ± 28.34
417006 ncm-dbt-01 576782 500 114 114 272 -0.0 ± 13.63 0 50 150 50 0 -0.0 ± 27.29
417005 ncm-dbt-05 584076 500 122 122 256 -0.0 ± 14.8 0 59 132 59 0 -0.0 ± 29.66
417004 ncm-dbt-03 583531 500 126 138 236 -8.34 ± 13.59 1 53 153 43 0 -15.3 ± 26.87
417003 ncm-dbt-02 584622 500 123 116 261 +4.86 ± 14.73 0 55 133 62 0 +9.73 ± 29.53
417002 ncm-dbt-04 569310 500 109 140 251 -21.57 ± 14.16 0 71 139 40 0 -43.3 ± 28.7
417001 ncm-dbt-01 579289 500 128 120 252 +5.56 ± 16.46 0 69 104 77 0 +11.12 ± 33.02
417000 ncm-dbt-05 581402 500 114 137 249 -15.99 ± 13.16 2 53 161 34 0 -29.25 ± 25.67
416999 ncm-dbt-02 585970 500 124 127 249 -2.08 ± 15.11 1 60 130 59 0 -2.78 ± 29.91
416998 ncm-dbt-03 581652 500 108 122 270 -9.73 ± 14.51 0 63 139 47 1 -20.87 ± 28.75
416997 ncm-dbt-04 567126 500 117 128 255 -7.64 ± 14.33 0 61 139 50 0 -15.3 ± 28.76
416996 ncm-dbt-01 577561 500 114 121 265 -4.86 ± 14.6 0 61 135 54 0 -9.73 ± 29.28
416995 ncm-dbt-05 582193 500 135 133 232 +1.39 ± 12.93 2 37 169 41 1 +4.17 ± 24.55
416994 ncm-dbt-02 587494 500 128 141 231 -9.03 ± 13.93 1 56 148 45 0 -16.69 ± 27.55
416993 ncm-dbt-03 582444 500 126 131 243 -3.47 ± 13.82 0 54 147 49 0 -6.95 ± 27.7
416992 ncm-dbt-01 578342 500 127 144 229 -11.82 ± 14.17 1 60 144 45 0 -22.27 ± 28.08
416991 ncm-dbt-04 570749 500 117 115 268 +1.39 ± 12.49 0 40 169 40 1 +1.39 ± 24.56

Commit

Commit ID 471d93063a8fc1803a4a34397fe39e2344a05d76
Author Topologist
Date 2022-03-28 20:43:52 UTC
Play more positional in endgames This patch chooses the delta value (which skews the nnue evaluation between positional and materialistic) depending on the material: If the material is low, delta will be higher and the evaluation is shifted to the positional value. If the material is high, the evaluation will be shifted to the psqt value. I don't think slightly negative values of delta should be a concern. Passed STC: https://tests.stockfishchess.org/tests/view/62418513b3b383e86185766f LLR: 2.94 (-2.94,2.94) <0.00,2.50> Total: 28808 W: 7832 L: 7564 D: 13412 Ptnml(0-2): 147, 3186, 7505, 3384, 182 Passed LTC: https://tests.stockfishchess.org/tests/view/62419137b3b383e861857842 LLR: 2.96 (-2.94,2.94) <0.50,3.00> Total: 58632 W: 15776 L: 15450 D: 27406 Ptnml(0-2): 42, 5889, 17149, 6173, 63 closes https://github.com/official-stockfish/Stockfish/pull/3971 Bench: 7588855
Copyright 2011–2025 Next Chess Move LLC