Dev Builds » 20260409-1953

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 06:44:47 583231 4000 1771 367 1862 +127.36 ± 4.54 0 23 591 1345 41 +289.03 ± 13.95
ncm-dbt-02 06:45:27 584788 4016 1767 361 1888 +127.01 ± 4.56 1 35 558 1385 29 +292.14 ± 14.42
ncm-dbt-03 06:44:47 584385 4000 1730 361 1909 +123.91 ± 4.68 1 42 577 1347 33 +280.74 ± 14.18
ncm-dbt-04 06:43:37 566822 3986 1726 380 1880 +122.11 ± 4.49 1 26 616 1326 24 +277.82 ± 13.66
ncm-dbt-05 06:45:10 583004 3998 1781 374 1843 +127.73 ± 4.6 1 31 565 1364 38 +291.54 ± 14.31
20000 8775 1843 9382 +125.62 ± 2.05 4 157 2907 6767 165 +286.2 ± 6.3

Test Detail

ID Host Base NPS Games WLD Standard Elo Ptnml(0-2) Gamepair Elo CLI PGN
463611 ncm-dbt-02 586731 16 6 3 7 +65.88 ± 61.23 0 0 5 3 0 +136.9 ± 147.45
463610 ncm-dbt-04 565587 486 196 54 236 +104.56 ± 12.92 0 5 92 145 1 +230.29 ± 35.38
463609 ncm-dbt-03 584790 500 211 40 249 +123.81 ± 13.81 0 4 80 157 9 +268.17 ± 38.21
463608 ncm-dbt-01 583782 500 213 45 242 +121.45 ± 12.4 0 2 81 164 3 +275.45 ± 37.75
463607 ncm-dbt-05 584748 498 222 41 235 +132.32 ± 12.75 0 4 64 177 4 +308.83 ± 43.08
463606 ncm-dbt-02 585000 500 220 46 234 +126.17 ± 13.42 0 5 72 167 6 +282.94 ± 40.5
463605 ncm-dbt-04 569270 500 218 38 244 +130.94 ± 11.96 0 3 65 181 1 +312.48 ± 42.67
463604 ncm-dbt-05 585379 500 232 50 218 +132.54 ± 12.91 0 3 68 173 6 +304.07 ± 41.65
463603 ncm-dbt-03 585590 500 219 45 236 +126.17 ± 12.3 0 2 75 170 3 +290.66 ± 39.4
463602 ncm-dbt-01 580655 500 221 53 226 +121.45 ± 12.95 0 4 78 164 4 +273.0 ± 38.74
463601 ncm-dbt-02 581860 500 212 53 235 +114.45 ± 14.05 1 8 75 163 3 +256.44 ± 39.69
463600 ncm-dbt-04 567839 500 225 48 227 +128.55 ± 12.43 0 2 73 171 4 +295.94 ± 39.99
463599 ncm-dbt-05 579413 500 222 42 236 +130.94 ± 12.76 0 2 72 170 6 +298.62 ± 40.29
463598 ncm-dbt-03 583614 500 211 42 247 +122.24 ± 13.12 0 4 78 163 5 +273.0 ± 38.74
463597 ncm-dbt-01 581111 500 224 44 232 +130.94 ± 12.76 0 3 69 173 5 +301.33 ± 41.33
463596 ncm-dbt-02 583656 500 220 48 232 +124.6 ± 12.71 0 4 73 170 3 +285.49 ± 40.16
463595 ncm-dbt-05 583698 500 214 43 243 +123.81 ± 13.46 0 7 69 170 4 +280.42 ± 41.44
463594 ncm-dbt-04 568156 500 221 42 237 +130.14 ± 12.59 0 2 72 171 5 +298.62 ± 40.29
463593 ncm-dbt-03 583112 500 220 45 235 +126.97 ± 13.94 1 6 65 173 5 +290.66 ± 42.73
463592 ncm-dbt-01 584076 500 222 49 229 +125.38 ± 12.7 0 2 78 165 5 +282.94 ± 38.55
463591 ncm-dbt-02 585590 500 224 47 229 +128.55 ± 12.82 0 6 63 179 2 +301.33 ± 43.45
463590 ncm-dbt-03 583698 500 205 43 252 +116.77 ± 13.01 0 5 81 161 3 +261.07 ± 38.02
463589 ncm-dbt-04 565981 500 219 49 232 +123.02 ± 13.29 0 6 72 168 4 +277.93 ± 40.53
463588 ncm-dbt-01 584832 500 224 46 230 +129.35 ± 12.8 0 4 68 174 4 +298.62 ± 41.71
463587 ncm-dbt-05 581444 500 219 47 234 +124.6 ± 13.09 0 3 78 163 6 +277.93 ± 38.66
463586 ncm-dbt-02 585211 500 217 39 244 +129.35 ± 12.8 0 3 71 171 5 +295.94 ± 40.69
463585 ncm-dbt-03 587367 500 223 44 233 +130.14 ± 12.59 0 5 63 180 2 +306.84 ± 43.45
463584 ncm-dbt-04 566060 500 214 56 230 +113.68 ± 12.68 0 3 89 155 3 +251.89 ± 35.92
463583 ncm-dbt-05 581111 500 224 51 225 +125.39 ± 14.47 1 5 73 162 9 +275.45 ± 40.24
463582 ncm-dbt-01 583196 500 225 49 226 +127.76 ± 13.57 0 3 77 161 9 +280.42 ± 38.93
463581 ncm-dbt-02 584874 500 219 46 235 +125.38 ± 12.51 0 5 68 176 1 +293.29 ± 41.75
463580 ncm-dbt-04 566967 500 218 49 233 +122.24 ± 12.76 1 2 76 169 2 +282.94 ± 39.21
463579 ncm-dbt-01 582736 500 223 38 239 +134.95 ± 13.03 0 1 72 168 9 +304.07 ± 40.16
463578 ncm-dbt-03 581361 500 216 56 228 +115.22 ± 13.88 0 11 70 167 2 +258.75 ± 41.02
463577 ncm-dbt-05 582987 500 223 43 234 +130.94 ± 11.76 0 2 67 180 1 +312.48 ± 41.9
463576 ncm-dbt-02 583698 500 224 30 246 +142.25 ± 11.28 0 1 56 191 2 +352.8 ± 46.04
463575 ncm-dbt-04 564723 500 215 44 241 +123.81 ± 12.73 0 3 77 166 4 +280.42 ± 38.93
463574 ncm-dbt-05 585253 500 225 57 218 +121.45 ± 12.77 0 5 74 169 2 +277.93 ± 39.92
463573 ncm-dbt-01 585464 500 219 43 238 +127.76 ± 12.45 0 4 68 176 2 +298.62 ± 41.71
463572 ncm-dbt-02 586477 500 225 49 226 +127.76 ± 13.21 0 3 75 165 7 +285.49 ± 39.5
463571 ncm-dbt-03 585548 500 225 46 229 +130.14 ± 12.97 0 5 65 176 4 +301.33 ± 42.75

Commit

Commit ID ead7e650da1dd07a2614ba4d8207470fe921a87b
Author anematode
Date 2026-04-09 19:53:08 UTC
Fix weird indexing bug [Passed STC](https://tests.stockfishchess.org/tests/live_elo/69cf6a5d1668971c9da23ae1) LLR: 2.93 (-2.94,2.94) <-1.75,0.25> Total: 51104 W: 13376 L: 13180 D: 24548 Ptnml(0-2): 134, 5555, 13963, 5781, 119 [Passed LTC](https://tests.stockfishchess.org/tests/live_elo/69d00a7ee2b443cb2670b5c6) LLR: 2.95 (-2.94,2.94) <-1.75,0.25> Total: 101256 W: 25867 L: 25732 D: 49657 Ptnml(0-2): 59, 10423, 29520, 10576, 50 When processing a promotion we add/remove nonsensical threats, namely, piece threats of pawns on the 1st or 8th rank. This is not only inefficient but also seems to lead to an obscure bug during a promotion–capture + `double_inc_update` that causes a lingering invalid feature. Currently – as an artifact of the training process – these parts of the net are filled with random values in [-1,1], which is why there is a bench change. @ces42 did a helpful analysis: > checking bench with depth 17 actually shows that the bug in has been there all the time since 8e5392d7 (when threat inputs was merged into master) > i.e. double_inc_update for ThreatFeatureSet has always been slightly bugged After this patch is applied we'll be able to better compress the net by filling the invalid features with 0, w/o changing bench. I expect a 15%-ish size savings. closes https://github.com/official-stockfish/Stockfish/pull/6701 Bench: 2700393
Copyright 2011–2026 Next Chess Move LLC