Dev Builds » 20170504-0246

Use this dev build

NCM plays each Stockfish dev build 20,000 times against Stockfish 7. This yields an approximate Elo difference and establishes confidence in the strength of the dev builds.

Summary

Host Duration Avg Base NPS Games Wins Losses Draws Elo
ncm-et-3 06:04:12 1986362 2430 769 109 1552 +96.79 +/- 8.0
ncm-et-4 06:03:24 1968500 2396 782 66 1548 +107.09 +/- 7.86
ncm-et-5 01:53:19 2024019 782 260 27 495 +106.76 +/- 14.11
ncm-et-6 01:53:36 2024999 759 279 34 446 +116.31 +/- 15.33
ncm-et-7 01:53:15 2012466 777 258 36 483 +102.11 +/- 14.5
ncm-et-8 01:53:20 1998516 768 265 31 472 +109.33 +/- 14.66
ncm-et-9 06:11:45 1961012 2456 828 93 1535 +107.26 +/- 8.07
ncm-et-10 06:02:22 1986247 2429 832 112 1485 +106.17 +/- 8.31
ncm-et-11 01:53:24 2016047 768 250 31 487 +101.9 +/- 14.28
ncm-et-12 01:53:31 1989413 767 248 36 483 +98.6 +/- 14.44
ncm-et-13 06:02:21 1990981 2454 838 114 1502 +105.64 +/- 8.26
ncm-et-14 01:53:18 2020266 768 267 23 478 +114.34 +/- 14.39
ncm-et-15 06:04:00 1985883 2446 807 96 1543 +103.99 +/- 8.03
20000 6683 808 12509 +105.16 +/- 2.83

Test Detail

ID Host Started (UTC) Duration Base NPS Games Wins Losses Draws Elo CLI PGN
60187 ncm-et-4 2018-11-25 13:36 00:22:37 1975672 152 48 2 102 +108.54 +/- 29.71
60186 ncm-et-3 2018-11-25 13:34 00:24:03 1965541 163 49 9 105 +87.04 +/- 30.95
60185 ncm-et-10 2018-11-25 13:34 00:24:20 1958019 158 58 6 94 +118.77 +/- 33.23
60184 ncm-et-13 2018-11-25 13:33 00:25:49 1971578 174 50 8 116 +85.55 +/- 28.83
60183 ncm-et-15 2018-11-25 13:32 00:26:12 1960771 174 61 8 105 +109.3 +/- 31.47
60182 ncm-et-9 2018-11-25 13:31 00:27:17 1973316 179 55 3 121 +103.92 +/- 27.28
60181 ncm-et-4 2018-11-25 12:20 01:15:07 1955694 500 153 18 329 +96.19 +/- 17.06
60180 ncm-et-10 2018-11-25 12:19 01:14:31 1973072 500 182 18 300 +118.33 +/- 18.48
60179 ncm-et-3 2018-11-25 12:18 01:15:22 1965996 500 147 22 331 +88.74 +/- 17.07
60178 ncm-et-15 2018-11-25 12:17 01:13:46 1968500 500 170 15 315 +111.37 +/- 17.65
60177 ncm-et-9 2018-11-25 12:17 01:12:47 1974205 500 159 11 330 +106.01 +/- 16.78
60176 ncm-et-13 2018-11-25 12:17 01:14:26 1969246 500 167 23 310 +102.97 +/- 18.12
60175 ncm-et-4 2018-11-25 11:01 01:17:00 1953825 500 176 10 314 +119.89 +/- 17.53
60174 ncm-et-9 2018-11-25 11:01 01:14:54 1972673 500 172 17 311 +111.37 +/- 17.91
60173 ncm-et-3 2018-11-25 11:01 01:15:41 1966461 500 161 22 317 +99.2 +/- 17.75
60172 ncm-et-15 2018-11-25 11:01 01:15:11 1971668 500 151 22 327 +91.71 +/- 17.26
60171 ncm-et-13 2018-11-25 11:01 01:15:02 1973910 500 175 24 301 +108.3 +/- 18.58
60170 ncm-et-10 2018-11-25 11:01 01:16:24 1960628 500 159 25 316 +95.44 +/- 17.88
59582 ncm-et-13 2018-11-19 20:03 01:13:30 1983767 500 175 30 295 +103.73 +/- 19.01
59581 ncm-et-4 2018-11-19 20:03 01:15:05 1977763 500 171 11 318 +115.23 +/- 17.37
59580 ncm-et-9 2018-11-19 20:03 01:23:35 1795721 500 195 32 273 +117.55 +/- 20.12
59579 ncm-et-15 2018-11-19 20:03 01:15:15 1972028 500 173 17 310 +112.14 +/- 17.96
59578 ncm-et-10 2018-11-19 20:03 01:13:42 1984416 500 177 25 298 +109.07 +/- 18.75
59577 ncm-et-3 2018-11-19 20:03 01:15:56 1967231 500 149 30 321 +84.31 +/- 17.74
9426 ncm-et-4 2018-03-18 00:28 00:36:33 2021815 244 74 12 158 +90.26 +/- 25.06
9425 ncm-et-6 2018-03-18 00:25 00:39:00 2024263 259 99 13 147 +119.91 +/- 27.07
9424 ncm-et-3 2018-03-18 00:25 00:39:03 2026554 267 88 8 171 +107.4 +/- 23.83
9423 ncm-et-10 2018-03-18 00:25 00:39:24 2020348 271 96 11 164 +112.77 +/- 25.04
9422 ncm-et-8 2018-03-18 00:25 00:39:48 2011696 268 81 13 174 +90.12 +/- 23.84
9421 ncm-et-14 2018-03-18 00:24 00:39:55 2020673 268 91 7 170 +112.69 +/- 23.9
9420 ncm-et-12 2018-03-18 00:24 00:40:19 2001857 267 86 9 172 +103.12 +/- 23.78
9419 ncm-et-13 2018-03-18 00:24 00:40:20 2023612 280 94 13 173 +103.46 +/- 24.31
9418 ncm-et-15 2018-03-18 00:24 00:40:30 2019372 272 87 9 176 +102.51 +/- 23.45
9417 ncm-et-11 2018-03-18 00:24 00:40:36 2016447 268 84 10 174 +98.49 +/- 23.64
9416 ncm-et-7 2018-03-18 00:24 00:40:34 2001161 277 94 12 171 +106.02 +/- 24.41
9415 ncm-et-5 2018-03-18 00:23 00:41:14 2022793 282 93 16 173 +97.34 +/- 24.52
9414 ncm-et-9 2018-03-18 00:23 00:41:30 2026063 277 93 13 171 +103.28 +/- 24.47
9413 ncm-et-7 2018-03-17 23:10 01:12:41 2023772 500 164 24 312 +99.95 +/- 18.05
9412 ncm-et-8 2018-03-17 23:10 01:13:32 1985336 500 184 18 298 +119.89 +/- 18.57
9411 ncm-et-3 2018-03-17 23:10 01:14:07 2026390 500 175 18 307 +112.91 +/- 18.13
9410 ncm-et-11 2018-03-17 23:09 01:12:48 2015647 500 166 21 313 +103.73 +/- 17.92
9409 ncm-et-10 2018-03-17 23:09 01:14:01 2020999 500 160 27 313 +94.7 +/- 18.07
9408 ncm-et-9 2018-03-17 23:09 01:11:42 2024099 500 154 17 329 +97.69 +/- 17.03
9407 ncm-et-5 2018-03-17 23:09 01:12:05 2025245 500 167 11 322 +112.14 +/- 17.18
9406 ncm-et-14 2018-03-17 23:09 01:13:23 2019860 500 176 16 308 +115.23 +/- 18.03
9405 ncm-et-12 2018-03-17 23:09 01:13:12 1976969 500 162 27 311 +96.19 +/- 18.16
9404 ncm-et-6 2018-03-17 23:09 01:14:36 2025735 500 180 21 299 +114.45 +/- 18.61
9403 ncm-et-4 2018-03-17 23:09 01:17:02 1926233 500 160 13 327 +105.25 +/- 17.0
9402 ncm-et-15 2018-03-17 23:09 01:13:06 2022960 500 165 25 310 +99.95 +/- 18.17
9401 ncm-et-13 2018-03-17 23:09 01:13:14 2023774 500 177 16 307 +116.0 +/- 18.07

Commit

Commit ID 8b15961349e18a9ba113973c53f53913d0cd0fad
Author joergoster
Date 2017-05-04 02:46:40 UTC
Fix multiPV issue #502 In general, this patch handles the cases where we don't have a valid score for each PV line in a multiPV search. This can happen if the search has been stopped in an unfortunate moment while still in the aspiration loop. The patch consists of two parts. Part 1: The new PVIdx was already part of the k-best pv's in the last iteration, and we therefore have a valid pv and score to output from the last iteration. This is taken care of with: bool updated = (i <= PVIdx && rootMoves[i].score != -VALUE_INFINITE); Case 2: The new PVIdx was NOT part of the k-best pv's in the last iteration, and we have no valid pv and score to output. Not from the current nor from the previous iteration. To avoid this, we are now also considering the previous score when sorting, so that the PV lines with no actual but with a valid previous score are pushed up again, and the previous score can be displayed. bool operator<(const RootMove& m) const { return m.score != score ? m.score < score : m.previousScore < previousScore; } // Descending sort I also added an assertion in UCI::value() to possibly catch similar issues earlier. No functional change. Closes #502 Closes #1074
Copyright 2011–2024 Next Chess Move LLC