Dev Builds » 20200110-0231

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 08:28:15 1963446 3319 2056 46 1217 +243.88 +/- 9.64
ncm-et-4 08:29:20 1943824 3308 2019 55 1234 +237.43 +/- 9.58
ncm-et-9 08:30:17 1966473 3364 2122 58 1184 +248.28 +/- 9.82
ncm-et-10 08:29:21 1962527 3337 2094 44 1199 +248.71 +/- 9.72
ncm-et-13 08:31:02 1972347 3341 2113 62 1166 +248.46 +/- 9.9
ncm-et-15 08:28:40 1968479 3331 2060 64 1207 +240.4 +/- 9.72
20000 12464 329 7207 +244.51 +/- 3.97

Test Detail

ID Host Started (UTC) Duration Base NPS Games Wins Losses Draws Elo CLI PGN
111229 ncm-et-4 2020-01-10 21:48 00:47:51 1902694 308 189 4 115 +241.18 +/- 31.48
111228 ncm-et-3 2020-01-10 21:47 00:48:49 1961994 319 198 1 120 +250.52 +/- 30.51
111227 ncm-et-15 2020-01-10 21:46 00:50:32 1973895 331 192 5 134 +222.39 +/- 29.01
111226 ncm-et-10 2020-01-10 21:44 00:51:33 1967540 337 220 6 111 +260.5 +/- 32.35
111225 ncm-et-13 2020-01-10 21:44 00:52:31 1970014 341 227 9 105 +263.0 +/- 33.43
111224 ncm-et-9 2020-01-10 21:41 00:55:10 1961688 364 218 9 137 +227.13 +/- 29.0
111223 ncm-et-3 2020-01-10 20:29 01:16:59 1965997 500 299 10 191 +229.12 +/- 24.45
111222 ncm-et-15 2020-01-10 20:28 01:16:50 1969402 500 317 7 176 +251.89 +/- 25.5
111221 ncm-et-10 2020-01-10 20:27 01:16:09 1960635 500 289 6 205 +222.92 +/- 23.34
111220 ncm-et-13 2020-01-10 20:26 01:16:27 1973281 500 315 7 178 +249.64 +/- 25.34
111219 ncm-et-4 2020-01-10 20:26 01:21:33 1852350 500 303 12 185 +231.21 +/- 24.94
111218 ncm-et-9 2020-01-10 20:24 01:15:42 1965996 500 322 10 168 +254.16 +/- 26.24
111217 ncm-et-10 2020-01-10 19:10 01:15:45 1959242 500 319 4 177 +257.59 +/- 25.3
111216 ncm-et-3 2020-01-10 19:10 01:18:06 1958470 500 311 9 180 +243.0 +/- 25.24
111215 ncm-et-15 2020-01-10 19:10 01:16:45 1962303 500 313 13 174 +240.82 +/- 25.81
111214 ncm-et-4 2020-01-10 19:09 01:15:38 1972650 500 299 9 192 +230.16 +/- 24.35
111213 ncm-et-13 2020-01-10 19:07 01:17:56 1971728 500 306 10 184 +236.51 +/- 24.97
111212 ncm-et-9 2020-01-10 19:07 01:16:17 1969402 500 303 9 188 +234.38 +/- 24.64
111211 ncm-et-15 2020-01-10 17:55 01:13:53 1970636 500 325 10 165 +257.59 +/- 26.5
111210 ncm-et-3 2020-01-10 17:53 01:16:01 1964917 500 333 2 165 +276.69 +/- 26.21
111209 ncm-et-10 2020-01-10 17:53 01:16:33 1969704 500 335 5 160 +275.45 +/- 26.8
111208 ncm-et-4 2020-01-10 17:51 01:16:39 1967546 500 318 9 173 +250.77 +/- 25.8
111207 ncm-et-9 2020-01-10 17:50 01:15:34 1963854 500 321 6 173 +257.59 +/- 25.71
111206 ncm-et-13 2020-01-10 17:49 01:17:11 1973581 500 331 11 158 +263.42 +/- 27.14
111205 ncm-et-15 2020-01-10 16:36 01:17:37 1973898 500 298 9 193 +229.12 +/- 24.27
111204 ncm-et-3 2020-01-10 16:36 01:15:49 1964147 500 308 6 186 +243.0 +/- 24.68
111203 ncm-et-10 2020-01-10 16:36 01:15:51 1960658 500 314 8 178 +247.41 +/- 25.37
111202 ncm-et-4 2020-01-10 16:34 01:15:45 1971132 500 301 9 190 +232.26 +/- 24.49
111201 ncm-et-9 2020-01-10 16:34 01:15:30 1969552 500 328 9 163 +262.24 +/- 26.65
111200 ncm-et-13 2020-01-10 16:33 01:15:17 1973900 500 321 9 170 +254.16 +/- 26.05
111199 ncm-et-15 2020-01-10 15:19 01:15:55 1963840 500 300 14 186 +226.0 +/- 24.91
111198 ncm-et-3 2020-01-10 15:19 01:15:59 1964301 500 306 9 185 +237.58 +/- 24.86
111197 ncm-et-10 2020-01-10 15:19 01:16:04 1953133 500 318 10 172 +249.64 +/- 25.91
111196 ncm-et-9 2020-01-10 15:17 01:15:33 1974671 500 318 9 173 +250.77 +/- 25.8
111195 ncm-et-4 2020-01-10 15:16 01:17:01 1972983 500 304 7 189 +237.58 +/- 24.5
111194 ncm-et-13 2020-01-10 15:16 01:15:28 1976538 500 303 7 190 +236.51 +/- 24.42
111193 ncm-et-3 2020-01-10 14:01 01:16:32 1964301 500 301 9 190 +232.26 +/- 24.49
111192 ncm-et-15 2020-01-10 14:01 01:17:08 1965380 500 315 6 179 +250.77 +/- 25.22
111191 ncm-et-4 2020-01-10 14:00 01:14:53 1967418 500 305 5 190 +240.82 +/- 24.35
111190 ncm-et-10 2020-01-10 14:00 01:17:26 1966782 500 299 5 196 +234.38 +/- 23.91
111189 ncm-et-9 2020-01-10 14:00 01:16:31 1960154 500 312 6 182 +247.41 +/- 24.99
111188 ncm-et-13 2020-01-10 13:59 01:16:12 1967389 500 310 9 181 +241.91 +/- 25.17

Commit

Commit ID 114ddb789bed2d74d6a786f5da6c9ce63d44de27
Author Joost VandeVondele
Date 2020-01-10 02:31:44 UTC
Update Elo estimates for terms in search This updates estimates from 1.5 year ago, and adds missing terms. All estimates from tests run on fishtest at 10+0.1 (STC), 20000 games, error bars +- 3 Elo, see the original message in the pull request for the full list of tests. Noteworthy changes are step 7 (futility pruning) going from ~30 to ~50 Elo and step 13 (pruning at shallow depth) going from ~170 to ~200 Elo. Full list of tests: https://github.com/official-stockfish/Stockfish/pull/2401 @Rocky640 made the suggestion to look at time control dependence of these terms. I picked two large terms (early futility pruning and singular extension), so with small relative error. It turns out it is actually quite interesting (see figure 1). Contrary to my expectation, the Elo gain for early futility pruning is pretty time control sensitive, while singular extension gain is not. Figure 1: TC dependence of two search terms ![elo_search_tc]( http://cassio.free.fr/divers/elo_search_tc.png ) Going back to the old measurement of futility pruning (30 Elo vs today 50 Elo), the code is actually identical but the margins have changed. It seems like a nice example of how connected terms in search really are, i.e. the value of early futility pruning increased significantly due to changes elsewhere in search. No functional change.
Copyright 2011–2024 Next Chess Move LLC