Dev Builds » 20170917-0844

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 06:01:42 2001342 2456 887 81 1488 +118.4 +/- 8.24
ncm-et-4 06:01:55 2000978 2442 879 87 1476 +116.9 +/- 8.3
ncm-et-5 01:53:23 2023854 774 277 21 476 +119.4 +/- 14.44
ncm-et-6 01:53:28 2026145 767 300 16 451 +135.06 +/- 14.95
ncm-et-7 01:53:21 1982747 763 281 24 458 +121.78 +/- 14.89
ncm-et-8 01:53:26 1989840 755 247 29 479 +103.26 +/- 14.37
ncm-et-9 06:01:55 2001406 2476 882 85 1509 +115.96 +/- 8.18
ncm-et-10 06:01:59 1977721 2422 845 80 1497 +113.62 +/- 8.17
ncm-et-11 01:53:16 1963284 760 257 32 471 +106.03 +/- 14.66
ncm-et-12 01:52:44 2017838 764 291 23 450 +127.28 +/- 15.09
ncm-et-13 06:01:39 1990576 2410 883 77 1450 +120.84 +/- 8.36
ncm-et-14 01:53:04 2018477 758 267 23 468 +115.96 +/- 14.58
ncm-et-15 06:02:03 1980491 2453 886 79 1488 +118.72 +/- 8.24
20000 7182 657 12161 +117.65 +/- 2.88

Test Detail

ID Host Started (UTC) Duration Base NPS Games Wins Losses Draws Elo CLI PGN
55332 ncm-et-10 2018-09-26 14:49 00:21:00 1979349 136 44 5 87 +102.51 +/- 33.68
55331 ncm-et-13 2018-09-26 14:49 00:20:50 1989560 141 54 3 84 +131.62 +/- 34.66
55330 ncm-et-4 2018-09-26 14:44 00:26:20 1987981 175 55 8 112 +95.66 +/- 29.85
55329 ncm-et-3 2018-09-26 14:43 00:26:52 1989874 178 70 4 104 +135.27 +/- 31.33
55328 ncm-et-15 2018-09-26 14:43 00:27:06 1902941 175 71 4 100 +140.16 +/- 32.16
55327 ncm-et-9 2018-09-26 14:42 00:28:23 1991455 195 75 6 114 +128.49 +/- 30.13
55326 ncm-et-13 2018-09-26 13:33 01:15:19 1967504 500 190 14 296 +127.76 +/- 18.55
55325 ncm-et-10 2018-09-26 13:32 01:16:37 1928468 500 169 15 316 +110.6 +/- 17.6
55324 ncm-et-3 2018-09-26 13:29 01:13:20 1990349 500 172 15 313 +112.91 +/- 17.75
55323 ncm-et-4 2018-09-26 13:28 01:14:34 1988139 500 193 16 291 +128.55 +/- 18.86
55322 ncm-et-9 2018-09-26 13:28 01:13:03 1989876 500 177 16 307 +116.0 +/- 18.07
55321 ncm-et-15 2018-09-26 13:27 01:15:14 1979822 500 187 18 295 +122.24 +/- 18.72
55320 ncm-et-13 2018-09-26 12:15 01:16:53 1949206 500 192 17 291 +126.97 +/- 18.89
55319 ncm-et-10 2018-09-26 12:14 01:16:19 1952487 500 183 19 298 +118.33 +/- 18.6
55318 ncm-et-4 2018-09-26 12:14 01:13:06 1989876 500 166 20 314 +104.49 +/- 17.85
55317 ncm-et-3 2018-09-26 12:13 01:14:32 1989558 500 177 11 312 +119.89 +/- 17.66
55316 ncm-et-9 2018-09-26 12:13 01:13:28 1986405 500 175 23 302 +109.07 +/- 18.51
55315 ncm-et-15 2018-09-26 12:12 01:13:25 1980446 500 175 12 313 +117.55 +/- 17.65
55314 ncm-et-3 2018-09-26 10:59 01:13:41 1988927 500 182 24 294 +113.68 +/- 18.93
55313 ncm-et-15 2018-09-26 10:58 01:12:54 1982948 500 185 16 299 +122.24 +/- 18.47
55312 ncm-et-10 2018-09-26 10:58 01:14:41 1963861 500 168 20 312 +106.01 +/- 17.94
55311 ncm-et-13 2018-09-26 10:58 01:15:13 1987351 500 173 15 312 +113.68 +/- 17.8
55310 ncm-et-9 2018-09-26 10:58 01:13:50 1989559 500 181 16 303 +119.11 +/- 18.27
55309 ncm-et-4 2018-09-26 10:58 01:14:46 1989874 500 194 21 285 +125.39 +/- 19.3
8490 ncm-et-7 2018-03-15 02:24 00:37:46 1986189 263 99 11 153 +120.91 +/- 26.24
8489 ncm-et-11 2018-03-15 02:23 00:38:15 1991640 260 96 10 154 +119.41 +/- 25.96
8488 ncm-et-14 2018-03-15 02:23 00:38:21 2017095 258 86 7 165 +109.91 +/- 24.21
8487 ncm-et-8 2018-03-15 02:23 00:38:29 2020673 255 84 8 163 +106.79 +/- 24.46
8486 ncm-et-4 2018-03-15 02:23 00:38:54 2024264 267 85 11 171 +98.88 +/- 24.05
8485 ncm-et-13 2018-03-15 02:23 00:39:09 2024100 269 86 10 173 +100.9 +/- 23.8
8484 ncm-et-12 2018-03-15 02:22 00:39:10 2020184 264 99 5 160 +129.37 +/- 24.87
8483 ncm-et-6 2018-03-15 02:22 00:40:01 2025900 267 96 8 163 +118.95 +/- 24.83
8482 ncm-et-9 2018-03-15 02:22 00:39:56 2024917 281 95 7 179 +112.59 +/- 23.23
8481 ncm-et-15 2018-03-15 02:21 00:40:27 2017097 278 101 10 167 +118.07 +/- 24.78
8480 ncm-et-10 2018-03-15 02:21 00:40:48 2020999 286 94 3 189 +114.52 +/- 21.89
8479 ncm-et-3 2018-03-15 02:21 00:40:48 2026064 278 99 6 173 +120.88 +/- 23.76
8478 ncm-et-5 2018-03-15 02:21 00:40:57 2023772 274 106 8 160 +130.01 +/- 25.38
8477 ncm-et-12 2018-03-15 01:07 01:13:34 2015492 500 192 18 290 +126.18 +/- 18.97
8476 ncm-et-4 2018-03-15 01:07 01:14:15 2025736 500 186 11 303 +126.97 +/- 18.11
8475 ncm-et-14 2018-03-15 01:07 01:14:43 2019859 500 181 16 303 +119.11 +/- 18.27
8474 ncm-et-9 2018-03-15 01:07 01:13:15 2026228 500 179 17 304 +116.78 +/- 18.25
8473 ncm-et-7 2018-03-15 01:07 01:15:35 1979305 500 182 13 305 +122.24 +/- 18.08
8472 ncm-et-11 2018-03-15 01:07 01:15:01 1934929 500 161 22 317 +99.2 +/- 17.75
8471 ncm-et-15 2018-03-15 01:07 01:12:57 2019696 500 167 19 314 +106.01 +/- 17.82
8470 ncm-et-3 2018-03-15 01:07 01:12:29 2023282 500 187 21 292 +119.89 +/- 18.95
8469 ncm-et-10 2018-03-15 01:07 01:12:34 2021162 500 187 18 295 +122.24 +/- 18.72
8468 ncm-et-13 2018-03-15 01:07 01:14:15 2025737 500 188 18 294 +123.02 +/- 18.77
8467 ncm-et-8 2018-03-15 01:07 01:14:57 1959007 500 163 21 316 +101.46 +/- 17.78
8466 ncm-et-6 2018-03-15 01:07 01:13:27 2026390 500 204 8 288 +143.89 +/- 18.74
8465 ncm-et-5 2018-03-15 01:07 01:12:26 2023936 500 171 13 316 +113.68 +/- 17.54

Commit

Commit ID ba4e215493de31263b9bd352af79d00193e545bf
Author syzygy
Date 2017-09-17 08:44:10 UTC
Let ss->ply denote the number of plies from the root to the current node This patch lets ss->ply be equal to 0 at the root of the search. Currently, the root has ss->ply == 1, which is less intuitive: - Setting the rootNode bool has to check (ss-1)->ply == 0. - All mate values are off by one: the code seems to assume that mated-in-0 is -VALUE_MATE, mate-1-in-ply is VALUE_MATE-1, mated-in-2-ply is VALUE_MATE+2, etc. But the mate_in() and mated_in() functions are called with ss->ply, which is 1 in at the root. - The is_draw() function currently needs to explain why it has "ply - 1 > i" instead of simply "ply > i". - The ss->ply >= MAX_PLY tests in search() and qsearch() already assume that ss->ply == 0 at the root. If we start at ss->ply == 1, it would make more sense to go up to and including ss->ply == MAX_PLY, so stop at ss->ply > MAX_PLY. See also the asserts testing for 0 <= ss->ply && ss->ply < MAX_PLY. The reason for ss->ply == 1 at the root is the line "ss->ply = (ss-1)->ply + 1" at the start for search() and qsearch(). By replacing this with "(ss+1)->ply = ss->ply + 1" we keep ss->ply == 0 at the root. Note that search() already clears killers in (ss+2), so there is no danger in accessing ss+1. I have NOT changed pv[MAX_PLY + 1] to pv[MAX_PLY + 2] in search() and qsearch(). It seems to me that MAX_PLY + 1 is exactly right: - MAX_PLY entries for ss->ply running from 0 to MAX_PLY-1, and 1 entry for the final MOVE_NONE. I have verified that mate scores are reported correctly. (They were already reported correctly due to the extra ply being rounded down when converting to moves.) The value of seldepth output to the user should probably not change, so I add 1 to it. (Humans count from 1, computers from 0.) A small optimisation I did not include: instead of setting ss->ply in every invocation of search() and qsearch(), it could be set once for all plies at the start of Thread::search(). This saves a couple of instructions per node. No functional change (unless the search searches a branch MAX_PLY deep), so bench does not change.
Copyright 2011–2024 Next Chess Move LLC