| Windows x64 for Haswell CPUs Windows x64 for modern computers Windows x64 Windows 32 Linux x64 for Haswell CPUs Linux x64 for modern computers Linux x64 | Author: protonspring
Date: Wed Mar 25 21:42:33 2020 +0100 Timestamp: 1585168953 Remove KNPKB endgame. This is a functional simplification that removes the KNPKB endgame. Testing on only KNPKB positions suggests that this removal actually gains Elo: Score of patch vs master: 3380 - 3035 - 33585 [0.504] 40000 Elo difference: 3.0 +/- 1.4, LOS: 100.0 %, DrawRatio: 84.0 % Score of patch vs master: 290 - 36 - 39674 [0.503] 40000 Elo difference: 2.2 +/- 0.3, LOS: 100.0 %, DrawRatio: 99.2 % removal also doesn't cause a regression with the standard book: STC LLR: 2.95 (-2.94,2.94) {-1.50,0.50} Total: 71376 W: 13794 L: 13719 D: 43863 Elo +0.37 Ptnml(0-2): 1066, 8092, 17290, 8181, 1059 https://tests.stockfishchess.org/tests/view/5e76c3d5e42a5c3b3ca2e8be LTC LLR: 2.95 (-2.94,2.94) {-1.50,0.50} Total: 28394 W: 3731 L: 3662 D: 21001 Elo +0.84 Ptnml(0-2): 167, 2339, 9116, 2408, 167 https://tests.stockfishchess.org/tests/view/5e76e5eae42a5c3b3ca2e8d3 closes https://github.com/official-stockfish/Stockfish/pull/2594 Bench 5480811 see source |
| Windows x64 for Haswell CPUs Windows x64 for modern computers Windows x64 Windows 32 Linux x64 for Haswell CPUs Linux x64 for modern computers Linux x64 | Author: Vizvezdenec
Date: Fri Mar 20 21:07:37 2020 +0100 Timestamp: 1584734857 Adjust singular extension search depth This patch applies a different singular extension search logic in case the position is ttPv && !PvNode. It changes the depth of this search, higher for this types of nodes, and lower for other nodes. passed STC http://tests.stockfishchess.org/tests/view/5e72bbaae42a5c3b3ca2e75e LLR: 2.94 (-2.94,2.94) {-0.50,1.50} Total: 12692 W: 2608 L: 2389 D: 7695 Elo +6.00 Ptnml(0-2): 238, 1414, 2839, 1601, 254 passed LTC http://tests.stockfishchess.org/tests/view/5e731c07e42a5c3b3ca2e770 LLR: 2.96 (-2.94,2.94) {0.25,1.75} Total: 145716 W: 19218 L: 18626 D: 107872 Elo +1.41 Ptnml(0-2): 1100, 13605, 42899, 14111, 1143 closes https://github.com/official-stockfish/Stockfish/pull/2590 Bench: 5398277 see source |
| Windows x64 for Haswell CPUs Windows x64 for modern computers Windows x64 Windows 32 Linux x64 for Haswell CPUs Linux x64 for modern computers Linux x64 | Author: Vizvezdenec
Date: Tue Mar 17 19:34:47 2020 +0100 Timestamp: 1584470087 Adjust singular LMR for positions seen in PV This patch continues work on altering search for ttPv nodes, using recent idea to alter it more in not PvNodes. Previous tweak based on this idea adjusted singularBeta - this one adjusts value of singularLMR, so they are both related to singular extension search. passed STC http://tests.stockfishchess.org/tests/view/5e700737e42a5c3b3ca2e659 LLR: 2.95 (-2.94,2.94) {-0.50,1.50} Total: 140608 W: 27053 L: 26659 D: 86896 Elo +0.97 Ptnml(0-2): 2425, 16337, 32439, 16625, 2478 passed LTC http://tests.stockfishchess.org/tests/view/5e7068eae42a5c3b3ca2e687 LLR: 2.94 (-2.94,2.94) {0.25,1.75} Total: 79318 W: 10463 L: 10064 D: 58791 Elo +1.75 Ptnml(0-2): 567, 7416, 23359, 7685, 632 closes https://github.com/official-stockfish/Stockfish/pull/2588 Bench: 4952322 see source |
| Windows x64 for Haswell CPUs Windows x64 for modern computers Windows x64 Windows 32 Linux x64 for Haswell CPUs Linux x64 for modern computers Linux x64 | Author: Stéphane Nicolet
Date: Tue Mar 17 19:20:37 2020 +0100 Timestamp: 1584469237 Anchored bishops Reduce the "bad bishop" penalty when the bishop is protected by one of our pawns, as it may indicate that the bishop has found a safe spot outside the pawn chain. STC: LLR: 2.94 (-2.94,2.94) {-0.50,1.50} Total: 176942 W: 34142 L: 33696 D: 109104 Elo +0.88 Ptnml(0-2): 3129, 20422, 40919, 20876, 3125 http://tests.stockfishchess.org/tests/view/5e6f61aae42a5c3b3ca2e62d LTC: LLR: 2.95 (-2.94,2.94) {0.25,1.75} Total: 42252 W: 5615 L: 5322 D: 31315 Elo +2.41 Ptnml(0-2): 308, 3881, 12500, 4084, 353 http://tests.stockfishchess.org/tests/view/5e701382e42a5c3b3ca2e661 closes https://github.com/official-stockfish/Stockfish/pull/2587 Bench: 4963440 see source |
| Windows x64 for Haswell CPUs Windows x64 for modern computers Windows x64 Windows 32 Linux x64 for Haswell CPUs Linux x64 for modern computers Linux x64 | Author: Joost VandeVondele
Date: Sat Mar 14 17:04:50 2020 +0100 Timestamp: 1584201890 Small cleanups closes https://github.com/official-stockfish/Stockfish/pull/2567 No functional change. see source |
| Windows x64 for Haswell CPUs Windows x64 for modern computers Windows x64 Windows 32 Linux x64 for Haswell CPUs Linux x64 for modern computers Linux x64 | Author: pb00067
Date: Sat Mar 14 16:30:17 2020 +0100 Timestamp: 1584199817 Simplify futility pruning parent node only continuation histories seem needed for this purpose. STC: http://tests.stockfishchess.org/tests/view/5e6b88dfe42a5c3b3ca2e4ab LLR: 2.94 (-2.94,2.94) {-1.50,0.50} Total: 113356 W: 21725 L: 21696 D: 69935 Elo +0.09 Ptnml(0-2): 1999, 13255, 26163, 13240, 2021 LTC: http://tests.stockfishchess.org/tests/view/5e6babbfe42a5c3b3ca2e4c2 LLR: 2.94 (-2.94,2.94) {-1.50,0.50} Total: 22164 W: 2917 L: 2821 D: 16426 Elo +1.50 Ptnml(0-2): 173, 2040, 6548, 2160, 161 closes https://github.com/official-stockfish/Stockfish/pull/2583 bench: 4839496 see source |
| Windows x64 for Haswell CPUs Windows x64 for modern computers Windows x64 Windows 32 Linux x64 for Haswell CPUs Linux x64 for modern computers Linux x64 | Author: protonspring
Date: Sat Mar 14 16:27:54 2020 +0100 Timestamp: 1584199674 simplify castling part of generate_all. somewhat more compact, generates same code. close https://github.com/official-stockfish/Stockfish/pull/2580 No functional change. see source |
| Windows x64 for Haswell CPUs Windows x64 for modern computers Windows x64 Windows 32 Linux x64 for Haswell CPUs Linux x64 for modern computers Linux x64 | Author: silversolver1
Date: Mon Mar 9 22:46:13 2020 +0100 Timestamp: 1583790373 Remove set statScore to zero Simplification. Removes setting statScore to zero if negative. STC: LLR: 2.95 (-2.94,2.94) {-1.50,0.50} Total: 84820 W: 16100 L: 16033 D: 52687 Elo +0.27 Ptnml(0-2): 1442, 9865, 19723, 9944, 1436 https://tests.stockfishchess.org/tests/view/5e654fdae42a5c3b3ca2e2f8 LTC: LLR: 2.94 (-2.94,2.94) {-1.50,0.50} Total: 57658 W: 7435 L: 7391 D: 42832 Elo +0.27 Ptnml(0-2): 441, 5397, 17104, 5451, 436 https://tests.stockfishchess.org/tests/view/5e657ce9e42a5c3b3ca2e307 closes https://github.com/official-stockfish/Stockfish/pull/2578 Bench: 5168890 see source |
| Windows x64 for Haswell CPUs Windows x64 for modern computers Windows x64 Windows 32 Linux x64 for Haswell CPUs Linux x64 for modern computers Linux x64 | Author: protonspring
Date: Mon Mar 9 22:16:26 2020 +0100 Timestamp: 1583788586 Equations for edges and corners. This is a functional simplification that removes the large arrays in endgames.cpp. It also fixes a recently introduced bug (960d59d54143d84aab26deae65279a611fc989f4) in KNBvK, now using flip_file() instead of ~. One fen added to bench to increase endgame coverage. STC LLR: 2.94 (-2.94,2.94) {-1.50,0.50} Total: 174724 W: 33325 L: 33404 D: 107995 Elo -0.16 Ptnml(0-2): 2503, 19607, 43181, 19608, 2463 http://tests.stockfishchess.org/tests/view/5e6448ffe42a5c3b3ca2e287 LTC LLR: 2.95 (-2.94,2.94) {-1.50,0.50} Total: 35640 W: 4679 L: 4621 D: 26340 Elo +0.57 Ptnml(0-2): 189, 2991, 11424, 3005, 211 http://tests.stockfishchess.org/tests/view/5e650b24e42a5c3b3ca2e2d8 closes https://github.com/official-stockfish/Stockfish/pull/2577 Bench: 5527957 see source |
| Windows x64 for Haswell CPUs Windows x64 for modern computers Windows x64 Windows 32 Linux x64 for Haswell CPUs Linux x64 for modern computers Linux x64 | Author: Gary Heckman
Date: Sat Mar 7 11:14:27 2020 +0100 Timestamp: 1583576067 Fix ambiguity between clamp implementations There is an ambiguity between global and std clamp implementations when compiling in c++17, and on certain toolchains that are not strictly conforming to c++11. This is solved by putting our clamp implementation in a namespace. closes https://github.com/official-stockfish/Stockfish/pull/2572 No functional change. see source |
| Windows x64 for Haswell CPUs Windows x64 for modern computers Windows x64 Windows 32 Linux x64 for Haswell CPUs Linux x64 for modern computers Linux x64 | Author: protonspring
Date: Sat Mar 7 11:09:38 2020 +0100 Timestamp: 1583575778 Remove KRPPKRPScaleFactors array Fucntional simplification that removes the KRPPKRPScaleFactors array. STC LLR: 2.94 (-2.94,2.94) {-1.50,0.50} Total: 47374 W: 9159 L: 9049 D: 29166 Elo +0.81 Ptnml(0-2): 707, 5325, 11560, 5341, 754 http://tests.stockfishchess.org/tests/view/5e5ff464e42a5c3b3ca2e156 LTC LLR: 2.95 (-2.94,2.94) {-1.50,0.50} Total: 31268 W: 4064 L: 3995 D: 23209 Elo +0.77 Ptnml(0-2): 173, 2734, 9764, 2777, 186 http://tests.stockfishchess.org/tests/view/5e61be6ce42a5c3b3ca2e1c1 closes https://github.com/official-stockfish/Stockfish/pull/2575 Bench 5123316 see source |
| Windows x64 for Haswell CPUs Windows x64 for modern computers Windows x64 Windows 32 Linux x64 for Haswell CPUs Linux x64 for modern computers Linux x64 | Author: protonspring
Date: Sat Mar 7 10:55:15 2020 +0100 Timestamp: 1583574915 Cleanup KBPsK endgame * Clarify distinction between strong side pawns and all pawns. * Simplify and speed-up determination of pawns on the same file. * Clarify comments. * more_than_one() is probably faster than pos.count. Passed STC: LLR: 2.95 (-2.94,2.94) {-1.50,0.50} Total: 40696 W: 7856 L: 7740 D: 25100 Elo +0.99 Ptnml(0-2): 584, 4519, 10054, 4579, 612 http://tests.stockfishchess.org/tests/view/5e6153b1e42a5c3b3ca2e1a9 closes https://github.com/official-stockfish/Stockfish/pull/2574 No functional change. see source |
| Windows x64 for Haswell CPUs Windows x64 for modern computers Windows x64 Windows 32 Linux x64 for Haswell CPUs Linux x64 for modern computers Linux x64 | Author: protonspring
Date: Sat Mar 7 10:48:16 2020 +0100 Timestamp: 1583574496 Small speed-up in BetweenBB A speed-up removing some comparisons. closes https://github.com/official-stockfish/Stockfish/pull/2571 No functional change. see source |
| Windows x64 for Haswell CPUs Windows x64 for modern computers Windows x64 Windows 32 Linux x64 for Haswell CPUs Linux x64 for modern computers Linux x64 | Author: protonspring
Date: Wed Mar 4 07:17:04 2020 +0100 Timestamp: 1583302624 Use equations for PushAway and PushClose A functional simplification replacing the corresponding arrays. Tested in two variants, also the simpler one performs well, even though differences to master should be minimal. STC LLR: 2.95 (-2.94,2.94) {-1.50,0.50} Total: 57864 W: 11092 L: 11001 D: 35771 Elo +0.55 Ptnml(0-2): 826, 6458, 14320, 6455, 873 http://tests.stockfishchess.org/tests/view/5e5da5b6e42a5c3b3ca2e05c LTC LLR: 2.95 (-2.94,2.94) {-1.50,0.50} Total: 7198 W: 982 L: 883 D: 5333 Elo +4.78 Ptnml(0-2): 33, 575, 2296, 650, 45 http://tests.stockfishchess.org/tests/view/5e5df13ae42a5c3b3ca2e077 LTC (This exact version. . . more simplified) LLR: 2.95 (-2.94,2.94) {-1.50,0.50} Total: 5392 W: 729 L: 631 D: 4032 Elo +6.32 Ptnml(0-2): 23, 405, 1751, 485, 32 http://tests.stockfishchess.org/tests/view/5e5ead99e42a5c3b3ca2e0e4 closes https://github.com/official-stockfish/Stockfish/pull/2570 Bench 5123316 see source |
| Windows x64 for Haswell CPUs Windows x64 for modern computers Windows x64 Windows 32 Linux x64 for Haswell CPUs Linux x64 for modern computers Linux x64 | Author: protonspring
Date: Wed Mar 4 07:11:48 2020 +0100 Timestamp: 1583302308 Consolidate Square Flipping Add a flip_rank() and flip_file() so that all of the square flipping can be consolidated. STC LLR: 2.94 (-2.94,2.94) {-1.50,0.50} Total: 57234 W: 11064 L: 10969 D: 35201 Elo +0.58 Ptnml(0-2): 822, 6562, 13801, 6563, 869 http://tests.stockfishchess.org/tests/view/5e5d2f2aafe6254521f2ffaa closes https://github.com/official-stockfish/Stockfish/pull/2568 No functional change. see source |
| Windows x64 for Haswell CPUs Windows x64 for modern computers Windows x64 Windows 32 Linux x64 for Haswell CPUs Linux x64 for modern computers Linux x64 | Author: Joost VandeVondele
Date: Sun Mar 1 09:31:58 2020 +0100 Timestamp: 1583051518 Small cleanups closes https://github.com/official-stockfish/Stockfish/pull/2546 No functional change. see source |
| Windows x64 for Haswell CPUs Windows x64 for modern computers Windows x64 Windows 32 Linux x64 for Haswell CPUs Linux x64 for modern computers Linux x64 | Author: Guenther Demetz
Date: Sun Mar 1 09:20:31 2020 +0100 Timestamp: 1583050831 Simplify lowply-history logic Don't restrict usage to ttPv nodes exclusively STC: http://tests.stockfishchess.org/tests/view/5e5634f284a82b4acd41499a LLR: 2.94 (-2.94,2.94) {-1.50,0.50} Total: 152796 W: 29146 L: 29178 D: 94472 Elo -0.07 Ptnml(0-2): 2590, 17792, 35628, 17836, 2552 LTC: http://tests.stockfishchess.org/tests/view/5e575d4984a82b4acd4149e8 LLR: 2.95 (-2.94,2.94) {-1.50,0.50} Total: 20078 W: 2688 L: 2587 D: 14803 Elo +1.75 Ptnml(0-2): 139, 1914, 5853, 1973, 160 closes https://github.com/official-stockfish/Stockfish/pull/2565 bench: 4923286 see source |
| Windows x64 for Haswell CPUs Windows x64 for modern computers Windows x64 Windows 32 Linux x64 for Haswell CPUs Linux x64 for modern computers Linux x64 | Author: Moez Jellouli
Date: Thu Feb 27 13:14:51 2020 +0100 Timestamp: 1582805691 Weak queen protection Extra penalty if weak piece is only protected by a queen. STC: http://tests.stockfishchess.org/tests/view/5e53c6ab84a82b4acd4148fa LLR: 2.96 (-2.94,2.94) {-0.50,1.50} Total: 44630 W: 8615 L: 8359 D: 27656 Elo +1.99 Ptnml(0-2): 746, 5156, 10323, 5276, 814 LTC: http://tests.stockfishchess.org/tests/view/5e54e05d84a82b4acd414947 LLR: 2.94 (-2.94,2.94) {0.25,1.75} Total: 175480 W: 23085 L: 22409 D: 129986 Elo +1.34 Ptnml(0-2): 1264, 16494, 51678, 16910, 1394 closes https://github.com/official-stockfish/Stockfish/pull/2564 Bench: 4923286 see source |
| Windows x64 for Haswell CPUs Windows x64 for modern computers Windows x64 Windows 32 Linux x64 for Haswell CPUs Linux x64 for modern computers Linux x64 | Author: AndyGrant
Date: Tue Feb 25 21:10:10 2020 +0100 Timestamp: 1582661410 Fix TT write in MultiPV case. fixes an error reported earlier as https://github.com/official-stockfish/Stockfish/issues/2404 by @AndyGrant. MultiPV at root shouldn't write to the TT for later lines, as that is neither the eval nor the bestmove for that position. Fixing this error doesn't matter for playing games (http://tests.stockfishchess.org/tests/view/5dcdbd810ebc590256324a11). However, it can lead to wrong mate announcements as reported by @uriblass. In particular the following testcase gives wrong results for the second search, prior to this patch: ``` setoption name MultiPV value 2 position fen 5R2/2kB2p1/p2bR3/8/3p1B2/8/PPP5/2K5 b - - 0 49 go depth 40 position fen 2B2R2/3r2p1/p1kbR3/8/3p1B2/8/PPP5/2K5 b - - 8 48 go depth 40 ``` fixes https://github.com/official-stockfish/Stockfish/issues/2561 closes https://github.com/official-stockfish/Stockfish/pull/2562 Only affects MultiPV search. Bench: 4697493 see source |
| Windows x64 for Haswell CPUs Windows x64 for modern computers Windows x64 Windows 32 Linux x64 for Haswell CPUs Linux x64 for modern computers Linux x64 | Author: 31m059
Date: Tue Feb 25 21:03:42 2020 +0100 Timestamp: 1582661022 Use single param for Outpost and ReachableOutpost. In November 2019, as a result of the simplification of rank-based outposts by 37698b0, separate bonuses were introduced for outposts that are currently occupied and outposts that are reachable on the next move. However, the values of these two bonuses are quite similar, and they have remained that way for three months of development. It appears that we can safely retire the separate ReachableOutpost parameter and use the same Outpost bonus in both cases, restoring the basic principles of Stockfish outpost evaluation to their pre-November state, while also reducing the size of the parameter space. STC: LLR: 2.96 (-2.94,2.94) {-1.50,0.50} Total: 47680 W: 9213 L: 9092 D: 29375 Elo +0.88 Ptnml(0-2): 776, 5573, 11071, 5594, 826 https://tests.stockfishchess.org/tests/view/5e51e33190a0a02810d09802 LTC: LLR: 2.94 (-2.94,2.94) {-1.50,0.50} Total: 14690 W: 1960 L: 1854 D: 10876 Elo +2.51 Ptnml(0-2): 93, 1381, 4317, 1435, 119 https://tests.stockfishchess.org/tests/view/5e52197990a0a02810d0980f closes https://github.com/official-stockfish/Stockfish/pull/2559 Bench: 4697493 see source |
| Windows x64 for Haswell CPUs Windows x64 for modern computers Windows x64 Windows 32 Linux x64 for Haswell CPUs Linux x64 for modern computers Linux x64 | Author: Günther Demetz
Date: Sat Feb 22 21:32:32 2020 +0100 Timestamp: 1582403552 Improve move order near the root Current move histories are known to work well near the leaves, whilst at higher depths they aren't very helpful. To address this problem this patch introduces a table dedicated for what's happening at plies 0-3. It's structured like mainHistory with ply index instead of color. It get cleared with each new search and is filled during iterative deepening at higher depths when recording successful quiet moves near the root or traversing nodes which were in the principal variation (ttPv). Medium TC (20+0.2): https://tests.stockfishchess.org/tests/view/5e4d358790a0a02810d096dc LLR: 2.94 (-2.94,2.94) {-0.50,1.50} Total: 100910 W: 16682 L: 16376 D: 67852 Elo +1.05 Ptnml(0-2): 1177, 10983, 25883, 11181, 1231 LTC: https://tests.stockfishchess.org/tests/view/5e4e2cb790a0a02810d09714 LLR: 2.95 (-2.94,2.94) {0.25,1.75} Total: 80444 W: 10495 L: 10095 D: 59854 Elo +1.73 Ptnml(0-2): 551, 7479, 23803, 7797, 592 closes https://github.com/official-stockfish/Stockfish/pull/2557 Bench: 4705960 see source |
| Windows x64 for Haswell CPUs Windows x64 for modern computers Windows x64 Windows 32 Linux x64 for Haswell CPUs Linux x64 for modern computers Linux x64 | Author: protonspring
Date: Thu Feb 20 08:32:17 2020 +0100 Timestamp: 1582183937 Updated KNNKP endgame. This is a patch that significantly improves playing KNNKP endgames: ``` Score of 2553 vs master: 132 - 38 - 830 [0.547] 1000 Elo difference: 32.8 +/- 8.7, LOS: 100.0 %, DrawRatio: 83.0 % ``` At the same time it reduces the evaluation of this mostly draw engame from ~7.5 to ~1.5 This patch does not regress against master in normal games: STC LLR: 2.94 (-2.94,2.94) {-1.50,0.50} Total: 96616 W: 18459 L: 18424 D: 59733 Elo +0.13 Ptnml(0-2): 1409, 10812, 23802, 10905, 1380 http://tests.stockfishchess.org/tests/view/5e49dfe6f8d1d52b40cd31bc LTC LLR: 2.95 (-2.94,2.94) {-1.50,0.50} Total: 49726 W: 6340 L: 6304 D: 37082 Elo +0.25 Ptnml(0-2): 239, 4227, 15906, 4241, 250 http://tests.stockfishchess.org/tests/view/5e4ab9ee16fb3df8c4cc01d0 Theory: KNNK is a dead draw, however the presence of the additional weakSide pawn opens up some mate opportunities. The idea is to block the pawn (preferably behind the Troitsky line) with one of the knights and press the weakSide king into a corner. If we can stalemate the king, we release the pawn with the knight (to avoid actual stalemate), and use the knight to complete the mate before the pawn promotes. This is also why there is an additional penalty for advancement of the pawn. closes https://github.com/official-stockfish/Stockfish/pull/2553 Bench: 4981770 see source |
| Windows x64 for Haswell CPUs Windows x64 for modern computers Windows x64 Windows 32 Linux x64 for Haswell CPUs Linux x64 for modern computers Linux x64 | Author: protonspring
Date: Sat Feb 15 15:13:14 2020 +0100 Timestamp: 1581775994 Simplify Futility Move Count remove two constants STC LLR: 2.95 (-2.94,2.94) {-1.50,0.50} Total: 62050 W: 11903 L: 11802 D: 38345 Elo +0.57 Ptnml(0-2): 1002, 7346, 14283, 7320, 1065 http://tests.stockfishchess.org/tests/view/5e41d73be70d848499f63c6d LTC LLR: 2.93 (-2.94,2.94) {-1.50,0.50} Total: 12850 W: 1679 L: 1572 D: 9599 Elo +2.89 Ptnml(0-2): 82, 1171, 3818, 1249, 96 http://tests.stockfishchess.org/tests/view/5e42bf07e70d848499f63cc0 Bench: 4762351 see source |
| Windows x64 for Haswell CPUs Windows x64 for modern computers Windows x64 Windows 32 Linux x64 for Haswell CPUs Linux x64 for modern computers Linux x64 | Author: Joost VandeVondele
Date: Sat Feb 15 15:10:07 2020 +0100 Timestamp: 1581775807 Fix for incorrect VALUE_MATE_IN_MAX_PLY usage. Fixes #2533, fixes #2543, fixes #2423. the code that prevents false mate announcements depending on the TT state (GHI), incorrectly used VALUE_MATE_IN_MAX_PLY. The latter constant, however, also includes, counterintuitively, the TB win range. This patch fixes that, by restoring the behavior for TB win scores, while retaining the false mate correctness, and improving the mate finding ability. In particular no alse mates are announced with the poisened hash testcase ``` position fen 8/8/8/3k4/8/8/6K1/7R w - - 0 1 go depth 40 position fen 8/8/8/3k4/8/8/6K1/7R w - - 76 1 go depth 20 ucinewgame ``` mates are found with the testcases reported in #2543 ``` position fen 4k3/3pp3/8/8/8/8/2PPP3/4K3 w - - 0 1 setoption name Hash value 1024 go depth 55 ucinewgame ``` and ``` position fen 4k3/4p3/8/8/8/8/3PP3/4K3 w - - 0 1 setoption name Hash value 1024 go depth 45 ucinewgame ``` furthermore, on the mate finding benchmark (ChestUCI_23102018.epd), performance improves over master, roughly reaching performance with the false mate protection reverted ``` Analyzing 6566 mate positions for best and found mates: ----------------best ---------------found nodes master revert fixed master revert fixed 16000000 4233 4236 4235 5200 5201 5199 32000000 4583 4585 4585 5417 5424 5418 64000000 4852 4853 4855 5575 5584 5579 128000000 5071 5068 5066 5710 5720 5716 256000000 5280 5282 5279 5819 5827 5826 512000000 5471 5468 5468 5919 5935 5932 ``` On a testcase with TB enabled, progress is made consistently, contrary to master ``` setoption name SyzygyPath value ../../../syzygy/3-4-5/ setoption name Hash value 2048 position fen 1R6/3k4/8/K2p4/4n3/2P5/8/8 w - - 0 1 go depth 58 ucinewgame ``` The PR (prior to a rewrite for clarity) passed STC: LLR: 2.94 (-2.94,2.94) {-1.50,0.50} Total: 65405 W: 12454 L: 12384 D: 40567 Elo +0.37 Ptnml(0-2): 920, 7256, 16285, 7286, 944 http://tests.stockfishchess.org/tests/view/5e441a3be70d848499f63d15 passed LTC: LLR: 2.94 (-2.94,2.94) {-1.50,0.50} Total: 27096 W: 3477 L: 3413 D: 20206 Elo +0.82 Ptnml(0-2): 128, 2215, 8776, 2292, 122 http://tests.stockfishchess.org/tests/view/5e44e277e70d848499f63d63 The incorrectly named VALUE_MATE_IN_MAX_PLY and VALUE_MATED_IN_MAX_PLY were renamed into VALUE_TB_WIN_IN_MAX_PLY and VALUE_TB_LOSS_IN_MAX_PLY, and correclty defined VALUE_MATE_IN_MAX_PLY and VALUE_MATED_IN_MAX_PLY were introduced. One further (corner case) mistake using these constants was fixed (go mate X), which could lead to a premature return if X > MAX_PLY / 2, but TB were present. Thanks to @svivanov72 for one of the reports and help fixing the issue. closes https://github.com/official-stockfish/Stockfish/pull/2552 Bench: 4932981 see source |
| Windows x64 for Haswell CPUs Windows x64 for modern computers Windows x64 Windows 32 Linux x64 for Haswell CPUs Linux x64 for modern computers Linux x64 | Author: Vizvezdenec
Date: Mon Feb 10 09:19:02 2020 +0100 Timestamp: 1581322742 Modify singular beta for ttPv positions. This patch lowers singular beta for positions that have been in pv and are not pv nodes. The idea of using ttpv && !PvNode improved scaling with TC and could be useful for other search heuristics. passed STC http://tests.stockfishchess.org/tests/view/5e3f6d7ce70d848499f63bbc LLR: 2.94 (-2.94,2.94) {-0.50,1.50} Total: 154953 W: 29688 L: 29272 D: 95993 Elo +0.93 Ptnml(0-2): 2616, 17912, 36037, 18210, 2673 passed LTC http://tests.stockfishchess.org/tests/view/5e405561e70d848499f63bfa LLR: 2.95 (-2.94,2.94) {0.25,1.75} Total: 70974 W: 9305 L: 8932 D: 52737 Elo +1.83 Ptnml(0-2): 466, 6658, 20920, 6826, 569 closes https://github.com/official-stockfish/Stockfish/pull/2550 Bench: 4932981 see source |