| 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 Sep 23 07:12:32 2019 +0200 Timestamp: 1569215552 Simplify connected pawn scoring When scoring the connected pawns, replace the intricate ternary expressions choosing the coefficient by a simpler addition of boolean conditions: ` value = Connected * (2 + phalanx - opposed) ` This is the map showing the old coefficients and the new ones: ``` phalanx and unopposed: 3x -> 3x phalanx and opposed: 1.5x -> 2x not phalanx and unopposed: 2x -> 2x not phalanx and opposed: 1x -> 1x ``` STC LLR: 2.95 (-2.94,2.94) [-3.00,1.00] Total: 11354 W: 2579 L: 2437 D: 6338 Elo +4.35 http://tests.stockfishchess.org/tests/view/5d8151f00ebc5971531d244f LTC LLR: 2.96 (-2.94,2.94) [-3.00,1.00] Total: 41221 W: 7001 L: 6913 D: 27307 Elo +0.74 http://tests.stockfishchess.org/tests/view/5d818f930ebc5971531d26d6 Bench: 3959889 blah |
| 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: Mon Sep 23 06:47:59 2019 +0200 Timestamp: 1569214079 Acknowledge fishtest authors Explicitly acknowledge fishtest authors. Their efforts are almost invisible, but essential for the project. Many thanks to https://github.com/glinscott/fishtest/blob/master/AUTHORS ! No functional change. |
| 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: noobpwnftw
Date: Mon Sep 16 15:09:45 2019 +0200 Timestamp: 1568639385 Raise stack size to 8MB for pthreads It seems there is no other way to specify stack size on std::thread than linker flags and the effective flags are named differently in many toolchains. On toolchains where pthread is always available, this patch changes the stack size change in our C++ code via pthread to ensure a minimum stack size of 8MB, instead of relying on linker defaults which may be platform-specific. Also raises default stack size on OSX to current Linux default (8MB) just to be safe. Closes https://github.com/official-stockfish/Stockfish/pull/2303 No functional change |
| 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: Mon Sep 16 01:37:39 2019 +0200 Timestamp: 1568590659 Scale down endgame factor when shuffling This patch decreases the endgame scale factor using the 50 moves counter. Looking at some games with this patch, it seems to have two effects on the playing style: 1) when no progress can be made in late endgames (for instance in fortresses or opposite bishops endgames) the evaluation will be largely tamed down towards a draw value. 2) more interestingly, there is also a small effect in the midgame play because Stockfish will panic a little bit if there are more than four consecutive shuffling moves with an advantage: the engine will try to move a pawn or to exchange a piece to keep the advantage, so the follow-ups of the position will be discovered earlier by the alpha-beta search. passed STC: LLR: 2.95 (-2.94,2.94) [0.50,4.50] Total: 23017 W: 5080 L: 4805 D: 13132 Elo +4.15 http://tests.stockfishchess.org/tests/view/5d7e4aef0ebc59069c36fc74 passed LTC: LLR: 2.95 (-2.94,2.94) [0.00,3.50] Total: 30746 W: 5171 L: 4911 D: 20664 Elo +2.94 http://tests.stockfishchess.org/tests/view/5d7e513d0ebc59069c36ff26 Pull request: https://github.com/official-stockfish/Stockfish/pull/2304 Bench: 4272173 |
| 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: Sun Sep 15 00:32:54 2019 +0200 Timestamp: 1568500374 Introduce midgame initiative This patch finally introduces something that was tried for years: midgame score dependance on complexity of position. More precisely, if the position is very simplified and the complexity measure calculated in the initiative() function is inferior to -50 by an amount d, then we add this value d to the midgame score. One example of play of this patch will be (again!) 4 vs 3 etc same flank endgames where sides have a lot of non-pawn material: 4 vs 3 draw mostly remains the same draw even if we add a lot of equal material to both sides. STC run was stopped after 200k games (and not converging): LLR: -1.75 (-2.94,2.94) [0.50,4.50] Total: 200319 W: 44197 L: 43310 D: 112812 Elo +1.54 http://tests.stockfishchess.org/tests/view/5d7cfdb10ebc5902d386572c passed LTC: LLR: 2.95 (-2.94,2.94) [0.00,3.50] Total: 41051 W: 6858 L: 6570 D: 27623 Elo +2.44 http://tests.stockfishchess.org/tests/view/5d7d14680ebc5902d3866196 This is the first and not really precise version, a lot of other stuff can be tried on top of it (separate complexity for middlegame, some more terms, even simple retuning of values). Bench: 4248476 |
| 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: Sat Sep 14 08:33:00 2019 +0200 Timestamp: 1568442780 Assorted trivial cleanups No functional change |
| 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: Sat Sep 14 07:47:05 2019 +0200 Timestamp: 1568440025 Use queens of either color in RookOnQueenFile The recently-added RookOnQueenFile evaluation term (36e4a86) provided a bonus for placing our rook on the same file as an enemy queen. Here, we relax a condition in this bonus, broadening its effect to any queen. It is also strategically desirable to place the rook on the same file as a friendly queen, so the restriction on the queen's color is removed. STC: LLR: 2.95 (-2.94,2.94) [-3.00,1.00] Total: 66856 W: 14847 L: 14815 D: 37194 Elo +0.17 http://tests.stockfishchess.org/tests/view/5d7b3c6a0ebc5902d385bcf5 LTC: LLR: 2.95 (-2.94,2.94) [-3.00,1.00] Total: 86786 W: 14264 L: 14248 D: 58274 Elo +0.06 http://tests.stockfishchess.org/tests/view/5d7b4e9b0ebc5902d385c178 Closes https://github.com/official-stockfish/Stockfish/pull/2302 Bench: 3703909 |
| 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: Sat Sep 14 07:34:19 2019 +0200 Timestamp: 1568439259 Update Makefile documentation Follow-up to previous commit. Update the documentation for the user when using `make`, to show the preferred bmi2 compile in the advanced examples section. Note: I made a mistake in the previous commit comment, the documentation is shown when using `make` or `make help`, not `make --help`. No functional change |
| 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 Sep 14 07:11:23 2019 +0200 Timestamp: 1568437883 Add sse4 if bmi2 is enabled The only change done to the Makefile to get a somewhat faster binary as discussed in #2291 is to add -msse4 to the compile options of the bmi2 build. Since all processors supporting bmi2 also support sse4 this can be done easily. It is a useful step to avoid sending around custom and poorly tested builds. The speedup isn't enough to pass [0,4] but it is roughly 1.15Elo and a LOS of 90%: LLR: -2.95 (-2.94,2.94) [0.00,4.00] Total: 93009 W: 20519 L: 20316 D: 52174 Elo +0.76 Also rewrite the documentation for the user when using `make --help`, so that the order of architectures for x86-64 has the more performant build one on top. Closes https://github.com/official-stockfish/Stockfish/pull/2300 No functional change |
| 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: Thu Sep 12 10:33:18 2019 +0200 Timestamp: 1568277198 Scale down complexity for almost unwinnable endgames This patch greatly scales down complexity of endgames when the following conditions are all true together: - pawns are all on one flank - stronger side king is not outflanking weaker side - no passed pawns are present This should improve stockfish evaluation of obvious draws 4 vs 3, 3 vs 2 and 2 vs 1 pawns in rook/queen/knight/bishop single flank endgames where strong side can not make progress. passed STC LLR: 2.94 (-2.94,2.94) [0.50,4.50] Total: 15843 W: 3601 L: 3359 D: 8883 Elo +5.31 passed LTC LLR: 2.96 (-2.94,2.94) [0.00,3.50] Total: 121275 W: 20107 L: 19597 D: 81571 Elo +1.46 Closes https://github.com/official-stockfish/Stockfish/pull/2298 Bench: 3954190 ========================== How to continue from there? a) This could be a powerful idea for refining some parts of the evaluation function, a bit like when we try quadratics or other equations to emphasize certain situations (xoto10). b) Some other combinaison values for this bonus can be done further, or overall retuning of weight and offset while keeping the formula simple. |
| 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: xoto10
Date: Thu Sep 12 10:05:35 2019 +0200 Timestamp: 1568275535 Bonus for rook on same file as their queen This patch creates a simple bonus for a rook that is on the same file as the opponent's queen. STC 10+0.1 th 1 : LLR: 2.95 (-2.94,2.94) [0.50,4.50] Total: 45609 W: 10120 L: 9733 D: 25756 Elo +2.95 http://tests.stockfishchess.org/tests/view/5d79895a0ebc5902d385484a LTC 60+0.6 th 1 : LLR: 2.96 (-2.94,2.94) [0.00,3.50] Total: 51651 W: 8606 L: 8288 D: 34757 Elo +2.14 http://tests.stockfishchess.org/tests/view/5d79a0850ebc5902d3854d27 Many thanks to @noobpwnftw for providing the extra cpu resources for fishtest, which led to me doing these tests. Closes https://github.com/official-stockfish/Stockfish/pull/2297 Bench: 4024461 |
| 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 Sep 12 09:44:40 2019 +0200 Timestamp: 1568274280 Simplify Weak Lever This is a simplification that integrated WeakLever into doubled pawns. Since we already check for !support for Doubled pawns, it is trivial to check for weak lever by just checking more_than_one(lever). We also introduce the Score * bool operation overload to remove some casts in the code. STC LLR: 2.95 (-2.94,2.94) [-3.00,1.00] Total: 26757 W: 5842 L: 5731 D: 15184 Elo +1.44 http://tests.stockfishchess.org/tests/view/5d77ee220ebc5902d384e5a4 Closes https://github.com/official-stockfish/Stockfish/pull/2295 No functional change |
| 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: Stefan Geschwentner
Date: Wed Sep 11 18:37:08 2019 +0200 Timestamp: 1568219828 Update reverse move stats For a good quiet non-pawn move consider the reverse move as bad and update the main history with a negative stat bonus. STC: LLR: 2.95 (-2.94,2.94) [0.50,4.50] Total: 19292 W: 4401 L: 4141 D: 10750 Elo +4.68 http://tests.stockfishchess.org/tests/view/5d7751d50ebc594e7864973c LTC: LLR: 2.95 (-2.94,2.94) [0.00,3.50] Total: 111952 W: 18762 L: 18275 D: 74915 Elo +1.51 http://tests.stockfishchess.org/tests/view/5d7771cf0ebc594e786498fa Closes https://github.com/official-stockfish/Stockfish/pull/2294 Bench: 3914238 |
| 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: Stefan Geschwentner
Date: Mon Aug 26 08:49:55 2019 +0200 Timestamp: 1566802195 Tweak Late Move Reduction at root Maintain best move counter at the root and allow there only moves which has a counter of zero for Late Move Reduction. For compensation only the first three moves are excluded from Late Move Reduction per default instead the first four moves. What we can further do: - here we use a simple counting scheme but perhaps some aging to fade out early iterations could be helpful - use the best move counter also at inner nodes for LMR and/or pruning STC: LLR: 2.95 (-2.94,2.94) [0.50,4.50] Total: 17414 W: 3984 L: 3733 D: 9697 Elo +5.01 http://tests.stockfishchess.org/tests/view/5d6234bb0ebc5939d09f2aa2 LTC: LLR: 2.96 (-2.94,2.94) [0.00,3.50] Total: 38058 W: 6448 L: 6166 D: 25444 Elo +2.57 http://tests.stockfishchess.org/tests/view/5d62681a0ebc5939d09f2f27 Closes https://github.com/official-stockfish/Stockfish/pull/2282 Bench: 3568210 |
| 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: VoyagerOne
Date: Sun Aug 25 11:09:08 2019 +0200 Timestamp: 1566724148 NMP Tweaks Tweak again the null move pruning preconditions. STC: LLR: 2.96 (-2.94,2.94) [0.50,4.50] Total: 19675 W: 4430 L: 4169 D: 11076 Elo +4.61 http://tests.stockfishchess.org/tests/view/5d52bc0e0ebc5925cf108300 LTC: LLR: 2.95 (-2.94,2.94) [0.00,3.50] Total: 73895 W: 12496 L: 12114 D: 49285 Elo +1.80 http://tests.stockfishchess.org/tests/view/5d52dcbc0ebc5925cf108552 Closes https://github.com/official-stockfish/Stockfish/pull/2268 Bench: 3690065 |
| 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 Aug 24 08:21:30 2019 +0200 Timestamp: 1566627690 Improve signature of evaluate_shelter() Remove one parameter in function evaluate_shelter(), making all comparisons for castled/uncastled shelter locally in do_king_safety(). Also introduce BlockedStorm penalty. Passed non-regression test at STC: LLR: 2.95 (-2.94,2.94) [-3.00,1.00] Total: 65864 W: 14630 L: 14596 D: 36638 Elo +0.18 http://tests.stockfishchess.org/tests/view/5d5fc80c0ebc5939d09f0acc No functional change |
| 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: Fri Aug 23 16:29:29 2019 +0200 Timestamp: 1566570569 Consolidate CastlingSide and CastlingRights This is a non-functional simplification that removes CastlingSide and implements the functionality in CastlingRights (thanks to Jörg Oster for a comment on the first version of this patch). STC LLR: 2.96 (-2.94,2.94) [-3.00,1.00] Total: 53854 W: 12077 L: 12019 D: 29758 Elo +0.37 http://tests.stockfishchess.org/tests/view/5d517b940ebc5925cf107474 Closes https://github.com/official-stockfish/Stockfish/pull/2265 No functional change |
| 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: Fri Aug 23 09:37:12 2019 +0200 Timestamp: 1566545832 Simplify futility equation This is a functional simplification. The 178 constant for the futility equation in master can be removed. STC LLR: 2.95 (-2.94,2.94) [-3.00,1.00] Total: 42626 W: 9508 L: 9428 D: 23690 Elo +0.65 http://tests.stockfishchess.org/tests/view/5d5d4e320ebc5925cf11254e LTC LLR: 2.95 (-2.94,2.94) [-3.00,1.00] Total: 26182 W: 4432 L: 4320 D: 17430 Elo +1.49 http://tests.stockfishchess.org/tests/view/5d5df70d0ebc5925cf112fee Closes https://github.com/official-stockfish/Stockfish/pull/2278 Bench: 3985701 |
| 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: Wed Aug 21 10:47:40 2019 +0200 Timestamp: 1566377260 Late move reduction, captures and CUT nodes Expand of Stefan Geschwentner's original idea: we always do LMR for captures at cutnodes. Passed STC http://tests.stockfishchess.org/tests/view/5d5b2f8e0ebc5925cf1111b8 LLR: 2.96 (-2.94,2.94) [0.50,4.50] Total: 36026 W: 8122 L: 7779 D: 20125 Elo +3.31 Passed LTC http://tests.stockfishchess.org/tests/view/5d5b40c80ebc5925cf111353 LLR: 3.22 (-2.94,2.94) [0.00,3.50] Total: 133502 W: 22508 L: 21943 D: 89051 Elo +1.47 Closes https://github.com/official-stockfish/Stockfish/pull/2273 Bench: 3494372 |
| 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 Aug 21 09:34:46 2019 +0200 Timestamp: 1566372886 Tuned Futility Equation @Vizvezdenec array suggested that alternate values may be better than current master (see pull request #2270 ). I tuned some linear equations to more closely represent his values and it passed. These futility values seem quite sensitive, so perhaps additional Elo improvements can be found here. STC LLR: 2.95 (-2.94,2.94) [0.50,4.50] Total: 12257 W: 2820 L: 2595 D: 6842 Elo +6.38 http://tests.stockfishchess.org/tests/view/5d5b2f360ebc5925cf1111ac LTC LLR: 2.96 (-2.94,2.94) [0.00,3.50] Total: 20273 W: 3497 L: 3264 D: 13512 Elo +3.99 http://tests.stockfishchess.org/tests/view/5d5c0d250ebc5925cf111ac3 Closes https://github.com/official-stockfish/Stockfish/pull/2272 ------------------------------------------ How to continue from there ? a) we can try a simpler version for the futility margin, this would be a simplification : margin = 188 * (depth - improving) b) on the other direction, we can try a complexification by trying again to gain Elo with an complete array of futility values. ------------------------------------------ Bench: 4330402 |
| 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: Jean Gauthier
Date: Wed Aug 21 09:11:17 2019 +0200 Timestamp: 1566371477 Slight speep up fetching the endgame table Replace calls to count(key) + operator[key] with a single call to find(key). Replace the std::map with std::unordered_map which provide O(1) access, although the map has a really small number of objects. Test with [0..4] failed yellow: TC 10+0.1 SPRT elo0: 0.00 alpha: 0.05 elo1: 4.00 beta: 0.05 LLR -2.96 [-2.94,2.94] (rejected) Elo 1.01 [-0.87,3.08] (95%) LOS 85.3% Games 71860 [w:22.3%, l:22.2%, d:55.5%] http://tests.stockfishchess.org/tests/view/5d5432210ebc5925cf109d61 Closes https://github.com/official-stockfish/Stockfish/pull/2269 No functional change |
| 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: Alain SAVARD
Date: Wed Aug 14 22:15:48 2019 +0200 Timestamp: 1565813748 Assorted trivial cleanups (July 2019) No functional change |
| 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: Stefan Geschwentner
Date: Wed Aug 14 20:59:04 2019 +0200 Timestamp: 1565809144 Tweak unsafe checks Remove mobility area for unsafe checks. Also separate the evaluation terms for unsafe checks and blockers for king with adjusted weights. STC: LLR: 2.95 (-2.94,2.94) [0.50,4.50] Total: 124526 W: 28292 L: 27504 D: 68730 Elo +2.20 http://tests.stockfishchess.org/tests/view/5d5138290ebc5925cf1070c3 LTC: LLR: 2.96 (-2.94,2.94) [0.00,3.50] Total: 84968 W: 14499 L: 14083 D: 56386 Elo +1.70 http://tests.stockfishchess.org/tests/view/5d527cfa0ebc5925cf107f93 Bench: 4139590 |
| 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: joergoster
Date: Thu Aug 1 09:28:34 2019 +0200 Timestamp: 1564644514 Revert "Improve multiPV mode" This reverts commit a8de07cc26999e2fef7298a63bfe349aaa4650fa. |
| 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: xoto10
Date: Wed Jul 31 09:35:20 2019 +0200 Timestamp: 1564558520 Tune search constants This is the result of a 200k tuning run at LTC: http://tests.stockfishchess.org/tests/view/5d3576b70ebc5925cf0e9e1e which passed quickly at LTC: LLR: 2.95 (-2.94,2.94) [0.50,4.50] Total: 12954 W: 2280 L: 2074 D: 8600 Elo +5.53 http://tests.stockfishchess.org/tests/view/5d3ff3f70ebc5925cf0f87a2 STC failed, but second LTC at [0,4] passed easily: LLR: 2.96 (-2.94,2.94) [0.00,4.00] Total: 8004 W: 1432 L: 1252 D: 5320 Elo +7.81 http://tests.stockfishchess.org/tests/view/5d407cff0ebc5925cf0f9119 Further work? No doubt some of these changes produce most of the gain and some are neutral or even bad, so further testing on individual/groups of parameters changed here might show more gains. It does look like these tests might need to be at LTC though, so maybe not too practical to do. See the thread in the pull request for an interesting discussion: https://github.com/official-stockfish/Stockfish/pull/2260 Bench: 4024328 |