Chess Engines Diary

  Stockfish 16 - official release version *Quality of chess play* Stockfish continues to demonstrate its ability to discover superior moves  with remarkable speed. In self-play against Stockfish 15, this new  release gains up to 50 Elo and wins up to 12 times more game pairs  than it loses. In major chess engine tournaments, Stockfish reliably tops  the rankings winning the TCEC season 24 Superfinal, Swiss, Fischer  Random, and Double Random Chess tournaments and the CCC 19 Bullet,  20 Blitz, and 20 Rapid competitions. Leela Chess Zero was the  challenger in most finals, putting top-engine chess now firmly in the  hands of teams embracing free and open-source software. *Progress made* This updated version of Stockfish introduces several enhancements,  including an upgraded neural net architecture (SFNNv6), improved  implementation, and refined parameterization. The ongoing utilization  of Leela’s data combined with a novel infe...

  💾   272 games from the tournament download   @chessenginesdiary 🕓 Time 3'+3" 💻HP ENVY i5-8250U 8GB RAM 🖬 GUI-Fritz 15 💾 All CEDR games  141.461 games download (3'+3") P Engine Score +/-/= 1: Swordfish 15.5-bmi2 24.5 / 32 (+17 -0 =15) 2: Stockfish, dev-20230620-6eaa1c 24.0 / 32 (+16 -0 =16) 3: SFT-Pro 24.0 / 32 (+16 -0 =16) 4: CatroPOLY V2.7GCC 20230617 24.0 / 32 (+16 -0 =16) 5: Stockfish, dev-20230619-a46087 23.5 / 32 (+15 -0 =17) 6: SpecTral GNX 2.40_b 23.5 / 32 (+15 -0 =17) 7: Polyfish 230621 23.0 / 32 (+14 -0 =18) 8: Clover 5.0 20.0 / 32 (+14 -6 =12) 9: Igel 3.5.0 64 BMI2 AVX2 19.5 / 32 (+12 -5 =15) 10: Superultr...

Rice - UCI chess engine Author - Slender  Rating CEDR=3180 v.7.0: New search features Better time control New net: 768 inputs, 4 buckets, 768 hidden layer. UCI stop support Elo difference vs 6.0.0: ~150 elo STC 6+0.06 Binaries: rice-7.0.0-v2 for POPCNT, SSE3, SSE4.1, SSE4.2, SSSE3 CPUs. rice-7.0.0-v3 for AVX, AVX2, BMI1, BMI2 CPUs. v2 has ~950K NPS, while v3 has ~1.5M nps. Using the v3 binary is suggested. Individual statistics:  Rice 5.0.0 - 190 games (+ 85,= 36,- 69), 54.2 % Critter 1.6a                  :   6 (+  2,=  4,-  0), 66.7 % Deep iCE 4.0 853              :   1 (+  0,=  0,-  1),  0.0 % Godel 7.0                     :   1 (+  0,=  1,-  0), 50.0 % Novice 3.1                    :   1 (+  1,=...

💾   180 games from the tournament download   @chessenginesdiary 🕓 Time 3'+3" 💻Lenovo Ideapad 330, i3-8130U, 12GB RAM 🖬 GUI-Arena 💾 All CEDR games  141.461 games download (3'+3") P Engine CEDR Score Perf Chg +/-/= 1: Little Goliath 3.17 3243 31.5 / 36 3315 +18 (+28 -1 =7) 2: Nalwald 17 3065 28.5 / 36 3140 +29 (+23 -2 =11) 3: Critter 1.6a 3200 27.0 / 36 3151 -22 (+22 -4 =10) 4: Texel 1.08 3188 27.0 / 36 3145 -18 (+22 -4 =10) 5: 4ku 3.1 2876 15.5 / 36 2807 -36 (+14 -19 =3) 6: Peacekeeper 1.50 2828 14.0 / 36 2757 -32 (+12 -20 =4) 7: Ice4 3.1 2400 13.5 / 36 2571 +72 (+10 -19 =7) 8: Pedantic 0.3 2400 11.5 / 36 2525 +50 (+10 -23 =3) 9: Myrddin 0.90 2629 10.5 / 36 2599 -11 (+7 -22 =7) 10: Cadabr...

Swordfish has a reasonable balance between strength and tactical abilities. It still has solid style of playing and achieves good results in test suites. It can be useful for analyzing sharp positions where regular Stockfish can overlook some tactics. Rating CEDR=3430 Swordfish 15.5 for Android download Our social media: Twitter        Facebook    Instagram    Pinterest    Reddit  Tumblr

BrainLearn - UCI chess engine Rating CEDR=3497 BrainLearn is a free, powerful UCI chess engine derived from BrainFish. It is not a complete chess program and requires a UCI-compatible GUI (e.g. XBoard with PolyGlot, Scid, Cute Chess, eboard, Arena, Sigma Chess, Shredder, Chess Partner or Fritz) in order to be used comfortably.  Read the documentation for your GUI of choice for information about how to use Stockfish with it.  Android builds were generated with CECSA (Chess Engine Compiler Script for Android). Chess engines are located in the libs directory and are available for arm64-v8a, armeabi-v7a, x86 and x86_64 based devices. If a chess engine supports NNUE (Neural Network Updated Efficiently) technology and the network is not embedded in the binary, you will find the missing network file in the networks directory. BrainLearn for Android 25.1 - download

  💾   144 games from the tournament download   @chessenginesdiary 🕓 Time 3'+3" 💻Dell G15 SE, i7-11800H RAM 16,0 GB 🖬 GUI-Banksia 💾 All CEDR games  141.461 games download (3'+3") Tech table: Engine KN/move NPS dep/mov time/mov mov/game time/game fails Brainlearn 25.1 18407 3869882 52.2 4.8 83.7 397.9 Cadie 1.5 5590 1010224 24.9 5.5 62.9 348.1 Caissa 1.10 29139 5771355 32.2 5.0 76.0 383.7 Critter 1.6a 43000 8299482 24.3 5.2 72.8 377.1 Fat Fritz 2 25711 5356569 45.1 4.8 77.0 369.6 Komodo 14.1 30677 6434355 31.2 4.8 79.3 378.1 Polaris 1.8.1 30230 6294950 41.8 4.8 69.7 334.8 2 SFT-Pro 16880 3594534 49.7 4.7 87.4 410.5 ShashChess 33.1 18457 3617522 44.7 5.1 70.2 358.1

Chess engine: Seawall , Rating CEDR=2736   Author - Patrick Heck This is an [UCI] chess engine. It started as an attempt to build a minimal program that was able to play chess. Over  time, it has gained features to make it stronger, but it still implements only a subset of UCI necessary to play  games, and there's almost no evaluation. Some ideas have been borrowed more directly than others; the obstruction difference algorithm is based on the version  from [Chess Movegen], while the use of a low pass filter and capture history for move ordering were borrowed from  [Stockfish]. Individual statistics:  Seawall 20230316 - 126 games (+ 40,=  6,- 80), 34.1   % Critter 1.6a                  :   8 (+  0,=  1,-  7),  6.2 % Tomcat 1.0                    :   1 (+  0,=  0,-  1),  0.0 % Komodo 13.02...

💾   48 games from the tournament download   @chessenginesdiary 🕓 Time 3'+3" 💻Dell G15 SE, i7-11800H RAM 16,0 GB 🖬 GUI-Fritz 17 💾 All CEDR games  141.461 games download (3'+3")  Stockfish: regression/progression test (Examiner: Stockfish 14.1) P Stockfish Score +/-/= 1: Stockfish, dev-20230531-d99942 16.5 / 32 (+1 -0 =31) 2: Stockfish, dev-20230604-06186b 16.0 / 32 (+0 -0 =32) 3: Stockfish 14.1 15.5 / 32 (+0 -1 =31)

  StockDr o y, the mark of strength and efficiency, emerges as a proficient chess engine reengineered in C++ from its popular and mighty C# predecessor, StockNemo. Striving to ascend the peaks of excellence, StockDory serves as a testament to the spirit of continuous improvement, eliminating bugs and performance issues that blemished its otherwise remarkable predecessor. Download the latest release and try it out for yourself! For maximum performance, compile the engine specifically for your hardware using the instructions below. 🛠️ Compiling StockDory is written in C++ and uses CMake as its build system. Requirements: 🏭 CMake >= 3.15 🐉 Clang (LLVM) >= 16.0.0 🥷 Ninja >= 1.10.2 Steps: 💾 Clone the repository git clone https://github.com/TheBlackPlague/StockDory.git 🔦 Setup Build Process cd StockDory cmake -B Build -DCMAKE_BUILD_TYPE=Release -DCMAKE_C_COMPILER=clang -DCMAKE_CXX_COMPILER=clang++ -G Ninja 🪛 Compile cmake --build Build --config Release 🏃‍♂️ Run ./Build/S...

Cadie - UCI Engine Author: Martin Wyngaarden Cadie is a stand-alone UCI chess engine written in platform independent C++. It is also the successor to my first UCI chess engine, Keele. Building requires a compiler supporting at least C++20. Being a stand-alone engine, Cadie requires a UCI compatible GUI to run. Compatible GUIs include Arena, ChessBase, and SCID, to name only a few. Cadie 1.5 download

Caissa - UCI chess engine,  Author: Michał Witanowski Rating CEDR=3510 v.1.10: This release introduces new neural network architecture. The accumulator is split into whites and blacks perspective. Trained on 1.9B positions. The biggest gains are visible in DFRC. Accumulator size is now 1024 neurons, so the net file is even smaller than in previous release (1536). Also includes a small fix in search (not updating best move if searched move doesn't beat alpha), worth around 15 Elo. Elo gain Book=UHO_XXL_2022_+100_+129.epd, TC=40+0.04 Score of Caissa 1.10 AVX2 vs Caissa 1.9 (AVX2/BMI2): 173 - 102 - 391  [0.553] 666 Elo difference: 37.2 +/- 16.9, LOS: 100.0 %, DrawRatio: 58.7 % Book=UHO_MEGA_2022_+110_+149.pgn, TC=8+0.08 Score of Caissa 1.10 AVX2 vs Caissa 1.9 (AVX2/BMI2): 267 - 157 - 390  [0.568] 814 Elo difference: 47.2 +/- 17.2, LOS: 100.0 %, DrawRatio: 47.9 % Book=chess960_3moves.pgn, TC=8+0.08 Score of Caissa 1.10 AVX2 vs Caissa 1.9 (AVX2/BMI2): 76 - 33 - 158 ...