Chess Engines Diary

v.1.0.5 Installation instructions The default previous ClassicAra model is included within each release package.  Moreover, the binary packages include the required inference libraries for each platform. The newer ClassicAra models can be downloaded in release 1.0.4.  You may choose alpha_vil_fx_models.zip and select a model size depending on your GPU/CPU and time-control.  At a very low time control (e.g. 30ms/Move), it is recommended to reduce the Batch-Size to 16. The models for CrazyAra and MultiAra the models should be downloaded separately and unzipped (see release 0.9.5). CrazyAra-rl-model-os-96.zip MultiAra-rl-models.zip (improved MultiAra models using reinforcement learning (rl) ) MultiAra-sl-models.zip (initial MultiAra models using supervised learning) For XiangqiAra you can download XiangqiAra-sl-model.zip (see release 0.9.9). Next, move the model files into the model/<engine-name>/<variant> folder. Stratego is only included in the Linux release fi...

v.1.0.5 Installation instructions The default previous ClassicAra model is included within each release package.  Moreover, the binary packages include the required inference libraries for each platform. The newer ClassicAra models can be downloaded in release 1.0.4.  You may choose alpha_vil_fx_models.zip and select a model size depending on your GPU/CPU and time-control.  At a very low time control (e.g. 30ms/Move), it is recommended to reduce the Batch-Size to 16. The models for CrazyAra and MultiAra the models should be downloaded separately and unzipped (see release 0.9.5). CrazyAra-rl-model-os-96.zip MultiAra-rl-models.zip (improved MultiAra models using reinforcement learning (rl) ) MultiAra-sl-models.zip (initial MultiAra models using supervised learning) For XiangqiAra you can download XiangqiAra-sl-model.zip (see release 0.9.9). Next, move the model files into the model/<engine-name>/<variant> folder. Stratego is only included in the Linux release fi...

  Combination: game ClassicAra 0.9.9 - Fritz 17 (0-1) Partia mi si� podoba�a (Fritz 17) [Event "MiniMatch.FritzAra, Szybkie 15.0min+10."] [Site "DESKTOP-AU8L67R"] [Date "2022.02.16"] [Round "9"] [White "ClassicAra 0.9.9 (Feb 14 2022"] [Black "Fritz 17"] [Result "0-1"] [ECO "D32"] [Annotator "0.28;0.28"] [PlyCount "40"] [TimeControl "900+10"] {11th Gen Intel(R) Core(TM) i7-11800H @ 2.30GHz 2304 MHz W=23.8 plies; 1kN/s B=23.2 plies; 9 469kN/s; 361 TBAs} 1. d4 {[%eval 28,21] [%emt 0:00:33] Obie Ostatni ruch z książki} d5 {[%eval 28,24] [%emt 0:00:20]} 2. c4 {[%eval 43, 24] [%emt 0:00:32]} e6 {[%eval 14,24] [%emt 0:00:19] (dxc4)} 3. Nf3 {[%eval 39, 20] [%emt 0:00:32] (Sc3)} c5 {[%eval 21,25] [%emt 0:00:39] (Sf6)} 4. cxd5 { [%eval 61,22] [%emt 0:00:33] (Sc3)} exd5 {[%eval 19,27] [%emt 0:01:01]} 5. Nc3 {[%eval 80,21] [%emt 0:00:33] (g3)} Nc6 {[%eval 41,26] [%emt 0:01:19]} 6....

P Engine Score ***** *** +/-/= 1: Fritz 17 6.5 / 10 =11===1=== (+3 -0 =7) 2: ClassicAra 0.9.9 3.5 / 10 =00===0=== (+0 -3 =7) Time 15'+10" Dell G15 SE, i7-11800H RAM 16,0 GB GUI Fritz 17 10 games P-15'+10" games CEDR 15.911 games Table: Scid vs PC

ClassicAra v.0.9.9 Features First experimental XiangqiAra release. Move generation back-end and Xiangqi ruleset is based on Fairy-Stockfish. Uses supervised neural network trained on 10k human Xiangqi games. Please refer to the thesis Evaluation of Monte-Carlo Tree Search for Xiangqi by Maximilian Langer, pdf for more information. UCI_Chess_960 support as introduced in https://github.com/QueensGambit/CrazyAra/releases/tag/0.9.8. (However, no official 960 network yet.) TensorRT API Update #164 Major bug fixes Handle flooding of UCI-commands (#167) CrazyAra Going To Infinite Analysis Mode On 1 Position (Can't Be Stopped) In Liground After Making Two Moves On The Board (#81) Avoid repeating positions in Xiangqi (closes #101) #166 TCEC This version has been submitted to the TCEC Season 22. ClassicAra 0.9.9 uses the wdlp-rise3.3-input3.0 model which was trained on the Kingbase2019lite data set as for release 0.9.5. The engine.json configuration file and update.sh shell script can be use...

ClassicAra This version has been submitted to the TCEC Swiss 2 event. ClassicAra 0.9.7 has a higher GPU and CPU utilization thanks to a higher batch size and more threads (#160). ClassicAra 0.9.7 - more and download

MultiAra This is the first release which features MultiAra, a version of Ara which supports all chess variants available on lichess.org. antichess atomic chess960 crazyhouse king-of-the-hill horde racing kings three-check The neural network models have been initialised by lichess.org game data from August 2013 until July 2020. All variant (except chess960) have been improved using reinforcement learning. Details can be found in the master thesis by Maximilian Alexander Gehrke which among others covers an ablation study between training from zero knowledge and supervised initialisation, as well as a preliminary strength comparison against Fairy-Stockfish. ClassicAra The model for ClassicAra uses a new input representation and a new WDLP value output head. The old model and input representation for chess is still supported as the input representation version is now inferred by the model name. CrazyAra&ClassicAra 0.9.5 - more and download

Promotion to Group E5: ClassicAra 0.9.0, Embla 2.0.7, Loki 2.0.0 . End of testing: Chessika 2.27 . Temp 3'+3" Intel Core i7 7,9 GB GUI  Arena 3.5.1 Book  Perfect 2021 All JCER games 94 .482 Table created  Scid vs PC JCER League Tournament:   612 games Amateur and new engines E1 Edition 45 Rating JCER Score 1 Cheng4 0.40 2788 47.5 / 60 promotion 2 MrBob 1.0.0 2753 42.0 / 60 promotion 3 Crafty 25.6 2711 41.0 / 60 promotion 4 NiCim 3.4.1 2799 38.5 / 60 5 BulldogVB 2.1 2795 38.0 / 60 6 Atlas 3.91 2728 36.0 / 60 7 Delocto 0.61n 2681 31.0 / 60 8 ProDeo 3.1 2656 29.5 / 60 9 Arminius 2018-12-23 2674 29.5 / 60 10 Olithink 5.9.2 2652 28.0 / 60 11 ...