ShashChess Versions Save

ShashChess 35 Aligned with Stockfish 16.1 Improved opening variety Improved Shashin theory Eliminated classical eval and handicap mode: there is another engine called Alexander for this.

ShashChess 34.6 Improved learning

Fix mcts UCI_ShowWDL option default to true A great thanks for their help to -Omar Khalid in correcting -Peter Martan in professionally testing the mcts code.

Fix avatart playing: now the winnable is for both mg and eg Fix mcts (eliminated GoldDigger and fixed a bug) UCI_ShowWDL to true by default

Fix native build on linux Eliminated Stockfish handicap mode and replaced with another better one, based on an idea of Tomasz Sobczyk and Michael Byrne Thanks to Tomasz Sobczyk https://github.com/official-stockfish/Stockfish/issues/3635 Michael Byrne https://github.com/MichaelB7/Stockfish/commit/18480caa5782ede902b2734798ee6f51f06a0d45

Fixed handicap mode and mcts Fixed native build (thanks to Marco Zerbinati)

Fix MCTS Added clang support and native build (see profilebuild.bat)

-Great improvement/fix to Shashin's theory -Bug fix for R-learning -Improved MCTS with the introduction of the new option MCTSGoldDigger -Improved handicap mode with the introduction of the avatar concept: possibility to play with the engine that not only, like Stockfish, makes mistakes with a frequency corresponding to the set elo, but simulates precisely the elo and thinking system of a player. To this end, the classic rating function (used only for handicap mode) has been preserved. As several research papers such as this one [url]https://www.ijcai.org/proceedings/2023/0541[/url].pdfalso show, the neural network is a blackbox that has revolutionized traditional strategic concepts in chess books. It is important, then, to try to interpret its results. To this end, comparison with the classical evaluation function is an excellent, at present, essential interpretive model. Finally, the avatar file, which can be generated with a private application, allows you to simulate exactly the playing style of a given player.(See the Readme for more information) -alignment with Stockfish 16 (latest official release).

Fixed a bug in learning: no more entries with depth=0 Old experience files can be reduced by this tool from Ibai Burgos. A great thanks to him! The engine is always aligned with official Stockfish release. From now on, I'll be always only aligned with Stockfish official release.

ShashChess 33.1 Fix learning. Now, the experience file has a .exp extension and not .bin as before. Improved Shashin theory, especially for mcts. So, simply, rename the file. Stockfish patch Update default net to nn-5af11540bbfe.nnue