ShashChess34

-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).