#chess #statistics #simulation #cheating #visualization

https://dorianquelle.github.io/blog/How-To-Catch-Smart-Cheaters/

#chess #statistics #simulation #cheating #visualization #stockfish

Красивые визуализации шахмат от Дориана Квелле. Интересная метрика average loss after novelty, видно, что он сам придумал, классная идея.

"Yet, the standout performer is David Howell, with an average score of 0.9071. Despite his tendency to withdraw or join late to tournaments, this strategy seems to benefit him as he avoids facing high-performing players in later rounds. With a record of 120 wins, 14 draws, and only 6 losses, Howell clearly dominates the field."

"Yet, the left plot exposes a glaring inconsistency in Kramnik’s performance. Despite his high accuracy in playing engine-recommended moves, he registers a higher average loss per move than one would expect for a player of his Elo rating. This suggests that while Kramnik plays the engine move in most cases, he is also prone to significant blunders. Playing the fools mate because you’re suspecting a player of cheating doesn’t help either."

https://dorianquelle.github.io/blog/Cheating-In-Titled-Tuesday/

#simulation #optimization

Интересная идея о том, как ускорять симуляции с помощью ML и тем самым помогать быстрой оптимизации. Действительно, почему бы и и нет! Сначала собираем входные параметры системы +итоги симуляций, а дальше симуляцию пробуем заменять быстрой ML моделью.

https://www.youtube.com/watch?v=ohGeGfUCV_A