There's a pretty finite set of tokens that represent valid chess moves, and a more than countable set of possible opening moves, with a very distinct probabilit

There's a pretty finite set of tokens that represent valid chess moves, and a more than countable set of possible opening moves, with a very distinct probability of moves; Nobody opens pg3, for example. Classical Markov chains played chess at some rate of success. ChatGPT is probably a lot better but not fundamentally different - It's predicting which moves to play based on sets of past games, not by memorizing it but by memoizing it.