On 16-11-17 18:15, "Ingo Althöfer" wrote: > Something like MCTS would not work in chess, because in > contrast to Go (and Hex and Amazons and ...) Chess is > not a "game with forward direction".
Ingo, I think the reason Petr brought the whole thing up is that AlphaGo Zero uses "MCTS" but it does not actually use Monte Carlo Playouts. I think your condition about "forward direction" only applies to the randomized playouts, yes? A neural network evaluation on the other hand is very much like a classical static chess evaluation. There are publications about Parallel Randomized Best First Search in chess, read them and notice how it compares to MCTS. -- GCP _______________________________________________ Computer-go mailing list Computer-go@computer-go.org http://computer-go.org/mailman/listinfo/computer-go