On Tue, Feb 17, 2009 at 8:35 PM, George Dahl <george.d...@gmail.com> wrote: > Really? You think that doing 20-50 uniform random playouts and > estimating the win probability, when used as a leaf node evaluator in > tree search, will outperform anything else that uses same amount of > time?
You'll probably find a variety of opinions on that. I think you can make a static evaluator that will give you a better estimate of the win probability estimate than 50 uniform random playouts. But... (there are a few big ones) implementing uniform random playout is a lot less work. On top of that, with some prudent exploration, you rarely spend 50 playouts all in one place. This is definitely something powerful in MCTS programs, that they can reject unpromising lines of play at relatively little cost. Mark _______________________________________________ computer-go mailing list computer-go@computer-go.org http://www.computer-go.org/mailman/listinfo/computer-go/
