> > Why should this value be static? Shouldn't the behavior change when there > is a certain win? >
I think it should. This is what I do. When I have a win rate above 80% I bias nodes by their average territory (counting prisoners). This might be the same as adding a small bonus to each large win, but I haven't thought about it. I also bias moves that will help clean up dead opponents stones, ie touch, atari, capture. And still in cases where win rate > 80%, I check whether the win rate and average territory of pass is close to those of the best move returned by the search. If they are, I just pass. I haven't spent a lot of time on this, so it could definitely be improved. But it does OK. I imagine some other engines are solving this problem more rigorously with a deterministic approach to deciding whether there are any moves left to increase the score before passing, but I haven't implemented anything like this yet.
_______________________________________________ Computer-go mailing list [email protected] http://dvandva.org/cgi-bin/mailman/listinfo/computer-go
