On Sat, 2008-12-20 at 16:04 +0100, Rémi Coulom wrote: > Still, some programmers tried null move in 9x9 alpha-beta go > programs. > As far as I remember, they got no strength improvement from it.
I tried null move once in my very weak alpha/beta go program. The issue with recursive deep search null move is that it is not that good and discovering long term threats, which is way more common in GO than in chess. However, it might work well for discovering certain classes of immediate tactical threats that must be responded to. As a general pruning mechanism, I fear it is not very workable - but I never say never. - Don _______________________________________________ computer-go mailing list [email protected] http://www.computer-go.org/mailman/listinfo/computer-go/
