computer-go: Random move generation

["Jeff Massung" <jmassung@xxxxxxxxxxx>]

Question for thought:

How good/bad is random move generation after move weighting is done?  If 
there are say 3-10 moves all with the same weight, of course a depth search 
could reveal 1 better than all, and that is ok.  But how about say at the 
beginning of the game when many moves (say corner moves) are essentially the 
same.  Should randomness be done to decide a move?

Wondering what other people have done here (aside from chess-like opening 
books)?

Jeff

______________________________________________________
Get Your Private, Free Email at http://www.hotmail.com