Re: computer-go: Hashing during minimax

[Peter Drake <drake@xxxxxxxxxx>]

Thanks for all the replies and research links.  Looks like I have some 
reading to do.

In response to your questions, yes, my static evaluation function is 
ridiculously simple at the moment.  I did improve it slightly so that 
it propagates influence across the board (an empty point belongs to the 
closest stone), and this makes it play much more reasonably.  The 
current move suggestor (try every legal move) is also, of course, 
absurd.  Since our project will focus on machine learning, I don't want 
to spend too much time hand-constructing good functions.

In other news, my GTP interface continues to work very well with 
several front ends (Goban, GoGui, and gGo).  Many thanks to the people 
who built these wonderful programs!

Peter Drake
Assistant Professor of Computer Science
Lewis & Clark College
http://www.lclark.edu/~drake/