Re: computer-go: Hashing during minimax

[Jim Levenick <levenick@xxxxxxxxxxxxxx>]

Hey,

Have you read the thesis by Burmeister?  I just printed it, and it 
looks pretty interesting (at a distance); I should read it this weekend 
(I'm travelling to CA and will have some down time).

On Wednesday, April 23, 2003, at 09:53 AM, Peter Drake wrote:

> What are people doing about this problem?  Limiting the size of the 
> table, and either ignoring additional boards or overwriting earlier 
> ones?  Skipping stored boards altogether?
I wonder if part of the problem is the enormous number of boards stored 
that are completely silly.  Like where the players are playing on the 
first line at the beginning of the game.  Obviously, we can't enumerate 
the game; as you pointed out long ago, the trick is to limit the number 
of moves that are actually considered.  Right?

> I have one other concern:  suppose I evaluate a board at a leaf node, 
> and later encounter it again higher in the tree.  Should I re-evaluate 
> it?
Um... not sure.  Tell me more.