Re: [computer-go] Re: Scalability

[Don Dailey <drd@xxxxxxx>]

   Date: Tue, 18 Jan 2005 16:11:09 -0500
   From: Chris Fant <chrisfant@xxxxxxxxxxxxxxxxx>

   > So maybe a good guide to what a quies search should include is any
   > move that helps resolve positions that your evaluation doesn't
   > understand on it's own.  This would encompass "responses to disturbing
   > moves" and "refutations of blunders."

   But what kind of blunder would not already be accounted for in the
   local tactics search during the full board evaluation of the leaf?

Don't think of it that way.  I'm trying to abstract the notion of
quiescence.  It's not REALLY about blunders or disturbing moves or
anything like this.  It's more along the lines of playing around moves
(via the q-search) that the evaluation doesn't understand.   


So if you define a blunder as any bad move that is immediately refuted
by local tactics in the global evaluation, then the answer is that
your quiesence doesn't need any help for refuting blunders but that's
just a way of thinking about it.  In this case it would fall back to
what Magnus Persson said, "responses to disturbing moves" and I would
add that a disturbing move is defined as a move that your evaluation
can't directly resolve without a bit of help.

It's not quite as simple as just going one more ply since Magnus
provides very specific responses to the "distrubing" moves.  In chess
those specific responses are mostly captures.  Actually good chess
programs don't look at all captures, they focus on the most "relevant"
ones like Magnus describes in Go.


- Don




_______________________________________________
computer-go mailing list
computer-go@xxxxxxxxxxxxxxxxx
http://www.computer-go.org/mailman/listinfo/computer-go/