RE: [computer-go] Scalability Odd-Even Effect

["David Fotland" <fotland@xxxxxxxxxxxxxxx>]

I wish it were narrow :(  I look at local forced responses (from patterns),
like connecting peeps.  I look at reversing the effect of the move that
leaves a group unsettled.  For these reversing moves, I generate all
reasonable moves, and stop searching when I find one that works.  I don't
have a fixed depth for Q search, so sometimes it's over 10 ply.

Version 11 MFGO does a 1 "ply" search, followed by quiescence, where the
first "ply" can be several moves deep if a pattern suggests a move sequence
rather than a single move.  For example for joseki, it will play out an
entire joseki line, then do Q search.

David

> -----Original Message-----
> From: computer-go-bounces@xxxxxxxxxxxxxxxxx 
> [mailto:computer-go-bounces@xxxxxxxxxxxxxxxxx] On Behalf Of Chris Fant
> Sent: Tuesday, January 18, 2005 9:21 AM
> To: computer-go
> Subject: Re: [computer-go] Scalability Odd-Even Effect
> 
> 
> > Quiescence search is really important in go.  many Faces 
> spends almost 
> > all of its full board search in quiescence.
> 
> I have given very little thought to a quiecence search in my program. 
> Mostly because I am not sure how to approach it.  Do you just 
> do a really narrow search (or just one line) with the goal of 
> stabalizing groups instead of the normal Go goals? 
> _______________________________________________
> computer-go mailing list
> computer-go@xxxxxxxxxxxxxxxxx 
> http://www.computer-go.org/mailman/listinfo/computer-go/
> 


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