RE: [computer-go] Computer Go hardware

["David Fotland" <fotland@xxxxxxxxxxxxxxx>]

Pros know all of the tactical move selection patterns that don't
exist in our input data.

Chess pros can also play extremely well without any apparent
look ahead.  Chess is a simpler game, and there are much larger
databases of games available.  If this approach works for go,
it should be even easier for chess, but I don't hear any chess
programmers suggesting that learning patterns has any hope of 
making a strong chess program :)

David

> -----Original Message-----
> From: computer-go-bounces@xxxxxxxxxxxxxxxxx 
> [mailto:computer-go-bounces@xxxxxxxxxxxxxxxxx] On Behalf Of Darren Cook
> Sent: Tuesday, October 19, 2004 9:47 PM
> To: computer-go
> Subject: Re: [computer-go] Computer Go hardware
> 
> 
> > I don't think it is possible to make a strong program by extracting 
> > millions of patterns from professional games and sorting 
> through them 
> > automatically. Go is too tactical, and the branching factor is too 
> > high to avoid heuristics in move sorting.
> 
> I always come back to that the pros seem to do it ("be a 
> strong program") 
> when they play lightening go (*). Of course they are not using just 
> black/white stones, as they've assigned meaning to chains as 
> the game has 
> progressed, but still I think search is an enhancement not 
> the fundamental 
> strength of a strong player.
> 
> Darren
> 
> *: I recently watched the last 30 or so moves of a pro 9x9 
> game played in 
> about 30 seconds total: both players were in serious time 
> trouble. That 
> included both a semeai and a ko, yet all the moves seemed 
> correct and some 
> were non-obvious (to my eyes). 
> _______________________________________________
> 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/