RE: [computer-go] Learning : was Chess programs versus go programs

["David Fotland" <fotland@xxxxxxxxxxxxxxx>]

Mark,

I tried limiting tactical search to 3 liberties instead of 4 on a set of
life and death problems.

On about 1000 problems, it solved 9 fewer, and spent about half the time
getting the solutions.  I expect some evaluation tuning will easily get
those 9 solutions back.  

So I think you are correct that using 4 liberties for tactical stability is
better than using 5 liberties once you have good life and death evaluation.
The next version of Many Faces will have this change.

David

> -----Original Message-----
> From: computer-go-bounces@xxxxxxxxxxxxxxxxx 
> [mailto:computer-go-bounces@xxxxxxxxxxxxxxxxx] On Behalf Of 
> David Fotland
> Sent: Saturday, December 11, 2004 10:25 AM
> To: tesujisoftware@xxxxxxxxxxxxxxxxx; 'computer-go'
> Subject: RE: [computer-go] Learning : was Chess programs 
> versus go programs
> 
> 
> I think that going to 4 liberties instead of 3 doesn't cost 
> much, since it is easy to prove that most 4 liberty groups 
> must get 5 liberties soon.  I'll try the experiment though, 
> and see what the speed/strength tradeoff is for defining 
> escape as 4 liberties instead of 5.
> 
> You make a very good point about wasted search.  I started 
> with tactics and added life and death later, so perhaps some 
> of my tactical search is now redundant.
> 
> David


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