Re: [computer-go] SlugGo v.s. Many Faces update

[steve uurtamo <apoxonpoo@xxxxxxxxx>]

> A random
> player would have an
> error signal of 1.0 and a perfect player would have
> an error signal of
> 0.  Has anyone else used anything like this?  Or
> know of another
> intrinsic performance metric?

i'm still trying to parse this function that you're
describing.  Can you describe it a bit more?

what's the expected value of any particular point
before any stones are put on the board?

i'm trying to correlate your 0.0 metric with the
idea of a perfect player.

presume that there exists a perfect human player.
likely, they wouldn't take all territory, even
(or especially) in a game against themselves.
getting more territory isn't "winning more", it's
just plain winning.

you know, the old "take all 4 corners, lose the game"
proverb...

s.


		
__________________________________ 
Yahoo! Messenger 
Show us what our next emoticon should look like. Join the fun. 
http://www.advision.webevents.yahoo.com/emoticontest
_______________________________________________
computer-go mailing list
computer-go@xxxxxxxxxxxxxxxxx
http://www.computer-go.org/mailman/listinfo/computer-go/