Re: [computer-go] Win rate and handicap

[David G Doshay <ddoshay@xxxxxxx>]

I guess I read the chart differently. If we assume that Many Faces is
perhaps an 8K and SlugGo is perhaps 4 stones stronger, then the
chart shows that between human players the win rate would be
about 20% for Many Faces when playing even. That is very close
to what we see for Many Faces playing first against SlugGo with
no komi.

I would not assign any ratings based upon this winning percentages
chart, but it is interesting to see that these programs show about the
same distribution in win/loss as do humans at about the same levels.

Also, I do not claim that SlugGo is really a 4 kyu player, only that it
plays even with Many Faces at a 4 stone handicap.

Cheers,
David

On 19, Nov 2004, at 7:44 AM, Jon Diamond wrote:

> The following link points to a page with a statistical analysis of 
> European
> tournament games.
>
> http://gemma.ujf.cas.cz/~cieply/GO/statev.html
>
> FWIW this indicates if these were human players then the weaker one 
> would be
> 1 dan! But I don't think these results can be extrapolated to non-human
> players.
_______________________________________________
computer-go mailing list
computer-go@xxxxxxxxxxxxxxxxx
http://www.computer-go.org/mailman/listinfo/computer-go/