On Fri, Dec 03, 2010 at 09:08:16PM +0800, Aja wrote:
> >Can it? Pachi also has 2-point semeai solver, which can solve the
> >situation well if the *opponent* reduced the group to two liberties;
> >it seems your feature has the similar properties.
>
> >I'd be curious how this is prevented in Erica?
>
> This is not so hard. I believe you must be able to figure out the
> solution easily by yourself. Attached is Erica's owner at move 157
> computed from 500 playouts without tree.
Well, my plan to fix this is to introduce a "self-2lib" routine akin to
the "self-atari" routine. It's just I could not map this to anything in
your paper, but perhaps it's so trivial you did not consider it worthy
of mentioning. :-)
Petr "Pasky" Baudis
_______________________________________________
Computer-go mailing list
[email protected]
http://dvandva.org/cgi-bin/mailman/listinfo/computer-go
