Re: RE: computer-go: Bad Tactical move in MFG

[David Fotland <fotland@xxxxxxxxxxxxxxx>]

I think you might be giving Many Faces too much credit :)  It understands that
after extending on the edge (a) that the two 2-stone groups have weaknesses,
but it does not find this tesuji.  Version 10 thinks that weakening the 4 
white stones
justifies taking a chance on dying.  The current version makes the eye.

David

At 03:16 PM 5/10/2002 -0400, you wrote:
> >Ok, I'm gonna ask the dumb question: Why does X live after a and c?
> >I could probably see it on a 9x9 board as it extends on the right toward the
> >corner (assumed empty) and make another eye there, but otherwise?
> >
>
> After a and c
> We get :
>
> |......
> |.OOO..
> |OXXO..
> |.XO...
> |XXOf..
> |OXXOOX
> |O.XXde
>  ------
>
> If O doesn't play 'd', X plays 'd', connects to the new stone, because O 
> won't be able to cut at 'e'.  Then, X can double-atari the two white
> groups. If O took time to prevent this, X has free space to the right to 
> make a second eye.
> If O plays 'd', X plays 'f' and then is able to capture one of the two 
> white groups, making a second eye.
>
> The best I see for O after a and c is :
>
> |......
> |.OOO..
> |OXXO..
> |.XO...
> |XXO.OO
> |OXXOOX
> |O.XXX.
>  ------
>
> With X to play. This is bad.  O was better enclosing X than letting it get 
> out like that.
>
> I'm quite surprised that MFGO spotted that sequence and made the right 
> move (good work David).  I'm usually able to beat it, and unless
> another poster had mention this tesuji, I would not have spotted it.
>
> Jeff.
David Fotland