Valkyria prunes a lot of move and do a lot of computations for doing
so. It also is faster with the code than without since games become
shorter when stupid moves that kills groups are removed. If a large
group is killed by light playouts on 19x19 then a light playout will
play a lot of moves inside this area. In the worst case leading to
even more cases...
With proper pruning and good patterns for defense playout get more
civilized and shorter.
Valkyria also can call a function IsAlive(pos) that can tell if the
stones near position "pos" have unconditional eyeshape. It does not
handle all cases but some of them. Since it is only called when needed
deep after many conditions have been fulfilled (for special patterns
where life is likely and is important for if a move is relevant or
not). I think it is more effective than running a fullboard eval of
life and death after each move.
-Magnus
Quoting fastgo <[email protected]>:
if the region is marked as unconditionally belonging to either black or
white, then black and white are forbidden to play inside the region.
In the case we discussed, if both false eyes are identified as unconditional
eye, then it cannot be filled. So no side will play inside in it.
If a program has less blunder to fill eye, that definitely reduce the
playout length.
On Wed, Apr 6, 2011 at 12:05 PM, Brian Sheppard <[email protected]> wrote:
How does the Benson algorithm result in shorter playouts?
A more precise question: how does your playout apply the ability to
determine unconditional life to the question of which moves one should play
in the playout, or to determining whether to continue to play out.
Thanks
-----Original Message-----
From: [email protected]
[mailto:[email protected]] On Behalf Of Erik van der Werf
Sent: Wednesday, April 06, 2011 10:49 AM
To: [email protected]
Subject: Re: [Computer-go] living by two false eyes
On Wed, Apr 6, 2011 at 4:11 PM, fastgo <[email protected]> wrote:
> In StoneGrid, it has built in the Benson algorithm in the playout. It
does
> not seem to help much on overall strength. But I think it should solve
the
> case here.
>
> I think you can implement the Benson Algorithm in the late stage of the
> playout.
>
> It has another advantage finishing the playout in shorter length. In
> StoneGrid, the playout on average finishes in a little bit more than 70
> stones on 9x9.
>
> You need quite a lot of effort to get it fast as well as correct. So the
> return of the effort is low.
I also have it in my program. The funny thing is that if I disable it
the program actually becomes slower because of the longer playouts...
Erik
_______________________________________________
Computer-go mailing list
[email protected]
http://dvandva.org/cgi-bin/mailman/listinfo/computer-go
_______________________________________________
Computer-go mailing list
[email protected]
http://dvandva.org/cgi-bin/mailman/listinfo/computer-go
_______________________________________________
Computer-go mailing list
[email protected]
http://dvandva.org/cgi-bin/mailman/listinfo/computer-go
