Re: [Bug-gnubg] Beaver issue ?

[Christian Anthon]

There was no test for this very rare case. Hopefully I fixed it right,
but I can't test without the sgf file.

Christian.

On Wed, Apr 16, 2008 at 5:05 PM, Massimiliano Maini
<address@hidden> wrote:
>
> Hi all,
>
> from a post on BGonline forum:
>
>         http://www.bgonline.org/forums/webbbs_config.pl?frames;read=18634
>
> Position ID: t2xLgADa2wEDQA Match ID: cAkAAAAAAAAA
>
> Cube decision
> Rollout cubeless equity +0,1830
> Cubeful equities:
> 1. Double, take         -0,1315
> 2. Double, pass         +1,0000         +1,1315
> 3. No double                 -0,1732         -0,0417
> Proper cube action: Double, beaver
>
> As pointed out in a message of the thread:
>
> "Probably I'm confused, but this looks a gnubg bug to me.
>
> If double/take is -0.13, then double/beaver is -0.26. And if no double is
> -0.17, then this
> isn't double/beaver, it's no-double/beaver. (It's double/take if beavers are
> not allowed.)"
>
> I'm not sure if this is related to Item #15434: Wrong cube evaluations in
> games with beavers
> on bug tracker ... I would say no since here the issue is not in the
> analysis but just in
> the interpretation of the rollout results. If it's a separate issue, I can
> open a new item
> (I may even try to ask for the sgf file).
>
> MaX.
> _______________________________________________
>  Bug-gnubg mailing list
>  address@hidden
>  http://lists.gnu.org/mailman/listinfo/bug-gnubg
>
>