[Computer-go] MCTS and perfect endgame

Jean-loup Gailly jloup at gailly.net
Sun Jul 3 06:12:59 PDT 2011


Leon,

> One of problems (which I tested with gogui, thankyou very much)
> was losing points in endgame when program is winning.

This is by design. Pachi maximises the chance of winning, not the number
of points. But if you want Pachi to win by more points while increasing
the risk of losing, you can simply increase the parameter val_scale. See the
description in uct/uct.c: "How much of the game result value should be
influenced by win size. Zero means it isn't". The default value is 0.04,
which is the result of tuning. (If you increase val_scale above this it
starts
losing more.)

Jean-loup

2011/7/3 Leon Matoh <leon.matoh at guest.arnes.si>

> MCTS and perfect endgame
>
>
> Greetings.
>
> My name is Leon Matoh. I am go player, 5d, from Slovenia.
>
> After long time I took time to work on go.
>
> I downloaded couple of free programs and was surprised of
> strength of pachi. Since it is open source i checked the code
> and tested some hacks.
>
> One of problems (which I tested with gogui, thankyou very much)
> was losing points in endgame when program is winning.
>
> I solved this problem and plan to contribute to project.
> I wonder if there is anybody else interested in this improvement?
> If so please tell me where to post the solution. (I change it for
> goodwill).
>
> Sorry if this is already solved.
>
>
> Leon Matoh.
>
> ______________________________**_________________
> Computer-go mailing list
> Computer-go at dvandva.org
> http://dvandva.org/cgi-bin/**mailman/listinfo/computer-go<http://dvandva.org/cgi-bin/mailman/listinfo/computer-go>
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://computer-go.org/pipermail/computer-go/attachments/20110703/b1ad55b4/attachment.html>


More information about the Computer-go mailing list