[Computer-go] Gomill - GTP strength comparison and parameter tuning tools

Olivier Teytaud teytaud at lri.fr
Mon Nov 15 16:19:00 PST 2010


At least there are upper complexity bounds like that :-)
>

I should precise that the upper complexity bound is for the case in which
the success rate is 100% at the optimum, which is clearly
not the most realistic setting; I would say that the result also holds
without this assumption, but I don't find a reference showing this
(showing this mathematically I mean)...
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://computer-go.org/pipermail/computer-go/attachments/20101116/3b95b97c/attachment.html>


More information about the Computer-go mailing list