On Wed, 2012-07-18 at 09:33 +0300, Guido Flohr wrote: > On Tue, 2012-07-17 at 21:56 +0200, Philippe Michel wrote: > > I don't really understand what the code there does, but looking at the way > > it does it, I'd be inclined to let everything related to stddevs unchanged > > but increase the 6.0 value (in your case, to 6.0 * sqrt(100/64)). > > Yes, I tried that as well, only with log2(100) instead of sqrt(100/64). > It "works" as well. Still not really satisfying. > > > FWIW, there is an email address for David Montgomery here : > > http://www.bgonline.org/forums/webbbs_config.pl?noframes;profile=david+montgomery > > You could try to ask him directly.
That helped, thanks. The solution is obvious, when you look at the resulting values. When the assertion fails, the probability can be set to zero. Committed as revision 1.85 of matchequity.c. Guido -- Империя ООД | Imperia OOD ул. „Княз-Борис-I“ № 86, София 1000 | ul. "Knyaz-Boris-I" № 86, Sofia http://www.imperia.bg/ _______________________________________________ Bug-gnubg mailing list Bug-gnubg@gnu.org https://lists.gnu.org/mailman/listinfo/bug-gnubg