Thanks a lot for the replay, I did not find this old discussion. move probability is gamma/(sum of all gammas), so my formula would be equivalent to using
log(probability) (as 1/(sum of all gammas) results in a additive constant after log) so the question is: why should I add winrate + probability/n_i and not using log and still why not any of the other decaying which I thought might be correct and why not using playouts of parent :) Thanks Detlef Am Sonntag, den 23.02.2014, 12:00 +0100 schrieb Harald Johnsen: > Hi, > > I think that bots that computes gamma values are not using the gamma but > the probability of the move for the progressive or constant bias. > > see : > https://groups.google.com/forum/#!searchin/computer-go-archive/gamma$2050/computer-go-archive/Zi3Rw0ePUw8/jzfOKyU9bssJ > > hj. > > _______________________________________________ > 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
