The code is based on Tom Keith ideas about computing the Match Equity Table (http://www.bkgm.com/articles/met.html)
The idea is to get the cubefull equity from the cubeless by assuming the equity functions is as a straight line between the take points (adjusted for gammons etc.) So this is not Janowksi's formulae ... -Joseph On 5 January 2012 22:19, Massimiliano Maini <[email protected]> wrote: > That would be interesting indeed. > Do you have any description of what your own cube code does ? > At the time I just went as far as verifying that the current cube code is. > OK with respect to R.Janowksi's formulae (and it was). > > Not sure I can help in integrating your code in gnubg, but I'm definitely > curious in understanding what's different with respect to the current one. > > MaX. > > On 5 January 2012 09:35, Joseph Heled <[email protected]> wrote: > > I wonder if this is just the net or the cube code contributes as well. > > As you know, the cube code in gnubg-nn is different from the one used in > > gnubg. (my own dewvelopment). > > I always thought it did better, but it will be interesting if someone > > verifies it, and offers it as a gnubg option (especially with regard to > the > > odd/even issue). > > > > -Joseph > > > > On 5 January 2012 21:17, Massimiliano Maini <[email protected]> wrote: > >> > >> On 5 January 2012 02:20, Joseph Heled <[email protected]> wrote: > >> > Seems like you got a very very slightly better race net, but I would > be > >> > surprised if it makes a difference in real life. > >> > > >> > Would be much more interesting to > >> > - get a better contact or crashed net > >> > - expand the roll-out database for all categories (should be easy > with > >> > the > >> > current availability of cycles) > >> > - improve cube decisions (this is a hard one) > >> > - improve back game evaluation and play (very hard one) > >> > > >> > -Joseph > >> > >> One thing that has always puzzled me is the strange behavior of the > >> strength > >> of gnubg at different plies. The last large scale study (done to compare > >> the > >> existing bots to the new extreme gammon), is resumed here: > >> > >> http://www.extremegammon.com/studies.aspx > >> > >> Checker play is fine, error goes down as plies go up. > >> However, for cube actions, the situation is very strange: > >> > >> 3ply does much better than 2ply and 4ply on missed doubles and wrong > >> takes. > >> But 3ply does terribly worse than 2ply and 4ply on wrong doubles and > >> wrong passes. > >> > >> It has always been accounted to the so called "odd-even" effect, but > >> no other bot > >> seems to be affected as much as gnubg. > >> > >> MaX. > >> > >> _______________________________________________ > >> Bug-gnubg mailing list > >> [email protected] > >> https://lists.gnu.org/mailman/listinfo/bug-gnubg > > > > >
_______________________________________________ Bug-gnubg mailing list [email protected] https://lists.gnu.org/mailman/listinfo/bug-gnubg
