Hello Theodore, thanks for your feedback and for the bug report. Indeed, I had inverted two variables and it gave wrong equities in match play. This is now fixed.
Thanks again, ciao, Alessandro On Fri, Apr 25, 2025 at 3:56 AM Theodore Hwa <[email protected]> wrote: > Awesome work! > > But I notice some discrepancy between the results from the gnubg-core-demo > website and my local version of GnuBG. For example, this position: > > XGID=---B-bD-C--AcC--bb-db---B-:0:0:1:22:0:4:0:7:10 > > At 2-ply, for the move 24/22 24/22 13/11 6/4 > > gnubg-core-demo gives -0.638 > my local GnubG gives -0.269 > > (For what it's worth, XG at 3 ply gives -0.307) > > Ted > > On Tue, 22 Apr 2025, Alessandro Scotti wrote: > > > > > Hi everyone, > > > > I've built a minimal and portable version of GNU Backgammon that > includes only the core logic — everything else has been stripped > > away, including the dependency on GLib. > > > > This version is specifically designed for the web: it loads fast, > performs very well even on mobile devices, and is extremely easy > > to use. Of course, it can also be used directly in C projects if needed. > > > > You can try a very basic web interface here: > > > > https://ascottix.github.io/gnubgcore/gnubg-core-demo.html > > > > The repo is here: > > > > https://github.com/ascottix/gnubg-core > > > > I'd be happy to share more details if there's interest, and I'm open to > feedback or suggestions. > > > > Cheers, > > > > Alessandro > > > > > >
