| |
BGonline.org Forums
comparison of all the new bots vs the old bots
Posted By: Frank Berger In Response To: comparison of all the new bots vs the old bots (Elizabeth (erichktn @ DG))
Date: Thursday, 12 November 2026, at 5:02 p.m.
A week ago, Marc asked if I’d like BGBlitz to be included in the figures (yes, of course), but it didn’t work out for him in terms of time integrating BGBlitz (and I would have had to do a bit of work on it as well).
His approach is fine in principle, but you should take computing time into account as well.
Using the same ply makes the nets comparable, and a lower play is better because larger search depth mitigates evaluation deficiencies. IMHO 1-ply and ABT-clock settings makes the most sense.
What I don’t like about the statement ‘XY is the strongest bot’ is for two reasons: - Firstly, because it implies that bot XY is *always* better, and that’s nonsense. - Furthermore, I haven’t yet seen a measurement method that has really convinced me (and unless I’ve miscalculated badly, the number of games at higher ply levels is still far too small here)
A large number of matches with ABT settings (or a larger number of hand-picked positions) would be the best approach in my view.
| |
BGonline.org Forums is maintained by Stick with WebBBS 5.12.