| |
BGonline.org Forums
BOT COMPARISON 350 / 500 Games
Posted By: Daniel Murphy In Response To: BOT COMPARISON 350 / 500 Games (Michael Depreli)
Date: Thursday, 18 March 2010, at 11:58 p.m.
Just to note, from your 350/500 stats, Michael, that:
(1) Gnubg with 3-ply checker play and 2-ply cube has a total error of 27.994, which is 2.866 better than 2-ply/2-ply and 7.623 better than 3-ply/3-ply.
(2) again with Gnubg, going from 3-ply to 4-ply checker play is a 5.415 improvement, but going from 2-ply to 4-ply cube is only a 1.055 improvement. Also, it's interesting that 4-ply checker is so much better -- 8.281 -- than 2-ply checker. That suggests to me that 2- and 4-ply disagree more often than one might have thought. Although, it could also just mean that 2-ply occasionally gets caught in a very large blunder.
| |
BGonline.org Forums is maintained by Stick with WebBBS 5.12.