|
BGonline.org Forums
synopsis of bot opinions
Posted By: Maik Stiebler In Response To: 10-10/11, they started with 11 and 21 seconds clocktime, and played 30 minutes (Sean Garber)
Date: Monday, 9 February 2015, at 7:20 p.m.
This is a synopsis of bot opinions just for the last game. Mochy has 92 decisions to Lorero's 53. So even while Mochy's errors are more costly in terms of total MWC according to the bots on most analysis levels, Mochy's PR is sometimes lower nonetheless.
PR(Mochy) PR(Lorero) MWC loss(Mochy) MWC loss(Lorero) Mochy's VRR (*) gnubg supremo, standard luck 57.2 ((5)) 23.3 ((5)) 525.759% 123.18% 117.77% XG 1-ply luck n.a. n.a. n.a. n.a. -286.74% XG World Class (4-ply luck) 14.43 15.14 132.72% 80.26% 142.26% XG Extensive(3-ply luck) (**) 13.16 16.2 121.06% 85.88% 96.18% XG RO-based VR (***) 3.97(****) 14.96(****) 36.503%(****) 79.303%(****) 92.80% (*): VRR = result - luck, result being 100% for win, 0% for loss
(**): used full rollouts, 1296 trials with XG 3-ply
(***): Errors compared to XG 3-ply. First pass 3-ply, each difference is rolled out by full rollouts with 3-ply, 1296 trials.
(****): relative to XG 3-ply
((5)): Gnubg doesn't do PR. This is Gnubg-style error rate multiplied by 0.5, which should be a good approximation.
|
BGonline.org Forums is maintained by Stick with WebBBS 5.12.