| |
BGonline.org Forums
A better approch? [LONG]
Posted By: Maik Stiebler In Response To: A better approch? [LONG] (MaX)
Date: Monday, 9 November 2009, at 10:59 p.m.
(equities, stdevs, number of MC trials)
First move had an equity of 2.5 (whatever the unit), other moves 0.0. Stdev was 1.0 for all moves. I'm sorry I did not read your description of your test program carefully enough: I did not do the rollout simulation step that it does, so my numbers are meant for the case that those equities I just wrote down are the rollout results, not the real equities. Anyway, the way I did it is the only way to demonstrate my point: If you get a rollout result that has one move 2.5 stdev ahead, and many moves nearly tied for second, MinCI will be off. But that point I tried to make is sort of irrelevant now: Imho, MinCI, and moreso CI_LB, is fine as a stopping criterion.
| |
BGonline.org Forums is maintained by Stick with WebBBS 5.12.