| |
BGonline.org Forums
eXtreme Gammon question
Posted By: Maik Stiebler In Response To: eXtreme Gammon question (Bob Koca)
Date: Thursday, 27 August 2009, at 9:44 a.m.
I'm also for a more Bayesian viewpoint in backgammon statistics.
Your technique of asserting the best play based on a 95% confidence RO actually gets it right more than 95% of the time.I wonder if and how this can be generalized.
To further confuse the issue, suppose we rollout two plays which we just know can only be a smidgen of equity apart (say, they represent different ways of shuffling checkers in your homeboard) and a 216 trials rollout without VR tells us the jsd is 0.100 and the number of jsd's separating the plays is 2.4. I would not bet that the winning play is actually better at 1/100 odds.
| |
BGonline.org Forums is maintained by Stick with WebBBS 5.12.