| |
BGonline.org Forums
eXtreme Gammon question
Posted By: Bob Koca In Response To: eXtreme Gammon question (MaX)
Date: Friday, 28 August 2009, at 5:09 p.m.
"Suppose you rollout 100 moves until you have 99% confidence that play1 is better than play2, play3, .... play100 (individually). I think the user would be surprised to see that the confidence of play1 being better than all the others is .99^99 = 0.37 ... "
It is not correct to multiply becuase of lack of independence. Suppose that play 2 is actually better than play 1. This is an indication that the true value of play 1 is lower thatn the rolled out value and taking this into account we no longer think that the other plays are still 99%.
| |
BGonline.org Forums is maintained by Stick with WebBBS 5.12.