[ View Thread ] [ Post Response ] [ Return to Index ] [ Read Prev Msg ] [ Read Next Msg ]

BGonline.org Forums

a bias in rollouts

Posted By: bob koca
Date: Thursday, 1 May 2008, at 6:24 a.m.

Suppose one finds the equity of a position by using the common technique of rolling out the top candidates and taking as the equity estimate the result for the best play. That technique is biased in the direction of giving too high an answer. If the top play is clear then it doesn't matter but suppose that there are two or more plays whose equities are very close. Winning the RO means there is more than 50% chance of the RO overstating the equity. Tus overall the technique is baised.

As an example consider a situation where there are 3 plays and they all actually give equity of exactly 0. The position will probably roll out as having positive equity since one of the 3 will probably RO better than it actually is.

Has this been pointed out before?

Bob Koca

Messages In This Thread

 

Post Response
Your Name:
Your E-Mail Address:
Subject:
Message:

If necessary, enter your password below:

Password:
 

 

[ View Thread ] [ Post Response ] [ Return to Index ] [ Read Prev Msg ] [ Read Next Msg ]

BGonline.org Forums is maintained by Stick with WebBBS 5.12.