|
BGonline.org Forums
15552-game rollout of 41Z-65R-43
Posted By: Timothy Chow In Response To: 15552-game rollout of 41Z-65R-43 (Chuck Bower)
Date: Monday, 28 December 2009, at 7:33 p.m.
I mentioned one example (buried deep in another thread, I admit) to Bob Koca, regarding transpositions. Suppose one rolled-out game starts with the dice 41-65-43 and reaches position P1, and another rolled-out game starts with the dice 43-65-41 and reaches position P2. When calculating the equity of the original position, you'll find yourself doing the computation equity(P1) + equity(P2) at some point. The standard assumption is that equity(P1) and equity(P2) are independent random variables, so that when you add them, their variances add: Var(equity(P1)) + Var(equity(P2)). However, perhaps P1 = P2 (i.e., the variations transpose). Then Var(equity(P1) + equity(P2)) = Var(2*equity(P1)) = 4*Var(equity(P1)), so your estimate of 2*Var(equity(P1)) is an underestimate.
|
BGonline.org Forums is maintained by Stick with WebBBS 5.12.