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

BGonline.org Forums

Long / Close Rollouts

Posted By: David Rockwell
Date: Thursday, 4 June 2009, at 1:08 a.m.

In Response To: Long / Close Rollouts (Neil Robins)

That's the technical issue. When you correct one error term, you are left with another. The question is whether or not an adjustment brings you closer to the mathematical value of a move. I've given this thought without attempting a proof. I think the 1 ply adjustment is an improvement because:

1. The residual error term is alternating (the sign reverses every ply) and probably decreasing.

2. I'm only suggesting doing this when there are an unusually large number of one ply errors. In that instance, the one ply errors should be relatively large compared to the errors at later ply. We can assume that later ply errors will be a "normal" size in most instances.

So, for example, I believe that a sequence of errors by ply might look like this (in total!):

.0050 -.0010 .0005 -.0002 .0001 etc. The first term will correct the majority of the error since the later amounts quickly become insignificant. In this totally fictional example, the total net error is .0044. The .0050 term overcorrects the error by approximately 15%. And, of course a .0006 residual error is too small even for me to care about.

And, I agree with Richard's comments (below).

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.