BGonline.org Forums

Opening rollouts in DB?

Posted By: Nack Ballard
Date: Thursday, 29 April 2010, at 5:17 p.m.

In Response To: Opening rollouts in DB? (eXtreme Gammon)

If anyone (Nack?) has that in a format I could use (i will need actual equity not only the differences)

No, sorry. For positions this early in the game, I only transcribe the margins (relative equities), and in straight thousandths (i.e., I even discard the zeros and decimal places).

One question that will arise is how to combine RO from different bot/level? should a XG 4-ply render all XG 3-ply RO obsolete?

I believe it shouldn't, especially as the 4-ply RO takes longer and therefore typically fewer trials are obtained than the 3-ply.

You watch. When the (shorthanded) 5-ply RO results first come out, people will too quickly abandon their belief in the 4-ply ROs, even though the latter will still be at least as reliable just on the strength of the number of trials.

As you may have noticed, I list results from different plies for the same bot separately, though I will sometimes also temporarily combine them for an average "value" (weighting the higher ply result and/or the greater number of trials more). As I have no program to do this for me automatically, it's a lot simpler to deal with margins than it would be with twice as many absolute equities, and in any case it is much easier to interpret the essential information.

However, if we take Neil's latest RO using XG 4-ply ... 64 is now [R=S P3].

For XG 4-ply, I've marked down [R=S P4] 46k. The one-thousandth difference in the P margin doesn't concern me as such, but the number of trials might, and I would certainly want to know of any new rollout, so please inform me if you know of any result that is extended or independent from the one Neil posted here.

Nack