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

BGonline.org Forums

Opening 21 [$ 7.9S]46+

Posted By: Nack Ballard
Date: Friday, 7 January 2011, at 5:40 a.m.

In Response To: Opening 21 [$ 7.9S]46+ (eXtreme Gammon)

Well researched, Xavier!

It doesn't alter your net adjustment much, but based on subsequent extensions by various bots, FYI some aggs have (evidently) moved by .001 or .002.

    21$-11 [N H23], prev. [N H21]
    21$-21 [$ S4], same
    21S-21 [U S3], prev. [U S4]
    21S-43 [X Z3], prev. [X Z2]
    21S-63 [S R3], prev. [S R4]

XG2's agg errors are underlined. This means that for the responses you checked:

    21$ equity needs to be reduced by 0.004/18 = 0.00022
    21S equity needs to be reduced by (0.003+0.003)/18 = 0.00033

Therefore, to account for known reply errors, the margin of Neil's rollout is only slightly narrowed from [$ S7.9] to [$ S7.8].

The biggest error effect is .023/36 = .00064 from 21$-11H, but XG2 correctly plays N, so only the previously compiled XG1 3-ply and 4-ply results get narrowed: the value of $ is inflated by the greater part of .001. So those two might get adjusted from [$ S5] and [$ S3], to [$ S4] and [$ S2], respectively (we wouldn't know without checking all equities to the fourth decimal place).

I believe Snowie had no reply errors to either 21$ or 21S (or if they did they canceled), so it is a solid [$ S8]. GnuBG has errors to 21$ exceeding those to 21S by .005 (by David Rockwell's research), which means we need to mentally adjusting Gnu's large [$ S19] margin to what still seems to be a rather inflated [$ S14]. This is a bit of evidence that Gnu's results in general seem to be slot-happy even after factoring out immediate misreplies.

Nack

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.