BGonline.org Forums

Opening 21 and analysis of XG 4-ply RO data

Posted By: Nack Ballard
Date: Wednesday, 7 April 2010, at 3:02 a.m.

First, here is data for opening 21:

Money
Snowie .. [$ S8] 80k
GnuBG .. [$ S16] 46k (same as site)
XG 3-ply [$ S6] 140k
XG 4-ply [$ S3] 46k

DMP
GnuBG .. [$ S13] 15k (same as site)

GS
GnuBG .. [$ S1] 61k
BGBlitz .. [S $14] 7k

GG
GnuBG .. [$ S32] 15k

The GS and GG Gnu results on the (bgonline) site are subsets but close in this case: s[$ S3] 46k, g[$ S31] 10k.

Other scores
XG 4-ply [S $8] 28k at -2-4 (2 away, 4 away)
GnuBG .. [S $13] 20k at -2-7 (MET = Gnu11)
GnuBG .. [S U9 V17 $17 D21] at -1-2 post-Cr (MET = Gnu11)

I had already long been confident 21S is best at GS (despite what is up at the site) based on Gnu's typical slot bias and other data, and your XG 4-ply RO at -2-4 of [S $8] further affirms it.

Nactation key: $ = Slot 13/11 6/5, S = Split 13/11 24/23. As usual, the play that wins the rollout is listed first, followed by the other play with its error size in thousandths of a point, while the "k" number outside the brackets is thousands of trials (rounded down). For example, [$ S6] 140k means that $ (slotting) beats S (splitting) by .006 after 140,000 iterations.

Secondly, here are (money, bot RO average) values for the other positions you mentioned (top plays only):

21S-21 [U S4]
21S-43 [X Z2]
21S-63 [S R4]
21$-11 [N H21]
21$-21 [$ S4]
21$-62 [S Z6 N11]

The Gnu rollout margins on the site are within .002 of value for 21S-21, 21S-63 and 21$-21. However, the site is somewhat off value for 21S-43 [X Z10], 21$-11 [N H31], and 21$-62 [Z N1 S2].

Your comments imply to me that XG 4-ply misplays 21S-21S, 21S-43Z, and 21S-63, by .004, .002 and .004, respectively, a total of .010, then divide by .018, for an effect of .00056 on the rollout margin.

OTOH, XG 4-ply misplays 21$-11H and 21$-21S, by .004 (divide by 18) and .021 (divide by 36), respectively, for a countereffect of .0008 on the rollout margin.

The net of these two effects is that your XG 4-ply rollout result of [$ S3] should probably be rounded down just a tad -- to the high side of [$ S2], given that the absolute equities of .013 and .011 already suggested a margin on the underside of .003.

Lest we forget, there is still the matter that the rollouts of XG 3-ply, Snowie and Gnu all have higher slot-favored margins (.006, .008 and .016, as shown at the beginning of this post) than XG 4-ply's mere .002 (after the slight adjustment). Make of this what you will.

Nack

Messages In This Thread

• Opening 21 and analysis of XG 4-ply RO data
  Nack Ballard -- Wednesday, 7 April 2010, at 3:02 a.m.
  • Opening 21 and analysis of XG 4-ply RO data
    Neil Robins -- Wednesday, 7 April 2010, at 9:44 a.m.
  • Opening 21 and analysis of XG 4-ply RO data
    Neil Robins -- Wednesday, 7 April 2010, at 10:04 a.m.
  • Opening 21 and analysis of XG 4-ply RO data
    neilkaz -- Wednesday, 7 April 2010, at 2:12 p.m.