BGonline.org Forums

51$-32 and 51$-52, data

Posted By: Nack Ballard
Date: Thursday, 6 May 2010, at 7:10 p.m.

In Response To: 51$-32 XG 4 ply RO (neilkaz)

Blue to play 32 or 52

In the 51$-32 ROs, Snowie has $ worse than S by .010, Gnu has $ .012 better, while XG has the two plays within .001 (half-way in between). By contrast, there is much closer $/S bot agreement in the 51$-52 ROs.

Here is money data for 51$-32 and 51$-52, for comparison:

51$-32:
Snowie .. [S=Z $10 D15 U28] 28k 17kU
GnuBG .. [$ S12 Z20 D28 U45] 15k 5kDU
XG 3-ply [$ S1 Z7] 77k
XG 4-ply [$ S1 Z7] 46k

51$-52:
Snowie .. [$=S D26] 20k 10kD
GnuBG .. [$ S4 D27] 46k 5kD
XG 3-ply [$ S2.5] 31k
XG 4-ply [S=$ D17] 46k 5kD

Other scores (d = DMP, s = GS, g = GG), with GnuBG:

51$-32: d[$ S4.6] 5k, s[S Z4.6] 15k, g[$ D14 Z23 S29] 10k 5kZS
51$-52: d[$ S4.4] 5k, s[S $0.4] 15k, g[$ D13 S23] 5k

Key: For 51$-32, $ = 13/10 6/4, S = 24/21 13/11, Z = 24/22 13/10, D = 13/11 13/10, U = 24/22 24/21. For 51$-52, $ = 13/8 6/4, S = 24/22 13/8, D = 13/11 13/8.

Error sizes are in thousandths. The number of trials (thousands, rounded down) appears after the brackets. The score appears before the brackets: d = DMP, s = GS, g = GG, else assume money.

For example, "[$ S2.5] 31k" means that (for money) $ is best and S is -.0025 after 31000+ trials. Or, "g[$ D13 S23] 5k" means that at gammon go $ is best, D is -.013 and S is -.023 after 5000+ trials.

Nack

Blue to play 32 or 52

Messages In This Thread

• 51$-32 XG 4 ply RO
  neilkaz -- Thursday, 6 May 2010, at 6:29 p.m.
  • 51$-32 and 51$-52, data
    Nack Ballard -- Thursday, 6 May 2010, at 7:10 p.m.
    • 32$ - which slot?
      Gregg Cattanach -- Friday, 7 May 2010, at 7:06 p.m.
      • 32$ - which slot?
        Nack Ballard -- Friday, 7 May 2010, at 7:56 p.m.