BGonline.org Forums

32Z-65 data

Posted By: Nack Ballard
Date: Tuesday, 16 November 2010, at 4:44 p.m.

In Response To: 32Z-65^10 RO clear result (Claude Landry)

Thanks for the ^10 rollout, Claude. Here is data for 32Z-65:

[R S10] ~31
[R S8] /15
[R S19] <62
[R S21] ^10

d[R S9] /5
g[S R21] /5

Key: R = Run (24/13), S = Split (24/18 13/8). Error sizes are in thousandths. Before the brackets, d = dmp, g = (gammon) go; else assume money. After the brackets, bot symbols are: ~ Snowie, / GnuBG, < XG 3-ply, ^ XG 4-ply. This bot symbol is followed by the number of trials, in thousands (rounded down).

For example, "g[S R21] /5" means that at (gammon) go, S is best and R is -.021 in a GnuBG rollout of 5000+ trials.

Agg (bot aggregate value) for 32Z-65 is [R S13]. Contrast with 52S-65, where agg is [R S57]. White's 10pt builder gives Blue a relative incentive of .044 for staying on the 18pt (with S); not enough to cross over zero, but enough to reduce the margin from .057 to .013.

This effect is spotlighted below at gammon go. According to Gnu's rollouts, S is best by .021 in the left-hand position, whereas R is best by .006 in the right-hand position; a swing of .027.

32Z-65 g[S R21] /5

52S-65 g[R S6] /15

Nack

Messages In This Thread

• 32Z-65^10 RO clear result
  Claude Landry -- Sunday, 14 November 2010, at 7:29 p.m.
  • 31S-65S Beats 31S-65R
    Paul Weaver -- Sunday, 14 November 2010, at 11:30 p.m.
    • 31S-65
      Nack Ballard -- Tuesday, 16 November 2010, at 4:56 p.m.
      • 31S-65, Snowie rollout
        Nack Ballard -- Tuesday, 16 November 2010, at 5:03 p.m.
        • 31S-65, Snowie rollout
          Nack Ballard -- Tuesday, 16 November 2010, at 5:09 p.m.
  • 32Z-65 data
    Nack Ballard -- Tuesday, 16 November 2010, at 4:44 p.m.
  • 4-ply rollouts
    Nack Ballard -- Tuesday, 16 November 2010, at 5:31 p.m.