| |
BGonline.org Forums
43-D-st Take
Posted By: Nack Ballard In Response To: 43-D-st Take (David Rockwell)
Date: Thursday, 3 February 2011, at 11:35 p.m.
Thanks for posting Stick's XG 3-ply rollout, David. I had only the nacbrac: d[D S1 U1 Z2] <31. Win% for D, Z, S and U, respectively (to be consistent with the order of Neil's money plays I listed earlier), are 50.06, 49.95, 50.02, and 49.98.
I found the text file from an opening 43 money rollout (done by Ken), which btw is [D Z6 S10 U14] <62. It gives win% of D = 50.03, Z = 50.04, S = 50.09, and U = 50.18. Summarizing the XG 3-ply data, then, we get the money vs dmp win% differentials in the right-hand column below.
...money...dmp...Diff
D..50.03..50.06..+.03
Z..50.04..49.95..-.09
S..50.09..50.02..-.07
U..50.18..49.98..-.20This is somewhat different from GnuBG's differentials, which are:
...money...dmp...Diff
D..49.81..49.85..+.04
Z..50.02..50.07..+.05
S..49.85..49.81..-.04
U..49.83..49.95..+.12If we (naively?) average XG 3-ply's and GnuBG's differentials (and round towards XG) and project them onto Neil's XG2 Roller+ rollout's money win%s, we "predict" the dmp column below:
...money...dmp...Diff
D..49.82..49.85..+.03
Z..49.95..49.93..-.02
S..49.95..49.89..-.06
U..49.77..49.73..-.04If the projected "+ (XG2 Roller+) rollout dmp win%s (middle column) match closely with Neil's actual dmp win%s, that would suggest that one should play Z at –p–2 but then drop anyway. Then again, the money/dmp comparison might act differently at the higher ply.
It is worth noting that the XG 4-ply and XG2 roller+ results both considerably dogged U for money (-.027 and -.023), though only 10k and 7k trials (respectively) were done; variance is still an influence. Compare with the 62k result on XG 3-ply: U is -.014.
Nack
| |
BGonline.org Forums is maintained by Stick with WebBBS 5.12.