[Bug-gnubg] Strange Rollout Result DMP41c_62.htm

[Ian Shaw]

 <<DMP41c_62.htm>> Can anyone explain why the winning percentage does
not match the MWC in this DMP rollout?

 DMP41c_62.htm

An identical rollout for a 63 roll works OK.
-- Ian

Game number 1

The score (after 0 games) is: Leader 0, Trailer 0 (match to 1 point, post-Crawford play)

---

Match Information

Date	January 13, 2005
Comment	Responses to initial 41 played 13/8

---

Move number 8: Trailer to play 62

Pip counts: Leader 162, Trailer 167
Position ID: 4PPgATDgc/ABMA Match ID: cAkrAAAAAAAA

#

Ply

Move

MWC

1

R

24/18 13/11

49.66%

27.3%	8.9%	4.8%	-	72.7%	27.0%	3.9%	49.66%
116.5%	15.2%	1.3%	-	116.5%	8.9%	0.8%	1.40%

Full cubeless rollout (trunc. at one-sided bearoff) with var.redn.

1296 games, Mersenne Twister dice gen. with seed 952953803 and quasi-random dice

Play: 2-ply cubeless prune

keep the first 2 0-ply moves and up to 3 more moves within equity 0.1

keep the first 0 1-ply moves

Cube: 2-ply cubeful prune

2

0

24/16

49.36% ( -0.31%)

49.4% 11.4% 0.5% - 50.6% 15.0% 0.4%

3

0

13/11 13/7

48.29% ( -1.38%)

48.3% 14.0% 0.7% - 51.7% 15.6% 0.9%

4

0

13/5

47.98% ( -1.69%)

48.0% 13.6% 0.7% - 52.0% 15.4% 0.9%

5

0

24/18 6/4

47.77% ( -1.89%)

47.8% 11.7% 0.6% - 52.2% 17.3% 0.6%

---

Output generated Fri Jan 14 17:18:44 2005 by GNU Backgammon 0.14.2-devel (HTML Export version 1.163)