BGonline.org Forums

Why those settings?

Posted By: Chris Yep
Date: Wednesday, 13 July 2016, at 4:22 a.m.

In Response To: Why those settings? (Bob Koca)

Any idea as to why the output explicitly included "cubeless" for my RO but not for Phil's RO?

You did a cubeless rollout. Phil didn't do a cubeless rollout.

You explicitly selected "cubeless" in your rollout settings. Phil did a standard 3-ply Large cubeful rollout. Since the position has a cube cap of 1 (maximum cube value = 1), it's true that the rollouts were essentially identical (ignoring the different seeds; Phil used seed 1), but technically Phil still did a cubeful rollout.

As for why your cubeless rollout has different gammon numbers than Phil's cubeful rollout, that may be related to the low cube cap. I don't have much experience in looking at low cube cap positions, but XG may be doing something strange in Phil's rollout (cubeful rollout but cube cap of 1; Jacoby rule). This is arguably a bug. Note that even if the Jacoby rule is not used, a cubeful rollout (cube cap = 1) still results in a lower gammon rate than a cubeless rollout or a DMP rollout.

I was able to duplicate Phil's result. 20/13 is a tiny bit different from Phil's 20/13, but that's probably because I'm using a slightly different version of XG.

		White is Player 2 score: 0 pip: 94 Unlimited Game Jacoby Beaver pip: 112 score: 0 Blue is Player 1
XGID=a--BCBB-BA---AA---ceAbbb--:0:0:1:61:0:0:3:0:0
Blue to play 61

1.	Rollout1	13/7 8/7	eq: +0.672
	Player: Opponent: 83.59% (G:7.33% B:0.26%) 16.41% (G:6.09% B:0.34%) Conf.: ± 0.003 (+0.669...+0.675) - [100.0%] Duration: 3 minutes 41 seconds
2.	Rollout2	20/13	eq: +0.608 (-0.064)
	Player: Opponent: 80.38% (G:5.02% B:0.22%) 19.62% (G:2.98% B:0.05%) Conf.: ± 0.003 (+0.604...+0.611) - [0.0%] Duration: 3 minutes 31 seconds
1 1199 Games rolled with Variance Reduction. Moves and cube decisions: 3-ply Search interval: Large 2 1198 Games rolled with Variance Reduction. Moves and cube decisions: 3-ply Search interval: Large

eXtreme Gammon Version: 2.19.211.pre-release

Messages In This Thread

• Don't agree with xg...need explanation please
  Phil Simborg -- Monday, 11 July 2016, at 10:11 a.m.
  • Don't agree with xg...need explanation please
    Casper Van der Tak -- Monday, 11 July 2016, at 11:13 a.m.
  • Don't agree with xg...need explanation please
    Rick Janowski -- Monday, 11 July 2016, at 12:51 p.m.
    • Don't agree with xg...need explanation please
      Rick Janowski -- Monday, 11 July 2016, at 12:54 p.m.
      • Why those settings?
        Bob Koca -- Monday, 11 July 2016, at 3:51 p.m.
        • Why those settings?
          Rick Janowski -- Monday, 11 July 2016, at 4:30 p.m.
          • And what actually were his settings?
            Bob Koca -- Monday, 11 July 2016, at 5:44 p.m.
            • And what actually were his settings?
              Rick Janowski -- Monday, 11 July 2016, at 6:12 p.m.
        • Why those settings?
          Timothy Chow -- Monday, 11 July 2016, at 10:55 p.m.
          • Why those settings?
            Rick Janowski -- Tuesday, 12 July 2016, at 7:46 a.m.
          • Why those settings - binary logarithm of cube value.
            Rick Janowski -- Tuesday, 12 July 2016, at 8:11 a.m.
          • Why those settings?
            Bob Koca -- Tuesday, 12 July 2016, at 8:23 p.m.
            • Why those settings?
              Timothy Chow -- Tuesday, 12 July 2016, at 10:59 p.m.
            • Why those settings?
              Chris Yep -- Wednesday, 13 July 2016, at 4:22 a.m.
  • Don't agree with xg...need explanation please
    neilkaz -- Monday, 11 July 2016, at 7:45 p.m.