                        |
| ||||
| XGID=-A---ED--aa---------abdcc-:1:-1:1:21:2:1:0:15:10 | |||||
| Blue to play 21 | |||||
| 1. | Rollout1 | 5/3 1/Off | eq: +0.9100 | |||
| ||||||
| 2. | Rollout1 | 6/4 1/Off | eq: +0.9093 (-0.0006) | |||
| ||||||
| 3. | Rollout1 | 5/4 5/3 | eq: +0.9078 (-0.0022) | |||
| ||||||
| 4. | Rollout1 | 6/4 5/4 | eq: +0.9069 (-0.0031) | |||
| ||||||
| 1 41472 Games rolled with Variance Reduction. Dice Seed: 83788161 Moves: 3-ply, cube decisions: XG Roller | ||||||
Rollout by Taper_Mike |
eXtreme Gammon Version: 2.03, MET: Kazaross XG2
| Thanks, HiGoneFive, for your reply. No doubt, you are right on the mark in you comments. I seem to have misled you, however. I meant to imply, In the position above, ... With that caveat, I think my remarks are accurate, too.
It is ironic, therefore, that both of us were wrong in this position! True, our error was only 0.0006, but the XG rollout above gives a 0% chance that 6/4 could be correct. The takeaway for me is that a gap on the 4pt is less harmful when the the 1pt and 2pt are also empty. That is because a roll of 4 can be played productively by moving a checker onto one of those points. Only when you roll a third 4 in the position above, before you also roll a 1 or 2, will you find that you are stacking a second checker on the 1pt or 2pt. In this light, it can be seen that keeping 4 checkers on both the 5pt and 6pt (by playing 5/3), is better than creating the uneven distribution given by 6/4. Mike |