This position proves why XGR++ can not be relied upon to be entirely accurate. This is a position where the difference in plays must be 0.
Even when you roll it out 1296 there is still a .001 difference in the plays, so even that is wrong. And if you use XGR+ it is a BLUNDER to make the second best play. (It gets even funnier, as XGR+'s play is 9/3 6/1!)
Now, my goal is not to criticize XG, but to simply remind everyone that:
a) use XGR++ over XGR+;
b) you need to roll things out to be more certain, and
c) even rollouts are not 100 percent.
                     |
| ||||
| XGID=-BB------a-ab---bb-cbb----:1:-1:-1:65:0:1:0:3:10 | |||||
| Blue to play 65 | |||||
| 1. | XG Roller++ | 16/11 13/7 | eq: -1.936 | |||
| ||||||
| 2. | XG Roller++ | 13/8 9/3 | eq: -1.953 (-0.016) | |||
| ||||||
| 3. | XG Roller++ | 13/2 | eq: -1.953 (-0.016) | |||
| ||||||
| 4. | XG Roller++ | 16/10 14/9 | eq: -1.953 (-0.016) | |||
| ||||||
| 5. | XG Roller++ | 9/3 6/1 | eq: -1.953 (-0.016) | |||
| ||||||
eXtreme Gammon Version: 2.10, MET: Kazaross XG2