At DMP, 3/5/7 is going to be better, sometimes significantly so, when White has two checkers on the bar, you're massively ahead in the race, and getting hit is fatal (or at least very bad).
That would describe your reference position itself, wouldn't it? At DMP, 10/3 is worth .011, which is still pretty small (at my level of play at least) -- and I'm having a hard time coming up with a variation worth more than .020 . . .
Isn't the 6/5/4 distribution only "ideal" for bearing off fastest (when the gammon is in question, or when your opponent has a bunch of checkers borne off)? But I'd think with the race essentially gin and gammons irrelevant, we generally play to clear the 6-point asap rather than putting a spare on it, so leaving the 3/5/7 distribution would be the intuitively appealing "safe" play?
      
 
          
 | | White is Player 2
score: 3
pip: 139 | | 5 point match | pip: 64
score: 3
Blue is Player 1 |
|
| XGID=bBBBBCBA--A-c-----bbbbb---:1:-1:1:16:3:3:0:5:10 |
| Blue to play 16 |
| 1. | Rollout1 | 10/3 | eq: +0.956 |
| Player:
Opponent: | 97.79% (G:87.01% B:0.28%)
2.21% (G:0.00% B:0.00%) | Conf.: ± 0.002 (+0.954...+0.958) - [100.0%]
Duration: 16.4 seconds |
|
| 2. | Rollout1 | 10/4 7/6 | eq: +0.945 (-0.011) |
| Player:
Opponent: | 97.26% (G:88.25% B:0.51%)
2.74% (G:0.00% B:0.00%) | Conf.: ± 0.002 (+0.943...+0.947) - [0.0%]
Duration: 18.3 seconds |
|
| 3. | Rollout1 | 10/9 7/1 | eq: +0.938 (-0.018) |
| Player:
Opponent: | 96.89% (G:86.12% B:0.29%)
3.11% (G:0.00% B:0.00%) | Conf.: ± 0.002 (+0.936...+0.940) - [0.0%]
Duration: 13.1 seconds |
|
| 4. | 4-ply | 10/4 5/4 | eq: +0.895 (-0.061) |
| Player:
Opponent: | 94.73% (G:84.88% B:0.54%)
5.27% (G:0.00% B:0.00%) | |
|
| 5. | 3-ply | 7/1 5/4 | eq: +0.896 (-0.060) |
| Player:
Opponent: | 94.81% (G:84.80% B:0.45%)
5.19% (G:0.01% B:0.00%) | |
|
| |
1 1296 Games rolled with Variance Reduction.
Moves: 3-ply, cube decisions: XG Roller
|
eXtreme Gammon Version: 2.10, MET: Kazaross XG2