I (too) quickly played the hitting 8/7* 3/1 thinking that the extra net g's/extra losses ratio was > 2/1. The RO shows that ratio to be 19.4/8.99 which is 2.16 and that the hit gives you an extra 0.12% bg's. Hitting is nonetheless a small mistake. Why is that? It has to be because White owns the cube which increases his cubeful winning chances just enough to make not hitting the slightly better play.
Hitting loses more extra games than I thought at a glance. There's the immediate 52 of course. If White dances, 7 numbers blot next time and 54 leaves some flys. If White rolls a 5 to anchor, you'll have 3 blotting numbers.
The lesson here is that the ratio of extra net g's/extra losses has to be > 2/1 when the opponent owns the cube in order for a play that gives you more of each to be correct.
         
 
          
 | | White is Player 2
score: 0
pip: 101 | Unlimited Game
Jacoby Beaver | pip: 51
score: 0
Blue is Player 1 |
|
| XGID=-BCEBaBaA----a-----ccbbb--:1:-1:1:21:0:0:3:0:10 |
| Blue to play 21 |
| 1. | Rollout1 | 8/6 1/Off | eq: +0.919 |
| Player:
Opponent: | 95.67% (G:2.30% B:0.01%)
4.33% (G:0.00% B:0.00%) | Conf.: ± 0.001 (+0.918...+0.921) - [66.8%]
Duration: 1 minute 25 seconds |
|
| 2. | Rollout1 | 8/6 3/2 | eq: +0.919 (0.000) |
| Player:
Opponent: | 95.49% (G:2.45% B:0.01%)
4.51% (G:0.00% B:0.00%) | Conf.: ± 0.002 (+0.917...+0.921) - [33.2%]
Duration: 1 minute 28 seconds |
|
| 3. | Rollout1 | 8/6 2/1 | eq: +0.916 (-0.003) |
| Player:
Opponent: | 95.44% (G:2.30% B:0.00%)
4.56% (G:0.00% B:0.00%) | Conf.: ± 0.001 (+0.915...+0.917) - [0.0%]
Duration: 1 minute 26 seconds |
|
| 4. | Rollout1 | 8/7* 3/1 | eq: +0.910 (-0.010) |
| Player:
Opponent: | 86.68% (G:22.29% B:0.13%)
13.32% (G:0.47% B:0.01%) | Conf.: ± 0.002 (+0.907...+0.912) - [0.0%]
Duration: 1 minute 22 seconds |
|
| |
1 1296 Games rolled with Variance Reduction.
Moves: 3-ply, cube decisions: XG Roller
|
eXtreme Gammon Version: 2.10