Since I would have made the trapping play (7/4, 5/3) without thinking much I decided to roll this position for money and explore it a bit further with some variants, some are clear but together provide a better understanding of the position.
The rollout (although not long) confirms that there is no need to trap and 6/1 is best.
| | White is Player 2
score: 0 pip: 86 | Unlimited Game Jacoby Beaver | pip: 82 score: 0
Blue is Player 1 | |
XGID=--cBBDCBB--------------eg-:1:-1:1:32:0:0:3:0:10 |
Blue to play 32 |
1. | Rollout1 | 6/1 | eq: +1.088 |
| Player: Opponent: | 95.47% (G:18.69% B:0.37%) 4.53% (G:0.02% B:0.00%) | Conf.: ± 0.002 (+1.086...+1.090) - [100.0%] Duration: 3 minutes 58 seconds |
|
2. | Rollout1 | 6/3 5/3 | eq: +1.072 (-0.016) |
| Player: Opponent: | 95.12% (G:18.09% B:0.35%) 4.88% (G:0.04% B:0.00%) | Conf.: ± 0.003 (+1.069...+1.074) - [0.0%] Duration: 4 minutes 08 seconds |
|
3. | Rollout1 | 7/4 5/3 | eq: +1.057 (-0.031) |
| Player: Opponent: | 93.63% (G:19.97% B:0.50%) 6.37% (G:0.07% B:0.00%) | Conf.: ± 0.003 (+1.054...+1.060) - [0.0%] Duration: 4 minutes 34 seconds |
|
4. | XG Roller++ | 7/5 7/4 | eq: +1.037 (-0.051) |
| Player: Opponent: | 94.17% (G:16.79% B:0.47%) 5.83% (G:0.02% B:0.00%) | |
|
5. | XG Roller++ | 8/3 | eq: +1.032 (-0.056) |
| Player: Opponent: | 92.65% (G:19.89% B:0.53%) 7.35% (G:0.05% B:0.00%) | |
|
|
1 1296 Games rolled with Variance Reduction. Dice Seed: 76510110 Moves: 4-ply, cube decisions: XG Roller
|
eXtreme Gammon Version: 2.10
Variant 1 - I moved all the defensive anchor and prime 1 pip back. The best trapping play (7/5, 7/4) is not far but still keeping the 6 prime is best.
| | White is Player 2
score: 0 pip: 83 | Unlimited Game Jacoby Beaver | pip: 97 score: 0
Blue is Player 1 | |
XGID=---cBBDCBB-------------eg-:1:-1:1:32:0:0:3:0:10 |
Blue to play 32 |
1. | Rollout1 | 7/2 | eq: +0.920 |
| Player: Opponent: | 89.55% (G:15.83% B:0.43%) 10.45% (G:0.24% B:0.00%) | Conf.: ± 0.005 (+0.915...+0.925) - [59.0%] Duration: 37.1 seconds |
|
2. | Rollout1 | 6/1 | eq: +0.919 (-0.001) |
| Player: Opponent: | 89.87% (G:14.98% B:0.37%) 10.13% (G:0.21% B:0.00%) | Conf.: ± 0.005 (+0.915...+0.924) - [40.9%] Duration: 32.4 seconds |
|
3. | Rollout1 | 7/5 7/4 | eq: +0.911 (-0.009) |
| Player: Opponent: | 87.65% (G:19.37% B:0.70%) 12.35% (G:0.47% B:0.00%) | Conf.: ± 0.005 (+0.906...+0.916) - [0.1%] Duration: 45.5 seconds |
|
4. | Rollout1 | 7/4 6/4 | eq: +0.909 (-0.011) |
| Player: Opponent: | 89.21% (G:15.28% B:0.42%) 10.79% (G:0.21% B:0.00%) | Conf.: ± 0.005 (+0.903...+0.914) - [0.0%] Duration: 29.9 seconds |
|
5. | XG Roller++ | 7/5 4/1 | eq: +0.898 (-0.022) |
| Player: Opponent: | 88.79% (G:16.08% B:0.32%) 11.21% (G:0.27% B:0.00%) | |
|
|
1 1296 Games rolled with Variance Reduction. Dice Seed: 76510110 Moves: 3-ply, cube decisions: XG Roller
|
eXtreme Gammon Version: 2.10
Variant 2 - I took variant 1 and moved all the spares 1 pip forward such that we get the bear-in cross-overs only by breaking the prime. The results remain the same.
| | White is Player 2
score: 0 pip: 83 | Unlimited Game Jacoby Beaver | pip: 94 score: 0
Blue is Player 1 | |
XGID=---cBDCBBB-------------eg-:1:-1:1:32:0:0:3:0:10 |
Blue to play 32 |
1. | Rollout1 | 7/2 | eq: +0.914 |
| Player: Opponent: | 87.95% (G:18.90% B:0.65%) 12.05% (G:0.41% B:0.00%) | Conf.: ± 0.005 (+0.908...+0.919) - [71.0%] Duration: 39.1 seconds |
|
2. | Rollout1 | 6/4 5/2 | eq: +0.912 (-0.002) |
| Player: Opponent: | 89.58% (G:15.20% B:0.35%) 10.42% (G:0.20% B:0.00%) | Conf.: ± 0.005 (+0.907...+0.916) - [28.2%] Duration: 33.3 seconds |
|
3. | Rollout1 | 6/1 | eq: +0.906 (-0.007) |
| Player: Opponent: | 89.35% (G:15.14% B:0.36%) 10.65% (G:0.24% B:0.00%) | Conf.: ± 0.005 (+0.902...+0.911) - [0.7%] Duration: 32.1 seconds |
|
4. | Rollout1 | 5/2 4/2 | eq: +0.904 (-0.009) |
| Player: Opponent: | 88.99% (G:15.80% B:0.43%) 11.01% (G:0.26% B:0.00%) | Conf.: ± 0.005 (+0.900...+0.909) - [0.1%] Duration: 27.8 seconds |
|
5. | XG Roller++ | 8/6 5/2 | eq: +0.889 (-0.024) |
| Player: Opponent: | 87.18% (G:18.85% B:0.49%) 12.82% (G:0.33% B:0.00%) | |
|
|
1 1296 Games rolled with Variance Reduction. Dice Seed: 76510110 Moves: 3-ply, cube decisions: XG Roller
|
eXtreme Gammon Version: 2.10
Variant 3 - Going back to the original position and adding 1 more checker to the advanced anchor. Now any play that breaks the prime is a big error.
| | White is Player 2
score: 0 pip: 108 | Unlimited Game Jacoby Beaver | pip: 82 score: 0
Blue is Player 1 | |
XGID=--dBBDCBB--------------ef-:1:-1:1:32:0:0:3:0:10 |
Blue to play 32 |
1. | Rollout1 | 6/1 | eq: +1.348 |
| Player: Opponent: | 97.76% (G:38.80% B:1.18%) 2.24% (G:0.01% B:0.00%) | Conf.: ± 0.003 (+1.345...+1.351) - [98.2%] Duration: 36.6 seconds |
|
2. | Rollout1 | 6/3 5/3 | eq: +1.343 (-0.004) |
| Player: Opponent: | 97.74% (G:38.52% B:1.07%) 2.26% (G:0.01% B:0.00%) | Conf.: ± 0.003 (+1.340...+1.346) - [1.8%] Duration: 34.8 seconds |
|
3. | Rollout1 | 7/5 7/4 | eq: +1.301 (-0.047) |
| Player: Opponent: | 98.08% (G:33.32% B:1.14%) 1.92% (G:0.00% B:0.00%) | Conf.: ± 0.003 (+1.298...+1.304) - [0.0%] Duration: 31.9 seconds |
|
4. | Rollout1 | 8/6 8/5 | eq: +1.296 (-0.052) |
| Player: Opponent: | 98.07% (G:33.01% B:0.98%) 1.93% (G:0.00% B:0.00%) | Conf.: ± 0.003 (+1.294...+1.299) - [0.0%] Duration: 29.2 seconds |
|
5. | Rollout1 | 7/4 5/3 | eq: +1.292 (-0.056) |
| Player: Opponent: | 96.59% (G:35.97% B:1.13%) 3.41% (G:0.04% B:0.00%) | Conf.: ± 0.004 (+1.288...+1.296) - [0.0%] Duration: 35.3 seconds |
|
|
1 1296 Games rolled with Variance Reduction. Dice Seed: 76510110 Moves: 3-ply, cube decisions: XG Roller
|
eXtreme Gammon Version: 2.10
Variant 4 - Same thing with only two checkers back. Now the trapping play is clear.
| | White is Player 2
score: 0 pip: 64 | Unlimited Game Jacoby Beaver | pip: 82 score: 0
Blue is Player 1 | |
XGID=--bBBDCBB--------------eh-:1:-1:1:32:0:0:3:0:10 |
Blue to play 32 |
1. | Rollout1 | 7/4 5/3 | eq: +0.906 |
| Player: Opponent: | 88.29% (G:17.24% B:0.44%) 11.71% (G:0.40% B:0.00%) | Conf.: ± 0.004 (+0.903...+0.910) - [100.0%] Duration: 20.1 seconds |
|
2. | Rollout1 | 6/1 | eq: +0.872 (-0.034) |
| Player: Opponent: | 89.27% (G:11.80% B:0.22%) 10.73% (G:0.20% B:0.00%) | Conf.: ± 0.004 (+0.868...+0.876) - [0.0%] Duration: 14.2 seconds |
|
3. | XG Roller++ | 7/5 7/4 | eq: +0.848 (-0.059) |
| Player: Opponent: | 87.97% (G:12.85% B:0.25%) 12.03% (G:0.17% B:0.00%) | |
|
4. | XG Roller++ | 5/3 4/1 | eq: +0.844 (-0.063) |
| Player: Opponent: | 86.26% (G:16.68% B:0.43%) 13.74% (G:0.66% B:0.00%) | |
|
5. | XG Roller++ | 8/3 | eq: +0.839 (-0.068) |
| Player: Opponent: | 86.54% (G:15.26% B:0.38%) 13.46% (G:0.32% B:0.00%) | |
|
|
1 1296 Games rolled with Variance Reduction. Dice Seed: 76510110 Moves: 3-ply, cube decisions: XG Roller
|
eXtreme Gammon Version: 2.10
Variant 5 - Same original positions but now the roll is 43 such that by clearing the 7 pt you get a great inner-board distribution. The trap play or saefely clearing the 7 pt play are ~0.03 off and keeping the full prime is best.
| | White is Player 2
score: 0 pip: 86 | Unlimited Game Jacoby Beaver | pip: 82 score: 0
Blue is Player 1 | |
XGID=--cBBDCBB--------------eg-:1:-1:1:34:0:0:3:0:10 |
Blue to play 34 |
1. | Rollout1 | 6/3 5/1 | eq: +1.093 |
| Player: Opponent: | 95.58% (G:19.08% B:0.33%) 4.42% (G:0.02% B:0.00%) | Conf.: ± 0.003 (+1.090...+1.096) - [100.0%] Duration: 19.2 seconds |
|
2. | Rollout1 | 7/4 5/1 | eq: +1.062 (-0.031) |
| Player: Opponent: | 93.63% (G:20.43% B:0.49%) 6.37% (G:0.08% B:0.00%) | Conf.: ± 0.003 (+1.059...+1.066) - [0.0%] Duration: 22.0 seconds |
|
3. | Rollout1 | 7/4 7/3 | eq: +1.061 (-0.032) |
| Player: Opponent: | 95.11% (G:17.11% B:0.38%) 4.89% (G:0.02% B:0.00%) | Conf.: ± 0.003 (+1.058...+1.065) - [0.0%] Duration: 16.4 seconds |
|
4. | XG Roller++ | 8/5 8/4 | eq: +1.043 (-0.050) |
| Player: Opponent: | 94.69% (G:16.24% B:0.40%) 5.31% (G:0.01% B:0.00%) | |
|
5. | XG Roller++ | 8/1 | eq: +1.043 (-0.050) |
| Player: Opponent: | 92.98% (G:20.28% B:0.49%) 7.02% (G:0.05% B:0.00%) | |
|
|
1 1296 Games rolled with Variance Reduction. Dice Seed: 76510110 Moves: 3-ply, cube decisions: XG Roller
|
eXtreme Gammon Version: 2.10
To summarize - Before doing this exercise and against a 2 pt board for my opponent I would have trapped with 3 checkers back (or fewer) and kept the prime with 4 checkers back. After this exercise I learned that with 3 checkers back the trapping isn't best when you can keep a full prime because the trade-off between extra G% and win% doesn't justify it. Since it is close you should still be alert to match scores with significantly higher G value.
Note that this conclusion refers to a very narrow part of the ending game. It would be wrong to generalize. For example, it would be very wrong to think the same if all the spares were sent 10 pips back.