I think the opening 64 is the most intriguing opening roll. For money and most match scores any of the three plays are valid. It takes an very lopsided match score to eliminate even one of the plays. Only a few scores have only one valid play according to XG (within .020 equity); Gammon Go, -5 -2, -4 -2 most prominently. Of course, that play is making the 2 point. (Is there anyone out there who refuses to make the 2 point?)
I undertook these rollouts to try to prove a theoretical point: what is the difference between XGR+ and XGR++ rollouts. At least for this move it seems to be not very much.
| | White is Player 2
score: 0 pip: 167 | Unlimited Game Jacoby Beaver | pip: 167 score: 0
Blue is Player 1 | |
XGID=-b----E-C---eE---c-e----B-:0:0:1:64:0:0:3:0:10 |
Blue to play 64 |
1. | Rollout1 | 8/2 6/2 | eq: +0.0149 |
| Player: Opponent: | 49.67% (G:16.09% B:0.93%) 50.33% (G:14.26% B:0.66%) | Conf.: ±0.0020 (+0.0129...+0.0168) - [95.2%] Duration: 17 days 22 hours 44 minutes |
|
2. | Rollout1 | 24/18 13/9 | eq: +0.0124 (-0.0024) |
| Player: Opponent: | 50.31% (G:13.73% B:0.68%) 49.69% (G:13.72% B:0.59%) | Conf.: ±0.0020 (+0.0104...+0.0145) - [4.8%] Duration: 26 days 08 hours 59 minutes |
|
3. | Rollout1 | 24/14 | eq: +0.0086 (-0.0062) |
| Player: Opponent: | 50.53% (G:12.22% B:0.59%) 49.47% (G:13.25% B:0.49%) | Conf.: ±0.0020 (+0.0067...+0.0106) - [0.0%] Duration: 27 days 05 hours 08 minutes |
|
4. | Book2 | 24/20 13/7 | eq: -0.0589 (-0.0738) |
| Player: Opponent: | 48.49% (G:13.89% B:0.65%) 51.51% (G:14.29% B:0.65%) | Conf.: ±0.0074 (-0.0663...-0.0515) - [0.0%] |
|
5. | Book3 | 24/20 24/18 | eq: -0.0760 (-0.0908) |
| Player: Opponent: | 48.71% (G:11.29% B:0.48%) 51.29% (G:14.59% B:0.51%) | |
|
6. | Book3 | 13/9 13/7 | eq: -0.0812 (-0.0961) |
| Player: Opponent: | 47.89% (G:14.01% B:0.71%) 52.11% (G:15.13% B:0.95%) | |
|
7. | 3-ply | 13/3 | eq: -0.0882 (-0.1030) |
| Player: Opponent: | 47.52% (G:13.20% B:0.60%) 52.48% (G:14.40% B:0.76%) | |
|
|
1 46656 Games rolled with Variance Reduction. Dice Seed: 58918962 Moves and cube decisions: XG Roller++
2 Generated by GameSite 2000, Ltd on 2/26/2011 using eXtreme Gammon 2.00 10368 Games rolled with Variance Reduction. Dice Seed: 38613390 Moves: 2-ply, cube decisions: 3-ply Red
3 Generated by GameSite 2000, Ltd on 5/21/2011 using eXtreme Gammon 2.00 Analyzed in XG Roller++
|
eXtreme Gammon Version: 2.10
Sorry 2 point haters, but making the 2 point is the top play albeit not by very much. Note that the equities have changed very little. The point play gained .0039, the split gained .0003 and the run play gained .0001 in equity. I admit that I was expecting something more dramatic. One thing that I take from this and other first and second move rollouts is that computers can never solve the human element in the game. These are three utterly different moves, pushing the game in very different directions and yet all are equal at a broad range of scores. The right answer, of course, is you. What is your style and how you think your skill interacts with your opponent.
Another interesting aspect is the amount of time for each play. The run play took the longest at 3 weeks 6 days 5 hours 8 minutes, the split play a little shorter at 3 weeks 5 days 8 hours 59 minutes, and the point play 2 weeks 3 days 22 hours 59 minutes. That's over a week less. I assume that is just cube action. The total run time was 10 weeks 1 day 12 hours 51 minutes on a 10 core running a little about 15.2 million positions per second. I ran these one unit of rollouts (1296) at a time, taking a little under 2 days then restart because of memory issues and continue.
And for comparison here is XGR+ analysis by Neil Kazaross.
| | White is Player 2
score: 0 pip: 167 | Unlimited Game Jacoby Beaver | pip: 167 score: 0
Blue is Player 1 | |
XGID=-b----E-C---eE---c-e----B-:0:0:1:64:0:0:3:0:10 |
Blue to play 64 |
1. | Book1 | 24/18 13/9 | eq: +0.0127 |
| Player: Opponent: | 50.38% (G:13.61% B:0.64%) 49.62% (G:13.55% B:0.58%) | Conf.: ±0.0022 (+0.0105...+0.0149) - [87.0%] |
|
2. | Book1 | 8/2 6/2 | eq: +0.0110 (-0.0017) |
| Player: Opponent: | 49.76% (G:15.86% B:0.87%) 50.24% (G:14.29% B:0.65%) | Conf.: ±0.0020 (+0.0090...+0.0129) - [12.9%] |
|
3. | Book1 | 24/14 | eq: +0.0085 (-0.0042) |
| Player: Opponent: | 50.60% (G:11.94% B:0.57%) 49.40% (G:13.10% B:0.47%) | Conf.: ±0.0020 (+0.0064...+0.0105) - [0.1%] |
|
4. | Book2 | 24/20 13/7 | eq: -0.0589 (-0.0716) |
| Player: Opponent: | 48.49% (G:13.89% B:0.65%) 51.51% (G:14.29% B:0.65%) | Conf.: ±0.0074 (-0.0663...-0.0515) - [0.0%] |
|
5. | Book3 | 24/20 24/18 | eq: -0.0760 (-0.0886) |
| Player: Opponent: | 48.71% (G:11.29% B:0.48%) 51.29% (G:14.59% B:0.51%) | |
|
6. | Book3 | 13/9 13/7 | eq: -0.0812 (-0.0939) |
| Player: Opponent: | 47.89% (G:14.01% B:0.71%) 52.11% (G:15.13% B:0.95%) | |
|
7. | 3-ply | 13/3 | eq: -0.0882 (-0.1008) |
| Player: Opponent: | 47.52% (G:13.20% B:0.60%) 52.48% (G:14.40% B:0.76%) | |
|
8. | 3-ply | 24/18 6/2 | eq: -0.1232 (-0.1359) |
| Player: Opponent: | 47.36% (G:11.18% B:0.53%) 52.64% (G:15.26% B:0.68%) | |
|
|
1 Generated by Neil Kazaross on 12/17/2010 using eXtreme Gammon 2.00 46656 Games rolled with Variance Reduction. Dice Seed: 41387960 Moves and cube decisions: XG Roller+
2 Generated by GameSite 2000, Ltd on 2/26/2011 using eXtreme Gammon 2.00 10368 Games rolled with Variance Reduction. Dice Seed: 38613390 Moves: 2-ply, cube decisions: 3-ply Red
3 Generated by GameSite 2000, Ltd on 5/21/2011 using eXtreme Gammon 2.00 Analyzed in XG Roller++
|
eXtreme Gammon Version: 2.10