         
 
         
 | | White is Player 2
score: 0
pip: 159 | Unlimited Game
Jacoby Beaver | pip: 164
score: 0
Blue is Player 1 |
|
| 21$-62S-64 |
| 1. | XG Roller++ | 13/7* 11/7 | eq: +0.1617 |
| Player:
Opponent: | 53.68% (G:15.25% B:0.75%)
46.32% (G:12.17% B:0.65%) | |
|
| 2. | XG Roller++ | 24/20 11/5 | eq: +0.1490 (-0.0127) |
| Player:
Opponent: | 52.97% (G:14.00% B:0.51%)
47.03% (G:9.80% B:0.36%) | |
|
| 3. | XG Roller++ | 13/7* 5/1* | eq: +0.0791 (-0.0826) |
| Player:
Opponent: | 51.15% (G:17.51% B:0.64%)
48.85% (G:13.78% B:0.83%) | |
|
| 4. | XG Roller++ | 13/9 11/5 | eq: +0.0538 (-0.1079) |
| Player:
Opponent: | 50.45% (G:14.11% B:0.48%)
49.55% (G:11.27% B:0.51%) | |
|
| 5. | XG Roller++ | 11/5 8/4 | eq: +0.0479 (-0.1138) |
| Player:
Opponent: | 50.23% (G:14.55% B:0.50%)
49.77% (G:11.69% B:0.56%) | |
|
| 6. | XG Roller++ | 24/14* | eq: +0.0367 (-0.1250) |
| Player:
Opponent: | 51.25% (G:12.99% B:0.55%)
48.75% (G:13.38% B:0.77%) | |
|
eXtreme Gammon Version: 2.10
Thanks for checking, Mike. I'm often curious about some of the moves in my would-be well-played retro-sequences. For anyone who doesn't know, a "well-played" move is any move that is best or within .02 of best (in an unlimited/money game).
For my retro-sequence of this thread (21$-62S-64P-55B-64H-51H-44), Mike brought two moves into the spotlight:
(1) The third roll play by Blue, diagrammed above. For the top six plays, XG's evaluation is [P C13 K83 O108 J114 H125] "&e. That is, according to XG evaluation, the move played, P (Point, 13/7* 11/7) beats the second best move C (Cross, 24/20 11/5) by .013. This gives P an eval cushion of .013 + .020 = .033. For purposes of verifying that 21$-62S-64P is "well played," I haven't bothered to roll this position out (though I am interested if anyone has).
(2) The fourth roll play by White, diagrammed below. Here, XG's eval is [P B15] "&e. That is, the eval has best play P (Point, bar/20 13/3 8/3) ahead of the play made B (Both, bar/15* 13/8) by .015. Here, my cushion is only .005 (as the "well played" threshold is -.020). Fortunately, the rollout result shown below vindicates: [P B12] "<=20, which clears the threshold by .008.
As I recall, the positions in New Ideas in Backgammon (published in 1996) are taken from actual games played in the '80s and early '90s. At the time, there still remained a bit of a bias against making deep points (3pt or lower). Unless there was a fourth checker already on the 8pt, it was very common, even when hitting, to play 55 as B (Both, bar/10 13/8) rather than P (Point, bar/20 13/3 8/3) in such positions. That increases the odds that NIB #13 was actually reached by 21$-62S-64P-55B-64H-51H-44 (though we'll probably never know unless Kit or Hal still has records of the original games).
Nack
 
| | White is Player 2
score: 0
pip: 166 | Unlimited Game
Jacoby Beaver | pip: 154
score: 0
Blue is Player 1 |
|
| 21$-62S-64P-55 (White on roll) |
| 1. | Rollout1 | Bar/20* 13/3 8/3 | eq: +0.1955 |
| Player:
Opponent: | 53.18% (G:19.05% B:1.53%)
46.82% (G:11.53% B:0.58%) | Conf.: ±0.0043 (+0.1912...+0.1998) - [100.0%]
Duration: 4 hours 19 minutes |
|
| 2. | Rollout1 | Bar/20* 20/10 13/8 | eq: +0.1836 (-0.0119) |
| Player:
Opponent: | 53.69% (G:16.91% B:1.10%)
46.31% (G:11.75% B:0.52%) | Conf.: ±0.0041 (+0.1795...+0.1877) - [0.0%]
Duration: 4 hours 40 minutes |
|
| |
1 20736 Games rolled with Variance Reduction.
Dice Seed: 81920408
Moves: 3-ply, cube decisions: XG Roller
|
eXtreme Gammon Version: 2.10