I watched this match on the internet and later analyzed it. I don't know why the cube had not been turned before; probably a missed double a few moves ago.
And yes, in the position where 61 was rolled it was ~1.14 too good to double.
So, now for the strange bot eval that cought my intention when I looked closer to the analysis:
| # | Ply | Move | Equity |
| | 1 | 2 | 12/11 7/1 | +0,965 |
| | 0,656 0,308 0,007 - 0,344 0,053 0,001 | |
| | 2 | 2 | 11/4 | +0,951 ( -0,013) |
| | 0,713 0,354 0,008 - 0,287 0,045 0,001 | |
| | 3 | 2 | 12/6 11/10 | +0,945 ( -0,020) |
| | 0,704 0,347 0,007 - 0,296 0,045 0,001 | |
| | 4 | 2 | 12/6 7/6 | +0,930 ( -0,035) |
| | 0,683 0,330 0,007 - 0,317 0,032 0,001 | |
What I don't understand here is that 11/4 wins more games, more gammons and loses fewer gammons than 12/11 7/1 but it is ranked behind this move. Why?
| # | Ply | Move | Equity |
| | 1 | 4 | 11/4 | +0,951 |
| | 0,723 0,353 0,007 - 0,277 0,046 0,001 | |
| | 2 | 4 | 12/11 7/1 | +0,934 ( -0,017) |
| | 0,654 0,308 0,006 - 0,346 0,059 0,002 | |
| | 3 | 4 | 12/6 11/10 | +0,915 ( -0,036) |
| | 0,705 0,348 0,007 - 0,295 0,049 0,001 | |
| | 4 | 4 | 12/6 7/6 | +0,903 ( -0,047) |
| | 0,683 0,328 0,006 - 0,317 0,035 0,001 | |
What I don't understand here is that 12/6 11/10 wins more games, more gammons and loses fewer gammons than 12/11 7/1 but it is ranked behind this move. Why?
| # | Ply | Move | Equity |
| | 1 | R | 11/4 | +0,886 |
| | | 0,708 | 0,328 | 0,005 | - | 0,292 | 0,047 | 0,002 | +0,654 | +0,886 | | 0,001 | 0,002 | 0,000 | - | 0,001 | 0,001 | 0,000 | 0,003 | 0,005 | | |
| | Full cubeful rollout with var.redn. | |
| | 1296 games, Mersenne Twister dice gen. with seed 894061949 and quasi-random dice | |
| | Play: world class 2-ply cubeful prune [world class] | |
| | keep the first 0 0-ply moves and up to 8 more moves within equity 0,16 | |
| | Skip pruning for 1-ply moves. | |
| | Cube: 2-ply cubeful prune [world class] | |
| | 2 | R | 12/6 11/10 | +0,860 ( -0,026) |
| | | 0,696 | 0,316 | 0,005 | - | 0,304 | 0,051 | 0,002 | +0,617 | +0,860 | | 0,001 | 0,002 | 0,000 | - | 0,001 | 0,002 | 0,001 | 0,003 | 0,005 | | |
| | Full cubeful rollout with var.redn. | |
| | 1296 games, Mersenne Twister dice gen. with seed 894061949 and quasi-random dice | |
| | Play: world class 2-ply cubeful prune [world class] | |
| | keep the first 0 0-ply moves and up to 8 more moves within equity 0,16 | |
| | Skip pruning for 1-ply moves. | |
| | Cube: 2-ply cubeful prune [world class] | |
| | 3 | R | 12/11 7/1 | +0,838 ( -0,048) |
| | | 0,637 | 0,283 | 0,009 | - | 0,363 | 0,061 | 0,004 | +0,535 | +0,838 | | 0,001 | 0,001 | 0,001 | - | 0,001 | 0,002 | 0,001 | 0,005 | 0,006 | | |
| | Full cubeful rollout with var.redn. | |
| | 1296 games, Mersenne Twister dice gen. with seed 894061949 and quasi-random dice | |
| | Play: world class 2-ply cubeful prune [world class] | |
| | keep the first 0 0-ply moves and up to 8 more moves within equity 0,16 | |
| | Skip pruning for 1-ply moves. | |
| | Cube: 2-ply cubeful prune [world class] | |
| | 4 | R | 12/6 7/6 | +0,837 ( -0,049) |
| | | 0,670 | 0,301 | 0,006 | - | 0,330 | 0,042 | 0,002 | +0,575 | +0,837 | | 0,001 | 0,001 | 0,000 | - | 0,001 | 0,002 | 0,001 | 0,003 | 0,005 | | |
| | Full cubeful rollout with var.redn. | |
| | 1296 games, Mersenne Twister dice gen. with seed 894061949 and quasi-random dice | |
| | Play: world class 2-ply cubeful prune [world class] | |
| | keep the first 0 0-ply moves and up to 8 more moves within equity 0,16 | |
| | Skip pruning for 1-ply moves. | |
| | Cube: 2-ply cubeful prune [world class] | |
What I don't understand here is that 12/6 7/6 wins more games, more gammons and loses fewer gammons than 12/11 7/1 but it is ranked behind this move. Why?
| 1. | Rollout1 | 11/4 | eq: +0,892 |
| Player: Opponent: | 71,10% (G:33,05% B:0,71%) 28,90% (G:4,29% B:0,30%) | Conf: ± 0,007 (+0,885...+0,899) Duration: 30 minutes 55 seconds |
|
| 2. | Rollout1 | 11/10 7/1 | eq: +0,861 (-0,031) |
| Player: Opponent: | 66,04% (G:28,47% B:0,98%) 33,96% (G:5,50% B:0,43%) | Conf: ± 0,013 (+0,848...+0,874) Duration: 45 minutes 43 seconds |
|
| 3. | Rollout1 | 12/6 11/10 | eq: +0,860 (-0,032) |
| Player: Opponent: | 70,27% (G:31,28% B:0,75%) 29,73% (G:4,81% B:0,35%) | Conf: ± 0,011 (+0,849...+0,871) Duration: 31 minutes 27 seconds |
|
| 4. | Rollout1 | 12/6 7/6 | eq: +0,841 (-0,050) |
| Player: Opponent: | 67,05% (G:30,18% B:0,71%) 32,95% (G:3,90% B:0,18%) | Conf: ± 0,010 (+0,831...+0,851) Duration: 31 minutes 00 second |
|
| 5. | Rollout1 | 12/11 7/1 | eq: +0,824 (-0,068) |
| Player: Opponent: | 63,49% (G:27,27% B:0,77%) 36,51% (G:5,24% B:0,28%) | Conf: ± 0,010 (+0,814...+0,834) Duration: 5 minutes 18 seconds |
|
| |
1 1296 Games rolled with Variance Reduction. Moves and cube decisions: 3 ply Search interval: Large
|
And finally, as someone suggested, with the cube on 2 it is right to go for the gammons:
Any comments are appreciated. Thanks