How bot errors can skew rollout results. I expect I've run loads of rollouts where this sort of thing is happening and haven't spotted it. How do you?
Interestingly the difference between the rollouts seems to be caused by XG taking the recube on what's around a 1.2 pass after 21/13 dance, while GNU passes (they are both using G11). I guess that 21/13 is the best practical move.
| # | Ply | Move | Equity |
| • | 1 | R | 21/16 9/6 | +0.7056 |
| | | 0.6339 | 0.3345 | 0.0070 | - | 0.3661 | 0.1030 | 0.0069 | +0.5122 | +0.7056 | | 0.0004 | 0.0006 | 0.0002 | - | 0.0004 | 0.0006 | 0.0003 | 0.0012 | 0.0017 | | |
| | Full cubeful rollout with var.redn. | |
| | 7776 games, Mersenne Twister dice gen. with seed 861628715 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 | 21/13 | +0.7042 ( -0.0014) |
| | | 0.6136 | 0.3035 | 0.0058 | - | 0.3864 | 0.0994 | 0.0052 | +0.4424 | +0.7042 | | 0.0003 | 0.0005 | 0.0002 | - | 0.0003 | 0.0005 | 0.0002 | 0.0010 | 0.0015 | | |
| | Full cubeful rollout with var.redn. | |
| | 7776 games, Mersenne Twister dice gen. with seed 861628715 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] | |