The main question was was 24/18(2) correct. It also through them off in the chouette that Snowie didn't even evaluate it on 3 ply so they thought it may be atrocious. The actual fault lays in Snowie's move filter, they weren't aware of the workaround.
I felt it was clearly wrong to play 24/18(2), giving up the long term equity of the acepoint game and though I didn't want to get too close to statistical significance because then that would crush Chuck's hopes & dreams for his play, I feel as confident as I did before this rollout that giving up the acepoint is dead wrong.
| Stick (Green) vs. Patrick Gibson (White) Money Game Session |
| Move 2 Green |                                  | Pip: 91 |
| Game 1 Money session Green-White: Score 0-0 | ||
| Pip: 182 |
| # | Ply | Move | Equity | |
|---|---|---|---|---|
| 1 | R | 21/18(2) 13/7 | -0.251 | |
| 0.3% 5.2% 42.5% 57.5% 27.8% 3.2% | ||||
| 95% confidence interval: - money cubeless eq.: -0.406 ±0.020. Rollout settings: Full rollout, 648 games (equiv. 14925 games), played 3-ply (precise), random seed, with race database. | ||||
| 2 | R | 13/4 10/7 | -0.287 (-0.036) | |
| 0.2% 4.1% 42.1% 57.9% 29.1% 3.4% | ||||
| 95% confidence interval: - money cubeless eq.: -0.440 ±0.018. Rollout settings: Full rollout, 648 games (equiv. 20353 games), played 3-ply (precise), random seed, with race database. | ||||
| *3 | R | 24/18(2) | -0.323 (-0.072) | |
| 0.2% 4.8% 39.1% 60.9% 29.4% 0.3% | ||||
| 95% confidence interval: - money cubeless eq.: -0.466 ±0.017. Rollout settings: Full rollout, 648 games (equiv. 23332 games), played 3-ply (precise), random seed, with race database. | ||||