                |
| ||||
| XGID=-A---------------------b--:0:0:-1:00:0:0:0:13:10 | |||||
| White on roll, cube action? | |||||
| Analyzed in 2-ply | No double | Double/Take |
| Player Winning Chances: | 72.22% (G:0.00% B:0.00%) | |
| Opponent Winning Chances: | 27.78% (G:0.00% B:0.00%) | |
| Cubeless Equities | +0.444 | +0.900 |
| Cubeful Equities | ||
| No double: | +0.444 (-0.456) | |
| Double/Take: | +0.900 | |
| Double/Pass: | +1.000 (+0.100) | |
| Best Cube action: Double / Take | ||
eXtreme Gammon Version: 2.10, MET: Kazaross XG2
Thank you for the clarification.
Consider the variant above. According to both XG (which employs the Rockwell/Kazaross table) and Michael's table, it is an easy take, the winning chances being 27.78%, clearly higher than both XG's and Michael's thresholds at pretty much any even or trailing score.
Imagine I just found some table entitled "Obscura" that tells me that the threshold at even and trailing scores is higher than 27.78%. Could that table be right? Well, let me see. If I pass all eight times, I'll be at 𤂿3, which is 14%. If I take all eight times, I'll win 8*27.78% = 2.22 games. If I win 2 games, I'll be at 9c, which is 5.6%. If I win 3 games, I'll be at 𤪙, which is 23.8%.
Interpolating 22% of the way between 5.6 and 23.8 gives me 9.6% (though it's actually lower, if the curve is accounted for properly). That's worse than 14%, so it is better to pass all eight times. And the Obscura table agrees that it's a pass (based on repeatedly needing more than 27.78% to take); yup, that's confirmation.
Would you say that this reasoning, which supports the Obscura table's numbers (and suggests a flaw in the other two tables), is correct? If not, why not?
Nack