I would have liked an opinion on the fact that the hit despite the difference in roll odds is correct by a huge margin. Where is the additional 10% coming from? is it the better bear of structure, the improved Gammon chances, the double jeopardy factor and more perhaps.
I believe this refers to the question of why 55 is better played 13/3* 7/2(2) instead of 7/2(2) 6/1(2), even though it gives White 11/36 hitting numbers instead of 9/36.
The score is: White 2, Blue 5 (match to 7 points). Blue to play 55
| White | 104 | ||
|---|---|---|---|
                        | |||
| Blue | 68 | ||
| Position ID: 27YBgAl35g0EAA Match ID: QYn2ACAAKAAE | |||
• Blue moves 7/2(2) 6/1(2)
Alert: very bad move ( -0.568)
Alert: unlucky roll! ( -0.508)
| # | Ply | Move | Equity | |||||||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 1 | R | 13/3* 7/2(2) | -0.2442 | |||||||||||||||||||
| ||||||||||||||||||||||
| • | 2 | R | 7/2(2) 6/1(2) | -0.8126 ( -0.5684) | ||||||||||||||||||
| ||||||||||||||||||||||
First because merely counting nonhitting and hitting rolls does not identify the good and bad responses in the problem position, most obviously because 13/3* 7/2(2), in exchange for allowing 2/36 additional hitting numbers (bad for Blue), turns 16/36 nonhitting but playable responses into 16/36 fanning rolls (good for Blue).
Second because counting good and bad rolls is a shortcut that works well in simple positions where good = win and bad = loss, but what really matters, always, is not how many good and bad numbers there are, but their total equity, and that depends on how good or bad each of them is.
In the problem position, all hitting responses are obviously equally bad. But the nonhitting responses are not equally good. After 13/3* 7/3(2) and any of the 25/36 nonhitting responses, Blue would stand so well that if he had not doubled he would be too good to double. But after 7/2(2) 6/1(2) and any of the 27/36 nonhitting responses, Blue would not even be good enough to double.