Black to play down 1-0 in match to 7.

Clearly the best play of 5-2 is to maximize the chances for gammon. So black should fight for control of the ace point instead of the obvious 13/8, 9/7 which completes the six point prime. Assuming this is correct theory, then playing 13/11, 9/4 would miss with six combinations while playing 13/6 would miss with seven combinations. If red re-enters next roll, then 13/11, 9/4 also completes the prime with more combinations than 13/6.

However, if you knew red was going to re-enter, then 13/11, 13/8 only fails to complete the prime with 5-3 whereas 13/11, 9/4 fails with nine combinations. There is some additional risk of allowing the miracle 1-1 for red, but that still requires parlay of a bad follow-up roll from black to seriously open the door.

I'm definitely playing 13/11, 9/4 here, but have a question anyway. If the advantage of 13/11, 13/8 is more chances to complete the prime assuming red does re-enter, why would Gnubg rate this third best play significantly better than 13/8, 9/7 which does so immediately? I'm finding it hard to believe it is really better to leave the door open a crack to the miracle 1-1 for a few more chances to hit, rather than to simply complete the prime.

ASSUMING he enters, you came to a certain conclusion. But this only happens 11/36 of the time. That means that 25/36 of the time, red is stuck with a blot, and you're much better off in those situations having more ways to hit loose on the ace point instead of fewer. You gammon chances go down significantly when he anchors, so any time you're able to hit loose, it's worth quite a bit to you.

After 13/8 9/7, you hit on 6x, 51, 42, 33, 22, 43, 52 (21 shots), but after 13/8 13/11, you've got all those rolls PLUS 55, 44, 53 (25 shots), which is almost a 20% increase in the number of shots.

So it would appear that adding even one extra chance to hit loose in such situations is worth the exposure to a fluke 1-1 because red is only about 30% to re-enter immediately. This makes more sense now, thanks.

Over the board I would play 13/6 without much thought, despite the 1 fewer hitting combinations. It doesn't pay off to 1-1, and the distribution after 9/4 13/11 looks a little funky. No big deal though. The prime doesn't matter here because you'll have to take it down so quickly. The idea is to hit loose on the ace if you can, otherwise get your builders to the 4, 5, and 6-points for the bearoff.

With this in mind, is 9/4 7/5 a viable option?
Opponent still has to roll two 1's or 1/1 to cause me any concern for the 3 checkers on the 13 pt.
[my copy of GNU tends 2 luckbox and hit miracle 1-6's and stuff ]

I'd like to at least keep the opportunity to complete the prime if red does establish an anchor, even though it can break down quickly. If I wind up having to leave a blot on the midpoint with a 6-1 at the wrong moment, being primed can make this much less awkward. Playing 13/6 or 13/11, 9/4 keeps the possibilities for making the bar point alive. With 9/4, 7/5 you are turning the hits with a six, into hits with a four which is no better than bringing one attacker in from the nine point.