This has happened to me many times. A great suggestion +1 for me. It ties in to a requested enhancement I made a couple of years ago for a focus quality factor, which I will summarize here:
Why not display a focus quality measurement. This is most likely already a number being calculated internally by SGP when it determines the best fit for the intersecting legs of the focus V curve. The enhancement would:
a) display this Focus Quality Number for each focus run
b) allow a minimum focus quality value to be set, default to none.
c) if the actual focus quality value is worse than the minimum chosen, rerun focus X times.
I would think that this could be implemented very easily and be fairly easy to debug.
This is less of problem for me now that I am only running 1 refractor, since almost all of my focus runs are, what you might call, perfect. A usual 10 hour night of imaging gives me about 7 close to perfect focus runs, none of which I would choose to rerun. However, even now, about every other night I have one that if I am watching it, I rerun using the manual (currently the only available option) process described by @CCDMan.
However, when I was using the RC12 it was really a big problem. Probably 1 out of every 3 focus runs were bad enough to require a rerun. Since I am adding a 12" Newton to my rig this spring, that issue may return.
This would be a very welcome and simple enhancement.