No, it is not nearly as predictable as you think it is. There are a ton of other possibilities - for example, in none of your scenarios do you raise the possibility of Louisville missing the extra point, but that's exactly what happened. There could be a safety. A defensive touchdown on the ensuing kickoff. Who knows? There were still eight minutes left, a lot could and did happen.
And your description of what happens if Louisville gets a touchdown is flawed. If Louisville got a touchdown and extra point to get to 28, it wouldn't matter if USF took the two or one - if anything, it is better to take the one. In both cases, USF would need to score a field goal and another touchdown but would need only one two-point conversion. They could take that on the first touchdown, but then if they miss, that means they and the Louisville defense KNOW they have to have two more touchdowns. Taking the one means 11 more points would be needed, leaving a field goal in play. I have always have heard it argued by coaches and analysts that you put the two off till last for this very reason.
And of course, all of this we only know via hindsight. At the moment, all Holtz knew was that there was no reason to assume Louisville's offense wasn't going to score again, so there's no guarantee that getting within a field goal right at that moment would even mean anything, so he initially decided to take the free point. Obviously a yard and a half was enough to change his mind, but USF would have had the lead in the exact same scenario at the end, just leading by one instead of two, which changes nothing.
You forgot the possibilty that Louisville forfeited the game. Seriously, I understand there are other possibilties but if they occur some small percentage of the time that the coach can't plan for them. PATs are like 98% in college, the assumption is always 7 not 6. Just because it happened (we did block the PAT) doesn't mean that it is a good idea to plan on such a thing happening. Same with a safety, I have no idea the odds of it occuring but it's much much lower than the 3 possibilities that I named. Notice in my assumptions I give Louisville 7 for a TD, which is the same assumption that every "go for 2 chart" is based.
I'm not following you here. You would be down either 10,11, or 12. Being down 10 is ideal (got the 2PC), but with so little time on the clock being down 11 or 12 either puts me in a situation where I need 2 TDs or 1TD/2PC and FG which means I'm back to the same thing I'm trying to avoid (going for the 2PC) so I'm really no better off. Unless you are saying that UL knowing that I have to get 2 TDs vs 1TD/2PC/FG would cause them to play a different type of D that would be to USF's advantage?