Think about this though, taking into account the stat you quoted about the coin flip. This is what I'm talking about with the coin flip being determinative. If the coin flip is the most determinative - to the tune of an 8% advantage - the "odds" of winning in overtime come down to the coin flip, which is the 50-50 proposition. In other words, you list out a bunch of factors which you say come into play before the coin flip - but then if the coin flip is the most determinative factor, do the odds before it mean anything?
Basically, all that other stuff you listed out sounds nice and rational, and makes a certain amount of sense. I'd wager that's the sort of thing most coaches are thinking about when they make the "play for OT" decision. But they just don't matter all that much.
That is a good point. Since in the college game the coin-flip makes a pretty small difference, I would guess that other factors may be more important when making the decision. In the NFL it is certainly more influenced by chance, which may influence decision making.
Ultimately I think a lot of coaches use the mindset of "which decision will get me raked over the coals less," or "which decision will prolong the loss for the greatest duration of time." I don't like this approach, but it is easy to criticize when you are not the one getting raked over the coals when aggressive decisions go wrong.
An aggressive mind-set with a great risk-benefit analysis is the best of all worlds (see Bill Belichick). However, there are not a lot of Belichick's out there.