With all due respect, I think that everyone on the board knows the difference between what people COULD do vs. what they ACTUALLY do. The question is how do you go about answering the question of the OP, which Kaitainen restated quite precisely as:
"what percentage of men are born with the talent to be able to run a 5 minute mile under ideal conditions?"
The key thing is, how do you go about answering that question? What counts as evidence?
Mandingo, A-10, the 5:00 female runner, some others, and I all opted for conjectures based on an empirical approach derived from direct experience, observation, anecdote, vo2 max requirements, the relative incidence of sub 5 high schoolers, data from the 5th Ave mile, and some elementary statistical reasoning.
Our estimates ranged from 1% to 14% (ballpark figures), or 1/100, 1/35, 1/20, 1/7. That's still a pretty wide range if you put it in those terms. Could very well be wrong but I don't think it's "loony-speak". It's certainly a long way from the 50% - 80% that you and others advocate.
I'm sorry if it has escaped me but I don't recall any specific evidence that you have brought to bear on this discussion? You might (probably) have perfectly good judgment on this subject; but reasons and evidence are better than appeals to (personal) authority, or subjective memories of what high school classmates could have done. And, to repeat, I am perfectly aware of the distinction between actual and potential.
Which leads me to a possible way out of this impasse. Rather than argue from experience or estimates of innate potential, which can verge on the metaphysical, perhaps we need to identify the physical and physiological attributes that go with running a 5:00 mile. Things like vo2 max, ft/st muscle fibers, running economy, height to weight ratio, etc. I don't know; I'm not a scientist. (For the sake of argument, we can assume ideal motivation, although in real life we know that mental toughness is important.)
Then it becomes a question for the physiologists like Jack Daniels: What attributes are necessary for running a 5 minute mile OR what attributes have been correlated with a sufficiently representative sample of 5 minute milers (different but related).
Once these traits are identified, then simply count the proportion of the population that has them, fit a regression line, and you will have a reasonably close estimate answering the original question.
Surely there is some research out there already. I'm open to changing my mind.