Ok, not 4th down decisions exactly, but he hits the nail on the head here (via David Brooks):
People who have information about an individual case rarely feel the need to know the statistics of the class to which the case belongs.
Coaches want to know all about the particulars of certain situations: how fatigued the defense is, how the left guard has been playing today, how well the team has done in short yardage so far (sample size of 3 plays probably), how good the turf's traction is, and so on. But many coaches remain generally (and willfully) ignorant of the statistics of their larger situation.
Sure, all those little things matter, but not nearly as much as the 'class to which the case belongs', as Kahneman puts it.
Case in point.
I tend to take the cynical view that experts don't like data, because the data can show how little of an 'expert' they actually are.
Take the the 4th down case. Head coaches and co-ordinators are miles better than you or I at preparing the players, designing plays and organising their teams and I'd never suggest that I could do a better job overall than they do. None of this, though, qualifies them to make a call that is a simple balancing of expected values.
If you want to know whether you should go for it on 4th down, you should ask the bookie, not the coach.