The following is, unfortunately, a very common experience'o'mine:
Chess book. Grandmaster game. Diagram.
I (after 1 microsecond): Play Nxe6, willya?
scrolldown, author: "Nxe6 was the decisive mistake." Oh bother.
Now, it is a very obvious explanation that mistakes made even by a GM must be very tempting mistakes for a FM patzer like me too, but "obvious" doesn't mean "proven" (and moreover, my style could be a critical factor here too). Thus this is clear science fodder: Collect 100 GM and 100 patzer blunders, give the positions to a GM and a patzer group, count how many of them get repeated. Has such an experiment (or, more general, a treatise on what makes a blunder tempting) ever been done? (Books specializing on blunders exist, for starters, but rather tend to research the reasons.)