THE PARABLES OF ECONOMICS. Professor Farrell, now operating under the banner of Gallowglas (he explains better than I can -- test site here) has some thoughtful responses to my shilling for economics.
First, an invitation to any reader from any other social science discipline: please provide me with fundamental parables for your field. (If you wish to argue that the field is too complex and subtle to be so treated, read and understand this and recall an aphorism attributed variously to P. Samuelson, D. McCloskey, or R. Solow: if you can't explain it to your mother-in-law, you don't understand it well enough.)
Now, to the game theory. Professor Farrell cites this paper as studying "certain aspects of social interaction that simply cannot be represented using conventional game theory." I beg to differ. It is an interesting paper, and check some of the names in the references, but modeling the evolution of this convention as a game has potential as a dissertation topic (or perhaps a Larry Samuelson student, or a Larry Kranich student -- two different approaches to evolutionary game theory within economics -- is working on a related problem.)
Turning to the topic of "restrictive equilibrium refinements" for the most part those are somewhat more advanced parables (not quite rocket science, where have you gone, Darius Gaskins?) but still pretty simple parables. Take "subgame perfection." Think about it this way. You're playing chess. You calculate some variations. Looks like a pretty good move except for one nasty case. Do you make the move and hope your opponent doesn't see the counter? Not if you expect to be paired up in the next round. Or take those "alternative equilibria." Please. Not if you expect to get it published. You don't leave $100 on the sidewalk and hope nobody notices. That's yet another simple parable: arbitrage profits do not go unexploited. If you write it with an unexploited arbitrage opportunity or a non-credible threat, and you don't see it, and the referee doesn't see it, and none of your readers see it, then perhaps it has gotten through as a non-science.
Where there might be an art form to game theoretic modelling is in recognizing when to disregard one of the parables. But there's a parable for that, too: sometimes, you don't pick up the blind with a hand full of trumps.