This may have circulated widely already....
But I do want to share the link for this nice game theory analysis of toilet seat position.
It seems to me there are two costs left out here.
The first is the fact that if wife "Marsha" doesn't notice the seat is up, in the dark, in the middle of the night, she either sits on the cold porcelain (bad) or drops several inches into the water, with her legs sticking out in front of her (much worse). So, an asymmetry based on risk.
The second cost is the cost of "John" being celibate for at least a week if "Marsha" incurs the first cost, above. Forget "she yells at him." It's more like "she reminds John that she controls valuable easements, involving access."
(Nod to BN, who I'm sure always puts the toilet seat down. But then he sits down for #1, so it doesn't cost him much...)
UPDATE: lisalogic's comment is worth posting directly:
Very nice, the model might still be extended, though. Let's say, C(t) is constant, but D(t) is something like a mean-reverting process or else a poisson process (I call this the riposte model). How would the equilibrum behave in time, then?
Anyway, while it is logic to model a cooperative game and argue that one of the equilibrums is more efficient, and thus increases public welfare, it is not very much so for a non-cooperative game. As a woman would put it: "More welfare - for You".
Two good points there: design the game so it is in the male's advantage to choose the more globally efficient outcome. And, the dynamic properties of the game are interesting. If there is not a single equililbrium, what do we expect for the time path of behavior?