I first encountered the Prisoner’s Dilemma in a middle school seminar, where it was presented as a morality problem. Do you potentially doom your unknown partner in order to improve your own odds? After seeing it presented with a similar framing soon afterward in Knights of the Old Republic, and as a case of social dynamics in Roger Highfield’s The Science of Harry Potter, I figured it was just a moral dilemma in the sense of questions like “do you let yourself die to save someone else?” and didn’t think of it much afterward.

The next time I encountered the Prisoner’s Dilemma was many years later, in an essay about how the Prisoner’s Dilemma should decidedly not be viewed as a moral issue in the traditional sense. You’re supposed to think of your partner either as someone entirely morally irrelevant, or assume that any bad karma points that you would get for defecting are reflected in the payout. Viewed this way, the problem is still at least as interesting.

Meanwhile, I was taking a microeconomics class, and the Prisoner’s Dilemma was introduced as an illustration of the idea of a Nash equilibrium (named after mathematician John Nash, who died last week). This was the first time I’d really grasped the paradoxical nature of the situation–no matter what your opponent does, you’re better off defecting, but at the same time, both players defecting is a preventably bad outcome. Without any sort of coordinating force, the most stable outcome manifests in a state that’s bad for everyone, one that no one would consciously choose given that a better one exists, but one that is very hard to get out of.

Around the same time, I ran into the problem in real life–this time in my college dorm.

I had a roommate who liked to watch TV at his desk with the speakers on, and while he was in our room, he would usually be watching TV. I, meanwhile, tended to play video games at my desk while wearing headphones (I didn’t own a TV, man). The headphones got uncomfortable after a while, and I would have preferred to play with my speakers on instead. However, if I did this, both of our sounds would partially block out the other’s, instead of just his blocking out mine. This would have been a better state of affairs, but I didn’t want to get into a battle of going back and forth raising our respective volumes forever.

I realized that this sort of worked as a version of the Prisoner’s’ Dilemma:

Headphones Speakers
Headphones Decent outcome Bad outcome, good outcome
Speakers Good outcome, bad outcome Bad-ish outcome

In retrospect, it probably would have made sense to start “defecting” once in a while, and not use my headphones–since this was an iterated version of the game, he might have gotten the hint and started using headphones himself. (Or I could have just said something, I guess.)

The dynamics of this scenario also change when you assume that the headphones are partially noise-blocking. Under these conditions, everyone using speakers at once is less preferable than you using headphones, even though they’re still uncomfortable and you’d rather use speakers. The payoff matrix for this version of the situation looks like the one for Chicken:

Headphones Speakers
Headphones Decent outcome Bad outcome, good outcome
Speakers Good outcome, bad outcome Worst outcome

Both players would prefer not to yield, but the worst outcome occurs when neither yields. This version may actually have been closer to the actual scenario, since I had these big ol’ over-the-ear heapdhones, and filling our room with a cacophony of American Dad combined with Uncharted 3 might have been more annoying than just having to hear one in the background of the other.

I may add more scenarios with similar models as I think of them; hat tip to Stephen for the title.