How the Prisoner’s Dilemma applies to my behavior

I have often called myself an amplifier. If I’m on a good team, I make the team better by bridging communication gaps, figuring out what everybody does well and steering that work to them, etc. If I’m in a bad situation, though, I can use those same skills to make things worse by undermining the decisions I don’t agree with, and stirring up discontent.

A connection I made recently is that this behavior might be analogous to game theory. The prisoner’s dilemma is a simplistic game theory construct that nevertheless reflects choices we are often asked to make in life. In short, for a one-off interaction, it makes more sense to “defect” (look out only for yourself), but over the long-term, cooperating leads to a better result for all. A societal equivalent might be that while it would be in our interest to shoplift from the local store (getting something for free while the store takes the loss), but if we did that repeatedly, the store would go out of business, and none of us would be able to buy goods there any more. The way we would behave if there was no tomorrow (and no future interactions) is different than the way we do behave when we know we will see the same people again tomorrow.

This is simulated by the iterated prisoner’s dilemma, where the players have to continue playing while knowing what the other players did previously. The surprising thing is that the winning game theory strategy in an iterated prisoner’s dilemma is the very simple tit-for-tat. The tit-for-tat player starts off by cooperating, and then just does whatever the other player did in the previous round. If the other player cooperates, they continue cooperating and building up points together. If the other player defects, tit-for-tat punishes them by defecting, but is open to “forgiveness” by returning to cooperating if the other player does first. It’s a very robust strategy, and was the winner in the famous Axelrod tournament.

The analogy I see to our everyday behavior is that tit-for-tat is an appropriate response to non-zero-sum thinking. I believe that cooperating leads to “growing the pie” (expanding the set of possibilities), which leads to more for all in the long run, as opposed to zero-sum thinking where my gain is your loss (as the defector in the Prisoner’s Dilemma would think).

Tit-for-tat also teaches us a way to address those who are stuck in that zero-sum thinking. We punish them by also defecting, so that we both lose, as in the ultimatum game. Ideally, this leads to them seeing the downside of their self-centered behavior, and increasing their cooperation in the future. But that also depends on us being generous and forgiving them when they do start cooperating.

So the connection I made recently was that my tendency to act as an amplifier is analogous to a tit-for-tat strategy on steroids. I start off treating everybody I meet as interesting and able to contribute to the team/goal. When they do so, I encourage that behavior and look for ways to amplify it. However, if they “defect” and act in a way that I consider detrimental to the team, I slam to the other extreme and look to take them down. But as I’ve gotten more experience, I have learned to also try to understand what they’re thinking so I can persuade them to try cooperating. If that works, I ideally start building on that cooperation to get us back into the non-zero “grow the pie” mode. But I admittedly sometimes hold grudges based on their past behavior (not my best trait – if that happens, I generally have to remove myself from the situation because it just gets toxic).

Anyway, I thought this was an interesting connection between game theory and my tendencies, and thought I’d share it to hear what others thought.

P.S. I’m going to try to get back into the habit of posting more of my thoughts now that I’m done with my summer of insane physical challenges, and so this is a first step in that direction. Encourage me by responding! 🙂

6 thoughts on “How the Prisoner’s Dilemma applies to my behavior

  1. Game theory makes for good frameworks but… I find that it tends to model strictly rational agents, which we humans are most definitely not. What if someone starts out defecting? How will they ever discover cooperation if all they’ve seen modelled is defecting?

    I wonder if the insight and takeaway here are more about feedback than acting. Sounds like an appropriate human version of this is to give clear and reasonably immediate feedback on ‘defecting’ and only defect yourself if the behaviour continues. Which leads us to ratings and metrics and peer reviews. Ouch.

  2. Very thought-provoking. I find it interesting that self-introspection would lead to a nice theoretically-clean analysis….I’m not as good at that yet. Thanks!

  3. Arvind, I don’t think that game theory is definitive as a way of modeling people by any means – I just thought it was an interesting lens through which to view my behavior. And, yes, part of the challenge is to teach people there’s a better way than min-maxing your way through life to accumulate the most for yourself, despite the teachings of Tim Ferriss, The Game, et al.

    Also, an interesting point about feedback – I think there’s a follow-up post here where I discuss the various ways to constrain action. I still like Lessig’s framework of code/architecture (design both digital and physical spaces to make it hard to do wrong behavior), law (outlaw wrong), social norms (shun wrong-doers) and the market (make it expensive to do wrong). We can choose to use any of these and I’m not quite sure how to apply that, but it seems like there’s something there.

    Eli, I don’t know how theoretically clean it is, but I thought it was a fun perspective on something I already do. Thanks for the comment!

Leave a Reply

Your email address will not be published. Required fields are marked *