A Prisoner’s Dilemma (and why “trust” is usually a math problem)

Happy Sunday Friends!

Here’s one quote I’m musing on this week, two ideas, three favorites, and one question to take with you into the week ahead.

One Quote I’m Musing

“The foundation of cooperation is not really trust, but the durability of the relationship.”
| Robert Axelrod

A friend of mine, Jim, asked me, “Why do you keep talking about game theory?”

Because game theory is what’s happening under the conversation. Everywhere you care about.

In your marriage. In your business. In your team. In any future where people have to choose between doing what’s honorable and doing what feels safe.

When nations and corporations posture, escalate, and burn fortunes to avoid looking weak.

And it’s all hiding inside one simple game introduced to us by the RAND Corporation’s research: The Prisoner’s Dilemma.

Before we get academic, I’ll give you the same scene I gave him.

A narrow hallway that smells like bleach and burnt coffee. Fluorescent lights buzzing overhead like they’re tired too. Somewhere down the corridor, a door shuts with a clean, final clang.

You’re in a small room with a metal chair bolted to the floor. The table wobbles by a millimeter every time you shift your weight. A cheap fan pushes warm air in slow circles. An investigator flips a folder like it’s already decided.

You know the other prisoner is in the building. Close enough to be real. Far enough to be useless.

The setup

Two prisoners. Separate cells. You both know the other is there, but you cannot talk to each other.

Each day, you are separately interrogated. You each choose to:

• Stay silent (cooperate)

• Rat (defect)

The rules:

• If both stay silent: the system can’t do anything and you each get a symbolic reward of 3 coins.

• If one rats and the other stays silent: the rat gets 5 coins and the other gets 0 coins.

• If both rat: both get a small reward of 1 coin each.

Pretty Simple:

• (Silent, Silent) = (3,3)

• (Rat, Silent) = (5,0)

• (Silent, Rat) = (0,5)

• (Rat, Rat) = (1,1)

The trap (why smart people still lose)

Jim instantly said what every sharp person says:

“So always rat. Five coins beats three... And you will get at least one.”

In a one-shot game, he’s right. “Rat” is the safe move.

But that safety creates the trap.

Because if both people think that way, you get (1,1).

Not because either person is evil, but because the system punishes the one who cooperates alone.

When being the “good one” makes you the sucker, people stop being good.

The twist is the repetition

This doesn’t happen once, it happens again and again.

After each round, the prisoners learn whether the other person stayed silent or ratted.

Then it starts again.

Now the math changes.

Reputation becomes currency.

The best strategy is simple, not soft

A generous Tit for Tat looks like this:

1. Start by cooperating.

2. Then copy what the other person did last round.

It wins for one reason:

It cooperates without being exploitable.

It’s nice first.

It punches back immediately.

And it forgives quickly when the other side returns to cooperation.

That’s the part most people miss. They think the lesson is “be nice.”

No.

The lesson is: be predictable, be fair, and be willing to respond in kind.

What’s the difference between generous tit for tat and regular tit for tat? Be about 10% more forgiving.

Now zoom out: the third player you weren’t introduced to

Welcome to the meta-analysis.
Most people tell this story like it’s a two-player game.

It isn’t.

There’s a third player you never see in the cells: the System, the interrogator, the institution, the rule-maker.

The System doesn’t care about your honor. It cares about efficiency: get the most information for the least effort.

In coin terms, look at what the System pays versus what it gets:

• If you both stay silent: the System pays out 6 coins and gets 0 information.

• If one of you rats: the System pays 5 coins and gets 1 stream of information.

• If you both rat: the System pays 2 coins and gets 2 streams, and you did the pressure work for it.

It’s about efficiency. Not necessarily malice; just optimization.

That’s the meta-lesson: when defection is everywhere, stop blaming “trust.” Audit the incentives.

Mutual defection (selfish, lack of cooperation) is usually the System’s favorite outcome.

Mirror Moments

If you’ve ever said any of these, you’ve been living inside this payoff grid:

• “I’m not sharing that yet. It’ll make us look bad.”

• “If I raise the issue, I’ll be the one blamed.”

• “I’ll play nice, but I’m keeping receipts.”

That’s not a character flaw. That’s a rational response to a one-shot environment.

When organizations say “collaborate” but reward:

• credit-hoarding,

• blame-avoidance,

• punishing messengers,

…they’re designing defection.

Then they act surprised when they get it.

A Final Reframe

Most trust problems are not trust problems.

They are incentive problems.

If you want more cooperation, don’t preach values, engineer incentives that make values rational, then live them.

Change the payoffs. Extend the time horizon. Make it a repeated game.

A tiny protocol (how I actually do this with real humans)

Here’s the thing about “Tit for Tat” in the real world: people aren’t robots. They miss emails. They have bad days. They read tone into a period.

So I run a slightly kinder rule:

• Assume noise once.

• Forgive the first slip.

• If it keeps happening, respond in kind, fast and clean, then come right back to cooperation.

That’s the “generous” part. About 10% more forgiving than your ego thinks you should be.

Now, to make this practical, here’s the move I use when work crosses people or teams.

The 60-second handoff

Before you end the meeting, the call, the hallway chat, ask one question:

“What’s one thing we can do right now to make the next handoff stupid-easy?”

Then do exactly one tiny thing that takes under a minute:

• drop the link

• name the owner

• write the one sentence of context you wish you’d had

That’s it. No manifesto. No culture deck. Just friction removed.

And when someone does it well, say it out loud. “Good catch.” “That helped.” “Appreciate you.”

That little acknowledgement is how cooperation stops being an aspiration and becomes a habit.

Two Ideas From Me

1. A culture is a payoff matrix people can feel.
   Watch what gets rewarded, not what gets written on the wall.

2. “Never trust again” sounds strong, but it’s expensive. Grudge-strategies protect the ego and bankrupt the future.

Top 3 Things This Week

1. 3 Questions: How AI could optimize the power grid
   MIT researchers are exploring how AI can optimize power grids for improved efficiency, increased resilience, and better integration of renewable energy sources. The focus on democratizing AI development and aligning it with real-world applications is a valuable point. | MIT

2. Virginia Woolf would remind us to “Arrange whatever pieces come your way.”
   A great reminder that while we can’t control what happens, but we can work with what we have.

3. A reminder on choosing our pursuits well from Pliny the Younger: “An object in possession seldom retains the same charm that it had in pursuit.”
   A frequent truth in my current chapter. Make sure we’re enjoying the journey towards accomplishment, or we might not know what to do once we’ve achieved it.

One Question to Take Into Your Week

Where in your life have you accidentally created a one-shot game… then acted surprised when people started protecting themselves?

Journal prompt:
“The place I ask for trust, but reward defection, is…”

If this hit you in the chest, forward it to one person you’re trying to build with.

Until Sunday, my friends.

Think Dangerously.
–e