Week 7: The Cost of Asking

Why someone has to move first, what it costs them, and what game theory says about waiting the other person out

10 April 2028

Open the slides

Everything so far has assumed the hard part is finding and evaluating candidates. It isn’t — it’s moving. Initiating reveals interest, and revealed interest is information the other side can use against you, which is exactly the structure of a war of attrition: both sides gain from someone acting, both prefer the other side pays the cost, and the equilibrium can be a standoff that makes both sides worse off than if either had simply gone first.

The uncomfortable part is that “both wait” isn’t a failure of the model — it’s a genuine equilibrium. Neither side benefits from unilaterally deviating given that the other is staying silent, which is exactly what makes an equilibrium an equilibrium. It’s simply a worse one than “someone moves,” for both of them, at the same time it’s stable. That combination — individually rational, collectively costly — is the actual content of the week, more than the standoff itself. We’ll also ask what happens when the two sides don’t value the outcome equally: the war of attrition’s exit-timing result says the side with more to lose from further delay moves first, which is a testable claim, not just a story.

There’s a direct line back to week 3, too. A costly signal was a claim that would be expensive to fake; moving first is exactly that, aimed at a person instead of a claim about yourself. Revealed interest is this week’s name for a signal you can’t take back once it’s sent — everything week 3 established about why costly signals are credible applies here, just with the risk pointed outward instead of inward.

This is the hinge week. Everything before it was about finding and reading information; everything after it is about what happens once two sides commit to act on what they’ve found. The midterm, sat this week, tests the first half by giving you the moment itself and asking what you’d do with it — not for a “right” opening line, but for a defensible read of the situation.

Outline#

• coordination games and war-of-attrition under incomplete information
• why “both wait” is a genuine equilibrium, and why it’s still the worst outcome for everyone sitting in it
• what determines who actually moves first when the two sides don’t value the outcome equally
• how this week’s “revealed interest” is week 3’s costly signal, aimed at a person instead of a claim
• why the model can predict a standoff but not who breaks it

One thing this framework won’t explain#

A war of attrition can have more than one equilibrium — both sides wait, or one side moves — and the model doesn’t say which one two real people land in. Something outside the model breaks the tie: a script, an etiquette convention, a gendered expectation about who’s “supposed to” go first. Game theory calls that tie-breaker a focal point, and it’s doing real work in every actual standoff this week’s model describes, while the model itself has no account of where a focal point comes from or why one wins out over another.

That’s a different shape of gap again: Weeks 1, 3, 4, 5, and 6 each found something the model couldn’t price, distinguish, or account for on one side of an interaction. This one is a model with multiple valid answers and no internal way to pick between them — the missing term lives outside the game entirely, in convention.

Further reading#

• Bulow, J., & Klemperer, P. (1999). “The Generalized War of Attrition.” American Economic Review, 89(1), 175–189. The paper this week’s standoff structure is drawn from — private information about each side’s value, and exit timing that depends on that value rather than on any direct signal of it.
• Schelling, T. C. (1958). “The Strategy of Conflict: Prospectus for a Reorientation of Game Theory.” Journal of Conflict Resolution, 2(3), 203–264. The journal-length precursor to Schelling’s book of the same name, and the source of the “focal point” idea behind this week’s won’t-explain section — how two sides coordinate on an outcome with no communication and no way to enforce an agreement.