Game Theory · GT-14
Both players want to coordinate, and disagree only about which coordinated outcome is better.
A coordination game where both players strictly prefer being together over being apart, but each prefers a different joint activity — producing two asymmetric pure equilibria, neither obviously fairer than the other, plus a worse mixed equilibrium.
Introduced by R. Duncan Luce and Howard Raiffa in their 1957 text Games and Decisions, using the (now dated) illustrative scenario of a couple choosing between a boxing match and a ballet.
The Mechanism
Coordinating on a shared preference, despite differing tastes
Both corners of agreement (top-left, bottom-right) beat both corners of disagreement — the tension isn't over whether to coordinate, it's over which of two good outcomes to coordinate on, and each side has a mild preference for a different one.
01 · UNLIKE CHICKEN, BOTH SIDES ACTUALLY WANT TO AGREE
A much friendlier coordination problem
Every equilibrium here strictly beats every disagreement outcome for both players — this distinguishes it sharply from Chicken, where the players are adversarial about who yields. Here, the conflict is mild and secondary to a shared, stronger interest in coordinating at all.
02 · TURN-TAKING AS A REAL SOLUTION
Repetition changes what 'the equilibrium' even means
In a one-shot version, there's a genuine coordination problem with no obvious tiebreaker. In a repeated version, alternating between each player's preferred outcome (this week your pick, next week theirs) becomes an equilibrium of the larger repeated game — a solution unavailable to the single-round analysis alone.
03 · COMMUNICATION AND CONVENTION USUALLY RESOLVE IT
Cheap talk works here — unlike in adversarial games
Because both players benefit from any coordinated outcome, pre-play communication (simply stating a preference) is credible and effective at selecting an equilibrium — unlike games with conflicting interests, there's no incentive to lie about your preference here, so talking genuinely helps.
Where It Fails / Inversion
Where it fails / inversion
Treating every mild-preference-conflict coordination problem as requiring elaborate mechanism design or negotiation over-invests effort in a low-stakes disagreement — the theoretical mixed-strategy equilibrium of Battle of the Sexes is actually worse for both players than either pure equilibrium, meaning over-analysis or excessive hedging can produce an outcome strictly inferior to just picking one and going.
How To Use It
Worked example · product roadmap disagreements between co-founders
Two co-founders each prefer a different next feature to prioritize, but both strongly prefer shipping something together over each building separately (splitting focus, confusing the team). The productive move — as with the underlying game — isn't optimizing which feature is 'objectively' better, since both are reasonable; it's establishing a fast, low-cost tiebreaker (alternating priority, a coin flip, deferring to whoever has more relevant customer data this cycle) so coordination happens quickly rather than through prolonged, costly negotiation.
How to use it
When you and a counterpart both want the same thing (to coordinate) but disagree about the specific version, don't over-invest in resolving the substantive disagreement — invest instead in a fast, legitimate tiebreaker (alternation, a coin flip, deference rules) that gets you to any shared outcome quickly, since the cost of disagreement usually exceeds the value difference between the options.
See Also