Coordination game

A coordination game is a strategic interaction in which all players have a shared interest in reaching the same outcome, but may disagree about which outcome to coordinate on. Unlike the prisoner's dilemma, where individual rationality produces collective failure, coordination games present a different puzzle: rationality alone is insufficient to determine behavior, because the best choice for each player depends on what they believe others will choose.

The canonical example is choosing which side of the road to drive on. It does not matter whether everyone drives on the left or the right — what matters is that everyone drives on the same side. If you expect others to drive on the right, driving on the right is rational; if you expect others to drive on the left, driving on the left is rational. The equilibrium is determined not by the structure of payoffs but by shared expectations. This makes coordination games a foundational case study in common knowledge: players must not only know the game, but know that others know it, and know that others know that they know it, ad infinitum.

Thomas Schelling, in The Strategy of Conflict (1960), showed that coordination games often have focal points — solutions that stand out by virtue of salience, convention, or shared cultural background. If two people must meet in New York City without communicating, 'Grand Central Station at noon' is a more likely focal point than 'the southeast corner of 73rd and Broadway at 3:47 PM.' Focal points solve coordination problems without requiring explicit agreement, and they explain how societies achieve spontaneous order without centralized planning. Schelling's insight was that rationality underdetermines behavior and that real-world coordination depends on extrarational cues: history, symbolism, and shared narrative.

Pure Coordination and Battle of the Sexes

In a pure coordination game, all equilibria are equally preferred by all players. Driving conventions are approximately pure coordination: left and right are symmetric in payoff. In a battle of the sexes — the classic mixed-motive coordination game — players still want to coordinate, but they prefer different equilibria. Two friends want to attend the same concert, but one prefers Bach and the other prefers Stravinsky. The worst outcome is attending different concerts; the second-worst is attending the less-preferred concert together; the best is attending the preferred concert together.

The battle of the sexes reveals that coordination is not merely a technical problem of aligning expectations. It is a political problem of resolving conflicting preferences while preserving the relationship. The tools that solve it — precommitment, turn-taking, bargaining, and social norms — are the basic machinery of social life. Every standard, every protocol, every etiquette rule is a solution to a coordination game that someone, somewhere, had an incentive to disrupt.

Coordination Failures and Institutional Design

Coordination games can fail. When network effects create positive feedback, the market may lock in on a suboptimal standard simply because it arrived first. The QWERTY keyboard, VHS video format, and Internet Explorer 6 are all cases where coordination settled on an equilibrium that was not the best available. The failure is not one of rationality but of path dependence: once enough agents have coordinated on one standard, the cost of switching exceeds the benefit for any individual, even if a collective switch would be preferable.

Institutional design addresses coordination failures by changing the game. Government mandates (driving on the right), industry standards (USB-C), and platform architecture (default settings) are all mechanisms that solve coordination by removing choice rather than facilitating it. This is a darker reading of coordination than Schelling's focal points: instead of spontaneous order emerging from shared expectations, order is imposed by actors with the power to dictate the equilibrium. The question of who gets to set the default is inseparable from the question of who benefits from the default.

The persistent fantasy that coordination problems solve themselves through spontaneous order ignores the power asymmetries built into every focal point. Schelling's strangers meeting in New York assume equal ignorance; real-world coordination almost never does. The person who names the meeting place first is not solving a coordination problem — they are winning a battle of the sexes disguised as one.