Global Games¶

Pick a single prediction out of a coordination game's many self-fulfilling equilibria by giving each player a private noisy signal of a common fundamental and deleting dominated strategies inward until one threshold cutoff survives.

Version: v2 · 2026-09-19 · History
Domain-specific #: 451
Domain group: Social Sciences
Origin domain: Economics & Finance
Subdomain: Game Theory and Equilibrium Selection → Economics & Finance

Core Idea¶

A class of incomplete-information coordination games (Carlsson–van Damme 1993; Morris–Shin 1998) that selects a unique equilibrium where complete information leaves many. Each player gets a private noisy signal of a common unobserved fundamental; at extreme signals one action is dominant, and iterated deletion of dominated strategies inward from those regions pins down a single threshold strategy — act if and only if your signal falls on the calculable side of a cutoff.

How would you explain it like I'm…

The Blurry Clue Game

Imagine a bunch of kids deciding whether to all push open a heavy door together—it only opens if enough push. Each kid gets their own slightly fuzzy guess about how heavy the door is. If your guess says it's super light, you push; if super heavy, you don't. Starting from those easy cases, everyone can work out one clear rule: push if your guess is below a certain line.

Noisy Clues, One Cutoff

Some games work only if lots of people do the same thing, like everyone joining in at once. If everyone knew all the facts, the group could end up either all joining or all staying out, and nobody could predict which. Global games add a twist: each player gets their own slightly fuzzy clue about the true situation. At the extremes, the right choice is obvious no matter what others do. Reasoning inward from those obvious cases, everyone settles on one rule: act if your clue is past a certain cutoff. That gives a single predicted outcome instead of two.

Noisy Signals Select Equilibrium

Global games are coordination games in which each player gets a private, noisy signal about a shared hidden 'fundamental,' such as a central bank's reserves. With complete information, such games often have multiple self-fulfilling equilibria. Everyone attacking can be rational, and so can no one attacking. The noisy signals fix this. At extreme signals one action is best no matter what others do, which creates 'dominance regions.' Players then reason inward. If others attack below some signal, I should attack a bit beyond it too, and so on, until a single threshold strategy remains. Each player takes the action if and only if their signal falls on the right side of a cutoff. That cutoff is set by the noise and the payoffs, not by any shared focal point, which gives a unique, testable prediction.

Global games are incomplete-information coordination games introduced by Carlsson and van Damme (1993) and developed for crisis modeling by Morris and Shin (1998, 2003). Each player observes a private noisy signal of a common unobserved fundamental. That informational structure selects a unique equilibrium in a game that would have multiple self-fulfilling equilibria under complete information. The mechanism is iterated deletion of strictly dominated strategies. At extreme signals one action is dominant. For instance, attacking is dominant if reserves are certainly zero, and not attacking is dominant if reserves are certainly unlimited. These dominance regions anchor beliefs at the boundaries, and inward iteration pins down a threshold strategy over the whole signal space. The cutoff depends on the noise structure and payoff parameters rather than on exogenous coordination devices, which yields sharp comparative statics. In the Morris–Shin currency-attack model, the threshold depends on the noise variance, the cost of attacking, and the reserves needed for a successful defense. The approach extends to bank runs (Goldstein and Pauzner 2005), sovereign-debt rollover, technology adoption with network externalities, and regime-change participation.

Scope of Application¶

An equilibrium-selection method that lives wherever three ingredients recur: a base coordination game with strategic complementarities, a common unobserved fundamental, and a private-signal structure.

Currency attacks — the originating case, with the attack threshold set by reserves, attack cost, and noise variance.
Bank runs — a unique run threshold as a function of bank fundamentals (Goldstein–Pauzner).
Sovereign-debt rollover crises — each creditor's decision keyed to beliefs about the others.
Technology adoption under network externalities — unique adoption thresholds replace chicken-and-egg multiplicity.
Political regime change — participation thresholds in protest keyed to beliefs about others.

Clarity¶

Global games dissolve the question that haunted coordination-game analysis: which self-fulfilling equilibrium do agents actually play? By perturbing the game with private noise, they supply the first mechanism-grounded selection — a calculable cutoff rather than an unsystematic focal-point guess. They also refute the fundamentals-versus-sentiment dichotomy, since the noise ties the outcome to the fundamental.

Manages Complexity¶

A coordination game under complete information is a swamp of self-fulfilling equilibria across a band of fundamentals. The framework collapses that swamp to one procedure with one output: solve for a single cutoff, a function of just the signal noise and the payoffs. A family of intractable crisis models becomes one tool deployed repeatedly.

Abstract Reasoning¶

The framework licenses equilibrium selection by construction (deducing rather than guessing the outcome), solve-for-the-cutoff reasoning that reads the crisis boundary off two inputs, and comparative statics on the threshold. A subtler move tracks public versus private information separately, since a public signal moves the threshold disproportionately.

Knowledge Transfer¶

A technical framework, so it transfers as method within the strategic-interaction substrate: the whole apparatus carries without translation across currency attacks, bank runs, debt rollovers, and regime change, which are co-instances of one object rather than analogies. It does not reach non-strategic systems — physical noise-induced-order phenomena share only a family resemblance. The portable selection insight belongs to the parent prime Coordination Problem and Equilibrium Selection.

Relationships to Other Abstractions¶

Current abstraction Global Games Domain-specific

Parents (4) — outgoing typed relations

Global Games includes Bayesian Nash Equilibrium Domain-specific

A Global Game contains a Bayesian equilibrium over signal-conditioned strategies; its uniqueness result selects one threshold member of that solution class.
Global Games includes Dominated Strategy Domain-specific

Global Games contains iterated deletion of dominated strategies as the engine that moves inward from extreme-signal dominance regions.
Global Games includes Threshold Prime

The selected equilibrium contains a cutoff that separates the signal region in which each of the two actions is optimal.
Global Games is a decomposition of Coordination Problem and Equilibrium Selection Prime

Removing private-signal game machinery leaves a method for selecting one outcome from several self-sustaining coordination equilibria.

Hierarchy paths (22) — routes to 14 parentless roots

Global Games → includes Bayesian Nash Equilibrium → is a kind of Nash Equilibrium → is a kind of, typical Equilibrium → is a kind of, typical Fixed Point

Show alternative paths (21)

Neighborhood in Abstraction Space¶

Global Games sits in a crowded region of the domain-specific corpus (28^th percentile for distinctiveness): several abstractions share nearly its structure, so a description that fits it tends to fit its neighbors too.

Family — Game-Theoretic Models & Paradoxes (38 abstractions)

Nearest neighbors

Computed from structural-signature embeddings · 2026-10-25