Potential Games

Games where all unilateral payoff changes are captured by a single real-valued function.

Ask the Game theory assistant 1 min read · Updated September 9, 2026

Definition

$G$ is an exact potential game if there exists $\Phi:S\to\mathbb{R}$ with $u_i(s_i',s_{-i})-u_i(s_i,s_{-i})=\Phi(s_i',s_{-i})-\Phi(s_i,s_{-i})$ for all $i,s_i,s_i',s_{-i}$ .

Ordinal, weighted, and generalized-ordinal potentials relax the equality requirement.

Intuition

Players jointly hill-climb a common function — best-response dynamics always converge.

This gives potential games strong computational and learnability properties.

Worked example

Congestion games are exact potential games with $\Phi(s)=\sum_e\sum_{k=1}^{f_e(s)}c_e(k)$ (Rosenthal 1973).

Coordination games on graphs are (weighted) potential games.

The math

Monderer-Shapley (1996): every finite potential game has a pure Nash equilibrium — the maximizer of $\Phi$ .

Better-response and fictitious-play dynamics converge to Nash in potential games.

Where it is used

Distributed sensor networks use potential-game formulations to coordinate channel selection: each sensor's local decision to switch channels corresponds to climbing the potential function, guaranteeing convergence to interference-minimizing configurations.

Bridge to convex optimization via $\Phi$ .

More in Game theory

Assembled from the ReLU.chat curated knowledge base. These explanations are concise on purpose; check the sources for anything important.