Skip to content

Mixed Strategy Equilibrium Finding

Why It Matters

Many games have no pure strategy Nash equilibrium - no combination of deterministic actions is stable. Yet players must choose something, and outcomes display statistical regularities. Mixed strategy equilibrium explains these patterns: players randomize in specific proportions that make opponents indifferent, so no pure strategy dominates. This is crucial for situations involving inspection (tax audits, drug testing), competition with unpredictability value (penalty kicks, poker), and hide-and-seek scenarios (patrol routes, ad timing). Without mixed strategy analysis, these situations appear chaotic but are actually structured probabilistically.

The Key Move

  1. Check if pure strategy Nash equilibrium exists - if yes, no need for mixing.
  2. If no pure equilibrium, assume players randomize.
  3. Let p = probability player 1 chooses action A (versus 1-p for action B). Let q = probability player 2 chooses action C (versus 1-q for action D).
  4. For player 1's mixture to be optimal, player 2 must be indifferent between C and D. Calculate: player 2's expected payoff from C equals expected payoff from D, which gives an equation in terms of p.
  5. Solve for p.
  6. Repeat for player 1: set their expected payoffs from A and B equal, solve for q.
  7. The mixed strategy equilibrium is (p, 1-p) for player 1 and (q, 1-q) for player 2.

Classic Application

Penalty kicks in soccer. Kicker chooses left or right. Goalie chooses which direction to dive. If kicker goes where goalie dives, usually saved. If different directions, usually goal. No pure strategy equilibrium - if kicker always goes left, goalie dives left; but then kicker wants to switch to right. Mixed strategy equilibrium: kicker randomizes with probabilities that make goalie indifferent (expected payoff from diving left = expected payoff from diving right), and vice versa. Empirical studies show professional players approximate these equilibrium probabilities - not through calculation but through learning.

Surprising Application

Bluffing frequency in poker. You hold weak hands sometimes and must choose: fold (certain small loss) or bluff (risk big loss if called, but win if opponent folds). Pure strategy "never bluff" is exploitable - opponent always folds to your bets, so you win only tiny pots. Pure strategy "always bluff" is exploitable - opponent always calls, you lose big. Mixed strategy equilibrium: bluff with probability that makes opponent indifferent between calling and folding. This depends on pot odds. Math shows optimal bluffing frequency increases with pot size. Explains why good poker players bluff more in big pots - not psychology but equilibrium mixing probability. Non-obvious because bluffing seems like psychological deception but it's actually mathematical randomization.

Failure Modes

Assuming players consciously randomize - humans often approximate mixed strategies through intuition or learning, not deliberate calculation. Not checking for multiple equilibria - some games have multiple mixed equilibria. Ignoring dominated strategies - must eliminate dominated strategies before calculating mixtures. Treating all randomization as equilibrium - sometimes observed randomness is just noise, not strategic mixing. Over-applying - many situations with apparent randomness don't have strategic structure requiring mixed equilibrium.

Go Deeper

Gibbons, Game Theory for Applied Economists, Chapter 1.3.B; Walker & Wooders, "Minimax Play at Wimbledon," American Economic Review 97(5), 2001