Matching Pennies, explained.
Matching Pennies is a two-player zero-sum game: one player wins when the choices match, and the other wins when they differ.
Why it happens
No fixed choice is safe against someone who can predict it. Randomization prevents systematic exploitation under the symmetric payoff rules. A mixed strategy is a probability distribution over actions, not indecision after choosing.
Mixed strategies randomize over actions. Matching pennies has no pure Nash equilibrium: after any fixed pair of choices, one player wants to change.
Read the result
Vary the opponent's choice frequency and compare expected payoffs. At a balanced mix, the symmetric game gives no predictable advantage to either pure choice. Changing the payoff table would change the relevant mix.
A worked example
A predictable defender
A defender always guards the left side, while an attacker benefits from choosing the other side.
Once the pattern is known, the attacker can exploit it even if left was a reasonable choice initially.
An unpredictable strategy can be valuable when your opponent's gain depends on anticipating your action.
OPTIONAL DEEPER DETAILGo deeper: inside the model
Inside this model
Matching choices pay A +s and B −s; different choices reverse the payoffs. A's expected heads payoff is s(2p−1), and tails is its negative. At a 50/50 opponent, both pay zero in expectation. In equilibrium both randomize equally.
Where this idea is useful
A practical use
A predictable inspection schedule can be exploited. Randomization can make a pattern harder to anticipate.
A common misconception
“Random play means ignoring strategy.”
Strategic randomization deliberately uses probabilities to prevent an opponent from exploiting a pattern.
What this explanation leaves out
- This is a zero-sum game with a fixed opponent policy. It does not imply that random behavior is best in general.
Should every mixed strategy use fifty-fifty odds?
No. The equal mix fits the symmetric Matching Pennies payoffs. In other games, optimal probabilities can differ because the gains and losses are unequal.
Would an opponent who knows your pattern be able to profit from it?
Associated thinkers
Further reading
Explore the original research or the teaching reference behind this experiment.