In many strategic situations, there's no single, stable choice that guarantees the best outcome. Consider the simple game of Matching Pennies. You and a friend each secretly place a penny on the table, either heads up or tails up. If the pennies match, you win. If they don't, your friend wins.

If you always choose heads, your friend will quickly learn to always choose tails, and you'll always lose. If you always choose tails, they'll counter with heads. There is no pure strategy Nash Equilibrium here. Any predictable action can be exploited. The only logical approach is to be unpredictable—to randomize your choice.

When players randomize, they are using a mixed strategy. A Mixed Strategy Nash Equilibrium (MSNE) occurs when each player's chosen probability mix makes the other players indifferent between their own available pure strategies. This is the core of the Indifference Principle.

You don't randomize to improve your own payoff directly. You randomize to make your opponent's best response uncertain. By keeping them guessing, you prevent them from exploiting your strategy. Your goal is to choose probabilities for your actions (say, playing Heads 50% of the time) that make your opponent get the exact same expected payoff no matter which of their pure strategies they choose.

Let's find the MSNE for Matching Pennies. Here's the payoff matrix, where your payoffs (Player 1) are listed first.

Let's say Player 2 chooses Heads with probability $q$ and Tails with probability $(1-q)$. For Player 1 to be indifferent between playing Heads or Tails, their expected payoffs from each action must be equal.

Solving for $q$ gives us:

This means Player 2 must play Heads 50% of the time to make Player 1 indifferent. Now, let's do the same for Player 2. Let Player 1 choose Heads with probability $p$ and Tails with $(1-p)$. To make Player 2 indifferent, their expected payoffs must be equal.

The only Mixed Strategy Nash Equilibrium is for both players to choose Heads or Tails with a 50% probability. This is the only unexploitable state of the game.

Mixed strategies aren't just theoretical. They appear constantly in competitive situations where predictability is a weakness. Think of a penalty shootout in soccer. If a kicker always aims for the same corner, the goalie will eventually learn to anticipate it. The kicker must randomize their shot placement, while the goalie must randomize which way they dive. The calculated probabilities depend on the kicker's accuracy with each corner and the goalie's reach.

Similarly, in business, a company might randomize its promotional offers (e.g., 20% off one week, buy-one-get-one the next) to prevent competitors from perfectly timing their own counter-promotions. Security is another key area. The US Coast Guard uses game theory to randomize its patrol routes, making it harder for smugglers or terrorists to predict where they will be at any given time.

Calculating mixed strategies reveals the mathematical foundation of unpredictability. It shows that in many competitive scenarios, the best strategy is to not have a single, discernible strategy at all.