In many strategic situations, being predictable is a weakness. If your opponent knows what you’ll do, they can easily counter it. This is where mixed strategies come in. Instead of choosing a single 'pure' action, players randomize their choices according to a specific probability distribution. The goal isn't just to be random, but to be random in a very precise way.

The core concept for finding the right mix is the Indifference Principle A player will only be willing to randomize their actions if each of those actions yields the same expected payoff. If one action were better, they'd just choose that one every time. So, to find a player's mixed strategy equilibrium, we calculate the probabilities that make their opponent indifferent to their own choices.

Let’s work through an example. Consider a non-symmetric game of 'Matching Pennies.' Player 1 wants to match Player 2 (Heads-Heads or Tails-Tails), while Player 2 wants to mismatch.

To find Player 1's mixed strategy, we'll make Player 2 indifferent. Let's say Player 1 plays Heads with probability $p$ and Tails with probability $(1-p)$.

Player 2's expected payoff from playing Heads must equal their expected payoff from playing Tails.

Expected Payoff for P2 (Heads) = $p(-1) + (1-p)(1) = 1 - 2p$ Expected Payoff for P2 (Tails) = $p(1) + (1-p)(-1) = 2p - 1$

Setting these equal to each other gives us the equilibrium.

Since the game is symmetric, the calculation for Player 2 is identical. If Player 2 plays Heads with probability $q$, they must also choose $q=0.5$ to make Player 1 indifferent. The Mixed Strategy Nash Equilibrium (MSNE) is where Player 1 plays (0.5 Heads, 0.5 Tails) and Player 2 does the same.

Another way to visualize this is through best response functions. A player's best response is the strategy that yields the highest payoff, given the other player's strategy. In mixed strategies, this 'strategy' is a probability.

Let's reconsider Player 2's best response to Player 1's choice of $p$ (the probability of playing Heads).

If Player 1 chooses $p > 0.5$, then Player 2's payoff for playing Tails ($2p-1$) is greater than for playing Heads ($1-2p$). So Player 2's best response is to play Tails for sure ($q=0$). If Player 1 chooses $p < 0.5$, Player 2's best response is to play Heads for sure ($q=1$). If Player 1 chooses $p = 0.5$, Player 2 is indifferent and can play any mix of Heads and Tails ($0 \le q \le 1$).

The MSNE is where these best response functions intersect.

This graphical method is useful for 2-player, 2-strategy games. For games with more players or strategies, the algebraic approach using the Indifference Principle is more practical. The logic extends directly: for an N-player game, you must find a probability mix for each player that makes all other N-1 players indifferent among the pure strategies they are mixing.

Not every game has a neat mixed strategy equilibrium. Sometimes, a situation called degeneracy occurs. A degenerate game is one where a player has more pure strategies in their equilibrium mix than are needed to make their opponent indifferent.

For example, if Player 2 has three choices (Rock, Paper, Scissors), but Player 1 only has two (Rock, Scissors), Player 1 might only need to mix between their two choices to make Player 2 indifferent between two of their three choices. The third choice for Player 2 might be strictly worse regardless of Player 1's mix. This can lead to multiple equilibria or situations where some strategies are never played.

Calculating mixed strategies transforms game theory from a discrete choice problem into a continuous one. By allowing players to use probability, we can model and predict behavior in situations where unpredictability is a core feature of the strategy, from financial markets to military tactics.

Ready to test your understanding? Let's see if you can apply these principles.

Mastering these calculations provides a powerful tool for analyzing complex strategic interactions.