Defined by John F. Nash Jr. (Equilibrium Points in N-Person Games, 1950).
Finds the outcomes where no player can do better by changing course alone, predicting where competing choices will settle even when everyone would prefer another result.
Two building-supply distributors deciding whether to cut their contractor price
Each cell gives both firms' quarterly profit in millions of dollars; Firm A is ours, Firm B the rival. A Best reply tag marks the better choice for a firm given what the other does; the cell where both are best replies is the equilibrium. Here it is a price war, though both would earn more holding. Change the two 7s to 4s (customers who no longer switch for a small discount) and holding becomes a second equilibrium. Illustrative.
Firm B chooses
Firm A chooses
Zero equilibria in pure strategies.One equilibrium in pure strategies.Two equilibria in pure strategies.Three equilibria in pure strategies.Four equilibria in pure strategies.
- Hold price, Hold price: an equilibrium: neither gains by switching alone.not stable: Firm A gains by switching.not stable: Firm B gains by switching.not stable: either gains by switching.
- Hold price, Cut price: an equilibrium: neither gains by switching alone.not stable: Firm A gains by switching.not stable: Firm B gains by switching.not stable: either gains by switching.
- Cut price, Hold price: an equilibrium: neither gains by switching alone.not stable: Firm A gains by switching.not stable: Firm B gains by switching.not stable: either gains by switching.
- Cut price, Cut price: an equilibrium: neither gains by switching alone.not stable: Firm A gains by switching.not stable: Firm B gains by switching.not stable: either gains by switching.
An example
A building-supply distributor's leadership team was ready to answer a rival's rumored contractor discount with one of its own. The strategy director drew a two-by-two on the whiteboard first: each firm holds its price or cuts it, with each cell's quarterly profit estimated from last year's lost-bid records.
Holding while the rival cut lost the most; cutting while the rival held won the most. Whatever the rival did, cutting paid each firm better, so both cutting was the only cell where neither would move alone, even though both holding beat it for both. The table told the team that matching discounts would not end the fight, and that the move worth making was one that changed the payoffs. They spent the money on next-morning delivery to job sites, so a small discount elsewhere would no longer pull their contractors away, and left the price alone.
Why it works
Each player's best choice depends on the other's, so "what should we do?" asked alone has no answer. The equilibrium answers a narrower question that does: where would things settle if each side kept choosing its best reply to the other? A cell where both are best replies is stable, because neither can do better by switching alone, and that is where competition tends to end up.
The uncomfortable lesson is that the stable cell is often worse for everyone than another cell they cannot reach, as in a price war. That turns strategy away from outguessing the rival and toward changing the game: commitments, information, or products that alter the payoffs, so the stable outcome moves somewhere better.
The thinking behind it
John F. Nash Jr. John F. Nash Jr., an American mathematician, defined the equilibrium in Equilibrium Points in N-Person Games (1950), in the Proceedings of the National Academy of Sciences: a set of strategies, one per player, in which no player can do better by changing strategy alone. He also showed that any game with a finite number of players and strategies has one or more such points, in mixed strategies (choices made with set probabilities) if not in pure ones. Because the definition holds for any number of players, it reaches well beyond two-person contests.
How to run it
- Name the players whose choices interact. Keep it to the two or three whose moves change each other's results.
- List each player's real options. Two or three strategies each, stated as actions: hold price, cut price, add a service.
- Estimate the payoffs for every combination. Use your data for your side and your best estimate for theirs, and note where each number came from.
- Mark each side's best reply. For each choice the other might make, circle the option that pays you most, then do the same for them.
- Find the cells where both are best replies. Those are the equilibria; with none in pure strategies, expect mixing or a moving target.
- Compare them with the cell both would prefer. The gap between the stable outcome and the better one is the prize for changing the game.
- Look for a move that changes the payoffs. A commitment, a contract, a new service, or new information that makes the better cell stable.
Where it goes wrong
The rival's payoffs are written from the wrong side of the table. Teams estimate their own numbers carefully and guess the rival's from their own view, then predict moves the rival would never make. Write the rival's payoffs as its leadership would see them, with its costs and pressures.
The game is played once on paper and many times in life. A one-shot table predicts a price war that two firms meeting every quarter may avoid. If the players will meet again, model the repetition.
The equilibrium is read as advice. It predicts where things settle and says nothing about whether that is a good place to be. Its most useful output is often a reason to change the payoffs.
When to use something else
- Repeated Games. When the same players meet again and again, and the prospect of future rounds can hold cooperation in place.
- Backward Induction. When moves come in turn and each player sees the other's move before choosing.
- Game Theory. When the team first needs the basic terms (players, strategies, payoffs) before building a table.
Where it sits in SOLVE eX
Go to the source
- Equilibrium Points in N-Person Games, Proceedings of the National Academy of Sciences, 1950