The moment a decision becomes strategic
Everything you have done so far in this domain was, in a quiet way, a solo activity. A consumer weighed her own satisfaction against her own budget. A firm in perfect competition just read the market price off a board and chose a quantity. Even a monopolist, alone in its market, only had to consult its own demand curve. In each case you could find the best choice by staring hard at *your own* numbers. Nobody was staring back.
Now picture two petrol stations on the same corner. If you cut your price, your rival might cut hers — wiping out the customers you hoped to steal. If you raise it, she might hold steady and take your whole street. Suddenly there is no "best price" you can compute in isolation, because the right answer to *what should I do?* is *it depends on what she does* — and her answer depends right back on you. This circular, looking-each-other-in-the-eye situation is what economists call a game, and the formal study of it is game theory.
Three ingredients: players, strategies, payoffs
Every game, however grand or trivial, is built from the same three elements of a game. First, the players: the decision-makers — the two petrol stations, two nations, two roommates. Second, each player's strategies: the full list of choices available to them, from "charge $1.50" to "build the missile" to "do the dishes tonight." Third, the payoffs: the score each player receives for *every possible combination* of choices, capturing how much they care about each outcome.
The payoff is the subtle one, so dwell on it. A payoff is not just money. It is a number standing in for everything a player values about an outcome — profit, yes, but also prestige, survival, fairness, or a quiet life. It is the same trick as utility from earlier rungs: collapse all of a person's wants into a single score so we can ask which outcome they prefer. And crucially, your payoff usually depends on *both* players' choices at once. That is exactly why you cannot pick a strategy without first imagining theirs.
The payoff matrix: a whole game on one grid
When two players each have two strategies, we can lay out the entire game on a small grid called a payoff matrix. One player's strategies label the rows, the other's label the columns, and each of the four cells holds a *pair* of payoffs — one number for each player — for that combination of choices. Reading it is a habit worth building now, because the payoff matrix is the workhorse of this whole rung. Let us build one for our two petrol stations, who each choose only between a High price and a Low price.
Station B
High price Low price
+------------+------------+
High price | 10 , 10 | 2 , 14 |
Station A +------------+------------+
Low price | 14 , 2 | 5 , 5 |
+------------+------------+
In each cell: (A's payoff , B's payoff) = daily profit, $000s
If both go High: 10 each. If both go Low: 5 each.
If A undercuts (Low) while B stays High: A gets 14, B gets 2.Read the bottom-left cell as a worked example: A chooses Low while B chooses High, so A wins the price war and earns 14 (thousand dollars of daily profit), while B, stuck at the high price with no customers, earns only 2. The diagonal tells the real story. The friendly outcome — both High, 10 each — is better for *both* than the cut-throat one — both Low, 5 each. Yet look what tempts each player: whatever B does, A earns more by going Low (14 beats 10 if B is High; 5 beats 2 if B is Low). The matrix has quietly captured a dilemma, which the next guides unfold.
Reading a matrix, and a first taste of the logic
There is a simple way to think your way through such a grid, and it is the heart of strategic reasoning: put yourself in each player's shoes and ask, "For each thing my rival might do, what is my best reply?" Do this honestly and patterns leap out. In the petrol game, A's best reply is Low no matter what B picks. When a single strategy beats every alternative against every possible move by the other side, it is called a dominant strategy.
- Pick a player and pretend to be them, ignoring the other's payoff numbers entirely.
- Suppose the rival plays their first strategy. Scan your own payoffs in that column (or row) and mark your best reply.
- Repeat for the rival's other strategy: again find your single best reply.
- If the same reply won both times, you have found a dominant strategy. Now switch sides and do it all again for the other player.
Run that procedure on both stations and you find they *each* have a dominant strategy of Low. So both rationally cut prices and land in the bottom-right cell, earning 5 each — even though the top-left cell would have given them 10 each. A pair of strategies that are each a best reply to the other, so that nobody regrets their move given what everyone else did, is a Nash equilibrium. Here the equilibrium is the gloomy 5-and-5. That self-defeating logic is the famous prisoner's dilemma, and meeting it properly is the next guide's job.
Why so much of economics is secretly a game
This is not a parlour trick reserved for petrol stations. Recall the missing piece from the last rung: oligopoly, a market with just a few large firms. Unlike the lone monopolist or the crowd of perfect competitors, an oligopolist's profit hinges on the rivals' moves — that is precisely interdependence, and it is why the supply-and-demand machinery quietly broke down for oligopoly. Game theory is the tool built exactly to handle it: the price game you just read *is* the oligopoly problem in miniature.
Once you have the lens, you see games everywhere. Two countries deciding whether to build weapons face the very same matrix shape as our petrol stations — both would prefer to disarm, yet each fears being the lone disarmer, so both arm: an arms race. A union and an employer bargaining over wages, each weighing whether to hold firm or concede, are playing a game. So are firms deciding whether to advertise, fishers sharing a depleting stock, and governments choosing tariffs. Wherever your best move bends to someone else's, the three ingredients are lurking.
One honest caveat before we go deeper. The neat numbers in a payoff matrix are a model, not a measurement — they assume each player knows the payoffs, ranks outcomes consistently, and reasons coldly toward the best reply. Real people misjudge what others value, act on spite or trust, and rarely have a tidy grid in front of them. As with every model in this domain, the matrix earns its keep by clarifying the *logic* of a situation — why disarming feels so dangerous — not by predicting any single decision to the dollar. Used that way, it is one of the most illuminating ideas you will meet.