When dominant strategies run out
In the last two guides you got lucky. The prisoner's dilemma had a tidy answer because each player had a dominant strategy — a move that was best no matter what the other did. When confessing beats staying silent against *every* possible choice by your partner, you do not even need to guess what they will do. But that tidiness is rare. In most real games — pricing a product, choosing a meeting spot, deciding whether to invest — your best move genuinely *depends* on theirs, and theirs on yours. That is the whole point of interdependence: there is no single move that is best against everything.
So we need a new kind of answer — one that does not lean on a dominant strategy. The question to ask is no longer "what is best against everything?" but "is there a combination of choices that is *stable* — one where, once everyone can see what everyone else picked, nobody wishes they had chosen differently?" A young mathematician named John Nash gave that question a precise, general answer in 1950, and it reshaped economics so thoroughly that he shared the 1994 Nobel Prize in Economics for it. The answer is the Nash equilibrium.
The best response to the best response
Start with one building block: a best response. Fix what everyone else is doing, then ask, "given exactly that, what is the single move that earns *me* the most?" That move is my best response to their choices. It is a humble idea — it does not try to outguess the whole world, only to answer one frozen situation optimally. A Nash equilibrium is then just a combination of choices in which *every* player is simultaneously playing a best response to all the others. Each person, looking at what everyone else did, finds that their own move was already the best they could have made.
The defining test is wonderfully concrete, and you can apply it yourself to any payoff matrix. Point at one cell — one combination of choices — and ask each player in turn: "holding everyone else's choice fixed, could you do better by switching your own move alone?" If the answer for *every* player is no, that cell is a Nash equilibrium. If even one player could profitably deviate by changing their move while the others stay put, the cell is not stable — that player will jump, and the situation unravels. Equilibrium is exactly the place where no single, one-sided deviation pays.
Working one out, by hand
Let us pin it down with a tiny example from a real market. Two coffee shops, Ava's and Ben's, each choose a Low or High price for the week. The grid below shows their weekly profits — the first number is Ava's, the second is Ben's. This is just a payoff matrix, read exactly the way you learned earlier. We will hunt for the equilibrium by the deviation test, one cell at a time.
BEN: Low BEN: High
AVA: Low Ava 30, Ben 30 Ava 50, Ben 20
AVA: High Ava 20, Ben 50 Ava 45, Ben 45
Deviation check on (High, High) = 45, 45:
Ava alone switches to Low -> 50 > 45 : Ava WILL deviate
So (High, High) is NOT a Nash equilibrium.
Deviation check on (Low, Low) = 30, 30:
Ava alone switches to High -> 20 < 30 : Ava stays
Ben alone switches to High -> 20 < 30 : Ben stays
Nobody gains by moving alone -> (Low, Low) IS the Nash equilibrium.Look hard at what the table just told us. Both shops would clearly prefer the (High, High) world — 45 each beats 30 each. Yet that better world is *not* an equilibrium, because from it Ava can sneak to a Low price and grab 50 while Ben is left with 20. Ben, anticipating exactly this, will not sit at High either. The only resting point is (Low, Low), where both earn the mediocre 30 — and neither can improve by moving alone. If that shape feels familiar, it should: this is the prisoner's dilemma wearing a price tag, and "both price Low" is its Nash equilibrium.
Stable does not mean good
Here is the lesson that students find genuinely startling, and it is worth holding onto. A Nash equilibrium is *stable*, but it need not be *good* — not for society, and not even for the very players stuck in it. In the coffee example, (Low, Low) is the equilibrium and it leaves both shops poorer than (High, High) would. The equilibrium is not a happy destination the market climbs toward; it is merely the place where the logic of self-interest comes to rest, even when that resting place is a trap everyone wishes they could escape.
This is why Nash's idea explains so much that looks irrational from the outside. An arms race is a Nash equilibrium: both nations would be safer and richer disarmed, but given that the *other* side is armed, each one's best response is to arm too — so neither dares stop. Overfishing, traffic jams, ad spending wars, and price wars all share this skeleton. Each actor is behaving perfectly sensibly given everyone else's behaviour, and the collective result is one nobody wanted. The equilibrium is the diagnosis; it tells you precisely why good intentions are not enough to escape.
Sometimes more than one, sometimes none in sight
A Nash equilibrium need not be unique. Many games have *several*. Picture two friends who lost their phones and must each decide, separately, which of two cafés to wait at. If both go to the North café they meet (great); both to the South, they meet (also great); split up and they each sit alone (bad). Run the deviation test: (North, North) is an equilibrium — if your friend is at North, your best response is North too. But (South, South) is *equally* an equilibrium by the same logic. Two stable outcomes, and the theory by itself cannot tell you which one the friends will land on.
This is a coordination game, and multiple equilibria are its signature. When there are many stable outcomes, the puzzle shifts from "what is stable?" to "which stable point will people actually pick?" Often the answer comes from something *outside* the cold payoffs — a shared landmark, a habit, a custom, the one option that simply stands out. ("We always meet at the North café.") That salient choice is called a focal point, and you will meet it properly in the next guides on coordination. For now, just register the shift: with multiple equilibria, history, culture, and expectations start to matter as much as the numbers.
And here is Nash's deepest contribution. You might worry that some games have *no* equilibrium at all — think of rock-paper-scissors, where for any single fixed move your opponent has a winning reply, so no pure choice is stable. Nash proved that if players are allowed to *randomize* — to choose a probability mix over their moves, like "rock one-third of the time" — then *every* finite game has at least one equilibrium. In rock-paper-scissors it is "play each option one-third at random," which nobody can exploit. That existence theorem is what made the concept universal, and it is the mathematics for which the Nobel was awarded.
What the concept can and cannot promise
It is easy to oversell Nash equilibrium, so let us be honest about its limits. The concept tells you which outcomes *could* persist, but not always how players *get* there, nor — when there are several — which one they will settle on. It is a prediction about resting points, not a recipe for play. Economists argue, fairly, about how often real people actually reach the equilibrium of a game, especially complicated ones, on the first try. The concept's power is descriptive and diagnostic; it is weakest as a crystal ball.
There is also a built-in assumption worth naming: the standard analysis leans on rational choice — players who understand the payoffs and reliably pick their best response. Behavioural economics, which you will reach later in the ladder, documents many ways real humans depart from that ideal: they misjudge odds, value fairness, make mistakes, and care what others think. None of this makes Nash equilibrium useless — it remains the indispensable first lens. But treat it the way a sailor treats a chart: a faithful map of the stable harbours, not a guarantee of the voyage, and never a substitute for looking at the actual water.