What game theory studies
Game theory is the study of decisions where the outcome depends on what other decision-makers do. In a coin toss, nothing you choose matters. In chess, poker, an auction or a negotiation, your best choice depends on what your opponent is likely to do, and their best choice depends on yours. Game theory gives that interdependence a precise form.
The word “game” is used broadly. It can mean a literal game, but also pricing decisions between two shops or how two drivers approach a narrow road. The aim is not to predict how people feel but to describe what sensible choices look like when everyone is reasoning about everyone else.
The ingredients
Most simple game-theory problems are described with the same few parts:
- Players: the decision-makers.
- Strategies: the choices each player can make.
- Payoffs: the result for each player for every combination of choices. Payoffs can be points, money, wins or any measure of how good the outcome is.
When two players each choose once and without seeing the other’s choice, the situation can be drawn as a small table. That table is the standard first tool.
A worked example: two players, two choices
Here is a made-up game with arbitrary points. Each player secretly chooses Cooperate or Defect. The numbers show the points for the row player first, then the column player.
| Column cooperates | Column defects | |
|---|---|---|
| Row cooperates | 3, 3 | 0, 5 |
| Row defects | 5, 0 | 1, 1 |
This is the structure of the classic prisoner’s dilemma. Read it like this: if both cooperate, each gets 3. If one defects while the other cooperates, the defector gets 5 and the cooperator gets 0. If both defect, each gets 1.
Dominant strategies
A strategy is dominant if it gives a player a better payoff than their alternative no matter what the opponent does. Test the row player:
- If the column player cooperates, the row player gets 3 by cooperating or 5 by defecting. Defecting is better.
- If the column player defects, the row player gets 0 by cooperating or 1 by defecting. Defecting is better.
Defect wins in both cases (5 beats 3, and 1 beats 0), so it is dominant for the row player. The table is symmetric, so it is dominant for the column player too.
If both players reason this way, both defect and each receives 1. That is worse for both than the 3 each they would have received by cooperating. This is the central lesson of the example: individually sensible choices can lead to a collectively worse outcome.
Equilibrium
An equilibrium, most often called a Nash equilibrium, is a set of strategies where no player can do better by changing only their own choice while the others stay put. In the table, (defect, defect) is an equilibrium: if the other player defects, switching to cooperate drops you from 1 to 0.
An equilibrium is not necessarily the best outcome, as the example shows. It is a stable one, in the sense that nobody has a reason to move away alone.
Some games have no such stable pure choice. In rock-paper-scissors, if you always pick rock, your opponent can learn to play paper. The stable answer is to randomise, choosing each option an equal share of the time, so that nothing can be exploited. This idea of randomising is the core of mixed strategies, which only matter when no single fixed choice is safe.
Why it matters in poker and other games
In poker, game theory underlies the idea of balancing strong hands and bluffs so that opponents cannot easily exploit your play. The specific arithmetic of how often a defender must continue to make a bluff unprofitable is covered under minimum defense, and the logic of misdirection is covered under bluffing. Game theory also explains why simply copying what worked against one opponent is risky: another opponent may respond differently.
Game theory is also used in video game design, where designers try to ensure that no single option dominates all others, and in competitive settings that change over time, as discussed under metagame shifts. The same thinking shows up in decision tools such as decision trees, which map choices and outcomes for one decision-maker facing uncertainty.
Game theory versus probability
Probability describes uncertainty about chance events, such as the roll of a die. Game theory describes interaction with other decision-makers who respond to you. Many real situations contain both. A poker hand includes random cards and a human opponent who adjusts, so each tool covers part of the picture: probability basics handle the chance side, and game theory handles the strategic side.
Limits to keep in mind
Game theory is a model, and models simplify.
- Real people are not perfectly rational. Opponents make mistakes, so an “equilibrium” play is not always the most profitable one against a specific weak player.
- Payoffs are hard to measure. The numbers in a table are assumptions. Real payoffs include feelings, reputation and long-term effects.
- Games repeat. In repeated play, players can reward cooperation or punish defection, so the one-shot result does not always apply.
- Many games are huge. Solving a full game exactly can be impractical, and only approximations are possible.
Nor does game theory make gambling profitable. In casino games with a house edge, no strategy removes the edge, and gambling can lead to real losses.
Takeaway
Game theory gives a tidy vocabulary for strategic choices: players, strategies and payoffs, then dominant strategies and equilibrium. Its main lesson is that the best action depends on what others are likely to do, and that sensible individual choices do not always lead to the best shared result. Use it as a way to think more clearly, not as a formula for guaranteed wins.