“Game theory is a branch of applied mathematics that studies strategic interactions among rational decision-makers, where the outcome for each participant depends on the choices made by all. It structures these interactions as ‘games’ consisting of players, their available strategies, and the resulting payoffs.” – Game theory – Applied mathematics
Strategic interdependence changes the logic of decision-making. When one person’s result depends on what others do, choosing well requires more than comparing private costs and benefits. It requires anticipating responses, recognising incentives, and assessing how an entire pattern of choices may produce an outcome that no participant intended. This is why game theory sits between mathematics, economics, politics, biology, psychology and computer science: it offers a disciplined way to analyse situations in which actions are mutually connected.1
A formal game normally identifies players, the actions or strategies available to them, the information they possess, the timing of decisions, and the payoffs associated with possible outcomes. A player may be an individual, a firm, a state, an algorithm, or a population of organisms. A strategy is not merely a single move. In an extensive game, it is a plan specifying what a player would do at every decision point that might be reached. Payoffs represent preferences over outcomes and need not be limited to money; they may include security, reputation, market share, votes, survival or fairness.1
The distinction between a decision problem and a game is practical. A firm with a guaranteed monopoly can often select output by comparing prices and costs without modelling a rival’s reaction. In an oligopoly, by contrast, a price cut may provoke retaliation, a capacity expansion may deter entry, and a public commitment may alter competitors’ expectations. The relevant question becomes not simply which action is best, but which action remains best after other players respond. Game theory therefore examines best responses and the feedback loops generated by interdependent choices.
How the mathematics works
A strategic-form game can be represented by a set of players, a strategy set for each player, and a payoff function for each player. For player i, the payoff can be written as u_i(s_i,s_{-i}), where s_i is that player’s strategy and s_{-i} denotes the strategies chosen by everyone else. A Nash equilibrium is a profile s^* satisfying u_i(s_i^*,s_{-i}^*) \geq u_i(s_i,s_{-i}^*) for every player i and every feasible alternative s_i. In plain language, no player can gain by changing strategy alone while the others keep theirs fixed.9,12
This condition is important but limited. It defines stability against unilateral deviation, not social desirability, fairness or truth. Several equilibria may exist, and some may be inefficient for everyone. In the prisoners’ dilemma, individually rational choices can generate mutual defection even though both players would prefer mutual cooperation. Mixed strategies extend the framework by allowing players to randomise, with probabilities chosen so that opponents are indifferent or best responding. Nash proved that every finite non-cooperative game has at least one equilibrium when mixed strategies are permitted, a result that established a broad existence guarantee rather than a universal prediction of behaviour.9,15
Major traditions and competing assumptions
Non-cooperative game theory studies situations in which agreements are not automatically enforceable. Its central tools include Nash equilibrium, subgame-perfect equilibrium for sequential decisions, and refinements that eliminate implausible threats or implausible beliefs. Cooperative game theory instead focuses on coalitions, bargaining and how gains can be distributed among groups. In a cooperative model, the analytical problem may concern whether a coalition can form, whether an agreement is stable, and how its surplus should be allocated rather than which isolated action each player selects.
Evolutionary game theory relaxes the idea that every player calculates perfectly. It studies how strategies spread when their relative success affects reproduction, imitation or persistence. A strategy can survive because it performs well against the population, even if no organism consciously understands the underlying incentives. Behavioural game theory goes further by testing formal predictions against laboratory and field evidence. People frequently use heuristics, respond to framing, value reciprocity, punish unfairness, and make systematic errors. These findings do not make mathematical models useless; they clarify which assumptions are descriptive approximations and which are normative benchmarks.
Applications and practical meaning
In economics, game theory helps explain competition, auctions, bargaining, contracting and market entry. Regulators use it to assess how firms may respond to rules, disclosure requirements or penalties, although the quality of any conclusion depends on the specified information and payoffs. In politics and international relations, deterrence, negotiation and arms control involve credibility: a threat matters only if carrying it out remains rational when the relevant moment arrives. In public policy, the framework exposes free-rider problems, in which each actor benefits from a shared resource while preferring others to bear the cost of maintaining it.
Technology has expanded the range of applications. Network users affect congestion, platforms design rules for interacting sides of a market, and online systems must anticipate strategic manipulation. Mechanism design reverses the usual direction of analysis: rather than asking what players will do under fixed rules, it asks how rules can be constructed to produce desired outcomes despite self-interested behaviour. This approach informs auction design, matching markets and elements of digital governance. Computer science also studies algorithmic games in which computation, information asymmetry and automated agents alter the meaning of speed, commitment and strategic adaptation.15
Where the framework can fail
The hardest modelling choice is often not solving the game but defining it. Payoffs may be uncertain, incomplete or socially contested. Players may misunderstand the situation, lack the capacity to calculate, or care about norms that are difficult to encode. A model that treats reputation as irrelevant can misread repeated interaction; a model that assumes perfect information can overlook bluffing and signalling; a model that assigns a single objective to a government or firm can conceal internal conflict. Predictions are therefore conditional: they describe what follows from a specified structure, not what must happen in reality.
Equilibrium also says less than its technical authority can suggest. Existence does not guarantee uniqueness, stability over time or empirical accuracy. If several equilibria are possible, institutions, history, focal points and communication may determine which one emerges. If the game changes as players learn, then a one-shot solution may be inappropriate. Repeated games, reputation models and learning dynamics address some of these problems, but they introduce additional assumptions. The practical discipline is to compare alternative specifications, conduct sensitivity analysis, and distinguish an equilibrium prediction from evidence about actual conduct.
Why it still matters
Game theory remains valuable because it makes hidden dependence explicit. It forces analysts to ask who can act, what each participant wants, what information is available, which commitments are credible, and how incentives change after each response. Its greatest contribution is not the promise of perfect prediction. It is a common language for analysing conflict, cooperation and coordination across fields that otherwise describe similar problems in incompatible terms. Used carefully, it reveals why individually sensible choices can create collective problems and how altered rules, information or payoffs can change the set of available outcomes.
References
1. Game Theory – Stanford Encyclopedia of Philosophy – 1997-01-25 – https://plato.stanford.edu/archives/spr2016/entries/game-theory/
2. Game Theory – Stanford Encyclopedia of Philosophy – https://plato.stanford.edu/archives/fall2009/entries/game-theory/
3. Game Theory – Stanford Encyclopedia of Philosophy – https://plato.stanford.edu/archives/win2007/entries/game-theory/
4. Game Theory – Stanford Encyclopedia of Philosophy – https://plato.stanford.edu/archives/fall2011/entries/game-theory/
5. Game Theory (Stanford Encyclopedia of Philosophy/Fall 2021 Edition) – 1997-01-25 – https://plato.stanford.edu/archives/fall2021/entries/game-theory/
6. Nash’s Game-Changing Idea Marks 75 Years – https://www.cmu.edu/math/news-events/articles/2025/1010_nash-equilibrium.html
7. History of Nash Equilibrium: Discovery and Use Today – 2023-12-12 – https://blogs.cornell.edu/info2040/2023/12/12/history-of-nash-equilibrium-discovery-and-use-today/
8. Game Theory – https://plato.stanford.edu/archives/win2004/entries/game-theory/
9. John F. Nash – Econlib – The Library of Economics and Liberty – 2023-08-31 – https://www.econlib.org/library/Enc/bios/Nash.html
10. Game Theory and Behavioral… – 1997-01-25 – https://plato.stanford.edu/entries/game-theory/
11. Game Theory and Ethics – Stanford Encyclopedia of Philosophy – 2021-09-27 – https://plato.stanford.edu/entries/game-ethics/
12. 3 Nash’s Equilibrium–Game theory’s foundation – 2006-09-21 – https://www.nationalacademies.org/read/11631/chapter/5
13. US game theory specialists win Nobel prize in economics – 2020-10-12 – https://www.theguardian.com/science/2020/oct/12/us-game-theory-specialists-win-nobel-prize-in-economics
14. PNAS Classics — Game Theory – https://www.kellogg.northwestern.edu/faculty/weber/decs-452/pnas%20classics%20–%20game%20theory.htm
15. Nash, Jr., John F. – https://www.informs.org/Explore/History-of-O.R.-Excellence/Biographical-Profiles/Nash-Jr.-John-F
