Every strategic decision you make depends on what someone else decides to do — and their decision depends on what they think you'll decide. This recursive loop, where outcomes are determined not by individual choices but by the interaction of multiple rational agents, is the territory game theory maps. It's the mathematics of interdependence, and it governs everything from nuclear deterrence to airline pricing to whether you should bluff in a poker hand.
John von Neumann and Oskar Morgenstern formalised the field in 1944 with Theory of Games and Economic Behavior, a 641-page treatise that reframed economics from the study of isolated agents maximising utility to the study of agents whose payoffs depend on one another's strategies. Von Neumann — a polymath who also contributed to quantum mechanics, computer architecture, and the Manhattan Project — had published the foundational minimax theorem in 1928, proving that in any two-person zero-sum game, there exists an optimal strategy for each player that minimises their maximum possible loss. The theorem was elegant but narrow: it applied only to situations where one player's gain was exactly another's loss. Real economic life was messier than that.
Morgenstern, an Austrian economist who had fled the Anschluss for Princeton, provided the economic intuition. His insight was that classical economics had a blind spot: it assumed each agent could optimise independently, as if their competitors' behaviour were a fixed feature of the landscape rather than a strategic response to their own moves. Adam Smith's invisible hand, Alfred Marshall's supply and demand curves, the entire neoclassical apparatus — all of it treated other market participants as environmental constants, like the weather. Morgenstern saw this as a fundamental error. A firm setting its price doesn't face a fixed demand curve. It faces a demand curve that shifts based on what competitors charge, which in turn depends on what they expect the first firm to charge.
The two men's collaboration produced the first rigorous framework for analysing situations where your best move depends on someone else's best response to your best move — an infinite regress that the mathematics was designed to tame.
The field's most celebrated result came a decade later, from a 21-year-old mathematics doctoral student at Princeton named John Nash. In 1950, Nash proved that every finite game with any number of players has at least one equilibrium — a set of strategies where no player can improve their outcome by unilaterally changing their own strategy, given what everyone else is doing. The Nash Equilibrium didn't require zero-sum conditions. It applied to cooperative and competitive settings alike. It was general enough to model arms races, price wars, auction design, evolutionary biology, and traffic congestion.
The proof was 27 pages long. Its implications reshaped half a dozen disciplines. Nash received the Nobel Prize in Economics in 1994, forty-four years after the proof, by which point game theory had colonised virtually every social science — from political science (voting systems, coalition formation) to evolutionary biology (John Maynard Smith's concept of the Evolutionarily Stable Strategy, published in 1973, was a direct application of Nash Equilibrium to natural selection).
The power of the framework lies in its core reorientation: stop asking "what is my best move?" and start asking "what is my best move given that my opponent is also trying to find their best move?" That shift — from optimisation to strategic interaction — is what separates game theory from decision theory. Decision theory works when you're playing against nature. Game theory works when you're playing against other minds. And in business, markets, negotiation, and politics, you are always playing against other minds.
The taxonomy matters for practical use. In zero-sum games, one player's gain is exactly another's loss — poker, military engagements, market share battles in fixed-size markets. In positive-sum games, cooperation can expand the total payoff — trade agreements, technology standards, platform ecosystems. In repeated games, the same players interact multiple times, and reputation, trust, and punishment become strategic variables that don't exist in one-shot encounters. Robert Axelrod's famous 1980s computer tournaments showed that in repeated Prisoner's Dilemma games, the simplest cooperative strategy — Tit-for-Tat — outperformed every sophisticated exploitative approach. The lesson for business: if you're in a game you'll play repeatedly, cooperation often dominates exploitation because the long-term cost of destroyed trust exceeds the short-term gain from defection.
The distinction between simultaneous games (where players choose without knowing the other's move, like sealed-bid auctions) and sequential games (where players move in turn and observe previous moves, like chess or market entry decisions) determines which analytical tools apply. Simultaneous games require mixed strategies and probability; sequential games require backward induction — reasoning from the final move backward to determine the optimal first move.
Nearly all business strategy unfolds as a sequential game, which is why "working backward from the endgame" is one of the most valuable habits a strategist can develop. When Bezos considered entering the cloud computing market in 2005, the critical question wasn't "is this a good business?" — it was "if we enter, how will IBM, Microsoft, and Google respond, and does our strategy still win given those responses?" The answer depended on backward induction from the eventual competitive equilibrium, not on a static market analysis.
What makes the field genuinely useful for founders, investors, and strategists isn't the mathematics itself — most practical applications don't require solving systems of equations. It's the mental discipline of modelling the other side's incentives, constraints, and likely responses before committing to a course of action. The founder who launches a price war without considering the competitor's cost structure and willingness to absorb losses is making a decision-theoretic move in a game-theoretic world. The investor who buys a stock without considering what the seller knows is bringing a calculator to a chess match.
Section 2
How to See It
Game theory operates wherever outcomes depend on the strategic interaction between two or more agents. The signature is any situation where your optimal choice changes depending on what the other party does — and they know it. Once you learn to see the game beneath the surface of competitive decisions, you'll find it operating in domains far beyond economics — from evolutionary biology to dating markets to corporate boardrooms.
Business
You're seeing Game Theory when two ride-sharing companies burn billions in subsidies to capture market share, each betting the other will exhaust funding first. Uber and Lyft spent a combined $8.6 billion in driver and rider subsidies between 2014 and 2019.
Neither could afford to stop unilaterally — the first to raise prices would lose drivers and riders to the competitor. This is a war of attrition, one of game theory's canonical models: both players suffer losses, and the winner is whoever can sustain the pain longest or credibly signal that they will. Uber's larger capital base and global diversification were strategic assets precisely because they made the "I can outlast you" signal credible.
Investing
You're seeing Game Theory whenWarren Buffett announces publicly that he will never sell a core holding. This isn't sentimentality — it's a credible commitment that changes the game for potential acquirers, competitors, and management teams.
When Berkshire acquired BNSF Railway for $34 billion in 2009, Buffett's known reputation for permanent ownership gave BNSF's management confidence to make 30-year capital investments in track, rolling stock, and terminals without worrying about a future owner stripping assets for short-term returns. The signal changed the behaviour of the other players in the game — management invested more aggressively, employees showed higher retention, and sellers increasingly preferred Berkshire as a buyer — which is exactly what credible commitments are designed to do.
Geopolitics
You're seeing Game Theory when nuclear-armed nations maintain arsenals they never intend to use. Mutually Assured Destruction — the doctrine that emerged from RAND Corporation game theorists in the 1950s, notably Herman Kahn and Thomas Schelling — is the highest-stakes Nash Equilibrium in history.
Neither superpower can gain by launching a first strike because the guaranteed retaliatory response makes the payoff negative for both sides. The weapons exist to make the game unwinnable, which is the point. During the Cuban Missile Crisis in October 1962, both Kennedy and Khrushchev were navigating a game-theoretic landscape where the rational move for each player depended entirely on the other's willingness to escalate — a "game of chicken" that Schelling had analysed abstractly just two years earlier. Schelling won the Nobel Prize in 2005 for demonstrating how the threat of mutual destruction, paradoxically, stabilises the system.
Technology
You're seeing Game Theory when Apple and Google distribute mobile operating systems for free (or near-free) to hardware manufacturers. The platform game is a coordination game: developers write apps for the platform with the most users; users choose the platform with the most apps.
Once a platform achieves critical mass, switching costs create a Nash Equilibrium where neither developers nor users benefit from moving unilaterally. Microsoft learned this in mobile: despite spending over $7 billion on Nokia and years of development, Windows Phone couldn't break the iOS/Android coordination equilibrium because the app gap made switching irrational for both developers and users. Apple's closed ecosystem and Google's open-licensing model are different strategies in the same coordination game, each optimising for a different payoff: Apple captures hardware margins, Google captures advertising data. Both are equilibria. Neither player can profitably deviate.
Section 3
How to Use It
Decision filter
"Before committing to a strategy, ask: what will my competitor, counterparty, or opponent do in response — and does my strategy still make sense given their most likely response? If your plan only works when the other side does nothing, it isn't a strategy. It's a wish."
As a founder
Map the competitive landscape as a game before entering it. Identify each player, their incentives, their constraints, and their likely responses to your moves.
The most common founder mistake is analysing the market as a static pie — "the TAM is $50 billion" — without modelling how incumbents will respond when you take a slice. A $50 billion TAM where the dominant player will match your price on day one is a fundamentally different opportunity than a $50 billion TAM where incumbents are structurally unable to respond.
When Peter Thiel co-founded PayPal in 1999, the online payments market had over a dozen competitors, including Billpoint (backed by eBay) and Citibank's c2it. Thiel's game-theoretic insight was that the market would consolidate around a single network — a winner-take-most coordination game — and that the critical variable wasn't product quality but adoption speed. PayPal spent $60–70 million on customer acquisition bonuses ($10 per signup, $10 per referral) to race past the tipping point where network effects made switching irrational. The bonuses were a strategic investment in changing the game's equilibrium, not a marketing expense.
The same logic applies to market entry timing. If you're entering a market with strong network effects, game theory tells you that being first to the coordination equilibrium matters more than being the best product. If you're entering a market with differentiated niches and weak network effects, the opposite is true — quality and positioning matter more than speed. Identifying the game's type before choosing your strategy is the first and most consequential decision a founder makes.
As an investor
Every market price is the output of a game between buyers and sellers. When you buy a stock, someone is selling it to you — and they have a reason.
The game-theoretic question an investor must always ask is: "What does my counterparty know or believe that I don't?"
George Soros built a $30 billion fortune by identifying games where one player was trapped. His 1992 bet against the British pound was a game-theoretic masterpiece: the Bank of England was committed to maintaining the pound's peg to the Deutsche Mark, but Britain's recession made the peg economically unsustainable. The Bank was a player with a dominant strategy (defend the peg) that was also a losing strategy (the defence required interest rates that deepened the recession). Soros identified the forced move — the Bank would eventually have to abandon the peg — and positioned accordingly. He made over $1 billion because he understood the other player's constraints better than the other player publicly acknowledged them.
The broader investment principle: every mispriced asset reflects a game where at least one player is constrained — by regulation, by mandate, by political pressure, or by legacy commitments — from making the rational move. The game-theoretic investor's edge isn't better information about the asset. It's better understanding of the constraints that force other players into predictable, suboptimal behaviour.
As a decision-maker
Use commitment devices to change the game's structure before it's played. Hernán Cortés burning his ships in 1519 is the canonical example — by eliminating his army's option to retreat, he changed the payoff matrix from "fight or flee" to "fight or die," which made his soldiers' commitment credible to the Aztecs and to themselves.
In business, credible commitments take the form of public announcements, contractual obligations, and irreversible investments. When Amazon built fulfilment centres across the country in the early 2000s — $2.5 billion in capital expenditure by 2005 — it was a commitment device. The infrastructure made Amazon's promise of fast, cheap shipping credible in a way that no press release could. Competitors could see the capital deployed and calculate that Amazon couldn't walk it back. Jeff Bezos's decision to keep prices low and accept thin margins was a similar signal: it told potential entrants that competing on price against Amazon would be a war of attrition against a player with demonstrated willingness to sacrifice margins indefinitely.
The key insight: a commitment is only credible if it's costly to reverse. A CEO who announces "we will never raise prices" makes a cheap commitment — words cost nothing and can be retracted. A CEO who builds $2.5 billion in logistics infrastructure that only makes financial sense at high volume makes an expensive commitment — the investment is sunk and the strategy is locked in. The game-theoretic value of commitment devices is proportional to their irreversibility.
Common misapplication: The most dangerous misuse of game theory is assuming all players are perfectly rational. The models produce elegant equilibria under the assumption that every participant correctly identifies their optimal strategy and executes it. In practice, people are emotional, poorly informed, overconfident, and frequently irrational. A CEO who models a competitor as a rational profit-maximiser may be blindsided when that competitor makes a spite-driven, value-destroying move to "punish" a perceived slight. Game theory is a map of how rational agents should behave, not a prediction of how actual humans will behave. The gap between those two things is where most strategic surprises originate.
Behavioural game theory — the empirical study of how people actually play games — consistently shows that humans cooperate more than rational self-interest predicts, punish unfairness even at personal cost, and systematically miscalculate probabilities in mixed-strategy situations. Use the rational model as a baseline, not a gospel.
A second misapplication is equally common: treating a cooperative game as zero-sum. Founders who view every negotiation with investors, partners, or early employees as a fixed-pie division game systematically underinvest in relationship-building and value-creation, because they're optimising for their share of the existing pie rather than expanding the pie. The best dealmakers — Buffett's partnership structures, Bezos's supplier relationships — understand that most business interactions are positive-sum games where the total payoff depends on the level of cooperation, not just its distribution.
Section 4
The Mechanism
Section 5
Founders & Leaders in Action
The founders and leaders who deploy game-theoretic reasoning don't typically draw payoff matrices on whiteboards. They think in terms of incentives, commitments, and opponent responses — the practical vocabulary of strategic interaction. What connects these cases — spanning payment networks, retail empires, Cold War diplomacy, semiconductor wars, and Napoleonic battlefields — is the same discipline: before acting, model the other side.
The pattern is remarkably consistent across centuries and domains: the decisive advantage goes not to the player with the best resources but to the player who best understands the game's structure and their opponent's constraints within it.
The online payments wars of 1999–2000 were a coordination game with a dozen players and a single equilibrium: one network would dominate, and the rest would die. Thiel understood this structure before his competitors did. PayPal's $10 sign-up bonuses and $10 referral bonuses — which cost the company an estimated $60–70 million — were not marketing. They were a game-theoretic investment in reaching the tipping point of a coordination game.
The critical insight was about eBay. PayPal embedded itself into eBay's auction marketplace by making it trivially easy for sellers to accept payments. Once a critical mass of eBay sellers used PayPal, buyers were forced to adopt it, which attracted more sellers — a positive feedback loop that made the equilibrium self-reinforcing. eBay's own payment product, Billpoint, couldn't compete despite eBay's platform advantage, because PayPal had reached the coordination equilibrium first. eBay's eventual acquisition of PayPal for $1.5 billion in 2002 was an acknowledgement that the game was over.
Thiel later generalised this reasoning in Zero to One: the most valuable companies are monopolies that have escaped competition entirely. His argument is game-theoretic at its core — competitive markets are games where no player earns excess returns, while monopolies are games with a single player and no strategic interaction required. The PayPal experience taught Thiel that winning a competitive game is expensive and uncertain; the far superior move is to find a game with only one player — or to make your game unplayable for anyone else.
Bezos's pricing strategy at Amazon is a textbook credible commitment. From the late 1990s onward, Bezos publicly and repeatedly stated that Amazon would prioritise low prices over margins — a signal designed to change the competitive game's structure. By accepting low margins and reinvesting aggressively, Bezos made it clear that entering Amazon's markets would require a willingness to fight a war of attrition against an opponent with no intention of raising prices.
The game-theoretic effect was deterrence. Potential competitors had to calculate not just whether they could match Amazon's prices today, but whether they could sustain those prices against a competitor with demonstrated willingness to sacrifice profitability for decades. Most rational entrants concluded the game wasn't worth playing. Between 2000 and 2015, Amazon's operating margins averaged under 3%, while the company invested over $100 billion in infrastructure. The low margins weren't a weakness. They were a strategic choice that made Amazon the dominant player in a game where the entry cost was unsustainable losses.
Bezos's 2005 launch of Amazon Prime extended the commitment device. The $79 annual membership made customers' shopping consolidation rational and competitors' customer acquisition more expensive — changing the payoff matrix for every retailer in the country. By 2023, with over 200 million Prime members spending an average of $1,400 per year versus $600 for non-members, the programme had achieved what game theorists call a "dominant equilibrium" — a state so stable that no player benefits from deviating.
Kissinger's approach to Cold War diplomacy was explicitly game-theoretic — he had studied the field at Harvard under Thomas Schelling and applied its logic to the most consequential strategic interactions of the 20th century. His doctoral thesis, A World Restored (1957), analysed the Congress of Vienna as a multi-player negotiation game, and the analytical framework he developed there shaped his approach to every major diplomatic initiative that followed. The opening to China in 1971 was a masterstroke of game-theoretic repositioning.
The Cold War was a two-player game between the United States and the Soviet Union — or so the conventional analysis assumed. Kissinger recognised that this framing itself was a strategic error. The Sino-Soviet split, which had been deepening since Khrushchev's denunciation of Stalin in 1956 and escalated to border clashes along the Ussuri River in 1969, meant there were actually three significant players. Kissinger transformed the acknowledged game by establishing relations with Mao Zedong's China through secret diplomacy beginning in July 1971.
The move changed every payoff matrix in the Cold War. The Soviet Union now had to calculate Chinese reactions to any move against the United States, and vice versa. Both Communist powers feared the other aligning with Washington.
The result was détente. The Soviet Union became more willing to negotiate arms limitation treaties (SALT I, signed in 1972) because the alternative — continued confrontation with both the U.S. and China — was strategically untenable. Kissinger didn't change the players' preferences. He changed the structure of the game by adding a player, which altered every equilibrium in the system.
The broader lesson extends beyond geopolitics: when you're losing a two-player game, consider whether a third player can be introduced to change the dynamics. In venture capital, a startup negotiating with a single potential acquirer is in a weak bilateral game. Introducing a second interested buyer transforms it into an auction — a fundamentally different game with a fundamentally different equilibrium price. Kissinger applied the same principle at civilisational scale.
Grove's competitive strategy at Intel was a sustained exercise in game-theoretic thinking. When AMD cloned Intel's x86 architecture and began selling compatible processors at lower prices in the late 1980s, the naive response was a price war. Grove recognised this as a Prisoner's Dilemma: mutual price cuts would destroy margins for both firms, benefiting only customers and OEMs.
Instead, Grove played a different game. He accelerated Intel's product cycle, releasing new processor generations (486, Pentium, Pentium Pro) faster than AMD could clone the previous one. This turned the game from a price competition — which AMD could win on cost structure — into an innovation race, where Intel's R&D budget ($1.8 billion in 1995, roughly four times AMD's revenue) was an insurmountable advantage. Grove also launched the "Intel Inside" branding campaign in 1991, a $500 million investment that created consumer-level demand for Intel chips, making OEMs reluctant to substitute AMD products even at lower prices. The campaign changed the game by adding a new player — the end consumer — whose preferences favoured Intel.
When AMD launched the budget Cyrix-competitive K5 processor, Grove responded not with across-the-board price cuts but with the Celeron — a deliberately limited low-end chip that competed at the bottom without cannibalising Pentium margins. The segmentation preserved Intel's pricing power where it mattered most. Grove's strategic genius was recognising that the game against AMD had multiple dimensions — price, performance, brand, ecosystem — and that the winning move was to compete aggressively on the dimensions where Intel held structural advantages while refusing to engage on the dimension where AMD had the edge.
Napoleon's military campaigns demonstrate game-theoretic reasoning applied under the most extreme time pressure and information constraints. His strategy at Austerlitz in December 1805 — widely considered one of the greatest tactical victories in military history — was a deliberate manipulation of his opponents' beliefs about his strategic position.
Facing a combined Austro-Russian force of approximately 85,000 men with only 73,000 of his own, Napoleon deliberately weakened his right flank near the villages of Telnitz and Sokolnitz and feigned retreat from the Pratzen Heights, creating the appearance of vulnerability. He even sent a diplomatic envoy to the Allied camp requesting an armistice — a signal of weakness designed to reinforce the illusion.
The Allied commanders, believing they could envelop Napoleon's weakened right, committed the bulk of their forces to the attack — exactly as Napoleon intended. This created a gap in the Allied centre on the Pratzen Heights, which Napoleon struck with his reserves under Marshals Soult and Bernadotte, splitting the enemy army in two and routing both halves. The battle was over within hours. Allied casualties exceeded 36,000; French losses were approximately 9,000.
The game-theoretic principle: Napoleon didn't just choose the best strategy given his opponent's position. He chose a strategy that manipulated his opponent into choosing a response that was optimal for Napoleon. He was playing a meta-game — selecting his own apparent weakness to exploit his opponent's pattern-matching. Schelling would later formalise this as the strategic use of information asymmetry, where controlling what the other player believes you know is as valuable as what you actually know. The broader pattern across Napoleon's campaigns was consistent: he won by understanding the decision calculus of the opposing commander and designing situations where their "rational" response to the visible information led them into a trap defined by the invisible information.
Section 6
Visual Explanation
The Prisoner's Dilemma — The most studied game in theory. Individual rationality leads to collective irrationality. Both players defecting is the Nash Equilibrium, but both cooperating produces a better outcome for each.
Section 7
Connected Models
Game theory doesn't stand alone — it connects to frameworks that address incentive structures, competitive dynamics, and strategic reasoning. Some amplify its insights, some push back productively, and some represent where game-theoretic reasoning naturally leads.
Here's how it connects to the broader lattice of models:
Game theory's foundational premise is that agents respond to incentives. Incentive-Caused Bias shows what happens when those incentive structures produce systematically distorted behaviour. Charlie Munger's observation — "Show me the incentive and I will show you the outcome" — is game theory compressed into a sentence. When a real estate agent's commission structure incentivises quick sales over optimal prices, the agent is playing a rational game whose equilibrium happens to harm the client. Understanding game theory sharpens your ability to identify misaligned incentives before they produce misaligned behaviour. The two models together form a complete toolkit: game theory maps the strategic landscape, incentive-caused bias flags where the landscape is rigged.
Game theory is inherently second-order: your move depends on their response, which depends on their expectation of your move. The entire framework collapses without recursive reasoning about consequences. Second-order thinking provides the cognitive habit; game theory provides the formal structure. When Andy Grove traced the consequence chain from a price war with AMD through margin compression through reduced R&D through lost performance leadership, he was applying second-order thinking within a game-theoretic frame. The two models reinforce each other because game theory without second-order thinking produces shallow analysis (just the first move), and second-order thinking without game theory ignores the strategic responses of other players.
Tension
[[Competition](/mental-models/competition) is for Losers](/mental-models/competition-is-for-losers)
Section 8
One Key Quote
"Real life consists of bluffing, of little tactics of deception, of asking yourself what is the other man going to think I mean to do. And that is what games are about in my theory."
— John von Neumann, as recounted by Jacob Bronowski in The Ascent of Man (1973)
Section 9
Analyst's Take
Faster Than Normal — Editorial View
Game theory is the mental model I reach for whenever I see a founder treating their competitive environment as a spreadsheet problem rather than a strategic interaction. The spreadsheet tells you about your own costs, margins, and growth rate. It tells you nothing about what the other side will do when you make your move. And in markets, it's always the other side's response that determines the outcome.
I've sat in hundreds of strategy sessions where teams analyse their own position in exquisite detail — unit economics, cohort analysis, CAC payback periods — and spend zero time analysing the competitive dynamics that will determine whether any of those numbers survive contact with reality. The analytical sophistication is directed entirely inward. Game theory redirects it outward, toward the interaction that actually determines results.
The single most common strategic error I encounter is the failure to model the opponent. A startup announces it will undercut the incumbent on price. I ask: "What does the incumbent do next?" The founder looks confused, because in their mental model the incumbent stands still while customers migrate. That has never happened. The incumbent responds — with price cuts, bundling, exclusive contracts, or acquisition offers — and the startup's plan, which assumed a static environment, fails on contact with a dynamic one. Game theory doesn't guarantee you'll win. It guarantees you won't be blindsided by the obvious countermove.
The founders who deploy game-theoretic reasoning most effectively share a trait: they think in terms of the game's structure, not just their position within it. Bezos didn't just compete in retail — he restructured the game by making low margins a commitment device that deterred entry. Thiel didn't just build a better payment product — he raced to the coordination equilibrium that made competing pointless. Grove didn't just respond to AMD — he changed the dimension of competition from price to innovation velocity. In each case, the insight wasn't about making a better move within the existing game. It was about changing the game itself.
The limitation I flag most often: game theory assumes the game is known. In stable, well-defined competitive environments — airlines, telecom, commodity markets — the model is powerful. The players are identified, the payoff structures are observable, and the strategy space is constrained.
But in genuinely novel markets — the early internet, the current AI landscape, emerging biotech platforms — the game isn't yet defined. The players are unknown, the payoff structures are shifting, and the strategy space is unbounded. Applying rigorous game-theoretic analysis to an undefined game produces precise answers to the wrong question. This is the domain where Thiel's advice to avoid competition entirely becomes most relevant — when you can't even identify the game, the safest move is to create one where you're the only player.
Section 10
Test Yourself
These scenarios test whether you can identify game-theoretic dynamics — the strategic interactions, equilibria, and commitment mechanisms that separate interactive reasoning from isolated decision-making. The ability to spot the game beneath the surface of a business or policy decision is a skill that develops with practice.
Is this mental model at work here?
Scenario 1
Two coffee shops open on the same block. Each considers whether to lower prices. Both reason: 'If I cut prices and they don't, I capture their customers. If they cut and I don't, I lose mine.' Both cut prices. Both end up with the same market share as before, but at lower margins.
Scenario 2
A pharmaceutical company spends $800 million developing a drug, then prices it at $50,000 per treatment. A generic manufacturer announces plans to produce a bioequivalent version at $5,000 once the patent expires. The pharmaceutical company files additional patents on delivery mechanisms and dosage forms, extending its effective exclusivity by seven years.
Scenario 3
A venture capital firm offers a startup a term sheet at a $50 million valuation. The founder immediately shares this with two competing VCs to generate a bidding war. One competitor offers $65 million. The original firm walks away entirely rather than participate in an auction.
Scenario 4
A software company releases its core platform as open source, giving away the code that took five years and $200 million to develop. Revenue comes from enterprise support, hosting, and premium features built on top of the free platform.
Section 11
Top Resources
The best resources on game theory span the original mathematics, the strategic applications, and the behavioural evidence that complicates the elegant models. Start with Schelling for practical strategy, then build formal foundations with von Neumann and Dixit.
The most practically useful book on game theory ever written, and it contains almost no equations. Schelling introduced focal points, credible commitments, and the strategic value of limiting your own options — concepts that transformed how diplomats, military planners, and business strategists think about competitive interaction. His analysis of nuclear deterrence remains the canonical treatment. If you read one book on game theory, make it this one.
The founding document of the field. Dense and mathematical, but the first chapter — which explains why economics needed a theory of strategic interaction — remains one of the clearest articulations of the problem game theory solves. Von Neumann's minimax theorem and the expected utility framework established here still underpin decision theory and strategic analysis eighty years later.
The best accessible introduction to applied game theory for business and everyday decision-making. Dixit (Princeton) and Nalebuff (Yale) cover commitment strategies, signalling, brinkmanship, voting, and bargaining with real-world examples and minimal mathematics. Updated in their 2008 follow-up The Art of Strategy. Start here if Schelling feels too Cold War-focused.
Axelrod's computer tournaments — where simple strategies competed in iterated Prisoner's Dilemma games — produced one of social science's most celebrated results: Tit-for-Tat, the strategy of cooperating first and then mirroring the opponent's previous move, outperformed every complex strategy submitted. The book transformed thinking about cooperation, trust, and reputation in competitive environments. Essential reading for anyone managing partnerships, alliances, or repeated competitive interactions.
Thiel's central argument — that competition destroys value and the goal is monopoly — is the most provocative practical application of game-theoretic reasoning in modern business writing. His framework reframes competition not as a virtue to be pursued but as a game-theoretic trap to be escaped. Read alongside Schelling and Axelrod for the complete picture: Schelling teaches you to play the game, Axelrod teaches you when to cooperate, and Thiel teaches you when the best move is to stop playing entirely.
Peter Thiel argues that the rational move isn't to play the competitive game better but to escape competition entirely — to build a monopoly where there are no other players. This directly contradicts game theory's premise, which is that strategic interaction between players is the fundamental condition to be managed. The tension is productive: game theory helps you win competitive games, but Thiel's insight is that the highest-value move may be to avoid playing altogether. The resolution depends on the game's structure. In winner-take-all markets, Thiel is right — competition destroys value. In markets with differentiated niches or cooperative equilibria, game-theoretic strategy creates value.
First principles thinking decomposes problems to fundamental physical or economic truths. Game theory depends on modelling other agents' behaviour — which is inherently uncertain and non-derivable from first principles. You can derive the cost of rocket fuel from commodity prices. You cannot derive your competitor's risk tolerance, emotional state, or information set from any fundamental truth. The tension: first principles gives you confidence about the physical world, but game theory deals with the social world where other minds introduce irreducible uncertainty. The best strategists use both — first principles for what physics allows, game theory for what competitors will do.
Leads-to
BATNA
Your Best Alternative To Negotiated Agreement is a game-theoretic concept in disguise. Every negotiation is a game where each party's power depends on their outside option. BATNA formalises this: the player with the better alternative to making a deal has more leverage because they can credibly walk away. Game theory explains why — a player who can't walk away is trapped in a game where the other side sets the terms. Kissinger's opening to China improved America's BATNA in negotiations with the Soviet Union, because the Soviets now faced the possibility that the U.S. would simply redirect its diplomatic attention to Beijing.
Leads-to
[Moats](/mental-models/moats)
Game-theoretic reasoning leads directly to the concept of economic moats — durable competitive advantages that prevent competitors from profitably entering your game. A moat changes the game's structure: it makes the entry cost so high that rational players choose not to play. Network effects, switching costs, economies of scale, and regulatory barriers are all moat types that game theory explains mechanistically. Buffett's insistence on investing only in companies with wide moats is an implicitly game-theoretic criterion: he wants to own businesses where the competitive game has been structurally resolved in the incumbent's favour and new entrants face a payoff matrix that discourages participation.
The behavioural dimension matters more than most practitioners acknowledge. Classical game theory assumes rational actors. Real competitors are driven by ego, fear, legacy commitments, and internal politics. The most dangerous opponents aren't the rational ones — they're the irrational ones who will destroy value to avoid losing. When Yahoo rejected Microsoft's $44.6 billion acquisition offer in 2008, it was an irrational move by classical game-theoretic standards — the offer represented a 62% premium to the stock price. But the decision was driven by Jerry Yang's emotional attachment to the company, not by payoff maximisation. The rational model would have predicted acceptance. The outcome was rejection, followed by Yahoo's continued decline. Model the incentives, yes. But also model the egos.
This is where the academic theory and the practical application diverge most sharply. Academic game theory produces beautiful equilibria under assumptions of perfect rationality and common knowledge. Practical game theory — the kind Kissinger, Grove, and Bezos practiced — accounts for the fact that your opponent may not know their own payoff matrix, may have internal politics that override rational strategy, and may act on information you can't observe. The gap between the textbook game and the actual game is where the most profitable strategic insights live.
There's also a temporal dimension that static models miss. The game changes as it's being played. Amazon's competitive game in 2000 was fundamentally different from its game in 2010, which was different from its game in 2020. The players changed (Walmart entered e-commerce seriously around 2016), the payoff structures changed (AWS became Amazon's profit engine), and the strategic variables changed (from price competition to logistics speed to ecosystem lock-in). The best game-theoretic thinkers don't just solve the current game — they anticipate how the game itself will evolve and position for the game they'll be playing in five years, not the game they're playing today.
One underappreciated dimension: game theory's value increases dramatically with the stakes of the decision. For routine competitive moves — a minor feature launch, a small pricing adjustment — the overhead of full strategic modelling isn't justified. For market entry decisions, platform strategy, major acquisitions, or pricing architecture, the failure to model opponent responses is genuinely reckless. The Uber/Lyft subsidy war burned $8.6 billion in combined capital. The founders who understood this as a war of attrition — a game with a known structure and known endgame — could plan for it. The ones who thought they were simply "acquiring customers" were playing a game they hadn't bothered to identify.
My strongest conviction about game theory's practical value: it's most useful as a pre-commitment discipline. Before you make a strategic move, force yourself to write down the three most likely responses from each significant competitor. If your strategy survives all three responses, proceed. If it only works when competitors do nothing, stop. That thirty-minute exercise, performed consistently, is worth more than a semester of formal game theory. The math is secondary. The habit of modelling the other side is everything.
