apply-nash-equilibrium

Apply Nash Equilibrium

Identify the stable strategy profile in a multi-party situation — the point where no party gains by changing their strategy unilaterally — then either play to it or design the game so the equilibrium lands where you want.

Why This Is Best Practice

Adopted by: McKinsey, BCG, and Bain apply game theory frameworks in competitive strategy engagements. Standard curriculum at Harvard, Wharton, Stanford, and INSEAD MBA programs. Used by the FCC for spectrum auction design, by investment banks structuring M&A negotiation strategy, and by energy companies (OPEC, BP, Shell) for production quantity decisions. Milgrom and Wilson's auction designs — direct applications of Nash equilibrium analysis — are now the global standard for spectrum, electricity, and emissions permit markets.

Impact: FCC spectrum auctions designed using Nash equilibrium auction theory have raised over $100 billion in US spectrum sales since 1994. Milgrom and Wilson received the 2020 Nobel Prize in Economics for "improvements to auction theory and inventions of new auction formats" — work that rests entirely on Nash equilibrium analysis. Dixit and Nalebuff's business applications have been assigned reading at over 200 MBA programs worldwide since 1991.

Why best: Intuition-based strategy, first-mover heuristics, and zero-sum thinking all fail under interdependence — what you should do depends on what others will do, which depends on what you do. Nash equilibrium is the only framework that provides a stable fixed point for multi-party reasoning. Unlike minimax (zero-sum only), it applies to non-zero-sum games where partial coordination, mixed motives, or repeated interaction may be stable. Unlike pure negotiation frameworks, it predicts where rational parties will actually end up, not just where you want them to be.

Sources: Nash (1951); Milgrom (2004); Nobel Prize Committee Scientific Background (2020); Dixit & Nalebuff (1991)

Steps

1. Map the game

Define the structure before solving it:

Element	Question	Example
Players	Who are the rational decision-makers?	Competitor A, Competitor B, Regulator
Strategies	What can each player choose?	Price high / Price low; Bid X / Bid Y
Payoffs	What does each player receive for each strategy combination?	Profit matrix, market share
Information	Does each player know others' payoffs?	Complete vs. incomplete information
Timing	Do players move simultaneously or sequentially?	Simultaneous = normal form; Sequential = extensive form

Use a payoff matrix for 2-player simultaneous games. Use a game tree for sequential games.

2. Check for dominant strategies first

A dominant strategy is always best regardless of what others do — no equilibrium analysis needed.

• Strictly dominant: strategy X always yields higher payoff than Y for every possible opponent move → play X, expect opponent to do the same
• Strictly dominated: strategy X always yields lower payoff → eliminate it; a rational opponent will never play it

Eliminate dominated strategies iteratively (IESDS) to simplify before solving.

3. Find the Nash equilibrium

For each strategy profile (combination of one strategy per player):

"Given everyone else's strategy, does any player want to switch?"

• If no player wants to switch → it is a Nash equilibrium
• If any player wants to switch → it is not

Mark Nash equilibria in the payoff matrix by checking each cell: underline the best response for each player. A cell where all players' payoffs are underlined is a Nash equilibrium.

Multiple equilibria: If more than one exists, use focal points (Schelling), communication, or repeated-game history to predict which one emerges.

No pure-strategy equilibrium: Compute mixed-strategy equilibrium — players randomize between strategies at probabilities that make the opponent indifferent. Common in zero-sum games (poker, penalty kicks).

4. Assess your current position

Compare where you are now to the Nash equilibrium:

Current position	Implication
Already at equilibrium	Stable — no rational pressure to change
Not at equilibrium	Expect drift; determine direction and speed
Prisoners' Dilemma equilibrium	Stable but suboptimal — explore repeated-game cooperation or binding contracts
Multiple equilibria	Coordination problem — who moves to which? Signal your intention early

For Prisoners' Dilemma structures (e.g., price wars, arms races): the Nash equilibrium is mutually bad. Escape routes: repeated interaction with credible punishment, enforceable contracts, or regulatory intervention.

5. Design the game (mechanism design)

If you control the rules — pricing structure, auction format, contract terms, platform incentives — design payoffs so your desired outcome is the Nash equilibrium.

• Make defection unprofitable: increase payoff for cooperation or add penalties for deviation
• Change information structure: reveal information that shifts payoffs (e.g., publish price lists to deter undercutting)
• Sequence moves: if you can commit first (Stackelberg), you become the leader and shift the equilibrium in your favour
• Create outside options: a credible BATNA changes your payoff structure and shifts the equilibrium

Ask: "If all parties play rationally given the rules I'm designing, where do they end up?" Adjust rules until that endpoint is where you want them.

Rules

• Map the game before solving it — skipping Step 1 leads to solving the wrong game
• Only eliminate strictly dominated strategies (weakly dominated strategies require caution in multi-player games)
• A Nash equilibrium is a prediction, not a prescription — rational play converges here, not necessarily where you want to be
• Repeated games change everything: cooperation can be sustained as equilibrium via credible punishment (Folk Theorem) — apply single-shot analysis only to one-shot interactions
• When in doubt about payoffs, use scenario ranges rather than point estimates

Examples

1. Spectrum auction design (US FCC, 1994) Problem: government needed to allocate radio spectrum efficiently. Prior lotteries and hearings wasted allocation. Milgrom and Wilson designed a Simultaneous Multiple Round Auction (SMRA) where Nash equilibrium analysis showed bidders would reveal true valuations and complementary licences would aggregate correctly. Result: $617M raised in first auction (vs. lottery baseline of $0 market value discovery). Global standard since adopted.

2. Airline pricing (oligopoly) Two carriers serve a route. Each can price High ($300) or Low ($200). If both price High, both earn $10M. If one cuts to Low while other stays High, cutter earns $15M, other earns $3M. If both cut, both earn $6M. Nash equilibrium: both price Low (classic Prisoners' Dilemma). Escape: signal commitment to High pricing through public fare announcements — if credible and reciprocated within 24h, coordination on High equilibrium holds.

3. Salary negotiation (sequential game) Candidate moves first (states reservation price or lets employer anchor). Sequential structure means first-mover can commit to a position, constraining the other party's best response. Analysis: stating a high anchor shifts the Nash equilibrium of the negotiation upward. Counter-offer strategy is determined by working backwards from both parties' outside options (BATNA).

Common Mistakes

• Solving the wrong game: defining strategies too broadly ("compete aggressively") makes payoffs unmeasurable. Strategies must be discrete and payoffs estimable.
• Ignoring repeated interaction: applying one-shot equilibrium to ongoing relationships. Repeated games sustain cooperative equilibria that one-shot analysis says are unstable.
• Assuming complete information: real markets often have private information (costs, valuations). Switch to Bayesian Nash equilibrium when payoffs are privately known.
• Stopping at equilibrium identification: Nash equilibrium tells you where things stabilise, not how to get there. Add a transition / mechanism design step.
• Confusing equilibrium with optimum: the Nash equilibrium is often Pareto-suboptimal (Prisoners' Dilemma). Identifying the equilibrium reveals the problem — it does not solve it.