Competitive Game Theory¶

Game theory's epistemic status is complex: it excels at revealing strategic structure in conflict situations but often fails at precise prediction. Most real-world "games" violate its core assumptions (perfect rationality, common knowledge, known payoffs), yet the framework remains invaluable because it exposes hidden interdependencies that intuition misses.

The case for extracting from game theory despite these limitations: humans systematically fail at strategic reasoning. We treat competitive situations as independent optimization problems, ignoring how others' responses reshape our options. We confuse "doing better" with "winning." We fail to recognize when cooperation becomes impossible or when competition is wasteful. Game theory doesn't fix human irrationality, but it provides tools for seeing strategic structure that remains invisible to direct intuition.

The core insight: competitive situations have exploitable patterns. When your outcome depends on others' choices AND they're reasoning about your choices, specific structural regularities emerge. These patterns recur across domains - from geopolitics to product pricing to evolutionary biology - because the underlying strategic logic is identical.

Extraction principle: what survives even when the math fails? The mental operations for mapping strategic interdependencies, identifying stable outcomes, recognizing when first-mover or last-mover advantage applies, spotting situations where individual rationality produces collective disaster. These reasoning moves remain useful even when we can't calculate precise equilibria or when real humans violate rational actor assumptions. We're extracting the conceptual scaffolding, not the computational machinery.

Foundational Strategic Primitives¶

These tools reveal the basic structure of strategic interdependence - applicable across almost any competitive situation.

Strategic Structure Recognition¶

These tools identify types of games with characteristic strategic patterns - applicable when you need to diagnose what kind of interaction you're in.

Dynamic and Evolutionary Tools¶

These tools reason about how strategic interactions evolve over time and through repetition.

Applied Strategic Tools¶

These tools guide decision-making and intervention in competitive situations.

Quick Reference¶

Decision Type -> Tool Mapping¶

When you need to...	Use these tools
Map strategic interdependence in competitive situations	Payoff Matrix Construction, Best Response Mapping
Identify stable outcomes where no one wants to deviate	Nash Equilibrium Identification, Dominant Strategy Identification
Determine if cooperation is possible or pure competition applies	Zero-Sum vs. Positive-Sum Recognition, Coordination vs. Anti-Coordination Distinction
Solve sequential games with multiple stages	Backward Induction, Subgame Perfect Equilibrium Identification
Understand why cooperation persists despite incentives to defect	Repeated Game Recognition, Evolutionary Stability Checking
Design incentives to achieve desired outcomes	Mechanism Design Inversion, Commitment Device Design
Analyze situations with incomplete information about opponents	Complete vs. Incomplete Information Distinction
Recognize when timing of moves matters strategically	Sequential vs. Simultaneous Move Recognition
Find equilibria in games without pure strategy solutions	Mixed Strategy Equilibrium Finding

Usage Notes¶

Domain of Applicability¶

These tools work well in situations with:

Clear strategic interdependence: Your outcome depends on others' choices and vice versa.
Identifiable players and actions: You can specify who decides and what options they have.
Consequential choices: Actions have meaningful payoff differences that motivate careful choice.
Some stability: Repeated interactions or long-enough time horizons for equilibrium thinking to apply.

They work poorly when:

True preferences are unknowable: Not just incomplete information (which game theory handles) but fundamental preference uncertainty.
Actions are continuous and high-dimensional: Game theory can handle these but analysis becomes intractable without computational tools.
Radical innovation: Game theory assumes fixed game structure, but transformative innovation changes the game itself.
Players are fundamentally irrational: Behavioral economics shows systematic deviations from game-theoretic predictions.

Limitations¶

What these tools CANNOT do:

Predict behavior precisely: Game theory identifies equilibria but doesn't guarantee players reach them, especially in one-shot games.
Handle genuine cooperation: Game theory models cooperation as sustained by incentives, not intrinsic motivation or solidarity.
Resolve fundamental value conflicts: Game theory takes preferences as given; it doesn't adjudicate whose preferences should prevail.
Replace domain expertise: Strategic structure matters, but so does domain-specific knowledge about payoffs, constraints, and feasible actions.
Account for bounded rationality fully: Real humans satisfice, use heuristics, and can't compute complex equilibria.

Composition¶

Tools that work well together: - Payoff Matrix Construction + Best Response Mapping → systematic way to find Nash equilibria - Backward Induction + Subgame Perfect Equilibrium → solving sequential games credibly - Repeated Game Recognition + Evolutionary Stability → understanding when cooperation persists - Commitment Device Design + Sequential Move Recognition → creating first-mover advantages

Tools that are substitutes: - Nash Equilibrium vs. Dominant Strategy (dominant strategy is simpler when it exists) - Mixed Strategy vs. Pure Strategy equilibrium (different solution concepts for different game structures) - Backward Induction vs. Nash Equilibrium (sequential games use backward induction instead of simultaneous Nash)

Common tool sequences: 1. Map situation → Payoff Matrix Construction 2. Classify game type → Zero-Sum/Positive-Sum, Coordination/Anti-Coordination, Sequential/Simultaneous 3. Apply appropriate solution → Dominant Strategy, Nash Equilibrium, Backward Induction, or Mixed Strategy 4. If repeated → check for Repeated Game structure and Evolutionary Stability 5. If designing system → use Mechanism Design Inversion

Integration with Other Domains¶

Complements to these game theory tools: - Decision theory under uncertainty: Game theory assumes strategic opponents; decision theory handles non-strategic uncertainty (nature, randomness). Use both when facing mixed strategic and non-strategic uncertainty.

Behavioral economics: Game theory assumes rationality; behavioral economics documents systematic deviations. Combine to predict when game-theoretic reasoning applies and when behavioral patterns dominate.
System dynamics: Game theory is static or comparative static; system dynamics models how equilibria shift over time with changing parameters.
Probability and statistics: Many game theory applications require calculating expected values, updating beliefs (Bayesian updating in incomplete information games).

Conflicts to watch for: - Optimization vs. equilibrium: Individual optimization (choose best action) differs from equilibrium analysis (best response to others' best responses). Game theory requires equilibrium thinking, not just personal optimization.

Normative vs. positive: Game theory is descriptive (what equilibria exist) not normative (what should happen). Don't confuse Nash equilibrium with optimal or fair outcomes.
Transfer limitations: Game theory tools transfer across strategic situations, but specific equilibria don't transfer - you must recalculate for new payoff structures.