brycewang-stanford/Auto-Empirical-Research-Skills / skills/11-James-Traina-compound-science/skills/game-theory
game-theory
VerifiedThis skill covers game-theoretic methods in structural econometrics and industrial organization. Use when the user is working with strategic interactions, equilibrium analysis, or game-theoretic structural models — including entry games, conduct testing, auction models with strategic bidding, bargaining, or matching markets. Triggers on "Nash equilibrium", "subgame perfect", "best response", "strategic interaction", "entry game", "conduct testing", "auction", "mechanism design", "matching market", "bargaining", "BNE", "Bayesian Nash", "static game", "dynamic game", "repeated game", "multiple equilibria", "equilibrium selection", "discrete game", "oligopoly", "game-theoretic", "player", "payoff", "strategy", "dominant strategy", "Bresnahan-Reiss", "Ciliberto-Tamer", "partial identification", "set identification", or markup test.
| File | Kind | Size |
|---|---|---|
| SKILL.md | skill md | 14.5 KB |
| references/equilibrium-computation.md | markdown | 6.2 KB |
| references/equilibrium-concepts.md | markdown | 4.5 KB |
| references/estimation-diagnostics.md | markdown | 10.1 KB |
| references/identification-in-games.md | markdown | 3.2 KB |
| references/io-applications.md | markdown | 16.2 KB |
| references/multiple-equilibria.md | markdown | 3.6 KB |
references/multiple-equilibria.md
View on GitHub# The Multiple Equilibria Problem — Detailed Reference
The multiple equilibria problem is the central identification challenge in empirical games. When a game has more than one equilibrium, the econometrician needs an additional assumption to determine which equilibrium is played — or must relax point identification.
## Why Multiple Equilibria Arise
**Coordination games:** Multiple equilibria by design (Stag Hunt, Battle of the Sexes). Players coordinate on one of several Pareto-ranked equilibria.
**Entry games:** The game in which firms simultaneously decide whether to enter can have equilibria where firm A enters and B stays out, B enters and A stays out, or both enter — all consistent with Nash. The observed outcome depends on unmodeled coordination mechanisms.
**Symmetric games:** Any symmetric game has a symmetric Nash equilibrium (players randomize identically), but may also have asymmetric equilibria.
## Equilibrium Selection Approaches
| Selection Rule | Basis | Applicability |
|---------------|-------|--------------|
| Risk dominance (Harsanyi-Selten 1988) | Robustness to opponents' mixing | 2x2 games; computationally difficult for large games |
| Payoff dominance | Pareto ranking of equilibria | Only applies when one NE Pareto-dominates all others |
| Trembling-hand perfect (Selten 1975) | Robustness to small mistakes | Refines away weakly dominated strategies |
| Sequential rationality (Kreps-Wilson) | Consistency at off-path information sets | Extensive-form games |
| Quantal Response Equilibrium (McKelvey-Palfrey) | Bounded rationality, logistic choice | Generates unique equilibrium; testable |
| Ordered equilibrium (Berry 1992) | Exogenous ordering by profitability | Entry games with asymmetric firms |
**Quantal Response Equilibrium (QRE):** Firms respond probabilistically — more profitable strategies are played more often, but not with certainty. QRE is indexed by a precision parameter λ → ∞ (QRE converges to Nash), making it useful for equilibrium selection via model fit.
```python
def quantal_response_equilibrium(payoff_matrix, lambda_param=2.0, tol=1e-10, max_iter=5000):
n = payoff_matrix.shape[0]
p = np.ones(n) / n
for _ in range(max_iter):
eu = payoff_matrix @ p
log_p_new = lambda_param * eu - (lambda_param * eu).max()
p_new = np.exp(log_p_new) / np.exp(log_p_new).sum()
if np.max(np.abs(p_new - p)) < tol:
return p_new
p = p_new
raise RuntimeError("QRE iteration did not converge")
```
## Set Identification (Ciliberto-Tamer Bounds)
When no selection rule is imposed, the model is set-identified. The sharp identified set contains all parameter values θ such that the observed data is consistent with *some* equilibrium selection mechanism under θ.
**Practical approach:**
1. For each θ on a grid, compute all Nash equilibria of the game
2. Check whether the observed outcome distribution can be rationalized as a mixture of Nash equilibria
3. The identified set = {θ : observed distribution ∈ convex hull of Nash outcome distributions}
**Inference:** Use Chernozhukov, Hong, Tamer (2007) for confidence regions, or Romano-Shaikh (2010) subsampling. These are more demanding computationally than point-identified models.
## Using Multiplicity as Identifying Variation
A clever alternative: exploit the fact that different markets may play different equilibria, and use observable correlates of equilibrium selection as instruments (Sweeting 2009, Ellickson-Misra 2011). This requires a model of equilibrium selection, but allows point identification of structural parameters.
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| finishing-a-development-branchobra/superpowersUse when implementation is complete, all tests pass, and you need to decide how to integrate the work | Stars: 323k | Updated | Claude CodeCodex | Verified | Install |
| systematic-debuggingobra/superpowersUse when encountering any bug, test failure, or unexpected behavior, before proposing fixes | Stars: 323k | Updated | Claude CodeCodex | Verified | Install |
| test-driven-developmentobra/superpowersUse when implementing any feature or bugfix, before writing implementation code | Stars: 323k | Updated | Claude CodeCodex | Verified | Install |
| skill-creatoranthropics/skillsCreate new skills, modify and improve existing skills, and measure skill performance. Use when users want to create a skill from scratch, edit, or optimize an existing skill, run evals to test a skill, benchmark skill performance with variance analysis, or optimize a skill's description for better triggering accuracy. | Stars: 198k | Updated | Claude CodeCodex | Verified | Install |