What problem does it solve? Standard risk models like Value at Risk and portfolio optimization fail catastrophically in domains with fat-tailed distributions and complex payoffs, yet practitioners have no systematic way to know when to trust their models. This Skill provides a diagnostic framework for identifying when statistical models become dangerous rather than merely wrong. ## Core Features & Use Cases - Fourth Quadrant Diagnosis: Classifies any risk domain along two dimensions — payoff type (simple vs. complex) and distribution type (thin-tailed Mediocristan vs. fat-tailed Extremistan) — to determine whether models are trustworthy. - Model Failure Explanation: Explains why VaR, stress tests, and portfolio theory structurally fail in finance, epidemiology, geopolitics, climate, and technology risk. - Exposure Redesign Guidance: Directs the response away from better modeling toward changing exposure — exiting the Fourth Quadrant or becoming convex rather than concave to uncertainty. - Use Case: A risk analyst evaluating whether a credit risk model can be trusted for a leveraged derivatives portfolio uses this framework to determine the portfolio sits in the Fourth Quadrant, concluding the model's error can be off by orders of magnitude and that exposure must be restructured instead. ## Quick Start Ask the AI to apply the Fourth Quadrant framework to evaluate whether a specific risk model or domain is safe to rely on.
