Lecture 7/10: HRP – Fixing Markowitz's Curse (The OOS Portfolio Fix)

🚀 Lecture 7/10 of our Advances in Financial Machine Learning series by Marcos López de Prado! Mean-Variance and Risk Parity portfolios look perfect in-sample… but they crash out-of-sample (even worse than the dumb 1/N portfolio). Why? Markowitz’s Curse: inverting ill-conditioned covariance matrices destroys diversification exactly when you need it most. What you’ll learn: • Why quadratic optimizers (CLA) and Risk Parity fail OOS • Markowitz’s Curse explained (condition number + matrix inversion) • From geometry to topology: Hierarchical Risk Parity (HRP) in 3 stages • Tree Clustering, Quasi-Diagonalization & Recursive Bisection (with full Python code) • OOS Monte Carlo proof: HRP beats CLA & IVP by a huge margin • 2026 practical lesson: Build truly robust portfolios without forecasting returns Direct follow-up to Lecture 6 (backtest stats + DSR) and Lecture 5 (bet sizing + CPCV). This is the edge that separates surviving quants from the rest. 📕 Recommended Book (affiliate link): Advances in Financial Machine Learning – Marcos López de Prado 👉https://amzn.to/4dPPMyH (supports the channel!) 👉https://amzn.to/3PSScCJ (Kindle Format) Subscribe + hit the bell for the full 10-episode Cornell ORIE 5256 series! #quantfinance #HierarchicalRiskParity #HRP #MarkowitzCurse #portfoliooptimization #financialml #lopezdeprado #quantitativetrading #AFML #RiskParity