Back to Blog
Tutorial
tutorial
portfolio-construction
risk-parity
optimization

Portfolio Optimization: From Markowitz to Modern Risk Parity

DR
Dr. Sarah Chen
March 3, 2026
1 min read
4,321 views
Portfolio Optimization: From Markowitz to Modern Risk Parity

Portfolio Optimization: From Markowitz to Modern Risk Parity

You've got great signals. Now how do you size your positions? Portfolio construction is where most quants leave alpha on the table.

The Evolution

1952: Markowitz Mean-Variance

The original. Maximize expected return for a given level of risk. Beautiful in theory, terrible in practice — it's extremely sensitive to estimated expected returns.

1990s: Black-Litterman

Combines market equilibrium returns with investor views. More stable than raw Markowitz, but still requires return forecasts.

2000s: Risk Parity

Don't forecast returns at all. Instead, allocate risk equally across assets. Simple, robust, and surprisingly effective.

2020s: Hierarchical Risk Parity (HRP)

Uses machine learning (hierarchical clustering) to build more robust portfolios that avoid the pitfalls of traditional covariance inversion.

Implementing Risk Parity

import numpy as np

def risk_parity_weights(cov_matrix: np.ndarray) -> np.ndarray: """Equal risk contribution portfolio.""" n = cov_matrix.shape[0] # Start with inverse-volatility weights vols = np.sqrt(np.diag(cov_matrix)) weights = (1 / vols) / np.sum(1 / vols)

# Iterate to equalize risk contributions for _ in range(100): port_vol = np.sqrt(weights @ cov_matrix @ weights) marginal_risk = cov_matrix @ weights / port_vol risk_contrib = weights * marginal_risk target = port_vol / n weights = weights * (target / risk_contrib) weights /= weights.sum()

return weights

What Wins on AlphaNova

Our top performers typically use:

  • Volatility targeting — Scale positions so portfolio vol stays at ~10-15%
  • Position limits — No single stock exceeds 5% of the portfolio
  • Sector neutrality — Net exposure to each sector is near zero
The boring portfolio construction beats the clever one. Every time.

Leave Feedback

/blog/portfolio-optimization-markowitz-risk-parity