Portfolio Optimization.
Turning noisy forecasts into allocations that survive out-of-sample.
4 articles · each one with a runnable notebook
Mean-Variance Optimization & the Efficient Frontier: A Six-Asset Portfolio in Python
Don't pick assets — pick the combination. We build the efficient frontier across six asset classes (2015–2024), see why every single ETF plots inside the cloud, and solve for the two portfolios everyone quotes: minimum variance and maximum Sharpe.
The Black-Litterman Model: Blending Market Equilibrium with Your Own Views
Reverse-optimize the market's implied returns, state one view — Germany at 10%, half confidence — and watch the Bayesian blend tilt the whole book sensibly while naive MVO lurches 75% into one country. Includes the no-views-equals-market sanity check.
Risk Parity from Scratch: Allocating by Risk, Not Capital — A Futures Portfolio
The 60/40 secret: capital weight ≠ risk weight. We compute marginal risk contributions across five futures, solve the risk-parity weights from scratch, match Riskfolio-lib to 1e-6, and scale the book to a 10% vol target with 1.45x leverage — the All-Weather mechanism in miniature.
60/40 in 2022: When the Stock-Bond Correlation Flipped
For two decades the 60/40 portfolio rested on one quiet assumption: when stocks fall, bonds rally. In 2022 that assumption broke — stocks fell ~18%, long Treasuries fell ~31%. We reconstruct what happened from SPY, AGG and TLT, watch the rolling correlation cross zero, and measure the damage against the GFC.