Financial Backtesting Skill

Best practices for backtesting ML-based trading strategies, based on methodologies from "Advances in Financial Machine Learning" (Marcos Lopez de Prado, 2018) and modern quantitative finance research.

Core Principle

A backtest is NOT a research tool. Feature importance is. — Marcos Lopez de Prado

Backtests are easily overfitted. Treat them as a final sanity check, not as an optimization target.

1. Triple-Barrier Labeling

The standard method for creating labels in financial ML. Instead of fixed-horizon returns, use three barriers: