Research
Hypothesis testing, backtesting pitfalls, overfitting, and how to build (and lose) edge.
Key Concepts
- Hypothesis-driven testing — Start with a thesis, not a pattern match.
- Overfitting — When your model memorizes noise instead of signal.
- Out-of-sample testing — Validating on data your model hasn't seen.
- Edge decay — Why strategies stop working and what to do about it.