Trading strategies and backtesting

I think that good quant investment managers … can really be thought of as financial economists who have codified their beliefs into a repeatable process. They are distinguished by diversification, sticking to their process with discipline, and the ability to engineer portfolio characteristics. — Cliff Asness of AQR (2007)

Broadly speaking, quantitative investing tries to turn an investment thesis into a repeatable process. The thesis is often model driven, perhaps related to some economic rationale that repeats itself. For example, maybe stocks with certain characteristics generate positive risk-adjusted returns, on average.

This process means using historical or simulated data to test a hypothesis. This is called backtesting – there’s an entire literature on the best way to do this. We’ll look at two complementary approaches: building and testing trading strategies with the bt package, and evaluating alpha factors with alphalens.

Note

Backtesting is seductive. It’s easy to find strategies that “work” in historical data but fail going forward. A healthy skepticism of backtest results is one of the most important things you can develop.

What you’ll learn

By the end of this chapter, you should be able to:

• Build and backtest trading strategies in Python using the bt package

• Construct signal-based strategies like SMA crossovers and momentum

• Evaluate strategy performance using Sharpe ratio, drawdowns, and other metrics

• Understand the pitfalls of backtesting: overfitting, look-ahead bias, survivorship bias

• Use alphalens to analyze whether a factor (like momentum) actually predicts returns

• Interpret factor tear sheets: returns by quantile, information coefficient, and turnover

Chapter structure

Section	Topic	Key Ideas
Backtesting with BT	Building and testing trading strategies	Equal-weight, SMA crossover, momentum, transaction costs, in-sample vs. out-of-sample
Factor Analysis with Alphalens	Evaluating alpha factors	Momentum factor, long-short portfolios, quantile returns, information coefficient, tear sheets

Key references

• Efficiently Inefficient by Lasse Pedersen – Chapter 3.3 introduces quantitative equity strategies

• bt package documentation – the backtesting framework we use

• alphalens-reloaded documentation – factor analysis toolkit from the Quantopian ecosystem

• Jegadeesh and Titman (1993) – the original momentum paper

Using AI for trading strategies

AI tools are particularly helpful here:

• Writing bt strategies: “Help me build a bt strategy that buys stocks above their 200-day moving average”

• Interpreting results: Paste backtest output into Claude and ask “Is this strategy any good?”

• Debugging: “Why is my backtest throwing an error about NaN prices?”

• Factor construction: “How do I calculate a 12-1 momentum factor for a universe of stocks?”

Tip

When asking AI about trading strategies, be specific about your universe (what stocks?), your signal (what triggers a trade?), and your rebalancing frequency (how often?). Vague prompts lead to generic answers.