Strategies

Build pairs trades, rank a universe, or combine several strategies into one portfolio. These worked examples bring VBT's data, indicators, signals, and analytics together on real market data, with code and plots you can adapt to your own ideas.

Choose a strategy workflow

features/strategies pages

Pairs trading

Screen cointegrated pairs, trade spread z-scores with shared cash, and validate them

Cross-sectional strategies

Rank a universe each period, rotate into the leaders, and test factors by quintile

Multi-strategy portfolios

Stack strategies as one portfolio, weight them, and compare recurring investing plans

Build on the examples

The examples use Python functions, pandas tables, and VBT portfolios, so you can replace their symbols, scores, signals, and allocation rules with your own. Use signal backtesting when you have entry and exit arrays, or event-driven backtesting when a decision depends on the positions and cash already in the portfolio.

The same research tools work across these strategy families. Compare lookbacks, thresholds, and position sizes with parameter optimization, then check the selected settings on later data with walk-forward testing. Inspect trades and drawdowns to understand what drives the result.

Compare each idea with a baseline, inspect its trades, and see the effect of costs and allocation rules. The examples give you a complete starting point for the next experiment.

Start here

Follow the Pairs trading tutorial to build a pairs strategy from screening to a custom simulator. For a first signals-based example, the Basic RSI tutorial goes from indicator values to a backtest. The Portfolio optimization tutorial develops allocation functions and connects them to portfolio simulation.