pair-trade-screener

Cointegration testing for pair trading uses the ADF test to confirm a long-run equilibrium between two stocks, ensuring the spread is stationary and mean-reverting. This statistical validation prevents trading spurious correlations by verifying the hedge ratio stability over the lookback period.

Pair trading screening requires daily adjusted price history data sourced via the FMP API, along with Python libraries including pandas, numpy, scipy, and statsmodels for statistical computations. This environment supports the cointegration testing, hedge ratio estimation, and z-score calculations needed to generate signals.

The half-life of mean reversion estimates the time required for a cointegrated spread to revert to its mean, indicating how long a pair trade takes to become profitable. It filters out pairs with slow reversion speeds, ensuring generated entry and exit signals have practical holding period expectations.

You generate market-neutral position sizing and stop rules from cointegration analysis by applying the computed hedge ratio to balance the long-short exposure of the validated pair. The rolling z-score thresholds define specific entry, exit, and stop levels for backtesting and execution.