Long-Only Statistical Arbitrage Using a Modified Z-Score Spread

Summary

This script outlines a long-side statistical arbitrage approach for a pair of instruments. It retrieves the second instrument’s price series, standardizes each instrument’s price against its own rolling mean and standard deviation, and subtracts the resulting z-scores to form a spread. It then centers that spread using a rolling median and scales deviations by the median absolute deviation to calculate a modified z-score. A configurable threshold is intended to identify an unusually low spread as a potential long entry.

The available document ends during the plotting section, before showing the order conditions, exits, or any performance results. It therefore does not establish exactly how a trade is executed or closed, and the title’s long-only framing cannot be evaluated against full strategy logic. The method also does not show tests for whether the pair relationship is stable or mean-reverting. A trader would need to assess pair selection, synchronization, transaction costs, and out-of-sample behavior before drawing conclusions from this signal construction.

Key ideas

• The method compares rolling standardized prices for a main instrument and a paired instrument.
• It defines the spread as the difference between the instruments’ individual z-scores.
• A rolling median and median absolute deviation are used to calculate a robust modified z-score.
• A configurable negative threshold is intended to flag a potential long entry.
• The available text omits trade exits, complete execution logic, and performance evidence.

Tags

This summary was written by Stratmill's research agent from the original; it is not a copy of the source.