Distance-Based Pair Selection and Spread Threshold Signals
Summary
This article describes a basic distance approach to pairs trading. During a formation period, asset price series are normalized so their scales are comparable, then candidate pairs are selected using squared Euclidean distance. The spread’s historical volatility is also measured. During trading, the same normalization is applied, a spread is formed from the paired series, and signals are triggered when it moves beyond a threshold, often set around two standard deviations. The strategy buys the lower-priced leg and shorts the higher-priced leg when the spread is sufficiently wide, anticipating reversion.
The article compares four selection criteria: smallest distance, membership in the same industry, frequent zero crossings, and higher historical standard deviation. It cites research reporting a significant relationship between zero crossings and pair returns, and argues that distance alone can favor spreads with little variance and limited profit potential. The approach is simple, but it fixes pairs before trading and considers only two assets at a time. It also omits transaction costs and does not establish that historical co-movement or convergence will persist.
Key ideas
- Normalize prices before comparing candidate pairs with squared Euclidean distance.
- Pair selection can incorporate industry membership, spread zero crossings, or historical variability.
- Trading signals arise when the paired spread crosses a volatility-based threshold.
- Frequent zero crossings are presented as evidence of repeated divergence and convergence opportunities.
- Fixed pair selection and a two-asset focus can miss opportunities and expose the strategy to changing relationships.
Tags
This summary was written by Stratmill's research agent from the original; it is not a copy of the source.