Threshold, Correlation, and Volatility Filters for Spread Trading
Summary
The document describes three ways to refine spread trading signals. A threshold filter enters or maintains a long or short spread position only when the predicted spread change crosses a chosen boundary; an asymmetric version allows different boundaries for each direction, with coefficients that may come from a threshold autoregressive model. The correlation filter permits a trading rule’s signal when the change in correlation between the spread’s underlying legs is below a cutoff, aiming to avoid stagnant spread periods and early moves away from fair value.
The volatility filter classifies market regimes using RiskMetrics-style exponentially weighted variance and rolling historical averages and standard deviations. It proposes avoiding very high volatility and favoring lower-volatility periods, while noting that inverse-risk position scaling is a distinct approach to time-varying exposure. The page reports that filters often improve out-of-sample Sharpe ratios, but gives no detailed test results or universal parameter choices. Outcomes depend on fitted thresholds, volatility estimates, data, and implementation; the examples are documentation illustrations rather than independent evidence.
Key ideas
- A threshold filter stays out of a spread when its predicted change falls inside a chosen band.
- Asymmetric thresholds can reflect different long and short response coefficients.
- A correlation filter can suppress signals during periods when the underlying legs’ correlation is rising.
- Volatility regimes can be estimated from exponentially weighted variance and rolling historical statistics.
- Filter performance depends on parameter estimation and requires out-of-sample evaluation.
Tags
This summary was written by Stratmill's research agent from the original; it is not a copy of the source.