Trading Rules Based on Skew and Kurtosis
Summary
This document outlines three futures trading rules built from skew and kurtosis: a standalone skew signal, skew conditioned on kurtosis, and kurtosis conditioned on skew. Signals are normalized by a robust volatility estimate and smoothed; conditioned signals use the sign of the other factor. The rules can be measured against an instrument’s own history, the cross-asset average, or the average within an asset class, using lookbacks from one week to one year.
The evidence shown is mainly rule-return correlations and a proposed portfolio allocation. Skew rules at nearby horizons can have very different correlations, while the two conditioned rules are strongly correlated with each other in the displayed examples. The portfolio tree groups skew, kurtosis, carry, and moving-average momentum rules. However, the excerpt does not provide full performance results, a complete fitting discussion, or enough detail to establish profitability. Correlations and an allocation tree alone do not demonstrate that these signals will generalize or survive trading costs.
Key ideas
- Skew and kurtosis can be used as standalone or mutually conditioned trading signals.
- The signal is normalized by robust volatility and then smoothed over time.
- Signals can be compared with an instrument’s own history or with cross-sectional averages.
- Correlations vary across lookback horizons and signal definitions.
- The allocation example combines these rules with carry and momentum signals, but does not establish net profitability.
Tags
This summary was written by Stratmill's research agent from the original; it is not a copy of the source.