Using the Hurst Exponent to Switch Between Momentum and Mean Reversion

Summary

This study uses the Hurst exponent to classify a time series as trending or mean reverting, then selects a strategy intended to suit that behavior. Momentum is matched to trending conditions, while mean reversion is used when prices exhibit mean-reverting behavior. The central idea is to adapt strategy choice to an estimate of the market regime rather than apply either approach continuously.

The document reports that this Hurst-based selection can produce higher returns, accompanied by higher risk. It does not specify assets, sample period, thresholds, trading rules, transaction costs, or quantitative performance figures, so the result cannot be independently judged from the summary. It also identifies limitations in the study and proposes Q-learning as a possible way to improve the strategy and implementation of its component algorithms. That proposal is prospective; no Q-learning results are reported here. The approach therefore illustrates a regime-selection concept, while leaving its robustness and practical performance unresolved.

Key ideas

• The Hurst exponent is used to distinguish trending from mean-reverting behavior.
• The method selects momentum for trends and mean reversion for mean-reverting conditions.
• The study reports higher returns alongside higher risk for Hurst-based strategy selection.
• The document proposes Q-learning as a possible improvement but gives no results for it.
• Details needed to assess costs, thresholds, and out-of-sample robustness are not provided.

Tags

# Optimizing Returns Using the Hurst Exponent and Q Learning on Momentum and Mean Reversion Strategies


# Optimizing Returns Using the Hurst Exponent and Q Learning on Momentum and Mean Reversion Strategies









Momentum and mean reversion trading strategies have opposite characteristics. The former is generally better with trending assets, and the latter is generally better with mean reverting assets. Using the Hurst exponent, which classifies time series as trending or mean reverting, we attempt to trade with each strategy when it is advantageous to generate higher returns on average. We ultimately find that trading with the Hurst exponent can achieve higher returns, but it also comes at a higher risk. Finally, we consider limitations of our study and propose a method using Q-learning to improve our strategy and implementation of individual algorithms.

Shown in full with attribution under the source's licence. Licence: abstract CC0

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