Estimating the Hurst Exponent with Rescaled Range Analysis for Moving-Average Signals
Summary
The article presents a custom MQL5 signal that estimates the Hurst exponent with rescaled range analysis and combines it with moving-average positioning. It uses the exponent as a regime guide: values above 0.5 are associated with persistence or trending, while values below 0.5 suggest mean reversion. A faster moving average is assigned to mean-reverting conditions and a slower one to trending conditions, with the intent of interpreting price position in light of the detected regime.
The estimation procedure partitions a sample into segments of varying sizes, adjusts each segment by its mean, accumulates deviations, measures rescaled ranges, and uses a log-log relationship to estimate the exponent. The article includes MQL5 implementation details and discusses sample-size and zero-variance checks. Its trading tests are described as limited, with the conclusion noting compute intensity, excessive trade filtering, and concerning drawdown relative to a raw autocorrelation signal. The author recommends broader testing with real-tick data over more years; the presented evidence does not establish robust profitability.
Key ideas
- Rescaled range analysis estimates the Hurst exponent from how range scales with sample size.
- The article interprets values above 0.5 as trending and values below 0.5 as mean-reverting conditions.
- It pairs regime classification with separate fast and slow moving averages for signal interpretation.
- Segment mean adjustment, cumulative deviations, and log-log regression form the estimation procedure.
- The reported tests are limited and raise concerns about computation, trade filtering, and drawdown.
Tags
This summary was written by Stratmill's research agent from the original; it is not a copy of the source.