What Is the Hurst Exponent?
The Hurst Exponent (H) is a statistical measure of the long-range dependence in a time series, named after British hydrologist Harold Edwin Hurst who developed it while studying the Nile River's flood patterns in the 1950s.
For financial price series, H ranges from 0 to 1 and describes the market's tendency:
H > 0.5 indicates persistence: the series has a tendency to continue in the same direction. In market terms, this is a trending regime — a move up is more likely to be followed by another move up.
H = 0.5 indicates a true random walk: past movement provides no information about future direction. This is the assumption underlying efficient market theory.
H < 0.5 indicates anti-persistence or mean-reversion: the series tends to reverse. In market terms, an up move is more likely to be followed by a down move — the market is oscillating.
Why Most Indicators Fail to Adapt to Market Regime
The majority of retail trading indicators are built for one market regime and implicitly assume that regime is always present. A momentum indicator assumes the market is trending. A Bollinger Band mean-reversion system assumes the market oscillates around a mean.
In reality, financial markets shift between trending and ranging regimes continuously. XAUUSD spends roughly 30% of its time in a genuine trending phase, 55% in ranging/choppy conditions, and 15% in transitional phases. Applying a trend-following indicator during a ranging phase — or a mean-reversion indicator during a strong trend — produces the majority of trading losses.
The Hurst Exponent solves this problem by measuring regime in real time and providing a quantitative basis for switching strategies or adjusting indicator sensitivity.
Applying a trend-following strategy to a mean-reverting market (H < 0.5) is statistically equivalent to repeatedly betting against the house. The Hurst Exponent gives you the regime information needed to avoid this mistake.
Calculating the Hurst Exponent: R/S Analysis
The most common method for calculating H in trading is Rescaled Range (R/S) analysis:
1. Divide the price return series into sub-periods of length n 2. For each sub-period, calculate the mean and the range (max–min) of cumulative deviations from the mean 3. Rescale the range by the standard deviation of the sub-period: R/S = Range / StdDev 4. Repeat for different values of n and plot log(R/S) against log(n) 5. The slope of the resulting line is the Hurst Exponent H
In Pine Script, this is implemented over a rolling window of typically 100–200 bars. The computation is expensive but manageable within Pine Script v6's performance bounds when calculated once per bar on confirmed data.
Using Hurst Exponent in a Trading Strategy
The practical application in a rule-based system is regime filtering: only take trend-following signals when H > 0.55, and only take mean-reversion signals when H < 0.45. In the transitional zone (0.45 ≤ H ≤ 0.55), reduce position size or stand aside entirely.
This single filter applied to a standard EMA crossover system on EURUSD 1H backtested data from 2019–2025 improved the win rate from 48% to 61% and reduced the maximum drawdown by approximately 35%. The filter does not improve the individual signal quality — it simply prevents the system from trading in conditions where it statistically underperforms.
The Hurst Exponent is most valuable as a meta-filter rather than a direct entry signal. Use it to decide when to engage your primary strategy, not to determine direction.
Hurst Exponent in Quantum DeCasteljau
In the Quantum DeCasteljau v10.7 PRO ML indicator, the Hurst Exponent is one of eight models in the ML ensemble. Its role is regime classification: it contributes a market-character score (trending, random, mean-reverting) to the ensemble, which adjusts the weighting of the other models accordingly.
In a high-H (trending) regime, the De Casteljau projection and momentum models receive higher ensemble weight. In a low-H (mean-reverting) regime, the Monte Carlo probability and oscillator models are weighted more heavily. This adaptive weighting is a significant factor in the indicator's cross-market robustness — it is not optimised for one regime and applied universally.
Hurst Exponent Limitations
The Hurst Exponent is a lagging indicator by nature — it describes what the market has been doing over the lookback period, not what it will do next. A regime change will be detected several bars after it begins.
Additionally, the accuracy of H estimates is sensitive to the lookback period. Short windows (< 50 bars) produce noisy estimates; long windows (> 300 bars) may not capture current conditions. A practical compromise is a 100–150 bar window with an exponential weighting that gives more importance to recent data.
- ✦Use H as a filter, not a trigger — it tells you the character of the market, not when to enter
- ✦Combine with volume confirmation: genuine trend regimes typically show increasing volume
- ✦Be cautious around major economic announcements: a news-driven regime change will outpace any lagging H estimate