The Hurst exponent is a measure for the behaviour of the market. It shows if the market behaves in a random, trending or mean-reversion manner. This can be used to select the right trading strategy for your market.
Hurst Exponent
The hurst exponent describes the self similarity of a market. Self similarity describes how similar past market snippets are to current ones. A Hurst exponent of 0.5 means that the market, over the long term, follows a random walk. In this case, in the long run, any trading strategy would be a zero sum game (excluding commissions).
If the Hurst exponent is above 0.5, the market shows a trending behaviour. Past moves are similar to current moves. Markets with a high Hurst exponent are perfect for trend following strategies. If the market went up in the past, there will be a better than 50% chance that it also moves up in the future.
If the Hurst Exponent is below 0.5, the market shows a reverting behaviour. If it went down in history, there is a good chance that it will reverse its direction in the future. These markets are markets for mean reverting strategies and short term reversal pattern analysis.
The picture above shows the Hurst exponent for several time-frames of the SPX. On intraday data (the last 1000 bars have been used for the calculation) the Hurst exponent is below 0.5. So be careful with trend following strategies on intraday data.
On weekly and monthly data the market clearly shows a trending behaviour, If there is a rising week or month, there will be a good chance that also the next week or month will move up.
Hurst Exponent market overview
Running a scanner over different markets and time frames gives an interesting insight in different markets. On the picture above 1000 bars of daily data is used on the left side, and 1000 bars of hourly data is used on right side.
As it can be seen, Bitcoin surely is the best market to do trend following strategies. It has got a Hurst exponent above 0.5 on both time frames.
A completely different market behaviour is shown with EUR/USD. This market shows a reverting behaviour on daily and hourly charts.
See the chart below: A mean reverting strategy surely would have had some edge with EURUSD (left), while a trend following strategy would have been the right thing for Bitcoin(right)
Tradesignal Coding
To use the Hurst analysis with your own markets you can use the code shown below. It is for private use and research only!
The code uses the python Hurst package to do the calculation. This code is embedded within the Tradesignal Equilla coding, which does the graphical output and takes care of the data.
When you intend to use the code with your Tradesignal make sure to have python and the needed packages installed.
To customise the indicator it has got several settings.
Use the period setting to define how much data should be used for the calculation (number of bars). If set to 0 all data on the chart will be used.
The Plot R/S.. setting can be switched to true to show the R/S analysis. It describes how self similar the data is over different time intervals. Please see Wikipedia for further information
The recalc setting defines after how many bars the calculation is updated.
Get the code
Please acknowledge that you are a private investor and will use the code only for research by typing “hurst” as password on the code download page
Thanks to www.tradesignal.com for the great software.
Free Webinar on Hurst
See more on the Hurst exponent and R/S analysis in my free webinar in February. Register and see dates
Pingback: The coastline paradox and the fractal dimension of markets - Quantitative Analysis and Trading
Most elaborate explanation that I could never find elsewhere. Really helpful for my masters thesis. Thank you so very much.