Power Law Local Hurst Exponent by G. Santostasi

This is a special version of the Hurst Exponent. It is based on observing local power-law properties for Bitcoin. Using a moving window the spectrum of the price is calculated using wavelets and then the spectrum is fitted with a log-log graph (given it shows power law properties). The regression parameters are then used to calculate the Hurst coefficient. This method is specific for Bitcoin and it shows its orderly and fractal properties at the same time.
The coefficient allows us to determine the market behavior of Bitcoin (trend above 0.5, mean returning below 0.5 or random around 0.5).