Hi Robert,

I am glad that the explanation of how to do rescaled range analysis was easy to understand for you. Yea!

Thank you for your question. I am no statistician, however, I do know that both hypothesis testing and confidence intervals rely on knowing the shape of the distribution of the data. After all, these techniques work because we know the mathematics that describe a distribution. We then make comparisons to these distributions, assuming they are the benchmark. So not knowing the shape of the distributions makes these techniques impossible.

The most common distribution is of course the standard normal distribution, which supposedly describes random processes. In rescaled range analysis only a Hurst exponent of 0.5 describes a random sample of data. Anything other than that is considered non-random. So the usual hypothesis testing and confidence intervals are not going to apply except in that special circumstance. Knowing the shape of the distributions of non-0.50 Hurst exponent samples is too tough for me to answer.

Does anyone in the audience have any information on distribution shapes for non-0.50 Hurst exponent data?

Yours, in service,

Jason