I was working on historical data looking for anomalous patterns that we would not expect to occur at random. I'd like to create a scheme to analyze data and markets to test for statistical significance and consistency over time.

Question : Are there some tools we could use to evaluate the nonrandomness of a database? What do we use usually to say that something is completely random or not (statistically)? Figure we work with change in some value over time.


Computer programmers use the NIST test Suite to evaluate the quality of pseudo-random number generators. I've been interested in using it to test market data for periods of stability (represented by high randomness) and periods of state transition (represented by low randomness). Multiple implementations exist.

A discussion of using the NIST suite to test randomness of financial data can be found in this article along with the author's details on his own implementation of the test suite.

| improve this answer | |

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.