Editor-in-Chief Courtney Smith: This is a rather interesting question.

The short answer is that no statistical test can prove causality, only intuition and logic can do that. However, statistics can provide a good indication of a relationship that can be used for prediction or trading. Remember, trading is the art of making decisions with imperfect knowledge. What I do is to either start with a hypothetical causal relationship and then do a statistical test, such as linear regression or correlation analysis, and see if it proves out. Alternately, I test various indicators and then see if I can see an underlying reason for them to be related or that one is caused by the other.

Hope this helps!