I know that the hedge fund Renaissance Technologies use computer-based models to predict price changes in financial instruments. These models are bases on analyzing as much data as can be gathered, then looking for non-random movements to make predictions. Could anyone have time to tell me how is it possible to find 'non-random movements'? I know it is essentially statistics and probably a bit of probability. What theory do I have to look for? Any examples maybe?
I doubt that the randomness of the stock markets can be overcome entirely by any kind of modeling, but with lots of data, clever assumptions, and time-series modeling, I think it might be possible to reduce the effect of random noise--enough to tempt some over-zealous PR guy to claim detection of 'non-random movements'. Ad copy is cheap, big-data time-series analysis is a little more expensive and less certain.
P.S. I'd like if someone could build a full (complete) answer on the subject in such a way I could understand better.
Thanks!