Friday, November 21, 2014

Stochastically Jumping without a clue

Recent research from NYU (1) that apparently demonstrates that the modeling techniques used to forecast stock market fluctuations could be used to predict animal behavior – in this case, movements of zebrafish – ignores many fundamental aspects of modeling. First, unknown to most people in the financial industry, stock market fluctuations cannot be predicted.In spite of the proclamations of a recent Nobel laureate, who claims he could smell a bubble anytime anywhere, he is still to demonstrate a usable prediction ex.ante. And, high flying hedge fund managers, without insider knowledge, could never create alpha – risk adjusted excess returns, systematically. And, second, equating animal behavior modeling to stock market predictions shows a level of incomprehension in both areas.

Stochastic jumps do occur – the problem with such things is that they are not predictable. Perhaps what the NYU team is missing is the right language – the characteristics of the underlying process of the movement of the zebrafish appear to have stochastic jumps in it. But that has nothing to do with stock market modeling – an oxymoron. The reason the zebrafish is jumping stochastically is the same why stock markets do at times – arrival of new information. And, by definition these are not predictable.

As trillion $ slosh around an industry with no value added to society, further research toward predictions of stochastic jumps seems unwarranted.
(1) http://esciencenews.com/articles/2014/11/13/stock.market.models.help.nyu.researchers.predict.animal.behavior