Humans vs. algorithms - Who follows Newcomb-Benford's Law better with their order volume?

Abstract

Newcomb-Benford's Law (NBL) is a well known regularity in the distribution of first significant digits (FSD) and therefore research in this field is manifold. As of 2012 research in the domain of financial markets is quite scarce, especially in the field of algorithmic trading. We pose the question whether order submission volumes of algorithmic traders and human traders follow NBL. Results in this context might help regulators to detect suspicious market activity and market participants to quantify the amount of algorithmic trading. Our findings indicate that the submitted order volumes of both groups follow NBL more than the uniform distribution. Comparing these two groups, we give a proof that algorithmic traders match NBL better than human traders, as human traders tend to overuse the FSD five. © 2013 Springer-Verlag.