The two-dimensional small ball inequality and binary nets (Q2409198)

It is mathematical folklore that there is a connection between irregularities of point distributions (measured in terms of the discrepancy), and the small ball inequality from probability theory (see for example [the first author and \textit{M. Lacey}, in: Monte Carlo and quasi-Monte Carlo methods 2012. Proceedings of the 10th international conference on `Monte Carlo and quasi-Monte Carlo methods in scientific computing', Sydney, Australia, February 13--17, 2012. Berlin: Springer. 23--38 (2013; Zbl 1304.11074)]). In the present paper, the authors use methods from Fourier analysis to establish a formal connection between these two topics in dimension 2. Their approach also gives a new proof of the small ball inequality in dimension 2. Interestingly, it turns our that the ``extremal'' sets with respect to this inequality are exactly the so-called binary nets, which form a well-known class of low-discrepancy point sets. The proof of the main result is based on an interpretation of the small ball inequality in terms of some sort of lacunary Fourier series, in the spirit of a classical theorem of \textit{S. Sidon} [Math. Ann. 97, 675--676 (1927; JFM 53.0252.02)]. The authors also discuss a possible extension of their approach to the higher-dimensional case.