Lemke Oliver and Soundararajan bias for consecutive sums of two squares (Q2089709): Difference between revisions

Let \(E = \{a^2+b^2\mid a,b\in \mathbb Z\} = \{E_n \mid n\in \mathbb N\}\) be the set of integers that are the sum of two squares written in an increasing order. The authors study consecutive elements of \(E\) in arithmetic progressions, and develop a heuristic model based on the Hardy-Littlewood conjecture for sums of squares, which explains the biases in the experimental data. They also prove several results related to averages of the Hardy-Littlewood constant in the context of sums of two squares. For a prime number \(q\equiv 1\pmod 4\) and integers \(a, b\), let \[ N(x;q,(a,b)):= \# \left\{E_n\leq x\mid E_n\equiv a\pmod q\;\text{and}\;E_{n+1}\equiv b\pmod q\right\}. \] The authors propose a heuristic model predicting an asymptotic formula for \(N(x;q,(a,b))\), based on the heuristic of \textit{R. J. Lemke Oliver} and \textit{K. Soundararajan} [``Unexpected biases in the distribution of consecutive primes'', Proc. Natl. Acad. Sci. USA 113, No. 31, E4446--E4454 (2016; \url{doi:10.1073/pnas.1605366113})] for the case of primes. They provide many other results too, so the interested readers can find a lot of useful ideas and information in this article.