On sequences with exponentially distributed gaps (Q7027474)

The authors compare the following assertions for a sequence \({\mathbf x} = (x_n)_{n \in {\mathbb N}}\) with values in the interval \([0,1]\):\N\begin{itemize}\N\item[(1)] \({\mathbf x}\) has Poissonian correlations of all orders;\N\item[(2)] \({\mathbf x}\) has exponentially distributed nearest-neighbor gaps.\N\end{itemize}\NIt is known that (1) implies (2). The authors show that the converse does not hold, by proving the following theorem:\N\NThere exist two sequences \((x_n)_{n \in {\mathbb N}}\) and \((z_n)_{n \in {\mathbb N}}\) with values in \([0,1]\) such that:\N\begin{itemize}\N\item[(i)] \((x_n)_{n \in {\mathbb N}}\) and \((z_n)_{n \in {\mathbb N}}\) have the same gap structure and their gap distribution is exponential.\N\item[(ii)] \((z_n)_{n \in {\mathbb N}}\) is uniformly distributed.\N\item[(iii)] \((x_n)_{n \in {\mathbb N}}\) is not uniformly distributed.\N\end{itemize}\NIn particular, there exists a sequence \((x_n)_{n \in {\mathbb N}}\) with exponentially distributed gaps that is not uniformly distributed, hence that does not have Poissonian correlations of any order \(k \geq 2\).\N\NThe construction is ingenious, a bit too technical to be explained in a few lines: we will only cite a remark of a referee who indicated that the methods reminded them of \textit{substitution schemes}, as in a paper of \textit{Y. Smilansky} and \textit{Y. Solomon} [Proc. Lond. Math. Soc. (3) 123, No. 6, 517--564 (2021; Zbl 1484.52018)]. The readers will appreciate the rich bibliography that puts all these questions in context.