Measures of affinity of a sequence for a number (Q4802190)

Let \(S\subset N= \{1,2,\dots\}\). Put \[ \underline m(S)= \liminf_{n\to\infty} {|S\cap N_n|\over n},\quad\overline m(S)= \limsup_{n\to\infty} {|S\cap N_n|\over n}, \] where \(N_n= \{1,2,\dots, n\}\), and \(|A|\) stands for the cardinality of the set \(A\). Let \(a\in R\), and \((x_n)\) be a sequence of real numbers. Define the strong and the weak affinity of the number \(a\) for \((x_k)\) by \[ \begin{aligned} L(a,(x_n)) &= \inf_{\varepsilon> 0} \{\underline m(\{k:|x_k- a|< \varepsilon\})\},\\ U(a,(x_n)) &= \inf_{\varepsilon> 0} \{\overline m(\{k:|x_k- a|<\varepsilon\})\}.\end{aligned} \] The authors prove that a number \(a\) is a statistical limit point of \((x_k)\) if and only if \(U(a,(x_k))> 0\). (Recall that \(a\) is a statistical limit point of \((x_k)\) if \[ a= \lim_{n\to\infty} x_{k_n}, \] where \(K= \{k_1< k_2<\cdots\}\subset N\) and \(\overline m(K)> 0\). Similarly, if \(L(a,(x_k))> 0\) then \(a\) is a statistical limit point of \((x_k)\), but the converse is not true.) The set of all \((x_k)\in\ell_\infty\) for which \(\sup_a L(a,(x_k))> 0\) is a nowhere dense subset of \(\ell_\infty\). The set of all \((x_k)\) such that \(\sup_a U(a,(x_k))> 0\), is a dense subset of \(\ell_\infty\) of the first category in \(\ell_\infty\).