An application of Shoenfield's absoluteness theorem to the theory of uniform distribution (Q1314862)

The author gives two different proofs, one in the language of set theory, the other in the language of probability theory for the following result: If \({\mathcal C}\) is a Polish probability measure space, \(W\subseteq \mathbb{N}^ \mathbb{N} \times{\mathcal C}\) a Borel set whose sections \(W_ x\), \(x\in \mathbb{N}^ \mathbb{N}\) have measure one and are decreasing \((x\leq x'\to w_ x\geq x_{x'})\) then \(\bigcap_ x W_ x\) has measure one. As a consequence one obtains an interesting result on uniform distribution: Let \(s\in \mathbb{N}^ \mathbb{N}\) be a sequence of positive integers, \(t\in \{0,1\}^ \mathbb{N}\) a sequence in \(\{0,1\}\) then \(t\) is called \(s\)-uniformly distributed if \(\lim_{n\to\infty} D_{s(n)} (t,n)=0\), where \[ D_ k(t,n):= 2^ k \max_{w\in \{0,1\}^ k} \Biggl| {{|\{i<n: (t(i),\dots, t(i+k-1))= w\}|} \over n} - {1\over {2^ k}} \Biggr|. \] It is shown that almost all sequences are \(s\)-uniformly distributed iff \(\log n- \log\log n- s(n)\) tends to infinity.