Haar null sets without \(G_{\delta}\) hulls (Q891152)

Let \(G\) be an abelian Polish group that is not locally compact. The main results of the paper are as follows: (1) there exists a Haar null (=shy in another terminology) subset of \(G\) that is not contained in any \(G_\delta\) Haar null subset of \(G\) (Theorem 1.3); (2) there exists a generalised Haar null subset of \(G\) that is not Haar null (Theorem 1.8). They answer some questions due to J. Mycielski and D. H. Fremlin, respectively. Under some set-theoretical assumptions, (2) was earlier obtained by \textit{R. Dougherty} [Fundam. Math. 144, No. 1, 73--88 (1994; Zbl 0842.43006)] and \textit{T. Banakh} [Commentat. Math. Univ. Carol. 45, No. 1, 119--137 (2004; Zbl 1098.03057)]. The main ingredients of the proofs of (1) and (2) are some classical results from descriptive set theory (universal sets, uniformisation) and some techniques developped by \textit{S. Solecki} [Fundam. Math. 149, No. 3, 205--210 (1996; Zbl 0887.28006); Fundam. Math. 170, No. 1--2, 197--217 (2001; Zbl 0994.28006)]. Recall that the difference between the notions of a Haar null set and that of a generalised Haar null set consists in using Borel sets in the former case and universally measurable sets in the latter case in their definitions. The generalised Haar null sets were first defined (under a different name) by \textit{J. P. R. Christensen} [Isr. J. Math. 13, (1973; Zbl 0249.43002)]. Reviewer's remark: There is some analogy between (2) and a result of \textit{E. Grzegorek} and \textit{C. Ryll-Nardzewski} [Bull. Acad. Pol. Sci., Sér. Sci. Math. 28, 229--232 (1980; Zbl 0496.28002), Proposition a].