Comparison theorem for some extremal eigenvalue statistics (Q2212612)

The paper under review provides some comparison results for the maximal gap between consecutive eigenvalues in the bulk (and certain other extremal statistics of a similar flavour) for generalized Wigner matrices. For example, if two generalized Wigner matrices have the property that the first four moments of their entries coincide, the mean values of a smooth test function evalued on the maximal gap are asymptotically equal, in the limit as the size of the matrices tends to infinity. Thanks to this type of comparison results the universality results follow and allow to link the probability distribution for the maximal gap between eigenvalues of generalized Wigner matrices to GOE and GUE random matrices of the same symmetry class for which such results are known, see [\textit{G. Ben Arous} and \textit{P. Bourgade}, Ann. Probab. 41, No. 4, 2648--2681 (2013; Zbl 1282.60008)] and [\textit{R. Feng} and \textit{D. Wei}, ``Large gaps of CUE and GUE'', Preprint, \url{arXiv:1807.02149}]. The class of the extremal statistics for which this type of results remains valid is quite large and allows, for example, scaling the gaps by the semicircle density of the eigenvalues.