A universal result in almost sure central limit theory

• Published in: Stochastic processes and their applications Vol. 94; no. 1; pp. 105 - 134
• Main Authors: Berkes, István, Csáki, Endre
• Format: Journal Article
• Language: English
• Published: Amsterdam Elsevier B.V 01.07.2001; Elsevier Science; Elsevier
• Series: Stochastic Processes and their Applications
• Subjects: Almost sure central limit theorem; Almost sure central limit theorem Logarithmic averages Summation methods; Exact sciences and technology; Limit theorems; Logarithmic averages; Mathematics; Probability and statistics; Probability theory and stochastic processes; Sciences and techniques of general use; Summation methods; primary 60F15; Logarithmic averages; secondary 60F05; Almost sure central limit theorem; Summation methods; Almost sure convergence; Central limit theorem; Summation; Probability theory; Limit theorem
• ISSN: 0304-4149; 1879-209X
• DOI: 10.1016/S0304-4149(01)00078-3

The discovery of the almost sure central limit theorem (Brosamler, Math. Proc. Cambridge Philos. Soc. 104 (1988) 561–574; Schatte, Math. Nachr. 137 (1988) 249–256) revealed a new phenomenon in classical central limit theory and has led to an extensive literature in the past decade. In particular, a.s. central limit theorems and various related ‘logarithmic’ limit theorems have been obtained for several classes of independent and dependent random variables. In this paper we extend this theory and show that not only the central limit theorem, but every weak limit theorem for independent random variables, subject to minor technical conditions, has an analogous almost sure version. For many classical limit theorems this involves logarithmic averaging, as in the case of the CLT, but we need radically different averaging processes for ‘more sensitive’ limit theorems. Several examples of such a.s. limit theorems are discussed.