On the strong law of large numbers (Q5917804)

The connection between general moment conditions and SLLN for identically distributed random variables is examined. Let \(f(x)\) be an positive even continuous function strictly increasing in the region \(x>0\) and \(f(+\infty)=\infty\), \(a_n=f^{-1}(n)\) where \(f^{-1}\) is the inverse of \(f\). The following two statements are proved. 1) Let \(\{X_n\}\) be a sequence of pairwise i.i.d. random variables \(S_n=\sum^n_{k=1} X_k\). If \(S_n/a_n\to 0\) a.s., then \(Ef(X_1)<\infty\). 2) Let \(\{X_n\}\) be a sequence of identically distributed random variables (without any independence condition). If \(\sum^\infty_{k=n}1/a_k=O(n/a_n),\) then \(Ef(X_1)<\infty\) implies \(S_n/a_n\to 0\) a.s.