Precise rates in the law of logarithm for the moment convergence of i.i.d. random variables (Q860604)

Let \(\{X, X_n,n\geq 1\}\) be a sequence of independent and identically distributed random variables with mean zero and finite variance \(\sigma^2\). Let \(S_n= \sum^n_{k=1} X_k\) and \(M_n = \max_{1\leq k\leq n}|S_k|\), \(n\geq 1\). Let \(r> 1\). The authors prove that \[ \lim_{\varepsilon\downarrow\sqrt{r- 1}}{1\over -\log(\varepsilon^2- (r- 1))} \sum^\infty_{n=1} n^{r-2-1}E\left[M_n- \sigma\varepsilon\sqrt{2n\log n}\right]_+= {2\sigma\over(r- 1)\sqrt{2\pi}} \] if and only if \(E(|X|^{2r}/(\log|X|)^r)< \infty\).