Almost sure limit points of and the number of boundary crossings related to SLLN and LIL for record times, inter-record times and the number of record values (Q791986)

For a sequence \((X_ n,n\geq 0)\) of i.i.d. continuous random variables let (L(n),\(n\geq 0)\) and (\(\Delta\) (n),\(n\geq 0)\) be the corresponding sequences of record times and inter-record times, respectively. Let \(\mu\) (n) denote the frequency of record values among \(X_ 0,X_ 1,...,X_{n-1}\), \(n\geq 1\). Define the following random variables: \(Z_ 1(n)=(\log L(n)-n)(2n\) loglog n)\({}^{-1/2}\), \(Z_ 2(n)=(\log \Delta(n)-n)(2n\) loglog n)\({}^{-1/2}\), \(Z_ 3(n)=n(L^{1/n}(n)/e- 1)(2n\) loglog n)\({}^{-1/2}\), \(Z_ 4(n)=n(\Delta^{1/n}(n)/e-1)(2n\) loglog n)\({}^{-1/2}\), \(Z_ 5(n)=(\mu(n)-\log n)(2\) log n logloglog n)\({}^{-1/2}.\) The following result is established: Theorem 1: The set of all almost sure limit points of each of \(Z_ i(n)\), \(i=1,..5\), is the closed interval [-1,1]. There are also several theorems on the moments of the number of boundary crossings, e.g. E[\(\sum^{\infty}_{n=1}I_{(| \Delta^{1/n}(n)-e|>\delta)}]<\infty\), where \(\delta>0\), \(I_{(B)}\) is the indicator function of the set B and [\(\cdot]\) is the greatest integer function.