Numbers of success-runs of specified length until certain stopping time rules and generalized binomial distributions of order \(k\) (Q5936213)

The authors introduoe a new way of counting success runs, which they call \(\mu\)-overlapping, and a new distribution of order \(k\), which they call generalized binomial distribution of order \(k\) and denoted by \(B_k(n,p, \alpha)\). For \(a=1\), \(B_k(n,p, \alpha)\) reduces to the binomial distribution of order \(k\) of \(K\), see \textit{K. Hirano} [in: Fibonacci numbers and their applications. Math. Appl., D. Reidel Publ. Co. 28, 43-53 (1986; Zbl 0601.62023)], and \textit{A. N. Philippou} and \textit{F. S. Makri} [Stat. Probab. Lett. 4, 101-105 (1986; Zbl 0584.60024)]. Let \(k\) and \(l\) be positive integers satisfying \(l<k\). For the iid Bernoulli case, the authors show that the number of \((l-1)\)-overlapping occurrences of success-runs of length \(k\) until the \(n\)th overlapping occurrence of success-runs of length \(l\) is distributed as \(B_{k-l} (n,p,2)\). They also treat the same problem for the case of higher-order dependent trials. The distribution remains essentially unchanged.