Random walks whose concave majorants often have few faces (Q2576375)

Let \(X_{1},~X_{2},\dots\) be independent random variables with a common continuous distribution function \(G,\) let \(R_{n}=\max \{k:X_{i_{1}}<\dots<X_{i_{k}},~1\leq i_{1}<\dots<i_{k}\leq n\},\) and let \(F_{n}\) denote the number of pieces in the smallest piecewise linear concave majorant of the set of points \(\{(0,0),(1,S_{1}),\dots,(n,S_{n})\},\) where \(S_{k}=X_{1}+\dots+X_{k},~1\leq k\leq n.\) \textit{C. M. Goldie} [Math. Proc. Camb. Philos. Soc. 106, No. 1, 169--177 (1989; Zbl 0671.05001)] observed that \(R_{n}\overset{d}{=}F_{n}\) for any \(n\geq 1.\) Since \(R_{n}/\log n\rightarrow 1\) a.s., the authors look for a possible analogous strong law for the sequence \(\{F_{n}:n\geq 1\}.\) They construct a distribution function \(G\) such that \(P(F_{n}=m\) i.o\()=1\) for each \(m\geq 1,\) and so conclude that the almost sure behaviour of \(\{F_{n}:n\geq 1\}\) may differ radically from that of \(\{R_{n}:n\geq 1\}.\)