A simple derivation of a complicated prophet region (Q1599069)

The paper deals with the problem considered by \textit{R. P. Kertz} [J. Multivariate Anal. 19, 88-112 (1986; Zbl 0598.60044)]. A straightforward proof of the theorem concerning the complete comparison of the values to a statistician, observing sequentially a sequence of i.i.d., \([0,1]\)-valued random variables, and the value to a prophet, observing the same sequence, and picking the largest observation to occur, is given. The proof is based on a balayage technique [see \textit{T. P. Hill} and \textit{R. P. Kertz}, Proc. Am. Math. Soc. 83, 582-585 (1981; Zbl 0476.60044)] and the properties of the recursively defined sequence of the following functions. For \(q\geq 0\), \(i=1,2\ldots,n\) and \(n=2,3,\ldots\) define \(f_{i,n}(q)= (\frac{n}{n-1}f^{n-1}_{i-1,n}(q)+q^n)^{1/n}\), where \(f_{0,n}(q)=q\).