Re: [math-fun] secretary problem: "harmonic strategy" conjecturally approximately optimal

correction: in my proof sketch, globally replace P[k-1] by S[k-1] where S[j]=sum(P[i] for i=1,2,...,j). This was just a typo, fixing it does not alter the alleged validity of the argument. There is hopefully a considerably simpler way to clean up the argument (assuming it was not just mistaken somewhere).