I was never a fan of contrived problems with arbitrary solutions. I agree that it offers a certain amount of insight into the interviewees thought process, but I tend to be a fan of problems with provably correct answers that require some rigor. Perhaps that's my problem?