Guessing who will win a Nobel Prize is a bit like forecasting the stock market: Experts don't seem to do it any better than laymen.