Has the same problem as all of the other lists: There’s way too much noise estimating the base rate of a Poisson distribution from like one or two events (prizes won), setting aside the fact that the prizes have specific taste that is not coextensive with good taste.