Roger, that is not fair to the point of being ridiculous: to expect 99% before saying someone beats chance.

Where I say something is wrong is: how could Amit beat random 92% compared of the time? He should get 50% (or maybe 40 to 60%). This is not a tiny amount of predictions we are talking about. Whether the method does not ake sense or not should have nothing to do with the statistical evaluation.

Chris