Not sure where you pulled that from, but that's a wildly inaccurate blanket statement to make about stats in general - 12 years out of a possible 15 or so is absolutely statistically significant when considered in context. That doesn't mean there's a curse, but those numbers definitely fail to reject the null hypothesis that cover athletes have a down season the same year that they appear on the cover.
With that said, I vote Cam. Swag trumps all.
I have written a pseudo-random number generator for a custom cryptographic suite. One of the stringent demands on such a system is the ability to create numbers that are as absolutely random as possible. What people don't understand about randomness is that if you flip a coin 20 times and it comes up heads 17 times, that is a perfectly logical random result. The 50/50 distribution only has to hold true when dealing with a massive number of permutations (on the order of millions) but a random number generator that produced a steady series of heads/tails combinations in order would be viewed as functionally broken.
In the same regards the Madden curse, which has multiple "bad" definitions but very few good ones (so the probability of a bad result is closer to 90% than 50%) having a run of 9-12 "bad" years out of 12 is not only statistically insignificant, it's entirely flawed by the faulty weighting of the probabilities involved (I mean, who considers Drew Brees Pro Bowl season to have been cursed? Idiots, that's who.)