Yep, too many folks out there that can't seem to grasp the fact that correlation does not equal causation. Of course, you should also take into account the duration of the 'experiment' when looking at such correlations. For example, if something's happened 10 times, and each time something else also happened, that doesn't mean much. Now what if the Superbowl had been going on for the last 100 years, and every year it had correctly predicted the direction of the market? The odds of that happening by some statistical fluke are pretty remote, so you'd almost have to start looking to see if perhaps there wasn't some sort of causational factor in there, such as the fans of one conference being wealthier than the other or something like that.
Forget all that CAPM, sound business practices BS, I'm picking my stocks based on the number of medals won by Algeria this year in the Winter Olympics darn it
