The Theory of Over-Analysis

I've been kicking around the theory that the excessive analysis of fights is counterintuitive. UFC 80 seems to be proving my theory. I spent hours reviewing fighters stats, watching fights, etc., to make those predictions. I managed to guess every one of the scheduled broadcast fights wrong! Whether that validates my theory, I have no idea. But it is very interesting.

Also, interestingly, at least two of the fights I mis-predicted went counter to my gut. Gouveia and Penn. Two of the others I basically mailed in: Werdum and Rivera. The first fight I actually guessed right was a gut call, and that was Paul Kelly. The second fight I guessed right was Hardonk.

My recent event prediction accuracy has been: UFC 76 = 70%, UFC 78 = 90%, UFC 79 = 70%, and UFC 80 = 20%. And the amount of analysis is "reciprocal" to to my accuracy! The more effort I put into analyzing, researching, watching fights, the worse I have done. Wild. (My accuracy average for the four UFC events above was 62.5%.)

Statistically, predictions will average out to 50% over time when fights are consistently well-matched, as they are in the UFC. So, I'm not too disappointed. You can see how my earlier good averages were humbled by this very bad set of predictions, bringing down my overall average closer to 50%. When fights are as well-matched as are Joe Silva's, eventually the overall average must approach 50%. (Ask a stats major; they'll tell you.)

Great experiment.

For UFC 81 and UFC 82, I will change up my predictions and pre-event analysis. I will shift back to more "gut" calls, to see how things go. I have predicted far too few major MMA events to make any conclusions, but at this juncture I think I am seeing a pattern.