On his show on Monday, Bill Polian made some comments about statistical analysis.
I think it was (CBS broadcaster and former New York Giants quarterback) Phil Simms on a broadcast I heard who said, 'Statistics are virtually meaningless in football unless you can match them up with videotape and say this is what this statistic specifically means against this defense in this situation against this offense in this situation. Standalone statistics such as exist in baseball – sabermetrics, if you will – have no validity in the NFL. Zero. I was glad to hear Phil say it publicly. It's not said enough. There is a growing – albeit small, but noisy – cadre of people who try to convince you that standalone statistics are really meaningful. They're not. As (Colts linebackers coach) Mike Murphy is fond of saying, 'There are two statistics that count on defense: points surrendered and turnovers.' That's all that counts. Everything else is meaningless. There are some we use where you do match it up with a front or a coverage. We do some pretty sophisticated analysis of it, but you have to match it with fronts and coverages. You have to know what you're looking at. You have to understand why a particular statistic has validity. I'll give you an example. We're now talking about receivers being “targeted” as opposed to how many balls they catch. All that tells you is: a, that the ball was thrown his way – it doesn't tell you why. It doesn't tell you what kind of a ball it was – whether it was good, bad or indifferent. It tells you whether he caught it. It creates a totally meaningless picture. I guess fantasy people look at it. As I've said many times, I have no idea what that's about. I'm sure it's meaningful to them, but in terms of winning games, most of them are not meaningful.
There's more to the quote, and you can look it up and read it. The comment touched off a firestorm over at the Footballoutsiders. The comments echoed similar dismissive comments that Polian made at the MIT Sloan Conference last year. As someone who relies heavily on the Outsiders numbers and on numbers in general, let me make some observations.