Recently, our editor-in-chief engaged in a discussion of stats versus the eye test with Phil Simms and Boomer Esiason. I was a little put out by the way it all went down in the end, particularly when Phil Simms said “the stat guys are idiots.” My issue is this. Why does it all have to be about straw men and broad brushes? Who are these so-called stat guys with no eyes?

The conversation started with a question involving Peter King, and Blaine Gabbert’s third down completion percentage. Is Peter King a stats guy? I suppose by strict definition he is. He cited a “stat.” I wouldn’t consider Peter King and myself very similar. I wouldn’t cite to a third down completion for a single player in a vacuum, without having some sense of how it related to others. I would feel the need to look further. What was his completion percentage on other downs? What was the average distance to go? Were they unevenly distributed against certain pass defenses? What did he do on other downs? All these things come to mind.

So I actually agreed with much of what Phil said early on (very early on, I mean like the very beginning). But I’m a stats guy (I think) and so I feel conflicted, because I must be an idiot. Here’s the thing about stats, though. Anything can be a stat. Most of them are recorded by people with eyes. When we throw out bad statistics with good ones, we paint with too broad a brush. If a statistic is bad, we should use our eyes to observe better things, things that are positively correlated with success. If we have eyes trained to look for the right things, the good things that are related to winning, then collectively we can get better info than one set of eyes making a snap judgment based on 10 minutes of observation and all the human bias that comes with it.

I think that stat guys are broadly painted with some crazy non-existent brush that suggests there is some “stat” camp that is in lock-step agreement, and another completely separate camp that consists of staring at things with eyes forced open Clockwork Orange-style for days on end. This is not reality. Eye test guys can misuse statistics with the best of them when they want to make a point with stats. Stat guys can disagree, and frequently do. There are good stat guys and bad stat guys, and often the same person can be both at different times. I can’t pretend I’ve never made errors in how I’ve viewed something. The best ones, though, don’t view things in black and white and are always seeking more questions that lead to more questions. The question and the issue is just as interesting as the ultimate answer, which should never be predetermined.

Back in his 1981 Baseball Abstract, Bill James said this about good versus bad “sabermetrics”.

[G]ood sabermetrics respects the validity of all types of evidence, including that which is beyond the scope of statistical validation.

Bad sabermetrics attempts to end the discussion by saying that I have studied the issue and this is the answer. Good sabermetrics attempts to contribute to the discussion in such a way as to enable it to move forward on a ground of common understanding.

I’ve seen plenty of excellent research that fully acknowledges all those things that currently are not measured. I’ve seen research that uses analysis to nuance the eye test with things like how good Dick Butkus was despite how bad his defensive team’s rankings were. I’ve seen discussions about the relationship between passing and running, which, rather than settling at “we know there is no relationship”, moved on to findings like this about success rate correlation. The adjusted yards for offensive line play at Football Outsiders, and tweaks to it and other things on the team level they do with the DVOA ratings that help me understand how efficient teams are. That Matt Ryan interception thing? None of the “stat guys” I know that think looking at raw interception numbers is a good way to evaluate a quarterback going forward.

I think being good at it is not just looking at spreadsheets; it’s philosophy and art and creativity and actually watching the sport (but also learning to watch the sport with different eyes as you learn things that may challenge conventional wisdom). It’s about recognizing what you don’t know as much as what you do. If it was just about spreadsheets, I would be much smarter to look at company earnings by quarter. I spend far more time thinking while in my car or pondering while pacing around the house and frustrating my wife than I do actually putting data in a spreadsheet. That comes at the end, and just helps me get there faster than Bill James could back in 1977.

We’ve seen bad sportswriters and good sportswriters across the spectrum. We’ve seen good stats and bad stats, and good eye test results and many bad ones (don’t blame Ryan Leaf or JaMarcus Russell on “stats guys”). When you over-generalize, though, you are missing out. Idiots occasionally have some good things to say.

[photo via Getty]