The stat guy and the analyst

I think Gary Danielson is one of the best color guys in the business.  His opinions on other aspects of college football, though, often don’t click with me.  That being said, I found this exchange he had with Bill Connelly at the recent Sloan Sports Analytics Conference to be a fun read I thought I’d share:

Gary Danielson: To me, stats tell the story of what has happened, not what will happen. I find it interesting, but I just don’t use it a lot. I played for the Lions, and I thought we had a chance to win every game. I didn’t want to find out that we didn’t.

It’s hard to put in highbrow stats into a game. It’s not like the NFL game — it’s a lot different. So many players, such different talent levels. The stats I use are most closely associated with the credible stats that Cris Collinsworth gets in the NFL.

Let me ask you this: If a team, according to stats, gets inside the 20-yard line four times, and they don’t score any touchdowns, is that a good thing?

We actually chatted about this for a few minutes. His point was that creating scoring opportunities is a very positive thing (and potentially a sign that you’ll be creating more), but blowing opportunities is tough. Teams quite often lose because of blown chances (see: Iron Bowl 2013), but teams that generate opportunities are likely to keep generating opportunities. The bottom line: stat folks are often seen as searching for concrete, black-and-white conclusions. Yes, you should absolutely go for it on fourth down here. Yes, this is good, and this is bad. Et cetera. That’s the common perception. But really, it’s the exact opposite. Most stat lovers revel in the gray area, the total lack of concrete answers.

Both get some good points in.  A guy like Bill isn’t arrogant enough to suggest stats paint a black and white world, but there are people out there – shoot, there are commenters here – who will try to insist otherwise.  On the other hand, sometimes there’s more to learn about the sport in Bill’s gray area than Danielson seems ready to admit.

The real issue is that college football is a much harder sport to illuminate with statistical analysis than most others.  But that doesn’t mean the search doesn’t have its rewards.

About these ads

10 Comments

Filed under Stats Geek!

10 responses to “The stat guy and the analyst

  1. Cojones

    Stats are useful in the null hypothesis; i. e., stats demonstrated that the excuses used by the D Coach didn’t hold up.

  2. Ben

    At some point in the season, “the stats” indicated that our defense was holding most teams below their season average at that point. And in the AU game, we gave up about 70 yds more than their season average, and we all know where there 70 yds was.

    So you can read stats to tell you anything you want them to. I remember in a sociology class at UGA our prof said, “Ice cream sales go up in the summer, and so does the murder rate. Can we assume that those two are related?”

    • Sure, but half of the fun of statistics is trying to accurately determine when correlation actually equals causation. I feel like Bill C. and the Football Outsiders guys do a wonderful job of trying to marry those two. I have just as little tolerance for the guy that refuses to acknowledge any validity to statistical evidence (cough Joe Morgan cough) as the guy that believes statistics are the end argument. Statistical analysis certainly suffers from a lot of garbage in-garbage out, but again – that’s half the fun is figuring that part out..

  3. 69Dawg

    Stats are for losers.

  4. Mayor of Dawgtown

    “…shoot, there are commentators here–who will try to insist (that stats tell the whole story).” Yeah, like insisting that the number of fumble recoveries is only determined by luck rather than by coaching, hustle or any other controlled factor. :)

    • Well, you first have to figure out a way to quantify hustle and coaching into a meaningful measure to conclusively attribute fumble recoveries to anything other than luck. And therein lies the rub between objective and subjective analysis.

      • Mayor of Dawgtown

        Soooo……when a Georgia player recovers a fumble you just assume it’s luck rather than hustle, technique or coaching. Right? :)

  5. Spike

    “Stats are for losers and assistant coaches…” Bear Bryant.

  6. Ed Kilgore

    People who hate advanced statistical analysis typically rely on primitive statistics, or on Football Wisdom derived from primitive statistics. More information is always better; what you do with it is the real issue.

    All the “hustle” and “fundamentals” and “attitude” and GATA in the world don’t matter if they don’t produce results on the field. Stats are just a way of measuring that, and improving the odds of getting that ultimate stat of a W.