Sunday, June 03, 2007

Subjective Best Of Lists Are Impeachable

And now, the triumphant return of college football blogging!

A few weeks ago several blogdudes (I'm sure more, but it's been a while...) made note of The Sporting News' Tom Dienhart ranking all of the current BCS coaches. Mr. Dienhart gave no criteria or objectives upon which he based his list. I assume, it must be subjective. And whenever I smell subjectivity in college football, I get mad. Mad like I've a bellyful of cheap gin.

UPDATE: Bill of Eagle in Atlanta thought on these same lines a full month before I did. I didn't see it before, but much credit to him for being significantly ahead of the game!

See, as I've said before, no other sport is influenced by media subjectivity like college football, which necessarily relies upon the opinions of sportswriters and the opinions of people who rely upon sportswriters to base their opinions to crown its champion (or, if you'd prefer more precision, to select which teams play for the title). So when teams or personalities that are often synonymous with teams receive public praise or criticism, it necessarily affects how those teams are presented, and the initial impressions of particular programs directly affect placement in polls. I find this unfortunate, but even if you don't, it's a fact of the game in its current form.

The problem (well, one of the problems) with Dienhart's list is that it is a road map for "the benefit of the doubt" given to some coaches, but not others. If a team coached by a high rated team on his list loses a couple of early games to drop to 1-2, you'll likely see a column somewhere by someone saying "a Coach X coached team is too good to fall apart. Don't be surprised to see them winning the rest of their games..." or something like that. Teams coached by guys that rank high on lists like this are less likely to drop precipitously in the polls after a loss, or merely that they're more likely to be ranked highly (perhaps without reason) in preseason polls, providing the springboard needed to end up in title games or top bowls.

So this benefit of the doubt issue I see... well, some guys deserve the benefit of the doubt. That's certain. Some guys really are great coaches. So the idea of a list of great coaches isn't exactly bad in the abstract. What's bad is that we have no idea what Dienhart uses to come up with his list. Big wins? Recruiting rankings? Team discipline? Wins and losses? Who gives him exclusive interviews and access? We don't know. And that's why this list is bad.

OK. When I looked at the list I saw probably the same odd rankings most people saw. Houston Nutt ahead of Mark Richt? Chan Gailey ahead of Phil Fulmer? Rich Rodriguez in the top 3? Dennis Erickson in the top 10? But then I tried to come up with objective ways of rating these guys. I wanted to make sure my own opinions weren't wrong (what if Dienhart were right, based upon objective criteria?).

So I ranked the 66 BCS-automatic-bid conference coaches in a number of categories that I think provide some objective analysis:

  1. Longevity. The longer a guy has coached at a major college football program, that means he's probably capable. Guys who can't hack it don't stick around long, and there are few coaches who make the leap to the NFL. That said, I think this is one of the weaker objective categories for determining how good a coach is today. Who would you rather start a program with right now, Methuselah Paterno or Urban Meyer?
  2. Number of National Championships. Winning a title matters. This, I'd say is also one of the weaker objective categories for determining how good a coach is. Specifically, there are a host of major college football programs where, no matter how good a coach is, the resources just aren't there for a national title run (and even if they did win all their games, there are a number of programs which just don't have enough juice to get writers and poll voters to change preconceived notions). Also, not all national titles are equal. Nick Saban, Steve Spurrier and Urban Meyer all have a title in a season where they lost a game, while Tommy Tuberville has an undefeated season, but no title.
  3. Number of Conference Titles. Again, winning a title matters. The other teams in your conference are your closest competitors, so winning titles means you are besting your closest rivals. That said, not every conference title is equal. An undefeated run through a terribly tough conference is better than a three way tie for first with all teams at 5-3 in the conference (which has happened before). And, needless to say, the fact that there are teams included that haven't always played in conferences. Charlie Weis may never win a conference title. That doesn't mean he's a worse coach than Ty Willingham, who won one at Stanford. Joe Paterno coached for 25 years as an independent, so he has the same number of titles as Mark Richt.
  4. Winning Percentage. Winning matters. Winning more than losing. Beating the other team every single week. Of course, not every team plays the same schedule. Some schedules are much more difficult than others. And not every team has the exact same resources. Les Miles has a much easier road to a good winning percentage at LSU than Bobby Johnson has at Vanderbilt.
  5. Winning Percentage vs. Program's Historic Winning Percentage. This, I believe, is the most accurate objective measure to compare coaches. Basically, you look at the particular coach's winning percentage at BCS schools, and compare that to the winning percentage that the school (or schools, if the coach has appeared at multiple BCS programs) has had. The number should be a + or -. For example, if Ralph Friedgen has a winning % of .676, and Maryland's all time winning percentage is .533, Friedgen's number for this category is +.143. This attempts to isolate the effect of a coach upon a program without regard to inherent benefits or drawbacks of the program. Of course there are flaws to this - some programs aren't the same today as they are over a 100-year period. For example, Vanderbilt all time has a program winning percentage over .500, but that's because of good numbers back in leather helmet days.
So how do things shake out? Read on in the next few posts to find out... (and it might take a few days to get them all out)


Anonymous said...

Given the problem of change over time in program status, why not use the winning percentage of the (X + 10) years immediately preceding the coach's hiring, where X is the number of years the coach under scrutiny has worked at the institution?

I can understand why that may be too time consuming, but the entire task already seems daunting as it is. When you've already decided to undertake an incredibly ambitious project, what's another few hours on CFBdatawarehouse?

ATL_eagle said...


here was my take on Dienhart's nonsense with an ACC take.

LD said...

Great post, Bill! Seriously, I didn't see that before writing this whole bit, or else I definitely would've credited you (and I will anyway!)