College Football

The wrong way to pick College Football Playoff teams

The college football commentariat produces better and more interesting analysis than its NFL counterpart. Tony Barnhart's criteria for a playoff selection committee is an exception to that rule.

By Michael Bird@BravesAndBirds Jun 3, 2013, 9:32am EDT

Last week's discussion of the composition of a College Football Playoff committee – a discussion highlighted by Jeremy Foley setting off alarm bells in numerous fan bases by joking that he nominated Mike Bianchi for the committee – reminded me of a funny little juxtaposition that took place two weeks ago.

On the morning of May 13, Spencer Hall drew a series of huzzahs from various circles (along with a prickly, humorless response from CBS’s Pete Prisco that illustrated Hall’s point) for his piece on why the NFL media is so drab. Then, on the afternoon of May 13, Tony Barnhart, the self-proclaimed "Mr. College Football," wrote a piece on what a Playoff selection committee should consider when picking teams and illustrated that the college football media has its own "self-contained ecosystem of volume without substance."

More: Sports Illustrated roundtable on Playoff selection committee

In this case, however, Barnhart wasn’t so much using the "bloated language of derivatives traders" as he was the Washington insider’s platitudes that once would have come from David Broder and are now the province of Politico's Jim VandeHei and Mike Allen. In short, Tony is not big on data. Instead, you should just trust him that his friends can meet in a private room at the local country club (perhaps Augusta National, since Tony lists Condoleezza Rice’s membership there as a credential for separating elite college football teams) and make decisions for the rest of us.

Look at Barnhart’s guiding principles for the committee. First, he says that we should not get hung up on bias. Isn’t that pretty much what every middle-of-the-road deficit scold says? If the Democrats and Republicans could just put away their ideological preferences, then we could have the moderate budget that hews ever so closely to the preferences of the deficit scold. You know why that won’t happen? Because the political process exists to allow for the expression of those ideological preferences. We cannot wish bias away. We can only ensure that the biases balance out in the fairest way possible. That’s not going to happen when one-quarter of Barnhart’s proposed nominees are from the SEC. As bad as the current BCS formula is, at least the various biases are more likely to be balanced out when the number of voters is three digits instead of two.

Next, Barnhart says that the committee members have to go to games and that their live impressions of teams should be the tie-breaking consideration. Let’s count all the ways that this is a bad idea. It reduces sample sizes in a sport that already suffers from a small N. How often are the voters going to see the teams fighting for the last spot in the playoff? Once? Twice? Maybe they catch a team on its best or worst day. Maybe they watch the team in a particularly favorable or unfavorable matchup.

Third, Barnhart ignores the wealth of evidence that eyewitness testimony is inherently unreliable. Human beings are noted both for our inability to use our eyes to make accurate assessment and for our overconfidence that our interpretation of what our eyes tell us is correct. Think about the teams that have been involved in close should-they-play-for-the-title debates during the BCS era. Don’t you have one or two mental images of those teams in your mind’s eye? Do we really want playoff spots decided based on whether the mental image of Junior Rosegreen’s hit on Reggie Brown predominates over the Oklahoma defense gang-tackling Vince Young?

And then we get to the best part, which is Barnhart’s guiding principle that "nobody knows the game better than the people who played it and coached it." This is what connected Hall’s piece in my brain, specifically the line about The NFL Today on CBS being "the master class on bellowing nothing-ism." Why are NFL pregame shows in general and The NFL Today so unwatchable? Because they hire former players and coaches without paying the slightest attention to whether those men say funny, smart, or insightful things about football.

I’d wager that I can go into a random Taco Mac on a Sunday afternoon and the guy or gal to my right at the bar will be capable of more interesting commentary than Shannon Sharpe or Dan Marino. The abilities to run fast and catch or throw a ball are not necessarily correlated with the ability to analyze and communicate.

So why do we think that coaches and players would necessarily do a better job of determining whether Alabama or Oklahoma State are more deserving of getting the last spot in a playoff? Take Archie Manning, who is one of Barnhart’s 16 proposed committee members (and, not surprisingly in a paean to insider thinking, sits next to Barnhart in the CBS studio every Saturday in the fall). What exactly is on Manning’s resume that makes him more qualified than you, me, or hundreds of other college football fans to interpret data? Manning is a familiar figure and would therefore lend a patina of establishment credibility to the work of a selection committee, but outside of this political role, he does not have any credentials to justify Barnhart’s confidence.

Most notably, Barnhart omits any reference to advanced statistical analysis. With limited exceptions, pro sports teams use quants to aid in their decision-making processes. Vegas sports books and sharps use sophisticated models to set and gamble on lines. When there is actual money at stake, nerds are invaluable. But when a mouthpiece for the SEC is picking his friends and peers for the social event of the year, then there is no room for statistical experts.

This fallacy is evident when Barnhart lists his selection criteria – "record, strength of schedule, strength of opponent's schedule, quality wins, quality road wins, and the all-important eye test" – and omits the factor whose exclusion from the computer rankings after the 2001 season – margin of victory – turned the subsequent rankings into "nonsense math."

On the whole, Hall seems right that college football commentary is more interesting and not as serious as writing about the NFL. You need only contrast the oft-witty signs and occasional Lee Corso-uttered f-bombs of College GameDay with the forced guffaws of the NFL shows if you need an illustration.

However, the prattle about the NFL is just background noise. College football is going to have to turn its debate about process into an apparatus that will have a hand in deciding future national champions. Our prattle is better, but it has more tangible impact, which means that Barnhart’s thinking is dangerous if decision-makers take it seriously.