How Moneyball Could Deflate College Football
WHY YOU SHOULD CARE
Because sports are a lot more interesting when the same teams don’t win all the time.
If only it were still 1971 … or 1995. That could be a Nebraska state anthem, at least for sports fans who remember those undefeated seasons. Though still Bowl contenders, the Cornhuskers haven’t touched glory in a while. But let’s forget all that! The team whose fans wear giant ear-of-corn hats has a secret weapon (other than those fans). And it’s not some 10-pack of five-star players wooed from more glorious pastures with the promise of … pastures. No, it’s Tucker Zeleny, sports analytics director.
“I want to add some value here,” says the secret weapon, who finds himself running a newly created department just a year after joining the program as a part-time data analyst.
Welcome to yet another version of Moneyball — only this time it’s not Brad Pitt using a data-driven strategy to make winners out of the underdog Oakland A’s, it’s one of those high-powered NCAA Division I teams that occupy our Saturday afternoons bringing on the nerds. And that has some fans and sports economists sweating, because once powerhouses like Nebraska hit the computers, the gap between the haves and have-nots in college football may only grow — in both the win and money columns.
Not all SportSource clients get to drink from the same Datorade.
Could this Big Data approach tilt things even more strongly in favor of teams with the mostest? “The answer to that question is probably yes,” says Ellen Staurowsky, a sports management professor at Drexel University. But what happens next is a subject of fierce debate. Staurowsky thinks statistical analysis will mostly offer a huge advantage for early adopters with Big Bucks for Big Data, one that should fade as others catch up. But the reverse could also be true — for instance, if smaller college football teams enjoy “short run” success with statistical analysis that helps them identify the best plays, players and strategies, only to get crushed by their big-budget rivals over time, says Dennis Wilson, an economist at Western Kentucky University. (Which, by the way, is ranked No. 93 in the USA Today list of big-budget NCAA athletics programs — a full 67 places behind Nebraska.)
It’s not as though college football is all that well-balanced to begin with. Where the NFL is, in many respects, essentially socialist in nature — its teams pool and share about 60 percent of all league revenue, and the worst teams each season get a leg up with better draft picks — the college game is much more unbridled laissez-faire. According to the NCAA’s 2014 report on revenue and expenses, the biggest-spending team in the top-tier Division I-A laid out $146.8 million over the year, while teams in the bottom half of all teams in the division spent $62.2 million or less. Yes, the NCAA sometimes reshuffles the deck by moving teams between conferences, but that’s often just a Band-Aid. Intraconference discrepancies can still be enormous, Wilson says — just compare Mississippi State and Texas A&M.
Here’s how analytics and Big Data could just make things worse. Increasingly, teams have access to much more comprehensive data on both their rivals and their own team’s performance on the field. “Up until two years ago, everything we did was in-house,” Ball State head coach Pete Lembo tells OZY. Now the team uses SportSource, a data analytics company that serves 20 teams in Division I-A, from smaller programs like Ball State (4–4 in 2014) to behemoths like Oregon (8-1 in 2014). It’s all in hopes that mining the data in new ways can help with everything from recruitment to play-calling to game preparation.
But not all SportSource clients get to drink from the same Datorade. While pricing is “similar” for different programs, some teams pay for “add-on features” like more customized information, says SportSource founder Stephen Prather. Worse, smaller teams have way fewer opportunities to exploit analytic information. In recruiting, for instance, data might recommend particular recruits to a smaller-budget team like Vanderbilt — but also to a big-budget team like Alabama. And in such cases, there’s no draft to help Vandy get the player.
Smaller budget teams already have a hard time keeping high-performing coaches; for example, after James Franklin led Vanderbilt to a 9-4 season, Penn State quickly hoovered him up with a head coach offer. They could easily face a similar problem with their data geeks. But while programs ramp up their analytics usage, the numbers are still untested and it’s hard to say what tangible results they provide. Buying a large analytics platform might be like buying a Rolls-Royce when teams “could just be buying Volkswagens,” says sports economist Craig Depken. That’s because football’s been a tougher sport to model than others, like baseball. With 22 players on the field, “there’s so many moving parts,” Prather says.
At Ball State, Lembo says analytics make him “more aware of what reality is.” Last year, in the final game of the season against Bowling Green, Ball State used statistical analysis to figure out that running a slower offense might give them an edge — and they pulled off an on-the-road upset. But you can find skepticism about statistical analysis everywhere, even in the pros. “I still haven’t found anybody that can put a numerical value on heart, toughness, competitiveness,” Buffalo Bills GM Doug Whaley tells OZY.
If analytics do lead to greater team imbalance, fans will be the ones to suffer. When the teams aren’t competitive, “it’s a long, dark winter,” Wilson says.