Despite the failure of analytics to determine a collegiate football champion, I love the college football season as much today as I did when I was a kid and Grandpa North was the laundry manager for the Athletics Department at Brigham Young University (BYU).
This was back in the days when players only got new uniforms every few seasons; when the players knew the laundry manager and he knew them -- after all, he was the one who made the uniforms spotless and the helmets shine on crisp autumn Saturday afternoons. Grandpa loved watching and listening to the Cougars, and so I come by my passion for college football honestly. On pleasant fall "away game" afternoons, we’d mow the grass and then perch lawn chairs under the willow tree, munching crackers, sipping Coke, and listening to Paul James call the action on the radio. Sometimes we won, sometimes we lost, but we were always building lasting memories. As the 1970s gave way to the '80s, Grandpa retired, but we still kept up with the team, and, by then, the team was getting really good.
In 1984, BYU was the only Division I team in the nation to run through the season undefeated. On the strength of several consecutive successful seasons and an early win over a highly regarded Pitt team, the Cougars climbed high in the rankings, eventually reaching the top spot. After a thrilling victory over Michigan in the Holiday Bowl, the Associated Press and other organizations crowned BYU the national champions. Sadly, Grandpa North had made his final run for the end zone in October of 1983, but I’m sure he watched that championship year from a comfy cloud, probably swigging from an icy-cold Coke.
Alas, several major college football programs cried foul! The universities of Washington, Nebraska, and Florida had only one loss each. They had much more storied programs, they played tougher schedules, and had they been given a crack at the Cougars -- surely they’d have beaten this lowly team from Utah. Something had to be done.
There was talk of creating a playoff, but this would threaten the relevance (and even the very existence) of bowl games -- games that were becoming increasingly lucrative. After several years, those who stood to gain the most by protecting the bowls eventually created the Bowl Coalition, then the Bowl Alliance, and finally, the Bowl Championship Series (BCS), which is the system we’re using until 2014. Not surprisingly, through each of these iterations, analytics served an increasingly focal role in determining which team would eventually play for a chance to be named the best. Ultimately, none of these systems have survived because they have failed to yield acceptable results, regardless of what the BCS’ own biased website says.
Respected sports author Dan Wetzel has penned numerous articles and an excellent book detailing the shortcomings of using analytics, especially hopelessly flawed (even stupid?) analytics, to determine a champion. For example, the computer rankings have been criticized in one specific instance for erroneous data entry, and have consistently been scrutinized for a lack of openness.
In another head-scratcher, one computer ranked Arizona Western, a two-year junior college, as the 30th best team in the nation at one point last year. (They were good, but the Matadors still suffered one loss at the end of an all-junior-college slate of opponents, so a computer ranking ahead of the likes of Ohio State and Florida State was probably reflective of a flawed data model.) Even this year, Florida is ranked five hundredths of a percent higher than Alabama by the BCS computers, but the Gators never have a chance to prove their championship worthiness on the field.
Based on such examples, Wetzel quotes noted University of California-Irvine statistician Hal Stern in one of his articles: “I am advocating a boycott of the Bowl Championship Series by all quantitative analysts.”
That’s how bad the BCS’ analytics are. In 2014, the BCS will be the latest system to end as it yields to a four-team playoff system. Most experts remain leery of this system because the BCS brain trust designed it to still protect lucrative bowl games, and because the selected teams will still largely be determined by suspect analytics.
No so-called mid-major team since the ’84 BYU Cougars has achieved a national championship. Are there some problems in the world that analytics just can’t solve? Will we one day just settle it on the field?