Dec. 2, 2010
UI economist develops alternative college football ranking
The final Bowl Championship Series (BCS) standings will be released this weekend, and when they are, expect another chapter in the ongoing controversy over polls, playoffs and strength of schedule.
Stacey Brook, an economist in the University of Iowa’s Tippie College of Business, has his own contribution to the argument -- an exhaustive formula that tries to determine the best team by measuring each team’s overall productivity. After all, he said, economists, stock analysts and management experts measure a firm’s performance based on its production, so why not apply the same measures to college football teams.
Especially considering that improved performance by both firms and football teams can lead to millions of dollars in revenue, as universities can cash big checks if they get picked for the right bowl game.
“About 65 percent of the total bowl payout goes to the five BCS bowl games, while the remaining 35 percent comes from the remaining 29 non-BCS bowls,” Brook said. “Clearly, being eligible to compete in a BCS bowl is much more financially rewarding to the team and its conference.”
He admits his rankings can raise a few eyebrows, which he said demonstrates the limits of using polls to measure the best of anything, especially when nobody defines exactly what you should be measuring. For his own poll, Brook started with the presumption that productive teams —- and, hence, better teams —- score more points than their opponents and give up fewer points.
“Then I stepped back and asked, how do teams score points, and how do they keep their opponents from scoring points?” he said. After examining several measures, he finally settled on 17 statistics that best measure a team’s productivity on each side of the ball, including yards gained, number of first downs, touchdown scoring percent, number of offensive plays, missed and made field goals, and turnovers.
He then plugs those numbers into two lengthy formulas, one to measure offensive productivity and one to measure defensive productivity. Some statistics are weighted because his research has found that they are more important to a team’s success.
Subtracting defensive production from offensive, he arrives at a list of the most productive teams.
And that’s when the eyebrows start going up because his survey produces some curious divergences from the polls (he lists each week’s results and explains the numbers on his Team Sports Analysis blog at http://teamsportsanalysis.blogspot.com/).
For instance, Auburn, ranked at the top of the BCS poll this week, is only in 16th place according to Brook’s math, thanks to a less than stellar defense that lowers its overall production score. Boise State, which fell out of the BCS top 10 after losing to Nevada last week, remains at number 1 in his ranking.
Iowa’s stifling defense had the Hawkeyes at 10th in his survey, until the loss at Minnesota sent them to 15th. In other polls, the loss knocked them out of the rankings entirely.
And Hawaii, which just entered the AP poll for the first time this season at number 25, is in 8th place in the Brook ranking, thanks to a prolific offense that moves up and down the field like the marching band at halftime.
“There are some counter-intuitive results, to be sure,” he said.
But he said such divergences are the inevitable result of using polls to measure something without first adequately defining what should be measured. Brooks’ own formula takes into account 17 statistical measures and is weighted toward time of possession, turnovers and total yardage. But the people who vote in the USA Today or Harris polls that determine BCS rankings each have their own formulas that consider different factors. Some may favor total points or overall yardage, some may vote based on how the coach treats the voter during the post-game press conference or how cute the school mascot is. Whatever it is, it’s likely different from Brook’s, and different from the other poll voters.
“Voter bias and the degree to which they’re biased definitely has an impact on how the rankings turn out,” said Brook.
His research also has one other interesting finding —- that the strength of a team's schedule is overrated, and teams like Boise State and TCU should be considered for the national championship game as those from a BCS conference. Brook’s strength of schedule ranking analyzes how each team performed against other teams and where those opponents are ranked. In every analysis, most every team has the same strength of schedule score, rendering the factor statistically insignificant. Total production, he said, seems to be little impacted by a team’s schedule.