Published by the Students of Johns Hopkins since 1896
May 22, 2022

Professor creates score model for Super Bowl

By Elizabeth Arenz | February 7, 2013

Two weeks ago, Daniel Naiman, Professor and Chairman of the Department of Applied Mathematics and Statistics at the Whiting School of Engineering, predicted that the 49ers would win the Super Bowl by three points.  Describing himself as a statistician since 1980, Naiman said this was simply a fun exercise he had wanted to investigate for a while.

“I was asked if I had anything to say regarding the possible outcome of the Super Bowl so I replied, ‘Well yes, there’s been something I’ve been meaning to try,’” he said.

Naiman created two different graphs using data collected from results within the National Football League since 1970. That year marks the league’s division into the American Football Conference (AFC) and the National Football Conference (NFC).

“I’ve played with the numbers before but this is the first time I’ve gotten this much attention,” he said. “Of course, this is Baltimore, and I’m a big Ravens fan so I imagined there would be interest.”

By determining the average point differences resulting at the end of every season in which an AFC and an NFC team played against each other, Naiman was able to evaluate which conference had been more successful from season to season. With this knowledge, he created a graph detailing whether the seemingly stronger conference of one year actually won the Super Bowl in that year.

“I explored questions I had thought about for a while but I truly only

focused on the low hanging fruit, the easy answers,” Naiman shared.

Naiman found that over the past 22 years, the AFC has had a stronger season and in those seasons, has won 13 times, a 59 percent success rate. During the same time period, the NFC has had stronger seasons for a total of 20 years, winning 14 times in these years for a success rate of 70 percent.

Last year, the NFC’s teams beat those of the AFC by an average of 5.6 points, their best year since 1970. Naiman compared this year’s situation to the last time the NFC had a similar advantage over the AFC. He found that San Francisco’s 49ers had an advantage over Baltimore’s Ravens and should win by three points.

“Doing this well is very difficult, a full-time job,” Naiman said. “This was not a really careful or detailed analysis.”

Naiman maintained that the Ravens could still come out with a surprising win as the AFC’s Colts did in 1971, the same year the NFC had a similar advantage over the AFC beating those teams by an average of 7.8 points. Though the actual results were a flipped version of those found by Naiman, the Ravens still only won the Super Bowl by three points, following Naiman’s score prediction.

Naiman also used a second graph to determine his predictions, this one organizing data regarding how well people who report spreads have performed in the Super Bowl’s past. This analysis, published on a local television station, indicated that in the past 46 years, the game has had 15 underdog wins and 31 favorite wins, meaning that though a team may enter with the upper hand, the probability that it walks away with a victory is not as high as one might think.

Although results at large sporting events are variable, predictions are one way that those fond of betting can do so with a higher success rate.

“I’ve been examining how, historically, people don’t act rationally when betting and they don’t often follow point spreads to a high degree,” Naiman explained. “These types of prediction are interesting for people who bet because when they are quantitatively savvy, it gives them the opportunity to make more money.”

Despite his hopes for an accurate prediction, Naiman is simply happy the Ravens came away from last Sunday’s game victorious.

“I’ve been here in Baltimore since 1982 and now I feel like I grew up here.”

Comments powered by Disqus

Please note All comments are eligible for publication in The News-Letter.

News-Letter Special Editions