CHAMPAIGN, Ill. — For the average college basketball fan looking for an edge in a March Madness office pool, a University of Illinois expert in statistics and data analysis has some advice on how to pick winners: After the Sweet Sixteen round of play, ignore a team's seeding, which is a statistically insignificant predictor of a team's chances of winning.
According to Sheldon H. Jacobson, a professor of computer science and the director of the simulation and optimization laboratory at Illinois, picking the higher-seeded team to beat a lower-seeded opponent usually works only in the first three rounds of the tournament. Once the tournament enters the Elite Eight round, a team's seed in the tournament is irrelevant.
"In the Sweet Sixteen round, the rankings still hold – but just barely," Jacobson said. "From the Elite Eight round and onward, you might as well pick names out of a hat."
Jacobson, who along with graduate student Douglas M. King wrote an article titled "Seeding in the NCAA Men's Basketball Tournament: When is a Higher Seed Better?" published in the Journal of Gambling Business and Economics, said the impetus of the study was to see if a team's seeding was a good predictor of how far the team ultimately would go in the Big Dance.
"You would expect once you get deeper in the tournament that the higher seeds would continue to dominate," Jacobson said.
"But after the Sweet Sixteen, top seeds stop dominating. For just about any team they play, no matter what their initial seeding was, the odds of either team winning is reduced to a coin flip."
Tournament seedings, which are determined by a ten-member committee of NCAA basketball athletic directors and conference commissioners from across the country, are an easy, convenient predictor for people with little knowledge of the current college basketball scene, but are ultimately ineffective in predicting the final three rounds of the six-round tournament, Jacobson said.
"People often overvalue seedings," he said. "The best advice is, pay attention to them early in the tournament, but as the tournament gets going, remember that their usefulness as a predictive measure fades."
In last year's tournament, the outcomes fell in line with Jacobson's predictions.
"Last year was a classic year when all of our statements about statistics and probability came true, which is that the high seeds were able to get to the Elite Eight very easily, but after that you could flip a coin in terms of who is going to win," Jacobson said.
All four of the number one, two and three seeds made it through to the Sweet Sixteen, "which is exactly what our research says, that seeding does make a difference in the early rounds, and the top three seeds are going to be pretty predictable until the Elite Eight round," Jacobson said.
For a team to makes its way into the Elite Eight round, they try to avoid playing a one-seed as much as possible, Jacobson said.
"Eights and nines play ones in the second round, when their ranking still has value," he said. "Fours and fives potentially could play a one in the third round, when one still has value. So if you want to go far in the tournament, I would rather be a 10-seed than an eight or a nine, as paradoxical as that may sound."
When that happens, upsets are more likely to occur.
"If you have a six- or seven-seed playing a one-seed in the Elite Eight, that's a prime upset candidate," Jacobson said. "If all else fails, pick a name out of a hat or flip a coin. Statistically, it won't make a difference."
Jacobson said that other qualitative factors outside of a team's initial seeding such as player match-ups, a team's style of play and its relative "hotness" or "coldness" prior to the game have a greater effect on the outcome of contests in the later rounds of the tournament.
"By the Elite Eight, you have to study the more qualitative aspects of a team," he said. "You have to pay attention to intangibles such as match-ups, injuries, how close they are to their home and how many home fans are going to be there. Those factors make more of a difference than seeding."
What do the researchers hope happens in this year's tournament?
"Hope isn't the issue – we just observe it," Jacobson said. "With over 25 years of data, you start to see trends, and all we're doing is observing those trends. But I think everyone loves the underdog, the Cinderella. The hard part, of course, is figuring out who the Cinderella is. It's very difficult to predict statistically who that is."
Although it may be difficult to predict who will wear Cinderella's glass slipper, and who will turn into a pumpkin at the stroke of midnight, Jacobson said that the tournament hasn't seen any real outliers in significant numbers for a couple of years.
"The last time we saw anything like that was in 2006 when we had a four-seed and George Mason, the 11th-seed," he said. "In 2005, we had a four-seed and a five-seed. So we're due for a year where we'll have a lot of early upsets. I've been saying that for a few years, but we really are due for one, statistically speaking."
The Journal of Gambling Business and Economics