March 16, 2011
Engineering professor's lecture illustrates effect of multivariate statistics on basketball match-ups
Submitted by Erinn Barcomb-Peterson
The subject of much discussion, even some worry for Wildcat fans, is the match-up between the fifth seeded Kansas State University's men's basketball team and No. 12 seeded Utah State Aggies. The game is Thursday, March 17.
According to Shing Chang, associate professor of industrial and manufacturing systems engineering, the Wildcats don't much to worry about from their AP-ranked No. 19 foes -- statistically speaking, that is.
Chang will use research by Peter Tiernan of Bracketscience.com to illustrate the effect of multivariate statistics in a lecture to his Quality Engineering class March 16 -- giving his students plenty of time to fill out their brackets before the Wildcats tip-off.
"Over the past 26 years, five seeds have won nearly two-thirds, 69 out of 104 games or 66.3 percent, of their first-round match-ups with 12 seeds," said Chang, citing Tiernan's research. "However, a closer look at the data from all those games shows an upset is more likely if the No. 5 seed has any of seven qualities present."
These qualities include:
* They didn't go to the previous tournament.
* They have a pre-dance losing streak of two or more games.
* They've won fewer than five of their last 10 pretournament games.
* Their strength of schedule is ranked weaker than 90.
* They score fewer than 66 points a game.
* They don't have any junior or senior starters.
* Their combined rebounding/turnover margin is lower than two possessions.
Five seeds possessing any of these attributes are only 11 for 24, or 45.8 percent, in avoiding first-round upsets. All other five seeds are 81.7 percent proficient, going 49 for 60, in dispatching 12 seeds.
"The good news for K-State is we're part of that second group," Chang said.
Based on this research Arizona is the only No. 5 seeded team that falls into the victim group, he said.