Psychologists say overconfidence predicting upsets can hamstring NCAA pool participants
For those who've agonized over an NCAA tournament bracket only to be beaten in a pool by someone with zero knowledge of college basketball, an answer.
CHEYENNE, Wyo. - For those who've agonized over an NCAA tournament bracket only to be beaten in a pool by someone with zero knowledge of college basketball, an answer.
Psychologists Sean McCrea at the University of Wyoming and Edward Hirt at Indiana University-Bloomington have studied why this happens. Their conclusion: Deep analysis often does more harm than good.
Avid college basketball fans tend to predict more upsets than actually occur and moreover don't outperform random chance in predicting which underdogs win, the psychologists say.
Compared to those who play it safe - neophytes, maybe, who rely heavily on how teams are seeded - they end up hobbling themselves, say McCrea and Hirt, who published their findings recently in the Journal of Applied Social Psychology.
"We feel that our skill should mean something," Hirt said. "But I think most of the skill we have is already comprised in the seeding and we can't do better than that."
McCrea and Hirt are college basketball fans. Both said they were less motivated by the noble advancement of science than by repeated schooling by know-nothings.
"There always seemed to be someone who didn't watch a single game all season that would do really well," McCrea said. "It kind of got us scratching our heads."
The psychologists analyzed more than three million entries in ESPN.com's Tournament Challenge competition. The average success rate for each team picked was 75.2 per cent in 2004 and 72.9 per cent in 2005.
Had everyone simply picked winners based on how teams are seeded, their success rate would have been higher: 87.5 per cent in 2004 and 75 per cent in 2005.
So why do people choose certain teams to upset others?
Hirt and McCrea asked people who'd passed a test of college basketball knowledge to fill out NCAA brackets and rate each team for strength and plain-old likeability.
Again, the study participants didn't average better than the tournament seeding. And while strength and likeability offered guidance, that didn't explain many choices.
The psychologists noted that people sometimes picked upsets just to boost their perceived chances of being correct overall, a phenomenon known as "probability matching."
Thus the quandary of tournament pools: Going strictly by how teams are seeded is a statistically sound approach for beating others on average but a poor one for actually winning the pool - the whole point of even entering - because upsets are all but inevitable.
This year's topsy-turvy tournament proves that upsets are unpredictable and inevitable. So which upsets to pick?
"If you're going to pick some upsets, it's probably a good idea to avoid the ones that everyone else is picking," McCrea said. "You want to have a bracket that sets you apart from everyone else, given that no one is really that good at it."
Sometimes the NCAA selection committee takes heat for their seeding choices. But McCrea gives the committee credit for applying objective data that's difficult if not statistically impossible to improve upon.
At Jackson's Sports Grill in Cheyenne, Mike O'Donnell, who was watching the tournament Thursday, said he chose every 12th-seeded team to beat every fifth-seeded team in a pool he entered.
"Every year, it seems like two or three of the 12 seeds win," he said. "It's kind of unlucky for the fives."
He did all right with another first-round underdog he picked. Tenth-seeded St. Mary's beat seventh-seeded Richmond, 80-71.