Personalize your DealNews Experience
- Create an Account or Login
- Select your Interests
- Toggle your Interests On/Off
It's really simple to set up. Create an account or log in to get started.
Whether it's for bragging rights or maybe a cash prize at your office (we won't tell the IRS), everyone wants to win with their NCAA March Madness bracket. You can probably select a good percentage of winners; maybe you'll even get the Final Four correct, or the overall champion at least.
But could you correctly predict the outcome of all 67 contests? It might as well be magic. Depending on the year, the odds have been estimated anywhere from an easy one in a billion to as much as one in 9,223,372,036,854,775,808 (that's 9.2 quintillion). To help increase your chances of achieving this feat (or just winning), DealNews has enlisted the expertise of Dr. Tim Chartier, a Davidson College professor and bracketology expert.
Because so many people interested in the NCAA tournament are basketball nuts, they tend to pick their brackets with a certain amount of bias for their favorite teams. "The people who do well, they're not the sports fanatics," Chartier says. "The fans tend to overthink it and have attachments to who they think should do well." His advice: Make like a mathlete and use statistics to your advantage.
Chartier says that to make your bracket really hum, you have to go beyond a simple win-loss schedule. He cites recency as a major factor: that is, how well a team does late in the season, just before the tournament.
"When we first started, that was all we did: weight with recency," he says. "We had a student who did that in 2009, and he beat 97% of the more than 4 million brackets submitted to ESPN. In 2010, we had a student who beat 99% of 5 million brackets. And last year we had a cheerleader, and she beat 96% of the 8 million brackets."
Chartier identifies two other factors you can "upweight" to get your bracket more accurate. He has his students take a close look at a team's road record. "That's important because the whole tournament is away, if you think about it," he says. "But you can have an advantage based on where you play if it's close to your home court."
The second involves winning streaks; you'll want to favor teams who've proven they can string those wins together. "The way you win the tournament is that you have to have a five- or six-game winning streak. Obviously if you can't do that you can't win it all."
Just as baseball has the "Moneyball" phenomenon, basketball has its own version of statistical analysis, called APBRmetrics. And when it comes to basketball stats, Chartier gives his due to statistician and father of APBRmetrics Dean Oliver, who identified four factors that contribute to a basketball team's success: shooting (40%), turnovers (25%), rebounding (20%), and free throws (15%).
The exact method Chartier uses for putting together 350 linear equations and 350 unknowns that help him form the ultimate bracket is complex. But we believe him when he tells us that these linear systems, when you break them down, "are based on what you learn in middle school. There's a lot of math software that does it for you so you don't even have to know how to do it."
The equations can include variables such as the weighted factors mentioned above, and perhaps more importantly, the strength (or weakness) of a team's schedule over the season. "The teams with the highest winning percentages aren't always in the tournament, so you have to be really careful in thinking winning percentage is the most important thing," Chartier adds. "It isn't the fact that you can go in and win, but whether you can win against hard teams. That's what gives you the huge bump."
Readers, what are your tips for brackets this year? Let us know in the comments below.