Tim Chartier shines during the NCAA Division I College Basketball Tournament, AKA March Madness: He's one of the nation's top bracketologists.
Every year, as winter edges into spring and the nation’s top college basketball teams vie for a national title, America gets a fever called March Madness. It’s spreading now. You might have it.
On March 15, teams from around the country began playing what will amount to 63 NCAA Division I Championship tournament games, excluding play-in games. The last teams standing fight for the title April 4 in Houston. Millions of fans are following the games on television.
At the same time, in homes, offices and online, millions are also making game-by-game predictions for how the teams will fare and which will emerge as champ, plotting their guesses on tiered charts called brackets. In more than a few cases, money will be on the line.
No one has ever produced a perfect bracket: The odds of correctly forecasting the outcome of every contest have been calculated as 9.2 quintillion to one. That’s why Quicken Loans and Warren Buffett were comfortable offering a $1 billion prize in 2014 to a person who could.
Still, students of Davidson College professor Tim Chartier (PhDApMath’01) have fared exceptionally well in March Madness bracket pools — in 2010, one student’s bracket performed better than 99.9 percent among five million brackets submitted to ESPN. This has helped Chartier, an applied mathematician (and experienced mime), earn a guru-like status among the nation’s bracketologists.
The prediction formulas developed by Chartier and his students have become so accurate that they’ve landed the professor on ESPN and National Public Radio and in The New York Times, and opened opportunities unforeseen when he was a CU-Boulder graduate student in the 1990s.
The attention has led him deeper into sports analytics research and related projects, including a book with Princeton University Press called Math Bytes, which has a chapter on the mathematics of bracket predictions, work with the German national basketball team and an endless stream of requests for presentations at universities, conferences and companies.
“I’ll [still] be speaking about March Madness in June,” he marvels.
You're never going to be 100 percent correct. Those are really important lessons for students, in data analytics and life in general.
While at CU, from 1996 to 2001, Chartier was more of a football fan than basketball fan, despite the gleeful mania generated by Chauncey Billups (Soc ex’99) and the Buffs' 1997 NCAA tournament appearance.
He grew more interested in hoops after joining the faculty at Davidson, a 2,000-student liberal arts college in North Carolina that is the alma mater of 2014-15 NBA MVP Stephen Curry and the winner of last year’s regular season Atlantic 10 Conference title. There Chartier connected with basketball coach Bob McKillop, a believer in using inventive statistics to assess and manage performance. Chartier and the Cats Stats, a volunteer student group, have been supplying the team with voluminous data and analyses since 2013. (Davidson plays as the Wildcats.)
They analyze where players shoot from and how well, players’ scoring percentages when guarded versus unguarded and how certain lineups perform relative to others, for instance.
With 66 seconds to go in an Atlantic 10 conference quarterfinal game last year, and with Davidson trailing by a point, McKillop’s coaching squad picked a defensive scheme based on a Cats Stats analysis. As The New York Times reported at the time, the decision led to a possession, a score — and a win.
McKillop’s son Matt, an assistant coach, told the paper the Cats Stats’ work was a factor in “why we chose that play at the end.”
It was Chartier’s academic work that first brought him to the coaches’ notice. In 2008 he and a colleague were working on problems in predictive analysis, including ranking systems for Internet search engines. The rich and abundant data available for college basketball teams served as useful material for testing their algorithms and ultimately prompted their study of ways to predict teams’ performance against certain other teams with specific characteristics.
Eventually Chartier and interested students began applying these mathematical innovations to the NCAA tournament brackets.
“The nice thing is that in sports, generally there’s a lot of data that’s easily accessible,” he says. “But on the flip side, there’s skill and luck. You’re never going to be 100 percent correct. Those are really important lessons for students, in data analytics and life in general. Randomness is going to happen. If it’s random, you can’t predict it.”
Though Chartier teaches a variety of applied mathematics courses, his favorite is “Finite Math,” a lower-level course ordinarily taken by non-majors. In it he usually gives students a taste of bracketology-related math, which is primarily linear algebra and was the focus of Chartier’s graduate work with CU professor (now emeritus) Steve McCormick.
This type of math is commonly used in data mining, economic forecasting and developing computer graphics. In bracketology, it helps model teams’ relative strength.
“One of the reasons I love that course is that for most students, it is the last math class they will ever take,” says Chartier. “On the first day I tell them, ‘Someday many of you will sit at a table where a kid will ask why anyone should like math or science. Get one story in the next 16 weeks to tell them.’ Not all of them continue in math, but I love seeing them get to know just a little bit more why math is important.”
Chartier is no one-bracket pony. While at CU [which his parents, Myron (Hist’60) and Janet (A&S ex’61), also attended] he started integrating another of his interests into math instruction — mime.
He and wife Tanya both studied with the celebrated French mime Marcel Marceau. An administrator at the Boulder Public Library saw them perform a sketch involving math. The library had a National Science Foundation grant to support math-related performances, and invited them to perform there. Thus was born Mime-Matics, a side business in which the Chartiers explore mathematical concepts through mime at schools, colleges, conferences and other forums.
“You hear about STEM [Science, Technology, Engineering and Mathematics], but we like to talk about STEAM, which would be STEM with ‘arts’ added to it,” Chartier says.
In mid-January, with the college basketball season well underway and March Madness on the horizon, Chartier and his students were busy testing and tweaking their formulas and developing new methods — for identifying the “Cinderella teams” whose promise had been overlooked, for instance, and for varying aspects of their analysis from round to round of the tournament to account for the fact that some factors may become more or less significant as play progresses.
“We started a month ago,” he said in January. “March Madness is very real to us.”
During the tournament, Chartier is hunkered down in Davidson, watching a lot of games and taking a lot of calls from reporters.
One thing he isn’t doing is tracking the performance of his own bracket: He rarely fills one out.
“I am doing so much work and helping so many others,” he says, “I don’t get around to it.”
Opening illustration by Harry Campbell; photo courtesy Tim Chartier/Davidson College