by Bryan Betts ’13, Student Writer
In 2002, the Oakland A’s proved that a player’s worth isn’t always defined by his batting average.
The team won 20 consecutive games setting an American League record, despite having a payroll a fraction the size of big city franchises like the New York Yankees. Their success was largely due to signing players other teams did not want because of perceived defects based on outmoded thinking.
It’s this radically unconventional, mathematically-based approach to scouting, featured in the film Moneyball, that mathematics major Jordan Lyerly ’12 set out to analyze and improve.
Lyerly, a baseball player himself who has pitched for and coached Furman’s Club Baseball team, approached mathematics professors John Harris and Kevin Hutson with the idea of doing sports analytics research based on the Moneyball concept.
“I didn’t know what we were going to do or how we were going to make it work,” Hutson said, “but I thought it was a great idea.”
Following a weekend in Boston at the MIT Sloan Sports Analytics Conference, which featured well-known sports figures and analysts such as ESPN columnist Bill Simmons, former NBA coach Jeff Van Gundy, and owner of the Dallas Mavericks Mark Cuban, they “began to understand what mattered to these folks—the kinds of questions for which they wanted answers,” said Harris.
The research began that summer. Lyerly, along with Will Decker ‘14, Rob Picardi ‘13, and Aaron Markham ‘12, decided to go with a suggestion from Davidson College professor Tim Chartier and study how to use different ranking systems to assess the value of MLB players and teams.
The idea guiding their research was that players should be ranked not only by how well they performed but also by whom they played. Thus, if a batter got a hit off a good pitcher, it should count more than if they got a hit off a bad pitcher. The same principle went for pitchers facing good and bad batters.
Using an automated process called “web scraping,” the group pulled players’ 2010 and 2011 MLB season statistics from the web, and entered it all into a computer program. From there, they ran the data through formulas they developed to generate rankings for every batter and pitcher who had played professional baseball during those seasons.
Their work, however, failed to turn up any revelatory findings. They had the rankings, but they needed to figure out what to do with them. They needed to find the “misfit,” a player whose statistics belied his actual talent.
Later that fall, the summer portion of the research complete, Lyerly was watching SportsCenter and heard the analysts mention Doug Fister, a pitcher who had played poorly early in the year but phenomenally after being traded to another team late in the season.
Their research had ranked Fister as one of the top twenty pitchers in baseball despite having unremarkable statistics for most of his career. The reason? Fister had been pitching against exceptionally good hitters, which hurt his numbers. Now that he was with a different team, though, the competition was not as formidable and Fister could be seen for the great pitcher he was. Their research had recognized Fister’s talent, despite what his statistics said, and his trade to another team had proved it.
“All of a sudden I had this diamond in my hand because this is what I was trying to prove,” Lyerly said.
Even though he graduated Furman in June, Lyerly continued doing research with Hutson and Harris throughout the summer, and his sights are set on being a presenter, not just an attendee, at the place where it all began—the MIT Sports Analytics Conference in Boston this spring.
Sports analytics research has taken hold at Furman and spiraled off into a number of similar projects under the guidance of professors Harris and Hutson. Lyerly’s project was picked up by the Huffington Post, Furman professors worked with ESPN the Magazine on a football rankings project, and students have presented their research at regional conferences.