4 Analytics Lessons From Professional Sports
Nate Silver, Michael Lewis, Mark Cuban and other sports luminaries share analytics advice you can apply to your business.
Want a career in data-driven decision-making? Nate Silver, the famously accurate predictor of the 2012 presidential election, advises neophytes to "run a fantasy baseball team or an NCAA Tournament pool" to learn applied statistics.
"Sports is a pure laboratory, and there are very good data sets available, especially in baseball and basketball," Silver said during a panel discussion on Friday at the Seventh Annual MIT Sloan Sports Analytics Conference in Boston. "There is luck involved in who wins and who loses, but you have good criteria for measuring success, so you can test ideas, make predictions and see how well it works."
More Software Insights
- Strengthen Organizational Agility with the Latest Advances in Case Management
- Rising to the challenge of RMORSA - leveraging existing capabilities and building new ones
- Don't Get Stuck on Your Virtualization Journey: Where to Focus Next
- When It Makes Sense to Move to Desktop Virtualization: Seven Key Indicators
Silver is the author of the FiveThirtyEight blog on NYTimes.com. But those who have him pegged only as a politico don't realize that he got his start in statistical analysis through sports, as the developer of the Player Empirical Comparison and Optimization Test Algorithm (PECOTA) system for forecasting the performance and development of Major League Baseball players. He sold that business to Baseball Prospectus in 2003 and continued to manage it until his political blog took off.
[ Want more on the use of analytics in sports? Read Analytics Drives Next Generation Of Moneyball In Sports. ]
It was fitting that the panel moderator was Michael Lewis, author of the seminal 2003 book Moneyball, which chronicled the pioneering use of data analytics in baseball by the Oakland Athletics. Moneyball also helped shine a light on the use of analytics in other sports, as championed by co-panelists Mark Cuban, owner of the NBA's Dallas Mavericks; Paraag Marathe, chief operating officer of the NFL's San Francisco 49ers; and Daryl Morey, head of basketball operations for the NBA's Houston Rockets.
What advice did these analytics pioneers have to share on numbers-driven decision-making? Here are four themes that apply no matter what industry you're in.
1. Stick With It.
Major League Baseball has been steeped in statistics for generations, but that doesn't mean teams, coaches and scouts were receptive to numbers-driven decision-making. Lewis recounted how shocked he was by how controversial Moneyball was when it came out and by how angry baseball traditionalists were about the book.
"They were angry because it was costing people their jobs," Lewis said. "Moneyball made the argument that the general manager, for example, might be relatively more important than the coach." Silver recalled that when he attended the winter MLB meetings in 2003 right after Moneyball was published, "it was like the Sharks versus the Jets" – that is, two rival gangs: one group embracing analytics and the other clinging to their experience, instinct and intuition. As numbers have proved their worth, however, acceptance has grown.
"The business has gradually realized that almost all your return on investment in baseball comes from drafting and developing young players who then get priced substantially above the market rate," Silver said. "The scouts use stats as a baseline now on which you improve. All these players have stat lines, and it helps you know that one player is a lot more likely to be successful than another." Another sign that the baseball business has embraced statistics is the fact that Sabremetrics -- stats compiled by the Society for American Baseball Research and used by analytics advocates -- are "light years ahead of where they were even five years ago," Silver said.
The lesson here is that perseverance pays off, and that proven stats will breed demand for more breakthroughs in data analysis.
2. Gather More Data.
In any field, you encounter diminishing returns as you add complexity to analytic models, Silver noted. Yet even in baseball, with its seemingly endless array of stats, "we're not at the point where the possibilities have been exhausted," he said. Relatively new scouting data sets have yet to be exploited.
"For example, we now know all about the break on a pitching prospect's curve ball, and we have years of that data now so we can go back and see how predictive" those characteristics might be, Silver said.
Silver was referring to video-captured data showing the coordinates of the pitcher, his arm, the ball and ball movement in multiple dimensions in every pitch during every game. Baseball is in the process of adding similar multidimensional fielding data that will help show, for example, the difference between players who truly hustle and sometimes get charged with errors versus error-free players who are playing it safe and allowing more hits per chance.
The NBA is working on similar player-tracking data, and at least 13 teams are installing the six special cameras required throughout a stadium to track the position of each player, the ball and each and every shot, rebound and steal during every second of every game. That setup obviously adds up to the collection of really big data, and professional basketball has just begun to discover what it can do with that information, said Morey, the Houston Rockets GM.