It’s March, and TIP is part of Duke University, so we’re focused on one thing right now: college basketball. It’s time to watch some of the best athletes in the world, debate whether coaches are using the right strategies, and have friendly arguments about which teams and players are better.
But to win an argument, you need data. For example, if you’re arguing with your friends about the GOAT—not the animal, but about which player is the Greatest of All Time—you need to provide some reasons for your pick. Is it Michael Jordan (unfortunately, for Duke fans) because he was such a great scorer? LeBron James because he does it all? Bill Russell because he won eleven NBA championships? Do you think people are forgetting about how dominant Diana Taurasi was in college and the WNBA?
Whatever the case, you need information. You need ways to measure the impact a player has. You can start with points, assists, rebounds—but which of those statistics is more important? How do you compare players from today with players from fifty years ago?
Things quickly get complicated. But this is where the growing field of sports analytics can help.
From TIPster to data guru
In the 1980s, Sean Forman was a TIPster living in Iowa. His dad was a football coach, and Sean would help compile statistics from the games. He used to look at sports encyclopedias to find which players led the league in various categories. He got really into fantasy baseball, and he would make spreadsheets to rate players so he could build the best team.
In the 2000s, Sean was in graduate school for mathematics. He was still into fantasy sports, and he thought a website of sports data would be much easier to use than the old encyclopedias that were stored in his dad’s house. So he built one.
“I really was interested in stats and wanted something that I could use,” Sean told us in an interview. “I thought it would be useful, and I did it part time for about six years while I was in graduate school in mathematics. Then I had a job as a math professor, and quit that after six years to start doing the site full time.”
That website started as baseball-reference.com. Sean has since built other websites for basketball, football, hockey, soccer, and more at sports-reference.com. According to a New York Times profile of Sean’s company, the websites received one billion views last year.
The websites are so popular because they provide all the information fans want to know about their favorite teams and players. They fuel the arguments that we love to have with our friends. And they are changing the way we watch sports.
“I’m sure all of your readers have their phones with them all the time, and if they have a question they just look it up,” Sean says. “Fans are becoming more interested because the Internet has allowed all the Cleveland Cavaliers fans to get together and argue [for example] whether Collin Sexton is any good.”
Teams making predictions
It’s not just fans who are interested in data, though. It’s teams, too.
“Teams are collecting data that can help figure out what’s going to happen in the future,” Sean tells us. “Say that J. J. Redick is able to shoot 50 percent on open threes. That’s a valuable piece of information because now that they’ve added Tobias Harris, he’s probably going to get more open threes, so his scoring average is probably going to go up.”
To make better predictions, you need better data. That’s why teams are investing a lot of money in more advanced statistics and analytics.
For example, most basketball fans know about shooting percentages, points scored, and other standard measurements. But those only tell us “70 or 80 percent of the context of what’s happening on the basketball court,” Sean says.
If one player scores eighteen points a game, but only plays for ten minutes, and another scores twenty points but plays for thirty minutes, which is better? Is it more helpful to score thirty points or to get thirty rebounds? To try to answer those questions, analysts are creating new statistics like win shares, which attempt to “divvy up credit for team success to the individuals on the team,” according to Sean’s website. Teams can use that information to make better decisions.
An engineering problem
In addition to the mathematics problems teams must solve to create advanced statistics, they must also solve engineering problems.
For example, say a player takes ten three-pointers and makes five them. Using a basic shooting percentage, the player looks pretty good—50 percent is a great three-point average. But what if you could determine that the player missed every shot she attempted when a defender was within five feet? You would know a lot more: she’s great at hitting open shots, but not very good and making them when she’s guarded.
But to do that, you would have to be able to measure how close the defender was every time she took a shot. The referees probably won’t let your assistant coach onto the floor with measuring tape during play, so what do you do?
The NBA solved the problem in 2013, when it installed a six-camera motion-tracking system called SportsVU in every arena. It tracks each player’s position on the court, how far he travels, how far he is from other players, how fast he’s moving, and more. Now, if you look up Kevin Durant’s shooting statistics on basketball-reference.com, you can see exactly where on the court he took each shot.
Teams can use this information to improve their game plans. And they are. In sports all across the world, players, coaches, and general managers are using these advances in statistics to get a leg up on their competition. And in the process, they’re proving that sports are as intellectual as they are physical.