If you’re a National Basketball Association (NBA) fan, then you have probably heard of Daryl Morey, who most recently spent nearly 15 years with the Houston Rockets, 13 as general manager, before becoming president of basketball operations for the Philadelphia 76ers in 2020.
If you’re a basketball fan and you also love data analytics, then you definitely know Daryl. Throughout his varied career leading NBA teams, he has used analytics to recruit talent and maximize player and team performance. Morey’s basketball strategy, which relies heavily on analytics, favors three-point shots and layups.
During the first week of May 2021, Wharton’s Adi Wyner, a professor of statistics and the faculty lead of the Wharton Sports Analytics and Business Initiative, (which also runs our Wharton Moneyball Academy summer program for high school students) met up with Morey online during the Wharton Annual Analytics Conference for a Q&A about how he uses analytics to create impact on and off the court.
Below is an edited version of their conversation (The full video interview is available below).
Abraham (Adi) Wyner: The shot profile [3-point shots versus 2-point shots] was a market inefficiency that you were able to exploit very successfully… As a mathematician, I quickly calculate that three is 50% bigger than two and you would have to have an incredibly large gap in success probability to not make three points incredibly worthwhile. And yet, it was a long haul from when the three-point shot was introduced until we have Moreyball. What took so long [to adopt the three-point shot in basketball]?
Daryl Morey: I want to emphasize how obvious it is if you look at the data. That’s why it’s such a good question you’re leading off with. If you look at the data, which obviously I did (and others) when I first got into the league in 2001, it’s overwhelming. [Shots made eight feet from the basket go in almost identically percentage-wise as shots made from 24 feet]…Why did it take so long? The game was introduced without a three-point shot, and when it was added, all the game had been built around how to get the ball close to the basket and get open shots close to the basket. So, you can imagine years and years of skill and coaching and plays and players who played the game and how they trained themselves all building how to get shots close to the basket and open. That was the entire framing and mindset of the league. Along comes someone who’s putting a weird line on and saying you get three [points] now. Everyone viewed it as a gimmick, and nobody liked it.
Wyner: You really think it wasn’t until the data showed how clearly it worked? Or was it just the sense that people started to see other teams doing it and succeed?
Morey: The data showing it worked didn’t do it. [That only convinced] a few teams… The only reason we were able to do it in Houston and led the league in three-pointers for 15 years was [because of] a leader Leslie Alexander who owned the team and came from Wall Street. He had owned the team for about 10 years and was a smart guy and was basically like, what we’re doing doesn’t work. We need to do something different…Leslie was the first one to say: ‘I’m not going to hire someone who looks at data as the second in charge; I’m going to hire them as the person in charge.’
“The NBA is an extremely complex sport. You need to respect that complexity and know there’s a lot more to it than any sort of simple data.” — Daryl Morey, Head of Basketball Operations, Philadelphia 76ers
Wyner: [Does the data show] another market inefficiency in basketball that should be tackled?
Morey: There is one and it’s very relevant, but it’s nowhere near as big. Each innovation is a smaller and smaller edge that you’re eking out. One has been talked about quite a bit and the league hasn’t adopted it. There’s a strong cultural approach to how the game has always been played that when a shot goes up, it’s advantageous to make sure that you are setting your defense and making sure that the other team can’t get an easy runout or easy transition basket, compared to if the shot goes up, you could also choose to try to get the rebound. It turns out again that the math is pretty overwhelming that more players should go get the rebound and you should accept some negatives on the backend. This has been studied for years now and the league is still quite a bit behind that innovation. It’s taking just as long as the three-point ones took to make that change…It’s actually a way more complex calculation and because of the complexity, it’s taken longer for the league.
Wyner: How do you get the coaches to accept these analytically driven innovations? If these things really come out of advanced analytics, is that a real obstacle to getting it accepted because the coaches, the actual players, they don’t quite understand it?
Morey: One thing that’s great about the NBA is that it’s an extremely complex sport. Anyone who is using data and says it’s easy and you should do X or Y and the data is overwhelming – there are certain things like threes versus twos where that’s true – but most of it is very complex. Take, for example, one ‘made’ jump shot and the factors that go into that. Let’s say the ball goes in. Did the ball go in because of the training the player put in to improve his shooting, or how good of a shooter they are just coming into the pros? Was the defense bad? Was the pass to him good? Was the pick-and-roll timed correctly to create the advantage to create the pass? Even one simple play in basketball is actually extraordinarily complex, especially relative to a sport like baseball. You need to respect that complexity and know there’s a lot more to it than any sort of simple data. At the NBA level, coaches are extremely good and extremely hard-working. I’ve found that it’s quite easy to walk them through, here are some options and here is why there’s an advantage. You can get to a good result.
All of these things we’re talking about, like do you go for offensive rebounds or do you get back, those are small potatoes relative to the importance of players on your team and how good they are and how hard they play. You’ve got to understand the different levels of this and say, okay, even if we did this new idea perfectly well, that could mean a half a win to one-and-a-half wins a year. Getting the right players and getting them to play hard…that can mean 10 wins. As a coach, they have to always keep that in mind. They say, ‘I might lose this battle and not do something optimal here, but it’s keeping my best player happy.’
Wyner: How big are the analytics staffs in basketball across the league…and are there opportunities for working in analytics?
Morey: It’s grown fast. I always tell folks who are younger that you’ve got to catch whatever the next wave is. No team has less than one [data analyst]. Every team has at least one. I’ve seen staffs as big as 10 and others as small as one. If I had to guess the average around the league, you’re looking at three to five.
What does Adi Wyner mean when he says that Daryl Morey has exploited market inefficiencies in basketball?
Daryl Morey says, “Even one simple play in basketball is actually extraordinarily complex.” What implications does that have for data collection and analysis?
How do you feel about the use of data and analytics to improve player performance and game outcomes? Is this valuable, or does it take something away from the game? Let us know your thoughts in the comment section of this article.