Guaranteeing A Win: A Basketball Analytics Perspective
By Spike Eskin
PHILADELPHIA (CBS) – Vasu Kulkarni thought it was going to be easy. As the self-proclaimed biggest basketball fan in the world, playing basketball in India and playing for the University of Pennsylvania would be one in the same.
As it turns out, it wasn't so easy.
"I didn't realize that at 5'9" and 135 pounds that wasn't about to happen," Kulkarni said. "I was in for a rude awakening when I showed up at the Palestra."
Vasu Kulkarni is the founder of a sports analytics company called Krossover, which takes the film of college and high school basketball, football and lacrosse games and breaks it down into advanced analytics and no-nonsense video.
"I spent four years playing a lot of pickup, and taking a lot of protein shakes, and finally as a senior I was allowed to walk on," he said. Kulkarni was finally on the Penn JV basketball team.
"So I get through that season and I sort of saw as part of the basketball program how much effort went into preparation for game day. We had two, three coaches, they'd be scouting out opponents, they'd be watching a ton of film, putting together a power point deck, handing it out a couple of days before the game, handing it out to the team and saying 'ok, you need to study these tendencies for these players, this is the key to winning.' And it was just an incredibly manual and arcane process to me. And being probably the only computer guy to ever come through the basketball program, it was sort of like, there's got to be a better way than what they're doing right now."
Growing up a basketball fanatic, but without a basketball body, analytics was the best way to do what he loved for a living.
LISTEN: Spike Eskin interviews Vasu Kulkarni
"I always wanted to work in basketball. I figured if I got a 9-5 job I'd be terrible at it. It seemed like a no-brainer to try to mesh basketball, computer science and data with a real job," he said.
Advanced statistics made their way into the sports public consciousness first with baseball, but it's in basketball that the most recent waves are being made. Analytics have become a word used with much greater frequency in Philadelphia since the Sixers hire of Sam Hinkie as general manager and team president. Hinkie is known as a leader in the field, and a regular at the Sloan Conference at MIT, a meeting of the minds to discuss sports statistics.
Sports is just part of the big-data revolution. Kulkarni saw the connection while working in India during a summer he spent back home, rather than getting an internship in the US. He saw businesses using all sorts of data to make more informed decisions.
"Walmart, for example, was tracking every single item sold. What time it was, the demographic of the person who was buying it, how they were buying it. And then you could go back, and sort of run a search and look for 'let's find every white t-shirt that was sold in the Midwest between the hours of 5pm and 7pm.' And that was just absolutely fascinating to me. I think that also somewhat helped in crafting the product for Krossover. If you could do that kind of system for businesses, why wouldn't you want to run those searches using basketball data to come up with better decisions?"
While analytics are talked about in great lengths in professional sports, not every team buys in. Some of those teams may be paying the price.
"There's definitely some disparity, even at the top organizations. I mean you hear about teams like the Lakers and the Clippers not wanting to do any of this stuff. Maybe they're doing something behind the scenes that the public doesn't know about," he said. "But most teams that sort of have adopted statistics and analytics as a form of decision making to some extent have been pretty vocal about it. And the [Oklahoma City Thunder], the Houston Rockets, the Mavericks, those guys have been all-out talking about what they're doing. And then you hear about, guys like the Lakers, you look at them and you think they're just loading their roster with talent and they're willing to pay the luxury tax, and they think they're going to be competitive. If you look at this year, both the Lakers and the Clippers had teams on paper that on paper were sort of championship caliber teams that could make a deep run in the playoffs. And they both got their butts kicked. And it could be coincidental, there's more than just the data that's bad that results in a loss like that. However, it has been four or five years now since some of these other teams have really started to adopt analytics. And it could be coincidence, or it could just be that the time has come that these younger teams who are making better decisions are beating these other teams that historically have been going on a manual methodology, a gut feeling, if you will. And so now these teams are starting to beat them."
Part of the split within basketball and baseball, between the stat guys and the non-stat guys (for lack of better words) is a generation gap. That generation gap, along with the idea of doing things differently, results in resistance.
"There's always some percentage of these coaches who are old school guys. They believe in their gut feeling and how they've been doing things for decades now," he said. "You know, it could be a combination of, they're just not numbers guys to begin with. These are basketball coaches. They're good at what they do, but they just don't know what to do with all these numbers. And so that fear invariably results in, "ah, I don't need that stuff. I know what I'm doing when it comes to coaching. I don't need that stuff. " And then you start to see the younger guys, the guys that are in my generation who are now becoming assistants, and they've grown up with iPhones and iPads the last several years. And they had computers growing up. And so for them, sort of using tools like this is common. And they're, as far as my opinion, as they get into head coaching jobs, there will be a much greater shift in the amount of teams who are using these sorts of tools."
The tools that Krossover provides are more data, and more video, but delivered in a cleaner, more efficient package. Clients upload game tape to Krossover's website, and it is then analyzed and sent back to the team with a package of statistics, and the ability to easily click on different aspects of the games and quickly watch video of certain plays. It's more effective, and more efficient than watching an entire game tape, or spending hours with pens, paper, and spreadsheets tracking the data yourself.
"[Coaches] need to come to the realization that no matter how many guys you have on staff, at the end of the day, a computer program is going to beat them, especially this sort of work. You know, I'm not saying that we're going to give you a better decision than your coaches are going to make. But I'm saying when it comes to data collection, when it comes to putting all of this together in a form that is easily digestible, it's much better to use a computer program. At least it's more efficient and cheaper to use a computer program, than it is to use humans."
The video portion of what Krossover does is often times the most impressive to players and coaches.
"Numbers are great, numbers tell a story, but at the end of the day, the human mind likes to see things visually," he said. "And so, I think yo have to convert those numbers into graphs, and then you have to move from graphs into actually watching video. Again, with a very fluid sport like basketball where there are so many things happening, you know, with baseball there isn't much to watch, in the form of video. As a batter you can watch your at-bats against certain pitchers When it comes to basketball, there are a million things you can look at. How do players like to play on the left side of the floor and the right side of the floor? There's just no substitute to watching film, and I think that goes back to one of the things you asked me. When it comes to these old-school guys, how are they reacting to numbers?
Within the information that the team gets, is information they have never seen, and perhaps never even considered.
"We like to look at location data. Here's a simple stat that you don't really hear about anywhere that we track is assist distribution by location. So we try to tell a coach where his point guard likes to pass the ball. Is that at the rim, in the paint, in the mid-range, and beyond the arc. And we show them how many assists are coming in that region for each of their players," he said.
The advanced statistic that has gotten the most publicity in a mainstream sense is player and team efficiency. Instead of asking how many points a team scored or allowed, which can often times be result of pace rather than efficiency, the number of points per possession scored and given up is used. Still, it doesn't pain the entire picture.
"It's never just one stat. Anyone who kind of says 'oh, I just need to look at these one or two stats and that's the difference between a win and a loss,' that's never the case. It's always trying to look at a number of things, put together in context, and it's about what you're trying to achieve. You know, are you trying to self-scout, and make an improvement about how your team is playing? Are you trying to make an opponent scouting decision? Are you making a personnel decision on who to draft or who to play? Depending on what your end goal is, yo have to look at a number of different stats."
Krossover even tries to quantify things that were previously thought of as unquantifiable; like hustle, and sports IQ.
"A lot of coaches that go by this gut feel, they like to talk about hustle. They like to talk about these intangibles. Sometimes that's their answer to statistics, is 'oh well you can't really measure the intangibles. And that's what I look at when I'm talking about my team.' And we try to look at some of those things as well. So we measure deflections. Deflections are a good measure of how active your hands are, and how active your team is as a whole on defense. So if you can measure deflections, winning 50/50 balls on the floor, steals and rebounds, now you have possibly to put some sort of weighted average on these things and come up with a metric around hustle, which is one of those intangibles that coaches really like to see."
"Something that we've been actually working on for the last six months is a way to measure sports IQ. It always frustrated me when I'm watching a basketball game and a hear a commentator go, 'oh, he's a high IQ guy." What the heck does that even mean?"
So they went about developing sIQ, which is an app that operates as a game, and was debuted at the Sloan conference this year, where it was tried out by guys like Mavericks owner Mark Cuban, Hinkie, and Rockets GM Daryl Morey. The game plays the first few seconds of a basketball play, and then freezes the video. It gives the user the result of the play, and the user has to decide, as quickly as possible, how that result happened.
"We've tested hundreds of people at this point. We've tested pro athletes, we've tested pick-up basketball players like myself. We've tested front office guys in the NBA, and invariably when you administer this test, not only are we measuring whether you got the question right or wrong, but we're also measuring your response time down to the millisecond. And what we've found is guys who have spent their entire life around basketball, guys that know everything about basketball, a good example is a GM. They will not do very well at this test compared to an elite athlete. And it comes down to this idea of something called perceptual ability and perceptual learning. This idea that athletes that train at a certain level, that train every single day, their brains get wired in a certain way that they're kind of making reactions and not decisions. And so a guy like me, I'm making a decision when I look at that clip, and I'm deciding given the spacing, and given where these players are, I think this is what's going to happen. More often than not, I'm probably right. But it probably takes me four or five seconds to answer a question. Whereas an elite athlete, their response time is under one second and they're getting most of these questions right," he said.
So how did Hinkie do?
"He got to play the game, I don't remember how he did," he said. Adding that he thinks Hinkie did well, but not to the level of an elite athlete (Hinkie scored 40th out of 85 people who took the test at Sloan).
As for Sam Hinkie, the hire in Philadelphia has been as polarizing as the topic of analytics. Kulkarni, for one, things it will work out.
"He's got a great background, coming from Houston I think. He seems to have followed the analytical path, he's been a consultant, he's got an MBA from Stanford and he's done some of those things, always been a numbers guy. I think when you look at the new ownership of the Sixers, those are also all Wharton, Harvard guys. They're Wall Street investment, big time private equity investment banking guys. So they certainly understand the value of data and numbers and all that. So I think it's going to be a great fit."
At the end of the day, Kulkarni, Hinkie, anyone who is involved in advanced statistics is trying to do one thing and one thing only: win.
"The holy grail is, 'can you guarantee a win?' Right, I mean that is what this is all going towards. Can you guarantee a win? I mean, who knows. I would say, as a sports lover, hopefully no one's ever going to figure that out. Let the innocence of sports remain and let us all love sports so that there is no way to guarantee a win unless sports is no longer fun," he said. Perhaps it would be less fun, but something tells you Sixers fans would take it at this point.
"But that's really what these guys are trying to do. They're trying to create a formula for winning. And so they're getting the data, but nobody has really figured out how to take that massive amount of data and come up with the 10, 20, 30 things that you have to look at to create a formula to guarantee a W."
"These data guys are going to lose their jobs if they can't figure out a brand new way of looking at the game," he said. "And so the fact that they're being brought in, at such high positions, and they're being paid a lot of money, I mean GMs make millions of dollars, it means in my opinion that they're going to have to look at the game in a very very different way, and come up with something new and novel that gives them a reason to be around in a position of power."
Visit Krossover to find out more about their services.
Follow Vasu Kulkarni on Twitter @VasuKulkarni
Follow Spike Eskin on Twitter @SpikeEskin
for more features.