EDITOR’S NOTE: Before things get wholly playoff-tastic ’round these parts, for your infotainment, ruruland conducted an interview with Dean Oliver, a pioneer in sports analytics and the author of Basketball on Paper, the groundbreaking tome on basketball analytics. Oliver also spent five successful years working in the front office of the Denver Nuggets. He currently plies his trade at ESPN, where he led the team that developed Total QBR–a new method for evaluating quarterbacks–and is responsible for building sports analytics that are in use across the company. His work has been published in academic journals, highlighted in diverse popular publications and he appears in numerous televised segments.
Needless to say, we’re thrilled he agreed to chat with Nick about basketball, advanced analytics, the Knicks and other matters of great import.
Nick Ruland: I understand that you’ve either attended or been a panelist for each of the seven MIT Sloan Sports Analytics Conferences. What do you think is the most promising or interesting research?
Dean Oliver: Well, I think the player tracking and the ball tracking are the two most likely areas that will change the way we think about sports.I think there are a lot of steps involved to really turn that into information. Much of it is just data. There have been some research projects that have shed some light on it, but we aren’t even at the tip of the iceberg. There is a lot to be done to process that into useful information. But that’s where the promise is.
Nick: Did you read Zache Lowe’s story on how the Raptors use SportsVu? It talked about the Raptors use of isolations, which according to the video site Synergy, are low efficiency plays, but the Raptors use isolations to generate other efficient shots in the offense. What kind of affect do you think this kind of analysis has on NBA organizations?
Dean: It should have an effect. There is no doubt that it should. Of course there are a lot of ways to interpret the data that is there with analysis like that,and of course it depends on who is guarding you and your ability to run an isolation as well. A lot of teams put the ball in the hands of their best player and it limits the challenge of communication and coordination. Isolations in many way are easy plays for coaches to draw up, though there are variations of it, but it is giving a lot of power to your best player. That can be useful, but it can be predictable if you’re doing it a lot.
Nick: How did the Nuggets and/or you view isolation plays during your time in Denver?
Dean: Isolations force the player to make the right decision. There are guys who can make the decision with the right amount of time. Some of it is individual ability, but a lot of it is playing with a guy who understands where teammates are supposed to be and how the defense is supposed to react, like if the double team comes from the baseline or the wing. But defenses will mix those things up, especially in the playoffs, because you are playing the same guys seven times. In the regular season it’s harder to make adaptations. Teams tend to do things they are pretty good at. But in the playoffs, things change a lot, and players understand a lot better the other teams plays because they are not coming off back to back where you have to do ‘this or that’ against another team. They had to respond to Dewayne Wade last night and tonight it’s Melo. And they have to deal with the differences between those guys. The time to make adjustments and the ability to make adjustments is better in the playoffs, for both offense and defense. Then it becomes how well can you make adjustments. A defense will really force a player to potentially get out of his comfort zone.
Nick: So I guess playoff offenses that are varied in terms of ways they can attack a defense are better suited to making adjustments and being one step ahead of playoff defenses?
Dean: I don’t know that there is a general rule of thumb for offense and defense. It’s more the sophistication of the coach and the players executing that in terms of their ability to adapt. If there is a rule of thumb for offense or defense, I haven’t necessarily seen it. We all know the rule that defense wins championships. That just means in the regular season your offense better be pretty dang good, but in the playoffs you better be able to work hard and make those adjustments. That’s where you see that cliche come into play.
Nick: We have a lot of Wins Produced advocates on our site. According to their metrics, Ronnie Brewer is a more valuable and productive player than Carmelo this year in 1600 fewer minutes. Needless to say, while many find this notion highly implausible, Berri’s analysis has been source of quite a few lively debates. I’d love to hear your general take on having team stats being parceled out to individuals, and how useful that information is when it yields that kind of comparison?
Dean: Well, I don’t know, there are multiple ways of doing it, depends on how you are parceling out some of the defensive stuff correctly. By a lot of the metrics I look at, Melo has had one of his best seasons. Even on a per minute basis I don’t have Brewer better than ‘Melo. And one of the interesting things is you look at these, I mean consider the former Orlando Magic as they were constructed last year. Last year wasn’t even their best defensive season, they were dominant and with fairly similar personnel (from years past). You look at how their former players have been this first go-round in the league (outside of Orlando), Earl Clark, Ryan Anderson, Howard, Jason Richardson, and then just look at the guys that are on Orlando, and if you look at the sum of their parts last year and how good they were defensively, and again, they were not as good as they used to be, and I think every single one of them is worse this year (statistically.) So it’s not a straightforward thing. Defense is a cooperative, collaborative effort that involves the coach designing things and coordinated by the team. But it is the skills working together and the cooperation and coordination of those skills. So defense, you wish you could say could put a great defensive player on a team and he’d be great, but it’s not an additive process.
Nick: Is there a metric that picks up on those interaction effects that you see or use?
Dean: It’s not a simple metric, it’s a model of how you put people together. If you are looking for a single number it’s essentially impossible. It’s hinged upon communication and how skilled and how well it fits. It is a simulation of a team, not a single metric.
Nick: In your book, Basketball on Paper, you talk about a metric you developed called the Skills Curve, which shows that for every point of usage, a player adds .6 points to the individual per 100 possessions and .25 points to the team per 100 possessions. In regards to the Knicks, who have low usage players like Jason Kidd (who pretty much only takes spot up shots), Tyson Chandler, (the dive man who also limits his offense to alley-oops and offensive rebounds) and, on the other end of the spectrum, you have Melo, whose efficiency isn’t amazing. Where do you see guys who take a lot of shots playing with guys who can’t take many shots or choose not to?
Dean: It’s the system we are trying to develop, right? You can have single metrics but they won’t tell you how they’d perform in certain environments. We try to capture a little bit of that (usage/efficiency trade-off). And it is useful if you have those high efficiency low usage guys who can play with someone who has high usage and at least not terrible efficiency, and Melo is usually at least very solid at getting the ball in the basket, he is not as efficient as some of the others. And so it’s about that fit. You are creating the team. It is not the sum of its parts, it’s does that fit in there?
Nick: So it’s about the value of a player in a team context?
Dean: In my mind, yes, that is what you area really trying to capture. I think it is the best way to build a team. If we are trying to build a team, and we have parts that we generally want to keep, and the parts that come in that fit best with that, that is what is most important. There may be guys on another team and they are terrible, but they can come to you, I mean, Steve Novak has been on how many different teams, and he is in a situation where he can play off Melo and JR really well.
Nick: Are the analytics guys gaining more influence in the way organizations and coaches make decisions, obviously that trend-line is going up, but how do you think has changed from just three or four year ago?
Dean: Well, obviously I have been in the game for a long time, and its dramatically different than when I first entered, and over the last three or four years we have seen a rather significant boom in teams adopting it to some level. I don;t know the level to which things get accepted in the front office and with coaching. Frankly the Toronto story was somewhat a surprise. I didn’t know how much they were using it. I know Case, and he would always listen to the group, but you’d never know how much they were doing or anything like that. There are things beyond my knowledge, but I know from various people of course, that they are having more and more say. That doesn’t meant that they are getting their way, but that’s okay, because a lot of time it;s about the conversation and producing the right evidence. Sometimes the evidence is not convincing but it focuses the conversation so that you can get at the right information, whether it is quantitative, video, interview information with different coaches. And that is what you want, to make decisions more efficient, whether it is a single number or based upon some uncertain numbers focusing a conversation.
Nick: How much do you see the analytics side influencing scouts and how they do their job and player evaluation and what they look at? Do those two groups butt heads quite a bit or are they starting to come together?
Dean: At my time in Denver I thought it was a very productive relationship. I didn’t think we but heads in non-productive ways. There are ways you should butt heads, it’s productive. When you are trying to make decisions from a quantitative method, you are trying to bring in things related to the information you are already using. It’s productive because we are asking different questions. What you’re saying sounds opposite of what I’m saying, but maybe we just need to put them on the right playing surface. There is a translation from numbers to words, and there a lot of different ways to say things. Words important and good can be intermixed. They aren’t the same. That is the challenge, when things look opposed to each other when they are not.
Nick: Are their proprietary metrics/models/quantitative measures that organizations have that only they have and developed over the years?
Dean: I would say a significant number of teams, investing in more than one person, have proprietary measures. I have plenty of things I did that I haven’t released and probably never will. Most teams have people doing a lot of that. Basketball on Paper was a framework, with a lot of details based upon decisions, and those details can change and lead to different questions. It was also written before there was a lto of play-by-play data out there, so the framework still works, but there is a lot more detail you pour into it now.
Nick: For the average fan, what is the best metric to understand the game that is publicly available?
Dean: When you talk about an average fan a lot of times they are concerned about who is best, but I don’t know, I’m always of mixed feelings with single value metrics. They can be useful. I tend to use win shares because of the public things that are out there it is most closely aligned with things I have done. That said, there aren’t significant difference between that and some of the other things out there. P.E.R. and some of the linear weights things are pretty different. I tend to stray away from those, but I also use them in different ways, a lot of times for setting a market value. but it depends.
Nick: And market value because guys that score are seen to be more valuable?
Dean: Well, linear weights is kind of what is in the NBA mindset for many years. And you are having the evolution of how the market gets set. Dave Berri did something that showed the value of just scoring is coming down. The marketplace is changing and you always want to keep track of it.
Nick: Looking at the playoffs, you have a serious contrast of styles between Denver and New York. We talked a little bit about defenses being able to adjust, play harder and execute in the playoffs. But what about a fastbreaking team that offensively uses a lot of dribble-hand offs and doesn’t have a go-to player, I know this might be a tired question, but do you think that makes a difference in the half-court offense for the playoffs?
Dean: Yeah, I think the Nuggets rely a lot on transition. And transition is one of those things that when you get into the playoffs, defenses find ways to take away transition offense. You have to have a half-court offense that you can rely upon. And the Nuggets, losing Gallo, that takes away a weapon even though they have some diverse weapons, but I think that is going to be a challenge. But it is amazing right now to see them survive without him and Ty. But it is a different, George is a very good coach, having seen what he can do with an injured team, and he makes the most of it. I thin kit is easier to do that in the regular season. I discount some of the wins that are happening now in terms of predictive ability for the playoffs because they have established, I think Ty being out, if he returns, also simultaneously benefits the Nuggets, because the teams preparing for him right now are watching tape without him and forget what he can do.
Nick: How flexible do you think Karl is in terms of changing an offense and defense around different kinds of teams?
Dean: He has been flexible. I mean he has the star system to lean on. Does he prefer the balanced, the guys he has now? Um, yeah. In the sense that they all share the ball constantly, that is just his philosophical thing. But he also enjoyed having Melo. where you have situations where you need to have one guy pick it up, he enjoyed that aspect, too. But I think he comes from a background where the balanced way is the ideal way to play.
Nick: Do you think there is some validity to the criticism of Karl’s offenses lacking structure?
Dean: I don’t know. I haven’t thought enough about it. I also think it needs to be recognized how strong the West has been. I mean, 50 wins and we squeeze into the playoffs as the eight seed? Whew. It has been ridiculously strong the lpast however many years, and I actually think it has weakened some this year. The question is a valid one, but I can’t say I have a an answer for it.
Nick: What about this Knicks team, do they remind you of the 2011 Mavericks?
Dean: I do not know if any comparison to the Mavericks is fair. I don’t know if anyone saw that coming. That was a team that wasn’t supposed to win against the Heat, against the Thunder, the Spurs. I don’t think they were supposed to win all those games. You only make that comparison if the Knicks are beating the teams they aren’t supposed to. In some ways you can say they are comparable, a style comparison, Dirk did a lot of things and was a tremendous mid-range guy who could kill you. I did analysis and his mid-range jumpers were the most important in the league that I have seen in many years. Melo has stretched his game to the 3-point line and that has helped. But with Kidd and Chandler there are some similarities and they are essentially doing the same thing they were before. Both guys continue to contribute, even if Kidd has had a bit off a drop off.
Nick: And with regards to mid-range shots. Are we starting to come back full circle where, now that all offenses go for those shots every time down, defenses are doing everything they can to defend them, and less to defend mid-range shots? Could mid-range shots eventually become valuable again?
Dean: Again, it comes down to team context. Layups and 3-pointers are good because they force teams to spread out. If those are your two threats, it’s hard to deal with them, because you have to help and recover and the recovery distance is much longer. So the mid-range shots in most situations are not the best shots. But if there are good defenses — the good defenses the Mavericks were able to exploit — they are good defenses customized and adapted to take away three point shots and layups, they give up the mid-range. So if you are that great player facing defenses that are used to giving the mid-range and you can make the open mid-range, there is a match-up advantage you have. It’s not necessarily a long-term advantage.
Nick: How valuable do you think the guys are today that come in low on the skills curve but are efficient in very particular areas, as opposed to say, guys with a lot of skills but aren’t great at anything? Where is the market moving in relation to both kinds of players?
Dean: It still comes down to the teams that are the best fits setting the market. As long as there is one team willing to bid for someone’s service because you fit them really well, that is where the market gets set. The other teams that don’t have a fit, well, for them the guy might be a terrible player. From an economics perspective it is not the most efficient because it’s not about just money you’re talking about fit.
Nick: What about (generally speaking) guys who lack the skills to take shots most players at their position can take — does that hurt a team in ways that aren’t quantified in publicly available metrics?
Dean: Yeah, certainly, having guys on the floor who only have on skill set in terms of shooting, whether it is Chandler or Novak, which are completely opposite skills, you have to have the system in there to get value out of them because they can definitely hurt you offensively. It’s amazing because they both have essentially one great skill — one finishes at the basket and one finishes a long way from the basket.
Nick: What do you think about the idea of the Kobe assist, or the guys like Rose and Iverson who penetrate a lot and seem to create more opportunities for offensive rebounds for their teammates?
Dean: It definitely exits. When I wrote the book and looked at offensive rebounding off of twos and threes, and the whole myth that was busted that threes are easier for offense to rebound. At that time we didn’t have all the detailed information. But it was like, shoot, twos, these layups, you draw help defense and what does help defense do? It makes weakside rebounding available. It is definitely there, it is an important effect you should try to account for. So guys like Iverson deserve a little more credit for what they did in their careers. A little more. It’s not going to change you from one tier to the next tier.
Nick: Would you like to get back into an NBA front office one day?
Dean: I certainly miss the competition and I enjoy the team I have here at ESPN but I miss the competition. Whether I go back to an NBA team or a football team, I have certainly been doing a lot of football, too. I think the football side has helped make me think of basketball in different ways. The situations aspect of football, football is very situational, and I think in basketball it’s easy to get a lot of things right because there are so many plays. But you can overlook situations, but by having to deal with situations all the time in football, it forces me to understand the importance of situational stuff on the basketball side.
Nick: Last one–Where would you rank Anthony as a player?
Dean: Top 20, and I’m pleased to see it. The ability is there. So many of my old guys from Denver are in New York. Obviously Ken and Marcus and J.R.
Nick: Mark Warkentien must have a lot of influence on that.
Dean: I don’t think it’s a pure coincidence (laughs)….