Toronto Raptors v Golden State Warriors

Most basketball fans knows about the statistical/analytical revolution happening in the NBA right now, similar to what baseball went through years earlier that inspired “Moneyball” – both the term and the film. The advanced metrics unearthed during this revolution have helped us break down and understand both games better than ever, regardless of the traditionalist criticisms, and the NBA recently added all of the advanced stats you can ask for to its own website.

The NBA may actually be surpassing Major League Baseball in terms of how far this revolution will be taken thanks to “SportVU,” a program currently used by half of the Association’s teams which records every single second of action on the floor “and spits it back at its front-office keepers as a byzantine series of geometric coordinates,” writes Zach Lowe of Grantland.

Lowe’s fascinating feature on SportVU should be read by all, especially basketball savvy Raptors fans who want the organization to immerse itself in stats and analytics, as one of the main points the in-depth feature makes is that “The NBA is undergoing an analytical transformation, and the Raptors are one of the teams at the forefront.”

Sandwiched in the intriguing Grantland post which focuses on the Raptors are a few talking points for Raptors fans…

- The system is so incredibly in-depth and advanced that in addition to showing what every player on the floor is doing and where they are on every possession, it also shows the Raptors’ staff what players should be doing and where players should be moving with “ghost” icons (check out the clear circles with the same numbers as the white/Raptors circles in this video). Lowe writes that “the ghost players are consistently more aggressive on help defense than the real Toronto players,” but it is also pointed out that it’s tough to cover all of the ground the ghost says you should be covering unless your name is LeBron James.

- I love this quote from Raptors’ technical director of analytics Keith Boyarsky – “Anybody who is going to pooh-pooh this kind of analysis will say things like, ‘You can’t measure defense, because it’s all about the guy who doesn’t help or rotate.’ That it’s about what you can’t measure. But that’s exactly what we’re measuring.”

- Brian Kopp, the executive vice-president of STATS, says the Raptors “go a step beyond” what most other teams are doing and also states that “their visualizations are the best I’ve seen.”

- While Bryan Colangelo and his staff sound like they are fully on board the SportVU and analytics bandwagon, the coaches and Dwane Casey in particular don’t seem nearly as ready to rely on it. How about this quote from Casey on SportVU: “It’s a good backup for what your eyes see. It may also shed light on something else, but you can’t make all your decisions based on it, and it can’t measure heart, and chemistry, and personality.”

For people like me who thought Casey should be given more rope in Toronto than Colangelo at this point, Lowe’s feature certainly gives us something to think about. SportVU, advanced stats and analytics are the way of the future and it is encouraging to hear that the organization seems to agree. As for Casey, I completely understand why some more old school coaches may not be ready to let everything ride on numbers – in fact I agree with it – but I don’t think anybody’s trying to tell them to coach only by numbers. The point of an analytics-based revolution is to enhance traditional basketball evaluation and coaching, not to replace it all together.

Furthermore, are “heart,” “chemistry” and “personality” the deciding factors in Aaron Gray getting valuable fourth quarter minutes over Jonas Valanciunas or Andrea Bargnani over Ed Davis earlier in the season? Because if so, then I suggest Casey and his staff start taking these new, more advanced measures more seriously than those overrated intangibles.

- Despite my frustrations with his puzzling substitution patterns this season, I still believe in Casey’s defensive approach and his passion for the game, and I really do hope that he’s the one capable of leading the Raptors into a successful future, but when you recall the reported organizational split between Calderon and Lowry earlier this year and the apparent divide on the full value of analytics, the question has to be asked – is he the right guy for the job given the analytics-based direction the team is obviously going in?

- Here are a couple more passages from Lowe to consider in regards to the possible divide – “part of the team’s job is to sell a sometimes skeptical coaching staff on the value of all these new numbers and computer programs, says Alex Rucker, the Raptors’ director of analytics.” Later, when discussing a theory about jacking more three-pointers that the analytics team holds, Rucker is quoted as saying “That’s a conversation we’ve had with our coaching staff, and let’s just say they don’t support that approach.” For the record, Lowe himself downplays the divide later in the post and to be honest, I don’t think it’s at all isolated to Toronto, as most coaches around the league and around pro sports in general are probably hesitant to fully accept analytics.

- Given the potential for higher reward, three-pointers are highly valued by stats guys, and in general, advanced stats tell us that the best offence is one that gets to the free throw line regularly while knocking down threes at an above-average clip. Hitting more threes also depends on attempting more long-range shots, and one of the interesting tidbits from Lowe’s piece is when he writes that on a specific Raptors’ possession the did not end up with a three-point attempt, “the analytics team would have liked Gay, a below-average career 3-point shooter really struggling this season, to jack up a contested 3 at this moment — with about six seconds left on the shot clock.”

What I find so interesting about this is that we all rip on Rudy for taking too many threes when he’s not a consistent three-point shooter, which would seem to contradict a basic set of analysis, while the actual analytics people prefer he takes more! Maybe Gay’s frustrating decision to take what we consider to be “too many” three-point attempts is actually coming from the analytics team. If you’re a stats advocate like I am and you read through Lowe’s post, you may come to the realization that it’s time to stop criticizing Gay for jacking all of those threes, though our criticism of his long-twos should remain very much alive. In any event, whether it’s smart for Gay to be attempting more threes or not, Rudy has to improve his stroke from behind the arc, because one of him or DeMar DeRozan becoming even average three-point shooters could do wonders for this offence.

- Overall, Lowe’s fascinating Grantland post should be seen as encouraging by Raptors fans looking for their team to gain any advantage they can, and perhaps a few years down the line we’ll really start to see this investment in analytics pay off with sustained success on the court. But by reading through the column multiple times, the biggest obstacle still appears to be getting coaches and players to fully buy into what the analytics tell us, and I remain skeptical that you can teach so many old dogs so many new tricks.

Whether it’s the Raptors or a whole slew of other teams, we won’t start to see this investment in analytics truly pay off until the old school coaches, players and executives embrace the revolution, or perhaps more realistically, until those decision makers are slowly but surely replaced over the years by more forward thinking basketball minds.

With analytics at the forefront, the Raptors have been given a light to guide themselves out of the tunnel. How effectively they put that guiding light to use (and that includes how well Colangelo himself actually utilizes the analytics he has invested in) will go a long way in determining how quickly they emerge from the darkness that is mediocrity.

Comments (27)

  1. That was a great article by Lowe. It’s definitely encouraging that the Raps are invested in this analysis, but it’s not yet clear how they’re incorporating this into building a team identity. If you want to shoot more 3s, probably best not to invest so much into DeMar, Rudy and Fields. Maybe you want your best 3 point shooter (Lowry) shooting more, not less.

    Anyway, I’ve always complained about the lack of vision guiding the development of this team under BC. From offence only + Triano to defence + Casey, no continuity with players and no identifiable or at least plausible core group to build around. If this analysis can help solidify a vision based on reality, that’s a good thing. I just don’t believe BC is capable.

    • While I agree that the acquisition of Gay seems to go against the analytics, paying or trading for Fields, Lowry and Amir is probably at least partly because of analytics

  2. That’s interesting that Casey doesn’t want to use it as much. I thought he was a coach who relied on advanced stats and analytics a lot especially for his lineups.

    Also funny that Colangelo is apparently on board with all of this analytics stuff. Seems like most, if not all of his moves would go against any advice that advanced stats and analytics would give him.

  3. It seems to me that the tail end of the piece speaks less about how a team should only rely on the 3 and more about the value of the 3 compared to all the long twos that Gay and DeRozan likes to take. If you notice where they talk about how the Heat has solved the NBA offence, their offence revolves around only taking short twos and corner 3′s where you have wade and lebron often starting from the top of the key, driving to the basket sucking in the defence with their sharpshooters camping in the corners.

    Also, while I agree with Alex Rucker that the corner 3 is the most valuable shot in the game, he uses a poor example (a sub 35% 3-point shooter taking an open shot is MUCH more different than a sub 35% 3-point shooter jacking a contested shot). Again, I point to the Heat as well as Dwight Howard’s Magic, where their offensive schemes are predicated upon 5-feet and open 24-feet corner 3′s.

  4. This doesn’t mean much if they have a coaching staff that is not really open to giving credence to any of this data along with a front office that seems to pick and choose when they want to take it into consideration.

    What a clusterfuck of an organization.

    • Don’t hold back PBI, let us know how you REALLY feel about this organization

    • Bang on, man. Look at the teams W-L record. How much are they utilizing analytics? Very little I bet. If they did, I doubt they would have gone with a derozan-gay wing pairing. The only acquisition that was analytics driven was lowry’s. This is probably an attempt by the front office (BC) to save face and get some good PR im guessing. In the end, it means squat if the team sucks as much as it does currently and casey is the head coach who doesnt beleive in it.

  5. My favourite quote about stats: “Like people, if you torture numbers enough they’ll tell you what you want to hear”.

    But anyway, I think it would be interesting to see how playing exactly to the states would play out in practice as a lack of balance in offence might lead to intangible results. I mean assume Lowry is shooting 42% from 3 right now and taking 5 a game. That doesn’t mean that if you tell him to take 10 a game he still shoots 42%. Maybe 5 a game are the only good looks he is getting, maybe trying to get him 5 more good looks comes at the expense of someone elses offence or defenders guard him tighter out there and he starts making less of his orirginal 5.

    Taking this to the extreme the article mentions an offence where LeBron shot a 3 on EVERY possession wouldn’t be bad, but you have to think it would be a disaster in reality. It’s hilarious to consider though.

    • Ya but Gay is also shooting poorly on long two points attempts and while I’m sure what the Raps team is doing is alot more complicated this, some simple math.

      Based on basketball reference, this season Gay has attempted 390 two points from 10+ fee from the basket, on which he’s shooting 37.2%. Therefore, from this range, you’d expect Gay to score 0.743 points per possession, which is awful.

      From three, Gay has attempted 206 shots shooting 28.6%. This results in an expected points per possession of 0.858, which is somewhat less awful.

      Despite being an awful three point shooter, the Raps would be better off if Gay substituted a good portion of his mid-long rage twos to threes, especially if he could take more from the corner.

      • Yup, and on the basis of just those figures I agree, if he’s going to shoot from distance he should take threes instead of long twos.

        Where this gets interesting is considering if anything else would change as a result of him changing shooting habits. For example, does the fact the defense doesn’t know if he’ll shoot a 3 or 2 lead to him being more open when he does shoot 3s? meaning his 3 percentage would drop. Or does his 3 percentage then actually improve because he’s shooting more of them, therefore getting more practice and being in more of a rhythm for that particular shot in a game. Or maybe it stays the same and he just shoots more at 28.6%. Can the stat analysis tell us which?

    • That line from Rucker on Lebron shooting a 3 on every possession of course isn’t mean to be taken literally. He’s just making a mathematical argument where all else being equal (and it’s not), the percentages show this would be a more efficient offence. But even the stats geeks understand what Casey says in that article, in that there are chemistry issues etc. that also need to be considered (e.g. how happy are your bigs going to be if they never get looks). Also if you want to take higher quality 3s, it’s better to mix up your offence so it’s not so predictable. A contested 3 may be better than a long 2 for many players, but an open 3 is better at the end of the day.

      • I know, but it poses a very interesting argument as to where the extreme mathematical arguments such as that start becoming practical in reality. What if Lebron Shot a 3 on every 2nd possession or every 3rd? at some point it becomes valuable to the offence, it’s just finding the sweet spot.

        Also the point about all else being equal is really important because it wouldn’t be n this case, and I’d argue that for significant change based it stats it never will be, so predicting the outcome becomes more than maths.

  6. I wonder why Colangelo was never quoted in the article. Is it because he didn’t want to seem like he was disagreeing with Casey? I think Casey and colangelo are probably a lot closer with their philosophy on defence as opposed to offence.

    Hearing that Rudy gay should be taking more contested 3s reminds me of the football coaches who say that you should never punt.

    • I assumed that was based on his career % of 339 and not this season % .247. If his career pctg was the same as this season would they have said the same thing?

  7. Whether or not teams become reliant on these advanced statistics is besides the point; there is next to no parity in the NBA, and while as fun as teams like the Miami Heat are to watch, true fans want to see their teams have some glimmer of hope. I know your point wasn’t that baseball and basketball are similar but in the case of adv. stats in the NBA; individuals are much more important. Therefore I wouldn’t be as quick to be so reliant on them. Interesting article though.

    • I agree with your reply AJ. If superteams, teams like the Heat who stockpile stars, are the way to win a title, then I don’t see too much of a point of leaning too heavily on an analytical paradigm to make roster moves and coaching decisions. Of course the Spurs are perhaps an exception to this, but then again that’s premised on the idea that Duncan, Parker and Ginobli aren’t quite in the same stellar constellation as James, Bosh, and Wade.

      The funny thing about Moneyball influencing Basketball analytics is that the team at the forefront of it all, the Oakland Athletics, won jack shit with it. Not a single World Series. It was designed to make a small market team competitive.

      I’m not anti-analytics, its just that I think that believing analytics to be the be all and end all of basketball analysis is a somewhat flawed position. Without a strong conceptual understanding of the game – knowing what model works and what doesn’t – and knowing the strengths and weaknesses of your roster and how they relate to exxecuting that model, no amount of analytics will help you assemble an NBA championship team.

      And i don’t think its an old school versus new school question as Jerry West won all those championships with Lakers pre analytics, and I would venture he’s about as old school as it comes. The model any GM is working with is more important than any analytical tool as no amount of finessing analytical data will make a bad model work. It’s the Indian, not the arrow, who makes a good marksman.

      So, getting back to the Raptors, its not that I think Colangelo is purposely avoiding analytics because he is too old school, I just don’t think the model Colangelo is working with, if he even has one, works. And, no amount of analytical finesse and SportVU can correct and compensate for a poorly constructed roster. It’s just not possible.

      • Lowe also mentions, in this article and others, that the Heat play an analytical style. They’re near or at the top of the league in attempts at the rim and threes from the corner. Does their skill and athleticism help them do this? absolutely but the fact is they use what they have to play in the most efficient manner instead of relying on their athleticism and doing whatever.

        Other teams play with variants of this combination and are achieving success, such as Denver who almost do nothing but take shots from the paint while mixing in some threes.

        Also, To disparage the Oakland Athletics, I don’t agree with either. To slag on one team for not winning everything while being extremely competitive with well below average payroll and not slam all the above average payroll teams that don’t win is disingenuous.

        • Okay, disparaging the Athletics was unfair. Still, keeping the Athletics in mind, mining databases for nuanced data can only take you so far. You still need high ceiling talent to win.

          Still, my point stands. Any set of statistical data can become worthless if you are trying to graft it onto a model or paradigm that simply doesn’t work. If Colangelo’s model is to have an offense driven style of basketball, then why does he have some pretty inefficient players on it? If it really was true that Casey would rather have kept Calderon and traded Lowry, than it might imply that he probably understands the team better than Colangelo does because a pass first point guard can help bump up the relative shooting percentage by providing at least 8 or so easy looks a game.

          Again, I’m not anti-analytics, its just that it is a tool, like any tool, that is only as useful as the person using them.

  8. Here’s the thing, it all goes back to talent and player actually making he shot, because you can have advanced stats (which I think could revolutionize the game) telling the raptors to shoot more threes but at the end of the day you need players to knock it down. One thing I notice about this team now is that there in search for an identity. While last year the defense was great, this year it’s been bad and interesting thing is that there’s more talented players. So my question is that is Casey the right coach and is bc the right gm for us? I really think we need a new coach who has playoff experience and good coaching record because at this point I wouldn’t want to risk having a coach who is experiencing being a head coach (triano) and a coach that has never been to the post season as a head coach (dwane casey). Just get a coach that you can count on because next year raps HAVE to make playoffs

    • The point the analytics guys are making is that they understand that Gay or DeMar are well-below avg. 3 point shooters but that it’s not particularly relevant when the majority of their shots come from mid-long range twos where they also shoot poorly. The extra point the team gets from a three creates a higher expected return on the team’s possession than those mid-long range 2s.

  9. Fun fact from NBA.com

    Amir Johnson is somehow a plus 162 for a team that has been outscored by 109 points this season.

    • That’s not surprising at all.

      ex. Amir is on… raps outscore opponents 10-5. Amir is +5. Amir subs off, opponent goes on 20-10 run.

      Amir is still +5, whereas the team has been outscored by 5!

    • Therefor moneyball dictates play Amir for 48 mins a game and raps never lose :D

  10. It’s worth noticing how much of a shadow Colangelo is in this story and how much ink Ed Stefanski gets. Not hard to read those tea leaves, really.

  11. It seems Colangelo wants to use advanced stats for the coaching staff, but not when judging his own players. Either that, or he’s simply extremely poor at interpreting the data.

    • Yeah, this is part of the problem. When BC talks about Lowry, he talks about how great the advanced stats are for him. That’s fine. But then he signs DeMar to this crazy extension and I’ve never heard anyone, BC included, make an argument that advanced stats/analysis are remotely kind to DeMar. No need to even mention Bargs here but it at least looks like that ship will be sailing soon enough. If this analysis matters, shouldn’t it matter for every player?

      The other thing about this article that kills me is that they are just starting to show the players some of this analysis. If the coaching staff doesn’t embrace it, what’s the point? You need to identify the system you want, based on what you understand, and then go out and find the coaches and players that will buy in and make that system work as well as possible. This is what everyone loves about the Spurs. With the Raps, I still don’t think they know what system they want.

  12. Teams like the Spurs/Heat/Nuggets/Rockets are honestly revolutionizing the game right now. The Thunder really, really need to get on board – they should have found a way to keep Harden and dump Perkins. They would have had the best roster in the league for playing analytics-informed small-ball.

    Agreed that it’s great that the Raptors are looking at the data, which begs the “So what’s the system?” question all the more.

    The ONLY two teams who play a traditional roster and are having success with it are the Pacers and Grizzlies.

Leave a Reply

Your email address will not be published. Required fields are marked *