A look at how Demar Derozan and Rudy Gay are keeping the Raptor’s offense above water-level, but just barely so.

The Heat. The Spurs. The Grizzlies. The Raptors, The Thunder.

In order, those five teams featured the most successful five-man units last season (min 300 minutes played). I (Jason) kid you not; The Raptor’s quintet of Lowry-Derozan-Gay-Johnson-Valanciunas was more successful than the Thunder’s Westbrook-Sefolosha-Durant-Ibaka-Perkins (by Net Rating; 12.9 vs 12.3).

Now, this statistic is descriptive, meaning it tells us what happened, but we can’t be sure if it is predictive, meaning we don’t know if we can use it to anticipate future events. Certainly, it’s one number that is open to interpretation; is it just a mirage (data fluctuation), is it legitimate, or is it some of both? I don’t know. I’m not here to talk about that.

A closer look at the lineup reveals that the combination flourished on defense, while being merely average on offense. The defensive rating clocked in at a sterling 92.5, while their offensive rating checked in at a good, but not great 105. 4.

The offensive rating perplexed me. I thought the lineup would be more successful on offense. Kyle Lowry was an above average shooter, Amir Johnson was a great shooter and Jonas Valanciunas was one of the most efficient offensive players in the league (9th in NBA in TS% at 61.2% per Hoopdata). How could that lineup be merely average on offense?

Well, one obvious reason is the duo of Rudy Gay and Demar Derozan shot poorly (let’s leave out random fluctuation and personnel adjustment). You know the story already; Demar shot at a slightly below average efficiency and Gay shot a horrid 49.4 TS% (compared to 54% for league average SF).

Now, the problem isn’t necessarily that the wings shot poorly, it’s that they shot poorly AND often. Compare their usage rates (an estimation of how many possessions a player uses) to that of other wing combinations across the NBA (zero on the X-axis represents league average, standard deviation is just a measure of “how different”, see note 1 below):


Rudy+DD used over half (53%) of the Raptor’s offensive possessions, and they were only topped by the Knicks (Carmelo + JR Smith) and the Heat (some dudes named Lebron and Wade). Now, like I said earlier, disproportionate offensive loads is not necessarily a problem. It’s only a problem when they aren’t very efficient. Compare how Toronto’s duo stacked up to other wing combinations in the NBA:


Toronto’s wings shot a tonne, but they scored at a poor rate. Given that their combined usage was 53%, the offense, or at least as currently constructed, is pretty much doomed to be around average (and don’t tell me that DD’s usage will go down with Gay around, DD averaged MORE shot attempts per game with Gay in Toronto).

Well, that is unless the wings shoot better.

Demar’s biggest problem is that he shoots a lot of long-two’s and that he can’t sink three-pointers. Can he develop a three-pointer? Sure, it’s possible, but the historical prescient is against him.

On the other hand, Gay has a history of being an average shooter (career 52.5 TS%) and his percentages last season were certainly career lows, meaning he he’s probably due for some positive regression. However, the biggest unknown is how his vision-correction surgery will impact his shooting. Will it have any impact? If so, will it be positive or negative, and by how much?

If Gay and Derozan were to improve their jumpshooting, they would likely create some positive externalities for Johnson, Valanciunas and Lowry. As it currently stands, the startling lineup features almost no three-point shooting. If DD and Gay become threats from deep, this will give Valanciunas and Johnson more room to operate.

Certainly, there are unknowns, but one thing is known; if Gay and Demar shoot as much, and at the same rate they did last season, the Raptors’ offense is doomed to mediocrity. Here’s to hoping that something changes for the better.

Note 1: Only players who played over 25+ minutes per game and 30+ games were included. This left Sacramento, Minnesota, Utah and Phoenix with only one player, and thus they were eliminated. Several teams featured multiple players who fit the criteria. On that basis, I chose the two players who started the most games for their team. For example, Chicago had Luol Deng, Jimmy Butler and Marco Belinelli qualify. However, Belinelli started more games than Butler (27 vs 20), therefore Belinelli was chosen alongside Deng. The full list and data can be found here.

Thanks to Hoopdata, Basketball-Reference, ESPN, the NBA Geek and NBA Stats for the data used in this post.

Share this:
  • Casey Sherman

    They’ll both shoot better (and will have to shoot better; I’ll be the first to say that 49% TS is not going to get it done), and you can take that to the bank. By how much is the big question. One thing to consider is that volume shooters allow other players to take only the shots they want, which keeps their TS% high. Not a huge factor, just food for thought.

    • DDayLewis

      I don’t know if that’s a definite connection (the volume scorer thing). It’s not like DD and Gay are taking all the crap and leaving Jonas and Johnson with the best shots. For example, over a quarter of JV’s plays were post-ups, which I really fail to see how Gay/DD helps with (at least in a significant way).

      • Jonathan Mac Lean

        But if the clock runs down to 5 seconds, whose gonna take that shot? Probably DD or Gay. But I know, thats not the reason why their TS% is low, but it sitll counts. In situations like that, rather than chuck the ball they should drive the lane and look for the foul.

      • Casey Sherman

        Absolutely it is, the question is does it make a significant difference. Let’s say JV has the opportunity to run 10 post-ups in a game. Now, no two post-ups in the history of the nba are created equal–sometimes he’ll have good position, sometimes bad. Sometimes there’ll be a good defender, sometimes bad. Sometimes there’ll be lots of time on the clock, sometimes little. Sometimes there’ll be a help defender lurking, sometimes alone. Sometimes JV will be on a hot day, sometimes cold. And a million permutations of all these factors. Having players that can shoot a high volume at a somewhat respectable clip allows players like JV to call for the ball when he really likes the matchup, but not feel responsible if he passes it up and make them shoot, even if it puts them in a bad position.

        But I see what you’re saying–the volume scorer effect is probably pretty small. But that’s just post-ups, which are just one of the many ways to score in basketball

        • DDayLewis

          You’re assuming that more shots = less efficiency. The book is out on this issue. This will probably interest you:


          • bobloblaw

            Interesting article but I don’t like how it just bunches all players together. Those tables look at Ersan Ilyasova and Amare and say “ok, in 2012 Ersan started to take more shots and became more efficient, while Amare started to take less shots and became less efficient.” But what does that prove? Ersan improved his skills and Amare declined athletically.

            • bobloblaw

              PS: i meant the wages of wins article.

            • DDayLewis

              It’s a populational study, meaning it corrals a lot of data. The assumption is that the contextual factors all even out when they take into account a tonne of data.

              The one question I have for that study is that there are some biases to consider. Perhaps only the players who shoot really efficiently, and are capable of doing so over more shots, are given the opportunity to shoot more. Other players (think like a Kosta Kufos) are efficient, but coaches intuitively understand that more shots will not result in the same efficiencies for each shot, and so they’re not given more shots.

              Nevertheless, it’s a really cool study, IMO

              • bobloblaw

                Well, I think it’s a flawed assumption. If a player improves and can shoot better, he’ll often get more shots. All these examples would play right into the article author’s hands and show exactly what he wants to see. I don’t see how it would even out.

                Actually, I suspect that most of these shot fluctuations depend on simple player improvement or decline.

                I do think this article has value. It shows that NBA players usually get more shots when they deserve them. It also shows that shooters age better than non-shooters.

                But I don’t think we can make a general conclusion that a player won’t become less efficient if he’s asked to shoot more.

                • DDayLewis

                  Yeah, I too think that the conclusion is too strong, but it’s not a bad study. Certainly the book is out on the efficiency vs usage argument.

                • bobloblaw

                  Well, I disagree, I do think efficiency depends on usage, among other factors. I think this article is useless when it comes to that specific problem.

                  There are two very different questions:
                  1. How does efficiency and usage correlate historically in the NBA?
                  2. How does efficiency and usage correlate if skills level remains the same?

                  The article answers question no.1 but it thinks that it answers question no.2.

                  To look at no.2, we’d probably have to limit our data to “in prime” players, players who are not in rapid growth or rapid decline. But that’s also not a straightforward thing, there’d also be plenty of outliers, late bloomers, early decline. Just hopefully not as much.

                • DDayLewis

                  The article is certainly not useless in this debate, it specifically tries to quantify the debate with some data. What you have is a hypothesi, and people have tried to test that hypothesis. The article comes to the opposite conclusion, but that doesn’t make it useless.

                  2. The point of doing a population study is that for the most part, the outliers cancel out.

                • bobloblaw

                  There is a difference between outliers and wrong data set. If we wanted to know if 18-23 year olds like jazz, we would poll 18-23 year olds. We wouldn’t poll 15-40 year olds and hope that “outliers” would get cancelled out somehow. It’s a wrong data set.

                  Similarly, if the question is “how does efficiency and usage correlate, when skill level doesn’t change”, we need to be very careful with our data set and try to exclude players whose skill level did change. Most NBA players change skill wise, athleticism wise year to year. That’s not an outlier, and it won’t get cancelled out.

                  This isn’t an “I hate stats” argument. I’m just pointing out a glaring problem with that research. Stats are fine; but misusing them is a problem.

                  Now, my position is just a hypothesis, indeed. However, it’s based on common sense and observation. If a player has a big role, defense pays more attention and adjusts. Scoring is harder and takes more talent.

                  Now, maybe common sense is wrong in this case. But I don’t think that this particular research is any kind of proof. So for now, I’ll stick with that opinion. I think ‘book’ is clearly not out on that issue.

                • DDayLewis

                  Okay, so you don’t think the study can be applied to young developing players whose skillsets are still changing. That’s fine. It doesn’t make it the wrong dataset; it just means you can’t make certain inferences from that dataset.

                  It would incredibly subjective to parse out which players experienced a “skill change”. There’s no standard in which you could accurately and uniformly affix this onto the dataset.

                  Look, I’m no spokesperson for this article. I just think it’s worthy of everyone’s attention. Too often people just abjectly accept the notion that more shot attempts will lead to lower field goal percentage and that’s just simply not a fact.

                • bobloblaw

                  Why would the study be inapplicable to young players but applicable to older ones? That’s not my position at all. The study treats them all as one.

                  You think that the developing/declining players are “outliers,” but maybe “prime” players are the minority in that data set, maybe they are the ones who are misrepresented? That’s not clear at all. That’s my position.

                  Re limiting the data set, they could diversify it based on age groups, or years in the league, or minutes played. That’s not subjective. There have been multiple researches trying to answer the question of when does a “prime” start and finish in the NBA. It would be interesting to see that data combined with the data in the article. It would give us an idea of how efficiency and usage correlate when skill change is less of a factor.
                  Would that be perfect? Of course not. But as presently constructed, I think that research is misusing stats.

                  Like I said, the data in that research has other value. It kind of shows that shooters age better, for example. It kind of shows that coaches tend to reward improvement.

                • DDayLewis

                  You’d once again have to define “prime”. For example, if you define prime as ages 25-27, then yeah, they’d be the minority because it’s such a small window.

                  FWIW the WoW community has done research on player development by age. You can find that here: http://wagesofwins.com/2012/07/25/age-is-just-a-varible/

                  If you group by age/minutes/seasons in league, you’re assuming that skill is a function of age. That might be true to some extent, but it’s questionable.

                  Again, where are they “misusing” stats? Misusing stats would be me saying “Quincy Acy is a better player than Jonas Valanciunas because his PER is higher (15.9 to 15.6)”.

                • bobloblaw

                  Anyway, you see the problem with the article, right?

                  Right now you are kind of asking me to come up with the perfect methodology to fix that research — which is more than a little unfair. This is a big topic worthy of a PhD research, and it’s been tackled by many before, and we aren’t going to solve it here and now.

                  But as long as you understand my point, and why I think this article is way too simplistic and inconclusive, it’s all good with me. I wrote the last few posts only because you kept misunderstanding my position.

                  We can’t look at NBA players in a two year window and completely ignore development/decline aspect. Players develop, improve, get more minutes and more shots. Which is why those numbers in the article aren’t surprising at all, they confirm common sense. When players improve, NBA coaches give them bigger roles, and as long as they don’t fuck up (i.e. become inefficient) they keep those roles.
                  Does that mean that we can now give everyone bigger minutes and they will play well? No.
                  Unless someone does actual legit research and shows it, I don’t see a reason to believe that. That article is certainly very shaky and superficial.

                • DDayLewis

                  We both acknowledge that this isn’t nearly the be all and end all, and I share many of your criticisms, but why are you attacking the article with such phrases such as “misusing stats”, “shaky”, “superficial”, “bad study”? It concludes too strongly, but aside from that, it’s just an interesting case study.

                • bobloblaw

                  Maybe it’s a little strong, but you misunderstood my position several times, so I’m trying to be as definitive as I possibly can.

                  It’s an interesting case study as long as its results are used within reason. Does that data show that usage doesn’t affect efficiency? Or does it show that NBA players tend to get bigger roles when they become good enough to handle them? It would take a big leap of faith to claim the former.

                  The author of the article kind of wants to use that data on Harden that way but he chooses not to go there in the end. Which is good. If he used that data that way, my criticism would be too weak, not too strong.

  • themaven

    Yes. Yes they are. Doomed, that is. Which I think was part of the reason they hired Dwane Casey, wasn’t it? I mean, they tried to be Phoenix North already and it didn’t work out so well, right?

    Defence first is the right way for this team to go.

  • Marz

    Actually, I find this quite encouraging. I think the offense will improve because I expect:
    1) More touches for JV – I believe this will stem from a natural progression of his abilities
    2) More touches for Kyle Lowry – Contract year
    3) Less shooting from DD/Gay – New offensive assistant coach, plus statistical data from last year
    Whether or not their shooting improves, I don’t expect a big enough delta for it to matter.

    • DDayLewis

      Having DD and Gay use 50+ of all possessions leaves very little room for JV to flourish offensively. It’s a zero sum game (unless the Raptors increase their pace).

      • Marz

        My assumption was that they’d receive less possessions, thereby giving more to Jonas/Kyle.

  • Thimble

    One of the biggest issues with Gay and DeRozan was that they didn’t dish enough when they were double or even triple teamed. Perhaps with Val being groomed as Toronto’s future go to guy, we might see better ball distribution. I think Gay and DD’s numbers will improve if they shoot a higher percentage of their shots when they’re open.

  • BigV

    You also have to take in effect that dwayne caseys offense didnt help the cause.

  • Jonathan Mac Lean

    An increase in defense will help this team offensively. We have one of the most athletic starting lineups in the league (not THE most, but definitely high up there). If the defense improves we will see a lot more fast breaks, and our wings can run up there fast, resulting in fewer contested shots, which will obviously mean better TS%. Not every posession will be a fast break, but those few extra ones they can get per game will help (and our bench is pretty atheltic too with Ross, Acy, Buycks..). But once again, I think the best way for this clubs offense to improve is via defense.

    • DDayLewis

      This 5-man combo already had a DefRtg of 92.5, which is like Memphis/Indiana level of awesome on defense. You can’t really ask them to be any better than that.

      • Jonathan Mac Lean

        True, but considering they’ll have a full training camp together with a defense-oriented coach going back to his defensive ways, well, you get where im going. If they were on par with memphis/indiana on defense last season, then lets aim to be the best! And our current roster has 3 guys who were on both those teams last season..

        • DDayLewis

          Having a full training camp will be great, but how long will it take for them to learn Casey’s favourite play; iso-DD/Gay?

          Boom, roasted.

          In all seriousness, having the full training camp will help. Hopefully Nurse comes up with some better offensive schemes, or at least some more balanced lineups.

          And yeah, if they could be on par with Memphis/Indiana, the Raps will make the playoffs. The difference is that they both have elite defensive anchors (Hibbert. Gasol) and we have…Hansbrough? I dunno.

          • DDD

            Hopefully Nick Nurse can… Nurse our offense back into form. but seriously though i hope nurse and bayno take the reigns on the offensive side of the ball. in a recent interview, bayno emphasized floor spacing. i hope this is applied in the raps games, because that will mean more open lanes for our many athletic wings!

            • DDayLewis

              Right, but that would require Lowry/JV/Amir to stand farther from the basket and take shots from long-range. That’s not exactly a trade-off I want to make. I think the duo of DD and Gay needs to be broken up. Start Novak in DD’s place (I know they play different positions) and the offense will probably flourish. Defensively…not so much.

          • ItsAboutFun

            Your see-saw view of the “defense” of that unit is confusing.

            – In one post you’re saying “This 5-man combo already had a DefRtg of 92.5, which is like
            Memphis/Indiana level of awesome on defense. You can’t really ask them to be any better than that.”

            – Then the next post it’s “if they could be on par with Memphis/Indiana, the Raps will make the
            playoffs. The difference is that they both have elite defensive anchors (Hibbert. Gasol) and we have…Hansbrough? I dunno.”

            Which is it? You can’t ask for more, or you’re asking for more? Sorry, but I just find these two statements quite contradictory.

            • DDayLewis

              I see why you’re confused. The 92.5 they posted last season is descriptive, meaning it recounts history, but it’s not necessarily predictive, meaning it can predict future events in this regard.

              I’m not quite willing to state that this group is actually that good on defense. That’s why I said if they could be on par.

              They’ve done well. I just don’t know if that will continue.

              • Jonathan Mac Lean

                Thats what I wonder too, but thats also why I’m looking forward to the training camp results. We shall see starting Oct 30… And what you said about having Hibbert and Gasol as anchors is true, but they managed to be on par with them without that anchor, and that says something. Valanciunas is improving at a rate that all raptors fans can be proud of, but I’m not sold on his defense (yet).
                Man, this season is going to be fun to watch (and analyse)!! October 30 can’t come soon enough!!

  • DDon

    I expect Jonas to be a better defender this year than last! And I think that Psycho T, Acey and Buycks will all play improved defense over last years alternatives.

    • DDayLewis

      Psycho T is a really underrated defender. I can’t quite figure out why he’s so successful (he’s strong and mobile, but he’s also short and non-athletic), but he ranks so highly across all defensive stats. I’m not sure how much time Acy and Buycks will see without an injury occurring.

      • ardefen

        hansborough is 6′ 10”. hardly short, even for an nba forward

        • DDayLewis

          Ehh, he’s 6’95 in shoes. Okay, maybe he’s not short, per se, but he doesn’t necessarily possess great height for the position.

      • Casey Sherman

        Perhaps being on Indiana explains the stats? I haven’t looked at his stats in-depth, but I think it’s a general rule of thumb that elite defense is usually the product of being in an elite system rather than being an elite individual defensive player. With some exceptions, of course

        • DDayLewis

          True, and Hansbrough is unlikely to be an exception.

          He’s definitely not a bad defender. He’s incredibly strong and pushes his man out decent post-position. He’s also pretty mobile and able to hedge well on P+R’s.

          • DDon

            I think that Psycho T will be a good influence on Acy! I think Acy has springs in his legs that Psycho T does not have….and a growing variety of shots. I think that his offensive potential is often overlooked. But if he adapts some of Psycho’s attitude and tactics he will be a tougher defender than last year. Acy is very strong and elevates. I think he can take some of the foul pressure off Amir and Jonas.

  • Take it in

    Bs all bs let them play ball dam it

  • FAQ

    Surely it all comes down to Team IQ!

    Team IQ = IQdefense + IQoffense

    Then there’s the “Gel Factor” and when you can call this new bunch of bballers a “team”.

    Any guesses, or must we wait until Xmas to see if they are cake or coal?!

  • hig dale

    Gay and DD take up a lot of our possessions which I wouldn’t mind if they made better decisions and expanded other parts of their game. They have the ball in their hands a lot but have few assist. They don’t lead a fast break because their not comfortable doing that or being in a pick and roll they’re better finishing, but they also turn the ball over. If we’re going to go with this wing generated offense and they’re going to have the ball in their hands that much we need more from them than just a threat to shoot or we will play side line to side line instead of baseline to baseline. They’re the key to the offense but if they play with tunnel vision there may not be a light at the end of it.

    • DDayLewis

      Well said.

  • Wiley

    Another thing to consider is that there is no doubt that there will be better ball distribution this year with a healthy and ready to rip shit up Kyle Lowry and Valanciunas having a great sophomore year.

  • SR

    The offense *should* be better this year. They had so much roster turnover last year, they: a) were switching between point guards with two very different styles of play, and b) brought in a primary volume scorer part way through the year and tried to incorporate him on the fly.

    I’m hoping b) is partly why we saw so many iso’s for Gay instead of more complex plays. Casey admitted they basically took a couple plays for Gay from Memphis and learned them on the fly to run for Rudy in Toronto. This year they have the opportunity to build the offense from the ground-up to fit the personnel (which includes an improving Jonas).

    Also, the stability factor should help, if the players have the IQ to take advantage of getting to know each other. In Miami’s first year together, Wade and LeBron basically just happened to be on the same court at the same time, each doing his own thing. There was very little synergy there. It took an entire season and a trip to the finals for that team to figure out how to maximize playing together at both ends of the court – and those are some of the best basketball players in the world. Here’s to hoping there’s a similar developmental trend in Toronto.

  • Tom

    Having seen him hit 3s in warmups, and having seen him knock down a lot of long 2s in games, I’ve never understood why Amir doesn’t shoot 3s. I think he could do it, and it would dramatically improve our floor spacing.

    Also, as bad as Gay and DeRozan are from 3, I’d still rather see them taking 3s than long 2s.

    • DDayLewis

      I agree. Three’s are worth 1.5x as much (obviously) which justifies a drop in FG%.

    • Jonathan Mac Lean

      I’ve wondered the same thing… I think maybe he’s mentally stuck on his roll of being a defensive, rebounding, hustle player. How many times have we seen him hesitate before shooting? I say if Amir is open, shoot it. Don’t hesitate if the defender isn’t there, just shoot. Can you imagine if Amir has been a strech 4 this whole time?!

    • ppellico

      are you really suggesting Amir be a three point shooter?
      or is this a set up….?

      • Jonathan Mac Lean

        more of a hope than a suggestion… I was only half serious. I think he can hit the 3, but not consistently enough to take the shot regularly during games. I could trust him with a mid-range shot though, even a deep-2 if he’s open. I find his problem is that he will get the ball while wide open, do a fake shot (hesitation) which gives the defender time to get to him, rather than just confidently taking the mid-range shot, while still open.

        • DDayLewis

          I think his shot is just really slow.

  • Phat AlberG

    Casey got 10 to 15 games to prove himself, if he doesn’t George Karl will be next Raptors coach.

    • DDayLewis

      Karl has a bad reputation for not playing centers very much. Are you sure you want him to captain the ship with JV on-board?