An in-depth look at Demar’s passing game, flush with statistics and video breakdowns.

The conversation about Demar Derozan always starts and ends with three things; his post-game, his contract and his inability to sink a three-pointer. But with the Raptors employing such a wing-heavy offense, why doesn’t anyone talk about his passing game?

The obvious answer is that Demar doesn’t produce very many assists. His career high came last season when he dropped a mere 2.5 dimes per game. That’s not very good for a guy who played almost 37 minutes per game last season. His assist rates (% of made baskets assisted by Demar) haven’t been so hot either:

assist rate

However, low assist figures doesn’t mean Demar can’t/won’t pass – there are also contextual factors to consider. Demar is the finisher on most plays and he’s not really tasked with the responsibility of creating for his teammates. This suppresses both his assist rate (because he can’t assist on his own baskets) and assists per game (because that’s not his role). It’s a classic case of “the numbers don’t tell the whole story.”

More importantly, the question shouldn’t be “is Demar a good/bad passer?” That’s far to simplistic. Rather, the question I wanted to answer is “what passes can Demar make?” With that in mind, I dove into the synergy sports video archives to find my answer, and I came up with two areas of note: passes in transition, and passes on drive-and-kick plays.

Passes in Transition

With the help of Louvens Remy, video production wizard and loyal Raptors Republic reader, I was able to provide a TL;DR solution for the post. Presented below is a video “tele-stration” breakdown of Derozan passing in transition:

As you can see from the video above, Demar is a capable passer in transition. He’s usually the finisher on the fast break, where he uses his elite athleticism to score at a rate of 1.23 points per transition play (80th percentile in NBA), but he also has a bit of creativity. Over his four-year career in the NBA, Demar has tightened up his dribble and developed new moves which have allowed him to harness his athleticism. By no means am I suggesting that he’s now got Kyrie Irving’s handles or anything, but the ability to pass makes him far more dangerous than your prototypical finisher on the fast break.

Listed below are two more examples:

His court-awareness on the fast-break is crucial for the Toronto Raptors. Last season the Raptors scored a measly 9.9 points in transition per game, which ranked 26th in the NBA. Yes, a lot of that has to do contextual factors like the efficacy of Toronto’s defense (22nd in DefEff), defensive rebounding rate (16th in the NBA) and pace (23rd), but Toronto clearly failed to maximize the athleticism of their wings.

If Demar is able to shoulder more ball-handling responsibilities on the Raptors’ fast breaks, the team would be less reliant on getting the ball to Lowry before they can break out. Opposing defenses would presumably key in on stopping Demar from getting to the basket (see the aforementioned clips), which should theoretically leave Raptors’ teammates open for layups or jumpshots.

Drive and Kick

Demar is also capable of drawing extra defenders on his drives and finding the open shooters. Once again, with Louvens’ help, here is your TL;DR option:

I discussed Demar’s improved handles in the previous section and how it has allowed him to be effective in fast-break scenarios, but his ability to break down defenders is also key in half-court sets. Last season, Demar averaged 0.87 points per play as the pick-and-roll ball-handler which ranked 30th overall in the NBA. As you can see from the video above, as well as the video below, Demar’s preferred option is to drive into the paint before deciding to attack the rim or to find the open man. Either way, opposing defenders are forced to step up and stop his drive, which distorts the shape and configuration of opponent defenses, thereby creating opportunities for Demar’s teammates.

However, Demar is also able to create in other scenarios, namely in isolation, and more recently, from the post. In the clip below, Demar is able to beat his man, draw the help defender and kick it out to an open Kyle Lowry on the perimeter:

Similarly, he’s also capable of passing from the post. If he’s paired with a knock-down shooter like Steve Novak, opposing defenses will face a difficult decision in whether to allow Demar to attack one-on-one, or to leave Novak to double-down in the post.


Despite posting sub-par assist numbers, Demar is a capable passer in certain scenarios, namely in drive-and-kick scenarios and on the fast break. However, capable and willingness are two things (seriously, 24%+ USG, 12% AST%?), and he is undoubtedly lacking in the latter.

If the plan going forward is to have the offense run through Demar, be it in the post or in pick-and-roll situations, Demar has at least shown the ability to make correct decisions and decent passes in these scenarios, especially in transition and on drive-and-kicks.

Lastly, I would like to offer my sincerest gratitude to Louvens Remy for helping out with collecting the highlights used in this clip and for putting together the two telestrations in the post. Thanks for volunteering your time, offering your feedback on the post itself, and lastly, for putting up with my comically hurried narration. If you ever need guy for putting together a video project or two, drop Louvens a line at @LouvensRemy and make sure to check out his excellent work over at Black Sunrise Pictures.

Statistical support from Hoopdata, NBA Stats, Synergy Sports and Basketball-Reference.

  • theswirsky

    ” Demar is the finisher on most plays and he’s not really tasked with the responsibility of creating for his teammates. This suppresses both his assist rate (because he can’t assist on his own baskets) and assists per game (because that’s not his role).”
    here is some context to that context – he doesn’t get assists because he isn’t asked to do it, or is he not asked to do it because he can’t do it, or is he asked to do it but can’t do it and the team accepts that? What about all the ‘average’ SGs in the league, are they asked to create, or are they not asked to create yet able to anyways? Why would we assume one over the other?
    Randomly throwing out an unfalsifiable hypothesis is not context. Its conjecture.
    What those numbers tell us is one of three things – either he chooses not to pass, or he does not pass effectively or a combination of both those, all vs his peers. Nothing more nothing less.
    This is not a case of numbers not telling the entire story, its a case of not believing what the numbers are telling us, and those are not the same thing.

    • DDayLewis

      I was speaking more in generalities. Being the “finisher”, meaning shooting the ball, would theoretically lower your assist%.

      Assist% = DD’s assists/FGM while DD is playing

      Since he can’t assist on his own shots, if he makes a basket, that increases the denominator, which lowers assist%.

      • theswirsky

        Yes you can’t assist on your own shots – but neither can the ‘average SG’ who gets more assists. We still see plenty of players who get alot of assists while taking alot of shots. Or pass the ball instead of shoot it.
        What we do know is Demar has had equal opportunity to get assists vs his peers (on a % of possessions basis) and does not.

        • DDayLewis

          Okay, what I am pointing out is a characteristic of the AR formula, and how it’s impacted by a player’s role. If your role on the team is to shoot a lot, it spends possessions which thereby increase the denominator.

          Now whether this is the sole reason behind Demar’s low AR figures, or not, is certainly up for debate, but I am trying to give context to the data.

          • theswirsky

            One last stab at explaining this. Here is the equivalent argument with Rudy Gay

            Rudy is shooting 33% from the field vs his peers shoot an average of 45% (made up number). So we see Rudy just misses more shots. So if Rudy hits more shots he will shoot a higher %.

            That Rudy missed shots instead of made shots is not ‘context’, just like Demar taking shots instead of passing the ball is not context. What context would be is Rudy misses more shots than his peers because X or Y happens. (for example, he’s forced into last seconds shots which are bad shots). Next, we need to prove that ‘context’ actually even exists (Rudy actually takes more last second shots than his peers on a per shot basis) for it to apply to the argument. If we don’t do that we aren’t adding context, we are just adding conjecture. Until we do go about proving that Gay takes more last second shots, his fg% vs his peers is telling the entire story as well as know it (ie. he is not scoring as efficiently as his peers, because he is missing more shots on a % basis).

            So Demar has a lower assist % than his peers because he passes less? Yes. We know that. Thats not in question. If he passed more would he have a higher assist %? Yes. We know that to. If you do more of something on the same amount of possession you will have done a higher % of that thing. This applies not only because he will generally assist on more shots by passing more but because he will take less shots which inevitably count as less possession.
            But so would ALL of his peers. If they all passed more, they would also have an even higher assist %. These are just obvious truths like 1+1 = 2, if I put additional water in a bucket I have more water in my bucket, and if I drive faster I will get to my destination sooner.

            The context is the WHY? Why does he pass less than his peers? Your statement was:

            “he’s not really tasked with the responsibility of creating for his teammates”
            do we know this? Has this been stated anywhere? Is there evidence to support this? Just because he scores (shoots) more, doesn’t mean he isn’t asked, expected, would be more desireable if etc, he passed more instead. Does Casey/Masai actually desire him to pass less than his peers? And even if they do, why do they desire him to pass less? Is it because they think he’s an equal good passer as his peers, but a better shooter than them? Or do they simply think he is a better scorer than he is a passer? Or do they simply think he sucks at passing and the lesser of 2 evils is him shooting?
            The answer to the above is pure conjecture, and as such doesn’t change the ‘story’ the stats give us. Rather just adds an additional story that may or may not be true. The stats though couldn’t care less about stories that may or may not be true. The very reason they work is because they ignore stories that may or may not be true.
            The stats are telling us the only unbias truth we know. Demar, to date, does not pass as much and/or aswell as his peers. Period. Full stop. End of Story.

            • DDayLewis

              Your analogy with Gay is totally off.

              I am not saying: Demar will have a higher AR if he just produced more assists. I am saying that using more possessions (shooting, turnovers) will increase the denominator in the AR formula.

              And the “why” is very much up for debate. I put out a suggestion (not his role), and I specifically said “Now whether this is the sole reason behind Demar’s low AR figures, or not, is certainly up for debate, but I am trying to give context to the data.”

              • theswirsky

                “I am saying that using more possessions (shooting, turnovers) will increase the denominator in the AR formula”

                yes I realize this and I included that.

                But that doesn’t change a thing, its just an obvious truth. If you don’t shoot as much (or replace shots with passes) you will use less possession, and as such have a higher assist%. Just like if Rudy missed less shots or took less shots he’d otherwise miss, he’d shoot a better fg%. You change either the numerator or the denimonator in an equation you will naturally get a different number. Thats how fractions work.

                That is not context. Thats just the same math with different values, values that do not exist.

                “I’m simply suggesting that the problem might not be as bad as it seems because of how AR is calculated, and Demar’s role on offense”

                I’m pointing out that there is no evidence to support that assessement. How AR is calculated has nothing to do with the magnitude of the ‘problem’ Rather its Demar actions/decision making on the floor vs his peers to date (the actual input) that is effecting how he compares to his peers (the output).

                As such, claiming “It’s a classic case of “the numbers don’t tell the whole story.” is false.
                If people are going to use stats they need to not only use them right, but explain them right. If they don’t, they are simply offering disinformation.

                • DDayLewis

                  Here’s a little evidence towards the claim that more usage can result in lower AR:

                  Sample: SG who played more than 20 games, more than 15 min per game, USG>20

                  Findings: Average AR in sample: 15.3. Average AR in total SG population: 17.1

                  See the drop off? What’s one possible reason that drop off exists? Because if you use more possessions, your denominator increases.

                  I am not going more rounds on this thing. You’re really not getting my assertion. Re-read the article and my responses.


                • theswirsky

                  I understood “possession change the denominator” because I understand basic math.

                  Does ‘role’ have an effect on AR? Well if we assume ‘role’ (in this situation) = change in usage, and change in usage = (but not exclusive to) change in shots, and change in shots = change in AR, then the relative rate of change in ‘role’ should = the relative rate of change in AR.
                  Therefore AR will be relative to ‘role’

                  See I get it.

                  But here is a question – does that ‘role’ as we are assuming it to be even exist? And if it does, whats the expected relative rate in change of AR?

                  snap…… guess you forgot to, you know, prove existence.

                  …….Oh well who cares about that when we can repeat possession change the denominator over and over.

        • SR

          Would it be more accurate/revealing to compare his assist rate with wing players who have similar usage rates?

  • thedude

    Holy shit, always a tough crowd at RR.

  • HogyG

    While I believe that advanced statistics can add light to what a player can do, I also feel that most people get so lost in the number crunching that they ruin the purpose of the game (like how you and theswirsky have clearly ruined this thread with your fractal shootout… this is BASKETBALL not math class boys. haha). The problem with breaking down these numbers at this point of the year is that DeMar has not finished developing his game, and therefore last year’s numbers do not reflect today’s DeMar. Every season so far, he comes back with a stronger game and more tricks in his bag. As you are drawing from last years numbers (which I understand why, as you need a large enough sample size to draw your conclusions from), you are still talking about last year’s DeMar. My eyeball test, as his early numbers may suggest, that he has improved his distribution skills (along with other facets of his game) but it’s just too early in the season to be trying to use numbers to prove anything, especially using last year’s stats.

    DeMar is finding his teammates on the regular so far this season, and I love what I’ve seen from him on both ends of the floor. We’re on the same page that in the past he’s shown that he’s a capable passer but has come off at times as an unwilling one. I would suggest that this may be due to a coaches request for him to go for the finish when in position to do so. Sometimes, when players are playing with other players who are known for shooting, they find themselves surprised when that guy goes off-type and attempts a pass instead. Last year our team struggled to rebound competitively on a nightly basis, perhaps coach requested that our bigs cheat into the paint when certain teammates were in scoring positions in order to give our rebounders a competitive edge. I’m just guessing that this may have been a strategy previously (because for the most part I have no idea what coach Casey is thinking), but as an example outside the Raps, players on the Knicks generally go for rebounding positions not fill passing lanes when Melo or J.R. Smith have the rock. Also save for 40 games of his career, DeMar has played with Calderon, who was so proficient in finding the open teammate that DeMar could have felt his duty was to fill the basket (meaning it wasn’t due to a lack of desire to pass, but more from a duty to finish up strong), but I’m just throwing out some random thoughts of the moment right here.

    Now, back to bashing analytics. Something that seemingly became popular after everyone saw Moneyball and believed that a similar approach can be brought into other sports like basketball. The problem is that baseball (where all this became popular) is very numbers friendly and is ultimately a one-on-one encounter between the man at bat and the pitcher. This makes extrapolating conclusions much easier when comparing every players numbers up against the same pitchers etc. Basketball is a fluid rotation of five players and therefore poses many more variables to consider before being able to extract any truth from them. As a poor example, it’s generally true that east coast teams win through defense, while west coast teams win with offense. I believe a team plays two thirds of their games against their own conference, so perhaps it would be more accurate to compare DeMar’s numbers only to other east coast players with a similar usage rate as their competition is more likely a fair representation of his nightly competition. Do you add that variable into your breakdown of the numbers? Doubtful, and it wouldn’t be the only one to consider either.

    No disrespect to all the analytic lovers in the Republic, but I don’t believe any of you armchair Einsteins out there are factoring in enough variables to create a true and accurate picture of any player through your dabbling in advanced statistics. If you could, Billy Bean (or maybe the Memphis Grizzlies) would have hired you by now. If you can, then you are clearly wasting your time on this site as MIT may also be interested in your services as well. Jokes aside, numbers may never lie, but they can certainly be misinterpreted. By all means have fun looking at them, but don’t believe you can tell us the whole story with them alone.

    • DDayLewis

      This post is 95% video breakdown. It’s not about the numbers.

      • HogyG

        True, but by the time I read your discussion with theswirsky it was just a math problem to me. But don’t try to deny that you love your numbers.

        • DDayLewis

          Ehh, he wanted to debate one little excerpt. I seriously doubt that he actually read the piece itself.

          And yeah, basketball stats are fun. There’s so much room for interpretation.

          • HogyG

            Agreed, and I don’t want to come off like I don’t accept them at all. I only feel that some people stretch them into relevance. Lately, it feels like many of the discussions across the Republic boils down to analytics, but stem from a gut feeling or something from the heart that they are wrongfully trying to prove with numbers. Sometimes it’s just your opinion, you don’t have to prove it, you just have to give it. For me, math was a class I skipped to go shoot hoops instead, so generally I try to stick with the eyeball or just the classic stats to make my stands with. But stats certainly are fun never the less, without them we wouldn’t have fantasy sports! haha.

            I liked your blend of video, stats/numbers and feelings in this post, it’s nice to see what’s being talked about. The pop-ups on screen were both helpful to follow your thoughts and were really fun too. A few things that might add to the videos in the future (if this is something you plan on continuing), your title screen could use a jingle (think along the lines of sports center) that can brand it as a Raptors Republic piece, you could also lay a music track (like some jazz or an original piece even) or simply an audio track underneath (eg. crowd noise, an arena organ or maybe sound effects with the graphics and transitions) to add production value and help mask/eliminate the slight hollow sound from your recorded voice over. As an example in the first vid, the voice track sounds kind of dead on its own and then the piece comes to life when we see the play in real time with the addition of all the natural sounds we’re used to. Not sure where you recorded your V/O but you may want to consider a space with more sound dampening. If you’re doing it at home, try a closet filled with clothes or hang some moving blankets or thick quilts across the walls in your room and on the floor. Perhaps it only sounds that way because it’s riding all alone over the visuals. Not trying to knock what you guys did at all (it was fun), only trying to help you make it a better product in the future.

            • Louvens Remy

              Thanks for the constructive criticism, HogyG. This first batch of videos was a nice little test run. We are hopefully going to do more and add a little bit more production value to them as we move along. Some of the stuff you mentioned is already on the list but it’s nice to hear what people have to say.

              • HogyG

                Right on. Keep up the good work boys.

    • theswirsky

      Math took a man off the planet earth and put them on the moon, and then, since that wasn’t enough, brought them back. It created a network that transverses the globe and can transfer information in fractions of a second. It created devices so powerful it can not only split something as small as an atom, but can create more energy from those atoms than all other known sources in the world millions of times over. Or, if someone is in a sour mood, anhihilate square miles of life or matter in the blink of an eye instead.
      But there are too many variables in basketball for math to figure out?

      • HogyG

        I’m not questioning the mathematical abilities of Von Braun, Oppenheimer or Einstein, I’m sure for any of them it would be elementary. However, last I looked none of them talk analytics on the Raptors Republic forum. Are you suggesting that your grasp of numbers rivals any of them?