FanPost

Using SportVU Data to Take an In-Depth Look at Melo's Passing


Last year, the NBA installed special cameras called SportVU cameras in a few arenas that allowed the league to measure a ton of new and cool things. They captured data on every facet of the game: passing, rebounding, shooting, defense, even the speed at which players traveled. However, the data was not made available to the public, likely because not every arena had the cameras and the league was merely testing the technology. Over the summer, the league installed SportVU cameras in the remaining arenas and, as a result, has made this season's data publicly available, which is awesome news.

The other day, I was playing around with the data.[1] I was looking, specifically, at Carmelo Anthony's possession and passing numbers, and I realized that the data provided by SportVU would be helpful in answering the age old question: is Melo selfish?

In the interest of full disclosure, I'm a fan of his. I don't think he's a selfish player, and I feel like he gets unfairly punished in media narrative for not being LeBron James. Newsflash: no one is LeBron James. Not Kevin Durant, not Paul George, not Carmelo Anthony. I do think a team can win a championship with Melo as their best player; last year's team made that more clear than ever. But that's neither here nor there.

To the matter at hand, my cursory viewing of the SportVU data led me to ask myself three questions about Melo that will hopefully help determine whether Melo is really as much of a ball hog as some say he is:

1. Does Melo pass the ball more or less often than other players of his ilk?

2. Do Melo's passes lead directly to shots more or less often than other players of his ilk?

3. Who is more to blame for Melo's low assist totals: Melo or his teammates?

Before I can divulge the conclusions I've arrived at and what I think they mean for Melo as a player, I should explain what data I pulled and why. First, I found every player in the NBA that had played in at least 10 games[2] and passed the ball at least 30 times per game on average. I figured that any player who passed less than 30 times would be either a spot-up shooter (like Ryan Anderson) or a marginal player and therefore not very relevant for this exercise. Next, I removed any players from that list who averaged less than 30 front court touches[3] per game. The reason I chose front court touches rather than all touches is because it helped eliminate some big men that are mostly finishers but get a lot of touches in the back court as the ball is being brought up the floor. For example, Kevin Garnett (fuck that guy) averages 52.2 touches per game but only 27.9 front court touches per game. However, he averages 41.4 passes per game, which leads me to believe that he does a lot of passing in the back court; those passes are fairly inconsequential. The obvious issue here is that a player could average 30 passes and front court touches per game, but most of those passes could be in the back court and I would have no way of knowing. Unfortunately, there's no stat for front court passes. There's also not really an easy way to approximate this; the first thought that comes to mind is that I should say that since 27.9 is 53.4% of 52.2, Kevin Garnett probably averages 22.1 front court passes per game, or 53.4% of 41.4. But doing such a rudimentary calculation with no context is risky. Kevin Garnett could easily average 27.9 front court passes per game. He could also easily average 17.1 front court passes per game.[4] 22.1 is close to the average of those two extremes, but I'm not comfortable with making assumptions like the ones I'd have to make here, so instead I'm leaving the passing averages alone. As such, some of the players that remain average more passes per game than front court touches, but I'm not too concerned with those players as it is. Their numbers are relevant enough to the study to not be removed, but I don't take an in depth look at any of them.

Finally, I removed any players that I could easily classify as point guards, for obvious reasons. Melo is not a point guard; comparing him to point guards is a waste and would make his numbers look worse than they are. This left me with 62 players. The players, and the relevant data for each one, are listed below.[5]

Player GP MIN per game Touches per game Front Court Touches per game Close Touches per game Elbow Touches per game AST per game Passes per game Assist opportunities per game Passes per Front Court Touch Assist opportunities per Pass AST per Assist opportunity Player
Al Horford (ATL) 20 32.8 62.3 40.3 5.8 7.3 2.4 43.6 5.4 1.08 0.12 0.44 Al Horford (ATL)
Alec Burks (UTA) 20 26.7 51.3 45.5 0.5 2.7 2.5 36.6 6.4 0.80 0.17 0.39 Alec Burks (UTA)
Andre Iguodala (GSW) 13 37.4 54.5 45.8 1.9 2.9 6.3 40.9 9.9 0.89 0.24 0.64 Andre Iguodala (GSW)
Andrea Bargnani (NYK) 17 31.3 49.3 30.1 2 4.8 1.5 32.9 2.9 1.09 0.09 0.52 Andrea Bargnani (NYK)
Arron Afflalo (ORL) 18 37.3 67.1 54.7 0.7 3.4 4.2 46.1 7.8 0.84 0.17 0.54 Arron Afflalo (ORL)
Avery Bradley (BOS) 20 30.2 49.5 40.9 0.9 1.9 1.3 32.8 3.4 0.80 0.10 0.38 Avery Bradley (BOS)
Blake Griffin (LAC) 20 36.8 78 48.8 5.5 12.6 3.1 55.4 6 1.14 0.11 0.52 Blake Griffin (LAC)
Bradley Beal (WAS) 13 40.2 59.9 52.1 0.4 3.5 3.5 36 6.2 0.69 0.17 0.56 Bradley Beal (WAS)
Carlos Boozer (CHI) 17 30.2 49.3 30 6.2 6.4 1.9 31.4 3.4 1.05 0.11 0.56 Carlos Boozer (CHI)
Carmelo Anthony (NYK) 17 39.9 69.3 53.5 3.3 5.8 2.7 40.9 6.6 0.76 0.16 0.41 Carmelo Anthony (NYK)
Caron Butler (MIL) 11 30.2 54.8 40.5 1.1 3.9 1.6 38.5 3.6 0.95 0.09 0.44 Caron Butler (MIL)
Chandler Parsons (HOU) 18 37.4 58.6 49.2 1.8 2.1 3.9 41.4 6.8 0.84 0.16 0.57 Chandler Parsons (HOU)
David Lee (GSW) 18 33.4 53.6 30.2 3.9 4 2.5 34.3 4.2 1.14 0.12 0.60 David Lee (GSW)
David West (IND) 19 30.4 60.3 36.4 3.9 8.2 3.1 44.6 5.5 1.23 0.12 0.56 David West (IND)
DeMar DeRozan (TOR) 16 38.9 57.4 52.9 1 4.3 2.6 33.2 5.9 0.63 0.18 0.44 DeMar DeRozan (TOR)
DeMarcus Cousins (SAC) 14 31.6 60.1 38.9 8.1 5.1 2.6 35.2 4.9 0.90 0.14 0.53 DeMarcus Cousins (SAC)
DeMarre Carroll (ATL) 20 31.3 40.8 30.6 0.9 0.7 1.6 30.6 3.3 1.00 0.11 0.48 DeMarre Carroll (ATL)
Derrick Favors (UTA) 20 32.9 66 31.1 6.4 5.5 1.4 50.2 3 1.61 0.06 0.47 Derrick Favors (UTA)
Dirk Nowitzki (DAL) 20 32.5 63.4 41 1.6 5.8 2.5 42.8 5 1.04 0.12 0.50 Dirk Nowitzki (DAL)
Dwight Howard (HOU) 20 33.4 56.8 30.1 6.3 3.3 1.8 39.1 3.5 1.30 0.09 0.51 Dwight Howard (HOU)
Dwyane Wade (MIA) 14 33.9 61.1 49.2 3.3 5.5 5.4 39.8 9.3 0.81 0.23 0.58 Dwyane Wade (MIA)
Evan Turner (PHI) 19 37.1 66.8 56.7 1.5 4.2 3.9 42 8.3 0.74 0.20 0.47 Evan Turner (PHI)
Gerald Henderson (CHA) 19 32.5 59.2 51.4 1.6 4.3 2.5 41.6 5.6 0.81 0.13 0.45 Gerald Henderson (CHA)
Gordon Hayward (UTA) 20 37.1 72 64.8 1.2 3.9 5 50.5 10.8 0.78 0.21 0.46 Gordon Hayward (UTA)
Greg Monroe (DET) 19 34.3 54.2 31.2 4.4 8.4 2.3 37.7 4.8 1.21 0.13 0.48 Greg Monroe (DET)
J.R. Smith (NYK) 12 31.4 52.2 39.8 0.7 2.5 1.9 36.2 5.1 0.91 0.14 0.37 J.R. Smith (NYK)
James Harden (HOU) 16 38.9 67.1 59.7 0.9 3.3 5.5 41.8 11.4 0.70 0.27 0.48 James Harden (HOU)
Jeff Green (BOS) 20 34.4 48.2 38.7 1.1 2.4 1.6 31.1 3.6 0.80 0.12 0.44 Jeff Green (BOS)
Joakim Noah (CHI) 17 31.6 63.7 30.9 3.9 7.1 3.4 51.5 5.6 1.67 0.11 0.61 Joakim Noah (CHI)
Joe Johnson (BKN) 19 33.7 48.4 42.8 1.3 3.7 2.6 32.4 6 0.76 0.19 0.43 Joe Johnson (BKN)
Jordan Crawford (BOS) 20 29.3 70.7 60 0.4 3.5 5.2 54.8 10.7 0.91 0.20 0.49 Jordan Crawford (BOS)
Josh McRoberts (CHA) 18 28.3 72.6 41.5 0.9 7.4 4.1 62.1 7.8 1.50 0.13 0.53 Josh McRoberts (CHA)
Josh Smith (DET) 19 36.3 56.6 40.1 2.7 2.7 3.5 37.5 6.2 0.94 0.17 0.56 Josh Smith (DET)
Kevin Durant (OKC) 17 39.4 66.9 55.1 1.9 5.2 5.1 39.1 9.3 0.71 0.24 0.55 Kevin Durant (OKC)
Kevin Love (MIN) 19 35.9 88.2 50.1 5.2 11.5 4.1 63.4 8.5 1.27 0.13 0.48 Kevin Love (MIN)
Kevin Martin (MIN) 18 35.2 53.8 47.3 0.8 2.9 2.3 31.6 5.3 0.67 0.17 0.43 Kevin Martin (MIN)
Kyle Korver (ATL) 16 33.7 46.1 33.1 0.9 3.1 2.8 35.6 4.6 1.08 0.13 0.61 Kyle Korver (ATL)
LaMarcus Aldridge (POR) 19 37.4 69.3 43 4.5 9.6 2.6 43.3 5.8 1.01 0.13 0.45 LaMarcus Aldridge (POR)
Lance Stephenson (IND) 19 35.1 53.6 44.8 1.5 2.1 5 38.5 9.1 0.86 0.24 0.55 Lance Stephenson (IND)
LeBron James (MIA) 19 35.6 72.2 61.1 2.9 6.3 6.1 49.2 11.1 0.81 0.23 0.55 LeBron James (MIA)
Luol Deng (CHI) 17 38.6 55.4 46 2.9 3 4.2 34.1 7.4 0.74 0.22 0.57 Luol Deng (CHI)
Manu Ginobili (SAS) 18 23.6 42.7 37.7 0.8 2.7 4.5 30.5 8.4 0.81 0.28 0.54 Manu Ginobili (SAS)
Marc Gasol (MEM) 13 34.5 67.7 47.7 8.5 18 4.3 49.2 8.5 1.03 0.17 0.51 Marc Gasol (MEM)
Martell Webster (WAS) 18 33.9 45.2 31.9 0.8 1.1 2.3 33.8 3.6 1.06 0.11 0.64 Martell Webster (WAS)
Monta Ellis (DAL) 20 37.3 74.8 65.3 0.5 3.9 5.8 50.3 11.7 0.77 0.23 0.50 Monta Ellis (DAL)
Nene (WAS) 15 33.3 50 34 6.3 6.7 2.9 33.8 5.7 0.99 0.17 0.51 Nene (WAS)
Nicolas Batum (POR) 19 36 64.5 45.5 0.7 3.1 5 49.7 8.6 1.09 0.17 0.58 Nicolas Batum (POR)
Nikola Vucevic (ORL) 17 33.2 69.6 36.2 4.6 4.2 2.5 52.5 4 1.45 0.08 0.63 Nikola Vucevic (ORL)
O.J. Mayo (MIL) 17 31 49.9 42.5 1 5.2 2.4 32.6 5.5 0.77 0.17 0.44 O.J. Mayo (MIL)
Pau Gasol (LAL) 18 30.8 70.4 43.8 4.1 11.4 3.1 50.7 7.2 1.16 0.14 0.43 Pau Gasol (LAL)
Paul George (IND) 19 36.7 68.8 56.3 2.2 4.7 3.4 43.8 7.5 0.78 0.17 0.45 Paul George (IND)
Paul Millsap (ATL) 19 32.1 55.4 35.3 4.4 5.7 2.6 37.8 4.9 1.07 0.13 0.53 Paul Millsap (ATL)
Rudy Gay (TOR) 16 35.8 58.4 45.5 1.8 3.4 2.3 33.6 5 0.74 0.15 0.46 Rudy Gay (TOR)
Shawn Marion (DAL) 20 32.7 53.4 34.5 3 2.4 1.9 39.6 4.2 1.15 0.11 0.45 Shawn Marion (DAL)
Spencer Hawes (PHI) 17 33.3 70.6 40 3.9 5.3 3 53.8 5.8 1.35 0.11 0.52 Spencer Hawes (PHI)
Tayshaun Prince (MEM) 18 27.8 41.8 34.8 1.1 3.6 1.6 32.9 4 0.95 0.12 0.40 Tayshaun Prince (MEM)
Thaddeus Young (PHI) 16 33.2 63.2 37 4.8 4.5 1.5 44.2 3.7 1.19 0.08 0.41 Thaddeus Young (PHI)
Tim Duncan (SAS) 16 28.1 54.1 35.3 4.9 9.3 2.9 38.4 4.4 1.09 0.11 0.66 Tim Duncan (SAS)
Trevor Ariza (WAS) 13 35.3 51.6 39.5 0.5 1 2.2 35.5 4.5 0.90 0.13 0.49 Trevor Ariza (WAS)
Victor Oladipo (ORL) 18 29.9 63.4 54 1.1 2.2 3.8 43.7 8.4 0.81 0.19 0.45 Victor Oladipo (ORL)
Vince Carter (DAL) 19 24.3 44.7 32.1 1.1 2.4 2.4 31.6 5.2 0.98 0.16 0.46 Vince Carter (DAL)
Zach Randolph (MEM) 16 32.7 56.1 35.3 6.4 4.8 1.8 36.9 4.6 1.05 0.12 0.39 Zach Randolph (MEM)
Mean - - 59.3 43.1 2.7 4.9 3.1 40.9 6.2 0.98 0.15 0.50 Mean
Std Dev - - 9.8 9.5 2.1 3.1 1.3 8.0 2.3 0.23 0.05 0.07 Std Dev
Player GP MIN per game Touches per game Front Court Touches per game Close Touches per game Elbow Touches per game AST per game Passes per game Assist opportunities per game Passes per Front Court Touch Assist opportunities per Pass AST per Assist opportunity Player

The last three columns on the right are quick averages I did myself. Each column contains a statistic that should help answer one of the three questions I posed above. I also included the mean[6] and standard deviation.[7] Let's dive in. Don't forget that since this is only about 5 weeks into the season, the sample sizes are quite small and the data may be a bit unreliable.

Question #1: Where does Melo's passing rate among other players, quantity-wise?

To determine this, I figured the best two numbers to look at were passes per game and front court touches per game. The mean passes per front court touch for all 62 players is 0.98, or a little less than 1, with a standard deviation of 0.23. Melo falls within one standard deviation of the mean, as he averages 0.76 passes per front court touch and 0.98-0.23 is 0.75. Assuming this is a normal distribution (which it probably isn't[8]), 34.1% of players, or about 21 of the 62, should be within that same range of 0.75 to 0.98, so Melo's certainly not alone there. He passes less often than the average player in this sample, but not ridiculously so.

I'd also like to look at how Melo stacks up against a few specific players he's often compared to. I chose Kevin Durant, James Harden, and Paul George. I did not choose LeBron. Comparing anyone else to LeBron is a waste and I won't do it. I also chose Monta Ellis because he's having a great year and I was curious about him, and Victor Oladipo[9] because he was in the general data range of the others and fit the regression line, as you'll see in the graph below.

Graph1_medium

The large triangles are the data points for the players I've highlighted. I picked colors that would easily identify with the players, but just in case: Melo is orange, Durant is light blue, George is yellow, Monta is gray, Oladipo is white, and Harden is red. As you can see, other than Monta Ellis (who, like I said, I'm only looking at out of curiosity), all of the players are pretty close to each other. Melo actually averages the least touches per game of the players I looked at, but is in the middle of the pack passes-wise. The black line in the middle of all the data is a regression line.[10] The equation that starts with "y=" in the bottom corner is a way to determine how many passes a player should be expected to make based on the amount of front court touches they get. The linear equation suggests that, since Melo averages 53.5 front court touches per game, he should be averaging 44.2 passes per game. He only averages 40.9 (which, by the way, is exactly average for all 62 players). So based on this, Melo passes less often than he would be expected to. However, the number below the linear equation is the correlation coefficient[11], which basically tells us how predictive the linear equation really is. A number close to 0 suggests that the equation isn't very good while a number close to 1 suggests it is. This number is 0.14, which means that the equation doesn't help too much. So while Melo may pass less often than he should, we can't be sure. It's important to note, though, that Durant, George, and Harden also pass less often than they should (especially Durant and Harden) based on the linear equation. Therefore, I can say that Melo probably isn't too out of the ordinary when it comes to how often he passes.

Question #2: Do Melo's teammates take shots off his passes more or less often than average?

In this case, the two numbers I'm looking at are assist opportunities per game and passes per game. I chose assist opportunities rather than straight assists because what's important here is whether the other Knicks are taking shots off of Melo's passes, not whether they're making them. The mean in this case is 0.15 assist opportunities per pass with a standard deviation of 0.05. Melo averages 0.16 assist opportunities per pass, which is right at the average. Based on that, Melo's not out of the ordinary when it comes to passes leading to shots. Let's check out the graph:

Graph2_medium

I've highlighted the same 6 players. As you can see, Melo averages by far the lowest assist opportunities per game. However, Melo's right near the regression line while the other 5 players are way above it. This suggests that the other players are making passes that lead to shots much more often than they'd be expected to based on the amount of passes they make. The caveat here is that the correlation coefficient for this regression line is 0.21, which is obviously not very high, so again, we can't be too sure of the conclusions. On the other hand, we do know that Melo's not up there with the others. As Allzingers suggested to me in the comments the other day, this could mean that Melo's not getting his teammates the ball in great places to score, but since these numbers don't provide in-game context, there's no way to know. Either way, Melo's passes are leading to shots at an average rate.

Question #3: Are Melo's low assist totals his fault or his teammates'?

We already know that Melo's average when it comes to assist opportunities per pass. But what I'm now curious to know is whether or not his teammates are converting enough of his assist opportunities into assists. The mean assists per assist opportunity (which is kind of like FG%, but not exactly) is 0.5 with standard deviation of 0.07. That makes sense; an average team shoots close to 50% and you wouldn't expect too many teams to shoot higher than 57%[12] or lower than 43%. Melo averages 0.41 assists per assist opportunity, which is terrible. It's more than one standard deviation away from the mean, which means that Melo's in the minority here. Other than Melo, the team shot a bit under 43% for the season through Thursday, so the fact that they're converting only 41% of shots they take off of Melo's passes is interesting. Let's compare him to the other players:

Graph3_medium

Other than Kevin Durant, the players I've highlighted are all below the regression line. This regression line is actually very predictive, as you can see both by how the data seems linear and the correlation coefficient of 0.9. It makes sense that there wouldn't be too many outliers when measuring something so similar to FG%. I think, however, that a deeper look into the data is necessary. The Pacers as a whole shoot 45%. Paul George's assists per assist opportunity is 0.45. The Rockets shoot 47.9%. Harden's assists per opportunity is 0.48. The Thunder shoot 45.7%, but Durant's assists per opportunity is 0.55. The Magic and Mavericks both shoot 46.1%. Oladipo's assists per opportunity is 0.45, while Monta's is 0.50. This suggests that Monta Ellis and Kevin Durant are passing the ball to teammates that are wide open quite often, while Melo and Victor Oladipo aren't finding open teammates as often. This could mean a number of things: Melo's not taking advantage of passing out of double teams, he's not giving the ball to teammates in good spots, etc. No matter the issue, it's probably Melo's fault that his assist totals are low.

So, what have we learned? Melo is average when it comes to how often he passes and how often teammates take shots off of his passes, but those shots aren't going in as often as they should. It probably means that Melo isn't the ball hog that some people make him out to be, but that he also isn't setting up his teammates to succeed. Again, the sample size is small and there's not enough context here, so I don't want to make any concrete observations. Hopefully, I'll revisit this at the end of the season when I'll be able to make conclusions without as many sample size warnings. I can say that Melo's passing isn't nonexistent, but it's not great either. He seems to have made the same observation based on his quote about trying something new after the game last night, and probably without all the research I did.

[1] It can be found at http://stats.nba.com/playerTracking.html. The interface is really intuitive and the site does an excellent job of explaining everything (certainly a better one than I can do).

[2] 10 was a fairly arbitrary number, but I picked it because no team had played more than 20 games as of Thursday so this way any player I picked would have played in at least half of their team's games. By the way, all of this data is only updated through Thursday as I pulled it on Friday afternoon.

[3] Front court touches are exactly what you would imagine them to be. Any time a player receives the ball on his team's offensive side of the court, it registers as a front court touch.

[4] 52.2-27.9 is 24.3 back court touches per game. Therefore, if he passes the ball every time he touches it in the back court, the fewest front court passes he could average is 41.4-24.3 or 17.1.

[5] I included a few columns of data that I don't actually use for the study, such as close and elbow touches, just as a point of reference.

[6] Mean is just another way of saying "average."

[7] Basically, standard deviation attempts to measure how much the data varies from the mean. For more, check out the page for standard deviation on Wikipedia.

[8] I'd imagine more players average less than 1 pass per front court touch than vice versa, but let's keep it simple.

[9] I know that technically Oladipo's a point guard, but he really isn't. The Magic kind of don't want to win games.

[10] http://en.wikipedia.org/wiki/Linear_regression

[11] http://en.wikipedia.org/wiki/Coefficient_of_determination

[12] I wouldn't expect any teams to even shoot close to 57%, but: 1. Small sample size, and 2. As I said, the distribution probably isn't a normal one but I'm pretending it is.

X
Log In Sign Up

forgot?
Log In Sign Up

Forgot password?

We'll email you a reset link.

If you signed up using a 3rd party account like Facebook or Twitter, please login with it instead.

Forgot password?

Try another email?

Almost done,

Join Posting and Toasting

You must be a member of Posting and Toasting to participate.

We have our own Community Guidelines at Posting and Toasting. You should read them.

Join Posting and Toasting

You must be a member of Posting and Toasting to participate.

We have our own Community Guidelines at Posting and Toasting. You should read them.

Spinner.vc97ec6e

Authenticating

Great!

Choose an available username to complete sign up.

In order to provide our users with a better overall experience, we ask for more information from Facebook when using it to login so that we can learn more about our audience and provide you with the best possible experience. We do not store specific user data and the sharing of it is not required to login with Facebook.

tracking_pixel_9347_tracker