The Real and Fake Shooters of the NBA

What makes a basketball player a “shooter”?

I’d argue that, to qualify as a “shooter”, a basketball player has meet the following criteria:

they have to be good (ie: accurate on the shots that they take)
they have to like to shoot (ie: they shoot a lot)

Given that definition, it should be easy to find the NBA’s shooters, as well as the NBA’s “wannabe shooters”. To identify the shooters, all we have to do is find the players who both shoot and score at an above average rate. Similarly, we can identify the wannabe shooters by finding the players who take a lot of shots despite scoring inefficiently.

This season, the average NBA player has a True Shooting percentage (TS%) somewhere around 53.5%, given the following positional averages:

PG: 53.2%
SG: 53.5%
SF: 53.9%
PF: 53.1%
C: 54.4%
Average: 53.6%

This season, the average NBA players takes around 16.5 field goal attempts per 48 minutes (FGA/48), given the following positional averages:

PG: 16.9
SG: 17.5
SF: 16.4
PF: 16.1
C: 15.0
Average: 16.4

Let’s be a little bit more selective and move that FGA/48 figure up to 18. That leaves us with 100 NBA players this season (minimum of 400 minutes played). How many of these players who like to shoot are actually shooting better than average?

Thirty-seven, if we are generous and use 53% as our cut-off for true-shooting percentage. Topping the list are the following 15 players:

Player	FGA/48	TS%
Kevin Durant	22.3	65.1%
LeBron James	22.8	64.1%
Amare Stoudemire	18.5	63.2%
James Harden	21.7	60.7%
Tony Parker	22.5	59.8%
JaVale McGee	18.3	58.7%
Manu Ginobili	18.5	58.6%
Blake Griffin	20.3	57.6%
Stephen Curry	21.8	57.4%
O.J. Mayo	18.3	57.3%
Kobe Bryant	25.8	57.1%
Dwyane Wade	22.4	57.1%
Ryan Anderson	21.4	57.1%
Kyrie Irving	24.9	56.9%
Vince Carter	19.7	56.4%

Most of those names are not surprising, although I must admit to being surprised at the appearance of Amaré Stoudemire and JaVale McGee, albeit for different reasons (accuracy and shooting rate, respectively).

In contrast, the following 15 players are at the very bottom of the 100 player list:

Player	FGA/48	TS%
Michael Beasley	23.3	45.6%
Kevin Love	23.1	45.8%
Kevin Seraphin	21.1	45.8%
Byron Mullens	19.2	46.5%
Monta Ellis	22.2	47.3%
Rudy Gay	22.1	47.3%
Raymond Felton	20.6	47.5%
Terrence Ross	18.2	48.1%
Glen Davis	21.6	48.3%
Andrea Bargnani	20.9	48.3%
John Wall	20.5	48.3%
Richard Hamilton	21.1	48.4%
John Henson	20.5	48.5%
Dion Waiters	21.9	48.9%
Marreese Speights	21.2	49.2%

Once again, most of this list is not surprising to me, with the exception of Kevin Love and John Henson (accuracy and shooting rate, respectively). Kevin Love’s appearance on this list can probably be explained by the fact that he broke his shooting hand not once, but twice this season.

But I’d like to look at these lists in another way: by team.

Of course this analysis is rather shallow. First of all, it doesn’t account for differences in shooting rates. For example, Kobe Bryant creates more wins for his team than O.J. Mayo does for his — even though Mayo has the better TS% — because Kobe takes more shots while remaining very similar to Mayo in terms of TS%. Second of all, this analysis lumps all players in two broad categories and makes no attempt to differentiate within the two groups. For example, in our list, LeBron James and Kevin Durant are “above average”, but so are Damian Lillard, Klay Thompson, and Andray Blatche, even though those three barely meet our “above average” qualifications. Third of all, we decided to ignore positions when we averaged the positional averages, but it’s not really fair to compare guards with centres when centres shoot almost a full percentage point better. And fourth of all, this analysis does not control for minutes played. Having a wannabe shooter who doesn’t play many minutes — John Henson, for example — is much better than having a wannabe shooter who plays a ton of minutes — Monta Ellis, for example. But I think this analysis gives us a decent idea of the types of shooters each team employs.Dallas, Golden State, Indiana (who knew), and Miami don’t let their inaccurate shooters take a lot of shots. Indiana barely joins this first group, as their two players (Paul George, 53.7%; David West, 53.5%) are right around average. Toronto, Charlotte, Washington, Phoenix, Philadelphia, and Detroit don’t let their accurate shooters take a lot of shots. In particular, Toronto and Charlotte seem to have major problems with this, which shouldn’t really come as a surprise to dedicated NBA fans.

And of course, while it is the most important aspect of basketball, shooting is not the only aspect of basketball; there is also gaining possession, preventing loss of possession, and both individual and team defense. So a team could attempt to make up for playing many wannabe shooters by playing several players who are good in those other areas. A team could also attempt to make up for playing wannabe shooters by also playing several players who shoot efficiently, although less frequently. But it would be a lot more simple if teams just played shooters who are actually good at making the ball go in the basket.

- Devin

22 Responses to "The Real and Fake Shooters of the NBA"

John Algus says:

February 28, 2013 at 12:17 am

Hmm. I think the term “shooters” needs to be changed to “scorers” in this article. I know you miss out on the great backdrop of McGavin, but Blake Griffin is not a better shooter than Ryan Anderson. If you’re going to talk about shooters, talk about jump shooters. Information on jump shooters is easily available at basketball-reference.com.
Ed Herzog says:

February 28, 2013 at 6:49 am

Why are you using true shooting percentage instead of effective field goal percentage? True shooting percentage includes free throws and it sounds like the intent of the article is to only look at field goal attempts.
Al S says:

February 28, 2013 at 6:49 am

I agree with John Algus. In my opinion, the term “shooter” in common NBA parlance refers to a person who primarily shoots jump shots, and specifically long distance jump shots. JaVale McGee wouldn’t be considered a “shooter” under normal NBA parlance.

I acknowledge, of course, that making a lot of layups and dunks is a *good thing*, and thus McGee’s high TS% and shot attempts are positive contributions to his team. But they do not make him a “shooter” under common NBA parlance.

You may believe that your definition is more useful in determining who best helps a team, but it doesn’t accurately show who are the best and worst “shooters” as the term is commonly used.
Andres Alvarez says:

February 28, 2013 at 6:50 am

Interesting comments so far. Just want to point out, while everyone’s arguing the definition of “Scorer vs. shooter”, Devin’s written a pretty good piece examining which teams have players that get the ball through the hoop. According to the rules, I hear that matters, no matter what you call it.
Nathan Baker says:

February 28, 2013 at 7:12 am

Hey, who had the better game last night? Chandler or Curry?
Andres Alvarez says:

February 28, 2013 at 7:16 am

Nathan,
May have to do a post on how to “quickly” derive production from boxscore. Excellent question. Answer? EASILY Curry. His production from scoring alone is almost as good as Chandler’s was AND he’s a PG as opposed to a C. Both performances are insane. Curry’s though was unreal. Chandler was playing in superstar mode. Curry was playing in Game Genie mode.
Arthur North says:

February 28, 2013 at 7:37 am

I would not have looked at TS%, since that takes into account getting to the line, which is a separate skill entirely. Would rather see analysis of eFG% on jumpshots and then possibly lump a list of that with a list of the best free throw shooters.

Of course if you were gonna do only jumpshots you’d have to only look at players that shoot more jumpers per minute than the average at their position (so someone like Steve Novak or Kyle Korver could make the list), and I don’t know where you can easily find that type of data.
Chris D. says:

February 28, 2013 at 7:47 am

“Indiana (who knew)”

Pacers fans knows, although we will admit to being heartily surprised this season. However, their shooting performance would be even better if Roy Hibbert could actually handle the ball within 3 feet of the rim.
EntityAbyss says:

February 28, 2013 at 8:04 am

What Stephen curry did yesterday was unreal. It was like watching a basketball god shooting. Absolutely destroyed the knicks. They just couldn’t guard him.
Hasani Davis says:

February 28, 2013 at 10:01 am

this is the best post ever. because it plays into the major aspect of what I see in the nba
“STAR POWER” plays more into who is taking shots than who can put the ball in the basket.

Ray Felton has been built up all year in ny magazines and radio as a savior and Amare has been billed as a villan. I keep pointing out amare is getting really efficent shots and not taking bad ones, while JR smith melo and Felton keep jacking up bad ones. yet they keep refusing the give amare any type of shot and dumping more into the afore mentioned 3.

the one i CANT explain with numbers…. the raptors. I said rudy gay is a bad shot taker and Derozen is 5 times worse… yet they have been winning. it boggles my mind.
todd says:

February 28, 2013 at 10:20 am

Exposed some pretty poor perimeter defenders. Dang, I was just looking up dfg% stats. The Wizards are pretty good defenders.
isaac says:

February 28, 2013 at 10:45 am

When people who follow basketball say ‘shooter’ they don’t mean the same as a scorer. Steph Curry is a shooter, Amare is not. Also, for this article TS% makes significantly less sense than eFG%, even if TS% is usually the preferred stat.
Brett Candeias says:

February 28, 2013 at 10:52 am

Maybe I’m being too pedantic and misinterpreting the author’s point, but can it really be said that a player “creates” a certain number of wins for a team, implying causation? (i.e. Would it be more accurate to say that one “accounts for” more wins instead?)
Nate says:

February 28, 2013 at 11:19 am

“… Absolutely destroyed the knicks. They just couldn’t guard him. …”

I don’t think the Knicks mind getting ‘destroyed’ if they have more points on the board when time expires.

“May have to do a post on how to “quickly” derive production from boxscore.
…Chandler was playing in superstar mode. Curry was playing in Game Genie mode.”

Knuth’s back-of-the envelope calculation is (points won – posessions used) So:
Points + Rebounds + Steals – Shots Taken – Turnovers

Steven Curry -> 54 + 6 + 3 – 28 – 3 = +32
Tyson Chandler -> 16 + 28 + 2 – 9 -0 = +37

If you like can do a per 48 position adjustment of C + -2, PF – 1, SF+1, PG+1, SG + 2.

Roughly speaking, each ‘Knuth point’ is worth 1/30 of a ‘Win Produced’, so you could say that either of these players ‘took over the game’. Maybe there’s some kind of intangible that I’m not accounting for, but I expect that doing a formal calculation ‘Wins Produced’ will lead to the same result:

Both players had spectacular performances, but even accounting for position Chandler contributed more to the Knicks winning than Curry did for the Warriors. (And Chandler played slightly fewer minutes.)

The fact that Curry’s performance perceived as mind-blowing, while Chandler is just seen as very good, might speak to the low psychological impact of rebounds – especially offensive rebounds compared to their value.

P.S. To avoid possible confusion: I am not Nathan Baker
XavierQ says:

February 28, 2013 at 2:31 pm

Love the Knuth equation, though I’d hesitate on defensive rebounds really counting as a posession. For bar napkin work, it’s kosher.

As for the Pacers, they do not follow traditional shot distribution patterns because of their passing philosophy. Shooters, both good and bad, will be hard pressed to take enough shots to qualify over a season on the Pacers.
Devin Dignam says:

February 28, 2013 at 2:44 pm

All,

Shooters, scorers…we’re all basically talking about the same thing. When you’re actually playing basketball, “shooters” are really what we call the players who are accurate jumpshooters. It doesn’t really matter how often these players shoot, as long as they are accurate. But in the case of this article, I was literally looking for “shooters” — players who shoot the most! But yes, I had a picture in mind when I was coming up with the post :)

I thought about going with eFG% instead of TS%, but to me that misses out on a decent chunk. For example, Kobe really doesn’t shoot that well, in that he normally doesn’t shoot well from three and tends to have a mediocre FG%. But his production from free-throws almost always makes up for (and then some) his poor shooting elsewhere, and he gets so many free-throws because he takes so many field goal attempts. Generally, as your FGA go up, your FTA go up as well. So guys who shoot more shots from the field use free-throws to produce more wins for their teams. That’s why I decided to include it in this analysis. Feel free to disagree, but it was intentional.

It’s also easier (and quicker) to analyse than jumpshot data. Thanks to The NBA Geek, we’ve got positional averages in terms of FGAs and TS%. But we don’t have those averages for jumpshots (not yet, at least). This was an attempt to take a quick look at the type of shooters/scorers each team has. We know it’s not going to be perfect — in the post I included four problems I have with my own analysis!
Devin Dignam says:

February 28, 2013 at 3:34 pm

Nate,

That Knuth equation isn’t bad, but I use Win Score:

Win Score = PTS + STL + OREB + 0.5*DREB + 0.5*AST + 0.5*BLK – TO – FGA – 0.5*FTA – 0.5*PF

Now, it does take a little bit more time to do in your head, but it’s going to give you a better idea of productivity. Here are the lines from the Curry-Chandler game (yes, I like that better than Knicks-Warriors):

Curry: 54 + 3 + 0 + 0.5*6 + 0.5*7 + 0.5*0 – 4 – 28 – 0.5*7 – 0.5*3 = 26.5
Chandler: 16 + 2 + 10 + 0.5*18 + 0.5*1 + 0.5*1 – 0 – 9 – 0.5*7 – 0.5*1 = 25

As Dre said, Curry’s game was a bit better, especially when we consider that Curry is a PG/SG, and Chandler is a C. Of course, it’s a bit more complicated, because Curry played 48 minutes to Chandler’s 42. As it ends up, the Wins Produced data is as follows:

Curry: 48 minutes, WP48 0.727, Wins 0.73
Chandler: 42 minutes, WP48 0.815 Wins 0.71

Which gives us basically the same picture.
Nate says:

February 28, 2013 at 4:12 pm

“…Love the Knuth equation, though I’d hesitate on defensive rebounds really counting as a possession. For bar napkin work, it’s kosher…”

A defensive rebound is definitely taking possession of the ball. On the ‘How to calculate wins produced’ page they talk about crediting the team instead of the player, but in practice it seems to be a small factor. ‘Deconstructing the Rebound with Optical Tracking Data’ effectively suggests a different approach for crediting rebounds, but the data isn’t readily available.
Nate says:

February 28, 2013 at 5:11 pm

I don’t want to be adversarial, but this legitimately confuses me a little:
“[Win Score]’s going to give you a better idea of productivity [than approximate Knuth]”

Is that something we can say with confidence? Do we even know (or have a consensus on) what ‘productivity’ is?

“Win Score = PTS + STL + OREB + 0.5*DREB + 0.5*AST + 0.5*BLK – TO – FGA – 0.5*FTA – 0.5*PF”

The -0.5 (or 0.44) * FTA certainly makes sense to me, but some of the other stuff just seems strange.
For example, since this is the second time it’s come up, can someone elaborate on how discounting defensive rebounds by half give a more accurate notion of ‘productivity’ or why do you credit players with giving assists without charging them for using assists?
Devin Dignam says:

February 28, 2013 at 6:50 pm

Nate,

Win Score is the easy to calculate version of Wins Produced. It explains 98% of the variation in WP48 (so it’s basically the same). Dave’s old site had several articles up about Win Score, but it appears that most of them are now having malware problems. Here’s one that references Win Score. Note that the formula is has changed since this old post was written — the value of a defensive rebound has been halved to account for the diminishing returns associated with defensive rebounds — so those PAWS averages are no longer accurate.

As for diminishing returns on defensive rebounds, head here for more on that.
While We’re Waiting… All stats, all the time | WaitingForNextYear says:

March 2, 2013 at 7:00 am

[...] The Real and Fake Shooters of the NBA — [Devin Dignam/The Wages of Wins Journal] [...]
Rudy Gay lucks out again says:

March 8, 2013 at 12:44 pm

[...] assists, steals, blocks, fouls, and turnovers) The real problem Rudy Gay has? He’s a scorer who can’t shoot. Yes, he scores more points than an average player. The issue? He gets more misses than the [...]

February 2013
M	T	W	T	F	S	S
« Jan				Mar »
				1	2	3
4	5	6	7	8	9	10
11	12	13	14	15	16	17
18	19	20	21	22	23	24
25	26	27	28

Share this: