Long Shots: Part Deux

In a recent post I started to look at the 16-23 ft jump shot — aka the worst shot in basketball. The average FG% at that range is about 39%, but that includes shots that are assisted and unassisted. If we correct for that (by assuming that assisted shots have a +9.5% better chance of scoring), the average "unassisted FG%" is about 34%. That means that the average "assisted FG%" is 43.5%. It should be obvious that any player who can create his own shot at that distance and hit above that 43.5% rate is going to be very valuable, especially against tough defensive teams. There are currently (well, as of a week ago) 5 players in the entire NBA that meet this special criterion:

Player Name Tm Pos UFG% ATT AST%
Al Horford ATL C 47.30 4.93 91.6
Stephen Curry GSW PG 46.16 4.70 29.9
Dirk Nowitzki DAL PF 45.23 6.62 81.8
Chris Paul NOR PG 43.92 3.77 11.4
Steve Nash PHO PG 43.71 3.43 13.6

Three of these players are PG, so they are generally running the offense. That means that even though they are great shooters, it is often going to be more efficient for them to pass the ball to a worse shooter (according to UFG%), who will be (apparently) more efficient on the receiving end of an assist. Let's take a look at non-PG who shoot at an assisted rate above 46%, which is the efficiency of the best shooting PG in the league when taking an unassisted FG:

RANK Player Name Tm Pos AFG% ATT AST%
1 Al Horford ATL C 56.80 4.93 91.6
2 Dirk Nowitzki DAL PF 54.73 6.62 81.8
3 Pau Gasol LAL PF 52.36 2.54 85.7
4 Jason Terry DAL SG 50.92 5.15 58.7
5 J.J. Redick ORL SG 50.85 2.52 49.0
6 Nick Young WAS SG 50.57 6.36 51.9
7 Anthony Morrow NJN SG 50.37 3.14 64.5
8 Hedo Turkoglu ORL SF 50.18 2.68 24.4
9 LeBron James MIA SF 49.46 5.05 21.5
10 Jamal Crawford ATL SG 49.37 3.80 32.9
11 Elton Brand PHI PF 49.03 2.84 78.6
12 Carmelo Anthony DEN SF 48.79 6.57 39.1
13 Luis Scola HOU PF 48.75 5.33 92.1
14 Ray Allen BOS SG 48.70 3.40 71.6
15 Amare Stoudemire NYK PF 48.54 5.41 62.7
16 David West NOR PF 48.25 5.18 86.8
17 Ben Gordon DET SG 48.16 5.38 56.2
18 Andrea Bargnani TOR C 47.92 5.52 58.7
19 Ryan Gomes LAC SF 47.61 2.83 83.1
20 Nene Hilario DEN C 47.08 1.46 88.6
21 Anthony Parker CLE SG 46.89 3.37 59.1
22 Kevin Garnett BOS PF 46.87 5.79 90.8
23 Chris Bosh MIA PF 46.78 5.83 81.3
24 James Harden OKC G 46.45 1.45 32.1
25 Marc Gasol MEM C 46.12 1.26 88.2

There are a lot of interesting things to point out here. First, I am surprised by how many PF/C are on the list. Both Gasol brothers, Horford, Dirk, Brand, Scola, Stoudemire,  West, Bargnani (this has to be his best skill, right?), Nene, Garnett, and Bosh. As an aside, that injury to West is really going to hurt NOH. Someone is going to have to step up big time and take those 5 or so shots per game. Paul will have to take 1-2 more, but remember he is shooting those unassisted, so the efficiency won't be as high as West. The next best shooter at this range on the team is Ariza, but his AFG% is a lowly 35%. In fact, Ariza only attempts 1.5 shots at this range. NOH is the fourth best shooting team from 16-23 ft, so that was a real strength. West's injury is really a major (and possibly devastating) double hit — affecting both NOH post game and long jumpers.

Looking at the list, I see the presence of Jason Terry and Jamal Crawford, two guys who are known for bringing offense off the bench. To me, the "offense" that they bring is the long jumper. Atlanta and Dallas are 1-2 in the league in FG% from 16-23 ft. It's not just Horford and Dirk. Who might be the Terry and Crawford of the future? Well, judging by this list, my guess is guys like Redick, Young, and Morrow.

In fact, let's look a bit more closely at Nick Young. This is a young guy, great athlete, not so great defensively, who is often criticized for lack of efficiency. His TS% this season is 54.1%, which is actually a career high. In a vacuum, you might say he's taking too many bad shots (~6 per 40 minutes), but I would argue that his shot selection is actually where it should be. In fact, he could even be taking more. Here's what the Wizards put on the floor:

Player Name Tm Pos AFG% ATT AST%
Nick Young WAS SG 50.57 6.36 51.9
Kirk Hinrich WAS G 43.87 4.36 38.2
John Wall WAS PG 36.74 4.03 18.5
Andray Blatche WAS PF 32.61 4.42 72.5

Of course, Hinrich is no longer there, but even when he was, Young was by far the best option. Do you really want John Wall or Andray Blatche taking those shots? On a better team, Young might not take so many of those shots, and his TS% might be higher, but the fact of the matter is, that the guy who can shoot from that range is always going to be the best option. That guy, whether it is Dirk or Curry or Young or LeBron, is the guy that has to sacrifice his TS% for the sake of the team. Because somebody does really have to take those shots. When Dirk goes down, look who's left:

Player Name Tm Pos AFG% ATT AST%
Dirk Nowitzki DAL PF 54.73 6.62 81.8
Jason Terry DAL SG 50.92 5.15 58.7
Shawn Marion DAL SF 43.97 1.35 79.3
Jason Kidd DAL PG 36.75 1.41 50.0

When Dirk is not on the floor, and Terry is on the bench, Dallas is screwed! Oh, one last thing. Let's take a look at Melo, the team he used to play for, and the team he plays for now:

Player Name Tm Pos AFG% ATT AST%
Carmelo Anthony DEN SF 48.79 6.57 39.1
Ty Lawson DEN PG 48.66 1.93 19.4
Nene Hilario DEN C 47.08 1.46 88.6
Arron Afflalo DEN SG 43.26 1.67 76.2
Chauncey Billups DEN PG 41.08 3.13 36.0
Player Name Tm Pos AFG% ATT AST%
Raymond Felton NYK PG 49.96 3.48 26.7
Amare Stoudemire NYK PF 48.54 5.41 62.7
Wilson Chandler NYK SF 41.37 2.75 54.0
Danilo Gallinari NYK F 33.97 1.96 79.3

Melo was the best 16-23ft shooter on Denver. Now that he's on the Knicks, he's actually about the same as Amare. And none of the guys NYK sent over to Denver are better than Melo at that range (at least, this season). It's surprising to me that Denver is playing so well right now, but I have a feeling that in the long run, when teams scout them better, and in the playoffs, when defense is tighter, that long jumper is going to become more of an issue. I also think that Amare and Melo, both being efficient 16-23 ft jump shooters, can only help the team in the long run. They clearly haven't figured it out, as evidence by their current losing streak, but to me, the answer there is to get Amare back down in the post, and leave that long jumper to Melo. His TS% will never be stellar, because he takes so many of those shots, but he's almost aways going to be the best option the team has at that range.

One more thing...I left PG off the list above. Here are PG sorted by number of attempts, with the first column reflecting their UFG% rank (among fellow PG):

PG

RANK Player Name Tm Pos UFG% ATT AST%
15 Derrick Rose CHI PG 37.52 4.86 15.6
1 Stephen Curry GSW PG 46.16 4.70 29.9
20 Russell Westbrook OKC PG 34.94 4.56 11.2
5 Jose Calderon TOR PG 41.08 4.30 20.2
13 Darren Collison IND PG 38.11 4.06 9.4
32 John Wall WAS PG 27.24 4.03 18.5
22 Tyreke Evans SAC PG 32.27 4.02 18.2
21 Andre Miller POR PG 33.63 3.81 46.0
2 Chris Paul NOR PG 43.92 3.77 11.4
19 Tony Parker SAS PG 34.99 3.74 31.7
4 Luke Ridnour MIN PG 43.07 3.73 20.3
6 Raymond Felton NYK PG 40.46 3.48 26.7
16 Beno Udrih SAC PG 37.20 3.46 40.0
3 Steve Nash PHO PG 43.71 3.43 13.6
25 Derek Fisher LAL PG 30.91 3.39 74.6
27 Rodney Stuckey DET PG 30.66 3.30 35.2
8 Deron Williams UTH PG 39.83 3.18 54.4
24 Chauncey Billups DEN PG 31.58 3.13 36.0
18 Rajon Rondo BOS PG 35.29 3.11 39.1
28 Brandon Jennings MIL PG 30.38 3.02 17.1
30 Daniel Gibson CLE PG 29.49 2.91 47.5
11 Devin Harris NJN PG 38.27 2.90 18.2
17 Ramon Sessions CLE PG 35.78 2.89 44.4
23 Mike Bibby ATL PG 31.65 2.77 56.3
12 Jrue Holiday PHI PG 38.26 2.75 28.8
14 Jameer Nelson ORL PG 37.77 2.68 13.0
10 Mike Conley MEM PG 38.79 2.62 23.3
7 D.J. Augustin CHA PG 40.10 2.35 20.0
26 Baron Davis LAC PG 30.89 2.32 22.2
29 George Hill SAS PG 29.68 2.01 45.5
9 Ty Lawson DEN PG 39.16 1.93 19.4
33 Kyle Lowry HOU PG 24.84 1.83 33.3
31 Jason Kidd DAL PG 27.25 1.41 50.0

Rose takes the most shots from 16-23 ft, but is only middle of the pack, in terms of efficiency. This means two things: 1) He probably takes too many shots (Deng and Boozer are actually better options on that team) and 2) His overall shooting efficiency (TS%) is lower than what it could be (or should be), if he didn't take all those shots. Does that make him more valuable or less valuable? Tricky question. Look at the other end of the list. Jason Kidd clearly knows his role. He's a bad shooter, and knows that he has better options (Dirk and Terry), and gets them the ball. Does that make him more or less valuable than another PG who shoots more? These are tough questions. And I don't have any great insights to offer, but I do think that these are exactly the kinds of questions that need to be asked by GM's, coaches, and media folks (and heck, us fans). I'll leave it at that, I guess.

 

20 thoughts on “Long Shots: Part Deux”

  1. Basketballvalue only lists an APM value for one Dallas line-up that has neither Terry not Dirk, but it's a +3 line-up. It looks like either Terry or Dirk is virtually always on the court.

  2. Good stuff.

    I checked the 4 lowest assists per game teams and found that they all were right at 60% assisted on jumpers. They took 70% of their shots as jumpers, give or take a few %points. So these least assisted teams take about 28% of their shots are unassisted jumpers. Better teams will take less, but probably still at least twenty-some percent. Using your data you could allocate out "recommended unassisted jumper attempt levels" based on unassisted jumper FG% and some sharing of the load (one player probably can't realistically take 100% of them because of defense pressure, though they could take a large or very large share) and compare that distribution to the actual and ask further questions about player management, player decision-making and the assumptions.

    That could be done instead of talking about or trying to get to zero such shots or in some cases building a model that heavily penalizes for such shots that probably can't be completely eliminated and might deserve to treated more gently, as they sort of are in some cases thru implicit overall break-even points for shooting but usually not at the precise play by play level (you could argue that APM does). But if you know the shot clock part and the unassisted jumper FG% average of the league. team and / or player you could score a mud-range shot attempt differently (in some fashion) in the last 4 seconds of the shot clock than in the middle or the early part of the shot clock and I think a most advanced model would at least try. Your model is the the most likely one to step up to this challenge / opportunity.

    I assume your play by play database would be able to break out unassisted jumpers by part of the shot clock? That would be enlightening to see. If they probably can't eliminate all unassisted jumpers, which teams and players are taking them "at the right time" from a time sensitive efficiency standpoint?

  3. "I assume your play by play database would be able to break out unassisted jumpers by part of the shot clock? That would be enlightening to see. If they probably can’t eliminate all unassisted jumpers, which teams and players are taking them “at the right time” from a time sensitive efficiency standpoint?"

    Yes, certainly. I think that we should see USG as playing into this too. "Good" use of usage (lol) would be a player that takes a lot of those bad shots for a team that needs him to take those shots, while "bad" usage would be the opposite. Somehow, we need to credit players for their shots relative to the distribution that the team needs them to have, what they actually do, and how efficient they are.

    1. Yes, so perhaps one could create a table showing the average distribution of unassisted jumper attempts for teams by player shot frequency (related to roles of '#1 shooter" thru #8+ ) and the average unassisted jumper FG% and then apply an appropriate penality for exceeding your "role-share" of unassisted jumper attempts to the season level data and for underachieving on FG%.

      Some offsetting credit for taking such shots in the last 4 seconds of the shot clock could I guess by applied at PBP or season level.

  4. The unassisted FG% and attempt data is very useful to see.

    I have often been less or far less impressed by high FGA PGs, than many others in the media and fandom. This is the data to demonstrate the targeted issue, the unassisted attempts at poor or extremely poor FG%s. Again, every team will have to take some and some PGs will bear a large share of this load, voluntarily or out of necessity, but it is good to see who is estimated to shot less than 30% FG% on these attempts and who is taking a lot of them. Could / should the team shift at least some of those shots to other teammates? I think the answer is yes in most cases.

    Shifting unassisted low FG% jumpers from PGs to others who shoot such shots better on average will probably only gain fractions of a point or maybe 1 point or a bit more but you probably should sweat the tenths of points or 1 full point if you are serious about winning and winning a title against 29 other teams.

    Kidd's low shot attempts takes on a new light with this data.

    Wall probably needs the most adjustment to reduce low FG% unassisted jumpers. Westbrook and Jennings are other leading cases that also need some adjustment. Very few of their teammates will shot these particular shot opportunities worse on average than they did and most will be better. if you can chance the ball movement and decision-making.

  5. Whats the variance of this stat from year to year? Do players improve or stay consistent for the most part?

    Would be nice to see a few lists:

    1) potential GSW targets for the off season (free agents, trade targets, underrated realistically acquirable players)
    2)Leaders by position
    3) past leaders who "fell off" or were injured in 2010/11 (always looking for some sleepers)

    1. Good questions. I plan to look into some of these issues more in the future. First, I must come up with a good pun for "Long Shots: Part 3".

      Off the top of my head, though, I think Nick Young would be an interesting option to replace Ellis. He has one year left on his contract, but I have a feeling WAS will re-sign him. I actually think that Udoh - although he hasn't shot very efficiently thus far - should put a lot of work into developing that mid-range and long jump shot. Judging by the PF/C on this list, that part of the game is probably underappreciated and more valuable than it appears.

      1. How about "long shots the menage a trois?"

        I was thinking the exact same thing with Udoh, he almost has a turn around jumper (like a stiff Dirk). He has attempted it like 5 times this year, but a start.

        I wouldn't be surprised to see Washington give up Young for very little, they didn't want to play him from the beginning and may think his value is peaked on a poor club.

        My boy from this list, AA arron afflalo, Solid D from reputation ( I would like to see something back this up, Solid 16-23, plus 3 point shot and great size next to curry. He's restricted but gettable.

        1. Age 35, but Anthony Parker kinda fits the bill for a one year contract experience player.... I imagine if Reggie Will stepped in 9 inches he would be near the top of this list, I know you mentioned he was below average. that just seems counter intuitive to what my eyes tell me from watching pretty much every game he has played.

  6. Ryan Gomes stands out, how does he shot the 16-23 so well, but he's so bad at everything else? This year at least.

    1. Last season with MIN, his average was even higher and he took even more attempts. Perhaps, *this* is his value - at least, on a team that needs it.

      LAC needs it. Blake Griffin's UFG% is 26%. Gordon's UFG is 35%. Before he was traded, Baron's UFG% was 31%.

  7. A full list of unassisted jumper attempts per game by player for the league would be good. UFGAs of all types is not quite as good unassisted jumper attempts to me as I assume unassisted attempts at the rim are still pretty good- but it would be good to see that broken out too for players.

    Tables at team level for one or both would be good too to isolate the frequency and efficiency of the 'freelancing" offense of each, necessary to a degree but perhaps excessive in some cases. 82 games has the necessary data but it is not compiled in this manner.

    1. The PBP set I have doesn't have shot distance, but it usually specifies when a shot is a jump shot. For our purposes, it is probably good enough to make the three categories: a) close (dunks, layups) b) jump shots c) 3-pt.

      1. Yeah, seeing all unassisted jump shots as a group is better than not seeing them as a group, though I can see that ideally unassisted long jumpers separately would be most helpful.

        I guess it might be worth saying that if you have to take an unassisted jumper, closer than 16 feet might be better- but really the decision-making (and analysis) should include both distance and degree of "contest". If you can get a long but less effectively contested jumper that might be better than closer but more contested, especially if "closer" is just 10-15 instead of 5-8 feet. Synergy apparently has some accounting of contested.

  8. The average FG% of shots more than 3 feet from the basket up to 23 feet are all about 39% but I guess 16-23 feet shots are assisted 50% more so I can see the reason for your initial focus on the 16-23 foot shot (and not all UFGAs as I was slipping into thinking I was seeing in the lower part of the article).

    UFGA here means just 16-23 footers? If so, it might help to label it longUFGA and longUFG%. Total UFGA and total UFG% could be of interest too and it would help to know which one is viewing.

    1. Also, when does a 16-23 footer become a "detriment" compared to taking at 23 foot 9 incher plus? Or is this just a completely different possession shot than taking a 3 and more of a pick and pop attempt (so many PF's on the list) or late shot clock iso situation?

      This really reminds me of the difference between D wright and Morrow. Although he has gotten better at the 16-23 footer, D wright was horrid inside the arc at the beginning of the year. As opposed to morrow who can pretty much hit from anywhere. How many times have we seen Wright shot fake step in and miss? This is really a interesting stat the more I think about it.

      I would like to see it added to the standard slash line of FG%, 3pt%, and FT%.

      So instead of a great 48/40/85 (steve nash ) line you add in 48/43/40/85...

  9. Doesn't put us in a drastically better spot, BUT

    Ellis for Ben Gordon and Detroits first rounder?? Contracts the same (if i remember correctly) Ben could come off the bench. Sign a defensive 2, AA, Parker or trade for young...

Leave a Reply