Joined: 18 Aug 2005
Posts: 370
PostPosted: Wed Aug 12, 2009 9:12 pm Post subject: Exploring Shot Types at the Team Level Reply with quote
As some of you may already know, I've been doing a lot of work using the play-by-play data at BasketballValue, specifically with shot types. Today I decided to focus on just one team: the Los Angeles Lakers. I took at look at their shots at the per minute and per quarter levels. You can find the link here:
http://basketball-statistics.com/blog1/ ... es-lakers/
Up next I was thinking about doing something similar for an individual player. In order for the sample sizes to be large enough, it would probably have to be a player with a ton of shot attempts such as Kobe Bryant. Does anyone have suggestions for looks at players that would be informative/interesting?
Back to top
View user's profile Send private message AIM Address
Harold Almonte
Joined: 04 Aug 2006
Posts: 616
PostPosted: Thu Aug 13, 2009 1:31 am Post subject: Reply with quote
Can we say that teams's shot strategies start the game trying to be "fair" and towards the end are obligued to try the "optimum"?
Back to top
View user's profile Send private message
Crow
Joined: 20 Jan 2009
Posts: 829
PostPosted: Thu Aug 13, 2009 2:37 am Post subject: Reply with quote
Maybe another part is that players start the game thinking they can the mid-range well then they experience shooting 40% or less as most do and they trim back those attempts out of embarrassment of missing. And the defense has more energy early and the 3 ball is better covered.
As for target players maybe you could do Bryant, James, Pierce and maybe Ginobili or Carmelo. See how they compare.
Or Billups vs Parker.
Or D Howard vs Shaq.
Or Gasol vs Garnett.
Or Andre Miler vs Kidd.
Or just Odom. Or Roy. Or Iggy or Durant, Yao, R Lewis, Arenas when he was healthy, etc.
Last edited by Crow on Fri Aug 14, 2009 1:34 pm; edited 1 time in total
Back to top
View user's profile Send private message
Jon Nichols
Joined: 18 Aug 2005
Posts: 370
PostPosted: Thu Aug 13, 2009 12:57 pm Post subject: Reply with quote
I think those are both reasonable hypotheses, although it might be prudent to look at effective field goal percentage as the game elapses before we decide if anything is optimal.
Back to top
View user's profile Send private message AIM Address
DJE09
Joined: 05 May 2009
Posts: 148
PostPosted: Thu Aug 13, 2009 8:11 pm Post subject: Reply with quote
Jon,
at 82games we see that LAL have 3 of top 30 2pt Jumper shooters (many of their inside shots will be your "mid-range", so my comparison of the data is not necc. accurate). I went to 30, cos Kobe is at 30, and he shoots lots of 2pt jumpers - despite not being as accurate as Ray Allen Smile. So as a team they shoot these mid-rage shots "well". Therefore we might expect an elevated level of this type of shot, and used more as a weapon. So I am saying that LA tends to shoot more mid-range as they are better at it than most teams, and it is not a primary defensive strategy to deny open long 2.
In fact Fisher is 5th best in 2pt Jumper% on last season. Both he and Gasol are over 45%.
To offer an alternative hypothesis, we may see an increased frequency of mid-range shots initially - to create higher percentage shots later: Gasol shoots long 2s so defender has to come to him when he has ball at range, creating passing and driving opportunities later in game. Kobe and Fish also for that reason, and so they can potentially set up 3pt shots later as teams are focussed on denying the 2pt jumper as well - you can only take away so much that top players can do.
As evidence for this I would point out that all 3 play the first 8 minutes (usually) the period where there is a very high number / proportion of mid-range shots, coupled with the fact that this is not part of Odom's game so we see an increase in the number of inside shots when he checks in.
As a corollary to this I would point out that Houston is on record as saying they actively 'give' Kobe the Long 2 to deny his other (higher scoring value) options.
For 2pt Jumper % we can see Boston also has this as a strength, whilst Atl, Den, NYK and Phi are very weak (no player in top 40 - nb min 100 attepmts to qualify). Perhaps Den would be an interesting comparison?
Back to top
View user's profile Send private message
DJE09
Joined: 05 May 2009
Posts: 148
PostPosted: Thu Aug 13, 2009 8:36 pm Post subject: Reply with quote
Also, I just noticed, There seems to be a difference of about 1.5 shots per quarter between the halves. This is probably compounds the impact of 3s.
So the lakers seem to play the second half at a slower pace than the first? or perhaps they turn the ball over more in the second half Smile or more likely they get foulled more ... (sorry Jon I know you are sick of me and my shooting fouls)
Back to top
View user's profile Send private message
Jon Nichols
Joined: 18 Aug 2005
Posts: 370
PostPosted: Thu Aug 13, 2009 9:12 pm Post subject: Reply with quote
DJE09 wrote:
Jon,
at 82games we see that LAL have 3 of top 30 2pt Jumper shooters (many of their inside shots will be your "mid-range", so my comparison of the data is not necc. accurate). I went to 30, cos Kobe is at 30, and he shoots lots of 2pt jumpers - despite not being as accurate as Ray Allen Smile. So as a team they shoot these mid-rage shots "well". Therefore we might expect an elevated level of this type of shot, and used more as a weapon. So I am saying that LA tends to shoot more mid-range as they are better at it than most teams, and it is not a primary defensive strategy to deny open long 2.
In fact Fisher is 5th best in 2pt Jumper% on last season. Both he and Gasol are over 45%.
To offer an alternative hypothesis, we may see an increased frequency of mid-range shots initially - to create higher percentage shots later: Gasol shoots long 2s so defender has to come to him when he has ball at range, creating passing and driving opportunities later in game. Kobe and Fish also for that reason, and so they can potentially set up 3pt shots later as teams are focussed on denying the 2pt jumper as well - you can only take away so much that top players can do.
As evidence for this I would point out that all 3 play the first 8 minutes (usually) the period where there is a very high number / proportion of mid-range shots, coupled with the fact that this is not part of Odom's game so we see an increase in the number of inside shots when he checks in.
As a corollary to this I would point out that Houston is on record as saying they actively 'give' Kobe the Long 2 to deny his other (higher scoring value) options.
For 2pt Jumper % we can see Boston also has this as a strength, whilst Atl, Den, NYK and Phi are very weak (no player in top 40 - nb min 100 attepmts to qualify). Perhaps Den would be an interesting comparison?
Good points. Denver, Atlanta, or maybe someone like the Magic may be an interesting follow-up study.
Back to top
View user's profile Send private message AIM Address
Jon Nichols
Joined: 18 Aug 2005
Posts: 370
PostPosted: Thu Aug 13, 2009 9:16 pm Post subject: Reply with quote
DJE09 wrote:
Also, I just noticed, There seems to be a difference of about 1.5 shots per quarter between the halves. This is probably compounds the impact of 3s.
So the lakers seem to play the second half at a slower pace than the first? or perhaps they turn the ball over more in the second half Smile or more likely they get foulled more ... (sorry Jon I know you are sick of me and my shooting fouls)
If I had to guess I would say pace for the entire league slows down as the game progresses.
Back to top
View user's profile Send private message AIM Address
Ryan J. Parker
Joined: 23 Mar 2007
Posts: 711
Location: Raleigh, NC
PostPosted: Thu Aug 13, 2009 9:22 pm Post subject: Reply with quote
It would be interesting to see those graphs in terms of % of shots taken.
_________________
I am a basketball geek.
Back to top
View user's profile Send private message Visit poster's website
DJE09
Joined: 05 May 2009
Posts: 148
PostPosted: Thu Aug 13, 2009 10:28 pm Post subject: Reply with quote
Has anybody studied change in pace in-game?
Ryan, you can eye-ball the charts for the proportions, but without the shooting fouls the proprtions (of total shots) would be skewed. Smile
Back to top
View user's profile Send private message
Ryan J. Parker
Joined: 23 Mar 2007
Posts: 711
Location: Raleigh, NC
PostPosted: Thu Aug 13, 2009 10:55 pm Post subject: Reply with quote
You must have better eye-balls than me. Razz
_________________
I am a basketball geek.
Back to top
View user's profile Send private message Visit poster's website
Jon Nichols
Joined: 18 Aug 2005
Posts: 370
PostPosted: Fri Aug 14, 2009 11:39 am Post subject: Reply with quote
Ryan J. Parker wrote:
It would be interesting to see those graphs in terms of % of shots taken.
Here you go, Ryan:
Back to top
View user's profile Send private message AIM Address
Ryan J. Parker
Joined: 23 Mar 2007
Posts: 711
Location: Raleigh, NC
PostPosted: Fri Aug 14, 2009 12:17 pm Post subject: Reply with quote
Very cool. Thanks!!
_________________
I am a basketball geek.
Back to top
View user's profile Send private message Visit poster's website
schtevie
Joined: 18 Apr 2005
Posts: 415
PostPosted: Fri Aug 14, 2009 12:35 pm Post subject: Reply with quote
Very interesting picture. My instant reaction is a question. Assuming that mid-range shots suck on average, are the Lakers taking a lot of unforced sucky shots at the beginning of games or are opposing defenses compelling these somehow by better defending the 3pt line early on?
If you have the time and inclination, could you overlay eFG% or better yet the TS% for each type of shot?
Back to top
View user's profile Send private message
Jon Nichols
Joined: 18 Aug 2005
Posts: 370
PostPosted: Fri Aug 14, 2009 12:42 pm Post subject: Reply with quote
schtevie wrote:
Very interesting picture. My instant reaction is a question. Assuming that mid-range shots suck on average, are the Lakers taking a lot of unforced sucky shots at the beginning of games or are opposing defenses compelling these somehow by better defending the 3pt line early on?
If you have the time and inclination, could you overlay eFG% or better yet the TS% for each type of shot?
I believe my next post (perhaps around Sunday) will be all about effective field goal percentage. I may do it on just the Lakers, but I'll probably look at the league as a whole.
Also, one thing to keep in mind: my term "midrange" can be a bit tricky, because it also includes post shots such as fades, hooks, and bank shots.
schtevie
Joined: 18 Apr 2005
Posts: 408
PostPosted: Fri Aug 14, 2009 1:28 pm Post subject: Reply with quote
Right now, what your numbers imply (assuming that game pace does not vary by quarter) is that were the Lakers to select shots throughout the game the way they do in the 4th quarter, that nearly four more 3 pointers would be taken (at the expense of midrange shots).
Assuming that the success rate of such a trade-off would reflect overall averages, an approximation, using data from 82games and official stats (and fiddling with the numbers to account for definitional differences between Roland's Jumpers and your combined 3 pointer and Midrange) is that, on average, an extra 0.4 to 0.5 point per game might be gained by looking more for the 3, earlier on.
Maybe there isn't a counterfactual gain but maybe there is.
I look forward to your next post.
Back to top
View user's profile Send private message
DSMok1
Joined: 05 Aug 2009
Posts: 602
Location: Where the wind comes sweeping down the plains
PostPosted: Fri Aug 14, 2009 1:46 pm Post subject: Reply with quote
I suspect that during the early part of the game the teams are feeling out the opponent's flaws on defense, and there is some disorganization--thus the high number of midrange jumpers. The central part of the game should be about what is to be expected. At the end of the game, the team that is behind will shoot more threes in a high-variance strategy (I suspect the 3pt% late in the game would be lower than during the rest of the game...) Could you calculate the FG% of each type of shot throughout the course of the game? I guess that is next on your agenda.
Back to top
View user's profile Send private message Send e-mail Visit poster's website
Crow
Joined: 20 Jan 2009
Posts: 806
PostPosted: Fri Aug 14, 2009 1:56 pm Post subject: Reply with quote
I have asked before, as an aside, if the mid-range (using the 82games data on it) is somehow more important in the playoffs, speculating that the efforts to deny the inside game and 3 point game are higher in the playoffs.
(This despite my general critique of it, to listen to the other side that Dean Oliver raised when he recommended trying to make the mid-range gamer work for the Sonics against the Spurs in the playoffs a few years ago- despite it being exactly what the defense wanted and was "allowing" or daring to "try to beat us this way". I guess they couldn't get enough going from 3 point land or inside and turned to this strategy. Not quite successful enough with it though, as the Spurs bet.)
That Dallas, San Antonio, LA and Boston are 4 of the top 6 on mid-range accuracy in regular season is indirect suggestion the insiders think it is important ability to have.
I still think Orlando is right to emphasize inside and 3 point shot frequency, maybe even to the point of being the clear "leader" in least mid-range shot frequency but being 11th worst at that aspect of the game still hurts some.
Denver and Houston (with analytic shops) are essentially in a tie with Charlotte for next least on frequency. Denver was 7th worst on mid-range accuracy, Houston closer to average at 13th.
If you could get a team low on mid-range frequency but still top 10 on mid-range accuracy and good accuracy on inside and 3 point shooting then I'd think you'd really have something that would be tough to defend and mutually helpful and quite efficient. the Lakers might be the closest to that or at least one the closest. Portland and Boston are in that group too. Cleveland wasn't far too this pack but shot to many mid-rangers to my mind. Maybe the addition of Shaq will dampen that some and get them all the way to where they want to be- playoff champion.
Back to top
View user's profile Send private message
schtevie
Joined: 18 Apr 2005
Posts: 408
PostPosted: Fri Aug 14, 2009 4:41 pm Post subject: Reply with quote
My math bad. The aforementioned 0.4 to 0.5 points were actually percentage points. So, the notional actual points forgone are double that: 0.9 to 1.0. This is not a small extra return on four shots, assuming it is real.
As for it being real, the actual Laker numbers are hopefully forthcoming, but in the meantime, I find myself unpersuaded by DSMok1's accounting for the shot trends. His description is that a game begins with a "feeling out" process at the beginning, followed by normalcy, followed by shooting more threes as a high variance strategy. I see three problems with this.
First, assuming that "feeling out" implies that the offense is probing for better shots, it somehow instead finds itself settling for presumably inferior quality mid-range shots. It seems to me that there are two options here for what is going on. Either the offense is flaking out (i.e. making errors of judgment) or it is the defense that is doing the feeling, perhaps having extra energy to better cover to the three point line, energy which dissipates as time goes on.
Second, as to the middle being "what is to be expected", this isn't really an explanation but an assumption. Either the offense is making (near) optimal choices or not, or not. We await the data.
Third, though over the season, there would certainly be a few games when the Lakers (a 65 win team) needed both to play catch-up AND where a three point strategy was the optimal one (as opposed to the normal strategy where the better team plays conservatively and still is expected to catch up) but it seems unlikely to me that this would drive the trend on a chart of seasonal averages.
Lakers' 3 point shooting had an eFG% of 0.541. My estimate/guess of Jon's mid-range season average is 0.431. To the extent that these averages hold on the margin, taking 3s is a dominant strategy throughout.
Finally, responding to Crow's observation that Dallas, San Antonio, LA and Boston are 4 of the top 6 on mid-range accuracy, what this really means, I am guessing, is less that it is important to try and cultivate a disembodied team ability at mid-range shooting and more that it is really important to have a player named Nowitski, Duncan, Gasol, or Garnett. Just because a team is good at taking a below-average shot doesn't mean it is a good shot to take (as a first option).
Back to top
View user's profile Send private message
Crow
Joined: 20 Jan 2009
Posts: 806
PostPosted: Fri Aug 14, 2009 5:13 pm Post subject: Reply with quote
I agree Schtevie that strong individual players as mid-range are probably a key part of the overall mid-range efficiency, though I don't immediately know the actual share of above average efficiency impact at team level goes to the top guy himself.
One test:
Looks like in quick rough terms the Mavs take a bit over 40 mid-range shots by 82 games and Dirk takes 15 of them. He shoots 47.4 while the rest of team shoots 42.6 from mid-range. With the average team shooting just under 40% from mid-range looks like Dirk is responsible for 62% of the team's mid-range "superiority". But the mid-range remains far worse than any other shot and 74% of all the shots Dirk took were mid-range.
Back to top
View user's profile Send private message
Jon Nichols
Joined: 18 Aug 2005
Posts: 370
PostPosted: Sat Aug 15, 2009 12:58 pm Post subject: Reply with quote
"Tracking the Efficiency of Different Shot Types at the Team Level":
http://basketball-statistics.com/blog1/ ... es-lakers/
Back to top
View user's profile Send private message AIM Address
DJE09
Joined: 05 May 2009
Posts: 148
PostPosted: Sun Aug 16, 2009 4:55 am Post subject: Reply with quote
schtevie wrote:
Lakers' 3 point shooting had an eFG% of 0.541. My estimate/guess of Jon's mid-range season average is 0.431. To the extent that these averages hold on the margin, taking 3s is a dominant strategy throughout.
I am already on record as pointing out that every team, even Philly, shot better from 3 than from long 2s (in terms of eFG%). Actually, the most interesting aspect of Jon's charting is that inside shots / finishes at the rim are a better shot option than 3s. So, maybe some of us 3pt advocates have to view some of the mid-range 2s as attempts (failed or future creation) at generating inside shots.
I would also point to the drop in 3pt scoring in DO's "Desperation Time" to the level of midrange - at least for the Lakers - which is extremely interesting given the marked increase in 3pt frequency and is again evidence that it is not simply a case of teams shooting more 3s to generate more points.
In case anyone thinks I am making pointed comments at them, I am making them at myself Smile
Back to top
View user's profile Send private message
Harold Almonte
Joined: 04 Aug 2006
Posts: 616
PostPosted: Sun Aug 16, 2009 9:17 am Post subject: Reply with quote
Nichols, you maybe should try your next step with the most extremal teams (between 3p and inside): PHI and DAL. And a closer detailed look to their crunch times maybe.
Back to top
View user's profile Send private message
schtevie
Joined: 18 Apr 2005
Posts: 408
PostPosted: Mon Aug 17, 2009 12:14 pm Post subject: Reply with quote
Jon, thanks for data. A quick question about the TS%. How exactly is this defined? Do the data show the total points associated with each type of shot attempt (i.e. made shots, made fouls on missed shots, and and-ones) such that the only scoring excluded from the chart is made foul shots on non-shot attempts and technicals?
This aside, I am confused as to why you squashed the interpretation of these data: "Midrange shots are the most common attempt throughout despite being the least efficient. I would imagine Lakers players and coaches aren’t stupid, so there must be hidden benefits to these attempts."
There is no call to speculate about the collective intelligence of this or any other team, but there is every reason to believe that NBA shot selection is not optimal, especially when it comes to three pointers. And I would like someone to come up with one plausible argument as to how there can be a benefit to a shot attempt that is expected to lead to fewer points than another.
Regarding the under-attempting of 3 pointers however, NBA history could not be clearer. What began as an "above-average below-average shot", within a few years became an above-average shot and has remained so for about a quarter of a century. Yet the arbitrage one might have expected to have quickly occurred, equalizing the returns on the margin, has not yet come to pass. A quarter of a century! There are players in the NBA whose mothers weren't old enough to have children when this all began. Shares of three point shots have increased (while efficiency was increasing for many years!) but still no equalization. In this context, I find it hard to interpret your data as implying optimality. To the contrary.
My interpretation of your data is two fold (and somewhat tentative). There are quarters one through three, and then there is the fourth. In the first three, there is essentially no relationship between three point attempts and success rate (if anything, a positive one). Once can trade off mid-range shots for superior three-pointers.
Then the fourth quarter comes about and, apparently, defenses get a bit serious and/or garbage time leads to crappy shots because it doesn't matter when the game is in hand. (Prof. Rosenbaum formally identified the qualitative differences of crunch/garbage time in the context of APM. Now, this phenomenon in and of itself is interesting in that it speaks to another, global suboptimality. If indeed point differentials matter for probability of victory, then non-crunch time effort should be higher. But this is another matter.) But even within this context, there are still positive trade-offs between mid-range shots and three-pointers.
Anyway, it will be interesting to see what other team data show. Thanks for the work.
Back to top
View user's profile Send private message
Kevin Pelton
Site Admin
Joined: 30 Dec 2004
Posts: 978
Location: Seattle
PostPosted: Mon Aug 17, 2009 1:07 pm Post subject: Reply with quote
schtevie wrote:
And I would like someone to come up with one plausible argument as to how there can be a benefit to a shot attempt that is expected to lead to fewer points than another.
Wouldn't the most elegant answer be because it spaces the floor to improve the efficiency of another shots?
Back to top
View user's profile Send private message Send e-mail Visit poster's website
Jon Nichols
Joined: 18 Aug 2005
Posts: 370
PostPosted: Mon Aug 17, 2009 2:39 pm Post subject: Reply with quote
Quote:
Jon, thanks for data. A quick question about the TS%. How exactly is this defined? Do the data show the total points associated with each type of shot attempt (i.e. made shots, made fouls on missed shots, and and-ones) such that the only scoring excluded from the chart is made foul shots on non-shot attempts and technicals?
Unfortunately, shooting fouls are again not included. The data only says "shooting foul," so I can't associate what kind of shot attempt it was. One would assume "close" shots are the most common attempts in which fouls take place, so perhaps the efficiency of those attempts is the most underrated.
Back to top
View user's profile Send private message AIM Address
Jon Nichols
Joined: 18 Aug 2005
Posts: 370
PostPosted: Mon Aug 17, 2009 2:43 pm Post subject: Reply with quote
Kevin Pelton wrote:
schtevie wrote:
And I would like someone to come up with one plausible argument as to how there can be a benefit to a shot attempt that is expected to lead to fewer points than another.
Wouldn't the most elegant answer be because it spaces the floor to improve the efficiency of another shots?
That would be my guess. For example, Dwight Howard this past year and Shaquille O'Neal in his prime created a lot of open threes through their postups. And while a team feeding them the ball will likely have more close attempts than normal, they're also more likely to have more midrange/post attempts than normal. Therefore, there may be a correlation (just my hypothesis) between the amount of attempts of midrange shots and the success rate of three-pointers. Perhaps that is something I will dive into next...
Back to top
View user's profile Send private message AIM Address
johnschuhmann
Joined: 16 Jan 2008
Posts: 25
PostPosted: Mon Aug 17, 2009 3:07 pm Post subject: Reply with quote
A thought on the higher number of threes taken in the 2nd and 4th quarters...
It's a personnel thing.
The Lakers (in the regular rotation) that had the highest ratio of 3s to 2s were:
Radmanovic: 64% of his shots were 3s (before trade)
Vujacic: 56%
Fisher: 44%
Farmar: 35%
Ariza: 32%
Of the five players above, only Fisher was a regular starter (in the regular season). And since the reserves spend most of their time on the floor in the 2nd and 4th quarters, it makes sense that the team's # of threes is higher in those quarters.
_________________
John Schuhmann, NBA.com
Back to top
View user's profile Send private message
schtevie
Joined: 18 Apr 2005
Posts: 408
PostPosted: Mon Aug 17, 2009 4:05 pm Post subject: Reply with quote
Contested mid-range jump shots are to floor spacing like....Republican Party participation is to health care reform? Basically, we are talking about an event that subverts the intent of the process.
The purpose of spacing the floor, either by coaches' design or the IQ of players on the court, is to get high quality shots. Taking a bad shot is the subversion of this goal. It is an opportunity forgone, never to return. It is doing the defense a favor. The next time down the court when the ball is in a similar position, how can the offense be considered to be better off? How will defenders be put in a weaker position for the previous bad shot having been taken?
Bad shots are bad shots.
Back to top
View user's profile Send private message
Jon Nichols
Joined: 18 Aug 2005
Posts: 370
PostPosted: Mon Aug 17, 2009 4:37 pm Post subject: Reply with quote
Schtevie, I just ran a few regressions related to what you are talking about (midrange attempts vs three point accuracy, post attempts vs three point accuracy, close attempts vs three point accuracy, etc.). Basically, I didn't find any correlations whatsoever. Of course, I'm only using one season's worth of data. Anyway, I need to go through the process of writing an article and creating a presentable format for the data, so the article should be up in a day or two.