Team Strengths and Weaknesses
by Madison Dan on Jul 16, 2011 1:17 PM CDT
Email
PrintIn my deep dive into Basketball-Reference.com, I got tired of looking at player data and decided to dig into the team statistics. It led me to an investigation of the factors that lead to wins. To some extent, I’m just reinventing the Wins Produced wheel, but building it up myself provided me with insights I have a hard time getting from other people’s work.
What I will show are the areas in which the Wolves (and other teams) distinguish themselves. While the resulting recommendations (shoot better from 2-point land, improve defense, reduce turnovers) may be somewhat obvious, I found it interesting to break the Wolves down in this level of detail. I also include a link to a Google Docs spreadsheet that allows you to see a chart for every NBA team breaking down their relative strengths and weaknesses. Details after the jump…
Using data for the last five years (150 team seasons), I estimated a regression model of wins as a function of a number of factors, including:
- Share of field goal attempts that are 3 pointers (3PTShare)
- Free-throw attempts per game (FTA/G)
- 2-point field goal shooting percentage (2PT%)
- 3-point field goal shooting percentage (3PT%)
- Free-throw shooting percentage (FT%)
- Offensive rebounds per game, pace adjusted (ORB)
- Defensive rebounds per game, pace adjusted (DRB)
- Assists per game, pace adjusted (AST)
- Steals per game, pace adjusted (STL)
- Blocks per game, pace adjusted (BLK)
- Turnovers per game, pace adjusted (TOV)
- Pace, in possessions per 48 minutes
- Defensive rating, or the number of points allowed per 100 possessions (DRtg)
While I wanted to only include statistics that tied to specific aspects of play, I ended up needing to include DRtg to more fully explain defensive performance. That is, DRB, STL and BLK only explain about a third of DRtg, so separately including DRtg noticeably improves the explanatory power of the model. (It reduces the average absolute error from about 3 games to about 2 games.) In contrast, Offensive Rating (in points per 100 possessions) did not add anything to the more descriptive offensive statistics already included in the model.
Before I get to the statistical model, let’s look at where the Wolves have stood over the last five seasons. The table below shows the Wolves’ NBA ranking for each factor during each of the past five seasons. There are a few areas in which the Wolves stood out in 2010-11, and not always in a good way. They had the fastest pace in the league (which doesn’t end up being a good thing), were third in offensive rebounding, and fifth in 3-point shooting percentage. On the other end of the spectrum, they were 29th in 2-point shooting percentage, 30th in turnovers, 28th in assists, and 27th in defensive rating.
Just these simple rankings tell a lot of the story, but a statistical analysis can help determine the importance of each of these factors in winning games. I did this by estimating an Ordinary Least Squares model with number of wins as the dependent variable and the variables listed above as the explanatory variables. The resulting coefficients are estimates of the change in wins due to a one-unit change in each factor. I calculated an “importance” score for each factor by multiplying the estimated coefficient by the standard deviation of the factor values, or the number of wins a team could expect from improving by a “substantial” amount (one standard deviation) in a particular factor, all else equal. (The “all else equal” part is a little tricky. For example, DRB is a big part of DRtg, but the importance scores don’t reflect that, effectively making DRB look less important than it is.) The figure below shows the importance scores (click on it to make it bigger).
As the figure shows, 2PT%, DRtg, ORB, and TOV are the most important factors. Most of this does not bode well for the Wolves. AST, STL, and BLK are not very important factors in winning.
I next developed a set of charts to demonstrate what sets each team apart from a .500 team. That is, the model predicts that a team with average values for every factor will win 41 games. Deviations from the average value contribute to a team being better or worse than average. The figure below shows the resulting figure for the Wolves. (The vertical axis is the number of wins gained or lost by being different from the average value over the last five years.)
Not surprisingly, the Wolves are hurt most by three factors: low 2PT%, high TOV, and high DRtg. That is, they miss a lot of shots inside the arc, they turn the ball over a lot, and they don't play much defense. Note that the high pace of play has cost them about 2 wins. They actually pick up some wins by shooting well from 3-point land and gathering a lot of offensive rebounds.
There is an additional "factor" on the far right of the figure called "unexplained." This is the difference between the actual number of wins and the number of wins predicted by the model. The model predicts that the Wolves would have won 22.2 games with their statistical profile. This is consistent with Pythagorean Wins (23.9), Wins Produced (22.8), and Win Shares (24.9). In other words, the Wolves were bad in a way that a variety of statistical models can't explain. It could be due to bad luck, bad late-game coaching, late-game boneheaded play, etc.
I've made one of these for every team using 2010-11 statistics. It can be found in this Google Docs spreadsheet. (I had to re-format the figure to conform to the limitations of Google Docs, which are very annoying).
In the spreadsheet, use the highlighted cell as a drop-down list to select the team. Here's a short summary of some interesting (mostly obvious) things I found:
- The Cavs were bad across the board, which is not surprising.
- The Mavs and the Heat were good at both offense and defense.
- The Bulls were great at defense, but average at offense.
- The Thunder were quite good on offense (particularly free throws), but average defensively.
- The Bucks were just what you'd expect from a Skiles / Jennings team: inefficient on offense and great on defense.
I find it interesting to compare across teams, as the figures provide a handy snapshot of what they do well and poorly. I hope you have some fun with it as well.
65 comments
|
12 recs |
Do you like this story?
Comments
I tried to keep this as non-technical as I could,
given the nature of what I did. If you want to see the regression model and data, it’s on the Calcs sheet of the Google Docs spreadsheet. I would rather have used rebounding percentages, etc. (rather than the pace-adjusted per game values), but BR doesn’t have them all at the team level, and I didn’t want to keep digging for other sources.
Excellent stuff
Some thoughts: Could you have added some more specific defensive stats in order to avoid having to use DRTG and still get as good (or almost as good) a model? I’m thinking of opponent FTA/Gm, perhaps opponent 3PTShare, and others that might give us a more granular look at where the Wolves defensive breakdowns were most apparent. (Wolves had the 3rd most FTAs against, and the most 3ptAttempts against in the league last season).
With more data (which might or might not be freely available) we might be able to get more granularity on some of the glaring weaknesses. Is the horrific 2pt% a function of shot selection? Did the Wolves take more 20 foot jumpers (worst shot in the game) than the average, or were they simply worse at converting from everywhere inside the arc than the rest of the league?
This might give us more insight into how much of a factor coaching vs. talent was in the results. If it is a matter of shot selection, it might be more of a coaching issue, whereas a failure to convert at a reasonable level from anywhere suggests more of a lack of talent. Or not; one doesn’t want to go too far in assumptions. The lack of talent remains apparent either way.
Similarly, one wonders if there’s a way to correlate turnovers with poor defense. Were easy shots for the opponent as a result of Wolves turnovers a significant factor in the poor DRTG, high FTA against, etc?
This is all just ruminating, by the way, not a suggestion that you should go get these answers for me. You have done a lot of good work here. I get the feeling that a reduction in turnovers down to the league average is the single biggest thing they could do to increase wins. A back of the envelope (not even) calculation suggests that a league average number of TOs, all else being equal, might have resulted in roughly an extra 250 points scored, or about 3 extra points per game. That is an enormous number (though perhaps slightly exaggerated; fewer TOs would have probably meant a slower pace—still, it would have been significant). And such a reduction could only have helped defensively as well.
The Wolves are like the worst meal you've ever had--terrible while you're eating it and even worse later.
by Eric in Madison on Jul 16, 2011 2:09 PM CDT reply actions
Thanks
BR does have some more (but not a ton more) granularity on defensive stats, which I’ll probably look into in a following iteration. This is already about the 10th version I’ve played around with, so I got a little tired of playing around with it and just wanted to get something out there.
A bit more
BR has opponent’s eFG%, TOV%, ORB%, and FT/FGA. If I included these in the model instead of DRtg, they appear to do just as well as DRtg did in terms of overall explanatory power. (The R-squared on these models is around 0.95.)
The importance scores and Wolves deviation from mean values (respectively) are:
Opp eFG%: 5.7, -5.6
Opp TOV%: 2.5, -1.0
Opp ORB%: 1.3, 0.2
Opp FT/FGA: 1.8, -1.7
So the Wolves main defensive problem is, somewhat predictably, that they allow the opponent to shoot too efficiently. Getting deeper for this, or for the 2PT% part you mentioned earlier, requires data that I haven’t seen. I agree that it’d be interesting to break it down even further, though.
likely due to poor rotations and communications on defense
since the wolves (under rambis) crowded the paint, it made it easy for teams to drive, kick and swing to the open guy. The wolves just couldn’t recover and an opponent would get a wide open shot from a location they liked. Too much help D for no purpose or too early, thus allowing for the defense to be broken down.
No one is getting Rubio's rights unless they pry them from our cold dead fingers.
by TheEvilProfessor on Jul 17, 2011 12:12 PM CDT up reply actions
I found some interesting stuff on Hoopdata
which shows attempts, FG%, etc. by location: at the rim, 3-9 feet, 10-15 feet, 16-23 feet, and 3 pointers. The first interesting tidbit (from the 2010-11 season) is that 20 foot jump shots are not the worst shot in the game. Here are the FG% by location:
At the rim = 0.641
3 to 9 = 0.390
10 to 15 = 0.391
16 to 23 = 0.394
Now guess which shot the Wolves take more of than anyone in the NBA? Yup, they are more than 2 SDs above the mean on share of 2 -pointers from 3 to 9 feet. On those shots, they rank 21st on shooting percentage.
Here are the Wolves rankings of share of attempts and shooting percentage by location:
At the rim = 27, 26
3 to 9 = 1, 21
10 to 15 = 19, 20
16 to 23 = 21, 23
So the Wolves shot bad from most everywhere inside the arc, but took most of their shots from the worst spot possible. Anyone else thinking of Darko’s hook shots?
Do you have an opinion about coaching vs. talent given this? I guess the bad-from-everywhere part (except 3-pointers) suggests bad talent.
Even more worrying to me
is that in the most efficient FGA shot in the game, shots at the rim, the Wolves were 27th in the league in attempts.
Hopefully getting out on the break more will help with that. And, yes, I suspect that less baby hooks from Darko and Pekovic will go a long way towards fixing the 3 to 9 foot shots problem.
I am completely shocked that there is no appreciable difference between 3-9 feet and 20 footers
I am not shocked that the Wolves couldn’t throw it in the ocean from anywhere.
The Wolves are like the worst meal you've ever had--terrible while you're eating it and even worse later.
by Eric in Madison on Jul 17, 2011 4:42 PM CDT up reply actions
It's probably selection driven by two factors
1. the players taking the 3 to 9 foot shots are big men that are, on average, worse shooters, than the guards and SFs that take longer jump shots.
2. many of the 3 to 9 foot shots are failed drives to the basket (to attempt those juicy efficient shots at the basket. So the 3 to 9 foot shots are often not be design.
I haven't written an insightful post about beer.
Those both sound plausible.
I’ve been playing with some Hoopdata numbers and I think the results are worth a separate FanPost. They might help sort out some of this stuff. Basically, I’ll compare our beloved Wolves to average NBA players at their position, where the comparison is based on shot selection and shooting percentage by location. I just need to find the time to write it, which may be in the next week.
From what I’ve looked at so far, I’ve drawn a few conclusions:
1. Kevin Love is not very good at the rim (or anywhere except the 3-point line).
2. Beasley’s shooting percentages are OK (but not great), it’s his shot selection that needs fixing.
3. Ridnour has a really nice mid-range game.
4. Our guards and wings are terrible at getting to the rim and finishing when they get there.
5. Our team is not very good.
yeah
I looked at some Hoopdata stuff last year and Love is horrible at the rim. Not a surprise really. He hits the three well because he’s open. Hopefully in a better offense he’ll hit more jump shots because they’ll be less often contested.
I haven't written an insightful post about beer.
by littleboxes on Jul 18, 2011 11:26 AM CDT up reply actions
Free throws would be a factor here.
If a player fails to make the shot but gets two free throws and can shoot a high percentage of those it can make a difference. This is maybe Loves second best asset ahead of his 3 pt shots and passing.
That's true.
Love gets to the line at a good rate and is a very good free-throw shooter.
by Madison Dan on Jul 18, 2011 12:41 PM CDT up reply actions
love is one of most overrated players in nba
by jadedeed2327 on Jul 18, 2011 2:33 PM CDT up reply actions
If you're going to make a claim like that
you should at least try to say why you think it’s true. You have no idea how bad the Wolves would have been if Love hadn’t played for them. There are several different statistical methods that indicate that what Love does on the court IS really important and leads to wins, and your one-liners are not very convincing rebuttals.
by Madison Dan on Jul 18, 2011 2:38 PM CDT up reply actions 1 recs
i mean for all the stats and double doubles, he led them to the worst record in nba
by jadedeed2327 on Jul 19, 2011 1:17 PM CDT up reply actions
5 rep points for the first time you use the text box in addition to the title bar
No one is getting Rubio's rights unless they pry them from our cold dead fingers.
by TheEvilProfessor on Jul 20, 2011 8:55 AM CDT up reply actions
actually
I think I do have a cousin who is a cop…as well as some distant 2nd or third cousins.
No one is getting Rubio's rights unless they pry them from our cold dead fingers.
by TheEvilProfessor on Jul 20, 2011 10:52 AM CDT up reply actions
It's not on Darko
At least not fully. When you break down the TWolves player’s stats on Hoopdata, Darko is at (or slightly above) league average at 39.1% for the 3-9 ft range. In fact the Wolves players accounting for the majority of these close shots are all just slightly above the league average. The problem is the glut of players who combined for 6 attempts per game and were all way below the league average, even though none of them individually took nearly as many close shots as Darko, Love or Beasley. Pin it on Flynn, Telfair, Tolliver, Koufos, Wes, Lazar, Wayne, and especially Brewer. None of our other guys were good enough to compensate for how bad those guys were inside.
But the shot distribution is on Darko,
I looked at a list of centers, and Darko shot the highest percentage of any of them from 3 to 9 feet, and his share of shots at the rim was 20th highest of 33 I narrowed it down to.
But the shot distribution is on Darko,
I looked at a list of centers, and Darko shot the highest percentage of any of them from 3 to 9 feet, and his share of shots at the rim was 20th highest of 33 I narrowed it down to.
Well, the data looks like there are two places to shoot the ball
At the rim, and everywhere else. (OK, beyond the arc also, so three places). Darko playing soft is absolutely responsible for why we had so few good shots (at the rim) as compared to bad shots (everywhere else). However, I don’t think he takes the blame for why we shot so poorly away from the rim. He’s not the primary reason why the Wolves were 21st in FG% at that 3-9 ft. range. Certainly not the way you can attribute a huge proportion of OKCs struggles from 3-9 on Westbrook and Green.
It's interesting to
me that Beasley’s offense is actually spread pretty evenly amongst all of those categories.
3.7 attempts at the rim
3.0 attempts 3-9 feet
2.6 attempts 10-15 feet
5.5 attempts 16-23 feet
2.2 attempts 3-pointers
The only one that’s a significant amount higher than the others is 16-23 foot shots, and even then, it’s not that many more attempts. If he would just take a step back and shoot 1 more 3-pointer and 1 less 16-23 foot jumper, they’d all be close to 3 a game.
Those percentages are not much higher than 3pt percentages.
Other than at the rim it would seem wise to pass out to beyond the arc as much as possible. Again, other teams took more 3 pt shots than we did. considering 3 pt shooting is kind of a strength for us why didn’t Rambis insist on trying to take more?
Yeah, I guess that's right.
On average, everything between the rim and the 3-point line is a bad bet.
by Madison Dan on Jul 18, 2011 12:35 PM CDT up reply actions
You mean in the Google spreadsheet?
If so, there’s a yellow cell, B1, in which you can either type the three-letter abbreviation of the team you want to look at, or you can click on the cell and an arrow will show up just to the upper right of the cell. Clicking on the arrow will reveal a drop-down list from which you can select a team.
I hope that helps.
Can you tell me more about what you see?
Can you see the figure but not make changes? This is my first try with Google Docs, so I may need to change a setting. Has anyone else been able to make it work?
by Madison Dan on Jul 17, 2011 7:39 AM CDT via mobile up reply actions
I just found a setting that may do the trick
before people could only view, but now people should be able to edit it. I would have thought that people could have changed the drop-down selection even in the “no edit” case, but I guess not.
Thanks for the info & effort you put into it MD!
"It's tough to make predictions, especially about the future." -- Yogi Berra
just 20 gallons?
or are their some pints involved too?
No one is getting Rubio's rights unless they pry them from our cold dead fingers.
by TheEvilProfessor on Jul 17, 2011 12:13 PM CDT up reply actions
One handy way to thumbnail sketch a team's defense would be
the percentage of opposing field goals that are assisted. Somehow I’ve always thought that was a plausible quick glance at a team’s defensive mojo.
The Wolves are terrible at that
(second-worst in the NBA last year), but the stat isn’t that highly correlated with Hoopdata’s measure of defensive efficiency (0.28). In contrast, the correlation for opponent’s TS% is 0.93.
Yeah, it's just a very accessible quick stat that seems to scale pretty well.
“Defensive Efficiency” means a big ball of wax. At multiple scales, though – in a quarter, in a game, in a whole season – asking “How often does the opponent get a dime on baskets?” gives you some idea whether other teams are cutting you up with a pass or two.
(Hoopdata keeps this one handy both offensively and defensively, if I remember right.)
It definitely works in the right direction (by the eyeball test)
so it’s a pretty good single indicator. But there are some very good defensive teams that aren’t exceptional with the opponent’s assist percentage (e.g., Orlando, Milwaukee), so it’s not perfect.
by Madison Dan on Jul 17, 2011 10:05 AM CDT up reply actions
2 pts. either way
Is it better to have a high or low dime per point ratio?
What defensive function is being measured, help defense? BBall IQ?
I don’t see much there myself.
by WinTheLottery on Jul 18, 2011 8:26 AM CDT up reply actions
When you factor in Free throws I wonder how much difference that would make.
What percent of missed shots at 3-9 feet end up drawing fouls for instance.
But if they draw a foul,
they don’t count it as an attempt (unless the shot is made), right?
by Madison Dan on Jul 18, 2011 12:46 PM CDT up reply actions
Still though...
fouled attempts that end in (on average) 1.53 points aren’t counted in made shots. I don’t know if it’d make a difference, but I’d be curious to see the points per attempt (with fouled attempts, and the resulting free throws factored in) from every range, given that not many 20 foot jump shots end with a player getting to the line.
True but it might still make taking the shot worthwhile because you get points
so lets say you make 40 pct on a shot but draw a foull on about 1 in 4 attempts so lets say on 100 attempts (25 wont get counted but result in 50 free throws and 40 points lets say) and you make 30 of the 75 shots thats ( 40 pct ). the net is 30×2=60 and 40 which equals 100. 100 points on 100 shots not good but then lets say you get fouled much less at 20 feet out (say 10%) and still shoot 40pct. now 90 shots are counted resulting in 72 points and 10 result in 20 free throws for .8×20=16 points. this would produce only 88 points.
So what I'm saying is that for someone like Love it may not be so bad to attempt
that 3-9 footer because he does draw fouls often enough to make it worth while.
I don't have data on foul rates by location
but I have a hard time seeing how the foul rate from 3-9 feet would be so high as to make it a better than an at-the-rim-shot, which carries a much higher shooting percentage on average (e.g., 66% vs. 41% for PFs). You’ve made a case relative to longer jump shots, which is possible, but I don’t think it changes the idea that the Wolves need to be better at getting to the rim and converting when they get there.
Agreed on that. yes I was only trying to make the case to take a 5 footer over a 20 footer.
However, I can see the advantage other teams use of passing the ball back out to hit a three or to draw players away from the rim so someone can sneak in and get a dunk. Then like you said there is the frustration of the player who could get a dunk but simply doesn’t do it.
Plus you might add if a player gets to the rim he is also going to draw fouls.
probably more of these would result in 3 point plays then any other two point shot.
In summation it would seem to be
1.get to the rim , 2 take a 3 point shot, 3 (only by a small margin over 4) take a short 3-9 footer and 4 settle for a long 2 pointer. and the Twolves take too much of option 3 and Beasley option 4. but for those who do not draw fouls and shoot poorly near the rim option 3 might be the worst. How does that sound?
Excellent post.
I would guess that the “unexplained” factor has to do with inexperience in late game situations (from both the players and the coach).
Comments For This Post Are Closed
Secondary Sidebar
User Tools
NBA Playoffs Content Module
FanPosts
Community blog posts and discussion.
Recommended FanPosts
Recent FanPosts
In Association With
Ball Don't Lie - NBA - Yahoo! Sport
- Create-a-Caption: Jeff Pendergraph scares his Pacers teammates
- James Harden is a cake now
- Mike Miller loses shoe, plays two possessions without it, Pacers don’t score (VIDEO)
- Larry Bird calls his Indiana Pacers ‘soft’ following the team’s physical Game 5 loss to Miami
- Dexter Pittman’s elbow to Lance Stephenson’s throat is just one of three nasty hits in Miami/Indiana’s Game 5 (VIDEO)
