This is the question my latest for The Huffington Post addresses. Short answer… the attitude and abilities of the Pistons are not in alignment. And this disconnect has led to anger and frustration (which is being directed at the coach).
In addition to a link to my work for The Huffington Post, let me direct everyone’s attention to the following stories:
A Few More Links
- Ty Willihnganz: Best and Worst Scorers in the NBA East (apparently scoring isn’t as hard as it appears)
- Ty Willihnganz: Contested vs. Free Rebounds in the Bucks-Bulls game (apparently many rebounds aren’t free)
- Arturo Galletti: The Rebuild & the Rankings (among many other stories, Arturo apologizes to Darryl Morey and the Rockets).
- Henry Abbott: The Knicks could use a good geek (short comment below)
Quick Comment on the Value of Corey Brewer and Kobe Bryant
Henry’s story argues that Corey Brewer may be capable of making a positive contribution to a team’s success. The post sites statistics like adjusted plus-minus. Brewer posted a 1 year adjusted plus-minus (APM) of 0.85 with the Minnesota Timberwolves this year. And his defensive measure was in the negative range (which is “good”). Then again, by the same measure Kobe Bryant has a measure of -11.15 (which has to be one of the lowest APM marks in the NBA this season). And Kobe’s defensive measure is very far in the positive range (which is “bad”).
So is Brewer “better” than Kobe this year? Well, APM is very inconsistent from season—to-season. So like the weather in the Midwest (and this is something often said in places like Nebraska), if you don’t like a person’s APM just give it some time and it will probably change.
Of course, that is a problem for decision-makers. How are they supposed to know if a result seen from APM reflects a player’s ability or is simply the “noise” in the model? If you can’t answer this question (and I don’t think you can), I don’t think you can use this information in making decisions.
Just for the record… according to the automated Wins Produced website (from Dre Alvarez), Brewer has a WP48 of -0.022 and an Adjusted P48 of 0.120. Kobe’s marks are 0.207 and 0.344. Such marks are quite consistent with what we have seen from Kobe and Brewer across each player’s respective careers (ADJ P48 has a 0.85 correlation from season-to-season; so it is quite consistent). In other words, the box score says Kobe is much better than Brewer. And this has always been true.
Now maybe Brewer’s defense is just amazing. And maybe his defensive ability puts him on par with Kobe this season. But I suspect that isn’t true. So maybe the Knicks were wise to let Brewer find employment elsewhere.
By the way…. a statistical measure doesn’t have to be consistent with conventional wisdom. In fact, whether a measure is consistent with conventional wisdom or not is not relevant to whether the measure is “good” (at least, that has always been my argument). The problem with APM is not that it is defies conventional wisdom (like Wins Produced, this is often the case). The problem is that it is quite inconsistent over time. In addition, the results are simply hard to explain (i.e. can Brewer be such a great defender that he simply overcomes all of his remaining shortcomings?). Such problems suggest that APM probably doesn’t help people make better decisions.
– DJ
P.S. One last note… Henry doesn’t think the Knicks should have used the APM results to decide whether to keep Brewer. He thinks they should use it to go look at more film. But if APM tells you a player is “good”, and you then go look at film, what are the odds you will walk out of the film session thinking the player is “good”?
Wages of Wins Journal 
brgulker
March 1, 2011
Dr. Berri, our Pistons are getting hard to cheer for…
dberri
March 1, 2011
Ben,
I have been watching the Jazz more and more. Right now they are better than the Pistons (at least, I think by a little bit). And they have some hope for the future (Derrick Favors, Paul Millsap, multiple lottery picks, etc…)
brgulker
March 1, 2011
I don’t blame you. You are missing a nice game from Stuckey at the moment though (well, apart from 3 turnovers early).
dberri
March 1, 2011
I am watching Knicks and Magic tonight on NBATV. Melo is amazing. He has made eight shots (and taken more than 20).
brgulker
March 1, 2011
Yes, but remember that he (magically, by osmosis I think) makes his teammates better. Just look at Chauncey.
Italian Stallion
March 1, 2011
Come on guys, Melo wasn’t that bad of a “throw in” in the Billups deal.
dberri
March 1, 2011
Melo “created” 24 shots tonight. People should focus on this, not the 16 misses. Without those 24 shots, the Knicks would have dribble around in circles 24 different times. So he was a key addition to this team.
Italian Stallion
March 1, 2011
dberri
You have no idea how hard it is for people to grasp that concept. I’ve written e-mails to several NY Knicks beat writers and spoken to many friends/fans over the last few weeks. It’s like talking to a wall. At this point people just ignore me (or unfollow me on twitter – lmao) because I don’t agree with what they think they already know.
All people see is the athleticism and the ability get off shots that most players wouldn’t be able to create without considering how often they go in or whether other potential alternate opportunities would have been superior.
I’m not even extreme on this issue.
If there’s 5 seconds left on the shot clock, I’d prefer Melo to have the ball in isolation over Gallinari because I think he can get off a better look and knock it down more often too. But how often does that happen?
The rest of the time I’d rather have some combination of Gallo/Amare/Fields/Turiaf/Chandler taking the shots. Melo’s extra usage often comes right out of the hides of other more efficient scorers and hurts the team. It has been obvious both visually and in the numbers that in the first four games he hurt the Knicks offense despite accumulating a lot of points, but fans don’t get it.
Greyberger
March 1, 2011
Why didn’t you use 2-year APM? People who use APM usually go for more than one year’s worth of data.
2-year APM tells an entirely different story:
Lakers
Gasol +2.4
Kobe +3.0
Odom +4.1
Artest +2.15
Bynum +1.80
Fisher -0.74
Brown -4.77
As far as I understand it the more player minutes in the dataset the better the results.
marparker
March 2, 2011
re Melo in the last 5 seconds of the shot clock,
This made me think. I wonder if teams with alot of efficient scorer let the shot clock run down.
marparker
March 2, 2011
GreyBerger,
Whats the point of a number that takes two years to start getting close to giving a complete picture?
dberri
March 2, 2011
Greyberger,
The analysis of Brewer’s defense is based on one-year data. So that is what I why I focused on one-year APM.
About two-year APM…. as you add more years to APM it is true the results become more statistically significant . But that may be because as you increase the number of observations used to estimate a model, the standard errors of the estimated coefficients will fall. So if you get a data set that is big enough, we will find most (if not all coefficients) suddenly appear to be statistically significant. This may not mean your estimates are “better”. It is just reflects how the standard errors of the coefficients are estimated.
Italian Stallion
March 2, 2011
marparker,
“I wonder if teams with a lot of efficient scorers let the shot clock run down.”
It probably depends. I doubt a D’Antoni coached team is the same as others.
To me, the real issue is that sometimes there’s plenty of time left on the clock when Melo gets the ball, but he immediately looks for his own shot to often.
He often gets it, holds it, seems to be making a decision about how to try to score for a few seconds, and then either makes a spectacular move and scores, gets fouled, or throws up a terrible shot.
Unfortunately, when you net it all out, it’s not a great result given that there was still enough time on the clock to move the ball and perhaps get an open or better look that was a superior option.
The fans and media are enamored with the difficulty of the successful attempts and excuse the misses because they were difficult shots without considering the potential for alternatives.
In addition, when he has open looks they are often long/mid range 2s that tend to be inefficient shots for everyone or 3s where he is not even an average shooter.
To be clear, I don’t hate him as a player.
He adds value to an offense and has the talent to add a lot of value, but he is overrated in general and hurting the Knicks right now despite scoring a lot of points.
DH
March 2, 2011
IS,
I’ve been a Melo “fan” for years, but not an apologist; I think your comments are pretty much right on. I would only say that they are slightly exaggerated/shaded by Melo’s recent play (which has, obviously, highlighted the flaws educated NBA observers have always pointed out). The exaggeration (coming from lots of people talking about it right now, not just here) is not surprising either, given the recent media hype machine about Melo, the massive extension, and the many quality players the Knicks gave up for him. I’m only disappointed how any value Melo brings (albeit somewhat inconsistently) gets lost because people who understand real basketball value find themselves having to discuss how overrated he is.
Which brings me to my real point — if we could get over the argument about whether certain statistical measures are useful (why the world can’t is another topic altogether), I think applying the measures to Anthony specifically provides an opportunity to discuss the more important next step — how the measures should be a tool to use in shaping a player’s development.
I may get this a little wrong, but from my occasional reading over the years, it seems that DBerri’s research is easily useful for basketball evaluators, as players tend to be pretty consistent over their careers re: WP48. But at the same time, it shows coaches rarely matter for an individual player’s improvement because, again, players’ performance stays the same across different coaches and systems. Shouldn’t we be trying to figure out why that is? I mean, that seems crazy, right?
The way I see this applying to Anthony is that most people agree that he COULD be an extremely effective player, and most of the improvement could come simply from not doing some things he does too often, like shooting the long, contested 2’s. He could also focus a little more on rebounding (which he tends to do well, at least for stretches of his career), make the easy swing pass a lot more quickly and often, and focus on improving his fundamentals on defense. All of which he’s very capable of, but somehow has been incentivized his whole life not to do. So why can’t a good coach take these things, show him how it could result in better play and more wins, and make it happen? Why is a player’s psychology about his role in a basketball game so set in stone?
The other night, George Karl, in discussing Melo’s lack of defensive focus, implied Melo was too egocentric or had too much superstar attitude to teach. Karl was talking about being excited about the new team he had, and said something like Melo had “to be managed” while the new guys could just be told what to do and they would have to do it. That seems crazy to me for Karl to believe that, especially when I know Karl had some success with coaching Anthony at different times in terms of stopping the bad % shots and rebounding more.
Anyway, that’s a really long comment, but I mostly just wanted to point out that I believe a lot of the dissonance between the so-called “general public” view on Anthony as a superstar and the advanced statistics that show he is not has a lot to do with how Anthony has all the right tools and abilities to be a superstar, and even seems like he cares about getting better when he talks about basketball. He even puts it all together on some occasions, but not consistently and sometimes does the exact opposite, like the game last night. The question is why hasn’t anyone been able to teach him, or why can’t he learn, to do it all the time? That’s the question I think is the most interesting to move towards.
Greyberger
March 2, 2011
It’s not just a bigger dataset though – it’s a better one, with fewer problems of co-linearity.
87% of the time that Kobe has been on the court, Gasol has been on the court too. APM doesn’t know what to do about these two players, since they don’t play apart from each other enough to say who is really responsible for the duo’s success.
An extra year of data doesn’t just add weight to the model, it reduces that 87% overlap and gives the model a chance to ‘solve’ who is responsible for what.
There’s a similar problem with AP M for Chicago; 90% of Derrick Rose’s minutes this year has been in the company of Luol Deng. Deng’s 1-year APM is very high and Rose’s is low.
Usually another year’s worth of data is enough to untangle problems like this. An alternative that allows you to focus on one year has been developed here: http://stats-for-the-nba.appspot.com/
dberri
March 2, 2011
Grayberger,
There are still problems with statistical significance when you look at two years of data. In fact, for five years of data it is still the case that about half the results are not statistically significant. And I repeat, more and more data will decrease the standard errors, so you are not really sure that you have “better” data.
In the end, there are real problems with using APM in decision-making. You are better off looking at the box score data.
Italian Stallion
March 2, 2011
DH,
I rarely agree with anything as much as I did with your comments.
Mike
March 4, 2011
I wonder whether APM has any biases against a particular group of players. If APM is not biased, then it will probably be considered the best measure of a player’s career performance in a decade or so once we have more data. For a single season, APM really doesn’t tell much about an individual player’s performance because the number is so noisy, but for a career that noise should cancel out.
Philip
March 5, 2011
DH, Fantastic comment.