This past weekend – in the words of Henry Abbott – Michael Lewis (author of Moneyball) brought the Moneyball story to the NBA. On the pages of New York Times Magazine, Lewis told the story of the No-Stats All-Star, Shane Battier.
As Lewis notes, “Here we have a basketball mystery: a player is widely regarded inside the N.B.A. as, at best, a replaceable cog in a machine driven by superstars. And yet every team he has ever played on has acquired some magical ability to win.”
Lewis contends that the traditional basketball metrics tell us that Battier is not an effective player. Yet Battier’s teams tend to win. To explain the magic that is Battier, Lewis turns to Daryl Morey (the young general manager of the Houston Rockets). Morey argues that by looking past the box score – to a player’s plus-minus – the value of Battier can be measured. And that measurement tells us Battier is a good player.
In sum, the traditional box score in the NBA is simply inadequate if we wish to measure a player’s value. Only the “advance” approach of plus-minus can tell us which players are “good” or “bad”.
Repeating the Story
The story Lewis tells is well-written and quite interesting. It’s also not entirely original.
Here is what Matthew Yglesias observed on Monday:
Of interest to any fan of this blog will be Michael Lewis’ long New York Times Magazine article about Shane Battier, in which Battier becomes a jumping-off point for some discussion of analytic approaches to basketball. One thing about the article that bothered me probably had nothing to do with Lewis, but the piece has been given the headline “The No-Stats All-Star.” The implication being that statistics can’t measure Battier’s important contributions.
On the contrary, as Dave Berri observed in response to a similar claim back in November 2007 if you understand the statistics correctly they say Battier is very good. Stats say that Battier is an efficient scorer with his modest number of shots, and that his net possessions numbers resulting from steals and turnovers are very good. Battier also appears to be, as best as one can tell, an excellent on-the-ball perimeter defender. This last bit really is an aspect of the game that conventional statistics don’t do a good job of capturing, but certain statistical systems-including Wins Produced-indicate that Battier is a valuable player.
As Yglesias notes, the story told by Lewis has been told before. The original author was Jason Freidman, who wrote Rocket Science: Daryl Morey Brings Hard-Core Statistical Analysis to the NBA in October of 2007.
When Friedman’s story appeared I posted the following:
What the Box Score Data Says About Shane Battier.
This post built upon these two stories:
Walking in Memphis Back to the Playoffs (September 19, 2007)
Walking in Memphis (April 5, 2007)
Each of these columns detailed Battier’s contribution to the Memphis Grizzlies and Houston Rockets. And those details – based solely on the NBA’s box score – revealed that Battier is indeed “good”.
The past few days a number of people have contacted me – via e-mail and the comments section in this forum – to comment on the Lewis story. What I said back in 2007, though, is essentially what I would say today. So my comment is going to consist of an updated version of my earlier post. After this post, though, I will link to a few more stories on Battier and Plus-Minus (and also offer a few more thoughts).
The Shane Battier Story Via the Standard Box Score View
Let’ start with a bit of background on Shane Battier. With Battier the Grizzlies averaged 48 wins from 2003-04 to 2005-06. After Battier departed, the Grizzlies became the worst team in the NBA. This suggests that Battier might have had some value.
But when we look at the box score data, it’s hard to find this value. At least, that’s what we are being told.
Let’s start with scoring, the box score statistic that is most frequently cited when discussing an NBA player. The average small forward will score 19.9 points per 48 minutes. The best Battier has ever done in his career is 17.4, and that was in his rookie season. For his career he only averages 14.5 points per 48 minutes. So Battier is a below average scorer. And for many, that makes him a below average player.
Of course there is more to the box score than just scoring. Let’s turn to a summary measure like NBA Efficiency. The average small forward will post a per 48 minutes NBA Efficiency mark of 20.3. For his career Battier’s per 48 minute mark is 16.9. So NBA Efficiency says Battier is below average also.
Okay, NBA Efficiency is too simple. Let’s look at John Hollinger’s Player Efficiency Rating, a more complicated measure of player performance. The average player has a PER mark of 15.0. For his career, though, Battier only posted a PER of 13.6. Again he is below average.
If this were all you looked at with respect to the box score, I guess you would have to conclude the box score data is pretty worthless. Clearly we need to go beyond the box score data to figure out a player’s value.
A Different Approach
Well, maybe not. Let’s take a different approach. What we could do is regress team wins on offensive and defensive efficiency. Such a regression tells us that 94% of wins are explained by a team’s efficiency marks. To put that in perspective, only about 90% of team wins in baseball are explained by runs scored and runs allowed. In other words, the link between the current stats and current wins is a bit stronger in basketball than it is in baseball.
In a moment I will return to the comparison between basketball and baseball. But for now, I want to note that from our analysis of the link between the efficiency metrics and wins we can derive the value – in wins – of each of the box score statistics. And those values are used to construct the two measures cited in The Wages of Wins – Wins Produced and Win Score.
What do we learn when we look at Battier’s Win Score? For the answer we turn to Table One.
Table One: Shane Battier’s Career
Per 48 minutes the average small forward posts a Win Score of 7.3. Except for Battier’s rookie season he has bested this average his entire career. Let me repeat this point. Win Score, a measure based entirely on box score statistics, tells us that Battier is above average.
If we delve into the numbers we can see why. First of all, although Battier doesn’t shoot much, he is an efficient scorer. And the aforementioned regression is quite clear on this point. Shooting efficiency matters in the NBA. Or to put it another way, inefficient shooting definitely hurts a team’s chances to win (the valuation of shooting efficiency by NBA Efficiency and PER is a point I have made before in more detail).
Battier’s value, though, goes beyond shooting efficiency. When we look at steals and turnovers we see another area where Battier helps. For a typical small forward, if we subtract turnovers from steals we get -1.1. In other words, typically a player will commit more turnovers than he will get steals. Battier, though, is not a typical player. Steals minus turnovers for Battier in his career is 0.2. This is a 1.3 swing in possessions for Battier in his career. It’s important to note that Battier is not just a below average producer of points (in terms of totals, not efficiency) and a below average rebounder. But his ability to hit shots efficiently, generate steals, and avoid turnovers – all stats found in the box score – tell us that he is an above average player.
As we can see in Table One, the story we tell about Battier from the box score depends on how we view the data. When we rely on scoring – or scoring dominated metrics like NBA Efficiency and PER – we see a below average player. But when we consider Battier in terms of efficiency, we see a player that is above average and a key player in the success the Grizzlies had from 2003-04 to 2005-06 (and as Morey notes, the success of the Rockets today).
Back to Baseball
In closing this discussion of Battier I want to make two more comments about baseball data. As noted, the link between current stats and current wins is a bit stronger in basketball. It’s also the case that the box score statistics in basketball have a stronger predictive power than box score data from baseball. The year-to-year correlation in Win Score per minute in basketball is 0.82. In baseball the year-to-year correlation in a metric like OPS is only 0.57 (a similar story is told for linear weights).
Of course the box score data in basketball doesn’t capture all a player does on defense. But the same charge can be made against baseball data. On-base-percentage, slugging percentage, OPS, and linear weights don’t tell us anything about a baseball player’s defense. But these various baseball metrics still tell us a great deal about a baseball player’s value.
Summarizing the Story
So here’s the story I am telling (and it is the same story I told in 2007). Box score data in basketball is at least as good – and I think it’s better – than data in baseball. The problem is that box score data in basketball is not well understood. Too often the only stat people look at is scoring. And scoring, by itself, doesn’t explain much of wins.
Metrics like NBA Efficiency consider more statistics, but this measure is dominated by a player’s scoring. Inefficient scorers can increase their NBA Efficiency value by simply taking more shots. A similar story can be told about PER. When we look at the box score statistics via the measures, again we can be misled.
Faced with this problem we are told to ignore the box score statistics. But a simpler solution is to simply heed the lesson we have learned about wins and efficiency. It’s well understood that wins are determined by offensive and defensive efficiency. If we simply take this relationship and apply it to the analysis of individual players we can see that players like Battier are truly valuable. And we can see this in the very box score statistics reported in every newspaper.
More on Battier and Plus Minus
And that is where my original post ended. Although re-runs are easy, I feel compelled to say something new. That something new will begin with the words of Carl Bialik.
Bialik – the Numbers Guy at WSJ.com – recently offered two columns on both the Michael Lewis story and the plus-minus system.
Shane Battier and the Box Score
Mark Cuban’s Surprising Player-Performance Numbers
Bialik interviewed Justin Kubatko and some economist from Southern Utah University (that would be me). Although we employ different methods, each of us told Bialik that the box score data does indicate that Battier is a good player.
Bialik also offered a few more details on plus-minus. Earlier in the month Mark Cuban – owner of the Dallas Mavericks – posted a plus-minus evaluation of a number of NBA players. Bialik sent these numbers to Roland Beech – of 82games.com – and received this reaction:
Beech told me he considered plus/minus ratings, as adjusted by regression analysis, “one of the most over-hyped player rating systems.” He pointed to the incongruous finding that Telfair is more valuable than Nowitzki, and blamed other questionable results – San Antonio Spurs point guard Tony Parker as a below-average player – on a paucity of data on top players’ performance without their star teammates. Beech also argued that there is too much noise in the system, because a player’s value is determined by coaching schemes, injuries and their assigned roles.
Let me note that the adjusted plus-minus evaluation of Sebastian Telfair does not tell us about the merits of the system. Good analysis begins with evidence and moves to a conclusion. In other words, we don’t start with a conclusion (Sebastian Telfair is not good) and move back to an evaluation of a model.
That being said, Beech does raise an important point. Adjusted plus-minus does have some serious shortcomings. Still, as Wayne Winston (the creator of this system) notes, the system can give us an evaluation of a player’s defense. So it can provide some insights.
About two years ago I made some effort to incorporate what plus-minus said about a player’s defense in calculating Wins Produced. Ty at Buck’s Diary has also made an effort to marry plus-minus analysis with a Wages of Wins metric (in Ty’s case, Win Score).
Such efforts are interesting. But they don’t appear to change the fundamental story told in The Wages of Wins. Although scorers are often celebrated by fans, the media, and the NBA, the non-scoring aspects of the game matter. And one can see this simply by looking at the entire box score and relating what is seen to wins. In sum, the NBA box score – like Shane Battier – is quite good (even if neither is very magical).
– DJ
The WoW Journal Comments Policy
Our research on the NBA was summarized HERE.
The Technical Notes at wagesofwins.com provides substantially more information on the published research behind Wins Produced and Win Score
Wins Produced, Win Score, and PAWSmin are also discussed in the following posts:
Simple Models of Player Performance
What Wins Produced Says and What It Does Not Say
Introducing PAWSmin — and a Defense of Box Score Statistics
Finally, A Guide to Evaluating Models contains useful hints on how to interpret and evaluate statistical models.
Wages of Wins Journal 
JAW
February 18, 2009
Some comments that came up in reading your attempt to incorporate +/- into defense
1) fielding is just less of a part of baseball because of all of the events that don’t relate to fielding. Even so, we observe that Derek Jeter goes from being a great player to a merely good one because of his defense. Just because it doesn’t change our evaluation much, doesn’t mean it shouldn’t. In basketball, defense is 50% of the game, even though on the ball defense is not.
2) Your regression will capture much about a player’s on the ball defense in looking at their other stats I imagine. The Shane Battiers are the exceptions to this, but it doesn’t mean that it’s not better specified using +/- (even though the simplicity of being able to use the box score is very helpful for historical comparisons. Even if the model predicts 95% of wins, we also care about how well it does in predicting how teams will fare when they switch.
3) Again, might we not see vast differences among players in on-the-ball defense not because there aren’t large differences, but because they’re largely correlated with other skills. It seems to me that Battier being a slight outlier in your methodology and the +/- system (where you rank him as good, and they rank him as very good) suggests that in individual instances we learn more.
Clipfan
February 19, 2009
Dave, your comparison to baseball does not work because defense (pitching) can be measured with near the same precision as offense. Defense (fielding) in baseball is similar but is only at most 20% of the total picture of winning while defense in basketball is around 50% of the total picture of winning. It is pretty hard for a basketball boxscore, no matter how well analyzed to ignore 50% of the picture of winning and be accurate.
Lior
February 19, 2009
Clipfan: as JAW notes, it is quite possible that basketball defense ability correlates with boxscore measures. Regressing the box-score against efficiency differential would then capture defensive value as well. This is not exactly what Berri et al do, but it’s pretty close.
In fact, at the team level their record is downright amazing. Knowing only the total boxscore for the team over the season (no individual game or individual player data!), they can give a very good prediction of how many wins the team had. With a bit of extra team-level but league-wide information (the “team defense adjustment”) they can be even more accurate.
The conclusion must be that the boxscore numbers do correlate with defense. Indeed, you can start from the observation that 1 defensive rebound equals 1 defensive stop. It’s not obvious that the player who captured the rebound is most responsible for the stop, but it is enough that there is some correlation for this to work.
Michael
February 19, 2009
Sorry to refer back to your previous post but I think it is strange that you refer to Kobe as “quite good”, yet here you state that Battier is “good.”
If Battier is “good” then Kobe is clearly much much better than quite good!
Horsecow
February 19, 2009
A lot of the discussion has been about whether Battier is “good” or not, but a more interesting part of the article dealt with how Morey gives Battier detailed statistical scouting of opposing players — zones on the court where Kobe Bryant shoots well and less well. This info could presumably help any defender, not just Battier, and doesn’t really enter into the “value” debate.
There was also an interesting comment from Morey about how fouling is “the worst outcome” and how you can tell the statistically-oriented teams by two things: lots of corner threes and avoiding fouls.
Maybe we should divide the use of stats into two phases: 1) figuring out who to hire (evaluation) and 2) using stats to advise players in game (strategy). There’s less conflict there than people seem to think.
Daniel
February 19, 2009
“Clearly”. There’s an objective measurement. Superlatives do not facts make.
Alien Human Hybrid
February 19, 2009
Dave,
The issue is not merely that Battier is underrated by the box score, or that he is underrated by most people watching the game. The more controversial, central point of the NY Times Mag article is that Battier is one of the best players in the game. Not simply above average, but an all-star. In effect, he is the Anti-Answer, Michael Lewis’ Allen Iverson.
Neither win score nor general public perception supports that assertion. This is largely why it is an interesting article.
Horsecow
February 19, 2009
@Alien Human Hybrid – where does Lewis say he’s “one of the best players in the game”? Undervalued, yes, does positive things that don’t show up in the box score, yes, but not “an all-star”. The article was mostly about Battier’s defense, anyway, and in that respect, yes, Battier is an “all-star.” He made the all-NBA defensive 2nd team last year, so he’s a top 10 defender.
kevin
February 19, 2009
One comment I want to mention in the article. Someone asked Morey if there were any statisitcs that could be interpreted as lacking any selfish bias. He thought for a moment and said “offensive rebounding”, then took it back when he considered the fact some players might use that as an excuse for not getting back on defense.
I can think of one, though it is not kept as an official stastic: fouls drawn. I find it peculiar that the number of fouls a player commits is recorded, but not the number that a player is committed against. You can sort of infer it from the number of foul shots a player takes but that’s an imperfect measure.
to add to that, statistics should be kept that indicate how many fouls a player draws prior to the penaly and after the penalty. I think the results would be fascinating.
Tball
February 19, 2009
Clipfan,
The box score doesn’t ignore 50% of the picture. The box score captures steals, blocks and rebounds, all of which prevent the other team’s scoring, as well as fouls and turnovers, which can contribute to the other team scoring. The box score does not capture the opposing FG% resulting from one defender’s man-to-man defense, but it does capture opposing FG% from the team’s defense (and how many defenders are really left out on an island).
You can make an argument about which aspects of the game are ignored by the box score and how significant they are, but I think 50% is missing the boat.
Michael
February 19, 2009
“‘Clearly’. There’s an objective measurement. Superlatives do not facts make.”
Daniel, asinine and grammatically dubious objections aside, after 41 games Kobe was 14th in the league with a 0.237 wp48 and 7.4 wins produced. Shane Battier was 157th, with a wp48 of 0.107 and 1.3 wins produced. So ‘clearly’, by Professor Berri’s own statistics, if Battier is ‘good’, Kobe is much better than ‘quite good’.
todd2
February 19, 2009
Rodney McCray received the same kind of attention for his role with the Rockets back in the day. How does he compare?
kevin
February 19, 2009
Boxscores do an OK job if you’re tru9ng to figure things out on a team basis. Where it breaks down is when you try to parse who the key players were who won and lost the game.
Tom S
February 19, 2009
Dave,
While it’s true that your system rates Battier more highly than any of the others available to the public (Hollinger ect.), I feel that you’ve downplayed the degree of disagreement between yourself and Lewis/Morey on Battier in general and the explanatory power of box score statistics in general. While it’s true that winscore has always ranked Battier above average, Morey’s contention is that Battier is elite, and his greatness is evident only in non-boxscore stats (tipping rebounds to teammates, forcing opponents to take low percentage shots, etc).
If these stats (non-box score) were to become available to the public, would you find room for them in your system? And if so, how?
Tom S
February 19, 2009
“I feel that you’ve downplayed the degree of disagreement between yourself and Lewis/Morey on Battier in general and the explanatory power of box score statistics in general.”
I meant boxscore statistics in particular.
Unfortunate typo, I thought that was a solid comment otherwise.
Westy
February 19, 2009
I agree with Tom S and clipfan to some extent. I think the point here is twofold:
1. Yes, we can tell Battier is decent based only on the box score.
2. However, the box score is still lacking information in giving us good data on who are the best defenders. In the Bill Simmons podcast with John Hollinger I think they make the point well that the NBA could help make its box scores better by collecting better information on defense.
I think that the fact Battier is ranked as above average in WP does not mean it completely evaluates him correctly. I think the case can be made that he’s being punished for being below average in rebounding even though he’s responsible for some of those rebounds being available.
Alien Human Hybrid
February 19, 2009
@ Horescow
The title of the article is “The No Stats All-Star”. The author supports this notion throughout the article. Here is one quote- there are others:
In its crude form, plus-minus is hardly perfect: a player who finds himself on the same team with the world’s four best basketball players, and who plays only when they do, will have a plus-minus that looks pretty good, even if it says little about his play. Morey says that he and his staff can adjust for these potential distortions — though he is coy about how they do it — and render plus-minus a useful measure of a player’s effect on a basketball game. A good player might be a plus 3 — that is, his team averages 3 points more per game than its opponent when he is on the floor. In his best season, the superstar point guard Steve Nash was a plus 14.5. At the time of the Lakers game, Battier was a plus 10, which put him in the company of Dwight Howard and Kevin Garnett, both perennial All-Stars. For his career he’s a plus 6. “Plus 6 is enormous,” Morey says. “It’s the difference between 41 wins and 60 wins.” He names a few other players who were a plus 6 last season: Vince Carter, Carmelo Anthony, Tracy McGrady.”
Alien Human Hybrid
February 19, 2009
I also agree with Tom S and Westy that the good professor is downplaying the difference in their respective evaluations. This may be entirely unintentional, and I definitely give him the benefit of the doubt.
Clipfan
February 19, 2009
Lior, I think Dave is claiming that the boxscore can reasonably evaluate _individual player_ contribution to winning. This is not the case except for offense. At a team level of course a boxscore can do well for defense. But for individuals it will miss about 50% of the picture and this makes it very difficult to trust when making an evaluation of an individual player’s contribution to winning.
Demetrius Mcmillian
February 20, 2009
Against an average NBA team, what winning % would a team of 5 Battiers have? 5 Kobes? 5 LeBrons? 5 Marcus Cambys?
Jason E
February 21, 2009
“Against an average NBA team, what winning % would a team of 5 Battiers have? 5 Kobes? 5 LeBrons? 5 Marcus Cambys?”
This is, of course, something that we will never no, so it’s pure speculation. They’d all have undefined winning percentages (what happens when you divide by zero) since it’s an imaginary scenario.
If it is presented to show that the WP model is flawed, it’s tired material. A model performs best when used to evaluate within the realm of the observations made to help create it. It performs best within the scope of reality. Since the “5 Kobes” or “5 Battiers” lineup has never been observed, it is outside of what can be addressed by the model. The thought experiment conclusions are totally irrelevant.
My gut is reaction, and probably most people’s, is that the all Battier or all Camby lineup wouldn’t fair well, but that’s completely consistent with an assumption of the model: position matters. Dave was fairly explicit in this when presenting the model, that it would break down if you stuck centers in to do the job of point guards. In such a case, we can hypothesize that the squad with the most flexibility in the ‘position abilities’ of the players would do better, but that’s still a hypothesis. Some of the success that “small ball” lineups have enjoyed suggests that the basic 5-position lineup may not necessarily be required and, if so, the position adjustment may not function correctly in all instances, but that’s another study to be done altogether.
John W. Davis
February 21, 2009
I love Battier. He needs to come home and play for the Pistons when he is a free agent!
Alien Human Hybrid
February 24, 2009
Jason E–
I think the “same player at every position” question is an interesting issue, if for no other reason than it may give you an opportunity to see the value of a position player in a slightly different way.
Think of it this way- what if you fielded a team of Lamar Odoms? A team who at every position could rebound, shoot, dribble and distribute? Given that Lamar is somewhat undervalued in NBA circles relative to his WP (WP 48 of .187 .253 in 07 and 08 respectively) this could be a very interesting and informative exercise.
Westy
February 25, 2009
Kevin Pelton does a great job of providing a synopsis on this subject here.
He notes,
“Certainly, using advanced metrics based on box-score stats gets us a lot closer to Battier’s true value than looking at his per-game numbers, but it doesn’t get us all the way there. This makes sense, since so much of what Battier does in terms of making shots more difficult for players like Kobe Bryant and drawing charges isn’t reflected in the box score. We know this in part because these numbers are occasionally tracked, and Battier’s are impressive indeed…
“Even the ability to have this discussion represents the advance of APBRmetrics. When Battier entered the league seven years ago, no one had even conceived of adjusted plus-minus. No matter what Morey thinks of it, the box score is here to stay. However, it is misleading people less and less as we devise better ways to tease value out of it, track more things that once went uncounted and advance methods like adjusted plus-minus that operate entirely independent of individual statistics. The ultimate beneficiaries have been the world’s Shane Battiers, and deservedly so.”
Paul
February 26, 2009
I’ll join the other posters who suggest that Professor Berri’s assessment of Battier doesn’t come close to agreeing with Lewis’.
Win score ranks Battier as the 7th best payer on the rockets this season, and 157th in the league. Lewis seems to be suggesting he is an all-star level talent, and therefore that he ranks somewhere in the top 24-30 players.
That’s an order of magnitude difference, and represents a fundamental disagreement with Lewis’ thesis, not a demonstration that his conclusions can be replicated using box score analysis.