One of my favorite sit-coms from the 1970s is the Bob Newhart show. Last night my wife and I put in the DVD from the fourth season of this show and watched an episode titled: The Duke of Dunk.
Here is how this episode was described at allmovie.com: It seems that even before Dennis Rodman, the Chicago basketball world was plagued with showboaters who thought that they were the center of the universe. In this episode of The Bob Newhart Show, Anthony Costello is cast as Dwayne Granger, aka “The Duke of Dunk,” a star basketball player. Upset that Dwayne’s monumental ego has resulted in a drop in morale and a long losing streak, the team’s coach goes to Bob for help. Written by Douglas Arango and Phil Doran, “Duke of Dunk” originally aired on January 31, 1976.
The synopsis at allmovie.com chose to compare the character of Dwayne Granger to Dennis Rodman. A more apt comparison, though, is any scorer who dominates the ball. To this point, let me offer a bit more information on this episode.
The Duke of Dunk in Therapy
The story begins with Bob and his friend Jerry going to a basketball game and walking away amazed by the exploits of the “Duke of Dunk”. In discussing Granger’s exploits the next day at the office, Carol – Bob and Jerry’s receptionist – observes that despite Granger’s scoring the team still lost. Bob and Jerry insist Carol doesn’t know basketball. But Carol argues Granger is nothing more than a “hot dog”. Furthermore, the great Celtic teams of the 1950s and 1960s were not about just one player. Carol then proceeded to name all the players who helped Bill Russell win all those championships (yes, Carol understood the Wisdom of Red Auerbach).
Although Bob and Jerry scoff at Carol, we later learn that Granger’s coach essentially reaches the same conclusion. Soon Granger is lying down on Bob’s couch, as Bob tries and figures out why Granger isn’t much of a team player. In the course of their conversations we learn that Granger is clearly motivated by his own scoring totals. He is also so focused on himself and his stats (i.e. scoring) that he doesn’t know the names of his teammates.
Obviously this story – told more than 30 years ago – has a Wages of Wins theme. The Duke of Dunk is obsessed with his own scoring numbers. It’s these numbers that have made him into a “millionaire athlete” and he sees no reason why he should change his ways. After all, scoring got him fame, attention, and money. So why should he care about wins and his teammates?
Granger Today
As I watched this episode – which I obviously enjoyed (although my wife – who is not a big basketball fan –did not) – I thought about linking this to players today. Although there are some players who come closer to the character played by the late Anthony Costello, I decided to just go with today’s Granger. No, I don’t think Danny Granger of the Indiana Pacers is quite the “hot dog” we see in the Duke of Dunk. But he does share the same last name as this character. And Danny Granger has also earned many of the benefits of being a scorer without making a comparable impact on his team’s ability to win.
Granger is averaging 25.0 points per game this season, a mark that leads the Pacers and ranks 6th in the NBA. Not only does Granger lead the Pacers in scoring, he was the lone representative of Indiana at the mid-season all-star game. In addition, he just signed a contract extension last November that will pay him $60 million over the next five seasons. This will make him one of the highest paid players on the Pacers.
So Granger is clearly Indiana’s star. But he’s not the primary producer of wins on this team. This season Granger has missed 14 games. In those games the Pacers have posted a 9-5 record. With him the team is 19-33. Although there are problems with looking a team’s performance with and without a player (we are not holding things constant so we are not sure what we are seeing is truly due to the player), in this case the Pacers’ performance without Granger is consistent with the story told by Wins Produced.
Table One: The Indiana Pacers after 66 Games in 2008-09
The Pacers have won 28 games this season — and as Table One indicates — the team’s Wins Produced sums to 29.4. Looking at Table One we can see that the play of Troy Murphy is a big part of this team’s success. After Murphy the team’s roster of above average players – WP48 [Wins Produced per 48 minutes] above 0.100 – consists of Jeff Foster and Travis Diener. T.J. Ford and Danny Granger – the two leading scorers — are close to average, so each helps. But most of the teams wins do not come from their top scorers.
Granger is primarily a small forward, and at this position – despite being a below average rebounder – his overall production is above average. But Granger also plays power forward, where his inability to rebound hurts his team. And when we put the whole picture together his overall production is below par.
King and Granger
Yesterday Daniel Fitzpatrick – a WoW Journal reader and fan of the Knicks – asked if I could look at Bernard King. Although I had forgotten this when I read Fitzpatrick’s comment, it turns out I had looked at Bernard King last fall in a post on the very much underrated Clark Kellogg. That post revealed that Kellogg, a small forward for the Indiana Pacers in the early 1980s, was a far more productive player than King. King, though, was a much bigger scorer playing in a much bigger market. Consequently, King is still remembered as an All-Star while Kellogg is known as a commentator on college basketball.
Like Kellogg, Granger is also a small forward with the Indiana Pacers. But unlike Kellogg, Granger’s level of productivity is quite comparable to King.
Table Two: Comparing Bernard King and Danny Granger
Table Two reports what the average small forward did from 1977-78 to 1990-91, as well as the small forward averages from 1991-92 to 2007-08. As one can see, small forwards from the earlier time period posted somewhat bigger numbers. But the stories with respect to King and Granger are quite similar.
Both players are (or were) above average scorers, in terms of efficiency, shot attempts, and scoring totals. Once you move past scoring, though, both players come (or came) up short. Granger and King are (or were) below average with respect to rebounds, steals, turnovers, and personal fouls. Consequently, Granger and King – despite their ability to score – don’t (or didn’t) contribute as much to wins as their fame and relative salaries suggest. As noted, Granger is about average. When we look at King we see a player who produced 68.4 wins and posted a 0.112 WP48 across his entire career. In sum, just like we see with Granger this year, King wasn’t – if we look at his entire career — much different from an average NBA player.
Group Therapy and the Demise of the Pacers
At the end of the Bob Newhart episode the Duke of Dunk brings his entire team in for group therapy. The implication was that if the team simply “came together” the wins would soon follow.
Although such an approach might work, it seems far easier to just find more productive players. Turning back to the Pacers we see that the quest to find more productive players is hampered when you invest $60 million of your team’s limited funds in a player who is only about average. When this happens, I am not sure group therapy is going to help.
Even with the Granger deal, though, I don’t think the Pacers are a terrible team. At the start of this season I agreed with John Hollinger. The Pacers in November looked like a potential playoff team. However, the injury to Mike Dunleavy – the team’s most productive player last year – has probably derailed that prediction. If Dunleavy comes back next season, though, the Pacers should be able to eclipse the 40 win mark.
Unfortunately, it might be the case of too little too late. Right now the Pacers are hampered by low attendance, which is primarily due to the team’s below average record. Indiana’s failure to draw is impacting the team’s bottom line – and as Cliff Brunt of the AP noted – might cause the Pacers to move out of Indiana.
Brunt’s article indicates that if attendance improved, the Pacers “need” to exit would be reduced. But for attendance to improve, Indiana needs a better product on the floor. In fact, I imagine they need a much better product on the floor.
Dunleavy recovering, though, is not going to transform this team into a title contender. So this means the team is going to have to find something else. Given the financial constraints of the team, the free agent market is not a real possibility. Consequently this team is going to have get lucky in the 2009 draft. And if that doesn’t happen, there is a real possibility that all fans of the Pacers are soon going to need a great deal of group therapy.
– 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.
Italian Stallion
March 13, 2009
This is the issue I have trouble with.
By also considering Danny Granger a PF, you are punishing him because of his lack of rebounding. That may be an accurate assessment of his contribution to WINS, but it’s not a valid assessment of Granger as a player or his theoretcial contribution to WINS.
I assume Granger has been asked to play PF on occasion because he’s the player most suited to do so. If anything, that ability/versatility is a positive attribute relative to other SFs that couldn’t do it as well.
Versatility is a positive in situations where the alternatives are worse.
What Indiana really needs to do is find a PF that can rebound so that Granger can play all SF. But that’s not Granger’s fault.
The same was true of guys like Larry Bird.
IMO, it’s borderline preposterous to punish Bird statistically relative to other basketball greats because he was asked to play some PF.
I totally understand that he contributed less on the boards than some other PFs and that has something to do with wins. It just has nothing to do with Bird’s talent and ability. He was also a SF that was sometimes asked to play PF because the team needed him and he was the best and most versatile alternative .
I would actually love to see Bird’s stats if he was considered only a versatile SF. That is SUPERIOR to a less versatile SF with the same stats that was never asked to play PF because he couldn’t do it at all.
Ken
March 13, 2009
The problem with your assessment, IM, is that Bird averaged 10 rebounds a game for his career. In other words, he rebounded like a PF when asked to do so. Versatility doesn’t mean playing different positions, it means playing different positions effectively. Bird could do that, as reflected in his stats. Granger can’t.
Erich
March 13, 2009
Despite the heartache of lost potential, Indiana is still on pace to beat their harshest preseason critic’s forecast.
Both Hollinger and my monte carlo simulation suggest they should wind up around 36 wins, beating Vegas’s 34.5 win projection.
I’m sure a playoff spot would make fans feel better, but the chances are slim. I peg them at 11% and Hollinger’s odds show 19%.
brgulker
March 13, 2009
I totally understand that he contributed less on the boards than some other PFs and that has something to do with wins. It just has nothing to do with Bird’s talent and ability. He was also a SF that was sometimes asked to play PF because the team needed him and he was the best and most versatile alternative .
======================
That’s a really good point, and frankly, it’s a pretty big hole in this post’s argument.
Perhaps not specifically about Granger, because he doesn’t rebound as well as other SFs.
However, I do agree that punishing a player because he is asked to play out of position is a flaw in this system.
brgulker
March 13, 2009
edit:
As a Pistons fan, I would argue that Rip’s numbers this season will be lower than they would have been had he not been asked to 1) play SF and 2) play next to AI so much.
Moreover, I wonder if Granger’s efficiency would be higher if he had better talent around him. In other words, if his surroundings were better, defenses would not be able to key in on him, thus creating more open (as opposed to contested) shots for him.
Tball
March 13, 2009
db,
Wonderful story.
IS,
Bird also slid over to PF at times because he lacked the lateral quickness (particularly later in his career) to defend quick SFs. McHale and Bird would rotate defensive responsibilities dependent upon whom Bird could more effectively cover.
The other thought to keep in mind with this positional discussion is opportunity. A player playing the PF position is likely going to be in a better spot to rebound than a player playing the SF position. Granger will naturally end up closer to the basket and will naturally record more rebounds. This statement is more conjecture and hypothesis than rule, but I would bet Bird’s rebounding rate improved when he played PF.
Nick
March 13, 2009
We’re not talking talent.
And it’s noted Granger is an above average SF. But if most of his minutes are at PF, then he is costing his team wins. Would anyone disagree that by having Granger, a natural SF, who gets outrebounded by NBA average PF’s, is a negative?
That is the ultimate point here. If the Pacers pick up a solid PF next year, slide Murphy to the 4, get decent minutes out of Foster at C, and get Marquis Daniels out of the lineup, and maybe even get back an effective Dunleavy, they can be a good team next year.
Noone is “punishing” Granger. The question here, isn’t whether or not Granger is a good player. It’s admitted, he is a better than average SF. The question is if Granger is producing wins on THIS Pacers team. And because he’s playing out of position. He isn’t.
Which reminds me of the Andrea Bargnani principle. I know he’s considering a bust around these parts, but ever since he was drafted, I’ve always thought of him as a SF. I still think that’s his natural position. Maybe PF. How does he project there?
dberri
March 13, 2009
Ken, Tball, Nick…
Good job explaining the position adjustment. This does seem to be hard for people to understand.
Nick,
Even at small forward Bargnani is below average. Sorry.
Italian Stallion
March 13, 2009
Ken,
Bird was not as good a PF as he was a SF. The fact that he also played PF effectively is a tribute to his greatness. That was my point!
I REPEAT!
Bird was great SF that was versatile enough that he could even be effective at PF. However, he should not be compared to other PFs just because he played some PF. He should be compared to other SFs that were NOT similarly versatile and his versatility should be considered a PLUS.
Ken,
Even if Bird was in a better position to rebound when he played PF, he must have been disadvantaged in other ways otherwise we would see all sorts of players playing out of position all the time.
Chris Paul would be in a better position to rebound if he played Center and Yao Ming would be in a better position to get assists if he brought up the ball, but I don’t think we’ll see either of those very soon. :-)
There are some isloated instances where playing out of position can make a player MORE effective, but generally when it happens it happens because it’s good for the TEAM not because it’s good for the individual players stats or his statistical comparisons to other players.
Nick,
I agree that Granger is costing his team wins when he plays PF, but that has little to do with his talent and ability at his natural position of SF.
Costing your team wins because the coach thinks it’s in the best interest of the team to play you out of position is an entirely different issue than measuring ability, but that’s not the way things get framed here.
It gets framed as if Granger is not any good, when he’s actually an above average SF.
If we are just going to discuss WINS, I’m OK with that kind of analysis and think it’s great. But if you start translating WINS into talent, ability etc.. you are way off base when you start comparing SFs that played some PF to other PFs. It should be the opposite. They should get EXTRA CREDIT relative to other SFs.
By the way, I am obviously not comparing Granger to Bird, but the point is the same.
Matt Walters
March 13, 2009
“If we are just going to discuss WINS, I’m OK with that kind of analysis and think it’s great.”
Welcome to “The Wages of Wins Journal.” I think an excellent idea would be for you (Italian Stallion) to start your own wordpress blog (they are free) entitled “The Wages of Talent, Ability, Etc. Journal.” While it may be difficult to arrive at a model with superior predictive power to Dr. Berri’s if you are rewarding players with “extra credit” for playing out of position, I encourage you to do so if you feel up to the challenge. I expect it will be immensely satisfying if you do succeed.
On that front, I don’t believe anyone will really be able to dispute your findings, given that talent and ability (not sure about etc.) are not always guarantors of productivity and performance. For example, I had a 1510 SAT score, but because I failed half my college courses I now work in the mall for $8/hr.
Anon
March 13, 2009
in my opinion you can’t take a model like this that measures how many wins each player is contributing and make any sort of meaningful connection to “talent.” A lot of players are unbelievably talented, but are never able to put it together and be productive.
In this case DB essentially says that playing granger at PF hurts the team because as a PF he’s not as valuable as when he plays SF. This doesn’t mean he’s a worse player, or less talented, it just means that if they play him out of position he’s not contributing as many wins. This point seems like it should be obvious.
play him at his natural position and he produces more wins. Play him somewhere else and you get relatively less production.
Michael
March 14, 2009
Sorry Prof but I read the Clark Kellogg article and this jumped out at me;
“He was never named to an All-NBA team. And he never even played in an NBA playoff game. So how could he be considered a “great” basketball player?”
“But when we look at the numbers we see that Kellogg was indeed an amazing basketball player.”
So Kellogg was great and amazing basketball player for having 3 years over 0.2 wp48, yet in your King James and Kobe article Kobe is described as “quite good?.”
I realise I brought this up in the Battier article (who you describe as “good” ) but it seems to me that the level of praise you heap upon players actually has more correlation (and is inversely proportional) to the level of recognition you perceive those players to have from the mainstream. In other words you will interpret your numbers in such a way that best positions you as a “debunker” of “popular consensus.”
This is not a criticism of your work. I just think you should ease up on players like Kobe, who according to the standards set by articles like the one on Kellogg, is a great and amazing player (even if Lebron is better.)
Nicholas
March 14, 2009
Since this is yet another post that defies conventional thinking (rightly or wrongly) by stating that Granger is a below average performer, I’ve got to ask a basic question.
Specifically, I know WP has like a 95% correlation to wins at the team level, which is why it’s an interesting metric. But does that accuracy trickle down? I’m not sure that it would have a 95% correlation to the actual productivity by each individual player.
I guess I’m just saying it seems like there’s a fairly plain causation problem here once you move WP down to the individual level. Because while its persuasive that substantially all causes for team statistics can be attributed to the team, I’m not sure that one can make the jump and say that all causes for an individual player’s statistics can be attributed to the player.
Just a thought.
Andy
March 14, 2009
Along the same lines as Nicholas suggests, I thought of a way that would seem to provide a healthy counterbalance to a metric like WP, but feel free pick my logic apart.
First, I need to establish a couple of points I can’t prove (though some may be provable) but seem logically sound.
-Players will make a greater % of shots when uncontested while shooting by the defense
-If a player is intentionally left open by the defense, essentially daring him to shoot, that means there are now 5 players focusing on guarding the other 4 offensive players, and in a 4-on-5 situation it would be more difficult for the other 4 to score.
If both of these points are taken to be true, tracking the quantity and distribution of ‘uncontested’ shots would yield a very telling statistic. (This stat is not publicly available, but it has been suggested NBA teams’ proprietary statisticians do track such things.) This stat would help measure on offense the cliche of ‘making teammates better,’ a dominant player should attract more attention from the defense, leaving other teammates more likely to be open, if he can get them the ball. Note-the assist stat doesn’t adequately capture this aspect. For example Orlando’s entire team offense is based around exploiting shooters left open on the perimeter because of the need to double-team Dwight Howard, a player who racks up few assists.
Bringing it back to Nicholas’ point, Howard’s efficiency produces a great deal of wins, and the team’s collectively high 3 point percentage produces wins, so at the team level WP is complete. But at the individual level, Howard and his shooters provide additional space for each other, and their numbers could be expected to suffer if Howard were placed on a team unable to keep teams from double-teaming him, and the shooters would have less open looks on a team without a dominant player like Howard.
Peter
March 14, 2009
The one thing I’ve read in critiques of Wins Produced is that the metric correlates well to team wins only after a statistical adjustment for team efficiency is made, and that that adjustment basically will cause any metric (regardless of the weights given to rebounds, shot attempts, etc) to correlate well to team wins. This critique claims that you could weight the box score statistics with any numbers you want, and it will still correlate well to team wins after the team efficiency translation is computed.
Regardless of whether that’s true, I’m curious to know whether the weights are derived from regressions to team wins, or whether they’re (as some critics have claimed) based on the idea that a possession “should” be worth one point.
dberri
March 14, 2009
Peter,
It is definitely based on a regression. And different specifications of the regression yeild the same results.
The critque you cite has also been refuted previously.
mr. parker
March 15, 2009
this from Bill Simmons latest article:
“Quick note on Varejao: He finished with eight points, 10 rebounds, one steal and one block in 36 minutes but was Cleveland’s second-best player and made at least eight fantastic energy/defensive/intangible plays in the second half. This is why I don’t trust the statistical revolution in basketball yet. Get a tape of the second half and find me a way to measure what Varejao did and I’m happy to jump aboard. We are not there yet.”
I’m pretty sure win score thought this was a great performance. This guy talks from his bum
dberri
March 15, 2009
mr. parker,
do you have a link to this?
Faizan
March 15, 2009
http://sports.espn.go.com/espn/page2/story?page=simmons/090311&sportCat=nba
Scroll down to the “running diary” of the second half. Occurs at the 1:34 mark of the fourth quarter about 3/4 way down the page.
123
March 16, 2009
“The critque you cite has also been refuted previously”
A link to the refutation would be nice
mrparker
March 16, 2009
faizan beat me to it
JoeM
March 16, 2009
123,
https://dberri.wordpress.com/2007/11/01/what-the-box-score-data-says-about-shane-battier/
Check out the comments.