On Sunday, while most basketball fans were tuned to CBS to watch the NCAA Tournament, ESPN telecast a game between the Miami Heat and Detroit Pistons. As a fan of Detroit, I decided to look away from the tournament and look in on the Pistons basketball. The picture was almost good. Specifically the Pistons — without the services of Allen Iverson, Rasheed Wallace, and Richard Hamilton – managed to lead most of the way before faltering at the end.
Although the game was entertaining, the comments of Mark Jackson and Jeff Van Gundy were more interesting. At different points in the broadcast the conversation turned to why the Pistons have faltered. And once again the viewers were able to see “expert” broadcasters try and reconcile these three “facts.”
- Allen Iverson is one of the greatest players to ever play the game.
- Chauncey Billups is a very good player, but not one of the greatest players to ever play the game.
- The Pistons with Allen Iverson are clearly worse than the Pistons with Billups.
The simplest way to reconcile these three statements is to acknowledge that the first is not true. At least, if we focus on productivity, Iverson is not one of the most productive NBA players in history.
But if you can’t make that leap, then you have to resort to words like “team chemistry”. Unfortunately, since “team chemistry” can’t be quantified (unlike player productivity) we can never properly evaluate the merits of the “chemistry” argument.
As I have noted throughout the season – most recently last month – one does not need “chemistry” to explain the Pistons demise. No, much of the decline (although not all) is tied to the fact that Iverson is not as productive as Billups.
Another Reason Why Trading Billups was a Good Idea
Again, I have said this before. What I wished to add to the subject is something I observed in looking at the numbers Billups is posting in Denver this season.
Table One: The Denver Nuggets after 70 games in 2008-09
As Table One notes, Billups has produced more wins than any other player on Denver’s roster this season. But if we compare his production to what he did last year in Detroit, he clearly has declined. Last year he posted a 0.304 WP48. This year in Denver his mark is only 0.188. Yes, he is still above average (average is 0.100). But he clearly is offering less.
When we look at the individual numbers – posted in Table Two – we can see where he has declined. This season Billups has seen declines with respect to both his shooting efficiency and assists.
Table Two: Evaluating Chauncey Billups
The next question we should ask is why Billups is doing less. And one issue I would emphasize is his age. Billups is 32 and will turn 33 in September. This means he is rapidly approaching the age where playing basketball in the NBA is not possible. To illustrate, across the past 30 years, 95% of player seasons were played by players who were younger than 35 years of age. In sum, the clock is approaching midnight for Mr. Big Shot and when it hits 12, he won’t be of much use to an NBA team.
All of this suggests that the Iverson-Billups trade was a good move by Joe Dumars (Detroit’s GM). Yes, fans of Detroit are suffering this year. But as noted previously, Iverson’s contract expires and this gives Detroit hope for next year. Plus, Dumars got rid of a player that will be approaching 35 years of age in 2010-11 while collecting $13 million.
In sum, it looks like Dumars has done the same thing to Billups he did to Ben Wallace. He let a player depart whose production was destined to slip. And this is something every prudent general manager should be doing. As the late Cotton Fitzsimmons once said (and I can’t find the quote but I think he said something like this): “Please don’t let my great players retire on me.”
Forecasting Wins
The predictability of age brings me to the issue of prediction. If we look back at Table One we see two projections of the Nuggets. The first assumes that what the players did in the past (typically last year) is what they will do in the future. The second looks at how many wins a team should get given what the players are doing this season. When we look at these two projections we see that despite the decline we see in the production of Billups, Denver’s Wins Produced this season is consistent with what we would have expected if we believed that Denver’s players would keep doing exactly what they did in the most recent past.
The approach presented in Table One is the standard approach I take in evaluating a team. But I think it has led to some confusion.
To see this, consider what J.A. Adande said at ESPN.com this past week. In a wonderful article on Malcolm Gladwell’s latest book – Outliers (an excellent book as I have noted in the past) – Adande made the following observation.
The Wages of Wins data suggested the individual components of the 2007-08 Boston Celtics were good for 52 victories based on their production the previous season. The Celtics wound up winning 66 games and the NBA championship. Clearly, something was up that couldn’t be explained by the numbers.
The table Adande linked to was part of a column I wrote on the Boston Celtics last summer. In this column I noted the following:
As Table Two indicates, given what these players did in 2006-07, the Celtics should have expected about 52 wins last year. In other words, the team should have expected about a 28 game improvement in the standings.
One should note, though, that 2006-07 was a down year for Paul Pierce, Kendrick Perkins, KG, and R. Allen. If we look at the 2005-06 numbers for these four players – numbers that appear more consistent with each player’s career marks -Boston would have expected about 65 wins last year. When the season was over, the Celtics – led by KG and Pierce (and let’s not forget Rajon Rondo) – won 66 games. The 42 additional wins was the largest regular season turnaround in NBA history. And when the post-season ended, the Celtics had won the franchise’s first title since the days of Larry Bird.
As one can see, if we consider more than just what Boston’s players did in 2006-07, then the 2007-08 season is easy to explain.
Obviously Adande ignored the words and focused strictly on the numbers I presented. And this is understandable. It certainly looks like I am arguing that what Boston would do in 2007-08 is strictly a function of their players did in 2006-07.
Certainly it’s the case the most important factor determining player performance in this season is what the player did in the past. And that is a story I think needs to be told. NBA players – relative to what we see in other sports – are very consistent.
That being said, past performance does not explain all of the future. A proper forecast would note the issue of age, the productivity of teammates, injury, and in some instances, coaching. None of these issues are raised in my simple tables.
They are, though, often raised in my posts. What I typically do, as I did with my discussion of the Billups and the Nuggets, is first present the table. Then I look for changes in performance and make that the focus of the discussion. For most players, there is little change. But for a few, we do see differences. And it’s these differences that make for an interesting story.
In sum, the tables are not designed to be a formal forecast. They are designed to start a conversation about a team. Part of that discussion emphasizes player consistency. But part of that discussion emphasizes other causal factors like age, injury, etc…. (but probably not “chemistry”).
– 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 
thepeaceblogger
March 23, 2009
Hi DB,
Just a few word to clarify your post, no one will statistically disagree with you that Allen Iverson is not one of the most productive player the NBA have seen, but JVG and M. Jackson words were that AI is one of the best player ever to have played the game and this is quite different from the point you tried to prove here. I mean if you ask a group of…5000 NBA fans around the world about Allen Iverson legacy to the game I can guaranty you that more than half of them will agree with JVG and M. Jackson. Statistic, productivity has nothing to do with people feeling happy to pay whatever it take to have to privilege to see Iverson play and for me, who saw MJ played in Paris (McDonald Game), Iverson deserve to be mentioned in the same league as His Airness.
I mean your mathematics approach to the game and your belief that statistic should judge player status or stardom in the league is quite strange from a basketball fan point of view, I have nothing again that but according to me, most fan spend money to enjoy the game and see the superstars players do thing we couldn’t imagine possible no matter how productive they are or not. When I remember the Bulls/Celtics playoffs game in 85/86 and the 63 Pts Michael Jordan scored that night I don’t thing he had a productive game (from your standard) and his team lost the game and the series, but nobody will never forget this game and the impact MJ23 had this night on the rest of the league and the game all around the world, this type of unique performance is the reason why we love the game so much.
When I read some of your articles sometime and see how hard you tried to convince us how statistically some legends of the game are not that productive (according to your metric) or not that good, for me it look like a lack of respect for theses guys and I think most of them don’t deserve it, no matter how clever or genius you are.
I hope I didn’t touch your sensibility to high and that my comment will (finally) be posted on your blog.
Take care
Tball
March 23, 2009
peace,
Just a few minor points of disagreement:
1) there are plenty of people out there that believe Iverson is far more productive than db believes;
2) if Jackson and VJ thought Iverson was less productive than Billups, they would have used this rationale to explain Detroit’s fall;
3) I am having trouble finding the box score, but I feel fairly certain that Jordan’s 63 point playoff game was highly productive – see db’s comments on Kobe’s and LBJ’s MSG games this season as a point of reference (even one-dimensional efficient scoring is highly productive, and MJ was not one-dimensional); and
4) most fans care more about winning than showmanship, at least in their home/preferred teams. Home attendance figures correlate with winning, but do not correlate with having a player in the top ten in PPG.
If you watch for the athletic/artistic appeal, I think that’s great. I wouldn’t be critical of anyone’s reason for watching basketball. But most fans show up for winning and Iverson is not as productive as his reputation.
thepeaceblogger
March 23, 2009
Tball
Well I think I agree with you on the fact that I want my team to win a championship no matter how good are the player and obviously more productive players in a team help to win a title, but if you tell me that attendance figures is only link to a team winning %, I’m sorry but I do not agree, AI is one of the few players in this league (with LB, Kobe, DWade, Shaq, DSuperman, Nash…) who can fill an arena no matter how many time the team will loose, I mean, one of the main reason behind Phoenix talk of keeping Nash no matter how old he is and how slow he has been this year is definitely because of attendance purpose for next season, not because of his crossover skill and attendance was one of the reason Philly waited so long before trading Iverson.
About the 63 Pts game I don’t really know the exact box score number, but knowing that the game was in 85/86 I can see MJ having a one show game as he use to have before Pippen became an impact player for the bulls, but again I won’t debate on that, what I tried to say was that fan like you (I guess) and me love the game for stellar performance like this one against the Celtics or like game 1 of Lakers/Sixers 2001 NBA finals where Iverson killed LA on his own (48 Pts), productive or not!
My point to DB was not about productivity as an issue for AI, not at all, I agree with that, my point was no matter how unproductive he is, Allen Iverson is one of the best player ever to have played the game, and I think that this is fairly correct argument.
Finally, I’m too old to watch NBA game or pay a fortune to fly over sea to attend Miami game (my team) just because of the athletic/artistic appeal, no Tball, as I said I had the privilege to attend several MJ game in the 90’s and never did I look at the box score later to see if my money spent was worthy, because trust me on that every penny was worthy and I would do it again.
Last word, I was at the Chicago/Miami game a few weeks ago where DWade had one of the most amazing performance ever seen at the AAA and believe me or not, at my hotel later after the game me and some 50 others fans spent the all night talking about our team and the future ahead with DWade (locking to 2011) and everyone out there were just happy to have been at this game and productivity, statistic or metric WG48…. never came up….Never
Tim
March 23, 2009
You can find the box score for Michael Jordan’s 63-point playoff game at this link, and you will see that he did much more than score 63 points:
http://en.wikipedia.org/wiki/List_of_National_Basketball_Association_top_individual_game_scores
brgulker
March 23, 2009
My point to DB was not about productivity as an issue for AI, not at all, I agree with that, my point was no matter how unproductive he is, Allen Iverson is one of the best player ever to have played the game, and I think that this is fairly correct argument.
I don’t mean to be inflammatory, here, but that is a completely incoherent statement. How can someone who has average to below average productivity be considered to be one of the greatest players to play the game?
I can name multiple players who have made much more significant contributions to their teams and ultimately league history than AI will because they won championships. Think about a guy like Dennis Rodman or Ben Wallace. Were they stars? Sure they were, especially Rodman. Were they (in their prime) “better” than AI? I would say absolutely, and it has everything to do with their high levels of productivity on the basketball court, which is much more than just scoring.
everyone out there were just happy to have been at this game and productivity, statistic or metric WG48…. never came up….Never
Of course it didn’t! You don’t physically attend basketball games to talk stats. You go to games to take in the atmosphere, cheer with fellow fans, etc.
Moreover, just because hundreds of other fans didn’t bring up stats doesn’t mean they’re irrelevant. Most NBA fans only pay attention to 2-3 (maybe four stats), which is a very incomplete picture of productivity. Think scoring, assists, rebounds, and blocks.
Those numbers, in and of themselves, do not explain why teams like the Spurs (1-2 stars with a bunch of role players) or the Pistons (et al) continue to be dominant teams in the NBA year after year.
Hence, statistics like WP48.
AI is one of the few players in this league (with LB, Kobe, DWade, Shaq, DSuperman, Nash…) who can fill an arena no matter how many time the team will loose,
coming into this season, the Pistons had a sell-out streak that went about roughly five seasons.
It ended shortly after we traded for AI and started losing.
In short, you’re wrong about superstars selling out arenas. The reason they do is because real superstars win, and winning sells out arenas.
JoeM
March 23, 2009
thepeace,
Here is an excerpt…(from Februrary 2007)
“Last season, the 76ers were one of six teams that sold out every game on the road, and Iverson, with his high-scoring, reckless playing style, was a big reason for that. But the 76ers saw little benefit, because in the N.B.A., home teams do not share ticket revenue with the visiting team. So while Iverson helped fill the Charlotte Bobcats Arena and other poorly attended arenas, he had less of an influence at the Wachovia Center in Philadelphia. Last season, the 76ers were one of only three teams to play before less than 80 percent of their home arena’s capacity. ”
The basic idea is that Iverson is a gimmick and a gimmick that home fans get/got tired of. Away fans? Not so much.
I think that it is pretty obvious that expectations consistently sell home seats.
And, as another poster said, your grammar is pretty poor.
Thepeaceblogger
March 23, 2009
@brgulker
Sorry dude but taling about attendance I never argue about the Pistons or AI impact on it and as far as know my point on this was true, Gilbert Arenas in DC is a perfect example.
But comparing Ben Wallace to Allen Iverson, I think that we watch a different nba everynight because according 2 me AI is a superstar and B. Wallace is just a good role player, Rodman was a superstar thanks to his outcourt stories and sex affair….
I mean if you or DB come to play on my playground and try to sale to the teens around that AI is just another player in the league I think we all gonna hage a laught, trust me…..
Look I do not disagree with the fact that AI is an above average player in term of statistic or whatever, I disagree with the fact that DB, Tball and you would not consider Iverson as one of the best player to have played the game, that’s all ….. Finally don’t try to lecture me about my kids or yours reading in 10 year from now of the impact Ben Wallace had on the game and his futur induction to the hall of fame, because we both know that this will never happen, but I can assure you that AI crossover and 48 Pts game in the 01 NBA finals will a favourite youtube link for the next generation.
Take care
Thepeaceblogger
March 23, 2009
@joeym
Thanks dude to remind me of being stupid (I guess that’s what you tried to say politely) anyway try to type on an Iphone in the tube in London at rush hour and maybe we will see.
kevin
March 23, 2009
It should be noted, to tie in the last thread to this one, the Red Auerbach hated to trade away his elite players, even after it became obvious they could no longer perform up to the level they previously demonstttrated.
Here is a list of players who spent their entire careers as Celtics:
Bill Russell
Sam Jones
KC Jones
Bill Sharman
Frank Ramsey
Tom Heinsohn
John Havlicek
Satch Sanders
Larry Bird
Kevin McHale
You could add Cowens and Cousy to the list if you discount ill-advised out of retirement appearances.
He wouldn’t do it, he said, because it showed disloyalty. That if you expected loyalty, you had to give it in return. It would be hard to measure what that loyalty was worth in wins and losses, when weighed against the liability of carrying a veteran or two who can no longer perform and who could have been used to restructure the team with younger players and draft picks. But it’s pretty easy to measure the success Auerbach had as a coach and GM. Nobody else has come close.
kevin
March 23, 2009
Getting back to the AI thing, peaceblogger, on what basis can you say Iverson is one of the greatest players ever?
His shooting isn’t very good, his shot seelection is abysmal, he gets overpowered by other 2’s when on defense, he turns the ball over too much, he lacks the instincts and court vision to play a 1, his natural position, he’s a lousy rebounder and the teams he plays on have trouble winning.
His one obvious plus is the ability to get fouled. But alas, he’s an average foul shooter, despite his stroke.
The definition of a great player is one who’s presence enhances his teams’ chance to win. Iverson has never been able to do that. When a player has this many liabilities, he can’t be considered even a very good player, never mind a great one.
Michael
March 23, 2009
“But it’s pretty easy to measure the success Auerbach had as a coach and GM. Nobody else has come close.”
Except for Phil Jackson :-)
kevin
March 23, 2009
When Phil functions as his own GM and scouting department, maybe we can start this conversation again.
brgulker
March 23, 2009
But comparing Ben Wallace to Allen Iverson, I think that we watch a different nba everynight because according 2 me AI is a superstar and B. Wallace is just a good role player, Rodman was a superstar thanks to his outcourt stories and sex affair….
Ironically, I think you’re sort of making my point for me here. And my point is that you and I have very different understandings of what constitutes a superstar.
By your standards (at least from what I can ascertain from your comments), a superstar is someone who scores a lot of points, gets endorsement deals, and wows the crowd. I think it’s also fair to assume that the only statistic you’re considering is points scored. Finally, I think it’s fair to assume that you are also applying those standards to being a “great player;” in other words, if you’re a superstar, then you are a great player.
I agree with you about one thing: AI, Kobe, LBJ, MJ — all of them are superstars. They score lots of points, generate media hype, get big endorsement deals, and generate lots of revenue.
Where I disagree with you is how we define being a great player.
I define greatness first and foremost by winning and ultimately by winning championships. Of the list of superstars I threw out for an example, only Kobe and MJ are champions.
Second, I would agree with kevin and say that greatness is defined by one’s ability to enhance the team’s chances to win. Here’s where I throw former Pistons like Rodman and Ben Wallace into the discussion — because what they do on the court is just as important to winning as scoring is.
You’re right to question me about this, because I’m not comparing AI to these guys. In fact, just the opposite is the case. I am using Wallace and Rodman to demonstrate contrast, not comparison. Wallace and Rodman (and those like them) are great players because of their contributions to great teams, not because of their individual talent per say.
In other words, I think greatness — at least to some extent — is derivative: your greatness is derived from the success of your team and your contribution to that success. If you took Rodman from the Bad Boys, the Spurs, or the Bulls, you would have taken an inextricable part of that team’s success from them.
In sum, it is my opinion that true greatness, which I have outlined here, is often overlooked because we directly associate talent with being a great player.
There can be no disagreement: AI is one of the most talented little men of his generation. No one can question his athleticism, his passion for the game, and his ability to score the basketball. But on the other hand, no one can question that he has never been on a championship team, and in fact, he has never played for a legitimate contender. And after watching him play for nearly one full season night in and night out, it’s becoming clear to me that there’s a reason for that.
AI simply is not a great player. Yes, he does (did?) some things well. He breaks down the defense. He draws fouls. He scores points. He even gets some steals.
But he does too many things below average to poorly to be considered a great player. He’s a terrible on-the-ball defender, which exposes the team’s interior defense. If you have a young Ben + Sheed Wallace to make up for that, it’s not noticeable. But, the Pistons don’t, and it’s hurting us.
In addition to being a bad on the ball defender, he’s bad off the ball. He plays the passing lanes, which is good, but he gambles and comes up empty too often, leaving the rest of the defense to scramble.
Also, he’s undersized and automatically creates match-up problems for our defense — and this is literally every game.
On the offensive side of the ball:
He shoots a below-average percentage.
His shot selection is often questionable.
Finally, and most frustrating to Piston fans, he’s a ball-stopper. In order to score, he needs to ball in his hands. If you watched Detroit before AI arrived, then you know our offense was constructed in exactly the opposite manner. We relied on ball and player movement to score, and we did it with remarkable efficiency for years. And we were a great team that won a championship as a result.
I’m so weary of hearing about AI being one of the greatest guards of all-time, because he’s not. If you demonstrate that he’s not by pointing out his dismal statistics, you get absolutely no quantitative responses. All you get is, “But he led Philly to the Finals,” And “He crossed MJ,” And “he’s scored so many points!11111”
If someone makes the claim that a player is great, then back that claim up with some wins, some championships, and finally some statistics which demonstrate that player’s contribution to those wins + championships. If you can’t, then you don’t have a great player, plain and simple.
Caleb
March 23, 2009
Peaceblogger is obviously trolling pretty hard.
Thepeaceblogger
March 23, 2009
@brgulker
Well I guess we agree on some stuff here, first we are different as fan without any doubt, 2) I understand the value of dude like Wallace and Rodman for their team because of their productivity whichimpact the team result and with 4 rings Rodman is probably the best example, 3) I never said that AI was one of the best PG in the game because we both know that this is false and maybe going so far by saying that he was one of the best player to have ever play the game was just inadequat and quite far from the true.
I love the game and every trip in the US to attend a game has always been decided on having the privilege to see unique player play the game, I mean I never thought of attending a Pistons game (no disrespect here) or a Spurs one….but every Heat, Cavaliers, Lakers or Rockets games I attended were to see superstars in real and enjoy the unique experience, and with AI elected as a starter at the all star again this year, I guess a lot of people appriciate what he bring to the game since been drafted.
Anyway it’s pretty late evening in England guys, will come back to you tomorrow if you don’t mind.
Take care all….
Daniel
March 23, 2009
Problem with the Billups/Wallace analysis. Billups and Wallace (before broken leg) would have been on pace to produce 20 wins between them this season. Their minutes have been given to players that will produce 5-8 wins. This means the Pistons gave up being a 55-60 win team for a couple seasons and the chance at a title because they MIGHT be a 45 win team when the same players decline in two seasons. Worst case scenario, they start to age like San Antonio, but keep selling out the Palace, keep winning, and have space after 2010, when most teams project to make stupid decisions and tie up all their cap space for another half-decade.
Just stupid. There’s no upside and it’s extremely unlikely they’ll get that kind of production at the two hardest positions to fill– center and point guard.
kevin
March 23, 2009
“I guess a lot of people appriciate what he bring to the game since been drafted.”
Oh, I think it’s safe to say we all “appreciate” what AI brings to the game. It’s just that very few of us “like” what he brings to the game.
brgulker
March 23, 2009
Their minutes have been given to players that will produce 5-8 wins. This means the Pistons gave up being a 55-60 win team for a couple seasons and the chance at a title because they MIGHT be a 45 win team when the same players decline in two seasons. Worst case scenario, they start to age like San Antonio, but keep selling out the Palace, keep winning, and have space after 2010, when most teams project to make stupid decisions and tie up all their cap space for another half-decade.
Just stupid. There’s no upside and it’s extremely unlikely they’ll get that kind of production at the two hardest positions to fill– center and point guard.
For the most part, I agree. However, I think Joe D is thinking even more long term than that. I think he is thinking about how good we might be in five years, if he has the flexibility to build around a relatively young core of Stuck, Rip, and Tay.
Anyway, I think I’d rather still be really good right now with an outside shot of winning it all than completely re-tooling. I’d take the 55-win Pistons who are scary in the Playoffs to this rebuilding mess any day.
kevin
March 23, 2009
“Anyway, I think I’d rather still be really good right now with an outside shot of winning it all than completely re-tooling. ”
You really think they had a shot? With the Celtics and Cavs in the same conference? I don’t think so. I think they’d have trouble getting past Orlando, quite frankly.
Mountain
March 23, 2009
” Unfortunately, since “team chemistry” can’t be quantified (unlike player productivity) we can never properly evaluate the merits of the “chemistry” argument.”
You can quantify team chemistry by looking at team and lineup +/- raw or adjusted.
You can quantify the “impact” of team chemistry by taking the sum of player values by any individual based metric and assigning a rank or quality level and compare that to lineup or team efficiency rank or quality level. The lineups or teams with the highest ratios have the biggest chemistry improvement impacts.
dberri
March 23, 2009
Sorry Mountain. There are so many problems with +/- that what you propose is really not going to work.
And what you propose about quanitifying “impact” is also incorrect.
Mountain
March 23, 2009
Of course we can disagree and disagree but unless you explain why you feel that way I will continue to feel I am correct and you are incorrect.
Mountain
March 23, 2009
Raw +/- is exactly scoreboard. Is scoreboard wrong?
Mountain
March 23, 2009
There are many challenges and problems with adjusted. I’ve explored and highlighted them pretty continuously as I also try to find benefit, as one piece, an additional piece of information and tried to improve adjusted by change of data input to more fairly split credit and blame to the lead actors, combination with statistical, use of splits, application to pairs to lineups, etc.
Mountain
March 23, 2009
I did say raw or adjusted. Though maybe you missed that or wanted to miss that. Even throwing out adjusted you can use my approach with raw +/- and find a good deal out about team chemistry and team chemistry impact.
Mountain
March 23, 2009
Or really lineup. A team is just the weighted sum of the results of lineups / when used.
5 individuals ultimately are judged by the scoreboard based on their performance in a specific lineup, not on average as individuals summed.
Lineups are living creatures not just block constructions.
Mountain
March 23, 2009
Many coaches and GMs implicitly act like a player is a player and will give his performance with anybody, anywhere. They, I think, are wrong. Most metrics give off that impression too and are wrong. including of course simple, traditional, to date adjusted.
Mountain
March 23, 2009
But raw +/- , i.e. scoreboard? It knows that lineup and timing matters. And my ratio of scoreboard / summed individual ratings knows it too.
Mountain
March 23, 2009
To a degree. Admittedly plagued by small sample sizes because coaches and GMs don’t narrow down to the best lineups over time and scattershoot so widely. Adjusted lineups can help some with this. But there remains a fair amount unspecifiable with high confidence as matchup and situation player response to them moment will vary for reasons and chance.
Mountain
March 23, 2009
should be…
…as matchup and situation and player response to them in the moment will vary…
Mountain
March 23, 2009
I did also merely claim you can “quantify” team chemistry by looking at team and lineup +/- raw or adjusted. I didn’t say you can know it absolutely, simply.
dberri
March 23, 2009
Mountain,
The impression I get in reading your comments over the years is that your knowledge of statistics is quite limited. Without leaving a multitude of responses in a row (and I would rather you not do that anymore), can you tell me if my impression is correct or incorrect?
Mountain
March 23, 2009
Multiple response was not planned but I found your last comment and this one insulting. I got a A in a grad level statistics course a long time ago. Am I advanced in that subject field? No. Do I understand enough to converse productively with anyone in the field of basketball analysis? By and large I think yes.
The impression I get is that you can’t reasonably defend your statement ” Unfortunately, since “team chemistry” can’t be quantified (unlike player productivity) we can never properly evaluate the merits of the “chemistry” argument.”
dberri
March 23, 2009
Mountain,
Let me try and say this without being insulting. Like many people in the APBRmetrics community, your formal training in statistical analysis is very limited. One class, a long time ago, simply is not enough.
This creates a problem for me in trying to have a discussion. To understand the issues with your statement about chemistry requires that you understand more about how to properly construct and test a model. So what should I do? Teach you econometrics in the comments of a blog post?
Of course, if I refuse to make the effort, you declare that I am unwilling to discuss your “position.” And to make the discussion even more unpleasant, you decide to post multiple comments in a row as if to say “he who posts the most wins.”
Frankly this whole thing is tiresome. I understand you are interested in discussing basketball statistics. But to discuss statistical analysis requires both training and experience. You lack both and I can’t provide either in a blog.
Mountain
March 23, 2009
I guess the GM that has asked me to meet with him gets a different impression than you.
The multiple posts seems like a minor thing to me. A good thing? No. But not intentionally umpleasant or intended to win by number of posts. What is reading a few hundred words too taxing? How does whether it appears as 1 post or many shorter ones really matter?
I have treated you pretty fairly. more fairly than most. In my mind better than you treated me- largely not responding to apt questions and observations. Your choice, but I don’t need to comment on your blog. If you don’t want dialog, that’s fine.
Mountain
March 23, 2009
Oh, I forgot. I took an undergraduate level econometrics course from Martin Bronfenbrenner. I got an A+. I almost went for a Ph.D. in economics but took a different path.
Matt Walters
March 23, 2009
Mountain,
I don’t know if anyone else cares about this discussion at this point, but I thought you should know that yes, you are embarrassing yourself. Posting one comment after another, boasting about grades in two classes and secret GM meetings, and accusing Dr. Berri and critics of not treating you “fairly” (while pointing to your own magnanimity in even deigning to post here) all help to convey an impression of you as a childish and deluded person.
At the very least, the appropriate response to Dr. Berri’s question was, “Yes, my knowledge of statistics is limited.” While I’m sure it is more satisfying to point out your (extremely limited) accomplishments, you are only proving Dr. Berri’s point, and you are really making yourself a laughingstock when you write sentences like, “I almost went for a Ph.D. in economics but took a different path.”
Mountain
March 24, 2009
Matt, the responses you deride were only in response to pretty impolite comments by Dr. Berri attacking my knowledge and also my ability to add to it.
You are welcome to your view. I, however, think my response was reasonable under the circumstance.
I still maintain that looking at lineup performance by raw +/- and adjusted is an obvious and excellent way to learn about chemistry. If you or Dave don’t I find that surprising and hard to support.
Phil
March 24, 2009
Mountain,
In defense of Dr. Berri’s work, what Wins Produced attempts to measure is a player’s productivity given only the box score and nothing else.
Does this leave out certain intangibles which WP fails to capture? Of course. It does not capture that good on ball defenders cause their men to turn the ball over more and shoot less efficiently; one of the reasons that Shane Battier, for instance, is very good (despite Berri’s metrics depicting him as average). It doesn’t capture which players can spread a floor, which facilitate a complicated offense, which set good screens, etc.
I believe that Berri has argued implicitly (though I cannot remember explicitly) that WP is a framework to be taken within context, and that one must apply outside intangibles as noted.
Berri has never argued that WP is an end-all predictor of success. It does reflect it very well, and WP tends to be consistent.
Now, would a lineup of Jason Kidd/Ronnie Brewer/Trevor Ariza/Anderson Varajao/Ben Wallace be an unstoppable 70+ win team? If their WP remained consistent they would, but can anyone really see that lineup being that successful?
You can take WP, like any statistical measurement, with as big a grain of salt as you’d like.
kevin
March 24, 2009
“I almost went for a Ph.D. in economics but took a different path.”
Huh. I had a similar experience. I almost sthupped Halle Berry, but took a different path.
brgulker
March 24, 2009
I think they’d have trouble getting past Orlando, quite frankly.
Do you watch when the Pistons play Orlando? We have dominated them in the Playoffs for years, and we swept them this season.
Personally, I’m praying for a 3 vs 6 match-up with Detroit and Orlando so the Pistons can still get out of the first round.
Heck, even Stan Van Gundy has admitted time and again how tough it is for them to beat us.
kevin
March 24, 2009
OTOH, Orlando has no trouble beating teams Detroit has a great deal of trouble beating.
Orlando has been on an upward trajectory and Detroit on a downward one for several seasons. It would be completely unsurprising if the worm turned in the playoffs.
The Magic are on target for a 60 win season. Detroit, even without trading Billups, wouldn’t have come within 8 games of 60 wins.
dberri
March 24, 2009
I think this exchange with Mountain highlights the problems I have had with members of the APBRmetrics community for the past three years. Mountain makes a statement that he considers “obvious.” I have the temerity to disagree. I then go even further and note that Mountain – like many members of the APBRmetrics community — lacks training in statistics and research experience. This lack of training and experience is simply a fact, but Mountain — like many members of the APBRmetrics community — takes offense that I would dare note this issue. He then offers a number of responses where he complains that I am unwilling to take his idea seriously and insists that he is “obviously” right (one should note a point in Mountain’s favor… he did leave out the personal attacks that are also common when the APBRmetrics community discusses my work).
As noted, this is essentially the same pattern I have seen in my interactions with members of the APBRmetrics community. So this is getting tiresome.
Nevertheless, let’s give Mountain another shot. Perhaps I just don’t understand what his is trying to say.
So Mountain…clearly state how you think plus-minus can be used to measure chemistry. You should begin by defining the term “chemistry.” Then provide us a step-by-step guide to how you would conduct this study.
kevin
March 24, 2009
Dave, relax. You’re never going to please everyone, no matter how cogent your arguments are.
there is a religious element to most people’s belief systems. You can confront them with all the objective evidence you want, but they will always find a way to rationalize their way around it, if they have decided they don’t want their mind changed.
I remember a quote about new ideas in science , I think it was Bernard Davis who made it, who said a grand new idea that shakes the conventional wisdom in science is not accepted, no matter how compelling the evidence. It isn’t accepted until all the stakeholders of the status quo die off and are replaced by others who have no vested intereest one way or the other.
brgulker
March 24, 2009
I don’t disagree at all. Orlando should be the better team, and they are when they play everyone else.
They have a much better record; they have young players who are improving from year to year; and, they beat better teams more frequently. That said, they don’t match up with us well. We have several bigs to throw at Howard, and they all frustrate him. Lewis and Turkoglu are both headcases, and they both struggle against us. Moreover, we have beaten them so many times that we have overwhelming confidence when we play them, and they have just the opposite experience — they buckle under the pressure.
All that to say, if you’re a Pistons fan, and you’ve watched us completely dominate Orlando the past several years and again this year; and, when you consider how we’ve struggled against ATL and MIA; it only makes sense for us to hope for Orlando in the first round. Frankly, it’s just common sense.
brgulker
March 24, 2009
BTW, my above comment was in conversation with Kevin’s comments about ORL beating quality teams this season. Looks like there was significant lag between when I hit submit and when the comment actually went live.
Additionally, I’d like to express how cool it is to see our author interacting in the comments!!!
I absolutely love this blog, and I plan on buying the book soon. Great work!
Your work has finally quantified how I’ve felt as an NBA fan for years. I’ve always believed that great players are players who win, and greatness and stardom don’t always go hand-in-hand. Stats like WP48 help quantify those suspicions very effectively and give concrete tools for me to utilize as an avid fan.
Obvioiusly, I’m a ‘Stons fan, so my discovery of this blog has helped me better understand the success of my team over the past several years, and I greatly appreciate that.
Unfortunately, it’s also helped me understand why AI is not a truly great player… which is a bit of a double-edged sword.
Anyway, thanks for your work and continued support of this blog. Whenever I see a new post has been made here on my feed reader, it’s the first one I read. Amazing stuff!
Jason E
March 24, 2009
The largest problem I have with attempts to quantify ‘chemistry’ is that it is very, very difficult to distinguish between a tangible product and stochastic variance. Were combinations of players successful because they had great ‘chemistry’ (e.g. had aspects of their games that complemented each other) or, like any probabilistic outcome, do some just by chance do better than others in terms of producing more than the sum of the components would suggest? Is this variation greater than we would normally expect (e.g. well beyond the standard error)? If it is not significantly greater than the error, calling it ‘chemistry’ doesn’t make it more useful.
In terms of utility, I’d also argue that it doesn’t matter whether it’s ‘chemistry’ or stochastic effects if it is not a good predictor of future events. If the ‘chemistry’ can go away from season to season, it doesn’t matter if it was really something beyond chance.
Given the inconsistencies in plus-minus results over time and (rather importantly and consequently) the large standard errors on things like ‘adjusted plus-minus’, it has diminished utility as a tool to predict future results.
This is not to say that it’s impossible to find patterns (and you don’t need plus minus to find them either). If a particular player combination (e.g. Stockton and Malone) when together consistently resulted in greater team productivity than expected based on summed productivity of the players on the court but showed no such boost with only one of them on the court, this would be evidence of some added effect. If you want to call it chemistry, fine. But if it happens without regularity, without some discernible pattern that makes it of value to predict future outcomes, it doesn’t matter what you call it. In such and event, it has little utility. It’s a mere curiosity.
Mountain
March 24, 2009
Phil, I have no problem with your defense of Dave’s work. I read it and find things I agree with (including Iverson in regard to this thread’s topic) and help my thinking and things that help my thinking even when I don’t or see things beyond the analysis presented (Billups in Detroit and Denver is about more than just age it is also about team chemistry). But I was primarily responding to a specific statement for a specific short reply on that itself, not offering a general critique of a metric or author.
Dave you’ve chosen to view me and this exchange largely but not entirely as just another APBRmetric encounter. I am an individual who happens to have read, written and participant there but am not a partisan of anything or anyone but my right to participate in discussion as I choose and as I able.
I don’t intend to be a play toy for you and given the continuing tone of your responses probably won’t be trying to converse with you further because I don’t find it as useful or pleasant or mutually respectful as it should be. Call me wacky but if someone takes the time to comment in detail on my writing I’ll almost always reply. You’ve said you are generally are not interested in dialog due to time constraints and I really wasn’t setting out trying to change you on this anymore, you make your own choices. I was making a brief comment (not having done so in weeks, always on the edge of not doing it anymore) not expecting any further exchange. But given that you replied this time, I did back. Especially given the irony of being told on the one hand by you that what I briefly mentioned “isn’t going to work” and “is incorrect” and on the other I am being told that a set of several dozens ways to study team chemistry that I conceived and listed using lineup level data analysis more sophisticated than the simple measure I mentioned is a main reason the GM wants to talk to me. Now I could have and perhaps should have just pondered that irony privately but it slipped out in response.
I’ll say a few more things too since you asked for more.
Merriam-Webster dictionary defines chemistry as “the composition, structure, and properties of substances.”
Here, at least to me, the “substance” is the team on the floor, the lineup affecting the scoreboard, the interconnected production unit with players performing different roles on almost every play, with very few plays that can entirely the product of one player.
When you say “team chemistry”. it seems obvious to me to study the composition, structure, and properties of the team. “Good” or “bad” chemistry are expressions of the properties or behavior of the substance- the team- thru the scoreboard. Good chemistry is a measure of “fit” expressed thru performance, ultimately aggregate lineup performance.
Is that different than how you define it Dr. Berri? Given that definition do you still maintain it is unquantifiable? Or not possible to study productively? My original post merely said I think it can be quantified and meaningfully studied using raw and / or adjusted +/-.
I say it is different than the sum of the individuals because that is what the lineup level data often shows. In a way that isn’t just noise or at least doesn’t appear to just noise to me. I say it is obvious but I will step back from that in the spirit of discussion and because it would be better to say that team performance is the product of interconnected and semi-autonomous “limbs”. To know performance you need to know both individual and lineup performance. How individual performance is affected by being in various lineups that rebound and pass amongst each other and take turns shooting and play defense together, etc. Usually individuals behave about as expected but their performance varies enough in lineups to make their performance in lineups worth of further study and active management to maximize results.
I won’t describe all the steps (sorry Dave, that information is for the GM if he wants to pursue it further) but it involves looking at player and team shot distributions, positions and roles and player types, PG leadership, the functionality of perimeter and interior sub-units, 3 point and passing capability, player versatility and consistency and more.
I agree with much of what Jason just wrote- but think the wise thing to do is to research team chemistry exhaustively and then decide how much you know or don’t know or know compared to the next guy you are competing with. I don’t expect perfect enlightenment but I do expected you can get an edge.
Mountain
March 24, 2009
Sorry for any difficulties caused on a dropped or slightly off word.
Italian Stallion
March 24, 2009
Personally, I think a blind man should be able to tell that chemistry between players is a factor in sports.
However, I suspect that if you spend your whole life studying advanced statistics and never actually participate in competitive sports/games at a high level you are probably not qualified to make a judgment on that.
With all due respect Mr. Berri, I didn’t intend the above to be insulting. I intended it to make a point in a shocking way to stress the point.
But think about what you are saying.
You are basically complaining that you have difficulty discussing some of these issues with people that don’t have a similar mastery of statistics.
Why not flip it around and ask yourself why you are qualified to discuss basketball or criticize the views of former great NBA players given that you’ve never played the game at a professional level?
For some of us that have actually played a sport or game at a high enough level, certain things are fairly obvious. That we have an interest in statisitics without a similar mastery does not mean we are wrong. It means sometimes we can’t prove what we know to be true. That’s why we might look to a guy like you with that expertise to explore these things while we are tinkering ourselves. I don’t think you should reject some of these ideas because it’s difficult for anyone to prove them statistically. You should use the expertise of former athletes etc…. to gain insights in what you should be looking at statistically.
stephanie
March 24, 2009
People with personal experience on a topic over estimate how much they know about said subject, especially when compared to outside observers. I believe there’s a cognitive bias with a fancy name for this, but I do not remember it.
Michael
March 24, 2009
Mountain don’t waste your breath trying to argue here. Everyone knows that the professor’s numbers are based on regressions and are therefore gospel and obviously vastly superior to plus minus, adjusted plus minus, p.e.r, win shares, and anything else that dare attempt to quantify the contributions of players in the game of basketball! For God’s sake don’t you realise you shouldn’t even be reading this blog without an advanced degree in statistics? Jeez! I bet when the emperor got new clothes you thought he was naked didn’t you!
I kind of had this impression that the heart of this place was just pure intellectual snobbery but wow…the hubris!
Patrick S
March 24, 2009
I read this website a lot, this is my first time posting. Let me start by saying I really appreciate the work here, and while I don’t think the statistical models presented here are the end all be all (nor are they intended to be), they are extremely valuable.
RE “Chemisty” etc: Win Score et al seems to be an attempt to quantify the productivity of players independent of other players. I’ve read multiple posts by Berri stating that players productivity remains constant even when that player is traded to other teams, etc. Age and injury (and random drop-offs) can influence a decrease in productivity, but the surrounding players generally don’t.
But then we have the idea of diminishing returns (at position), meaning that while Al Jefferson may generate 13 wins on his own (or whatever the number is), a team of 5 Al Jeffersons wouldn’t win 65 games.
These two concepts seem somewhat at odds to me. On one hand, there is an argument that player production remains relatively constant independent of teammates, on the other hand, diminishing returns seems to concede that the players are somewhat interlocking pieces, otherwise you could grab 5 Chris Pauls and never lose a game.
So, about chemistry. Is it possible that “chemistry” exists like diminishing returns exists? I don’t really have an opinion, just diminishing returns already suggests that players are dependent on each other, so maybe a concept like “chemistry” does exist.
Let me try defining chemistry as “When two or more players regularly produce more in combination than their individual production would suggest.”
Like I said, I don’t really have an opinion, I’m just interested in what other people think.
-Patrick
Jason
March 24, 2009
Patrick, the wins produced model does take position into account. A team of 5 Chris Pauls would mean that 4 of them were playing ‘out of position’ and the model never addresses what happens when someone plays out of position in (e.g. Al Jefferson’s ability to play point guard).
The model was developed with observations of real situations, and the ‘5 Chris Pauls’ etc. is beyond the scope of observation. Judging a model by a hypothetical that does not happen isn’t a particularly useful endeavor. We know that models tend to function best when evaluating things within the parameters observed when formulating the model. Since there’s never been a “5 Chris Pauls” (or 5 Al Jeffersons) lineup and such peculiar compositions were not considered in the development, the model *shouldn’t* be expected to hold. Saying that the model says 5 Chris Pauls would never win a game is irrelevant. It is outside of the parameters of reality. The model shouldn’t be expected to address situations beyond normal scope of variation previously observed. To do so is akin to presenting a strawman of sorts. This is an important point, but one that critics seem to pass over in the (seemingly endless) hypothetical ‘rejections’.
Your definition of ‘chemistry’ is reasonable (“When two or more players regularly produce more in combination than their individual production would suggest”) however to be a useful tool, it has to have predictive value, else discerning between ‘chemistry’ and random statistical variation around probabilistic events is impossible.
Jason E.
March 24, 2009
Sorry, should have said “Saying that the model says 5 Chris Pauls would never LOSE a game is irrelevant.”
Phil
March 25, 2009
Jason and Patrick,
The “5 ______ player” hypothetical was what I alluded to with the lineup I listed earler. 5 Chris Pauls obviously involved players playing out of position, but you can make a similar point without putting players out of position. Real question: how good do you think this team would be?
1. Jason Kidd
2. Ronnie Brewer
3. Trevor Ariza
4. Anderson Varajao
5. Ben Wallace
None of these players are out of position, and all are very productive from a WP standpoint, easily twice an average player’s productivity. But do you really think this lineup would win 75+ games, as would be expected if they continued their same production? No one is a reliable outside shooter in that lineup, and all are productive largely because of above-average rebounding numbers. Does anyone really think they would, however?
Berri has oft asserted that WP generally remains consistent regardless of who they are playing with – I’ve never seen the figures to back this up, but my gut feeling is that he is correct. (Incidentally, have there been any posts/writings on this?). Partly, I expect it is that NBA players generally are who they are; most have been playing organized, competitive basketball for nearly a decade before they even enter the league, and few drastically improve their abilities. (I think we can all agree that injuries can cause a drastic decline).
The other part, and this is one more difficult to address, is that generally the intangibles that players offer are not drastically different. I’ll focus upon defense, where most people aruge that intangibles play a greater role. Consider that a superb defender like Shane Battier might decrease the shooting of whomever he guards by 7% below their average, and forces one more turnover (I do not know the real figure, this is an estimation). Spread across the board, is this really going to make a big difference? We’re talking two more rebounds and one more steal per game at the most, spread across an entire team. From a WP perspective, this is not a huge impact. But three posessions that Battier has essentially created? That’s huge, despite that he will go uncredited.
The opposite effect – a player like Steve Novak that is very efficient but opposing players have a field day against offensively – also exists.
But there are enough other players on the floor that generally, things balance out. (Especially considering that most coaches make efforts to cover up other players’ weaknesses). While there are some players that provide drastically different defense individually, most coaches don’t field lineup that are drastically different in terms of offensive/defensive effectiveness for extended stretches.
What I am arguing is that it is both possible and likely that players are consistent in terms production regardless of whom they are around, and that even players that do not have an apparently strong impact in the box score or cause a spike in individual teammates’ production, can have an important impact on a team.
Leon
March 25, 2009
How do you make team chemistry? It’s all very ambiguous to me, and it’s a term you americans seem to throw around a lot. It’s barely mentioned in football in Europe and rugby because in reality chemistry doesn’t really make that much difference in them sports.
If you pick 5 very good rebounders in a team you will find that their production will probably fall. That’s not because there isn’t as much chemistry, it’s because there is only a certain amount of rebounds a team can get (or stop the other team from getting). Loading a team full of ace rebounders doesn’t mean there is gonna be a vast increase in the rebounding opportunities (unless of course they force many more misses)
The same is the case in playing lots of outside shooters. They can’t all shoot all the time. Where I think you guys are talking about chemistry you should be talking about getting players who’s styles work together. A good example is the backcourt (+ pierce) at Boston or the triangle at LA. It works because the players aren’t all trying to do the same thing.
In football (soccer) it’s far more obvious to see. You can’t have a team full of crossers, or players who all want to be in the box, you need a bit of both. Charlottes good run of form isn’t down to better chemistry, it’s down to playing their most productive 8 with a gameplan that complements all the players on the floor.
WP isn’t perfect and it’ll have problems with quantifying players like Shane Battier, but I think the more important thing when it comes to measuring how good a player Battier is (or how bad a player like Kapono is) to look at his opponents WP’s against him. Surely if you look over the whole season his opponents would have around an average WP. But I’d be able to confidently be able to say their WP would be less than that (even considering they player better teams being in the Western). I think that would be the best way to measure how good a defender a player is. You’d find his opponents committing more TO, more missed shots and less assists.
axim
March 26, 2009
one thing I will say is I have seen many discussions Dave has had with members of the APBRmetrics board, and for him to declare that most don’t have statistical knowledge is absurd. While Mountain admittedly does not have expertise in the subject, many of the people who have argued with Professor Berri in the past are in fact statistical experts. While this does not prove who is right in their arguments, it should certainly be noted
dberri
March 26, 2009
axim,
Can you give me a list of “many.”? There is Rosenbaum and,….???
And Rosenbaum — despite being referred to as an economist with an expertise in the study of sports — has never published a single paper on the topic of sports in a refereed economics journal.
Many people in the APBRmetrics forum are like Mountain. Eli Witus, Kevin Pelton, and Ryan Parker have no real training in statistical analysis.
mrparker
March 26, 2009
O.k call me stupid but isnt this team:
1. Jason Kidd
2. Ronnie Brewer
3. Trevor Ariza
4. Anderson Varajao
5. Ben Wallace
about what the Boston celtics have built????
Mountain
March 26, 2009
Ok one more. Because I feel like it is called for.
The names Dave mentions are largely self-taught though I believe Ryan Parker is studying statistics in college right now and who knows how far he goes.
But if you only respects Ph.D.s (and not a Harvard Law grad like Eli Witus who picked up the stat chops he needed easy enough), and it is narrow-minded to do so
(read Ed Kupfer’s stuff), then you could have listed Ilardi, Barzali, Oliver, I believe Tamada and actually a few others who stop by (TH Wilson, Sandy Weil, Heather A, D Sparks is finishing his, etc.) if you bothered to l00k closely at the apbrmetrics board.
The self-taught can be dismissed by Berri if he wants to but the self-taught have more statistical training and analytic capacity than most coaches or GMs. Like cherokee,”davis wylie”, 94*50, back2newbelf, etc.. They can and regularly do put out good stuff that insiders and academics would be wise to at least consider read and then if you can take it farther then go right ahead and prove it.
There is nothing on this blog that has been over my head to understand. I may not have the technical ability to produce some details in the academic articles but I have the capacity to understand enough of their value and weaknesses. For those who flippantly think or say otherwise I’ll just dismiss your evaluation and poorly trained reading right back.
But we don’t all see / get the best of each other. We’ll all do what makes sense to and works for us.
Live and let live.
dberri
March 26, 2009
Mountain,
Not to make you any madder…
But much of what you see on-line would never see the light of day in an academic journal. And this is because much of the work from the self-taught school violates fundamental rules of statistical inference and analysis.
This blog does not seek to present any of the analysis my co-authors and I do for academic journals. What I present here are the stories. These stories, though, only have some validity because we have the experience and training to make sure the econometrics is done well. What I see on-line is often not up to those standards.
Mountain
March 27, 2009
I am not mad anymore. I got ticked at a couple things but I’ve processed it.
There will always be folks that chime in with a shot at someone without knowing enough to take the shot. (Matt Walters, Dave and others choosing to pass evaluation have you read my 2000+ posts at apbrmetrics under the names Mountain, Mark and Jambalaya? If you have, think whatever you want. I think it would be harder to think I haven’t added to the basketball analytic discussion in meaningful ways though. If not, then you don’t know what I know or don’t know about basketball or basketball analysis. And professionally I’ve helped monitor, analyze and project revenues & expenditures for a state agency with a half billion dollar budget and written thick financial analysis relied upon by Governors and state legislative leaders, settled a lawsuit that helped thousands of bus riders, helped elect numerous local officials, serve on over a dozen significant boards, etc., so while your shot was annoying ultimately it has no effect).
I understand your perspective on research standards Dave. I didn’t really intend to trace old ground.
And I should say I just named some folks with Ph.D.s or self taught ability to usefully analyze basketball data and was not necessarily implying they critiqued your work. I was just mentioning them because I respect what they can say or deliver within the time they have for this stuff. I still think a lot of what gets done by non-academics or not fully to academic journal standards can still stimulate thinking and discovery. But I also dig for academic articles on occasion and respect the training the level of analysis and know it is of a different order. I want both not just one but you can evaluate and use or not use them as you wish too of course.
I wish communication was friendlier or easier but I know I am not entirely blameless either.
Anyways I will probably still read the stories, academic articles and books. But I probably won’t comment here further. If it isn’t much interest or use to you and sometimes frustrating for me too I’ll finally stop thinking it might get better and finally stop.
Carry on and enjoy the game and search as it works for you.
Phil
March 27, 2009
Mr Parker,
The Celtics have several players that are good outside shooters and capable of creating their own shot. The hypothetical lineup I listed earlier has players which are all shaky in both respects.
At a fundamental level, it’s because most NBA offense is generated by either screen-roll, perimeter isolation or post ups; these are plays that you need creative players and outside shooters to run effectively.
Some teams do not follow this trend. The Jazz, for instance, were capable of maintaining a respectable record and offensive efficiency even without Carlos Boozer and Deron Williams, their primary players for the aforementioned plays. But even then, they had a ton of solid mid-range shooters, and players that certainly are more able to create their own shot (especially in the frontcourt). The lineup I listed is very lacking in both respects.
Perhaps such a lineup could have similar success. But I’ve watched teams play against these players, and against all of them (except Brewer) their defender was allowed to sag off, making it much more difficult to generate offense.
Watching the Cavs, for instance, they already struggle offensively whenever Wallace and Varajao are playing at the same time, and they have Lebron James complimented by capable outside shooters. The Spurs often find themselves dead in the water offensively with Duncan and Ginobli injured and Parker takes a rest, even though there are other productive players.
It’s not limited to offense. Would a Jose Calderon/Eddie House/Steve Novak/Zach Randoph/Amare Stoudemire lineup be in title contention, as WP suggests? I think they’d be very good offensively. On the other hand, I honestly think that team could give up 65% efg.
Do you see where I’m going with this?
Alex
March 29, 2009
Related somewhat to Phil’s point (and, shockingly, the post’s), is there a good way to tell if Chauncey’s production has dropped because of age as opposed to teammates? Denver this year is not as good a team as last year’s Pistons, going by point differential, and last year Chauncey (probably) sent most of his passes to 48% shooter Rip Hamilton instead of this year’s 44% shooter Carmelo Anthony. If the main loss in productivity is from fewer assists, could it be because Billups has worse shooters on the other end of his passes?
Italian Stallion
April 1, 2009
“WP generally remains consistent regardless of who they are playing with”
The reason this is true is that most coaches and managements are smart enough to know that chemistry matters a lot. To avoid chemistry issues, they try to build a team where the skill sets are complimentary.
As a result, a lack of “chemisty” is usually not a huge factor. When it is, it is usually limited to a single player or two who will quickly be benched/traded etc… for someone that fits the teams needs etc… better.
If you were ignorant of the issue and looked at a huge number of players statistically, you could easily conclude that “chemistry” does not exist or is wildly overrated.
The reality is that it does exist and matters a lot, but it is only a factor in isolated cases as a result of good coaching and management. If you analyze these things on a case by case basis, you can usually tell when chemistry IS an issue.
Italian Stallion
April 1, 2009
I have one other point to make.
It was PHDs and Nobel Prize winners that ran Long Term Capital management and almost brought down the entire financial system.
It was similar rocket scientists that invented all the risk management financial instruments that may bring down the financial system now or in the future.
The idea that brilliant and educated people are always using the correct model of thinking is preposterous and dangerous.
I never finished college, never took a class in accounting, took one class in economics, never took a class in finance etc… However, I’ve succesfully managed my own personal portfolio for over 22 years because I read the right book right from the start (Vale Investing by Ben Graham) and learned the right things!
Don Taylor
May 1, 2009
Dr. Berri,
I had no idea where I should post my question since was off the topic of any recent posts, so I searched for the most recent thing on the blog about Chauncy Billups. I regret that my question will have to keep the company of the unfortunate comments you received on this thread and thank you for not deleting the comments section (that’s what I would have done! Who wants to waste their life in cyber argumentation with blowhards?) I would like to know if it is possible to isolate the games the Nuggets have played with Chauncy as a playoff success forecaster? (They look really sharp and I think they might contend.)
P.S. I’m no economist; I’m only a youth basketball coach. If I sub in my 4 best shooters and my best offensive rebounder and my +/- goes through the roof, have I found “chemistry” or a sagging zone?
Joe
May 11, 2009
If you sub in the 3-4 best shooters and the best 1-2 offensive rebounders/defenders and the +/- goes through the roof once you may found a sagging zone. If it happens against lots of teams and different defenses and that lineup works better than other combinations you could play you have found chemistry. If you decide who to play based just on individual stats instead of lineup performance that would be a mistake.
Joe
July 9, 2009
Hey Don did you get a response from Berri? Nope.
The blowhards are Berri and his cadre of sycophants who personally attack anyone who disagrees with them and their sense of academic superiority.
Don Taylor
July 9, 2009
Joe,
While posting comments on a well trafficked blog the poster must assume that there is a good chance his comments will not be given response. The blogger must read all comments and e-mails, research and create new content and balance that against whatever is going on in their real life. Most triage responses in a way that serves more of their readership and is most in line with their current musings. That is why I apologized for being off topic.
I am a fairly successful youth coach and despite that fact I’ve been constantly second guessed by interested parties. Often I’ve thought to myself, “I wish this guy actually coached a team so I could blow him out by 40!” So I know how it feels to think you really know a thing expertly. Ideally you shouldn’t be made to fend of the attacks of people who haven’t “paid dues” and haven’t been shown how much they don’t know by the university of past failures. The problem with +/- is it, in any of its variants, has a hard time showing a statistical relevance to historical wins and losses.
Now, people have careers centering around their expertise of statistics. Just as you wouldn’t expect a doctor to follow the prescriptions of a lay healer even though his adherents seem healed. Of course the doctor would need any treatment he would consider to meet a certain scientific standard. Even with the healed screaming their testimonies in front of him. A person’s health is the only thing that could be considered as important as basketball (joke). Why do you expect less from Dr. Berri?
No doubt the Nuggets are legit though (I consider them the runner ups) But the drastic difference in the play of an other wise extremely consistent Billups vs LA vs what he showed against New Orleans and Dallas show a little bit of a blind spot in wins produced: Both Cris Paul and J Kidd get plus marks in wins produced for their excellent defensive work in obtaining steals (both do very well in WP) but they never get minus marks for their lack of ball pressure, failure to prevent high shooting percentages and forcing teammates in compromised help positions. Intro Derrick Fisher, Jordan Farmar, and esp. Shannon Brown…..Party over!
Joe
July 10, 2009
Yeah Don shot defense is important.
I gave you my response in part because I think almost everyone deserves one. It usually doesn’t take much time. It is supposed to be exchange. Wasn’t sure if you saw the earlier comment or got anything from it but I thought I’d check.
I think you undersell what you can get from +/- but don’t have time to debate it further right now. Gotta get some sleep.
Personal attacks on the web are common but it is not the way of adults or professionals.
Debate the facts, don’t try to belittle people. I guess it happened to Berri too.
Coaching is a multi-faceted and surely difficult. For youth basketball parents should chill and let you do your best.
But NBA basketball is entertainment. Fans pay the bills and have a right to their opinions. The insiders make so many questionable moves and mistakes because there is so much not fully pinned down. They know stuff fans don’t but the insiders usually exaggerate the gap. Same with pundits and the major media.
I wouldn’t expect a doctor to follow the prescriptions of a lay healer even though his adherents seem healed but I know from personal experience that some of my doctors have been arrogant, sloppy and made obvious mistakes, often because they don’t listen and think they know it all. They don’t. Listen more.
I don’t trust or accept everything my broker says. Or my contractor. Or almost anybody. I usually split the difference between what an expert lets me and what I can figure out.