1 on 1 with David Berri: Part 2

David Berri

David Berri

In Part 2 of my interview with Stumbling on Wins author David Berri, the economics professor breaks down Amare Stoudemire’s previous and projected future values and analyzes whether the Phoenix Suns adequately replaced him this offseason. Be sure to also read Part 1, where Berri discusses advanced stats, chemistry and the value the Suns have gotten from their roster during the Nash era.

Michael Schwartz: We have fervently debated Amare Stoudemire’s value for years on this site, particularly in the lead up to this summer’s free agency period. What kind of value do you feel Stoudemire possesses and were the Suns smart in declining to match the Knicks’ max offer?

David Berri: The Suns were wise to let the Amare leave for New York.  Here is why I think this was a good decision (much of this was said at The Wages of Wins Journal on July 2).

  • Stoudemire is generally considered one of the very best players in the NBA.   Stoudemire has been selected to five All-Star games and last year he ranked 10th in the NBA in points score per game.  This suggests Stoudemire should be paid like one of the very best players in the league.
  • When we look at Wins Produced — or when we look past scoring — we see a somewhat different story.  Yes, Stoudemire is good.  But his production of 10.1 wins in 2009-10 only ranks 29th in the league. That’s quite good.  But relative to other top free agents (i.e. LeBron and Wade), Amare offers quite a bit less.
NBA Wins Produced 2009-10

NBA Wins Produced 2009-10

  • Furthermore, Amare is relatively old.  As the following table indicates, Amare produced 63.4 wins across his first eight seasons.  Players, though, only generally improve up until their mid-twenties (Amare’s best season was at age 25).  Given how player performance declines with age — and how minutes change with age — we can project what Amare’s production across the next five seasons might be.
  • As one can see, Amare is only projected to produce 32.4 wins.  Part of this is due to a decline in per-minute production.  And part of this is due to the fact Amare has only played about 65 games per season.
Amare Stoudemire Wins Produced

Amare Stoudemire Wins Produced

  • So the Knicks will be paying $100 million for an above average player, but not a player who is going to substantially change the team’s outcomes.   And that means fans of the Knicks are probably going to be disappointed.

Schwartz: So it appears the Suns made a smart decision not to sign Amare to a long-term max deal, according to your models. But now the question is have the Suns adequately replaced him this offseason by acquiring Hedo Turkoglu, Josh Childress and Hakim Warrick?

Berri: Although Stoudemire is not worth $100 million, he did produce 10 wins for Phoenix last season.  And this production needs to be replaced.

This may come as a surprise to fans of the Suns, but it looks like Phoenix has come pretty close to accomplishing this objective.

So far the following players have departed the Suns this offseason:

  • Amare Stoudemire: 10.1 Wins Produced
  • Louis Amundson: 3.6 Wins Produced
  • Dwayne Jones: 0.0 Wins Produced
  • Alando Tucker: -0.1 Wins Produced
  • Leandro Barbosa: -0.2 Wins Produced
  • Jarron Collins: -0.9 Wins Produced

In sum, these players produced 12.4 wins last year.

The Suns then added these players:

  • Josh Childress: 9.8 Wins Produced with the Atlanta Hawks in 2007-08
  • Hedo Turkoglu: 4.2 Wins Produced
  • Hakim Warrick: 1.4 Wins Produced
  • Gani Lawai and Dwyane Collins in the second round of the draft.  Both of these players were unexceptional in college last season, and neither should be expected to play much in 2010-11.

If you add up the wins added, you get more than 12.4 wins. This number, though, is deceptive since Childress is not likely to play as much with the Suns as he did with the Hawks.  Still, I think one can make the case that Stoudemire has been mostly replaced.

To see this, consider the following depth chart (Wins Produced per 48 minutes [WP48] is reported after each player’s name and these WP48 numbers are adjusted for the position where the player is listed).


  • PG: Steve Nash [0.277]
  • SG: Jason Richardson [0.191]
  • SF: Grant Hill [0.148]
  • PF: Hedo Turkoglu [0.008]
  • C: Robin Lopez [0.075]


  • PG: Goran Dragic [0.088]
  • SG: Josh Childress [0.261 at SG in 2007-08]
  • SF: Jared Dudley [0.120]
  • PF: Hakim Warrick [0.043]
  • C: Channing Frye [0.029]

An average player posts a WP48 of 0.100.  So the Suns currently have five above average players, and two players — Nash and Childress — who are at least twice as good as average.

Player performance in the NBA — relative to what we see in football and baseball — is relatively consistent across time.  That being said, age does impact performance and at some point (despite what happened last year) Nash and Hill have to decline.  But if Nash and Hill can come close to what they did last year, and Childress actually gets to play some (and produces as he did for the Atlanta Hawks), it is possible the Suns are still a 50-win team.  In other words, although Turkoglu and Warrick are not as productive as Stoudemire, I think it is conceivable that the Suns are still a 50-win team.

With what happened in Miami, the Suns are not likely to win a title next season.  But I do think the Suns are still going to provide playoff basketball for the people of Phoenix.  And they are going to do this without Stoudemire (who is probably not going to the playoffs in 2011).

Tags: Amar'e Stoudemire David Berri Hakim Warrick Hedo Turkoglu Josh Childress

  • Nashty Zealot

    As much as a Kobe hater as I am, I can't believe he's absent from the " Wins Produced " table.

  • Zak

    Well, I only wonder how this "Wins Produced" stat was calculated. Where's the math to back this up? I'm sure it's out there somewhere but it's not here so I either have to take Berri's word that it is a significant stat or not. In my experience, statistics aren't the end-all predictor of anything involving the interaction of individual human beings in ANY endeavor. If you were to just go by the statistics alone, Turkoglu should have made Toronto a contender in the East last year. Individual personalities and team chemistry are very important to each individual's personal achievements.

  • Mike L

    Yeah, first thing I noticed was Kobe missing. Didn't notice Dirk didn't make the list, but I did see Jason Kidd ranked way up there. I agree that just looking at PPG is a flawed way of looking at impact, there is something fundamentally flawed with a list doesn't include Kobe and Dirk. I can't lend this system any credence at all, unfortunately. I think the only way it can makes sense is if you create a measurement that also takes the "system" in mind … obviously guys like Frye and Tim Thomas when he was with the Suns were seen as great because of the system. And that probably accounts for Nash's longevity, too, I've got to think. Nash would be good on any team, as was John Stockton. But put either of those guys in a slightly better situation on slightly better teams and they both win multiple titles. So if there is a way of looking at the system they played in as a factor you might have a better read because just as the "team" is better than a "player" so the system that runs the team affects all who play in it. Food for thought …

  • Phil

    The wins produced is interesting but clearly flawed – no Kobe and Jason Kidd in 4th. You can’t put together an NBA team via a spreadsheet

  • ralphredimix


  • duds

    I don't know what's worse: saying Lamar Odom is better than Kobe or saying David Lee is the 11th best player in the league.

    C'mon Schwartz, if you're going to do advanced stats, you might as well use the ones that make sense.

  • JB

    With Kidd, Murphy, and Lee in the top 20, I'm pretty sure that the most important stat in WP48 is being white.

  • KZ

    Is it only me or does this guy look like one of the actors from Inception?

  • KH

    Jason Kidd is not whitebtw. If he is the Barrack Obama is white. No statistical system is perfect I respect guys who stand by there's even though they sometimes seem crazy because some great players do not shine in them. How should players be ranked? I personally think he is reight on the money on Stoudemire. He is a poor defender and a slifghtly below average per minute rebounder and it takes a lot away from his sterling offensive game.

  • Brian

    Kobe isn’t on there because the stats don’t include whining to management and demanding a trade until they buy a good team around you.

  • greg

    How is Brandon Roy not on that list as well? I've personally watched him grab a W on at least 10 occasions. Sorry, but there's just no way Samuel Dalambert produces more wins than Kobe or Roy.

  • Suns68

    And so to summarize, sports fans, it was right and good to give up Amare because players decline in productivity after their mid-20′s. But the bestest thing we ever did was reacquire Steve Nash at age 30.
    I’m right with you, Al.

  • Suns68

    sorry, meant Phil.

  • Luke

    Kobe not being on that list was the first thing I noticed as well. Pau and Lamar are on there and Marcus Camby is number 6? That’s just retarded. That makes this “Wins Produced” list complete BS… I hate this John Hollinger crap.

  • Steve

    The flaws can appear to be somewhat obvious because some players just don’t fit a certain mold. However, I think it’s easy to write something off as being wrong without really taking the time to think about it and the possible merits of the situation. While Jason Kidd is obviously wrong, and not having Kobe on there is obviously wrong, certain players don’t fit the system, and I would say Kobe is one of those players. Another one of those guys would be Dirk (also mysteriously absent from this list). Kobe and Dirk are the types of players who will sit on their butt for three quarters and skate by in games simply because they are THAT good. Then, in the fourth quarter, after their team has hung around long enough to give them a shot, they will turn it on, dominate the game, and get their team the W. They don’t play hard for 48 minutes because they don’t have to. A guy like Gerald Wallace, on the other hand, was everything to his team every minute he was on the floor. The Bobcats needed Wallace to be big for all his minutes, so he was. Same thing with LeBron. If LeBron wasn’t being LeBron all the time, the Cavs would have been a mediocre team, rather than a 60-game winner. While the system has faults (as do all statistical models), I don’t think this one is as far off as many of us might think. Besides the omissions of Kobe and Dirk and the inclusion of Jason Kidd, this list is fairly accurate as far as being able to pick out the most important players for their respective teams.

  • LCD screen cleaner

    um, guys, kobe isn't there coz he was injured quite a bit last year

  • Brian

    It doesn't calculate winning games by making the last shot… it's winning games by total impact. Just because a guy makes the last shot doesn't mean he "won" the game.

  • HD

    Wait. He’s basing his argument for us dumping Amare entirely on a table that ranks Troy Murphy above Deron Williams. And don’t get me wrong, I love Marcus Camby, but I’m not sure he’s better than Dwayne Wade.

    I agree with letting Amare go, but Berri seems to be one of those determinedly counter-intuitive commentators who try to bludgeon reality into statistical models and distort it in the process. He does an injustice to the movement. Stats are great, but only when they’re used to add colour and explanation to what is a living, breathing sport full of intangibles. Not when they’re used to squeeze everything an athlete is into a mathematical model.

  • Alvy

    Both Roy and Kobe were injured much of the 2010 season, and to add, Kobe is now 31 years old, I believe. So, what I’m saying is, while neither SG had a great regular season campaign in 2010, it doesn’t mean they haven’t produced big in previous years (I mean, do you know what they produced in 2008?)

    Here is the calculations for wins produced:

  • Pingback: Part Two of a Q&A at Valley of the Suns (and a discussion of conventional wisdom and advanced statistics) « The Wages of Wins Journal

  • Joe

    This chart is total BS!!! The Top 3 players who produce the most wins for their teams are Lebron, Kobe and Wade. They’re the 3 best players in the league period! What a joke to have Jason Kidd in the top 5. He couldn’t make a game winning shot to save his life! These Stat guys over analize everything to death. It’s retarded. The only thing i agree with is that Amare is not worth max money

  • Devin

    “Kobe didn’t make it onto this list, so this list (and thus the formula that created it) are bullshit”

    Wow. Good job. This is your method for coming up with a formula that determines the value of a basketball player:

    “Hmmm. Kobe Bryant is clearly the best player in the game – look at all those difficult mid-range jumpshots he makes! Any formula I come up with must have Kobe at or near the top to be considered useful”

    Instead, this is how science and statistics work:

    “I wonder what makes a team win. Is it points? [checks correlation] Not really, efficiency matters more than total points, and pace is also a factor. Is it rebounds? [checks correlation] Yes, rebounds are valuable to x extent. What about steals? [checks correlation] They are valuable to y extent….”

    Only once you have created a model FREE FROM BIAS do you look at the names on the list. And then you should go, “hmm, I never thought this player was that good. What is it that he does that gets him on this list?” If you do that, you’ll find that you can uncover why each player is on the list where they are. Jason Kidd rarely misses shots, rebounds exceptionally well for a PG, and adds lots of possessions. Camby rebounds and blocks very well. Kobe turns the ball over a lot, is relatively inefficient with his shooting, and, as he ages, is losing the athleticism that allowed him to gain extra possessions for his team.

    It’s not rocket science, people!

  • szr

    Wins produced, when applied retroactively to 30 seasons of NBA basketball, predicts the outcome about 94% of the time (its in Wages of Wins, if you’re curious, and also the math is on Berri’s website).

    The model was derived from understanding what it takes to win games – it isn’t points but efficiency and possession factors. After all, every possession you have in a game results in the same thing – you opponent gets the ball. Either because you score, turn it over, or you shoot and miss and they get the rebound. So the idea is to examine those factors that lead to efficiency possessions (rebounds, steals, shooting efficiency, assists, etc). This makes an intuitive kind of sense too. Dennis Rodman didn’t need to be a scorer on the Chicago Bulls to make the team a ton better. Its also why trying to do what New York did and cram a bunch of scorers together was bound to fail.

    And the more you think about it, the more you realize David Lee is a very good player. He shoots the ball very efficiently. He collects more rebounds per 48 minutes at his position than the average PF. He turns the ball over at a lower rate. Marcus Camby was very good last season too, but he played on two teams that were non-factors so no one noticed.

    Still, for those of you that point to Marcus Camby and dismiss the whole model, I know. I know it challenges conventional wisdom. So arguing against it by citing conventional wisdom isn’t going to change my mind.

  • Brian

    As I recall, the Lakers were still a very good team last season even when Kobe was out with injury.

  • Mike L

    Brian – I’m a Suns fan in Portland and I can tell you from the Blazers games I’ve seen that there are times when Brandon Roy just plain wills that team to a win literally single-handedly. It’s the kind of thing we count on seeing from Steve Nash. His mere presence on that team definitely adds wins. So Roy not being on this list is yet another reason to disbelieve its validity.

  • DKH

    Look, Brandon Roy’s a good player, but he missed more than a fifth of last season, and probably didn’t come back at 100%. What is presented above is a counting stat, and it could be hard to accumulate counting stats when a player misses significant time. Maybe some of you that think Roy should be on the list should check out a rate stat like wins produced per 48 minutes (WP48). Of the players on the list (that I checked), Chris Paul (an insanely efficient player) missed about half the season, Noah also missed a fifth (but came back strong), Bogut missed nearly a fifth, and Garnett missed nearly a fifth. These are all players near the bottom where they are starting to bunch.

    As to Kobe’s absence, he was a somewhat inefficient shot-jacker this past season, making 45.6% of more than 21 shots per game, and less than a third of his three-pointers. His net possessions is probably above average for a guard, since he’s a good rebounder, but I doubt it’s by that much, given his turnovers. Defensively, he gives the most fouls of any Lakers perimeter player.

    So Kobe had a lot of highlight-reel shots, but he was only above average (not star-level) at the things that create wins (i.e. efficient offense, possession creation, etc.).

  • jbrett


    That is the single dumbest argument I've ever read. Creationists believe; evolutionists observe. Only an idiot would fail to understand the difference between finding data and formulating theory to explain it, and creating a theory and then shoehorning the facts into a flawed mold–but you then proceed to justify that totally biased method with the argument that anyone can see who the real 'winners' are, and it is right and proper to jigger the facts to conform. You should have put in an application with the Knicks; you might be the one person who could make Isaiah Thomas look like a competent GM.

    Players contribute to winning games not with athletic skills, but with discrete and largely measurable actions. It isn't called Pretty Wins; it isn't called Spectacular Moments ESPN Shows Over And Over While Ignoring All The Turnovers And Missed Shots; it's called Wins Produced. Foot speed, vertical leap, and a sweet-LOOKING jump shot are all essentially meaningless if they don't result in positive outcomes; if you can't get possession of the ball, or snag the rebound, or make the shot, form and degree of difficulty have a value to your team of zero.

    You use the word obvious a lot, which is a good way to end up with egg on your face any time–but I'll concede to your way of thinking on one subject: You are obviously an idiot.

  • mike

    Funny how the evidence against is conventional wisdom, when we all know that conventional wisdom is rarely correct.

    The best example was the AI for Billups trade. Wins produced predicted both side’s outcome in that (Detroit head down, Denver up).

    Fans talk about how fundamentals count, how box outs matter, how “playing right” matters, but when you have a stat that shows that fundamentals matter, people are shocked and dismiss it.

    Lastly, if anyone thinks that Amare, a player who relies on Athleticism, will age like Nash, a player who not only doesn’t rely on athleticism but thrives WITHOUT it, then have I got a deal for you in one Antoine Walker :)

  • Steve

    Anyone with basketball sense would have known that the Billups/Iverson trade was going to work out for Denver and not for Detroit. Billups is a winner, AI isn’t. That’s the way their whole careers had played out up till that point (minus AI’s finals appearance in what was quite possibly the weakest conference of all time).

    Anyways, that’s not the point. A lot of people are coming on here and defending this model now when the original reaction was backlash. I’m still going to stand strong that this model is nowhere near perfect. It attempts to explain how wins are made, but no model is ever perfect. Something that models don’t account for is the dependency of a team on a player and how that (negatively or positively) affects the player’s performance. It’s completely obvious who the five best players in the NBA are (in no particular order):

    Wade, LeBron, Kobe, Dwight, Durant.

    Four of those five are at the very top of this list, so it’s clear that this model isn’t totally broken. Someone argued that you can’t take your preconceived notions and try to build a model around that…. but I think that’s exactly what should be done in the case of something like basketball. In science, it doesn’t work. Creationists create theories based on what they believe about creation. Evolutionists base theories on what they believe about evolution. They come up with a model, then interpret their findings based on the model they made up. Then they call it fact! We know that’s obviously wrong.

    But with basketball, something that is totally objective, and something we can never determine FACT, I think we can strive to make models fit our perception of the game and what “greatness” truly is. The reason I think we can do this is because the list above is littered with losers. and by losers, I mean players who have no business being on a list that is supposed to be about contributing to wins. Camby, Lee, Igoudala, Randolph, Love, Dalembert… they’re not winners. The fact that Lee produced nearly 16 teams for a team that won 29 is just wrong.

    Winners win. If these guys are “producing” wins, then they need to be on winning teams. Jordan never did any losing. He wasn’t always winning the championship (he needed Pippen for that), but he was always making the playoffs and competing, from his rookie campaign. Jordan should be the gold standard, in my opinion. Other guys to consider for that would be Russell (who was very much a product of his team, at least moreso than Jordan was), Magic, Bird, Kareem, Duncan, Kobe, and Shaq. If any of those guys would fail to make the all-time list for any particular model that is supposed to determine wins produced, then the model is faulty.

    I like this model. Don’t get me wrong. It rewards guys for doing the little things. I suspect Jordan would rank very high on this list because he did the little things and everything else. But anybody can do the little things well and become a “superstar” with this model, and that’s just not true. You won’t find a single GM in the league who would prefer to have Jason Kidd over Dwyane Wade, or Andre Igoudala over Kobe Bryant. It IS valid to attach a statistical model when it obviously doesn’t follow conventional thinking and that conventional thinking is obviously right.

  • mike

    I assume you mean "Attack" not "attach".

    It is silly when you offer no reason. Why is Kobe great? Because he wins? So what? Is Sasha Vujacic a "winner"?

    Basketball is a game of 240 player minutes, that need to be distributed. How to best do that is the question faced, and anyone who thinks the 40 minutes X player players are all that matter, well, we'll just have to agree to disagree.

  • Icanadjustandberight

    Wins produced, when team adjusted retroactively to make things fit, makes things fit about 94% of the time. Brilliant.

  • Devin

    I'm glad jbrett called Steve out on that one, because that was indeed a stupid comment.

    Interestingly, Wins Produced does put pretty much all of the so called "greatest players" near the top of the career win totals list – except for a few guaranteed hall-of-famers, for example: Iverson (I hope no one wonders why Iverson was not an exceptional player) and Kobe (he's a good player, but not one of the very best. His career wins totals will pile up pretty decently simply because of the minutes and number of games he will end up playing). The "great" players who actually aren't so great are always scorer types, and also tend to be dynamic and exciting guards. Casual fans watching the game always notice the flashy plays and fail to remember the turnovers and missed shots.

  • Pingback: Rapcast #86: Talking Phoenix East + Andersen trade motivation | Slam Dunking

  • Steve

    @jbrett- I’m curious as to how you believe that evolutionists have come up with the only idea that makes any sense for the origin of species. Evolution has been based on two basic principles: the adaptation of life, and the similarity of beings as a result of a common ancestor. Adaptation of life isn’t disputable, and it’s not something that creationists would try to dispute. However, making the jump that adaptation will lead to rapid bursts of species mutations that result in the progression of life is not the only possible explanation. In fact, it’s not even a good one because it is statistically impossible. I won’t get into that with you though.

    As far as basketball goes, I’m not wrong. If you want to hear why, go ahead and keep reading. Basketball statisticians do not account for every single stat they could possibly account for. For instance, who keeps track of the pass before the pass (aka the guy who sets up an assist)? Who keeps track of good screens and bad screens set? Who keeps track of who actually boxes out versus who out-jumps their opponent (it matters come playoff time, just ask Amare)? These are just three quick examples of things that you can easily see when watching a basketball game, but they don’t get tracked statistically. Jason Richardson is my most frustrating player in the NBA to watch because he commits some of the most bone-headed mistakes of any player I’ve ever seen (missed dunk against the Spurs, flying into Jason Kidd in the waning minutes when he had the entire court to work with so he got a charge and we lost the game, not boxing out Ron Artest when obviously there would be enough time for Ron Ron to get another look). But if you look at this stat, Jason Richardson was the second-best player on the team, producing 7.9 wins for the team. But does that stat account for the three games I just mentioned that he directly and single-handedly lost?

    No statistical model (and no theory for the origin of species) can ever account for every possible scenario. And no statistical model can ever account for everything that is truly there. If you were a statistician, you would know the faults of such models.

    One of the great ironies of this day is that statistics developed out of the field of biology, oddly enough. But, statistically speaking, all of the things that biology “proves” today are impossible, whether you’re talking about creation or evolution. Statistics are used to support the theories, but in neither case can they actually do so. For instance, under certain parameters that are highly favorable to the random creation of the most basic protein (by highly favorable I mean all of the elements necessary for protein synthesis concentrated in a small area with perfect conditions for molecule synthesis, reactions happening once per second, etc etc etc) the likelihood of the most basic building block for a protein being formed by chance is 1:10^54. Statistical impossibility is reached at 1:10^15. For one, I hope you know that 10^54 is not just 39 times bigger. It’s far more than that. And besides that, that’s the likelihood of the most basic possible building block of life being formed, let alone the development of life we have currently today. If there were one of the aforementioned reactions happening every second in those highly favorable conditions, do you know how long it would take to form the most basic building block if it wasn’t already impossible? 10^54/31,536,000 (seconds in a year). I don’t have a calculator that can even count that high, but it’s far more years than scientists even believe matter has existed.

    Let me just say, creation doesn’t explain things in a way that satisfies statistical methods either. Both creation and evolution are impossible. But my point in saying all of this was that I don’t think you understand what statistics even is. Part of the purpose in the creation of statistics was to explain observations, not just to categorize and analyze collected data. Scientists developed statistical methods so they could explain the contradictions that they were finding between seemingly conflicting theories of evolution. It was created so that they could take what they believed was true and explain it.

    That’s my argument with basketball. We know what is true in basketball. We know who the greatest winners of all time are. Now, we have to develop a model that shows what they did that no one else is able to do quite like them.

    Thanks for insulting me, by the way.

  • jbrett


    I recommend you check out Khandor's Sports Blog; it's the refuge of some of the net's most die-hard fruitcakes, and you'll feel right at home there. Like you, the author is unwavering in his innate knowledge of what is 'true' is basketball–and lots of other subjects, I'm sure–and he also lets no troubling facts dissuade him. You said:

    1) It's obvious who the best players are;

    2) Basketball is totally objective;

    3) We can strive to make our models fit our perception of the game, and greatness;

    4) Conventional thinking is obviously right.

    And now I realize you said them completely unselfconsciously. In essence: "Your model contradicts my faith (as in 'belief without evidence or reason'), therefore it must be wrong."

    I should not have called you an idiot; that was way too simplistic. You are, in fact, a fatuous pedant, and arrogant for no discernible reason. You have taken your casual observations and half-baked theories and installed them as gospel, and from that platform you attack any dissent as if your opinion was etched in stone by a finger of fire. You should trade in your PC for a pulpit; you'd still be a kook, but you'd influence a smaller circle of sheep.

  • Steve

    1) True. You will not find any circle of individuals in any significant basketball corners who will say anything other than this: The top five players in the league are Kobe, Wade, LeBron, Howard, and Durant.

    2) Mistake on my part. I was typing pretty quickly and meant to say subjective, but my mind got ahead of my typing and I ended up writing objective. Subjective was clearly my meaning, anyways, if you had an inkling of intelligence or reading comprehension.

    3) Absolutely true. That's the point of statistics. Like I said, statistics was designed to explain what we see.

    4) Absolutely true again. The five best players on the planet are those listed above. Conventional thinking tells us that, and it's true. Conventional thinking tells us Jordan is the greatest of all time, Bill Russell was a true team player and a winner, Magic and Bird were the greatest of the '80s, etc etc etc. Our instincts on the subjective matter of basketball are not incorrect.

    I appreciate you taking the time to attempt to insult me, but I really don't get why you try. I told you the reasoning behind the development of statistical methods. I offered evidence to support my claim. I established the subjective nature of the game of basketball, and thus the need to bring "conventional thinking" into the process of statistical analysis. You offered no support for your claims against my own and criticized me for what was actually your own elementary reading mistake.

    I won't argue with you further on the topic, but I hope that in future encounters in your life, you will realize that you don't have all the answers. One thing I do know an awful lot about is statistics, and I know enough to say that Berri's methods of analysis are not unbiased, and they should not be given credit as such. His method is subjective, and he has subjectively chosen weights and measures. I think he has selected most measures correctly, but obviously there is a big flaw in his system.

  • jbrett


    1) Chris Paul, Pau Gasol, Steve Nash, Tim Duncan: Just a few players knowledgeable NBA people might consider for top five. Gerald Wallace, Jason Kidd, Carlos Boozer, Rajon Rondo: A few more worthy of consideration for those who value production over the sniff test for greatness. Oh, and what significant basketball corners do you lurk in?

    2) Basketball is neither totally objective nor subjective, and arguing for either is moronic. Grow up.

    3) Are you the guy who redacts scientific data from environmental reports because they don't fit your political agenda?

    4) The Sun revolves around the Earth–which is flat, by the way. Conventional wisdom, before subjected to fact and reason. To argue that we can't KNOW anything about basketball suggests you belong to a snake-handling church, or something equally intellectual.

    And you won't argue anymore? You aren't arguing NOW; you're making pronouncements based on prejudices, and you can't even be goaded into examining your rigid preconceptions. Enjoy your next tent revival or NASCAR autograph show, mouth-breather; I've wasted enough time on you.

  • Steve

    1) Front office polls. They all have the same top five (at least top 3 in Kobe, Wade, and LeBron). Throwing Jason Kidd in there ahead of Deron Williams and Steve Nash is a joke. Do you watch basketball?

    2) Judgment of basketball is subjective unless you're looking at the only thing that really matters (the W-L columns). Saying that winning all 82 games in the regular season is ideal is an objective statement. Determining how best to win those 82 games is completely subjective. If it wasn't, someone would have figured out how to do it by now. But the whole point of this convo is to talk about advanced metrics, and that is what I was arguing is subjective. I think you're losing focus. And again, stop with the insults. It's flattering that you find this worthwhile, but you're really not advancing this conversation.

    3) I don't have a political agenda, and I don't get why you feel the need to bring that in. Since you ask, I read dozens of scientific journal entries every year, and I have yet to find one scientific research study that does not form its conclusion based off its hypothesis. They all do it. Some of them are right, mind you. But they all do it, no matter which side of the political fence they're on.

    4) Most people didn't believe the earth was flat. That's a myth. Columbus wasn't an outcast for thinking so. It's obvious just from looking at the stars from two different points that are significantly far apart that the earth isn't flat. Btw, the oldest account of a round earth that I'm aware of was found in the Bible, in the book of Job, which was possibly written in the neighborhood of 8000 years ago. Other civilizations, such as the Egyptians had a great understanding of the stars. The ancient Greeks identified the poles and meridians. So, that was a bad and misinformed example, but I get what you're saying. Maybe you should get your facts straight first, though, before you insult me.

    You know nothing about me, but I enjoy reading your guesses.

  • mike

    1) It's funny, but every year one team wins, one finishes last and a lot finish in the middle. A "Consensus" of basketball people includes GMs that put together the Nets as well as the Lakers and Celtics. A "consensus" of greatness is biased by a lot of things – for example do you favour team wins, or individual results? In Stumbling On Wins, Berri talks about factors that affect draft position in the NBA, and you'd be surprised by what they are.

    2) No it isn't subjective. To win a game, you need to score more points than your opponent. Breaking that down, you need to:

    i. Shoot efficiently

    ii. Defend well.

    iii. Maintain possession

    iv. Gather possession (rebound, steal, force turnovers).

    All that can be measured, and the relative value of each determined, e.g. what is a missed shot worth, relative to a rebound? David Berri did that, and WP48 is the result.

    You can disregard this all you want, but you need a reason WHY. What is the flaw in WP48 that makes it wrong? Is it incomplete? Your argument proceeds with no evidence, and refutes nothing. It is very much like saying "Everyone knows the world is flat so your idea that we can sail around the world is wrong".

    3) The point he was trying to make, that was lost a little in the bile, is that we cant discard ideas that don;t fit our world view. If anything, we need to embrace these.

    4) Most people DID believe the world was flat, and you are only talking about Europe in your refutal, and a very specific moment in time.

  • Steve

    1) Were we talking about the draft? Did I miss something?
    2) I don’t get your point. Yes, you need to score more points, but you still have to figure out how to do that. Object means “expressing or dealing with facts or conditions as perceived without distortion by personal feelings, prejudices, or interpretations.” Figuring out how to do the things you listed is not an objective matter. You know you need to do those things (and more), but figuring out which player can best do those is far more complicated than looking at a stat sheet.

    You contradict yourself in your second paragraph, as well. The relative value of weights and measures is completely subjective. No person can say that a missed shot is worth -0.68 points to his team and not -0.67 points. That’s called an “interpretation,” which is the opposite of objective.

    That’s my reason WHY. I wasn’t arguing the accuracy so much as the absurdity of the claim that the system is unbiased. As for its accuracy, I don’t believe it’s perfectly accurate. I think that’s obvious in the overrating of David Lee (how can a guy be labeled a winner when he loses more than 50 games), Lamar Odom, Zach Randolph, Jason Kidd, Kevin Love, Samuel Dalembert, Andre Igoudala, Gerald Wallace, Marcus Camby, Joakim Noah… etc. Are you really going to tell me that any one of those guys is more valuable or “produces more wins” than Dirk? Marcus Camby is a great player, but he’s not the sixth best player in the league. All of those guys above (except no-pass Randolph and Samuel Dalembert) are great players (Love may not be yet, but he will develop into a great player), but they’re not top 30 in the league with the possible exception of Iggy. Those ratings indicate severe flaws in this system, and I don’t see why anyone would argue anything to the contrary. Do you REALLY beleive Gerald Wallace is the fifth best player in the NBA? Really?
    3) This is a good point, but I don’t think it’s applicable in the realm of statistics, especially not advanced metrics in basketball. Real quickly, the statistic I care most about is winning, which is what wins produced is supposed to be about. But I don’t think true winners lose the way some of these guys have lost on this list. I like this system. I’m not saying, nor have I ever said this is a terrible system. It’s good. But Berri tries to pass this off as objective and unbiased, and that’s a total farce. It’s a good system. It’s very reasonable. It makes sense, mostly. But it’s NOT unbiased.
    4) Google “did Columbus’ contemporaries beleive the earth was flat,” and you might be surprised by what you find. I skimmed through the first 10 results and each one of them reports the same thing: the contemporaries of Columbus did not believe the earth was flat. His whole campaign for money didn’t hinge on the spherical nature of the earth, it was the estimated distance of his voyage that had caused Portugal (and even Spain, initially) to reject his proposal to sail to Asia.

  • Pingback: A Phoenix Suns playoff Wins Produced analysis | Valley of the Suns

  • Pingback: Josh Childress shines in training camp scrimmage | Valley of the Suns

  • Pingback: Phoenix Suns: Missing Amare | Get Buckets

  • http://www.facebook.com/profile.php?id=100003452540919 Genesis

    Furniture is being moved in today, and next up is window tnattmeres! We are being very careful to walk the line between providing enough privacy, but also keeping the views so open because the outdoors is absolutely gorgeous at this house.