So I’m working on a lengthy analysis at work, a team review for Professor Berri (think freedom) and my next (and hopefully great) series (Andres already got a preview). So a short post for today.
First the best available Free Agents (yes , I know some of these guys are retired):
Dampier and Fazekas really should be on Phoenix’s roster.
I also updated the preseason wins produced numbers (See here). Top 25 below. Enjoy
Finally a special xkcd that made me think of Wins Produced:
Add Wins Produced to make it complete (Image courtesy of xkcd.com)
Posted in: Uncategorized
Arturo's Silly Little Stats v2.0 
Guy
10/21/2010
Arturo: I’m surprised to see you post that cartoon (which is great) in this context. It seems that Wins Produced is analogous to astrology here, since NBA teams currently operate in a massively inefficient way (according to Dr. Berri’s analysis) and any team that used WP could quickly realize a big competitive advantage. Thus, 32 teams are leaving millions of dollars sitting on the figurative sidewalk. So to believe in WP I think requires you to also believe that capitalism — at least as practiced in the NBA — is not very ruthless at all. Right?
arturogalletti
10/21/2010
Guy,
The nba is currently a very inefficient market IMO. There are some teams that are clearly better than others . I did a whole piece on it (see here).
Chicago Tim
10/21/2010
Guy is right, the premise of the cartoon is that markets are efficient and rational. In fact, many economists, including Prof. Berri, believe otherwise, which would spoil the point of the cartoon.
arturogalletti
10/21/2010
Markets can and do correct themselves (as Dave points out for baseball in Stumbling on Wins). That’s what I was going for. I forget my audience is incredibly high brow.
Guy
10/21/2010
Arturo: I think you are understating the magnitude of the inefficiency in the NBA that is suggested by Dr. Berri’s work. He finds: 1) there is almost no correlation between team payroll and wins, 2) teams pay for scoring, but scoring has little win value, and 3) draft order has no relationship to players’ real value. All of that seems to add up to massive inefficiencies. Moreover, he argues there is very little usage/efficiency tradeoff. That alone means that every single NBA team could improve enormously on offense just by having its more efficient shooters take more shots, and vice-versa.
And no, I’m not suggesting Dr. Berri believes the highest efficiency player should take 100% of the shots, or something silly like that. But he does argue that within the range of current NBA practice — let’s say, usage rates of 10% to 30% — there is no significant usage/efficiency tradeoff. So that leaves a lot of room for every team to dramatically improve its overall efficiency. And this dwarfs whatever small undervaluing of on-base ability may have existed in baseball.
So believing in Wins Produced does really require you to believe that all 32 NBA teams are walking away from millions and millions of dollars in free revenue, while Wins Produced and related analysis is made freely available to them. I wonder what theory could explain such a gigantic market failure in the NBA, while MLB operates quite efficiently?
nerdnumbers
10/21/2010
Guy,
I remember reading a great interview with a head of a record label. He said something akin to “All these youngins were talkin’ bout this new ‘internet’ thing but I didn’t understan’ it none. I was worried they were trying to steal my money so I ignored it!” (Possibly paraphrased and dumbed down. . . ) but Dr. Berri has often made the point that most GMs don’t understand stats. As a result it’s not a matter of turning away free good advice it’s going the route of “I don’t understand it, so I will ignore it and go with what I know.” Michael Jordan and Isiah Thomas are both considered great great players. However, when you see their understanding of constructing a winning team. . . doesn’t match up. Anyway it turned out to be the winning strategy for the record labels, who didn’t turn away millions of dollars based on free advice (you’re aware most P2P software was FREE when iTunes hit right?)
Chicago Tim
10/21/2010
The baseball market took a hundred years to correct itself, and a good thirty years after the proper correction was discovered. I think stats came first to baseball because it’s so much easier to statistically analyze offense in baseball — even though it is a team game, offensively each man’s stats are isolated.
Basketball is well on its way to incorporating stats the right way. Already several teams pay good money for statistical advice, even though many of them proceed to ignore that advice. We even have one owner and one GM who have statistical training. Plus, there’s still no consensus about the right statistical analysis, even among statisticians. But it will happen. In thirty years, everything will be different. So if you are going to make money betting the over/under, now’s the time!
arturogalletti
10/21/2010
Guy,
The usage question is an interesting one. I believe that small usage increases do not impact player efficiency and small usage decreases increase player efficiency. So if a coach could convince a high usage player like Carmelo to decrease his usage by passing the ball to his teammates both Carmelo and his team would improve (the assumption here is that he sees a usage decrease and efficiency increase and everyone else sees their usage increase slightly). In essence this is what someone like Larry Brown would call playing unselfishly or the right way. I need to build a data set to prove this at some point.
Guy
10/21/2010
Nerd#: I don’t think a revolutionary new technology is a great analogy. It’s not like NBA teams have never heard of “rebounds” or “FG%.” The question is whether and how basketball has managed to not value them correctly. Sports teams don’t need to understand statistics to act efficiently. Years of trial-and-error under conditions of competition usually produce highly efficient outcomes, even if the practitioners don’t have the statistical ability to know why the things they do are “correct.” Baseball is actually highly efficient overall — managers choose to have players steal a base, for example, in almost the perfect proportions based on the score of the game and number of outs. And salaries closely track the most advanced statistical assessments of player value. Why would basketball be so different?
And as Tim notes, lots of NBA teams employ statistical consultants. I’d say use of advanced statistics is at least as common in the NBA as in MLB, maybe more so. True, no team appears to have adopted Dr. Berri’s methods — and I think many of them wish some of their opponents would do so — but that doesn’t mean they aren’t using statistical analysis.
Evanz
10/21/2010
“So if a coach could convince a high usage player like Carmelo to decrease his usage by passing the ball to his teammates both Carmelo and his team would improve (the assumption here is that he sees a usage decrease and efficiency increase and everyone else sees their usage increase slightly).”
You’d think, right? I can’t imagine coaches don’t tell players these things. I’ve thought about this a lot, and the only conclusion I have reached is that decision-making in the NBA (when to pass, when to shoot, etc) – to make a baseball analogy – is like the mental “ability” to take pitches, whereas the ability to make the shots or the right pass is more like the physical ability to hit the ball. You can’t teach Miguel Tejada how to take pitches, so you go out and get Scott Hatteberg.
Evanz
10/21/2010
@Guy
you said: “… while MLB operates quite efficiently”
Didn’t you read Money Ball? MLB was not at all efficient until people realized what Billy Beane was onto. I think the NBA is headed in that direction, with GM’s like Presti, Cho, and Morey. The correct use of stats will correct the market inefficiencies. And the new CBA will probably quicken the trend.
Guy
10/21/2010
Evanz: I’m quite familiar with Moneyball. However, the idea that Beane discovered large inefficiencies is mainly a myth. The claim was that OBP was undervalued in comparison to SLG. While the evidence for this is mixed, as a practical matter it hardly matters because the two are so highly correlated that there were almost no significantly undervalued players (high OBP but low SLG). And in any case, my main point is that the inefficiency claimed in Moneyball is about 1/100 the size of the inefficiencies that Dr. Berri claims to have identified.
nerdnumbers
10/21/2010
“Why Most Things Fail” was a pretty interesting book (although was hard to read a lot of it.) Anyway a point I remember reading in it was the “Market will Correct” Game Theory thought. They said in studies of classic supply-demand curve scenarios where a change happened to one curve it could take up to 100 years to correct (I’d be interested in more on this, haven’t looked into it much). But I think most economics courses leave out the time to adapt point in those scenarios.
“Triumph of the Nerds” a great documentary on the computer talks about a “30 year rule” which is it takes 30 years for a new technology to be adopted. For example the “modern” desktop was invented in the 70s by PARC at Xerox. It wasn’t until the 90s it was common place.
Just saying classic math + economic theory of “New Strategy is immediately adopted and takes over the market” isn’t necessarily true. . .
Chicago Tim
10/21/2010
There are some unfamiliar names in the top 25 preseason players.
Erick Dampier is meeting with the Suns today.
Shawn Ryan
10/21/2010
Guy-
Here are some random thoughts that your statement inspired:
I think that there’s some sort of disconnect between different aspects of conventional wisdom that leads to this propensity for mis-valuing particular statistics.
A lot of fans, but perhaps not a majority, seem to have a reasonable understanding of the statistics, on a team level, that need to be achieved in order to win. They are able to look at the team statistics of championship teams and recognize that they tend to score efficiently, avoid turnovers, get rebounds, shoot a lot of free throws, etc.
For some reason, they don’t carry that knowledge over when examining individual players, however. Perhaps the bias stems from watching players like Iverson, seeing how talented they are, and deciding that whatever it is that they do is good.
I think another root of these bad judgments is that normal fans don’t really differentiate very much between positions as far as having an idea of average production in the various statistical categories. I think that they know that big men get blocks and rebounds, that little men get assists and steals, and that the guys in between get points. That really is not a nuanced enough understanding of typical production of the different positions to be able to understand how much any given player given certain stats is helping his team. Really, rebound totals are not even taken into consideration when most fans glance over the stats of a PG, SG or SF. Turnovers seem to be disregarded all together (there’s a conventional wisdom notion that players like Kobe, Iverson, Melo will just have a lot of turnovers because they always have the ball in their hands; i.e. it’s complicated, so it gets ignored).
Another reason that it is not surprising that faulty conventional wisdom infects the reasoning of NBA decision makers (and I think this argument may come from Berri, but I’m not positive) is that their personal incentives are as much to appear to be making sound decisions as it is to actually make sound decisions. For a GM, the penalty, in terms of public outcry that could affect his career, is much worse for decisions that go against conventional wisdom. As a GM, it’s always a safer bet to get a player with an established name than to get a productive, but unknown player. If a well known player fails, the fans will blame the player and/or the coach (e.g. “he doesn’t try hard enough”, “he doesn’t have the will to win”, “the coach doesn’t have control of the team”, etc.), but with an unknown player, if there is a problem, the GM is likely to take the brunt of the blame.
Guy
10/21/2010
Shawn: Interesting observations. I’m sure you are right that fans make various mental miscalculations in assessing the game. And GMs doubtless make some as well (though fewer). But what that misses the reality of how sports evolve over time. We now have decades of history of professional basketball, in which hundres of coaches and GMs have had the opportunity to experiment with roster construction, playing time, and strategy on the court. As teams succeed, they will keep doing what they are doing. Teams that fail will tend to imitate the winners. Over time, sports become more efficient. The people running the sports may not know the full statistical logic behind the conventions of their sport — why NBA teams have smaller perimeter players and big men who play inside, or why the best baseball fielders are at SS and CF — but those conventions largely end up being efficient simply through trial and error combined with imitation.
Imagine for example that good rebounders were far undervalued. It would be easy for a losing team to hire a few of these players, who are relatively cheap, and then post an unexpectedly high winning %. That would get noticed, and then immitated. And it’s not like it’s a secret that rebounds are useful. So it’s just very hard to see how there could be these opportunities for massive efficiency gains just sitting there in plain sight.
And they are massive. Last year, for example, there was no correlation at all (negative, actually) between FG% and FGA/minute. If Dr. Berri is right, that’s an incredibly huge inefficiency — teams are making no effort whatsoever to have better shooters take more shots. Shot distribution is literally worse than a monkey throwing darts. If this were true, any poster on this blog could greatly outperform every single NBA coach, simply by insisting that FGAs track with shooting efficiency. Do you think that’s true?
Chicago Tim
10/21/2010
NBA coaches have limited power in the NBA. How much success would a coach have telling Carmelo Anthony to shoot less? Shooting is how these players get paid, after all, so the players act rationally when they shoot as much as possible, even if there are more efficient shooters on the floor.
And coaches act rationally when they lose following conventional wisdom, rather than benching Carmelo Anthony because he refuses to shoot less — a move that would be far more likely to get them fired than losing a few games. And GMs act rationally when they pay the market price for shooters rather than letting, say Derrick Rose walk away because he’s not worth what the market will pay for him. And owners still make money whether they win or lose — otherwise the Clippers would have folded years ago.
It’s not an open market, and it’s not a fair market, and it’s not a rational market. And it takes a long time for people to risk ignoring conventional wisdom when their jobs are at stake.
This obviously has implications far beyond sports. We have learned in the last few years how irrational the market can be. Yes, over the long run there are corrections, but in the short run there can be deep recessions and double-dip recessions and depressions, even though that isn’t rational.
notherbert
10/24/2010
read this just a couple of days ago, thus compelled to add it here. it doesn’t prove against the generality but its great how it specifically is opposite of what ChicagoTim is saying.
“When George Karl first arrived in Denver there were several games where Carmelo found himself on the bench in the fourth quarter because Karl was trying to teach him that not every shot is a good shot. There was a great deal of consternation that Carmelo might get fed up with sitting out during crunch time, however, Karl’s teaching began to slowly immerse itself in Carmelo and the next season Carmelo Anthony shot what was then a career high 48.1%, five points higher than the previous season, and he also bumped his true shooting percentage up from 52.6 to 56.3.”
full article – http://www.roundballminingcompany.com/2010/10/19/carmelo-anthony-efficient/
Guy
10/22/2010
“Shooting is how these players get paid, after all, so the players act rationally when they shoot as much as possible”
Then why don’t all players do this, Tim? Why don’t all players end up with about the same usage rate of 20%. After all, every player can shoot whenever they have the ball. If every player tries to maximize their points, the only possible outcome is that every player will take the same number of shots. Any player who tried to take more than their share would quickly find that his teammates refused to pass him the ball. So your model of player behavior can’t possibly explain what we actually see.
Evanz
10/22/2010
“Then why don’t all players do this, Tim? ”
Not all players get paid the same. It seems that, if there is any one factor that controls shooting frequency, it’s salary. Has anyone done a regression on that?
Chicago Tim
10/22/2010
“As much as possible” varies from player to player. Some can get away with more, some cannot, often based on their reputations when they come into the league, and how much money the team has invested in them. But for every player there is a built-in incentive to shoot more than he should.
There is also some incentive to win championships. After all, that’s what gets you endorsements, etc. So in the right situation shooters might learn restraint. But far more often they will know that their teams have no chance to win a championship, and then, as you say, it’s just a matter of whether their teammates will give them the ball, and also whether anyone will care if their coach sits them on the bench.
reservoirgod
10/23/2010
Guy:
We wouldn’t see each player on the floor use 20% of the possessions because each player on the floor doesn’t handle the ball on every play. On a team of completely selfish, shoot-first players the ball handler would keep all of the shots to themselves. There would be no passing, just shots & turnovers.
Chicago Tim:
I also disagree w/ the notion that every player has an incentive to shoot. Not every player is compensated for their scoring. The Miami Heat aren’t paying Jamaal Magloire to launch shots when he comes into the game. Same goes for Joel Anthony. Coaches have wanted Mike Miller to shoot more but he didn’t. Do a lot of players want to shoot? Of course. Do all players think their income is driven by shooting/scoring? No. Some players like Mike Miller & Pau Gasol are more interested in running the offensive set as designed than shooting.
If you think about the difference between the anarchist scenario of a team filled with ball hogs & chuckers (and let’s be honest – we’ve all been in that pickup game) and the average NBA game, then I think it’s damn near a miracle how far the coaches have gotten this much talent to work as cohesively as it does. It may be less efficient than the optimal model laid out by Wins Produced, but perhaps it’s as efficient as it can be or efficient enough, given the circumstances.
Chicago Tim
10/23/2010
You are just saying what I did, that not every player can get away with shooting more than he should (e.g., Magiore, Anthony), and also that some, who have a realistic chance of winning a championship, have a counter-incentive to shoot efficiently (e.g., Gasol).
As for Mike Miller, I think we agree that he is grossly underpaid. Perhaps he is that rare player who just doesn’t care about a max contract, or perhaps there is something else going on. But if he shot more, he would get paid more.
Guy
10/24/2010
Reservoirgod: We’re both speculating, of course, but I think the only logical outcome in a world of selfish, point-maximizing players is what I described. If players receiving in-bound pass just kept the ball, no one would score (they’d be swarmed by all 5 opponents). So players would have to negotiate a system that allowed them to pass and distribute shots, but give each player an equal opportunity to maximize his own point total. I think the only sustainable informal agreement is everyone gets to take about 1/5 of the shots.
In any case, I think we agree that actual player behavior cannot possibly be explained by this theory of player behavior.
Chicago Tim
10/23/2010
No Jerry Stackhouse? He’s the one Miami picked up. Looks like a significant drop off from Miller.
arturogalletti
10/23/2010
Jerry Stackhouse Last 3 Years: 2429 minutes 0.98 wins. Significant dropoff is understating it.
Neal Frazier
10/24/2010
part of the NBA’s inefficiency could be that about half the teams are the play things of idle billionaires – ie run for fun/ego rather than for proffit