If I understand the story correctly (and that’s a big if) Stephon Marbury decided this week he was not happy coming to work anymore because he didn’t want to come off the bench. The Knicks responded by taking away his paycheck, which suddenly caused Starbury to come to work.
Given that this is an incident with a New York player, it’s not surprising that the media blew the entire event out of proportion. So I thought it would be a good idea to dispel three myths about Marbury’s personal strike.
Marbury Myths
First, this is not evidence that the New York Knicks are not a well-managed club. Well, at least, this event was not necessary for us to reach that conclusion. The courtroom drama this summer, coupled with years where the Knicks managed to place themselves at opposite ends of the payroll and wins rankings, is enough to convince us that the Knicks are not well-managed.
Second, this event does not bring down the rest of the league (and yes, I think I heard Bill Walton make this argument, although I wasn’t paying real close attention). I doubt anyone decided not to go to a game this week, or not watch a game on TV, because Isiah Thomas and Marbury were feuding. And if they did, that is more than just a bit silly.
And finally, and this is the big myth, Starbury is really not much of a star. At least he hasn’t been for most of his career.
As we often do here, let’s go to the numbers.
The first number I want to cite is salary. According to Basketball-Reference.com (who gets the numbers from Rodney Fort and Patricia Bender), Marbury has been paid about $111 million in his career. And according to Chris Sheridan of ESPN.com, he is scheduled to be paid $40.3 million more over the next two seasons. So he has certainly been paid like a star.
When we look at the stats, though, it’s hard to see what he does to earn this money. Consider Table One, where Marbury’s year-by-year performance in each statistical category is reported.
Table One: Stephon Marbury’s Career Numbers
Per 48 minutes, Marbury has been below average with respect to shooting efficiency, rebounds, steals, blocked shots, and turnovers. He’s above average with respect to assists and personal fouls. He has scored at an above average rate – hence the name Starbury – but that’s because he takes an above average number of shots. And he gets to do that because possessions on his team typically start with him having the ball and therefore he often gets to call his own number. In sum, it doesn’t appear that Marbury is outstanding with respect to any aspect of the game. Most of his numbers are around the average mark, which would lead one to conclude that he is just an average player.
When we turn to The Wages of Wins metrics — Win Score, Wins Produced, and WP48 [Wins Produced per 48 minutes] – we see (as the stat-by-stat analysis suggested) Marbury has generally hovered around the average mark in his career. In five seasons he was above average, in six others he was below average. His career WP48 stands at 0.107, which again is about average.
Table Two: Stephon Marbury’s Career Wins Production
The one outlier was 2004-05, where Marbury posted a 0.208 WP48 and produced 14.19 wins. How did this happen? It wasn’t a sudden change in rebounds (still below average), or steals (still below average), or turnovers (still below average). No, the key was a sudden – yet not sustained – increase in shooting efficiency. By increasing his points-per-shot to 1.011, Marbury was able to increase his Wins Production by nearly eight (relative to his career average). Such a result suggests, as was argued earlier in the week, that scoring can indeed produce wins (like that’s a surprise).
Unfortunately, Marbury was not able to sustain what he did in 2004-05. His last two seasons he has been paid $33.6 million for about 8.2 wins. That works out to four million a victory. If all NBA teams paid at this rate, the NBA would have to be borrow money just to pay its players. So one suspects, Marbury has been a bit overpaid these last two seasons.
To Bench or Trade?
So should Marbury be benched or traded? Looking at last year’s performances, it doesn’t look like this team has many better options. Both Jamal Crawford and Nate Robinson were below average last year. This year Crawford has a Win Score per 48 minutes of 4.15, which is below average. Robinson has a mark of 9.33, which is well above average. But Robinson has only played limited minutes. Consistent with Isiah’s past decision-making, he is giving Crawford 40.5 minutes per game while Robinson only plays 18 minutes each night.
One should note that Marbury’s per 48 minute Win Score is 5.32, so if he wanted to argue he’s a better option than Crawford he would have a point. In the end, it might not be his play that led to his benching. It could be that Isiah just doesn’t like Starbury.
And so should these two part company? That would require that another team make a trade. According to Chris Sheridan, a bumper crop of free agents in 2009 might make the acquisition of Marbury attractive. In Sheridan’s words:
I am not sure such a strategy is a great idea. Sacrifice two seasons so you can bid on free agents? Somehow I can’t see this paying off.
My sense (and I could be wrong) is that without a buy-out, Marbury and Isiah are stuck together. Such a scenario is not too good for fans of the Knicks. But it should provide the rest of us with some entertainment.
– DJ
Our research on the NBA was summarized HERE.
The Technical Notes at wagesofwins.com provides 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
Wages of Wins Journal 
Paulo
November 15, 2007
But wouldn’t trading for a 30-year old Marbury *now* for cap space by 2009 be, for lack of a better word, incomprehensibly stupid?
Excuse me for being a bit blunt here, but why not suspend/just let him sit (see Iverson, Allen 2006-07)?
T.G. Randini
November 15, 2007
RAJON RANDO: MVP?
November 15, 2007
Everyone is aware of the cohesion and effort put forth by Mr. Garnett, Mr. Allen and Mr. Pierce… the holy trinity composing the core of the newly re-constituted Celtics of Boston. But it is the unsung, unheralded Rajon Rando who is on-track to become the next MVP.
In the new, improved Wins Produced system, we realized several things. Basketball is much more similar to football than it is to baseball. In the game of football, the ball is advanced over the playing field until it crosses the goal line. All statistics measuring quality of performance are based on the number of yards the ball is advanced by each player.
Basketball is extremely similar. The ball is advanced over the playing court until it is deposited in the hoop. We believe that advancing the ball is just as important in basketball as it is in football.
And who does fully half the advancing of ball for the Boston Celtics? Mr. Rajon Rondo. He advances the ball at least 50% of the forward footage on the court.
While he is doing this advancing work, the other players are just running around in circles on the other side of the court. After Rajon gets the ball more than half way across the court, he gives the other players the opportunity to further advance the ball. But there isn’t nearly as much net advancing going on when Garnett, Allen and Pierce have the ball. They throw it backwards almost half the time!
Mr. Rondo has accounted for half the footage the Celtics have gained in their first seven wins. He has accounted for .5 Wins/Game, and at this rate will have 41 Wins Produced at the end of the season. This would be a new all-time NBA record for Wins Produced.
Our season-end projection based on the current trend:
PLAYER WINS PRODUCED
Rondo 41.0
Garnett 15.0
Pierce 12.0
Allen 10.0
All those other guys 4.0
Mr. Rajon Rondo? MVP!
kevin
November 15, 2007
Randini – haha. Very funny. But not really at all.
If you want to mock the Wages of Wins, please do so elsewhere. Maybe make your own blog and you and all of your finance-MBA-head pals can chuckle along together. Otherwise, filling the comments here with garbage like this where you avoid the topic or discussion at hand just wastes all the rest of our time.
WoW is not perfect but it certainly is fun to read and helpful for those of use who appreciate more than sportscenter highlights and the other garbage that the mainstream media shoves down our throats.
And thanks to dberri and the others on this site that do engage in thoughtful discussion here :)
Mike H
November 15, 2007
Don’t feed the troll.
– Intertubes Ecological Society
T.G. Randini
November 15, 2007
dberri,
(the troll be fed… they ask for ‘thoughtful’ comment… so I address the following in economics jargon you will understand even if some of them don’t…)
GUNS & BUTTER
You denigrate Dean Oliver’s Difficulty Theory in your Wages of Wins tome because you equate his theory with a Marxist/Ricardo/Smith labor theory of value.
You indicate that the labor theory of value does not hold in a capitalist/free market environment of flexible prices determined by supply/demand models.
But guess what? Your analogy is wrong.
The NBA scoring system is much more akin to monopolistic/oligarchist or, even, centrally-planned economies.
Prices are not flexible. Prices are fixed.
All free throws made are worth 1 point. All field goals are worth 2 points, unless the field goal is made beyond the 3-point line. Then it is worth 3 points.
These prices for achievements are always the same. They are fixed.
Thus, the labor theory of value is valid in a system where prices are fixed. And Dean Oliver’s labor theory of value (difficulty theory) is the more appropriate model to use when evaluating performance in the NBA.
always,
t.g. randini
dberri
November 15, 2007
TG,
Okay, we see you also don’t understand the Labor Theory of Value.
Pete
November 15, 2007
They seem to just be indiscriminately giving away HIGH HONORS in finance in the University of Chicago’s MBA program.
h
November 15, 2007
Typical MBA — claims to understand everything, but understands very little?
Pete
November 15, 2007
Harold, I agree. I wonder if the MBA program requires a statistics class. If so Tom Randini slept thru his.
Andy
November 15, 2007
Dave,
Not arguing with you on this one (I don’t think TG understands the Labor Theory either, from that comment), but in general a lot of your responses to criticism are “You don’t understand X theory.” I don’t expect a dissertation each time, but something a little more illuminating might be nice.
T.G. Randini
November 15, 2007
dberri,
Trite. ‘Cause you know I’m right.
You equate basketball with (flexible) prices in your book supposedly derived from variations in supply and demand. How is that at all pertinent to a fixed scoring system = fixed prices? It isn’t and you know it. You obfuscate and generalize thereafter in your book without ever specifically saying why a fixed (price) scoring system equates to market pricing. And you can’t say why it does… because it simply doesn’t. So, it’s easy to be trite and say I don’t understand labor theory when you know you don’t have an economic justification to support your performance model. And if you can’t support it through economic theory, then your model (as an economist) falls apart because it has no legs to stand on. So then, if I am so wrong, without being trite, just why is it that fixed prices/fixed scoring values have ANYTHING to do with free market pricing? Regardless of the quantity of supply OR demand, a two point field goal is always worth two points.
Mike H
November 15, 2007
I hear that imaginary degrees and ethereal honors are easy to come by in the tubes.
Kent
November 15, 2007
I already miss the days when Harold A was the resident iconoclast on the comments thread. At least he wasn’t shrill.
Kent
November 15, 2007
At first I didn’t find Randini’s argument to particularly trenchant. But the capitalization of “ANYTHING” is compelling, and I have adopted his view.
dberri
November 15, 2007
Andy,
History of Economic Thought is one of my favorite topics to teach. So I don’t need much encouragement to go off on this. If you want, I will try and put together a more complete comment on the Labor Theory of Value later today.
T.G. Randini
November 15, 2007
Well, yeah, I’m just trolling here but the majority of comments seem to consist of shallow thinkers spending ten minutes trying to come up with one halfway witty sentence. (Yawn… but a couple of them were pretty funny!) At any rate, sayonara… at the very least, your hit rate or q-score probably jumped a bit in the past 24 hours.
Adieu!
Alien Human Hybrid
November 15, 2007
Dave,
Isiah is planning to get Nate Robinson a lot more minutes this year- but he has been hampered by a hamstring injury he suffered in the second game of the season. He was the first point guard off the bench in the preseason and would certainly have been Marbury’s replacement if he had not been injured. He played a painful 3 minutes in the Miami game and came off the bench in the Phoenix game (his season-high in minutes played) only after Mardy Collins went down with a foot injury.
Mike H
November 15, 2007
Trolls notwithstanding – perhaps it would be useful for someone to compile a FAQ with links to the appropriate answers. This way the conversation on various topics can progress beyond the reinventing-the-question stage? Of course, I’m sure that I’ve just reinvented-the-suggestion here.
T.G. Randini
November 15, 2007
dberri,
One last thing, (while all your acolytes groan) …
There was a point to my little parody/satire naming Rajon Rondo the MVP. It seems farcical because we all know it is much less difficult to advance the ball to the half-court line than it is to advance further and convert.
Ahhh… there we go with the magic word ‘difficulty’… or ‘Difficulty Theory’.
Let me put a capitalist spin on this rather than your take that equates difficulty theory with the labor theory of value (and/or) Marxism.
Economics is about many things, but one of them is the allocation of scarce resources. Let’s just consider (simply, to make the point…) two resources.
Resource A: The ability to shoot and score on a wide-open lay-up with 20 seconds left on the shot clock.
Resource B: The ability to create a shot and score while being double-teamed with two seconds left on the shot clock.
I believe that Resource B is far more scarce in high school basketball, college basketball, and professional basketball than Resource A is.
All other things being equal, a free market economy (note that I’m saying a capitalist economy, not your Marxist economy that equates difficulty with labor theory)… a free market economy would value the scarcer resource more highly.
A free market economy would value the Resource B field goal more highly than the Resource A field goal.
Ergo: Difficulty Theory (by Dean Oliver).
I’m just explaining his theory… not in a labor/difficulty construct which you all pooh-pooh… but within your preferred capitalist construct via the allocation (and pricing) of scarce resources in a free market.
Adieu,
TG
T.G. Randini
November 15, 2007
So, dberri, in summary, and you being an economist in a capitalist society… if the ability to convert one type of shot is a scarcer resource than converting another type of shot… why are both shots priced the same in your system?
Jason
November 15, 2007
And there he says he’s leaving again. Why or why do I not actually believe him?
Will
November 15, 2007
What % of team wins are explained by Difficulty Theory? I know WoW explains 94%…
Gabriel
November 15, 2007
T.G.: I rather enjoyed your analysis of Rajon Rondo. I can see the “troll” responses, but come on, your piece was hilarious.
The Ghost of T.G. Randini
November 15, 2007
Thank you Will. A bright question that speaks of intellectual curiosity rather than lemming comment.
The Ghost of T.G. Randini
November 15, 2007
And congrats Gabriel for getting it. (capitalize getting…)
Joey C
November 15, 2007
TJ Randini is an IDIOT.
Kal the Barber
November 15, 2007
Or is he god? He seems to have become some holy ghost…
Ken
November 15, 2007
I just don’t get all of the negative postulation. Just because WoW says that Allen Iverson is a below average player and Tyson Chandler is an above average player, people try to use this to come up with their own arguments as to why the model doesn’t work.
It always seems to be the same 2-3 people that can’t seem to get the idea that maybe Iverson isn’t as productive as we all thought he was, and that having a great rebounder on your team actually helps you win games. I love Iverson; he’s one of my favourite players in the league for the one thing that cannot be judged in the statbook: heart. But that doesn’t excuse the fact that he’s a low percentage shooter who turns the ball over frequently. And anyone who watches Wallace, Chandler and Rodman can see how valuable they are, keeping possesions alive with offensive rebounds, protecting the rim, and gaurding (in Rodman’s case) the opponents’ best low post player. WoW doesn’t turn these guys into valuable players, it just gives us an opportunity to see value where we might not have looked before.
And one last note to T.G., who seems to be adamant that WoW punishes players who take difficult shots with the shot clock winding down: The best player at doing this was Michael Jordan, and WoW just so happens to have him ranked as one of the greatest players of all time. The key is, he does many other things to help his team win, not just shoot the ball 25 times a game, which is also why Kobe (who does other things) is ranked high, and Iverson (who kills his rating with poor shooting and turnovers) is not.
Jon
November 15, 2007
dave, if you could, go back to the deron williams/chris paul comments and post a response to Guy and his theory on missed rebounds. I think this is one of the best, simplest points made. I don’t know much about economics, and have no idea what TJ Randini is talking about, so I am not really interested in that argument. But Guy brings up a simple, interesting point, and I wonder what is is wrong with his thinking. Jason, if you could respond as well, because you clearly have a very good grasp of WoW.
Jason
November 15, 2007
One way to make sure that shots late in the clock with defenders on you (which will always be difficult) to have minimal impact on your WP score is to make sure that someone on the team takes a better shot earlier on in the possession. Easier said than done? Perhaps, but though it’s true that Jordan appeared to be great in these “clutch” situations, more often than not, his shots came earlier in a possession.
It may be true that bad teams exact a penalty on their higher volume shooters when someone on the squad fails to get a good look earlier in the possession. I suspect that this falls more often on a guard who is more likely to have the ball in his hands when this happens. This is likely true because guards handle the ball more often and a large part of the reason that a bad shot has to be taken as the clock winds down is that a big man wasn’t able to the ball down low earlier.
What WP measures is the shot and the make or miss. It doesn’t *directly* measure the entire situation. So is it fair to penalize guards for this? With the caveat that there is error inherent in any model (better models minimize error) and indirect proxy measures are often contribute to this error, I think it is reasonably fair. By comparing guards against guards through position adjustment, players are compared against players more likely to find themselves in this situation. It is also part of a guard’s job to get someone a good shot. In such a case, if a guard is forced to take a difficult shot late, it’s in part because he wasn’t able to help someone get a better shot earlier. Entirely fair and penalizing teams with inept forwards? Sure. Does it work well enough on average? That’s an empirical question and I suspect it does, else WP performance of guards would be highly volatile and would reflect their teammates’ abilities such that their own scores wouldn’t remain terribly constant.
(I’m curious if in fact guards do have more volatility and variance in their Win Scores or WP as this would lead to some credence in this argument.)
Where I see the greatest potential for this to break down in terms of relative player rankings are on teams where jobs are divvied up differently. Teams employing multiple ball handlers (e.g. a ‘point forward’ or three guard lineups) should perhaps have different weights to the position adjustment. I suspect that this is not a common case though, that more often than not teams employ (on average) something close enough to a 5-position lineup such that the adjustments provide an accurate enough gauge of relative contributions.
It may also break down for teams with *very* skilled interior scorers who may well decrease the number of late possession shots guards are forced to take by themselves taking and making many more good shots inside. If true, then the covariance of players like Shaq or Duncan in terms of the WP of their guard teammates will be greater than that of a nearly similarly productive big who contributes largely through defense and rebounds. This is again an empirical question though I suspect it’s difficult to actually address at the severe shortage of offensively gifted bigs means that these data are few and far between. While the hypothesis looks nice to me, I’d be curious to see real data.
Sam
November 15, 2007
Question about wins produced…do the per 48 minute numbers assume a linear relationship? If so, it probably overstates to productivity of players like Nate Robinson (which is acknowledged in the posting). It would probably be fairly simple to look at this relationship and come up with a good function to map observed stats to 48 minute stats
I also wanted to make a point about all of TGs rants. Anyone who has watched a basketball game knows that there is a distribution with respect to the degree of difficulty in a shot. TGs arguments implicitly assume that the distribution of shots is skewed depending on a player’s skill level relative to that of his teammates. That is, the best player always take the hardest shots while the supporting players predominantly take easy shots. Whether or not this is the case is an empirical question, but I suspect that data would not support this assumption. It is likely that the best players take more difficult shots because they take more shots in general. Since the best players control the ball a lot more, they’ll be taking a lot of easy shots too. The distributions may not be any different. If this is the case, than the equal valuation of all shots should not penalize the win scores of someone who shoots a lot. Its also worth noting that shot difficulty can be endogenous in many cases. If a player like Marburry or Iverson barrels into the lane and puts up a contested shot, it is very likely that they could have found a teammate with a higher percentage shot. The contested shot in the lane isn’t forced by circumstance, it is the product of a bad decision. Thus, the win score would justifiably penalize “volume shooters” if their difficulty distributions are skewed.
Even if a star gets ALL of the last second shots, I doubt this will bias things very much. Very few possessions actually end with a buzzer shot.
Kal the Barber
November 15, 2007
I haven’t had any Econ since high school and all I remember are those damn graphs with the intersecting lines. Supply and demand lines, I think, and where they intersected on the graph was the P for price. So if there was more demand for the same amount of supply, then the P or price would go up. Well, maybe that TG guy has got something. If I read him right, if the ability to hit a difficult shot is scarcer than the ability to hit an easy shot, then there will be more demand for the ability to hit the difficult shot and the P or price will go up. Because either the demand is higher for the more difficult shot or the supply is lower for that ability. If I remember correctly, either raising the demand curve, or lowering the supply curve will make the P or price go up.
That’s about all I remember about Econ because a really cute girl sat next to me and I wasn’t focusing all my energy on the econ on the blackboard.
At any rate, if the P goes up with relative scarcity of supply, or a higher demand for something that is harder to do, then the P or value or whatever it is that is rewarded to the player for the shot should also be higher. So I think that TG guy has something there but I’m just a barber, not an economist, so what do I know? Got to run. Got a customer.
PS… I ended up with that cute girl by the way and we have three kids. Thank you for your time and by the way I now appreciate Rajan Rondo more.
The Ghost of TG
November 15, 2007
Sam’s comment: “TGs arguments implicitly assume… the best player always take the hardest shots while the supporting players predominantly take easy shots. Whether or not this is the case is an empirical question, but I suspect that data would not support this assumption.”
It would be great to test this through empirically through detailed analysis of several games. (There are several Finals games available on DVD for start/stop.) I would think, however, that the data would support TG’s observation and conclusions. If Pax or Kerr on the Bulls… or DJ on the Celts… through up too many difficult shots, they’d be benched. Jordan or Bird tossed up the hardest shots. Likewise, Rodman would be benched if he took the shots the Microwave took.
Sam’s “It is likely that the best players take more difficult shots because they take more shots in general.”
Yes, they take more difficult shots because they take more shots in total. But you would also find that a higher % of their shots are graded difficult compared to the average player for the reasons cited above.
Until an expert conducts a detailed visual analysis, it is just a difference of opinion. I am only basing my opinion on decades of observation and a modicum of common sense.
But common sense, like beauty, is in the eye of the beholder.
Sam
November 15, 2007
Jason, great points above. I really like the idea of looking into covariance among teammates, but agree that this would be difficult to set up and interpret. Great idea though. Everyone always talks about how certain players make teams better and I would love to see some statistical evidence on this. Another approach might be to look at a player’s efficiency with and without someone in the lineup. As an example, Shawn Marion had a career FG% of .462 before Nash showed up, and a 0.508 FG% (2.8 SDevs higher!) over the years Nash has been on the team. I’m not sure where you could find data that would allow you to asses this within a season, and between seasons you could really only look at cases where players change teams.
Jon
November 15, 2007
i still think a test that would work great to either prove or disprove wins produced in regard to the rebound talk is to find out what is easier to replace. you can look at any number of teams where a great rebounder(wallace, rodman) is replaced by a not as good rebounder and see the effect on the overall team. for scoring, you do the same thing. look at a team when a great scorer leaves (iverson, 1st example that comes to mind) and see how their scoring/offensive efficiency is replaced. if one is replaced noticably easier than the other, than even if that statistic is more valuable, players who produce it are not as valuable because its easy to do anyways
Kal the Barber
November 15, 2007
I just had a thought. If TG says not all shots are the same because some are harder, well maybe he’s right. But maybe in a different way too. Is a shot that wins the game at the buzzer worth the same value as a shot in the first quarter in a blowout? Doesn’t the game winning buzzer shot create more win value based on the incremental probability of winning than a shot that is made in a forty point blowout? They went from loss to win in the buzzer beating shot, but the win in the blowout was basically sorta guaranteed and doesn’t add as much value, I think.
I ran this by my wife and she said I’m nuts. 2 points is all you got at the end of the game and 2 points is what you got in the blowout and 2 points = 2 points so they are the same.
But they don’t feel the same to me. And I said maybe the point at the end of the year isn’t how many points your team got but how many wins your team got. And so some of the points are more important for getting wins than some of the other points. She was sorta following me and then I said…like when Jordan sank that shot against Utah!!! Before the shot, they had a 50% chance of winning the game but after the shot they had an 80% chance or something like that so one shot alone increased their win probability 30%!! (I’m not good at math so I’m just using those numbers for instances.)
My wife then said I’m either an idiot or an idiot savant. I don’t know what that means but she’s still cute and I love her.
Kal the Barber
November 15, 2007
Oh, I forgot to say, sorry my grammar isn’t good or maybe my math either, so don’t make fun of me like you do with that TG guy. This is the first day I’ve ever written anything and I only wrote because that TG guy said some interesting things, at least the things I could understand, and I’m sorry he is gone and now there is only a ghost .
But I thank him for giving me the courage to write and don’t think he would mock me wherever he has gone.
Sam
November 15, 2007
Ghost, good counterpoints. My intuition from years of observation would lead me to similar conclusions, but intuition can be fairly misleading when it comes to statistics. What we think the distribution is based on observation is subject to recall bias, which probably influences how we perceive star players. In any case, you’re right, video analysis is the way to get at this question. Maybe I’ll take notes next time I watch a game and report back.
Owen
November 15, 2007
I agree with Jason and also with Sam, whose point is quite correct I think. Yes, Jordan and other high volume shooters take more difficult shots sometimes. However, the entire offense is usually built around getting them shots. So they get a lot more easy shots relative to the average player to go with the difficult shots.
It seems like nearly every day someone makes the argument that scoring doesn’t matter to WP. When that happens, I always think about Shaq. and the fact his WP would be nearly 30% higher, i.e. 5 wins per year over his career, if he were an average free throw shooter.
Sam
November 15, 2007
Kal, I think you’re on the right track. In terms of wins produced, points in a blow out are worth less than points in a tight game. In a blow out, you can lose a couple points and still win. So, between games, points may be of different value, but within games they are all the same. If you don’t make the baskets in the first quarter, you won’t even be in a position to take a game winning shot at the last second.
I think you and your wife are both right, you’re just talking about different things. One might account for this by looking at the margin of victory in each game and applying the appropriate weights to a player’s production. This seems like a lot of work though, and might not yield much more info if the blow outs and OT games offset each other over the course of a season.
The Ghost of TG
November 15, 2007
Sam, I agree there might be some hindsight bias. But I’ve seen numerous instances of average players getting yanked for taking bad shots. I never saw Michael or Bird get yanked. I did see, however, the coach call a time out right away. That’s why Michael respeceted Phil. As in that one LA game, “Who was open, Michael? Who was open?”
But then again, if you were Michael and had spent a few years passing the ball to Brad Sellers and watching him consistently miss layups and easy shots, you have a tendency of growing blinders.
When Michael had a) better teammates and b)learned to trust them… he never lost again for a long time.
But I still say Pax or Kerr never attempted a difficult shot. With two players collapsing on Michael (if not three), they were almost always wide open when they got the ball.
The Ghost of TG
November 15, 2007
Kal, you’re a lot more savant than idiot.
Kal the Barber
November 15, 2007
Thanks, Ghost. I told my wife and she didn’t believe me so I showed her the site and now I think I’m gonna get lucky tonight.
Jon
November 15, 2007
just because if shaq improved his free throw percentage to average and this improves his WP, that doesnt mean that it isnt bias against scorers. rebounders have a clear advantage in this system. the question isnt whether there is an advantage over scorers, but whether or not they deserve it. Berri and everyone who believes in the method believes rebounding is important. Others believe that inefficient scoring at a high level does have some value. That is the debate. And the proposed test I gave I believe gives a pretty good answer to the question. I just have not done the test, nor do I know how.
Joey C
November 15, 2007
I like cats.
Will
November 15, 2007
I never got an answer to my question…
I think I’ll stick with the metric that most accurately explains NBA wins. You can nitpick that 6% all you want, but 94% is good enough for me.
Joey C
November 15, 2007
Will, doesn’t wins produced “explain” 94% of of NBA wins only because 0.94 is the R^2 of a regression of win probability against points score per possession and points given up per possession. If so, then couldn’t we call a new measure wins produced 2.0 that simply takes those points and give THEM all to the starting center w/ everyone else getting 0 pts. Wouldn’t wins produced 2.0 still expalain 94% of wins?
Ghost of TG
November 15, 2007
Joey’s Wins produced 2.0:
TG would be proud of you. Well said.
Jason
November 15, 2007
This is a very simple request, Randini. Please behave like a civilized individual who is actually interested in discourse, else please take your act elsewhere.
Jim P
November 15, 2007
Actually, i was rather enjoying the discoursing going on between Depp and Moss.
It’s not often you have celebrities posting on this site.
Joey C
November 15, 2007
On the APBRmetrics thread there is this comment:
“There’s a fellow going by the name of T.G. Randini who is currently dogging Prof. Berri. in the comments on his blog. The there and back (mostly the there) is quite amusing, and I was just wondering if anyone here want to fess up to another alias.”
http://sonicscentral.com/apbrmetrics/viewtopic.php?p=18212
Steve
November 15, 2007
I’m confused. I’ve been following the train of thought in all these postings and enjoyed all the commentary. Joey C says something extremely pertinent. Ghost compliments him on his comment. And then “Jason” makes an extremely rude remark. Is Jason the god of this site or just a disher out of extremely punk remarks?
Pete
November 15, 2007
Steve,
I agree with you that Joey makes a good point. I think wins produced is a good statistic, but saying it “explains” 94% of wins is very disingenuous. What “explains” 94% of wins is points scored and points given up. How points scored and surrendered is then broken down into wins produced is one of an unlimited number of ways. Joey’s example is a strawman, but technically it seems accurate.
Pete
November 15, 2007
TG Randini is a real person. I’ve never seen him, but ironically he is the author of a very eloquent review of “Wages of Wins” that partly motivated me to read it last year.
Jason
November 15, 2007
“Steve,” it appears that a number of successive posts have been deleted. Prior to my post there was a series of short posts attributed to “Kate Moss” and “Johnny Depp,” all of which (based on content) appeared to be the product of whomever is posting as Randini. It appeared very much like someone decided that he was going to use the board as his own playground, spoiling any real discourse. I suspected strongly that the Randini personality was this individual. If you did not see those posts, then my comment no longer has the context.
Am I the god of this board? No. However, I recognized uncivil behavior in the form of some rather juvenile posts and wanted very very much for them to stop.
dberri
November 15, 2007
Some good comments today. I deleted a bunch of silly stuff. Hopefully I will have more time tomorrow to comment.
Kent
November 15, 2007
“Starbury” is a reverse aptronym b/c he’s not really a star.
Kent
November 15, 2007
Dean Oliver writes about the Wages of Wins debate: “This debate is tinged with more emotion than logic too much of the time. As though the world will end unless so-and-so converts to the Religion of Stat X.”
http://sonicscentral.com/apbrmetrics/viewtopic.php?p=18212
Kent
November 15, 2007
Incredibly Harold Almonte is posting the same comments on that thread as well. This guy is ubiquitous. I’ll tell you what you’ve got to admire his passion for this stuff.
Pete
November 15, 2007
Does anyone know of any studies on how much preseason projections should be shaded after the first week of the season? How much information is there in performance in the first 5 games? In other words, Was bill Simmons right to increase his projection from 49 wins to 62 wins for the Celtics?
Also, happy birthday Randini! :-)
dberri
November 16, 2007
Pete,
Simmons should have said 62 wins from the start. But he couldn’t see that until he saw them play.
Harold Almonte
November 16, 2007
Kent-Yeah, you need to call it passion, given that I neither am a statician (I’m engineer), nor I have bussinesses related to sports, nor even I’am an english speaker, just a lover of the game pushing for a correct evolution of the game math interpretation. But, do not deceive yourself, I was almost a professional player, I ‘ve been coach, a former basketball teacher, and one of my former kid pupils is actually a latin NBA player.
Kent
November 16, 2007
Harold A,
Thanks for sharing your background info. And keep up the enthusiasm and passion for statistics and basketball.
Patrick
November 18, 2007
I skipped a lot of the T.G. trash talking. But I thought this was funny:
“Resource A: The ability to shoot and score on a wide-open lay-up with 20 seconds left on the shot clock.
Resource B: The ability to create a shot and score while being double-teamed with two seconds left on the shot clock.
I believe that Resource B is far more scarce in high school basketball, college basketball, and professional basketball than Resource A is.”
T.G.: You are correct, resource B is far more scarce.
But while resource B is a scarce resource required for winning accolades at the playground, it isn’t a resource that winning basketball teams require very much (if any) of, so its scarcity isn’t very relevant to the economics of winning.