Teams often turn to the NBA draft to change their fortunes. A few days ago, Erich Doerr provided a wonderful analysis of how much talent teams will find at the end of this June. Although there does appear to be some players who can help next year, it’s important to remember one basic fact about NBA rookies: Most rookies suck!!!
“Suck”, by the way, is a technical term often employed by economists for resources that do not perform up to expectations.
Okay, I just made that up. But Seth Stevenson at Slate.com did provide a wonderful defense of this word, and I think it does apply in this case.
Given the inability of rookies to consistently make much of a contribution, teams might turn to other places to find talent. And one of those places is the free agent market.
The Unrestricted Free Agent Market
According to ESPN.com, there are 100 unrestricted free agents available this summer (included in this list are free agents with early termination options, player options, and team options and I am going to pretend all of these are actually available). Of these, 58 appeared in at least 41 games this past season and also averaged at least twelve minutes per game (which means 42 did not meet these minimum thresholds).
Table One reports how well each of these players performed with respect to Wins Produced per 48 minutes [WP48] and NBA Efficiency per 48 minutes [NBA48].
Table One: The Unrestricted Free Agents of 2008
NBA Efficiency -as The Wages of Wins argues – is not a great measure of actual productivity (despite being highly correlated with Game Score and PER). But NBA Efficiency is a very good representation of how talent is perceived (by both the media and NBA insiders). In contrast, WP48 is connected to team wins. But it’s not as correlated with player salaries and the voting for post-season awards. In sum, each measure gives a different perspective on performance and can be thus used to determine which free agents are likely to be overpaid or underpaid.
Before we get to that subject, though, let’s start at the top. At the top of Table One we see a name that’s likely to be worth what he’s paid. Shawn Marion is the top unrestricted free agent in terms of NBA48 and WP48. So although he will help somebody next season, he may not be much of a bargain. This same story also applies to Antawn Jamison and Baron Davis.
Looking a little further down the list we see DeSagana Diop, Bonzi Wells, Michael Ruffin, and James Posey. Each of these players is ranked as above average in WP48 but below average in NBA48. So unlike Marion, Jamison, and Davis, each could be a bargain. Of course, Wells, Ruffin, and Posey are 31 years old, so if these players are a bargain, it may not be for long.
Searching for bargains is not the only issue in the free agent market. Teams also want to avoid over-priced players. And at the bottom of Table One we see some players are probably going to command more than they are worth. At the top of the list of overpaid players is Jermaine O’Neal. A few days ago I noted that J. O’Neal was the most overpaid player in the NBA this past season. And although NBA Efficiency indicates he’s still a productive player, WP48 tells a different story.
Table Two: The Career of Jermaine O’Neal
That specific story is told in Table Two. Table Two is an update of a table I posted last January. And with this table I noted the following:
When we look at O’Neal’s career we don’t see a major star. His career WP48 is above average, but a mark of 0.143 (prior to the 2007-08 season) pales in comparison to the top players in the game. Still, O’Neal has generally been a good player and he did lead his team in Wins Produced for three seasons.
What makes O’Neal “good” is his ability to get rebounds and block shots. Shooting efficiency, though, has been a consistent problem. Except for the 2002-03 season – again, his best year – O’Neal has always been below average with respect to getting his shots to go in the basket.
And this season the inefficient scoring issue has worsened. In addition, O’Neal is now below average on the boards. As a consequence, his overall productivity is now well below the average mark.
As noted, these were my words in January. And these are the same words I would say about O’Neal today. With respect to shooting efficiency, rebounds, steals, and turnovers, J. O’Neal is below average. So although he’s still taking and blocking shots, his overall contribution is below par. Consequently, if J.O’Neal departs Indiana – or Indiana pays to keep him – it’s likely J. O’Neal will be paid more than he is worth.
O’Neal is not the only player who will be potentially overpriced. Stromile Swift, the second pick of the 2000 draft, has never produced as much as his draft position suggested. For his entire eight year career he has only produced 13.3 wins [with a 0.060 WP48]. His career NBA48 of 23.8, though, exceeds the average mark of 22.2. So it’s possible that another team will take a chance on Swift (and once again be disappointed).
Looking at Those That Didn’t Play
Table One only considers free agents who played significant minutes in 2007-08. What of the 42 free agents who didn’t play in 41 games (and/or 12 minutes per game)?
Topping that list would be Elton Brand and Gilbert Arenas. Brand is 28 years old and a veteran of nine NBA season. Due to injury, he only played 274 minutes in 2007-08 [with only a 0.058 WP48]. In his career, though, he has produced 107.2 wins and posted a WP48 of 0.219. If Brand is healthy in 2008-09, he can be a player that pushes an average team into contending status (or a below average team like the Clippers back to average).
Like Brand, Arenas has also been hurt. Unlike Brand, though, Arenas is not quite as productive as people seem to believe. For his career he has produced 47.1 wins with a 0.140 WP48. Yes this is above average. But hardly worth the maximum contract Arenas was demanding before he got hurt. So Arenas may not be a bargain.
For those still searching for bargains, consider the names Trevor Ariza and Chris Andersen. The merits of the latter I discussed last March. As for Ariza, his career WP48 stands at 0.163 and he is only 22 years old.
Of course if you don’t like young players, Dikembe Mutombo is once again available. Once again Mutombo stepped in for an injured Yao Ming and performed extremely well. In 619 minutes he posted a 0.271 WP48, a mark that is well above average. Of course, Mutombo is also past 40 years of age. So if you wish to see a free agent actually log major minutes on the court, Mutombo may not be your man.
Let me close by noting one last old guy. Brent Barry – who I labeled “The Better Barry” last March — is also listed as an unrestricted free agent. He posted a 0.215 WP48 this past season and could probably help someone in 2008-09. Of course Barry could have helped someone besides the Spurs this past season but refused to leave San Antonio. And this was after the Spurs traded him to Seattle. So Barry will probably not leave this summer either.
Okay, those are my thoughts on the unrestricted free agents. Next week I hope to look at the restricted class. And that examination will address the question: Who made a worse decision last summer – Luol Deng or Ben Gordon?
– DJ
The WoW Journal Comments Policy
Our research on the NBA was summarized HERE.
The Technical Notes at wagesofwins.com provides substantially more information on the published research behind Wins Produced and Win Score
Wins Produced, Win Score, and PAWSmin are also discussed in the following posts:
Simple Models of Player Performance
What Wins Produced Says and What It Does Not Say
Introducing PAWSmin — and a Defense of Box Score Statistics
Finally, A Guide to Evaluating Models contains useful hints on how to interpret and evaluate statistical models.
Wages of Wins Journal 
W.C.
May 30, 2008
How do you value the ability to create your own shot?
Some players get high efficiency scores because of their high FG%, but they couldn’t score more than a dozen or so points a night without their efficiency rating falling through the floor. They can’t create for themselves and have limited athletic ability . They mosly get their points off offensive rebounds, fast breaks, defensive breakdowns etc….
On the flip side, there are players that are wildly talented athleticly. They score a lot, but only earn mediocre efficiency scores because their FG % is not that high. That happens because they take a lot of tough shots.
IMO the players in the latter group could easily raise their FG% and efficiency rating dramatically by being much more selective. They don’t because on many plays the alternative scoring option would be even worse. Basically, their ratings take a hit for the simple reason that on many plays they are best scoring option of a number of poor options because of their superior ability to create.
IMO, some guys with high rating that are low scorers are not nearly as good as their advanced statistical ratings and others that are high scorers are much better than their ratings but are getting punished because of their role as the primary offensive option on the team.
If a guy like David Lee went to Denver and was told to score 25 a night, his ratings would fall horribly.
If a guy like Iverson was told we only want 10-15 points a night because David Lee is going to option #1, his ratings would be higher because he would only take the very best shots each night.
Of course the team would get a lot worse if they did that because David Lee trying to get 25 points a night would be total fiasco.
W.C.
May 30, 2008
Let me address this issue another way.
IMO, given the choice, Allen Iverson taking a slightly off balance 20 footer with a defender on him is a better offensive option than David Lee taking an open 15 footer.
In addition, the probability of Allen Iverson finding himself in that position with little time on the clock is much higher than David Lee getting a really good shot.
Therefore, AI is often forced to take that tough shot because it’s the best available option.
Yet his ratings suffer for it and David Lee’s do not because Lee sits around waiting for an offensive rebound or offensive breakdown to do his limited scoring. That’s all he can do!
There is something wrong with the wildly superior athlete and scorer being punished for having the ability to create and make tougher shots and therefore being “forced” to take them as the best scoring option on a lot of plays.
Daniel
May 30, 2008
Lame. Teams like San Antonio and LA have guys that can create their own shots, but are more effective when they go for the best shot possible, rather than waiting around for their stars to create something out of nothing.
If the “wildly superior athlete” is able to draw the attention of the defense and create an open shot or a layup for a teammate, the wildly superior athlete is more valuable when he actually takes advantage of the opportunity to pass the stinking ball to get a better shot.
W.C.
May 31, 2008
I agree that it is wisest for a team to always take the best shot available.
But when you actually watch basketball, you soon realize that very often the clock is running down but there is no good shot available because the defense is solid.
Someone has to shoot anyway!
Unfortunately, the best available shot is often actually a very mediocre or poor one. That shot is usually taken by one of the more athletic guys on the team as he tries in desperation to create something good or a guy with a decent outside shot shooting from 20 despite being covered fairly well.
Being one of “GO TO” guys automtatically translates into a lot of more bad shots. That reflects poorly on the player statstically relative to players that can’t get open, can’t create for themsleves, have no outside shot etc… The latter type of player is rarely the one taking the mediocre or bad shot because if they took it, it would be a horrible shot (even worse). It is the latter type of player, that is sometimes forcing the superior player to take so many of those bad shots.
Obviously, there are some super athletic players that force things when they don’t have to and that SHOULD reflect poorly on them.
However, I think it’s almost preposterous to reward high efficiency that results from a lack of athletic talent, lack of ability to create, lack of ability to get open, lack of ability to hit an open shot from outside and punish guys that are taking a lot of tough shots because they HAVE TO.
J. Prince Lawrence
May 31, 2008
As a Minnesota fan who watched the huge disparity between Al Jefferson’s effort on offense versus defense, I am forced to disagree with PER; he is definitely not 13th best player in the league. Maybe 13th best scorer, but not 13th best player. His defense is just not there.
WP’s 46th rating is a much better rating, since it reflects how Big Al is a second tier star. I think we can see how the different basketball metrics rate offense by seeing how they rate Big Al; the higher rating he gets, the more the metric overrates scoring and underrates defense.
Jason
May 31, 2008
W.C., you appear to focusing on half the game, that being those times when the team has the ball. You also appear to be engaging in the same hypothetical critique of an efficiency model that has been hashed and rehashed here. Missing from such a critique is whether or not a model that ignores the ability to ‘create one’s own shot’ is an ability that correlates with winning games. You postulate as if this has already been demonstrated to be so. Is it?
JB
May 31, 2008
Most rookies suck as rookies… especially compared to the best known starters and main subs.
But many get better.
Do they suck in year 1 vs the comparably paid? Maybe the gap isn’t as wide. I think on average the draft returns decent to good value over time.
There are different parts of the draft and the free agent market. Good points of each, varying year to year. The best GMs understand the different markets and shop where they can and try to max return.
Does Boston retain Posey and House? That will be interesting within budget constraints but they may also be able to find willing and suitable replacements if necessary.
W.C.
May 31, 2008
Jason,
I’ve tried to find high quality comprehensive defensive stats, but I’m not sure they exist. If they do, they apparently aren’t available publicly.
For me, it usually comes down to visual observation in combinbation with stats because IMO stats are almost always limited in some ways, but at times so are my observations.
I realize that things like rebounding, assists, turnovers, personal fouls etc… are important, but so is the shot creation and the “GO TO GUY” factor. IMO, these latter factors can actually count as negatives statistically even though they are actually positives.
Let me use David Lee and Jamal Crawford as an example.
David Lee is typically highly rated partly because of his scoring efficiency. Jamal Crawford is rated poorly because of his poor FG% and suspect shot selection.
IMHO, if the Knicks has 5 players with David Lee’s creativity and outside shooting ability (size adjusted to fit their position) they would be the worst team in basketball. The reason I say that is because they wouldn’t be creative enough to get many high quality shots and they couldn’t hit enough tough outside shots to score much either. IMO, it would actually be comical to see how bad they were. And it’s not just a PF thing. There are plenty of PF’s that can shoot and create.
If you did the same exact thing with 5 Crawfords, IMO they would have no problem scoring a 100 per night.
In reality, what happens is that a guy like Crawford can go to the hoop and draw Lee’s defender to himself, but if he dished it out to Lee for an open 15 footer, that would actually be a much worse scoring option than simply forcing up a tough shot that he’s only 35%-40% to hit. He would also be better off chucking up one of his silly 3 pointers (that make us squirm) than dishing for Lee for an open outside shot. Taking that option hurts Crawford’s stats, but helps his team more than going to Lee.
IMO, high scoring creative PGs, SGs, and SFs are often penalized statistically for their role as the GO TO GUY that is often forced to take a bad shot while their less talented and less skilled teammates accumulate solid efficiency stats BECAUSE they are inferior or INCAPABLE scoring options.
I am not sure how to capture this in “WINS” statistically because almost every team has a few creative or decent outside shooters that take all those tough shots and get penalized for making that contribution. The only way their value would be measureable would be if we actually got to see a team with 5 David Lees win 15 games for the season and a team with 5 Crawfords win 35.
I think this is one of those things that you have to intuit, use common sense to understand, and see visually because the math will not get you there.
David
June 1, 2008
W.C.,
First, it is not useful to think of hypothetical teams with 5 of the same player. A baseball team with just pitchers or a football team with just quarterbacks will do poorly.
What’s more useful is to ask questions like, “If you switched David Lee with an average PF, how will that team’s record change?” We do have a few historical cases of this sort that seem to indicate that WP correctly values (or de-values) the “go to” effect.
For example, when Allen Iverson was traded for Andre Miller, Prof. Berri correctly predicted exactly how much the 76er’s would improve by:
At the time of the trade, the 76er’s were 5-18. Prof. Berri correctly predicted that their remaining record would be 30-29. That is pretty remarkable if you ask me.
This post may also be of interest to you, as it directly addresses the issue you raise:
With all that said, I am in agreement with you that the ability to create shots should be counted for something. I proposed in the past that unassisted FG’s should count for more than assisted FG’s, and that the difference should be equal to the value of an assist. That way, a self-created made shot and an assist + shot both have the same contribution to team-total-WP. I believe this is a theoretically sound approach to rewarding shot creation.
W.C.
June 1, 2008
David,
I understand your point and agree totally.
Unfortunately, that’s the only way I can think of to illustrate my point that creativity and being the “go to” shooter often HURTS a player statistically even though they have clear cut and ESSENTIAL skill sets that some of their more efficient peers do not.
The bottom line is that someone has to be in the shooting and scoring role. Whoever it is will be hurt statistically for taking a lot of bad and tough shots.
For arguments sake, let’s say I am Crawford’s agent and have been told his next contract will be based entirely on his efficiency and other advanced statistics.
I am going to advise him to stop shooting so much even if it hurts the team.
I’ll suggest that whenever there only a few seconds left on the shot clock and the choice is to take a tough shot or dish to QRich or Lee from the outside, he should opt for passing and let those guys throw up bricks.
He should also refrain from taking some of the other tougher shots he takes and do the same thing.
The result would look something like this:
1. His PPG would go down a bit
2. His FG% would rise – perhaps as much as 5%
3. His assists would rise because some of the bricks those guys throw up would actually go in.
4. Lee’s efficiency would fall through the floor
5. QRich would have even worse stats (assuming that’s even possible) LOL
6. The Knicks would be a worse team.
7. Stats people would perceive Crawford to be better than they do now and Lee to be worse than they do now.
Now let’s assume the Knicks get a great outside shooting SF and PF.
The result would look something like this:
1. His PPG would go down a bit
2. His FG% would rise – perhaps as much as 5%
3. His assists would rise even more because those guys woulod hit a lot of open shots.
4. The new SF and PF would hit a higher % of their open shots than Crawford would taking the tough shots he is forced to take now. So the Knicks would be a much better team.
5. Stats people would have a much higher opinion of Crawford than they do now.
W.C.
June 1, 2008
I think the key is to differentiate between players that have a low efficiency rating because they take a lot of poor shots even though there were better scoring options on the court and those that are forced to take a lot mediocre shots because they often ARE the BEST scoring despite having only a mediocre shot to take.
At the same time, you have to discount some of the high efficiency ratings of players that are hitting a high percentage of their shots because they only take the easy ones that fall into their laps and because they are INCAPABLE of creating for themselves or hitting from the outside.
Anon
June 1, 2008
Jamal Crawford is not at all a good example of your point. He shoots too many shots and misses all the time because he has no idea what a good shot is. It’s not like the knicks lack players who are willing to shoot.
A better example of your point would be Dwyane Wade on the Heat when he was actually playing. You could make an argument that playing on the equivalent of a development league team forced him to take a lot of bad shots.
And I like David’s idea of crediting the player less for a shot if it was assisted.
MarkT
June 2, 2008
I also like David’ s idea of reducing the credit given a player for an assisted FG. One of the issues I have with WoW is it dispenses too many points to individuals: A rebounds or makes a steal to start the possession, B gets an assist and C makes a basket. Team gets two points, But Winscore dispenses 3.5 points. Doesn’t make sense to me. Conversely, if A rebounds, B makes a good pass but C misses, team gets 0 points and net individuals is 0 so that makes more sense, although it is a shame that B’s good pass does not result in some value while A’s rebounds gets too much credit perhaps (especially since A’s rebound may be the result of a defensive equivalent of an assist – i.e., someone else may have caused the shot to miss and A may just be in the right place and do little work for the rebound).
Another problem with assists though is that you don’t get an assist if the shooter is fouled and does not make the FG, even if it is a great pass and they make both FT.
Jason
June 2, 2008
W.C., you seem to be missing quite a bit in your ‘analysis’ of efficiency stats.
I’ll admit that my eyes glazed over when you presented the hypothetical “5 David Lees” argument. I’ve found that particular argument (usually 5 Rodmans) to be inane and meaningless and has so very, very little to do with reality as to be worthless. Models often do not perform well outside of the range of normal variation and observation. Since there has never been a “5 same player lineup” it’s a meaningless exercise designed to prove a point that it cannot prove. An analysis of a model that requires positing something that never happens to refute it is worthless.
Indeed at least in Dave’s metric, being a high percentage shooter alone does not result in exceptionally high ratings. One must hit a high percentage of his shots *and* take shots to gain points since the position adjustment assumes an average contribution at the position. Failing to shoot means no score. David Lee rated highly because he rebounded well *and* did not detract from his rating by missing many shots.
If you “create” bad shots, you are not helping your team. Perhaps someone does have to shoot, but I think the data indicate that most low percentage shooters aren’t overly influenced by teammates or the “need to shoot.” I know the tired argument that suffers because as the clock runs down someone has to shoot has been presented. In actual analysis fewer players, even remarkable scorers like Bryant, wind up taking a disproportional percentage of their team’s ‘shot clock about to expire’ shots. It’s certainly not what contributed to Jamal Crawford’s problems. The Knicks shot 15% of their shots in the final 3 seconds of a shot clock. Crawford similarly took 15% of his shots in this time window. The ball wasn’t finding him with a “need to shoot” any more than it was finding other players.
MarkT
June 3, 2008
I think it is interesting that Jamal Crawford is in the top 10 in adjusted plus/minus.
http://basketballvalue.com/topplayers.php?mode=summary&sortnumber=90&sortorder=DESC&year=2007-2008
Jason
June 3, 2008
Funny too that Chris Paul gets a negative score in adjusted plus/minus. There appear to be some bugs in such methodology.
Anon
June 3, 2008
They should trade Paul and build around Peja and Melvin Ely
Eli W
June 4, 2008
I’ve tried to replicate Dan Rosenbaum’s method of splitting adjusted plus/minus into offensive and defensive components.
http://www.countthebasket.com/blog/2008/06/03/offensive-and-defensive-adjusted-plus-minus/
By doing that Chris Paul rates as the second best offensive player in the league…but the third worst defensive player. The method has a high degree of uncertainty though when considering just one season of data. It would be helpful to look at more seasons to see how Paul’s defense rated in the past.
MarkT
June 5, 2008
That chart was interesting, Eli. But I wonder why almost all the bad defensive players are PG, and none of the good defensive players are, ignoring the Price fluke?
Eli W
June 5, 2008
It’s probably better to just compare players within their position. Since it’s rare that a team will take their PG out of the game and replace him with a big man, adjusted plus/minus isn’t ideal for directly comparing players at different positions.
Barlo
June 5, 2008
Looks like something that a geek not a bball player came up with. STAT RATS!
mrparker
June 10, 2008
I.e shot creation,
Most of the players who take bad shots that they have to create on their own are sg and pgs. They are all being compared to each other. Shouldn’t that mute the effect of shot selection.
Shooting guards seem to be the most highly over/underrated players in these discussions.
I thought about this so I went to make an all time list of retired shooting guards of my generation(anyone who played in the 80s up until now).
No.1 is obvious. After that you can’t make a definitive argument for any player to be in a top 3.
This doesn’t happen with any other position. Shooting guard is overall a pretty crappy position. Only one guy ever really executed it well over a long career.
Napha
July 19, 2008
Thank you for this post.