Today’s post is going to focus on the relative merits of Caron Butler and Carmelo Anthony. But before we get to those merits, though, I want to start with something noted by Henry Abbott at TrueHoop.
Maxing Out Arenas
Abbott posted the following story on Tuesday:
Gilbert Arenas Throws Down the Gauntlet to the Wizards
Abbott’s story links to a blog entry from Agent Zero where he makes the following statement about his pending contract negotiations:
I want to get a six-year deal. I want to be a max player. If my team decides that they don’t want me here any more and they’re going to go in a different direction, then I got to look elsewhere. For me to look elsewhere, I want to go find a championship team who’s a championship contender. I’m going to have to take less money, but I’m willing to do that to win a ring. If my team doesn’t want me, then I’m going to another team and I’m going to take less money to go there.
Abbott offered the following interpretation of this statement:
I’m sure he doesn’t mean it this way, and maybe we’ll have a clarification soon, but it seems like Arenas is saying that he’d play for passion, and to win in some other city. But if he has to play with these scrubs in D.C. …
Meanwhile, “those scrubs” just beat the Celtics twice in a row.
And, looking at it from the other side, if you had a championship quality team, would you use your mid-level exception on a guy who dominates the ball? Having Gilbert Arenas on the floor re-orients the stars in your offense. (If you’re Cleveland, I think the answer is yes. But San Antonio? Phoenix? Dallas? You can love Arenas from here to the moon, and still swallow hard before noodling with your roster in that particular way — especially when you consider that he’s little, not a great defender, and injury prone.)
More than anything, this might be an example of why players angling for maximum deals should probably have agents manage the process. It would be smart to avoid even the slightest chance of needlessly pissing off the only good team on the planet that can afford to give you big money.
When I read Abbott’s story I immediately thought that a post examining the value of Gilbert Arenas would be a good idea. But it turns out; I already wrote such a column four weeks ago.
Surviving the Loss of Agent Zero
After reading over what I said last month I realized that I didn’t have much more to say on the subject. As I noted in December, Arenas is definitely a “good” player. But he’s not a great player. And it has to give the decision-makers in Washington something to consider that without Agent Zero – the player demanding the maximum NBA salary –the team is posting the best efficiency differential seen in Washington since 1979. One would think the loss of a player who truly deserves maximum money would have a more severe impact on a team’s fortunes.
Agent Zero on Caron Butler and Carmelo Anthony
Not wishing to go over this topic again, I decided to go look at what else Gilbert Arenas had to say. In the midst of a very lengthy post was the following observation:
Caron Butler ain’t getting credit for what he’s doing. He’s never gotten credit for what he’s done for us. He’s been in the shadows a little bit of me and Antawn, but it’s the Big Three. There’s no order of priority. He’s just as important as the two of us are to the organization and he’s showing it right now how smart of a move it was bringing him aboard for Kwame Brown. He’s given us more than we expected from him, a lot more.
If you list the top three-men in the league, for me, it goes:
1. LeBron
2. Carmelo
3. Caron
Okay, I can agree that LeBron is the best small forward. But Melo ahead of Butler?
Butler vs. Melo
Let’s just go right to the data. Table One reports what these two players have done across their respective careers, what they did last season, and their performance in 2007-08.
Table One: Comparing Butler and Anthony
The statistics in Table One are separated into three categories: Scoring, Possessions, and Miscellaneous. I am not going to talk about the miscellaneous stuff (assists, blocked shots, and personal fouls), where each player is pretty equal. What I am going to focus on is scoring and net possessions.
Scoring Totals vs. Scoring Efficiency
When we look at scoring, one might quickly conclude that Anthony has the clear edge. Per 48 minutes, Anthony consistently scores more than thirty points. Although Butler is above average, he can’t match this output.
But in evaluating scoring, totals shouldn’t be our focus. The key is efficiency. In other words, players should not be rewarded for simply taking shots, but rather for getting the shots they take to go in the basket. Rewarding shot taking ignores what the Wizards have done this season. Without Agent Zero, a prolific hoister of shots, Washington has simply found other people to shoot. Again, taking shots is not the skill that needs to be considered in evaluating performance. Making shots is what matters.
Of course when it comes to making shots, Melo did have the advantage prior to this season. His level of shooting efficiency, although below average for his career, still bested Butler. This year, though, a different story is told. As noted last month, Butler has both increased his shot attempts and shooting efficiency in 2007-08.
I would emphasize that Anthony still has the advantage with respect to getting to the free throw line. And that’s an important skill. But his advantage with respect to efficiency from the field has vanished this season. Consequently his ability to produce more wins than Butler via scoring has also vanished.
The Net Possession Story
As I often note, player evaluation tends to stop with scoring. And as I often note, the other stuff in the box score does matter. When we turn to possessions – rebounds, steals, and turnovers — we see where Butler has the clear edge. Although Anthony has the advantage this year on the boards, Butler has the advantage with respect to steals and turnovers. When we look at net possessions (rebounds +steals -turnovers), we see that Butler has a 1.6 advantage per 48 minutes this season. Given that each possession is worth about 0.033 wins, per 48 minutes the net possession difference between Butler and Anthony is worth 0.052 wins. Or, if each player logs 3,000 minutes this year, Butler’s advantage with respect to possessions is worth 3.3 additional wins over an 82 game season.
When we look at Wins Produced and WP48 [Wins Produced per 48 minutes], we see that Butler in 2007-08 is on pace to produce 17.2 wins with a 0.257 WP48. Anthony’s WP48 stands at 0.115, which is actually a career high, but well behind what Butler is offering. And when we look at Wins Produced, we see that Melo is only on pace to produce 7.3 victories.
It’s important to note that Butler doesn’t just best Anthony this season. Turning back to Table One we see that Butler’s Win Score- a metric (like Wins Produced) that considers all box score stats (scoring, possession, and miscellaneous) – is consistently higher than the mark we see from Melo. This is true this season, last season, and across each player’s career.
Summarizing the Story
Agent Zero is right to be impressed by his teammate. Butler is playing amazingly well this season. But I would argue he should be even more impressed than he has indicated. When you look past scoring, you can see that Butler has clearly produced more than a Carmelo (a player currently on pace to start in the 2008 All-Star game).
Let me close by noting that this story may not be entirely welcome news for Arenas. If we focus strictly on scoring totals – as is often the case – then one might argue that Arenas is a maximum contract player. But if we go beyond points per game, and consider the impact of all the statistics on wins, then we conclude that
a. Butler is more productive than Melo (good news for the Wizards) and
b. Arenas might not be worth maximum money (bad news for Agent Zero)
The first statement might make Arenas happy. I am pretty sure the second statement, though, is something Arenas wouldn’t believe. But both the data — and what the Wizards have done this season — suggests that maximum money for Agents Zero may not be a good investment for the folks in Washington.
UPDATE: Commentator Dan noted that the numbers for Carmelo Anthony in Table One for 2007-08 were incorrect. Turns out I switched the labels on 2006-07 and 2007-08. When you get the labels right, it changes my analysis of net possessions. It is still the case the Butler has the advantage, but it is not as big as I originally stated. The WP48 and Wins Produced numbers, though, are still what I said originally. Butler is worth about 10 more wins than Anthony this season. Thanks to Dan for catching this error.
– DJ
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.
SHC
January 18, 2008
I think that every one of your ‘evaluation’ posts has had the same message – conventional wisdom is wrong, wins-produced is right.
What I find disturbing is that you’re giving people the impression that you’ve computed something that’s indisputably unique and correct (at least within the regression framework).
I still don’t understand why you wouldn’t simply publish exact mathematical details on the web (I’m pretty sure that these are far from trade secrets of technological companies.). I wouldn’t go about preaching some theory without publishing every technical detail.
Anon
January 18, 2008
aren’t we supposed to buy his book to know the details of wins produced?
also, his evaluation posts don’t always go against conventional wisdom, it’s just that no one cares when they don’t.
Rashad
January 18, 2008
The full calculation for wins produced IS up on the web. Here is the link: http://www.wagesofwins.com/CalculatingWinsProduced.html
Rashad
January 18, 2008
Oh, just noticed dberri posted a link to it in the post itself already under technical notes.
Pete23
January 18, 2008
Another great post. Cheers to dberri!
Oren
January 18, 2008
Just out of curiosity, is Butler statistically better then the other players that Arenas listed?
mrparker
January 18, 2008
dissing arenas and carmelo…this ought to cause a great comments thread
Kevin Broom
January 18, 2008
I don’t see the diss of either Arenas or Anthony. In fact, I don’t think Dr. Berri is the only person to offer this type of critique of either player. Wizards fans are debating whether the team might actually be better without Arenas (they’re not, but there’s still a debate). There have been doubts about whether Anthony is a true franchise player since he came into the league.
Owen
January 18, 2008
The Nuggets are providing some very interesting food for thought for basketball economists. Allen Iverson’s performance has been eye opening. His ts% this year is currently at 57.2, which is 3% higher than his previous high for a year, and is 5.5% higher than his career average. It’s interesting, his shooting numbers have been significantly higher in terms of ts% since signing his four year extension before 03-04. He is also taking 3 less shots per 36 (b-r numbers), his usage is down, his rebounding is down a bit, and his PER, despite this astonishing leap in shooting efficiency, is just 21.7, barely above his career average.
Marcus Camby meanwhile is posting the best rebounding numbers of his career.
The Nuggets are also playing the fastest pace in the league. This may have something to do with trying to take advantage of playing at altitude. To me though it seems like it might also be a rational response to the problem of how to get Carmelo and AI enough shots to keep them happy.
And despite having two offensive “superstars” they are only the eighth best team in the league. Even more surprisingly, their offense is only ranked 14th. The strength of the team has been its sixth ranked defense.
Altogether, quite a conundrum….
Costa
January 18, 2008
The Wizards certainly wouldn’t be “better” without Arenas. The question is whether or not he’s worth his asking price. dberri is arguing here that, while Arenas is certainly a very good player, he’s not worth max money when looking strictly from a standpoint of his production on the court.
However, one can definitely make the case that he makes up for it with the significant off-court factors in his favor. Few rival Agent Zero’s commercial and marketing appeal, and that’s a very important point to factor in.
Costa
January 18, 2008
Owen, if you or anyone else here that’s interested haven’t already seen it yet, Kevin Pelton wrote up a nice piece analyzing what Denver has been doing on defense to be so efficient so far this year. Click my name to check it out.
With his surprising rise in shooting efficiency, as well as the fascinating defensive tidbits brought up in this article, I’m wondering if this isn’t the best basketball AI has played in his entire career!
Brian
January 18, 2008
I’ve been watching the Wizards closely this year. It’s amazing the difference in the team with and without Arenas. I like him, but without Gilbert, the Wizards are clearly a better defensive team.
Is there a good measure of individual defensive performance in the NBA? What about an in-line-up vs. out-of-line-up comparison of team defense adjusted for opponent scoring strength? That would tell us a lot.
And SHC, what’s the point of a post that regurgitates conventional opinion? There’s enough of that out there already. And “disturbing” is a disturbingly strong word. Al Qaeda threats are disturbing. Sports stat blogs might be bad, dumb, mistaken, biased, but not disturbing.
Liam
January 18, 2008
There is one thing I never like about stat comparisons. Caron is better than Anthony. Who are they being defended by? Anthony has been considered a dangerous scorer since he came into the league and as such has almost always commanded the best defender on the team they are facing and has seen his share of double teams; can we say the same about Carron? I am not saying that Caron is not a great player but as with any competitive sport stats can NEVER tell the whole story. The only way to come up with this would be to create stats based on a defender by defender basis. For example what are the career statistics of both player when facing Bruce Bowen as a defender or Ron Artest. Yes Caron has done great for the Wiz but don’t make a decision based on a stat line that still leaves out many factors. Stats can be twisted and turned in any number of ways to bring the point across that we want.
Tball
January 18, 2008
In the end, Gilbert was being too kind to both Melo and Caron as neither belongs in the top 3 three men. Maybe he shouldn’t put Caron ahead of Melo, but he should put him behind at least a half dozen other three men.
Also, it may be worth overpaying Gilbert accepting that he is not a great player. If he flees, does Washington have the cap room to replace him? Do they anticipate having the draft position to find his replacement there? If Arenas fills the seats (as self-absorbed players seem to do) and gives them their best chance to contend over the next 3-5 seasons, there must be a value there that extends beyond his win score.
Owen
January 18, 2008
TBall – One of the interesting findings of the WOW is that self-absorbed playing doesn’t fill seats, winning does. So that really is not a reason to keep Arenas.
It’s very difficult to make the case that Arenas should be given a max contract, but undoubtedly someone will…
Owen
January 18, 2008
edit: Doesn’t fill seats in home arenas that is, they do fill seats on the road, but that doesn’t help your revenues….
antonio
January 18, 2008
Maybe I am wrong about this, but I thought that a team did get a portion of revenue from their road games. Can anybody clarify? It would not be much, but I thought teams got a little bit
dberri
January 18, 2008
Antonio,
As far as I know, the NBA does a 100-0 split. Home team keeps everything in the regular season. That is what I have been told by people in the NBA.
magicmerl
January 19, 2008
And in the playoffs too, I believe.
dberri
January 19, 2008
No, the playoffs are different. There is a split in the playoffs between the teams and also the league.
nate
January 19, 2008
Nobody in the league is sufficiently under the cap to offer Arenas a max contract. Philly will have $10M in cap room. Atlanta will have $11M if they dump Childress.
Washington is bidding against themselves. Ernie Grunfeld can offer the max (about $16M) if he wants, but he can probably get away with something in the $12-13M range.
Dan
January 19, 2008
I just have a question about the calculations you used, David. Looking at the season stats as posted on ESPN.com, it appears your 2007-2008 numbers for Anthony are off in all of the possession categories except turnovers. From what I can tell, his rebounds per 48 minutes should be 9.0, steals should be 1.7, and turnovers should be 4.5 (like you show). Is there something I’m missing? I don’t know if that would change the substance of your argument at all, but it certainly pushes the numbers in Anthony’s favor a little, at least for this season when he seems to have made some strides in expanding his game.
dberri
January 19, 2008
Dan,
Thanks for catching that. As I noted in the update, I switched the columns for 2006-07 and 2007-08. Given that I am writing these posts at 11pm, I am surprised this kind of stuff isn’t happening more often (or maybe it is and no one is checking my work).
Ryan
January 19, 2008
Just found the website in a link from TrueHoop. I think your research is very interesting and while I’m not economist or mathematician, I think it has a decent amount epistemological value. You make good points with regards to Arenas but you could argue that he’s worth a max contract for two reasons: The previously stated reason of factors off the basketball court- he has electrified DC and is the one sports figure this town can really cherish and call their own ‘superstar’. And two, Gilbert changes the game- when the game is on the line, he can win it singlehandedly. As great as Caron is, he can’t do that (yet).
As for the fact that the Wiz are playing better defense- some of that is attributable to the absence of Gilbert but the majority of the credit should go to Randy Ayers and Haywood’s and Blatche’s improved defensive intensity. You also can’t discount the fact that mentally, the Wizards ‘feel’ like that they need to play better D without Gilbert in the lineup.
On a completely unrelated note, I attend William and Mary and one of your coauthors, Professor Schmidt taught me macro last semester. I can’t believe I didn’t even know he was partly responsible for this work. (Good, easy going professor too). Keep up the good work!
Dan
January 20, 2008
Just glad I can help, David. I appreciate your work, most especially when it shows how some of my favorite players can and should be improving their game. I’ve been an Anthony fan since he was at Syracuse, and I think this year he may finally be paying attention to the parts of his game that will help bring some more wins. Maybe he’s been reading WoW?
Steve
January 20, 2008
Caron Butler is definitely top 3 among SF’s in the NBA this season. This is his breakout year as a true franchise player.
Lebron James, Paul Pierce and Caron Butler have been the 3 best allaround small forwards in the NBA this season (In no order). Carmelo is 4th, Tmac would be there if he played as he did last year and wasnt injured so much.
So that “Half dozen sfs that are better” comment is an absolute JOKE.
grover
January 20, 2008
I was just looking at the table for this post and I had a thought:
Would it be difficult to indicate standard deviations for each stat shown?
The tables already show whether the player is above or below the average player at his position. With the use of std dev, we could get an idea of just how exceptional he is. Perhaps italic for 1 dev and bold for 2 devs?
Regardless, keep up the good work! Thanks for helping to change the perception that points are all that matter when evaluating a player.
Kevin Broom
January 21, 2008
As it happens, I recently wrote a piece for RealGM.com about whether the Wizards are better without Arenas. The answer: no. To summarize, they’re worse offensively — about the same shooting (slightly worse, actuall), but their turnovers are up a bit and they’re not getting to the FT line as frequently.
Defensively, the improvement appears to be independent of Arenas. I say this on the basis of hand-tracking Wizards defensive data during much of the first quarter of the season, including most (if not all) of the games Arenas played. This season, (despite (or maybe because of) knee woes) he was a much improved defender over previous seasons. Specifically, he played within the team’s scheme — forcing penetration to prescribed areas, challenging shots in his area, pinching down to help on post play, and not getting out of position by gambling for steals. He was still tending to lose track of his man on the weakside, and over-committing at times on help D, but overall a much better defender.
I know that the standard basketball calculus (Arenas out + defense better = defense better because Arenas out), but attributing the improved defense to Arenas’ absence just isn’t correct. Credit for the improved defense goes in no small part to their new defensive assistant, Randy Ayers. Ayers installed a new defensive scheme, and the players have bought into it. It also doesn’t hurt that the team actually has its best defender (Haywood) on the floor for more than half the minutes.