Debating the Value of Carmelo Anthony on Christmas

On Christmas Day, prior to the New York Knicks game with the Oklahoma City Thunder, Chris Herring (of the Wall Street Journal) and I engaged in a friendly Twitter debate*.

It all started with the following from Chris:

OKC obviously has huge advantage w Melo out. But worth noting that NYK nearly won last seasons meeting at MSG when Melo sat out bc of injury

I took issue with the word “huge”.  And hence we had the following exchange:

 @HerringWSJ If Melo was really a superstar, OKC would have a huge advantage with Melo out.

@wagesofwins Not when he’s 11-1 against Durant for his career, and routinely outplays Durant. That’s a bit shortsighted to say, IMO.

@HerringWSJ Sorry. Melo has never been close to Durant in productivity. Sample sizes and controls matter in doing analysis.

@wagesofwins Also, look at the first five years of each player’s career. Their numbers are identical per 36 mins. No difference.

@HerringWSJ Durant was above average across his first five seasons. Melo was not. Not sure what you are looking at.

@wagesofwinsMelo: 46.2% FG% on 18.7 FGA, 7FTA, 24 pts, 6 rebs, 3 ast Durant: 46.8% FG% on 18.1 FGA, 7FTA, 24.8 pts, 6 rebs, 2.7 ast

@HerringWSJ Durant’s effective field goal percentage was 50.7. Melo’sEFG was only 47.9. That difference matters.

@wagesofwins Sure it does. But I already made my pt: To act as if Carmelo Anthony makes no impact or helps his team by not playing is silly.

@HerringWSJ I didn’t say that. I said his loss wasn’t “huge”. Never said he didn’t have a positive impact. Just not a superstar.

Okay, let me add a bit more to this conversation.  Chris focused on the first five years of each player’s career.  And across these seasons he argued these players were not different.  Wins Produced, though, tells us these players were not the same.

To see why, let’s look at each player’s career numbers after their first five seasons in the NBA.








Scoring Factors

Adjusted Field Goal Percentage




Free Throw Percentage




Field Goal Attempts




Free Throw Attempts




Points Scored




Possession Factors

Offensive Rebounds




Defensive Rebounds












Miscellaneous Factors





Blocked Shots




Personal Fouls




Wins Produced per 48 minutes




Box score statistics are the career averages – per 36 minutes – across the first five years of each player’s career.   Average small forward performance taken from 1994-95 to 2012-13.

An average NBA player will have a Wins Produced per 48 minutes (WP48) of 0.100.  So according to Wins Produced, Melo was below average across his first five years while Durant was above average.

To understand why Melo was not “good”, consider his performance relative to an average small forward.  Across his first five years, Melo was slightly above average with respect to offensive rebounds, but below average with respect to defensive rebounds, turnovers, and steals.  So when it comes to possession factors, Melo wasn’t much different from an average player.  A similar story is told with respect to assists, blocked shots, and personal fouls.  Melo was slightly above average with respect to assists, but didn’t help with respect to blocks or fouls.

That means that Melo’s only significant contribution to wins – again, relative to an average small forward – had to be via scoring factors.  And if we look at points scored, it looks like Melo was amazing.  But points scored are determined by shot attempts and shooting efficiency.  Although Melo was “good” at getting to the free throw line and converting (and that is good), he was not good at getting his field goal attempts to go through the basket.  That means his lofty scoring totals were really just about Melo “taking” shots.

Once again, the operative word is “taking”.  When Melo left the Nuggets the team’s shot attempts really didn’t change.  That means that Melo simply “took” shots from his teammates.  Since his conversion rate was below average, the shots Melo took weren’t actually helping the Nuggets win much.

Now contrast this production with Kevin Durant.  Like Anthony, Durant’s production with respect to possessions and the miscellaneous factors listed above was not impressive. So if Durant was going to help, it was via scoring where this had to happen.   Unlike Melo, though, Durant really did help with his scoring.  With respect to shooting efficiency from the field and the line, Durant was above average.  So because Durant could get his shot to go in the basket, his shooting helped his team win games.

All of this was based on the first five years of each player’s career (a focus that Chris took on Twitter).  What about this year?

This season – after 30 games – Carmelo Anthony is an above average player with a WP48 of 0.168. In fact, Anthony is leading the Knicks with 3.56 Wins Produced.  When we look at the individual stats –  at– we see that Melo has improved with respect to possession factors (i.e. rebounds and turnovers).  But with an effective field goal percentage of 48.2%, we see that Melo is still not helping with his shooting from the field.

Like Melo, Durant has also improved. After 29 games, Durant has produced 7.2 wins with a 0.317 WP48.  When we look at the individual stats, we see that Durant is above average with respect to defensive rebounds, steals, assists, blocks, and personal fouls.  And of course, he remains above average – from the field and the line – with respect to shooting.  Consequently, Durant is one of the most productive players in the game.

So let’s return to the original question.  Is the loss of Melo “huge” to the Knicks?  He is leading the team in Wins Produced.  But relative to other top players, Melo is not that productive. And this is primarily because his scoring – despite his lofty point totals – really doesn’t help that much.

All of this means I stand by my argument that Melo is not a “superstar” and his loss is not “huge”.  The Knicks were  bad team when Melo played.  They look bad when he doesn’t play.  And as long as Melo remains the most productive player on the Knicks, I think you can expect “bad” and “Knicks” to be used in future sentences.

- DJ

* – at least, I think it was friendly.  As an academic economist (i.e. someone lacking in social skills), I may not always know when my questions are not considered “friendly” :)

Comments are closed.