About two weeks ago Jeff Zillgitt argued in the USA Today that DeMar DeRozan is “stepping up” in his second season in the NBA. 

More comfortable and confident in his second year, DeRozan wants to stand out.

DeRozan has 14 games of 20 or more points — but 11 since Dec. 1. Those include 37 points against the Houston Rockets on Dec. 31 and 30 against the Miami Heat on Jan.22. He is developing into what the Raptors envisioned for the ninth pick of the 2009 draft — an athletic wing who can score near the rim or on mid-range shots.

“I know and understand the game a lot better than I did my first year. I was learning on the go,” DeRozan said.

Toronto Raptors coach Jay Triano challenged DeRozan to make the rookie-sophomore game during All-Star weekend this season. DeRozan didn’t play in the game last season.

“He’s on a little bit of a mission,” Triano said. “I told him ‘Your goal is to make the sophomore game.’ When you look at his numbers, it’s a pretty good argument as to why he should.”

Raptors president and general manager Bryan Colangelo said the team has been meticulous in helping DeRozan improve.

“When I say that, it wasn’t only the things he was doing with our coaches on the practice court and coming back to get shots up every night,” Colangelo said. “But it’s also investing in the process of developing a young player.”

DeRozan started 65 games in his rookie season.

“It’s a learning experience you can put a price on,” said Colangelo who has studied similar athletic wing players in the first two years of their careers, guys such as Los Angeles Lakers guard Kobe Bryant, Detroit Pistons swingman Tracy McGrady, Houston Rockets guard Kevin Martin and San Antonio Spurs forward Richard Jefferson.

While not saying DeRozan will be just like any of those players, Colangelo said, “His development is trending in the right direction.”

Let me summarize the argument made by Triano and Colangelo (a summary based on the statements in bold above): After much work by DeRozan and the coaches in Toronto, DeRozan is improving.  We know this because this is the story the numbers tell.

If you review the article, the only numbers mentioned are scoring totals.  When we look at all the box score numbers, though, a different story emerges.   The following table reports DeRozan’s numbers in each of his first two seasons in the league.  The numbers in red are marks that are below the average numbers we see from an NBA shooting guard.

Let’s start with what DeRozan does well.  Relative to an average shooting guard, DeRozan is above average with respect to taking shots from the field and the free throw line, turnovers, and blocked shots.  Because he takes so many shots, he is also above average with respect to scoring.  And since – as the above story illustrates – people focus so much attention on scoring totals, DeRozan appears to be a “good” player; or at least a player trending in that direction.

But when we look at the other numbers, a different picture appears.  In his second season, DeRozan is below average with respect to shooting efficiency (from the field and line), rebounds, steals, assists, and personal fouls.  His assist numbers are especially low, suggesting that once DeRozan gets the ball he thinks first (and second and third) about taking a shot.  And who can blame him?  This strategy has led to praise from both his head coach and general manager. 

When we put all these numbers together, we see that DeRozan has posted the following Wins Produced per 48 minutes [WP48] numbers at shooting guard (it appears DeRozan has also played some small forward, so his overall numbers – reported in the table below — are somewhat lower that what is shown right here):

• 2009-10: 0.054
• 2010-11: 0.026

So in both his first and second seasons, DeRozan has been below average as a shooting guard (average WP48 is 0.100).  But to the extent that two data points establish a trend, the trend line is not going up (as Colangelo contends).  No, it appears that DeRozan’s overall productivity is declining.  Certainly he is taking more shots and scoring more points.  But because he is not scoring efficiently, or doing anything else to help his team win games, DeRozan should not be thought of as a player the Raptors can build around.  One suspects, though, that this is exactly what the team is thinking.

The discussion of how the Raptors view DeRozan should remind people of past discussions of the Raptors and Andrea Bargnani.  Like DeRozan, Bargnani was a lottery pick by Colangelo and the Raptors.  And like DeRozan, Bargnani was

• below average last year and
• is offering even less this year. 

To see how much less, let’s look at where the Raptors are as a team this season. After 55 games in 2010-11, the Raptors have posted a -6.0 efficiency differential (offensive efficiency minus defensive efficiency).  Such a mark is consistent with a team that would win 17 of their first 55 games (the team has won 15 times).  When we move from efficiency differential to Wins Produced, we can see where these wins are coming from.

As the above table notes, the team is being led in Wins Produced by Amir Johnson, Jose Calderon, and Ed Davis.  Each of these players is posting a WP48 mark above 0.200 (or twice the mark of an average player).  In general, teams with three players who are twice as good as average tend to be very good.  Unfortunately this team is being held back by the players who lead this team in shot attempts.

If you look at the statistics for the Toronto Raptors, we see the following players leading the team in field goal attempts per game: Andrea Bargnani, DeRozan, Leandro Barbosa, Linas Kleiza, and Sonny Weems.  These five players are the only players to attempt at least ten field goals per game.  And these five players have combined to produce -5.7 wins, with only Barbosa posting a WP48 mark in the positive range (and Barbosa is only barely above zero).

The evaluation of DeRozan suggests the Raptors’ decision-makers focus on scoring.  And when we look at the productivity of all the players who lead the Raptors in shots attempts, it seems that the problem in Toronto – remember this team has only won 15 games – might be with how players are evaluated.  Players seem to know that if they take shots and score, the Raptors decision-makers will be pleased.  Taking the actions that lead to wins (i.e. shooting efficiently, rebounding, taking care of the ball), don’t seem as important.  

Once again, the Raptors have only won 15 games this season.  When we look at the individual players, it appears that the players holding this team back are the players who are leading the team in shot attempts.  And it appears the people making decisions in Toronto can’t see past the scoring these players provide.

- DJ

P.S. If you would like even more discussion of the Toronto Raptors, click on over to NBAEh?  This blog – from Devin Dignam – examines the Raptors from the Wins Produced perspective.  In addition to providing some great analysis of the Raptors, Devin’s site also offers a list of Required Reading for anyone interested in Wins Produced. 

Weekend Podcast! The link is at

http://www.nerdnumbers.com/podcasts/STAT_02_12_2011.mp3

but of course smart readers like you are using the RSS feed at:
http://www.nerdnumbers.com/podcasts/stat_feed.xml

(Advanced->Subscribe to Podcast in iTunes)

The cast of the podcast:

• Andres Alvarez — http://nerdnumbers.com/
• Arturo Galletti – http://arturogalletti.wordpress.com/
• Mosi Plott – http://miami-heat-index.blogspot.com/
• David Berri – http://dberri.wordpress.com/

The discussion ranged from why we blog to various basketball related topics. The following links related to some of what was discussed.

• How to trick science oriented people.
• Web Comics oddly have a lot to do with running a stats blog
• This is what it’s like when world’s collide –
• Refs are biased? Home-court advantage is real?  Scorecasting – by Tobias Moskowitz  and L. Jon Wertheim — presents evidence that home-court advantage and referee bias are both real and related.
• Statdom and Fandom in one book?
• Anyone Catch the Reference?
• Not only does Fields get boards, he sells Jerseys. . . . umm kinda –

Beyond the podcast links, here are a few more items that you might find interesting…

• My latest for the Huffington Post — If Deron Williams Was Chris Paul, Maybe Jerry Sloan Would Have Stayed – explores the role expectations might have played in Jerry Sloan’s unexpected retirement. 
• Perhaps in response to the Celtics (Arturo’s team) defeating the Heat (Mosi’s team), Neal Paine – at basketball-reference.com – has explored how a team’s record against elite teams forecasts a team’s performance in the playoffs.  Neil’s answer: “…knowing how a team performed vs. elite teams actually tells you less about who wins a playoff series than a team’s record against all teams, even non-elite teams. Team records vs. the cream of the crop certainly sound meaningful, but when it comes to predicting success or failure in the playoffs, you’d be better off knowing how they did against the entire league.”
• Would you like to watch a collection of the greatest shots in NBA history?  Well, then Hoopism is the place to be today.  Be warned…clicking on this link could cause  you to lose a good chunk of your day!
• Tom van Riper – of Forbes.com and Fields of Green – argued this past week that the Knicks should pass on Carmelo Anthony.  The longer this trade takes to happen the more one suspects that these kind of arguments are carrying the day in New York (then again, Melo could get traded tomorrow for all we know).
• Ken Pomeroy recently posted an interesting discussion of plus-minus.  The entire discussion is worth reading.  The second-to-last sentence, though, seems especially note-worthy:  In basketball analysis, we should be filtering out randomness, not embracing it.

Comment on Comments

When I saw Ken’s discussion I wondered about the comments this would generate.  Plus-minus fans tend to be passionate in defense of their preferred metric, so I suspected Ken’s critique would lead to some interesting responses.  Such comments, though, never appeared.  And that is because Ken Pomeroy – like Gregory Mankiw – doesn’t allow comments on his blog.  One can contact Ken via e-mail, but he also explains on his blog that this may not generate a response.  More specifically, this is what Ken says about efforts to contact him:

I get many e-mails. While I read each one that crosses my desk, I am not able to respond to all of them. Actually, I am able to, but in some cases it’s better for the both of us if I didn’t, since it would take away from doing other things that may serve humanity better. If you want to increase your odds of getting a response, proper salutations and punctuation help, especially if we do not have a previous relationship.

There’s also a sweet spot in terms of length. E-mails of 1500 words describing your philosophy on every matter pertaining to offensive rebounding are unlikely to get a response. With messages of such length, it is time consuming to thoughtfully respond to each of your points, yet it would also be rude to respond with “I agree!” to such a long message.

On the other end, messages such as “are u insane?” or “do u predict totals?” are not likely to get a response, either.

After reading that, if you still feel like I am the type of fellow you want to correspond with, I am reachable by e-mail at ratings@kenpom.com .

As readers of this forum know, comments are allowed in this forum. And I even participate in the comment forum from time-to-time.  Upon reading the policy adopted by Ken Pomeroy, though, I started to have second thoughts about my policy.  Certainly I would save a great deal of time if I just eliminated comments entirely. 

Although following Ken’s lead is tempting, I have decided against this idea.   Because I do read all comments, though, I have decided to change the process by which these are moderated.  Specifically, I have decided that all comments will now have to be approved before they get posted.  The purpose of this approach (an approach used by others) is to eliminate “trolls” (i.e. people who wish to leave the same or similar comments over and over and over and over again).  If you believe this new policy is “unfair” to you and “restricting your speech”, well… I could just go with the policies of Ken and Greg :) 

One last note… the moderation of the comments will still follow the WoW Journal Comments Policy (a policy that was updated today).

- DJ

Editor’s Note: About a week ago I posted a story detailing how the Houston Rockets were defying the Pareto Principle.   Teams tend to get between 70% to 80% of their wins from their top three players (i.e. about 80% of wins tend to come from 20% of the team’s roster).  The Rockets, though, have been getting much more than 20% of their wins from the “rest of their roster” (i.e. players not in the top three on the team in Wins Produced).  This discussion might have led someone to wonder, “how do other teams look with respect to the Pareto Principle?”  As one can see below, Arturo Galletti – from Arturo’s Amazing Stats – has provided a quick answer (which originally appeared at Arturo’s Amazing Stats).

Andres Alvarez changed my life.

Let me explain. Back in April of last year, Prof. Berri had a contest. Pick the best Win Producer for a day and get a prize. I participated and built what I thought was a nifty tool in excel to aid me. Dre went out and gave us the power of Automated Wins Produced.

It’s more addictive to me than meth and posting batman images.

Oh God. They're making it more powerful.

He’s the one who kickstarted this blog by reducing the complexity of analysis by a few order of magnitudes. So you can all blame him for anything I do in the future or have done in the past in this space :-)

So, since he get’s a kick out of them, I decided to put up a visualization of the Automated Wins Produced for this season (as of an hour ago). The charts are:

• Done by conference
• Projected over an 82 Game season (for ease of reference)
• Show the Top three players for each team and the rest of the team.

The results are interesting:

Dre, I hope you enjoy this one and it inspires some more cool tools for us to play with. Which might inspire some more posts. Which might inspire some more tools ….

A delicious cycle (Image courtesy of xkcd.com)

Tonight Ray Allen hit his 2,561 career three-pointer, breaking the all-time mark by Reggie Miller.  This event led me to wonder, how does Ray Allen compare to Reggie Miller?  In fact, I can imagine people spending some time this evening debating this issue (those some people don’t include me… my wife and I are watching the Lakers-Celtics game but my wife doesn’t really care to discuss this issue).  

So let me offer some thoughts.

These thoughts – as is often the case in this forum – involves some numbers.  Let’s start with Wins Produced and WP48 [Wins Produced per 48 minutes].  The following table reports each player’s career numbers (Ray Allen’s numbers are prior to Thursday night’s game).

As one can see, these players are quite similar.

• Each player was close to average his rookie season.
• Each player hit his peak with respect to Wins Produced and WP48 at the age of 25 (players tend to peak in their mid-20s).
• Each player had his second best season at the age of 28 (players don’t decline very rapidly after their peak).
• Each player had his third best season after the age of 35 (Reggie Miller at age 38 and Ray Allen so far this season; and yes, it is possible — but not likely — that a player can keep playing well into his thirties).
• Except for Miller’s rookie season, each player was always above average (average WP48 is 0.100; and good players in the NBA – unlike what we see in baseball and football — tend to be above average throughout their careers).
• Their career marks are quite similar.   Ray Allen currently has a career WP48 of 0.168 while Reggie Miller had a career mark of 0.164.  After the age of 35, Miller had a career mark of 0.169.

When we turn to the individual stats we again see that these players were very similar.  

Both their career Win Score per 48 minutes [WS48] and their WS48 mark in their peak season were essentially the same.  And this production was achieved in much the same fashion.  Obviously both were amazing shooters.  Miller, though, was a bit better.  Then again, Allen was a bit better on the boards.  The net effect is that both have a level of productivity that is essentially the same.

So there you go.  If you got involved in a fight tonight arguing that Reggie Miller (or Ray Allen) was clearly the better player (or worse player), well… you might have wasted some energy.  At least, I think the answer is both players are quite good.  Furthermore, I think both players are pretty much the same. 

- DJ

P.S. By the way, Ray Allen’s record is not the big story tonight.  Jerry Sloan departing the Jazz is definitely something I will comment upon.  Hopefully I can get those comments posted tomorrow.

Update: Received an e-mail that noted that there was a mistake in Ray Allen’s numbers.  Allen played for two teams in 2002-03 and his numbers with Milwaukee and Seattle from that season were listed as separate seasons.  This has now been fixed.  That change, though, changed the story somewhat with respect to the second and third best seasons of Allen’s career.  The basic story — Allen and Miller are quite similar and each peaked at 25 — remains the same.

Carmelo Anthony clearly wants to leave Denver.

Okay, when I started writing this story, this seemed clear.  Now we hear that Carmelo Anthony might stay in Denver.   And if that happens, this story doesn’t work as well.  So let me proceed as if I never saw Melo say he “would take a real hard look” at staying in Denver.

Okay, Melo wants to leave.  And one suspects this is because Carmelo doesn’t think he can win a title with his current teammates in Denver.  But what if he had LeBron’s teammates?  In other words, what if the Denver Nuggets could send their entire roster to Miami – except Melo – for everyone with the Heat not named LeBron? 

One might think such a move would mean Melo and the Heat’s current roster (minus LeBron) would be a title contender.  And therefore, Melo would want to stay in Denver.

It turns out, though, that even if this move was possible, I think Carmelo would be ultimately disappointed.

Melo with the Nuggets

To see this point, let’s start with what where the Nuggets stands after 52 games this season.  The team has won 30 games; and the team’s efficiency differential (offensive efficiency minus defensive efficiency) of 3.0 is consistent with a team that will win about 49 games across an entire season.  Yes, this is a good team. But not exactly a title contender.

When we move from efficiency differential to Wins Produced – reported below – we can see who is primarily responsible for Denver’s success this season.

As one can see, Carmelo is an above average player (average WP48 or Wins Produced per 48 minutes is 0.100).  But in terms of Wins Produced, Nene Hilario, Aaron Afflalo, and Chauncey Billups have been more productive this season.   Although these teammates have been “good”, their collective WP48 of 0.117 (i.e. the team’s WP48 without Carmelo) is not good enough to transform the Nuggets with Carmelo Anthony into a title contender.

Melo with the Heat

But what if Carmelo got to play with Dwyane Wade, Chris Bosh, and everyone in Miami not named LeBron?  To answer this question, let’s look at the Miami Heat after 52 games.

Miami has won 38 games this season and their efficiency differential stands at 8.4.  This mark is consistent with a team that would win about 63 games across an 82 game season (and that is a bit of understatement about the team’s quality since they have suffered some injuries).

When we turn to Wins Produced – reported below – we can see which Miami players are producing these wins. 

Not surprisingly, LeBron is the most productive player in Miami.  But King James is not a one-man team.  Dwayne Wade, Chris Bosh, James Jones, Mike Miller, Eddie House, and Udonis Haslem are all above average players.   Although LeBron has a number of good teammates, his teammates are not much better than Melo’s teammates.  Collectively LeBron’s teammates have posted a 0.116 WP48.  Yes, LeBron’s teammates are slightly worse (okay, not statistically different) than the teammates currently playing with Carmelo.  And that suggests that if Melo and LeBron switched teammates, LeBron would still be playing for a title contender.  And Carmelo would still be disenchanted with Denver.

The Point of the Story

Obviously this trade was purely hypothetical.  Although hypothetical scenarios can be interesting, one would hope such an exercise would have some purpose.  And here is what I think is the point of this story.

Carmelo wants a trade so he can contend for a title.  And he thinks if he had better teammates this would happen.  But the problem in Denver isn’t Carmelo’s teammates. It really is Carmelo.  Right now – before he signs any extension — Anthony is the highest paid in Denver.  He is clearly the player Denver has built around.  But although Carmelo has been given productive teammates, his team is still not contending.

The explanation for this is simple.  Carmelo Anthony is simply not that productive of a player.  Yes he can score.  But his overall production – as noted several times – is not consistent with his reputation.  So no matter where Melo plays, unless he gets some teammates that are even better than LeBron’s mates in Miami (and these are pretty good teammates), Melo is going to be disappointed.

- DJ