Deron Williams for MVP?

Kenny Smith, at Yahoo! Sports (and an analyst with TNT), posted a list of MVP candidates this week.  The list begins with Kevin Garnett, Dwight Howard, and LeBron James.  It also includes Steve Nash, Tim Duncan, Tony Parker, and the subject of yesterday’s discussion, Kobe Bryant. 

In the midst of these players, though, was the following name and argument:

3b. Deron Williams: Surprise. Some people think his teammate, Carlos Boozer, leads the way for Utah. Williams, however, is the engine that makes the team go. He’s the reincarnation of Jason Kidd, only with a jump shot.

Williams, along with Chris Paul, has already been the subject of the following two columns at the WoW Journal.

Chris Paul vs. Deron Williams, Again

Defending Chris Paul

When I saw Smith’s comment, though, I thought I would comment on where these players are 18 games into the 2007-08 season (through Monday’s games).

Deron Williams and Chris Paul after 18 Games 

Let me start with the numbers for each player.

Table One: Williams and Paul after 18 games in 2007-08

As Table One reveals, relative to the average point guard each player is quite good.  When we compare Williams and Paul to each other, though, it’s clear who the most productive player has been this season.  Although Williams has the advantage with respect to shooting efficiency from the field and blocked shots, Paul has the advantage with respect to free throw percentage, scoring, rebounds, steals, assists, turnovers, and personal fouls.

With such dominance in so many categories, it’s not surprising to see Paul post a Win Score of 13.5 per 48 minutes while Williams only has a mark of 8.3. In sum, Paul is more productive than Williams (and as the above posts note, this has been true since college).  Let me phrase this with a bit more force. Although Williams is good, his production doesn’t come close to what we see from Paul.

The Utah Jazz and New Orleans Hornets after 18 games

Of course, someone might note that the Utah Jazz – led by Williams – is a better team than Paul’s Hornets.  And that’s true.  After 18 games the Jazz had a mark of 13-5.  The Hornets only won 12 of their first 18 contests. 

The difference in won-loss record does not quite illustrate the true difference between each team.  The Hornets currently boast a 3.7 efficiency differential (offensive efficiency minus defensive efficiency), which suggests the Hornets should have only won 11 games this year (or 50.5 wins across an 82 game season).

Meanwhile the Jazz have a differential of 9.0. Such a differential suggests this team should have won 14 games (or 64 wins across an 82 game season). In sum, the Hornets look to be a solid playoff team this year.  But the Jazz appear to be serious contenders for an NBA title.

When we look at Wins Produced – which connects efficiency differential to the individual players – we see why the Jazz are so good.

Table Two: The Jazz and Hornets after 18 games in 2007-08

Thus far Williams has produced 2.6 wins with a 0.183 WP48 [Wins Produced per 48 minutes].  Although such a mark is excellent, it’s only the third best mark on this team.  Carlos Boozer has currently produced 4.8 wins with a 0.387 WP48 (yes Kenny, Boozer does lead this team).  And Boozer is not alone.  Andrei Kirilenko boasts a 0.315 WP48 with 4.1 Wins Produced.   Additionally Ronnie Brewer, Paul Millsap, and Matt Harpring are also above average performers.

Turning to the Hornets, we see a team that’s dominated by Paul.  Thus far, Paul’s WP48 stands at 0.388.  Yes, Paul has a higher WP48 than Boozer (if ever so slightly), as well as both Kobe Bryant and Andrew Bynum (the topic of yesterday’s column).  Paul has also produced 4.7 wins this season and is on pace to have 21.6 Wins Produced by the end of the season.

Unlike Williams, though, Paul is not getting much help.  Except for Tyson Chandler, no other New Orleans player is on pace to produce double figures in wins.

Now someone might argue that Williams has better teammates because he makes his teammates better.  When we look at the career performance of Boozer and Kirilenko, this is a hard argument to make.  Both players have been consistently above average performers in their career.  And Kirilenko was actually a more efficient shooter before Williams arrived.  Plus in 2003-04 he had a higher WP48.  Turning to Boozer, although he has become a more efficient scorer with Williams, he’s also a much better rebounder.  In other words, the leap Boozer has made in wins production is about more than just scoring.

Evaluating the Individual

The trick in evaluating individual performance in a team sport is to separate the player from his teammates.  This is fundamentally why statistics are kept in the first place.  We wish to understand the contribution made by each individual, independent of teammates.

When we look at the numbers that separate players from their team, it’s very, very clear that Paul is a much, much better player than Williams.  And thus it’s Paul who should be considered a legitimate contender for MVP.  Williams, though, does play for a better team.  And as noted, this is because Williams has better teammates. Consequently, it’s not surprising to see people confuse the outcome observed for the team with the actual accomplishment of the individual.   In other words, we can expect more and more people to side with Smith in this debate.

Ultimately, though, this is not a democracy. 

Well, then again, it might be.  It’s just that the data has many, many more votes.

– 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

Wins Produced vs. Win Score

What Wins Produced Says and What It Does Not Say

Introducing PAWSmin — and a Defense of Box Score Statistics

Posted in: Basketball Stories