NBA Stats: Why Defensive Rating Is The Worst Advanced Statistic

Feb. 19, 2012; Indianapolis, IN, USA; Indiana Pacers center Roy Hibbert (55) loses the ball as Charlotte Bobcats center Bismack Biyombo (0) and Charlotte Bobcats power forward Boris Diaw (32) defend at Bankers Life Fieldhouse. Indiana defeated Charlotte 108-73. Mandatory credit: Michael Hickey-USA TODAY Sports

I’m an NBA stats freak.

When I was a kid, I nerded out on Knicks box scores every morning, and as I’ve grown up, I’ve come to love websites like basketball-reference.com.

I use the Per 36 minutes stat in almost every long-term analysis I write for Roberto Gato, and I can’t get enough of almost everything on that site.

I say “almost everything”, because, until recently, I did really obsess over every advanced stat I knew about.

Then the Bobcats let Boris Diaw walk to the San Antonio Spurs after he quit on the team and his defensive rating went from 111 to 102.

Defensive rating is a stat that estimates the points a player allows per 100 possessions. So in his final year with Charlotte, Diaw essentially gave up 111 points per 100 possessions. Then in San Antonio, in the same season, after sitting on the bench for most of February for quitting on Charlotte, Diaw magically became a better defender and gave up nine less points per 100 possessions? I don’t think so.

For his entire career, Diaw’s defensive rating hovered closer to 110 than 100, at least until he got to San Antonio.

Diaw played his first two NBA seasons in Atlanta. In 2003/04, his defensive rating was a 107. In 2004/2005 it was 112. Respectively, Atlanta ranked 23rd and 29th defensively those years.

He moved on to Phoenix and his defensive ratings in 05/06, 06/07, 07/08 and 08/09 were: 106, 109, 109 and 114. Respectively, the Suns ranked 28th, 23rd, 25th and 27th defensively in those same seasons.

I have a feeling you’re noticing the trend, but I’ll keep going to drive home the point.

He moved on to Charlotte during that 08/09 season and posted a 107 defensive rating. That’s seven points fewer per 100 possessions than it was in Phoenix, and in 08/09 Charlotte ranked ninth defensively.

In 09/10, 10/11 and 11/12 Diaw’s defensive ratings were 105, 108 and 111. The ‘Cats ranked first, 11th and 27th respectively those years.

Isn’t it really, really weird that Diaw’s best defensive rating of his career to that point came when he played on the best defensive team of his career?

And isn’t it really, really weird that in the two seasons Diaw got traded, his defensive rating magically, and significantly improved when he got sent to better defensive teams?

It’s not weird. It’s also, in my eyes, why defensive rating is the worst advanced statistic out there.

There are good defenders on the Charlotte Bobcats; Michael Kidd-Gilchrist can already play very good NBA defense, and Gerald Henderson is known as an annoying player around the league, because of the solid wing defense he plays. But Charlotte as a team is horrible defensively, and MKG and Hendo have defensive ratings at 110+

Even this stat’s direct opposite, offensive rating, is somewhat smooth no matter where a certain player ends up; The Bobcats are equally bad offensively as there on on D, and players like Ramon Sessions, Kemba Walker and Gerald Henderson, guys I would consider solid on offense, all have offensive ratings above 100.

Seriously, Washington is the worst offensive team in the league, and they have seven players with offensive ratings at 100 or higher.

But the worst defensive team in the league, Sacramento, has just one player at or below a 106 defensive rating. He’s Cole Aldrich. He’s played 10 minutes this season.

So even for a guy like me who loves almost all things advanced-statistic-wise, there comes a point where you have to realize that this level of analytics is not the holy grail of measuring a professional basketball player.

Let’s be serious. You could put Tim Duncan (the league leader in defensive rating right now) on the Bobcats, and he’d struggle to get his defensive rating below, or even at 100 by the time the season ended.

It’s a team-reliant stat that, honestly, should have no bearing at all on determining how good a player is defensively.