The Problem with Defensive Stats

The statistic I’ve developed (which can be found at http://basketball-statistics.com/cs.html) was originally just supposed to be about defense. Composite Score was in fact originally called Defensive Composite Score. The reason I focused on defense is because it’s still the holy grail of statistics. There are plenty of good rating systems out there for measuring a player’s offensive contributions, but defense is a much different story.

Defensive Composite score has been essentially the same thing since it was first developed: a combination of plus-minus, counterpart PER, and defensive rating. The reason I find it valuable is because it’s a combination of systems that individually have some flaws but are useful when combined together. Still, Defensive Composite Score is far from perfect.

To show the weaknesses of defensive stats, let’s look at some correlations. If all of the different ways of rating a player agree, it’s a good sign that they’re pretty accurate. For example, the three aspects of Offensive Composite Score are highly correlated with “r” statistics of .65, .51, and .60. However, the three aspects of DCS have much lower correlations of .18, .22, and .24.

This is just another indication that we have a long ways to go before we can accurately evaluate a player’s defense. In my opinion, until we track defensive stats in box scores the way we track offensive stats, it’s going to continue to be very difficult. Even if we track things such as forced misses and charges drawn, there are still a lot of teamwork aspects that would be hard to measure. A more in-depth charting effort that includes a more subjective tracking of those aspects may be more useful, however.

If anyone is interested in charting a team’s defensive statistics, send me an e-mail. If I get enough positive responses, we could get a nice system going.

Picks
Indiana at Orlando: Magic
Atlanta at Washington: Wizards
Boston at Charlotte: Celtics
Golden State at New York: Warriors
Denver at Minnesota: Nuggets
Oklahoma City at Memphis: Grizzlies
San Antonio at Houston: Rockets
Cleveland at Milwaukee: Cavaliers
New Jersey at Utah: Jazz
Dallas at Sacramento: Mavericks
Miami at L.A. Clippers: Clippers

Prediction Record: 138-67

Did you enjoy this post? Why not leave a comment below and continue the conversation, or subscribe to my feed and get articles like this delivered automatically to your feed reader.

Comments

No comments yet.

Leave a comment