Is the Best Defense a Good Offense and the Best Offense a Good Defense? A Quick Look at History

A few months ago, I downloaded the team statistics for each season since the NBA-ABA merger in 1976-77.  I chose to get each team’s average age, SRS, Offensive and Defensive Rating, pace, and Four Factors for both offense and defense; with 36 seasons and 960 teams, there are approximately a gazillion things you can do with this data. Hopefully this will result in a new analytical era for the AGR team.

I was first motivated to embark on the thankless and tedious task of downloading and spreadsheet-ing by the desire to confirm (or disconfirm) the age-old adages that “the best offense is a good defense” and “the best defense is a good offense.”

Intuitively it makes total sense, right? Get stops on D to get out in transition for easy hoops. Likewise, by managing the ball into the bucket on offense, you’ll stop the other team from beating you on the break. Furthermore, it’s not unreasonable to think that some of the same skills that allow an individual to be good at defense—athleticism, length, basketball IQ, tenacity, etc.—might also allow a player to be good at offense, too. Even on the team level this is true for attributes like communication and chemistry.

My initial strategy to empirically answer this question was to run a simple correlation between each team’s offensive and defensive ratings since the merger. Obviously, it is crucial to use offensive and defensive ratings rather than points scored/allowed per because ORtg and DRtg are pace-adjusted.

I’m not a math guru, so I wasn’t prepared to do regression analyses (which might help with a lot, especially given that I have four factors data). I figured if there was any relationship, it would emerge given a dataset this enormous.

Well, indeed, a significant trend did emerge (p <.0000001). But not the one I expected. When I first graphed the data and checked the relationship, there was a positive correlation (R = .226), meaning that as a team’s offensive rating got higher (i.e., better), their defensive rating also got higher (i.e., worse).

This took me aback; why did, historically speaking, teams’ defense decline as their offenses improved? It works the other way around too, of course—why would a team’s offense decline as their defense improved?

My gut reaction, other than the shock-induced vomit, was that this funky finding was a product of the changing landscape of league offenses and defenses over the years. After all, if the league averages for ORtg/DRtg for teams in the ‘80s and early-to-mid ‘90s were higher (and as a rule, the league averages for ORtg and DRtg are identical in any given season), then perhaps my “finding” was corrupted by a third variable, season/era.

To control for era, I decided I’d compare each team to the other teams from their season. I converted each team’s DRtg and ORtg into a z-score using the season averages from that season. Now each team’s ORtg and DRtg had a baseline and I could fairly compare teams across eras.

When I reran the correlation using z-scores, a new trend revealed itself; just as we might have initially predicted, there was now a negative relationship between offensive and defensive ratings, meaning better defenses (i.e., lower defensive ratings) predicted better offenses (i.e., higher offensive ratings).

The relationship isn’t incredibly strong mind you, although it was statistically significant (p < .0001). Despite this significance, a correlation of R = -.125 means an R squared, or variance, of .0156. This number can also be interpreted another way (I think?); the variation of a team’s offensive rating explains 1.56% percentage of variation in defensive rating and vice-versa.

Basically, knowing how good a team is at offense (or defense) gives you diddly-squat-zilcha-bagel in terms of predicting their success on the other end of the floor. I would have expected the link to be a bit stronger, given the litany of reasons I mentioned earlier when explaining the hypothetical link between good offense and good defense.

Although this statistical procedure indeed showed us that there is indeed a real connection between offensive and defensive ability—namely that being good on one side of the floor predicts being good on the other side–perhaps the more telling result is the surprising weakness of this relationship.  Offense and defense aren’t that connected; the abilities, skills, and knowledge that translate to success on one end of the floor seem to be more different than similar to the abilities, skills, and knowledge that translate to success on the other side of the floor.

Comment on the article or e-mail us at AGRbasketball (at) gmail (dot) com. Don’t forget to follow @AGRbasketball on Twitter and to like us on Facebook.

This entry was posted in NBA analysis. Bookmark the permalink.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s