This is part of a series of posts examining whether offense or defense has more control of various aspects of a typical college basketball game. The introduction is here. A description of the methodology is here.

So far in this series, we’ve focused on a specific stat in each entry, but today we’re talking about the whole enchilada – points per possession. Obviously, it’s the most important thing on the board and includes everything that matters.

Consistently over the past decade, offense has shown more control over its own efficiency than the opposing defense. A good offense will tend to hold up against a good defense and a bad offense will tend to underperform against a bad defense. These tendencies are not as strong for previous traits examined here, but at 64% influence, offense is clearly in charge.

This isn’t surprising given that only three of the other 11 stats I investigated are clearly in majority control of the defense. At least at the college level, it’s not all that easy for the defense to dictate what the offense does with the ball.

So does offense win championships? That’s a slightly different problem. This past season’s Kentucky and Virginia teams give some evidence that a dominant defense is not sufficient, even with a pretty solid offensive unit to go with it. But over the course of history, one could find examples that support either side of the argument. A more rigorous study is needed to get a handle on the issue. That said, good offense is going to win the battle against good defense a slight majority of the time.

One thing that is more questionable is the idea from defensive-minded coaches that defense is more consistent than offense. The theory is that defense is about effort and we all have control over our effort, while offense is subject to the vagaries of a bouncing ball going through the rim. But given that defense is slightly more at the mercy of the offense than vice versa, it stands to reason that offensive performance will be more consistent over time.

It’s not as catchy, but offense and defense wins championships. And when you see a box score feature excellent offense, it was more likely the offense’s ability driving that performance than a breakdown by the defense.

For funsies, here are the ten coaches with the best average adjusted offensive efficiency since 2002 (minimum 10 seasons)

 1 Mike Krzyzewski 116.7
 2 Roy Williams    115.1
 3 Mike Brey       114.3
 4 Mark Few        114.2
 5 Billy Donovan   114.0
 6 Jamie Dixon     113.5
 7 Bo Ryan         113.3
 8 Bill Self       113.3
 9 John Calipari   113.1
10 Thad Matta      112.8

For most of last season, there wasn’t a single Notre Dame player on an NBA roster, so give Mike Brey a few bonus points for his efforts in consistently assembling one of the best scoring units in the land. He makes some sacrifices on the defensive end to get there—and that will be discussed in the next installment of this series—but given the results falling out of this study, it’s not a bad approach to load up on the offensive end especially when you don’t have the talent of other top teams.

The list of defensive coaches with the best numbers is below.

 1 Bill Self        90.4
 2 Rick Pitino      91.1
 3 John Calipari    91.8
 4 Jim Calhoun      92.3
 5 Mike Krzyzewski  92.4
 6 Thad Matta       92.6
 7 Roy Williams     92.6
 8 Bo Ryan          92.9
 9 Gary Williams    93.5
10 Tom Izzo         93.5

Here’s the year-by-year data. One puzzling thing about this analysis is the large value for home-court advantage which suggests something like a 4-5 point advantage in absolute terms.

Year  %Offense  HCA
2015     63     3.6
2014     55     3.4
2013     60     3.7
2012     64     3.7
2011     67     3.6
2010     67     3.6
2009     69     3.6
2008     69     3.9
2007     68     3.9
2006     64     3.7
AVG      64     3.7


Offensive Spectrum – Ordered by pct of offensive “control”

FT%  98%  (HCA=0.5%,  r(off)=.19, r(def)=.04)
APL  86%  (HCA=-0.1s, r(off)=.55, r(def)=.23)
3P%  83%  (HCA=0.7%,  r(off)=.12, r(def)=.06)
OR%  73%  (HCA=1.1%,  r(off)=.23, r(def)=.08)
3PA% 71%  (HCA=0.0%,  r(off)=.52, r(def)=.33)
A%   71%  (HCA=2.6%,  r(off)=.32, r(def)=.21)
PPP  64%  (HCA=3.7,   r(off)=.51, r(def)=.36)
???  59%
???  50%
???  49%
???  36%
???  30%
???  15%