Peer performance pressure, and the limits of coaching theory


Last season, this offense made its shots and took outstanding care of the ball. What could possibly have gone wrong?

Years ago, when rebound percentages were still seen as newfangled, the first question I was ever asked by a Division I coach was what the best target numbers are for both offensive and defensive rebound rates. I don’t remember my response, though I would guess I delivered, as requested (I was thrilled simply to have had my opinion solicited), two numbers spelling out what the team “should” be doing.

I now think that was a mistake. True, thanks to Dean Oliver, we know that the four factors in basketball are shooting, turnovers, rebounds, and free throws, and we rank offenses and defenses on each of those metrics.

But of late, however unconsciously, I find I no longer regard all of the factors as a must-watch sequence in and of itself. Instead, I’ve elevated shooting to a co-starring role above the title in this movie we call hoops.

Furthermore, I’ve collapsed turnovers and offensive rebounds into one quantity and elevated that into the other co-starring role, one I call shot volume. (In this two-factor amalgam, offensive rebounding is clearly the junior — or at least downstream — partner.) And, for better or worse, I now regard free throws as an occasionally dispositive but basically exogenous event, kind of like a power outage or visiting relatives. 

So I suppose I should have told that coach years ago simply: It depends. In fact, in this same spirit of incorrigible relativism, I’ll toss the following out as Gasaway’s first law of offensive rebounding:

Get as many offensive rebounds as your turnover rate dictates in order to record a healthy shot volume.

(Meantime, keep opponents from matching your shot volume. And, oh, by the way, make your shots and keep the other team from making theirs.)

Most crucially, a “healthy” shot volume is going to be determined by what 350 other teams are doing. And, of course, what 350 other teams are doing is completely arbitrary.

As a coach, you can have a beautiful theory about what the proper role of offensive rebounding should be, and, purely as a sport, basketball welcomes all styles. In theory, you could forego offensive boards entirely, and that could work fine. Indeed, the NBA may be hurtling down that very road.

But in practice, in college basketball in 2018, the price for falling too far behind the shot-volume herd is to saddle your shooting and your defense with a series of increasingly severe and, eventually (and rather quickly), impossible compensatory demands. Shot volume doesn’t care one bit what style you play, but the metric is a cruel and harsh tyrant when it comes to results. Take three paradigmatic examples from last season: Michigan, Oklahoma State, and Santa Clara.

John Beilein is a force-feeder when it comes to shot volume.

Turnover and offensive rebound percentages
Shot volume index
Conference games only
                 TO%    OR%    SVI
Michigan        14.5   24.1   97.0

The Wolverines finished last in the league in offensive rebound percentage in Big Ten play, yet the team’s shot volume was well above-average thanks to the best turnover percentage in the conference. Beilein thinks he’s saying no to offensive boards, and, after a fashion, he is. But he’s tracking how many shots his offense gets (I know because I asked him at Big Ten media day), and it turns out that force-feeding a massive number of turnover-less possessions into 40 minutes can more than offset a relative lack of second chances. (For the record, this paragraph can be applied to Iowa State last season, as well.)

Conversely, Oklahoma State achieved effectively the identical bottom-line result last season from pretty much the 180-degree opposite stylistic direction.

Conference games only
                 TO%    OR%    SVI
Oklahoma State  19.1   35.1   96.9

Even with Jawun Evans running the show, the Cowboys were merely average in terms of taking care of the ball. But OSU crashed the offensive glass, and then-coach Brad Underwood didn’t worry about possible negative impacts to his transition defense. (I know because I asked him at Big Ten media day: “When you’re going after offensive boards, the reality is the other team has to keep guys back to stop you from getting them.”)

Which brings us to Santa Clara. In theory the Broncos are just like Michigan, turning down offensive rebounds but also excelling at avoiding turnovers. In practice, however, it didn’t turn out nearly as well in Santa Clara as it did in Ann Arbor.

Conference games only
                 TO%    OR%    SVI
Santa Clara     15.6   16.2   92.1

With a shot volume as low as SCU’s, the demands placed upon shooting accuracy become, in effect if not in theory, insurmountable. Just ask Virginia Tech about last season.

The Hokies aren’t a clean shot-volume parable like Michigan or Santa Clara because, as it happens, Buzz Williams’ guys weren’t all that great at taking care of the ball.

Conference games only
                 TO%    OR%    SVI
Virginia Tech   18.0   19.9   91.2

You could tell Williams his team’s turnover rate was a bit too high relative to the ACC, and his reaction, quite rightly, would be well, duh. But what’s instructive about Virginia Tech nonetheless is that the team helpfully illustrated how a terrible shot volume can negate even historic shooting accuracy.

Last year, the Hokies shot better in ACC play than any team has for at least a decade. Put another way, out of the last 135 team-seasons recorded in Atlantic Coast Conference play, Virginia Tech 2017 rates out No. 1 in terms of shooting.

And all it got Williams was the league’s No. 7 offense.

eFG%: effective field goal percentage
PPP: points per possession
Conference games only

                    eFG%      PPP
1. North Carolina   51.9     1.16
2. Duke             54.3     1.14
3. Wake Forest      52.6     1.14
4. Louisville       52.9     1.13
5. Syracuse         53.3     1.12
6. Florida State    52.4     1.12
7. Virginia Tech    57.7     1.10        

In sum, the coaches at Michigan, Santa Clara, and Virginia Tech all thought they were saying no to offensive rebounds last season, and all three teams finished last in their leagues in that category. But it turns out not all last-place finishes are created equal. The Wolverines still had a very good shot volume, while the Broncos and the Hokies did not.

That’s all well and good, but, if you’re a coach, is there a way of telling when, exactly, last-place offensive rebounding is too much of a good thing? One thumbnail I’ve come to use is simply the ratio between a team’s offensive and defensive rebound percentages. Every alleged iron law of analytics was born to be broken, and this one will be too. Still, it’s true that, so far, I’ve never seen events go particularly well for a team with a ratio under 0.33 or so….

Lowest offense-to-defense rebound ratios, 2017
Tuesday Truths teams
Conference games only
                       OR%     DR%    OR/DR
1.  Santa Clara       16.2    76.2    0.21
2.  Northern Iowa     16.6    73.9    0.22
3.  Richmond          18.6    72.2    0.26
4.  Washington State  20.0    70.8    0.28
5.  Virginia Tech     19.9    67.8    0.29
6.  Loyola            22.1    75.1    0.29
7.  San Diego         23.2    78.1    0.30
8.  Boston College    23.4    75.4    0.31
9.  Davidson          23.1    74.5    0.31
10. Saint Louis       23.0    72.2    0.32
11. Vanderbilt        22.9    70.5    0.32    

This rule of thumb will float as the aforementioned 350 other teams change what they do. Right now the Tuesday Truths average ratio is 0.42 (a few years ago it was 0.45). Nevertheless, until further notice, this is what we think we know…

If you’re going to such great theoretical lengths to avoid offensive rebounds that your percentage on that side of the floor is less than a third of what it is on the defensive end, in practice your shot volume is virtually assured of being lower than your conference’s mean. Far more likely, it will be way, way lower.

Perhaps the team performance that results in the shadow of this abject disregard for what the competition’s doing — which on one level could be seen as rather principled and courageous — can be termed “died of a theory.” So far, at least, it’s been no particular fun for teams to wear the DOAT label.

Threes and offensive rebounds
Naturally, as teams shoot more threes, and, most particularly, as stretch-5s roam our hoops landscape in ever greater numbers, we would expect to see a decline in offensive rebound rates. If your tallest guy’s hanging around beyond the arc, you won’t get as many second chances. Understood, but, again, performance refuses to go gentle into that principled night. Vanderbilt shot more threes in conference play than even Santa Clara, and the Commodores posted an offensive rebound rate that was nearly seven percentage points higher than what we saw from the Broncos.

Davidson Wildcats, I salute you!
Bob McKillop’s team is playing some offense in 2017-18….

Davidson's first two games
    TO%    OR%    SVI    eFG%    PPP
    4.8   27.9   110.1   67.3   1.51

The opponents haven’t been what you’d call a murderers’ row of great defenses (Charleston Southern and UNC Wilmington), but I hope you got to see some of those 80 minutes. Purely in terms of what’s possible when an offense brings the ball down the court, Davidson has been frolicking happily on the sport’s performance horizon.