On rebounding


(Mike Carter, USA Today Sports)

We live at a time of televised analytic plenty, yet, somehow, you still see rebound margin numbers flung up on the screen during this or that telecast in 2019. That makes me grit my teeth at the blatant luddite behaviors on display, of course, and, well, I’m right to do so. Rebound margin really is meaningless, an ersatz and mislabeled tribute paid to teams that alter shots yet refuse to go for steals and/or charges (with all of the above, preferably, transpiring at a fast pace).

In partial defense of my well-intentioned graphic-making brethren and sistren, however, I will additionally confess to the following. I’ve been mulling just how peculiar rebounds really are for a while now, and (this may say more about me than about rebounding) I’m still not sure I’ve found solid ground on this particular subject.

Here’s what I think I think….

I’m not a fan of whole-season rebound percentages in college basketball
Leave it to the sport’s endearing and enduring idiosyncrasies to overturn perfectly sound axioms regarding sample size.

Whole-season rebound percentages in the college game are kind of like the RPI. They can be perfectly fine, and they’ll be sufficiently congruent with current events in a majority of instances. But you can’t trust them sight unseen, so why use something where you need to pop open the hood and check the connections every single time?

At this writing, Kentucky’s fourth in the country in offensive rebound percentage, and third in the SEC in conference play. Same for Duke, No. 5 nationally and No. 3 in its conference. Wake Forest’s been half a standard deviation better than the Blue Devils at getting offensive rebounds against ACC opponents.

Defensive rebound rates can translate a bit better from whole-season to conference (cf. Maryland, Colorado, and Michigan), but even here you can’t just assume you’re hitting statistical bedrock every time. Take Kansas State, No. 17 nationally in defensive rebound percentage, more than 100 spots higher than its nearest in-conference competitor (Oklahoma, No. 132). Actual Big 12 play, however, has swiftly devolved into a vicious egalitarian struggle where every team’s virtually identical on the defensive glass and the Wildcats nominally rank No. 4 in the league in that category.

These numbers are in motion, of course, but this isn’t primarily a sample-size thing. It’s more of a basketball thing, or, better still, one more peculiarity of a mass-audience sport wherein the teams themselves select a sizable portion of their own opponents. There are few bread-and-butter box score numbers that vary as much as rebound percentages due merely to non-conference scheduling philosophies and/or to how certain coaches choose to change their look for conference play.

There’s a new breed of inverse-Michigan rebounding teams, and it looks weird
For as long as basketball has been played, there have been teams that choose to stay away from offensive rebounds but that do very well on the defensive glass. When you think of this style, think of John Beilein. He has always preferred to generate his teams’ shots through a low-low turnover rate while keeping offensive rebounds scarce but not self-defeatingly Wyoming-scarce (see below). We see this type of team all the time, and one good thing that came with throwing around offensive and defensive rebound percentages as opposed to simply “rebounds” as a counting stat is that we have the ability to give this kind of team its due.

But now, in 2019, we’re seeing teams that flip this script, and it is very strange. Why would you want to be great at offensive rebounding but terrible at defensive rebounding?

Inverse-Michigan Factor (IMF)
Standard deviations above/below conference means
Conference games only

                          OR%     DR%     IMF
Michigan State    2019   1.75   -1.20    2.95
Rutgers           2017   1.62   -0.90    2.52
West Virginia     2019   1.11   -1.35    2.46
Providence        2019   1.07   -1.33    2.40           

These are the largest IMF numbers we’ve seen in the past five years in major-conference play. Purdue 2019 is not far behind.

One problem with defensive rebounding as a descriptive basketball metric, at least in discussions of who will win, is that opposing coaches so seldom act upon being in the presence of its weak variant. Michigan State to date, stereotypes be damned, has been awful at defensive rebounding, but, with the possible exceptions of Matt Painter and Fran McCaffery, there will be no opposing coach who attempts to gain an advantage from that fact. The culture of coaching at the college level says rebounding’s not really coaching because it’s not and cannot be planned in the same hermetically sealed fashion that a play call is.

Allen Edwards is conducting a really interesting experiment
Few teams in recent memory have been more steadfast in their refusal of offensive rebounds than Wyoming this season, which at this writing is pulling down just 14.9 percent of its missed shots in Mountain West play. Once you get down to 15 percent, you’re basically at the performance horizon.

Keep in mind even the most committed offensive-rebound misanthropist can never get to zero. For one thing, there will always be “team” offensive boards where the opponent mishandles a defensive rebound and the ball goes out of bounds.

A 14.9 percent offensive rebound rate places Wyoming beneath 13 individual Division I players, from Tyrique Jones to Mark Vital, all of whom are more likely to rebound a teammate’s miss than are five Cowboys on the floor working collectively.

This is actually a good time to look at Allen Edwards’ group, because it just used its zero-second-chances voodoo in a win. Wyoming rebounded 14.8 percent of its misses and won at home against Colorado State, 74-66. Jordan Naughton came off the bench and hauled in three offensive boards in 21 minutes. He was likely made to run the arena steps that night.

In Mountain West play, the Cowboys have been utterly, gloriously and wonderfully normal in terms of accuracy from the field (50.2 eFG percentage), making them a near-ideal test case. It turns out that in 13 years of major-conference play, no previous team ever scored anywhere near so few points with this level of shooting.

Wyoming now

(I’ve plopped Wyoming’s dot down amidst past major-conference seasons because, well, that’s what I have handy.)

Possessions where there is no turnover and where one shot is attempted enjoy an overwhelming heuristic privilege both in coaches’ advance planning and in even the best after-the-fact analysis, particularly in film breakdowns. If you did nothing but look at film of Wyoming’s turnover-less possessions up to the point at which the shot is a make or a miss, you would see zero that’s out of the ordinary. I’m weird, but, from my chair, Edwards is doing very interesting things, if only we will listen.