Numbers Game: How Much Does Team Rebound Percentage Matter?

Happy new year all. Welcome back to Numbers Game, where we investigate the math behind fantasy basketball.

"Start Myles Turner against the Nets on Thursday, because the Nets have the worst rebounding percentage the league" — a brilliant young analyst (and did I mention dashingly handsome) last week.

It sounded good, but was this analyst full of malarkey? Also, spoiler, I'm the analyst.

On paper, the recommendation makes sense, but it got me thinking — the Nets' Trevor Booker is one of the league leaders in rebounds. In fact, several of the worst rebounding teams boast top-20 individual rebounders, while the three best rebounding teams don't have a single forward or center on that list. Teams average more than 40 rebounds per game, while the elite players only account for nine to 15 of those. Maybe a team with poor rebounding numbers is more indicative of the supporting cast, and not the primary glass-cleaners down low?

Several times this season, we've investigated a commonly held assumption only to find that we were off base. The impact of pace and back-to-backs were both more nuanced than conventional wisdom would have you believe. In that vein, we'll continue our investigations to check if our assumptions are mathematically valid.

This week: is the performance of good rebounders related to the quality of the opposing team?

The Problem

Most teams grab close to 50 percent of the roughly 90 available rebounds each game. The best rebounders – all of whom are bigs, save Russell Westbrook and the positionless Giannis Antetokounmpo – only account for 10 to 15 percent of the available boards. The rest are distributed between guards, wings and less-talented big men.

Rebound percentage describes the portion of available boards grabbed by an entire team, guards, wings and backups included. With so few rebounds going to the best players, is team rebounding strength really a helpful guide? Or do non-bigs and backups muddy the data?

The Pelicans rate as one of the worst rebounding teams of the last two seasons – despite also boasting one of the league's best rebounders in Anthony Davis. It is easy to imagine, therefore, that his teammates are driving down the team's numbers, while Davis vacuums up whatever he can reach in the middle of the floor. This possibility has some statistical basis – Pelicans guards have an even worse rebound percentage than the team as a whole. It seems possible, therefore, that despite New Orleans' inviting team stats, fantasy managers should avoid rebound-dependent bigs against Anthony Davis and the Pelicans.

Of course, it is also possible that rebounding percentage is a useful stat to use. If this is true, then we can move forward with confidence, which is better than making an unchecked assumption.

How do you even approach this question?

I chose to focus on only the best rebounders in the league, and their performances against only the very best and very worst rebounding teams. For players, I limited our investigations to the league's top-20 rebounders, by per game average.

For teams, I focused only on the teams that obtained either fewer than 48 percent or more than 52 percent of the available rebounds. Two percent up or down from 50 may seem small, but it is actually quite extreme. In 2014-15, only three teams would have qualified; in 2015-16, only five teams were that far from average. This season, three teams average higher than 52 percent – the Nuggets, Thunder and Bulls – while five teams average lower than 48 percent – the Mavericks, Celtics, Pacers, Pelicans and Nets.

After identifying our target sample, I isolated the top-20 players' performance against each of the eight teams listed above. Those 163 games – 64 against the good teams, 99 against the bad – are the focus of this article.

So, what did we find?

Rebounding percentage matters. A lot. Much more than I expected, in fact.

I expected this article to end up another version of "you thought this mattered, and it sort of does, but not as much as we were expecting, and never on Tuesdays or every other Sunday." Nope. Not at all. As a former stats teacher, I'm practiced in couching my language, in expecting limits to any stat-based analysis and expressing results with appropriate caution. No need for any of that here. Rebound percentage matters.

In the top 15, every player except Davis grabbed fewer than his season average rebounds against the three good teams. Sixteen of the top 20 exceeded their season average against the bad teams – and two of the four who fell short did so by less than 0.4.

When facing the good teams, these players lost more than one rebound per game. Facing a bad team added more than one rebound per game.

Anything else we should know?

A few other tidbits I noticed, though none match the significance of the headline. First, the rebound percentage rankings don't change much when sorted by position. The same three teams remain on top of the league when sorted by guards, forwards and centers. The bottom shuffles slightly, but the same five teams never stray from the bottom seven. As a result, I can't make any claims about whether or how much the positional numbers matter.

Pace had no obvious impact on these results. It is possible that is because of the ways in which the sample size was limited. But it seems more likely, given my earlier findings about the significance of pace, that the difference in possessions is outweighed by the relative skill – or lack thereof – of the opponents. To put that less mathematically: sure, the Mavericks are one of the slowest teams in the league, but they are so bad at cleaning the boards that a good rebounder will have a good night against them anyway.

Finally – and this may be a coincidence, but I'm mentioning it anyway – the data against the Thunder was worth noting. Although the Thunder rank second in rebound percentage (the Nuggets are first), the best rebounders struggled most against them. Top-20 rebounders averaged more than two fewer boards against the Thunder, whereas they only lost close to one rebound against the Nuggets and Bulls. I'll revisit this on twitter in a month or two, to see if it holds up as the sample size increases.

How do we use this?

This has an easy and obvious impact on daily gamers: before selecting your big men, check the rebounding abilities of their opponents. A team that grabs more than 50.5 percent of its rebounds is cause for concern. A team that grabs less than 49.5 percent is ripe for targeting.

But this also has uses for traditional gamers. While most of the top rebounders are nightly starters, some are still available on waivers. Trevor Booker is available in many shallow leagues, while Andrew Bogut and Joakim Noah are available in most deeper settings. The Mavericks play the Celtics and the Pelicans on either side of the All-Star break, and the extra rebounds from those games will be enough to warrant making Bogut a short-term add in many settings.

The more frequently your league allows pickups and roster changes, the more helpful this information is. In many formats, a manager could simply stream the best-available rebounder facing the Mavericks, and it would quickly make a dent in his overall boarding numbers. In leagues that have more restrictive settings, this information helps determine when to use – and when not to use – a widely owned but not always used player like Tristan Thompson or Clint Capela.