Deep Dives: Investigating Rebound Tradeoffs

facebooktwitterreddit
NC Deep Dives
NC Deep Dives /

One of the moretrenchant criticisms of the three point-happy modern era[1. As opposed to the “soft jump shooting team” junk which confuses aesthetic preference and play in a past era with effectiveness.] is that all the long shots reduce offensive rebounding. Offensive rebound rate is at an all time low this season with only around 23% of available misses snared by the shooting team. As offensive rebounds tend to produce high value possessions, and that big men, traditionally the producers of the majority of offensive rebounds grab fewer OREBS the farther they are stationed from the basket, there is some truth to the accusation.

On the other hand, there is a strategic element at work here. Many coaches, among them Doc Rivers, first in Boston and now in Los Angeles, and Steve Clifford in Charlotte[2. The Hawks under Mike Budenholzer have also been a low offensive rebound team, but with a lot of deep shooters and more perimeter oriented bigs in Al Horford and Paul Millsap, this probably represents a combination of player skillsets and coaching choice.] have observed that going after offensive rebounds is somewhat of a trade off with getting back on defense well enough to prevent opponent transition play.

If we define a “transition play” as a shot in the first 7 seconds of the shot clock, there is an observable correlation between “chasing[2. As defined by being within 3.5 feet of the location of the ball when the rebound is secured per SportVU.]” offensive rebounds and allowing the opponent to get out on the break if they end up securing the defensive board. Thru games of January 23[3. The Day the SportVU Shot and Rebound Logs Died. Best information is they aren’t coming back in that format any time soon, sadly.] here’s what the comparison looks like for this year:

Dashboard 2 (31)
Dashboard 2 (31) /

Obviously, not a perfect correlation, things such as steals and shot locations not to mention team-specific abilities like floor awareness and plain old hustle play big roles in allowing or preventing the opponent from getting early offense. But the tradeoff is there. The question is, is it worth it? Earier this season, friend of the blog Brian Skinner discussed[3. Warning: wonky.] the issue at length and estimated that for a theoretical average team, a 1% increase in offensive rebound rate would be worth about .62 points of offensive rating. Which gets us partway to the answer of the transaction at hand. We know what teams are “paying” for the improved transition defense allowed by conceding offensive rebounds in terms of offensive efficiency. What remains to be determined is the gains on defense.

Of course, it’s slightly more complicated, because first we better have to better understand the mechanism by which offensive rebounds are generated. On a broad level, it’s fairly clear – simply go after more of them! Across the league, contested rebound battles on missed field goals[5. Missed free throws seem to behave very differently, likely because of how easy it is to organize a solid defensive rebounding scheme from the static start of a free throw.] are essentially a 50/50 proposition between offense and defense. Over time, every team is going to end up with around half of defensive rebounds chances and around 5% of offensive rebound chances just by way of the ball dropping into someone’s lap with no opposing player in the immediate vicinity. The exact parameters of this baseline aren’t totally clear[6. And may in fact be changing, as even over the three seasons of SportVU data, the proportion of rebounds which end up as “Uncontested defensive rebounds” has risen from around 57% to around 60% for the league as a whole.], but those exact points matter less than the conclusion that for every 2 of those remaining rebounds a team contests, it will end up with one of them. To put it another way, a 2% increase in “offensive rebound chase%” will lead to a 1% expected increase in OREB%. It isn’t quite as simple as just trying for two more boards to get one, however. Often times, a player will contest a rebound to no great net effect because a teammate is also going for the same rebound.

To account for this “doubling up” effect, I wanted to see how well an increase in a lineups overall “rebounding aggression” might affect the likelihood of ending up with the board. To do this, I went back to 2014/15 data so that I’d have a full season of SportVU to work with. To come up with an “offensive” and “defensive” rebounding aggression factor, I simply summed the individual offensive and defensive chase%’s for every lineup.[2. I borrowed this idea from Eli Witus, now a longtime member of the Rockets’ front office. Almost a decade ago, Eli demonstrated the benefit of having multiple scorers on the floor by proving lineups with a high aggregate usage between the five players tended to outperform expectations in terms of offensive rating. I also subtracted the average total, so that a five man unit which was perfect average in terms of both offensive and defensive chase % would have an aggression factor of 0 both offensively and defensively, while a lineup with players who tended to eschew the offensive glass but pound the defensive backboard would have a negative offensive aggression but a positive defensive aggression factor.] I also controlled for the general offensive rebound rate based on shot locations, as shots near the rim are far more likely candidates for offensive rebounds than jumpers and especially above the break three pointers.

Using a simple regression analysis, I found that an increase of around 3.1% in a lineup’s offensive aggression was worth a 1 point bump in OREB%. To put it in slightly simpler terms, for every three contests of offensive rebounds by individual players the lineup could expect one more offensive rebound. As on a team level, winning contested rebound battles is 50/50, this then means that about 1 out of those three offensive rebound contests is doubling up on an already contested board, one is contesting a “new” rebound and losing, while the last is that precious marginal rebound.[4. On the flipside, the ratio for defensive rebounds was about 7:1, meaning 5 double ups, one loss and one marginal win.]

All of which is a to say, that to get that one additional offensive rebound, a team would have to put players in the vicinity of an offensive rebound three time. Of course, the number of times the team would have to actively pursue an offensive rebound is somewhat different, as SportVU’s rebound chances stat already includes a certain measure of success in terms of getting near enough to the rebound in the first place. At least from public data, the number of times a player “goes for” an offensive rebound at the expense of getting back on defense, but never gets anywhere near the ball is largely unknown. On the other hand, a player is sometimes making no special effort to chase the ball, but the carom hits him in the face. But as far as a first approximation goes, I think it’s solid.

All of which brings us to halfway point in answering the question, leaving the open issue how much does a “missed” chase cost in terms of defense. That, along with the opposite side of the issue, investigating whether there are measurable tradeoffs between offensive transition play and defensive rebounding, is an discussion for a future post.