Thursday, August 21, 2008

Offensive Rebounding, Part I

Offensive rebounds are an important factor in offensive success, yet often get swallowed up in the general category of "rebounds." Further, when commentators do talk about offensive rebounds, they often compare raw totals for teams, so it always ends up looking like the team that shot worse was a better offensive rebounding team, when in fact they just had way more opportunities. Finally, moreso than a lot of other aspects of the game, offensive rebounding at the team level can be greatly affected by an individual player -- for instance, last year the Knicks regularly threw out a lineup of David Lee and a bunch of schmoes and still ended up 10th in the league in offensive rebounding percentage. Or, for another example, Utah grabbed a full 5.3% more of its offensive rebounding opportunities when Paul Millsap was playing than they did when he wasn't.

More generally, here's a study that shows that offensive rebounding is much more of a product of team strategy than many other aspects of the game -- meaning that a team's success in grabbing offensive rebounds is largely dependent on whether they themselves choose whether to crash the boards or get back on defense. And here's a study showing that an individual player who is a good offensive rebounder can have quite a significant effect on the team's offensive rebounding, as seen in the Knicks and Jazz examples above.

How valuable is it, then, to have a great offensive rebounder on a team? The Knicks example showed us that a team can still suck even if they are competent offensive rebounders. And choosing to have your offense crash the boards might leave the team vulnerable to leak outs and fastbreaks at the other end (the Bucks, for instance, were the 3rd best offensive rebounding team in the league last year, but were by far the worst defensive team in the league. About 37% of their opponent's shot attempts occurred in the first 10 seconds of the shot clock, and they were the 4th most dunked on team in the league last year). On the other hand, offensively, every extra opportunity to score is a huge advantage. The question I'm curious about here, though, is about whether offensive rebounds give the offense any extra advantage beyond the extra scoring chance.

In Chatper 6 of the book Basketball on Paper, Dean Oliver writes:

On the offensive end, a rebounder has to find space to get around defenses and get close to the basket, where he is frequently poorly guarded due to defensive confusion and inside position. That makes an offensive rebounder a potentially very efficient scorer. John Maxwell, public relations director of the Charlotte Sting, did a study on WNBA offensive rebounds supporting this assertion. He found that offensive rebounders improve their field goal percentage from about 41 percent to 48 percent and their points per play from about 0.80 to about 0.94. That improved the team's points per play from about 0.80 to 0.90. That kind of difference is huge! Not only does an offensive rebound preserve a possession, it really does provide an easier opportunity to score.

In the past, when I've discussed offensive rebounding at this site, I've treated the issue independently, only considering the extra scoring opportunity created without any regard to the quality of said scoring opportunity. So, do offensive rebounds lead to better shot attempts?

One way to look at this question is to see whether teams shot better in games where they got more offensive rebounds. If those rebounds are leading to open shots right underneath the basket, that should show up in the FG%.

Over the last three years, 3,690 games have been played in the NBA, giving us 7,380 observations (one for each team for each game). So I got cracking. First I compared offensive rebounding percentage to effective field goal percentage.

[Recall: Offensive rebounding percentage is the percentage of available offensive rebounds that a team retrieves. If a team misses 10 shots and gets 4 offensive rebounds, then the have an OReb% of 40%. If a team misses 20 shots and gets 6 offensive rebounds, their OReb% is only 30%, even though they got more rebounds. Effective field goal percentage is normal field goal percentage adjusted for the fact that three-pointers are worth an extra point].

The results were somewhat unexpected. I looked at correlations team by team -- the reasoning being that better shooting teams will often tend to have lower offensive rebounding percentages, because they'll be de-emphasizing that strategy in favor of getting back on defense if they expect their shots to go in. Looking at the team by team results, there were a handful of teams, notably including Detroit and Utah, which displayed the expected positive correlation between OReb% and eFG%, but there were many more teams who had a negative correlation -- that is, they shot lower percentages in games where they had a higher OReb%.

Not quite satisfied, I made a few tweaks to the study. First of all, instead of looking at eFG%, I decided to look at True Shooting Percentage. True Shooting Percentage (TS%) looks at not only 2- and 3-point shots, but also at free throws, to give an indication of scoring efficiency on all scoring attempts. The thinking here was that a lot of the benefits of offensive rebounds might have been hidden in the previous study because they were being realized at the free throw line -- it is quite common, after all, to see a big secure an offensive rebound and get fouled before he can go back up for a shot.

The second tweak was to look at teams year by year, since some teams will undergo drastic changes in personnel and coaching strategy between seasons.

Still, after the changes, only about 36% of all team-seasons exhibited a positive correlation between offensive rebounding percentage and true shooting percentage. I'm not completely prepared to drop the idea that offensive rebounds can lead to better shot attempts, though -- it's quite possible that the cause-effect is being muddled here. For instance, in a game where a team is shooting well, it would be reasonable for them not to concentrate on offensive rebounds, and so the OReb% might be lower for that reason. The answer, then, is to look at plays within games that occurred immediately after offensive rebounds, to see if they were more successful than other plays. That will be presented in Part II.

While inconclusive, there is still something to be gained from looking at the extremes from the results of this study. The team that showed the greatest correlation between OReb% and TS% was the Utah Jazz of 2006-2007. In that season, Utah led the league in Offensive Rebounding Percentage and, on average, for every extra offensive rebound they retrieved, their TS% was about .7% higher (which translated to a little under .7 points a game for them). That was the year Utah returned to relevancy, winning the Northwest division and making it to the Western Conference Finals after missing the playoffs for three straight years. That year was also Paul Millsap's first year in the league, and he was the most prolific offensive rebounder and had the third highest True Shooting Percentage on the Jazz. Meanwhile, Carlos Boozer was the second most prolific offensive rebounder on the team and had the highest True Shooting Percentage on the team. That year, the two of them appeared to be able to efficiently turn their offensive rebounds into points.

Meanwhile, the Denver Nuggets of 2007-2008 showed the least ability to turn good offensive rebounding games into good shooting performances. That year, their leading offensive rebounder was Marcus Camby, who played his first healthy season in years and set a career high in minutes played. However, Camby has never been a particularly efficient scorer for a center (he had one or two good offensive years with the Knicks), and had the lowest True Shooting Percentage (49.8%) of all of the regular rotation players on the Nuggets. Unlike Boozer and Millsap, he didn't turn his offensive rebounds into points very efficiently.

The other curious result that popped out at me was the story of the Clippers the last three years. In both 05-06 and 06-07, they had the lowest correlation in the league of OReb% to TS%, and then jumped to 8th highest in the league last year (07-08). The most notable difference in the Clippers last year was the absence of Elton Brand, who missed most of the year with a ruptured achilles tendon. Elton Brand was the Clippers' best offensive rebounder in both 05-06 and 06-07, but he was (and is) an efficient scorer, shooting around 58% in TS% in both years. I'm not sure how to explain what happened. One possible explanation is Chris Kaman's renaissance year last year. After being misdiagnosed with ADD at the age of 2, Kaman had been taking unnecessary medications throughout his life until the 07-08 season. A new doctor took him off the medications and Kaman had a breakout season, improving his TS% from 50.2% to 53.8%. In the absence of Brand, he also increased his offensive rebounding to compensate. Does Kaman's improvement counterbalance the absence of Brand? I'm not really sure. It should be interesting to see what happens with Brand in Philadelphia next year -- last year's 76ers were the best offensive rebounding team in the league, led by Sam Dalembert, Reggie Evans, and Thaddeus Young. Furthermore, Dalembert and Young had the team's highest True Shooting Percentage, but Young will be moving from the 4 to the 3 to make room for Brand. While exciting and athletic, Thaddeus Young has a ways to go in terms of shooting and ballhandling ability, so it's possible we'll see a decrease in his TS% next year.

In any case, in Part II I'll take a look at what happened immediately following offensive rebounds. Hopefully that will provide more clarity as to the true value of the offensive rebound.

BONUS: Charles Oakley was mentioned in the comments to a previous post (sidenote: who didn't love Charles Oakley? I miss him enough (except for the post-Knick years, which I didn't particularly care about) that I'd watch his celebrity cooking show). I don't have much to add, but figured it might be fun to take a look at how the Knicks did with him in the lineup. The Knicks' best year with him was probably the 93-94 season, when they made it to the finals against Houston. That year, Oakley led the team in offensive rebounding (and was fifth in the entire league in total rebound percentage), as well as made the first team all-defensive team and the All-Star team in the East. He led the Knicks to be one of the top offensive rebounding teams in the league that year (with a 34.4% OReb%) -- this is significant because those Knicks teams under Riley were all about defense. In 93-94, offensive rebounding was the only offensive facet of the game in which the Knicks were above average. Oakley's contributions at the defensive end are well-chronicled, but it might be that he also contributed just enough offensively to make the Knicks an elite team that year. Hooray Oakley!

No comments:

Post a Comment