yardbarker

Impact of TRB% and Impact Rebounding on Good Rebounders

Moderators: tsherkin, Doctor MJ, CellarDoor

Post#1 Impact of TRB% and Impact Rebounding on Good
Fri Jan 1, 2010 11:50 pm by TrueLAfan

I want to start by saying that I don't like using Four Factors Rebounding percentages. ORB% = ORB / (ORB + opp DRB); DRB% = DRB / (opp ORB + DRB); Team Reb % = (ORB% + DRB%)/2. I think it overvalues offensive rebounding and penalizes good defensive teams on the boards. But I'm using it here because it was how I'd set up my spreadsheets. The conclusions are the same if you look at rebound differential.

*

First, we have to realize something that seems counterintuitive. Players rebound the same way, unless they have a teammate on the team who is a high impact rebounder. In short, team rebounding has little impact on individual rebounding. Once a team gets a to a certain point, players cannibalize rebounds from their own teammates--but if you're a good rebounder, you still get your share. Proof? Here are spreadsheets that show 89 seasons from top rebounding C/PFs.

Artis Gilmore
http://docs.google.com/Doc?docid=0AQlwu ... bWZq&hl=en

Patrick Ewing
http://docs.google.com/Doc?docid=0AQlwu ... eGdr&hl=en

Kareem Abdul-Jabbar
http://docs.google.com/Doc?docid=0AQlwu ... a3Zo&hl=en

Kevin Garnett
http://docs.google.com/Doc?docid=0AQlwu ... c2Qz&hl=en

Dikembe Mutombo
http://docs.google.com/Doc?docid=0AQlwu ... dGN3&hl=en

Robert Parish
http://docs.google.com/Doc?docid=0AQlwu ... NmR0&hl=en

Shaquille O'Neal
http://docs.google.com/Doc?docid=0AQlwu ... Y2c3&hl=en

Wes Unseld
http://docs.google.com/Doc?docid=0AQlwu ... Jkbg&hl=en

Jack Sikma
http://docs.google.com/Doc?docid=0AQlwu ... RndA&hl=en

In those 89 seasons:
There were about 33 seasons where the player played on a team that rebounded better or worse than his usual teams, and the player's rebounding moved in the same direction. Great Rebounding Team = Great Rebounding Year; Bad Rebounding Team: Bad Rebounding Year. This shows a positive correlation between team rebounding and individual rebounding.
There were about 32 seasons where the player played on a team that rebounded better or worse than his usual teams, and the player's rebounding moved in the opposite direction. Great Rebounding Team = Bad Rebounding Year; Bad Rebounding Team: Great Rebounding Year. This shows a negative correlation between team rebounding and individual rebounding.
There were about 24 seasons where the player played on a team that was equivalent to his normal rebounding team, AND/OR rebounded almost exactly as he was supposed to. This shows a neutral correlation between team rebounding and individual rebounding.

You could go a step further and break down the neutral numbers, but you end up with the same thing. There's no correlation between team rebounding and individual rebounding.

So what happens when a team has all good rebounders? Who loses out? The rule is simple. Great rebounders get their rebounds. When the team starts to hit the high point of the ceiling, they cannibalize from their own players. Great rebounders take rebounds from lesser rebounders—the team is less important. Impact rebounders, the best of the group, always get their rebounds.

Let's look at a more recent player—Zydrunas Ilgasukas. He's a starting C that, since 2003, has played 65-81 games a year.

Code: Select all
Year     Reb %  Team Reb% 

2003     14.1     52.1
2004     14.2     52.0
2005     15.2     52.0
2006     15.3     52.0
2007     16.4     52.7
2008     17.5     53.1
2009     16.5     51.1
2010     15.5     50.8


Well...wtf? The Cavs have always had good rebounding teams. When they got better in 2007 and 2008, Big Z's numbers went up/ When the team rebounding numbers went down in 2009, his Reb % went down too (although it was still over his average of 15.5 for that period). This year, on the worst rebounding team he's been on in the last eight years, he'd bang on his average. So—team rebounds well from 2003 to 2006; Z's rebounding slowly increases. Team rebounds really well in 2007 and 2008; Z's rebounding goes up. Team rebounds worse in 2009 and 2010; Z's rebounding declines.

In this case, there's more of a positive correlation than a negative correlation. The better the team rebounds, the better Ilgauskas rebounds. There are a couple of players like this. There are a couple of others that show negative correlation. Taken as a group, they balance out. The team has little to no effect on individual rebounding.

Other than natural up and down movement—what affects rebounding? Impact Rebounders. An impact rebounder is a player that rebounds 5.5 to 6% higher than an average player at his position. Given that the average Reb % for NBA starters by position are something like

C/Primary Frontcourt Rebounder—15.5%
PF/Secondary Frontcourt Rebounder—13.5%
SF—9.5%
SG—6.4%
PG—5.5%

it's pretty rare to have an impact rebounder. Dwight Howard is there. Not many others.

To give you an idea of the impact of a player like this, a 5.8% difference in rebounding for a player that plays 37 mpg is huge. It's shocking. At 41.5 rebounds per game, it looks like this

(37/48.3)*(41.5+41.5)*.055*82=287
(37/48.3)*(41.5+41.5)*.113*82=589

One player is responsible for a difference of 300 rebounds a game...essentially, turning a poor rebounding team (-100 rebound differential) into an outstanding rebound team (+200 rebound differential). That's the impact that Magic Johnson, in this example, had.

The problem with rebounding—but, at the same time, the thing that makes it a fascinating stat to analyze—is that there is a fixed number of rebounds and the difference between being a good and poor rebounding team is so low. Let's say that your starting SF grabs 6.5 rpg, when the “average” SF gets only 5.7. And your backup PF/C, who is supposed to get 4.2 rpg, gets 4.6 rpg. Everyone on the team is average. The immediate assumption is “well, that won't make a whole lot of difference.” You are in the upper third of teams in rebounding because of those small differences. In other words, there';s not a lot of movement available for rebounding.

So what does this mean? Well, on the most basic level, it means you can't stockpile rebounding. Once you get to a certain point—about 53% of available rebounds--you're topped out. After that, players will cannibalize from one another. What this seems to indicate, is that the better rebounders will maintain their numbers at the expense of the lesser rebounders. So if Marcus Camby changes teams, he's still going to rebound at an elite level.

Let's look at a good rebounder who got brought down by a great rebounder. If Dennis Rodman joined the San Antonio Spurs in 1994—which, by coincidence, he did—we can speculate that he'll take rebounds from the rest of the team. The Spurs were already a good rebounding team. But there's an upper limit to how good you can be, and it's pretty much Darwinian among teammates. Rodman's Reb % didn't suffer with the Spurs. David Robinson's did. Here are Robinson's numbers and his team's numbers from 1990 to 2001.

Code: Select all
Year     Reb %  Team Reb% 

1990     18.5     51.5
1991     19.1     51.8
1992     17.7     52.0
1993     17.1     49.0
1994     14.8     53.0
1995     16.2     51.6
1996     18.4     48.7
1998     18.2     51.4
1999     17.8     49.9
2000     17.1     50.4
2001     16.5     50.4


The Spurs are one of those teams that using Four Factors Rebounding percentages really messes with. I think they're a better rebounding team than this. But let's look at this. The first Spurs teams David Robinson every played on were among the best rebounding teams he played on. And he was pretty much at his peak as a rebounder. In 1992, the team was again a very good rebounding team—and Robinson's rebounding numbers went down. Okay...negative correlation. Then, in 1993, the team's rebounding dropped significantly—but so did Robinson's rebounding performance. The 1996 team was as bad as the 1993—but Robinson rebounded great. The 1998 team was another good rebounding team—and Robinson's rebounding stayed the same and was above average for him. You get the point. 1990-3, and 1996-2001, there's no connection between how the Spurs rebounded and how Drob rebounded.

So why are things so nuts in 1994 and 1995? The Spurs were at their absolute peak as a rebounding team in 1994, and Robinson was awful that year by his standards. They were very, very good in 1995...but Robinsons's rebounding was astill very low for him. Much lower than it had been in 1991 and 1992, when the team rebounded better. Much lower than in 1990 or 1998, when the team rebounded just as well. What caused the odd drop in David Robinsons's rebounding in those two years? Dennis Rodman did. David Robinson's rebounding wasn't affected by how his team rebounded....it was only affected when he played with a guy who was a better rebounder than he was.
Image
TrueLAfan
Retired Mod
Retired Mod
Posts: 7,336
And1: 106
Joined: Apr 10, 2001
Top

Post#2 Sir Charles
Sat Jan 2, 2010 6:41 pm by TrueLAfan

Another interesting player to look at is Charles Barkley. The average team Barkley played on between 1986 and 1995 grabbed 50.4% of available rebounds. In two of those seasons—1987 and 1990—the team rebounded almost dead on that average (50.3% and 50.2%, respectively). Let's put those seasons aside and look at the other eight seasons.

In 1986, 1988, 1993, and 1994, Charles Barkley played on teams that rebounded measurably better than his average teams (51.7%). His Reb % was 18.182.

In 1989, 1991, 1992, and 1995, Charles Barkley played on teams that rebounded measurably worse than his average teams (49.1%). His Reb % was 17.006.

Good rebounding teams—better rebounding. The opposite of what we expect.

What about those other two seasons, though—1987 and 1990? In a way, they provide the most fascinating examples of all. Barkley's rebound rate was 18.665...better than the other./ Here's the wild things—one of those years was his best rebounding year of the period—1987. Reb % of over 20. The other was one of his worst rebounding years—1990. Reb % of 16.7. the two teams were virtually identical in rebounding, but Charles Barkley's numbers varied wildly.

It's surprising how often this happens...that a player is on one of his average rebounding teams an dhas a great rebounding year. Or a terrible one. Which again points to the conclusion that team rebounding has little to no effect on individual rebounding numbers of excellent rebounders.
Image
TrueLAfan
Retired Mod
Retired Mod
Posts: 7,336
And1: 106
Joined: Apr 10, 2001
Top

Post#3 Final Thoughts
Sat Jan 2, 2010 6:50 pm by TrueLAfan

So what does this tell us about why great rebounders get more or less rebounds in certain seasons?

Nothing. That's right, nothing. Not a damn thing.

What is does is eliminate a possible reason why players are often said to rebound well. That has value as well from an statistical analysis point of view. There have been about 375 seasons in the last 35 years where a player has played over 2200 minutes and had a Reb % of over 16.5%. I've gone through about 30 percent of those seasons here (and about 50% of the seasons with 17.5+ Reb %). It's more than enough of a sample size to draw a conclusion, especially when the lack of correlation is so evident.

There are all sorts of reasons why players may or may not rebound well...but unless his team has a better rebounder/impact rebounder on board, it has little or nothing to do with how well his team rebounds.
Image
TrueLAfan
Retired Mod
Retired Mod
Posts: 7,336
And1: 106
Joined: Apr 10, 2001
Top


Return to Statistical Analysis


Who is online

Users browsing this forum: No registered users