RealGM

According to 82games.com

the Net PER by position for the Celtics this year (ie. Celtics Position PER - opponents Position PER) works out to:

Position ******* Celtics PER **** Opponents PER ***** Net PER

Small Forward *** 20.1 ************ 12.8 ************ 7.3
Power Forward ** 19.3 ************ 13.0 ************ 6.3
Shooting Guard ** 17.1 ************ 12.5 ************ 5.6
Point Guard ***** 16.2 ************ 13.9 ************ 2.3
Center ********* 16.2 ************ 16.8 **** negative 0.5

Keep in mind that Kevin Garnett logs 11 mpg as a center, and
James Posey logs about 14 mpg at power forward.

If you want to see how it breaks down individually, the NET Production for each Player is:

+16.5 Kevin Garnett
+ 9.9 Paul Pierce
+ 5.8 Ray Allen
+ 2.6 Tony Allen
+ 1.8 Eddie House
+ 1.6 Gabe Pruitt
+ 1.4 Rajon Rondo
+ 0.6 Leon Powe
- 0.2 Glen Davis
- 4.8 Kendrick Perkins
- 5.2 James Posey
- 8.1 Brian Scalabrine
-11.4 Scot Pollard

KG is the obvious MVP.
How many victory cigars did some clowns award him during that 29-3 streak???

This would seem to indicate that the Celtics worst player, factoring offense and defense, is Scot Pollard, followed by Brian Scalabrine. They would seem to be the two that you want to replace the most; Pollard with a more productive center, and Scal by sitting him until he is hopefully traded this summer.

Posey's athleticism is essential in many matchups, so his lack of scoring compared to his opponent is somewhat offset by his giving up shots to other Celtics, spreading the floor with his 3 point shot, and just playing defense when no one other than Paul Pierce is as suited.

Perk, likewise, has size in the middle that no one other than Pollard can offer. At times, size is essential, especially for screening players out.

Part of the reason that Perk's net numbers are so bad is due to that he gives up shots to other Celtics. Also, Perk is not able to generate his own shot, so that partly explains why he scores so little for the time that he is on the court.

Garnett is amazing, Pierce also (his defense this season is overlooked).

Ray Allen is clearly a significant positive, despite taking 6 less FGA and 1.2 less FTA than last year.

Tony Allen doesn't play more because as Ray Allen's backup, he basically only gets on the floor when Ray needs a breather, and Tony is not as good as Ray. There are a few occasions where the two Allens are on the floor together, with Ray playing small forward.

House can't do what Rondo can, but in his limited minutes, he is effective. For example, if House doesn't come into the San Antonio game and score those 10 points in 4 minutes, the Celtics would not have won that game.

Rondo plays 31 mpg to House's 19.7 mpg; Rondo's essentialness (?) is emphasized by his mpg.

Davis and Powe are close, with Powe having a slight edge at this point.

And as crappy as Gabe Pruitt's offensive numbers are (shooting %, points scored); apparently he at least has been playing some defense during his 58 minutes of action.


The Link is below with all the details.

http://www.82games.com/0708/0708BOS5.HTM

And for the Celtics Team stat page

http://www.82games.com/0708/0708BOS.HTM

PogueMahone:

Hi:

If you read this, I'd be interested to know your thoughts on the relative merits of the data that 82games.com provides, particularly with regard to their calculation of NET PER by position, and NET production by individual.

Thanks.

-- Jammer

Jammer wrote:PogueMahone:

Hi:

If you read this, I'd be interested to know your thoughts on the relative merits of the data that 82games.com provides, particularly with regard to their calculation of NET PER by position, and NET production by individual.

Thanks.

-- Jammer

Other people have said it better than me but it's hard to quantify match-ups and defensive breakdowns solely from play-by-play data without running a regression on the numbers to separate the wheat from the chafe. That being said, another thing that needs to be kept in mind is that 1 unit of Usage Rate (Usg-R) is worth approximately .6 points of PER. IMO, what you are seeing on the positional PER comparisons is the fruit of that inequity.

Consider the following:

PG: -2.07
SG: 1.89
SF: 3.32
PF: -1.25
_C: -2.69

Essentially, I have converted the per48 numbers into per minute, then calculated Usg-R (estimation of possessions used per 40 minutes.) By position (and using 82games positional data, mind you), the centers use the least possessions on the team as compared to their counterparts.

Take Perkins, for instance. He uses almost 5 possesions less, per40, than his counterparts (going by 82games data.) As I stated above, 1 unit of Usg-R is worth ~.6 points of PER. Quick math puts Perk at ~16.4 PER after the usage adjustment. According to 82games, he gives up a PER of 16.2.

Or you could take it one step further and normalize Usg-R to league average to put everyone on equal footing. Once you get the difference between actual Usg-R and league average, the simplified PER adjustment up above gives you a pretty good idea of what exactly is happening as far as 82games data permits and within the limitations of PER.

For instance, even after a Usg-R/PER adjustment, Pierce has been absolutely filthy defensively. Conversely, R. Allen has been getting absolutely abused playing minutes at SF.

Pogue Mahone wrote:

Jammer wrote:PogueMahone:

Hi:

If you read this, I'd be interested to know your thoughts on the relative merits of the data that 82games.com provides, particularly with regard to their calculation of NET PER by position, and NET production by individual.

Thanks.

-- Jammer

Other people have said it better than me but it's hard to quantify match-ups and defensive breakdowns solely from play-by-play data without running a regression on the numbers to separate the wheat from the chafe. That being said, another thing that needs to be kept in mind is that 1 unit of Usage Rate (Usg-R) is worth approximately .6 points of PER. IMO, what you are seeing on the positional PER comparisons is the fruit of that inequity.

Consider the following:

PG: -2.07
SG: 1.89
SF: 3.32
PF: -1.25
_C: -2.69

Essentially, I have converted the per48 numbers into per minute, then calculated Usg-R (estimation of possessions used per 40 minutes.) By position (and using 82games positional data, mind you), the centers use the least possessions on the team as compared to their counterparts.

Take Perkins, for instance. He uses almost 5 possesions less, per40, than his counterparts (going by 82games data.) As I stated above, 1 unit of Usg-R is worth ~.6 points of PER. Quick math puts Perk at ~16.4 PER after the usage adjustment. According to 82games, he gives up a PER of 16.2.

Or you could take it one step further and normalize Usg-R to league average to put everyone on equal footing. Once you get the difference between actual Usg-R and league average, the simplified PER adjustment up above gives you a pretty good idea of what exactly is happening as far as 82games data permits and within the limitations of PER.

For instance, even after a Usg-R/PER adjustment, Pierce has been absolutely filthy defensively. Conversely, R. Allen has been getting absolutely abused playing minutes at SF.

Pogue:

Thanks for responding, and thanks for the education.

You offer interesting comments on Perk, Paul Pierce and Ray Allen.

Anything insightful to offer on Rajon Rondo, Posey and House??

-- Jammer

^ I'm not sure if you were alluding to this Pogue but I have always thought that usage was one of the sketchier parts of PER. Back in the day it seemed like Walker/AI was actually benefitting PER-wise from shooting 6-22. Hollinger would always attribute this too the fact that you would rather have AI take a contested tough shot than one of his teammates take any shot at all esp at the end of a shot clock. I think his intentions were good that is that he wanted people to understand the value of creating shots, but that his implementation seems to be off.