Avg. NBA NET (player vs opponent) PER by position rankings

[Admin]

[Popeye]

We do need a backup SF, Evans can not guard SF's. When Evans came in for Marvin last game Prince ent off on him.

[crimedog]

Dol, what you have to consider is that this is a PER differential. Neither Horford, Marv, nor Zaza, have especially high PERs. This season our highest PERs from top to bottom are Joe, Mike, Josh, Al, Marv which is the order those are in. Josh is having a down year based on PER because his blocks are down.

[AHF]

I don't really like the net numbers so much for direct rankings. Bynum > Duncan? Utah 16th in the league in PG (even with the injury, really?)?

[Admin]

Dol, what you have to consider is that this is a PER differential. Neither Horford, Marv, nor Zaza, have especially high PERs. This season our highest PERs from top to bottom are Joe, Mike, Josh, Al, Marv which is the order those are in. Josh is having a down year based on PER because his blocks are down.

I realize that it's PER differential, that's why I said that in the subject (LOL) although I used the term they described it as which is NET PER. I think it's a very useful tool because even if one of our guys has a low PER then it would behoove him to force the opposing player at his position into having an even lower PER.

Do you not believe that this is an accurate tool to measure our proficiencies/deficiencies?

[crimedog]

I realize that it's PER differential, that's why I said that in the subject (LOL) although I used the term they described it as which is NET PER. I think it's a very useful tool because even if one of our guys has a low PER then it would behoove him to force the opposing player at his position into having an even lower PER.

Do you not believe that this is an accurate tool to measure our proficiencies/deficiencies?

Well, like any statistical measure it is a pretty useful tool if you douse it with a lot of analysis. Consider though, that different players are going to be asked to do different things. We don't really ask our Cs or Marv to put up a lot of points, consequently, they don't have particularly high PERs. Bibby, being a scoring PG, has a high PER for his position but he tends to get abused with great frequency to the point that he usually doesn't cover his own man if the guy is a scorer.

Having a poor differential at SF and C doesn't really mean those are our weak spots, especially at SF, it just means we don't ask those positions to do the things that really show up big time in the PER formula. On the other hand, I'd say a backup point gaurd that can play a little bit of D (maybe Acie can grow into this role) is a big need but that isn't reflected int he formula.

Edited February 13, 2009 by crimedog

[Admin]

I don't really like the net numbers so much for direct rankings. Bynum > Duncan? Utah 16th in the league in PG (even with the injury, really?)?

Where are you looking to see Bynum > Duncan? Were they on another page? The tables I posted above are the full production by position so it includes the backups at that position as well.

[Admin]

Well, like any statistical measure it is a pretty useful tool if you douse it with a lot of analysis. Consider though, that different players are going to be asked to do different things. We don't really ask our Cs or Marv to put up a lot of points, consequently, they don't have particularly high PERs. Bibby, being a scoring PG, has a high PER for his position but he tends to get abused with great frequency to the point that he usually doesn't cover his own man if the guy is a scorer.

Having a poor differential at SF and C doesn't really mean those are our weak spots, especially at SF, it just means we don't ask those positions to do the things that really show up big time in the PER formula. On the other hand, I'd say a backup point gaurd that can play a little bit of D (maybe Acie can grow into this role) is a big need but that isn't reflected int he formula.

No it doesn't necessarily mean they are our weak spots but I have no doubt that we're weakest at SF since we only have 1 SF on our roster and we get hammered with Evans playing 30+% of the minutes there for us so it definitely adds up. My guess is that the reason C is slightly worse than PF for us is because we had Solo playing a lot of minutes at C when Horford/Zaza were out whereas Horford moved down to PF when Smoove was out so there wasn't as much drop off.

[Guest]

I really wish 82games would use a different metric than PER. Raw numbers would be outstanding, but that is wishful thinking, 82games would never release that information.

[Admin]

I really wish 82games would use a different metric than PER. Raw numbers would be outstanding, but that is wishful thinking, 82games would never release that information.

Well the 3 most common are +/-, PER, and efficiency and many seem to have some sort of issue with each so I don't think any rating would satisfy them.

I wish that there was a publicly accessible database for NBA stats so that we could keep our own and create our own views. You don't know of anything like that do you?

[AtLaS]

For C, we are actually doing well defensively. We have the 12th best defensive PER against opponent C's. Obviously we are struggling offensively at the position, which we already knew. If we start running the PnR more, Horford's offensive production would increase.

Regardless, these stats can definitely be skewed so you have to be careful with them, because they usually favor offensive players. A player who scores 20 ppg is going to outproduce his opponent on most nights even if he's a big defensive liability.

I bet Ben Wallace in his prime had a terrible rating on this scale for center, but he was still one of the most valuable centers in the league.

These ratings almost essentially tell you which positions you rely on the most for scoring, with of course a few exceptions.

[Guest]

I wish that there was a publicly accessible database for NBA stats so that we could keep our own and create our own views. You don't know of anything like that do you?

Basketball-reference is the closest, but 82games has actual data that tells you how a team performs with certain players on the court and off the court. They have mountains of information, all that is available in raw form to the public is the box score.

Some teams hire their own statisticians to gather the same (and even more) data that 82games.com has. I am sure some teams even pay 82games for the data they collect. I remember reading an article on some MIT grad student (maybe another university?) who was originally doing free work for the Celtics. They liked his stuff so much that they decided to pay him and pretty well. I think he started 3 years ago helping out. I do know the Mavericks paid some guy for his specific player rating formula (think PER but better because this guy actually has a graduate degree), but I forget what his name is.

[mrhonline]

Remember that Marvin is often playing at the 4 spot.

Marvin

According to that, he's responsible for only 41% of the SF minutes...Mo Evans is 36% of that, Joe is 19%.

Also, over half (56%) of the center minutes are going to Solomon, Morris, and Zaza.

Edited February 13, 2009 by mrhonline