Marcus Camby first day news thread

"Mean" means average player, not average starter. Perhaps we're both incorrect in what we think the '15' means. I'd like to see the median.

 

PER takes minutes played into account, though. If not, it seems like a statistical measurement that is slanted to give starters a higher number. I still await Hollinger's explanation. I'll throw in that Batum, Bayless, and Cunningham are all near or above 15 in PER. I'd like to know how that is possible given their low amount of minutes.

 

I noticed the same thing. I'm not sure what it means.

 

That isn't what I'm saying, that individual PERs are affected by minutes played. I'm saying that the league-wide PER much more reflects the play of the starters because the starters play the majority of the minutes in the league. So the 15.0 average PER is created largely by the starters.

 

My question is much more narrow. Does PER reflect the "average starter", or is it an "average player" statistic. If not, what are the parameters. It seems to me that this question is up for debate, and I hope to get an answer for it. I'm not wondering what sets the value, but rather, what the value means in terms of player efficiancy.

 

Efficiency is by definition "Per something", not total. That's why people with small minutes played can be over 15. Again, look at the math - you tart with "unadjusted PER" which is:

uPER = (1 / MP) *
[ 3P
+ (2/3) * AST
+ (2 - factor * (team_AST / team_FG)) * FG
+ (FT *0.5 * (1 + (1 - (team_AST / team_FG)) + (2/3) * (team_AST / team_FG)))
- VOP * TOV
- VOP * DRB% * (FGA - FG)
- VOP * 0.44 * (0.44 + (0.56 * DRB%)) * (FTA - FT)
+ VOP * (1 - DRB%) * (TRB - ORB)
+ VOP * DRB% * ORB
+ VOP * STL
+ VOP * DRB% * BLK
- PF * ((lg_FT / lg_PF) - 0.44 * (lg_FTA / lg_PF) * VOP) ]


And notice the first element - (1 / MP) * (which is another way of saying - take all the things we can measure that come below and divide them by the minutes played).

The next step is to adjust it to pace:

aPER = (pace adjustment) * uPER

and finally, normalize it a league average of 15

PER = aPER * (15 / lg_aPER)

Of course, lg_aPER is the same calculation as aPer only for all the players in the league based on the number of minutes they are playing.

So - basically, you calculate the individual efficiency, divide it by the composite league wide efficiency and multiple by 15 to normalize across seasons. The numbers are "per minute" as I said, because lg_aPER will also start the same calculation by dividing by minutes played - (1/MP)*

 

My understanding is that it reflects the "average minute" since it's a per-minute stat. Who plays that average minute? All the players in the league are part of that, but mostly it is the starters since they are playing most of the minutes. So the average PER is closer to "average starter" than to "average player," but it's not precisely either.

It'll be interesting to see what Hollinger says in answer to your question, but that is what I've drawn from what he's written before and my own understanding of the stat.

 

I do not think you can say "average starter" really - and maybe I should not have used this word - because Hollinger's "Pretty Good Player" is much better - it does not use the dreaded "average" word which has a vague statistical meaning as well.

Anyway you look at it - I doubt very much that the average player in the NBA is "PGP" when you measure in a closed group because it makes the use of the word "good" to be superficial.

 

You used it though, and you beat me over the head with it. I just want to see what Hollinger's answer is to the question!

 

Well, if 15 is the "average minute", couldn't that mean that Outlaw is an "average player"?

 

No, it means he produces the same amount as players produce, on average, each minute. Since by far the most of those minutes are being played by starters, his production would be closer to average starter than average player.

 

Then 15 isn't a real average, at least in terms of measuring a player's value. I wonder what the mean is, and I wonder what the median is for PER. That would be better to assess average, good, and bad.

 

Sure it's a real average. It measures the average player-minute. It's just that most of those player-minutes are played by the better players (since they deserve the most minutes), so if you reach that standard (15.0 PER), you're a good player. Thus Hollinger's own designation of 15.0 PER as a "pretty good player."

Whether it's useful in judging player value is up to opinion.

 

I was just wondering if it meant "average starter". Fuck. You guys win, whatever that means.

 

I don't think it was a slight to Miller. Most likely he was just thinking of players on the Blazers that he hasn't played with that he looks forward to playing with. That doesn't mean he's not happy to be playing with Miller again.

 

And my answer is/was: no, not exactly, but pretty close. (For reasons stated in previous posts.)

Nobody wins when someone gets upset...!

I'm not sure why you're agitated. You seemed to want to discuss it, so I was discussing it with you. If you didn't want to discuss it, no one's forcing you.

 

Why am I agitated? You're giving an unknown answer to combat my own unknown answer. Geez. You win. I'm fine with that.

 

Yes, that's what I was wondering too.

I don't know what that means. But okay.

 

So, we're done talking hoops? Or are you going to continue to psycho-analyze me? What the fuck.

 

PapaG? Is that you?