SOTD - Different NBA positions and PER differences

[MEDIC]

Excellent post DH.

You have to wonder if it's simply difficult to have a high PER as a SG/ SF based on the formula, or if this speaks to the very shallow depth of wing quality in the league today.

It really does support what many of us have been saying. If you have the opportunity to acquire a top 10 dynamic wing player, you have to pull the trigger. They are rare. When you have one, it makes it a lot easier to build/ design a successful team.

It's also why I've always said I'd rather build with 2nd tier star wing than a 2nd tier star PF. PF's are a dime a dozen in the NBA. They are easily replaceable.

It would be interesting to see what those numbers were in the Jordan era (pre hand check & 3 second).

Goes to show you, all individual stats should be compared in relation to position.

[Ted Lasso]

This is one of the common criticisms of PER. There are others; near total disregard for defence, rewarding what might be deemed inefficient scoring, etc. These are things one ought to keep in mind while forming an opinion on a player with contribution from PER.

Having said that, this isn't quite like the TS% disparity between positions. Historical data could help determine whether this has been a fixture or it's simply the product of a dearth of talent. Beyond that, what's really stopping wing players from scoring more, rebounding better, or creating for teammates more often? Or perhaps point guards and bigs are just more valuable in basketball?

Just throwing some questions out there.

[Fairview4Life]

Coming from someone who uses stats in their job, I find the recent stat development in sports to be severely flawed, especially baseball. Too many qualitative factors at play.

Especially in baseball?

[Rapcity_11]

Coming from someone who uses stats in their job, I find the recent stat development in sports to be severely flawed, especially baseball. Too many qualitative factors at play.

That may be true. However it's still far more useful than not to utilize all the statistics available.

[Basketball_Jones]

Coming from someone who uses stats in their job, I find the recent stat development in sports to be severely flawed, especially baseball. Too many qualitative factors at play.

Especially in baseball?

Probably a typo, im guessing he meant to say basketball.

[Double Helix]

Excellent post DH.

You have to wonder if it's simply difficult to have a high PER as a SG/ SF based on the formula, or if this speaks to the very shallow depth of wing quality in the league today.

It really does support what many of us have been saying. If you have the opportunity to acquire a top 10 dynamic wing player, you have to pull the trigger. They are rare. When you have one, it makes it a lot easier to build/ design a successful team.

It's also why I've always said I'd rather build with 2nd tier star wing than a 2nd tier star PF. PF's are a dime a dozen in the NBA. They are easily replaceable.

It would be interesting to see what those numbers were in the Jordan era (pre hand check & 3 second).

Goes to show you, all individual stats should be compared in relation to position.

I think it's a talent thing. The farthest back I can go is via ESPN and only to the 2002-2003 season. During that year there were 30 SGs with PERs above 15.0. This year there are 17.

We've all seen it in the NCAA this past decade. There just doesn't seem to be anywhere near as many super talented SGs coming out. Last year's draft crop of Beal, Waiters, Ross, seemed to be one of the higher concentrations of SGs near the top of the draft since perhaps the Harden/Evans/Derozan draft. There just hasn't been as many dominant SGs coming out lately for whatever reason.

[Rapcity_11]

By the way DH, I get the impression you like PER a lot. Or at least use it a lot.

I'm curious as to why? It's a pretty good box-score summary, but with so much other stuff available to look at I'm not sure why you want to focus in on PER? For me, it's really on good as a quick indicator and if you have time to do further analysis it can basically be ignored.

[Double Helix]

By the way DH, I get the impression you like PER a lot. Or at least use it a lot.

I'm curious as to why? It's a pretty good box-score summary, but with so much other stuff available to look at I'm not sure why you want to focus in on PER? For me, it's really on good as a quick indicator and if you have time to do further analysis it can basically be ignored.

Advanced stats are just starting to trickle down into mainstream media reporting, GM discussions with fans, and around the water cooler. PER is where it starts for most people. It's not my favorite but because it's growing so much in popularity it's an easy thing to utilize.

I'm actually becoming a pretty big fan of using some of the advanced +/- stuff out there. That work tends to fit the best with what I see when watching games and can actually highlight the value of glue guys where they'd normally be ignored elsewhere.

[dhackett1565]

Looking at the total data set (hoopdata.com - I believe they use unforced PER, which doesn't force the average to 15), the averages break down as follow:

PG: 12.2
SG: 12.22
SF: 12.11
PF: 13.69
C: 14.38

And with a minutes restriction of 10 MPG:

PG: 13.83
SG: 12.8
SF: 12.91
PF: 14.69
C: 15.2

And with a games played restriction of 10+, with no minutes restriction:

PG: 13.43
SG: 12.5
SF: 12.53
PF: 13.94
C: 14.36

And using players above the median minutes played for each position:

PG (21.5): 16.0
SG (23): 14.4
SF (19.1): 15.4
PF (18.5): 15.9
C (20.7): 17.1

All of these sets show a less extreme variation than the 20 MPG restriction. That last seems the most reliable to me - it does support the inflation of C's PER, and similar patterns on the wings, with a less extreme distribution. So, similar conclusions, but with a little less impact (18% versus 22% maximum difference, 4.5% vs 5.8% average deviation).

Not utilizing a games played restriction isn't wise, in my opinion. You now have Daniel Orton's 62.8 PER and Viacheslav Kravtsov 34.37 PER affecting things, as well as some negative and 0.0 data.

A games played restriction of some kind is a must, in my opinion. I appreciate your work on the last offering though. That's helpful. Is that WITH a games restriction of 10 games as well or no?

The minutes restriction thing is true - though it strikes me that I should try to compile a minutes-weighted average per position just for comparison's sake.

That last has no games played restriction.

[MEDIC]

What might even be more interesting (& telling) is looking at the median value of each position, then comparing it to the mean.

It might be a good indicator of talent distribution at each position.

[Double Helix]

Looking at the total data set (hoopdata.com - I believe they use unforced PER, which doesn't force the average to 15), the averages break down as follow:

PG: 12.2
SG: 12.22
SF: 12.11
PF: 13.69
C: 14.38

And with a minutes restriction of 10 MPG:

PG: 13.83
SG: 12.8
SF: 12.91
PF: 14.69
C: 15.2

And with a games played restriction of 10+, with no minutes restriction:

PG: 13.43
SG: 12.5
SF: 12.53
PF: 13.94
C: 14.36

And using players above the median minutes played for each position:

PG (21.5): 16.0
SG (23): 14.4
SF (19.1): 15.4
PF (18.5): 15.9
C (20.7): 17.1

All of these sets show a less extreme variation than the 20 MPG restriction. That last seems the most reliable to me - it does support the inflation of C's PER, and similar patterns on the wings, with a less extreme distribution. So, similar conclusions, but with a little less impact (18% versus 22% maximum difference, 4.5% vs 5.8% average deviation).

Not utilizing a games played restriction isn't wise, in my opinion. You now have Daniel Orton's 62.8 PER and Viacheslav Kravtsov 34.37 PER affecting things, as well as some negative and 0.0 data.

A games played restriction of some kind is a must, in my opinion. I appreciate your work on the last offering though. That's helpful. Is that WITH a games restriction of 10 games as well or no?

The minutes restriction thing is true - though it strikes me that I should try to compile a minutes-weighted average per position just for comparison's sake.

That last has no games played restriction.

Can you do a minutes-weighted AND games played restriction? That seems like it would be the most reflective of the things we're talking about.

Thanks for your contributions, dhackett. I'm not in a position to do much more than comment now and again this morning.

[Rapcity_11]

By the way DH, I get the impression you like PER a lot. Or at least use it a lot.

I'm curious as to why? It's a pretty good box-score summary, but with so much other stuff available to look at I'm not sure why you want to focus in on PER? For me, it's really on good as a quick indicator and if you have time to do further analysis it can basically be ignored.

Advanced stats are just starting to trickle down into mainstream media reporting, GM discussions with fans, and around the water cooler. PER is where it starts for most people. It's not my favorite but because it's growing so much in popularity it's an easy thing to utilize.

I'm actually becoming a pretty big fan of using some of the advanced +/- stuff out there. That work tends to fit the best with what I see when watching games and can actually highlight the value of glue guys where they'd normally be ignored elsewhere.

Fair enough. Glad to see you know your stuff.

I'm also really into the +/- family of statistics. It really sheds a different light on the game.

[dhackett1565]

Total minutes played (not per game) weighted, with no game restriction:

PG: 15.27
SG: 13.79
SF: 14.46
PF: 15.35
C: 16.45

Total minutes played (not per game) weighted, with 10+ game restriction:

PG: 15.35
SG: 13.82
SF: 14.48
PF: 15.35
C: 16.46

As you can see, weighing by total minutes seems to eliminate the need for a games restriction, but it could also unfairly weigh players who have missed time due to injury even though they are very effective in the games they have played. So I also did this:

Minutes per game weighted, with 10+ game restriction:

PG: 14.83
SG: 13.48
SF: 13.76
PF: 15.04
C: 15.98

So that yields the same ugly results for the wing positions, with a more reasonable result for the C position (though not by much, relatively speaking). The entire set centers better around the expected average of 15 as well. Funny how the total minutes averaging and per game minutes averaging yield such different results.

[Double Helix]

Total minutes played (not per game) weighted, with no game restriction:

PG: 15.27
SG: 13.79
SF: 14.46
PF: 15.35
C: 16.45

Total minutes played (not per game) weighted, with 10+ game restriction:

PG: 15.35
SG: 13.82
SF: 14.48
PF: 15.35
C: 16.46

As you can see, weighing by total minutes seems to eliminate the need for a games restriction, but it could also unfairly weigh players who have missed time due to injury even though they are very effective in the games they have played. So I also did this:

Minutes per game weighted, with 10+ game restriction:

PG: 14.83
SG: 13.48
SF: 13.76
PF: 15.04
C: 15.98

So that yields the same ugly results for the wing positions, with a more reasonable result for the C position (though not by much, relatively speaking). The entire set centers better around the expected average of 15 as well. Funny how the total minutes averaging and per game minutes averaging yield such different results.

I'm going to add this to the OP and credit you for the work. Nicely done.

[lucky777s]

Good information to have. But starters are what make or break an nba team, especially in the playoffs.

Raps have an abundance of bench talent, good role players, glue guys, etc. We all know that. But we have not had great starting talent. Thus, our record.

I would raise the minute cut off to 28 or 30.

[ForeverTFC]

PER, like all other stats, need to be looked at in a holistic set. It would be idiotic to compare the FG% of a C and a SG. Same philosophy for PER.

At this point, we're not even looking at the effects of player role, team system, and surrounding players.

A three point shooter, playing with a dominant big man and a slashing PG will always have a higher PER & TS%. Put them on a team with a perimeter PG and no low post threat, and watch both stats plummet.

The casual fan often gets caught in looking at the numbers, without properly analyzing the circumstances.

[dhackett1565]

What might even be more interesting (& telling) is looking at the median value of each position, then comparing it to the mean.

It might be a good indicator of talent distribution at each position.

Using pure means and medians (they will be low due to using a lot of scrubs in the set):

Position - Mean - Median
PG - 12.2 - 12.47
SG - 12.22 - 12.1
SF - 12.11 - 11.88
PF - 13.69 - 13.7
C - 14.38 - 15.32

Using 10+ games played means and 10+ games played medians:

Position - Mean - Median
PG - 13.43 - 13.54
SG - 12.50 - 12.34
SF - 12.53 - 12.78
PF - 13.94 - 14.21
C - 14.36 - 15.32

Using 10+ games played and 15+ minutes per game for both means and medians:

Position - Mean - Median
PG - 15.00 - 15.38
SG - 13.50 - 12.84
SF - 13.74 - 13.52
PF - 14.94 - 14.91
C - 16.90 - 17.81

So, all three sets show that the median players have slightly above average PER's for PG's and PF's; well above average PER's for C's, and all agree that the median SG is below average. There is disagreement on SF's, but I would tend to agree with the most filtered data, that suggests that the median SF is below average.

[dhackett1565]

Good information to have. But starters are what make or break an nba team, especially in the playoffs.

Raps have an abundance of bench talent, good role players, glue guys, etc. We all know that. But we have not had great starting talent. Thus, our record.

I would raise the minute cut off to 28 or 30.

Sure, why not.

Minutes per game cut-off of 30+ MPG:

PG: 17.63
SG: 16.49
SF: 17.67
PF: 18.66
C: 18.42

A little more parity there, but still supports the earlier conclusions - wing players just have lower PER's, no matter what data set you look at.

Just for fun, MPG 35+:

PG: 18.47
SG: 16.57
SF: 19.36
PF: 19.19
C: 19.20

Again, SG's look pretty bad relative to the rest. SF's take a jump, but that has a lot to do with LBJ and Durant now comprising 1/6th of that group.

[trick]

Double Helix and dhackett,

Just wanted to show my appreciation for the work you two did. This could really shed some light into the whole PER conversations. Personally I'm not a fan of PER, but this sheds some positive light onto it making it less fraudulent.

[Ell Curry]

Great work.

Can someone calculate ORTG by position? My hunch is that SG is even weaker than these PER adjusted stats indicate, since PER rewards volume scoring. Someone like DeRozan looks better by PER than he does by pure efficiency.