OT: Pistons' Andre Drummond could rank as 'greatest teenager' in NBA History

Picture of Jed Clampett goes here.

I don't think anyone seriously thinks Drummond is better than LeBron.  But rate stats generally hold up pretty well in the NBA when a player is given more extended minutes, and if the Pistons have a big man who puts up a PER in the 20s while playing decent defense, that's awesome for them.

Right. It's an efficiency rating. He's certainly a more efficient player than Lebron. 'Bron (in his 1 NBA season as a teenager) averaged 3x the points on over 3.5x the shot attempts, 1/2 the rebounds in 2x the minutes, A:TO much better, and stl+blk are ~equal in 2x min.

That Cleveland team finished 9th and 12 games under .500.  

It was never meant to identify the best player in a league like the NBA. I think it does help to assign a value to guys like Drummond.

He will likely never be MVP and he can't take over a game like Lebron so it's difficult to compare them. A better, and more relevant question: At this point would you take any other rookie, other than Davis (not yet in ROY comp due to injury and minutes himself), over Drummond? 

I would change that to say that maybe the author is getting ridiculous in his interpretation and application of the metric.  But the objective formula of PER is not ridiculous.

PER is an all-encompassing metric, but it is just a tool to be used to measure player performance on a per-minute basis.  And it should be used as just a tool, not a definitive declaration of value.  There is a lot of nuance and context needed to derive a player's true value and PER as well as other metrics like win shares/wins produced/WARP (i dont actually like these as much) are part of that.

The author Tom Haberstroh probably went too far in asserting that because he had the highest PER that he's had the best season by a teenager in NBA history.

A couple problems with that:  1st) its very selective in defining teenager.  LeBron was still 19 for part of his 2nd season in the league when he had a ridiculous for his age 25.74 PER.  If LeBron was born a few months later, he'd probably have been just as good and would've qualified for this made up best teenager title.

2nd) He does adjust for minutes by extrapolating his numbers to per 36 minute production to better gague full game production.  Studies have shown that most players DO maintain per-minute production over larger minute loads (see Harden, Ryan Anderson etc), but that does NOT mean that his value in terms of what he's given to the Pistons should be judged by extrapolated stats.  He might have done just as well given more minutes, but that can't be assumed.  Hollinger's PER converts into expected wins added once adjusted for minutes played and in that calcualtion, he's barely ahead of Damian Lillard (4.2 to 4.0)

3rd) Advanced basketball stats need to be combined with a subjective sense of both the players role on offense and his ability on defense (which advanced stats suck at measuring by and large).  He scores mostly out of ball screens and on putbacks and while he cant create his own shot, he does indirectly create shots for others as a roll man in the PnR by sucking defenders in to defend the lob. There is major value in a roll man who does these things, but it pales in comparison to elite wing creators whom PER rates as his equals.

4th) Among the reasons hes played so few minutes are spacing on offense and his own defensive aptitude.  Frank tries to limit the mintues Monroe and Drummond play together and without a stretch 4, things can get mucked up inside a bit.  And his help defense and sense of when to rotate/when to stay home could use some work.  I could expand on points 3 and 4 but this is getting pretty long.  The gist - advanced stats need a ton of context and thus shaky conclusions drawn solely from a spreadsheet should be more an indignation of the person evaluating than of the stats themselves - the stats are best used as a tool to help paint a broader picture of value.

This is not where PER gets ridiculous. PER is simply a measurement of a player's effectiveness on a per minutes basis based on a formula.

Andre Drummond hasn't played a tremendous amount of minutes, but his effectiveness during those minutes have been impressive, therefore he has a high PER (as he should).

The ridiculousness is the article pointing at per 36 minute stats (which is entirely different from player efficiency rating, P.E.R.), and assuming that if Drummond were in the game for a full 36 minutes, we wouldn't see any drop off in his efficiency or effectiveness, when we know for a fact that part of the reason he hasn't logged more minutes is due to his conditioning.

Where PER get ridiculous is that according to the formula, a player only has to shoot something like 30% from the field and 20% from 3 in order to break even in terms of PER. So someone like Demarcus Cousins can shoot an atrocious percentage from the field as a big man (like 41%) but keep racking up his PER, even though he's killing his team. One of the primary flaws in PER is that it rewards high volume scorers too much.