ESPN adds new overall stat...REAL PLUS/MINUS..Finally!

[Doctor MJ]

RPM: Player's estimated on-court impact on team performance, measured in net point differential per 100 offensive and defensive possessions. RPM takes into account teammates, opponents, coaches and additional factors

....wat

Chiming in here, apologies if I'm being redundant:

This is ESPN getting with the times, in the sense that NBA teams have been hiring stat guys to make stats like this for over a decade. Cool that they've taken it mainstream.

"Real Plus/Minus" is basically ESPN re-branding Engelmann's xRAPM, so the stat isn't really new, it's just being re-packaged.

And so you know...I don't like xRAPM in the way Engelmann intended it to be used, or as it's being used here.

This is a stat that factors in everything possible to make the best possible guess as to how much impact a player is actually having...and in doing so it factors in things that are nothing but correlation. For example, aging curves. Because older players tend to be more impactful than younger players once you adjust for box score stats, this stat will effectively give older players a boost. That might be useful for someone who doesn't know anything, but if you know what you're doing you already know which guys are savvy and which aren't and you'll want to factor that in in your own minds.

Additionally, the use of Bayesian priors here mostly means factoring in at least the previous season of play. To be clear, a prior isn't factored in the same way that the stats of the season in question are - the two aren't literally being conflated. This doesn't change the fact though that prior seasons color the results.

So when you read this:

LeBron has been accused of coasting at times this season, and RPM reveals where the charge may have merit.

While LeBron was aptly named to the NBA All-Defensive Team last season, his defensive impact this season has been surprisingly mediocre, at least as measured by RPM (-0.21 DRPM). His rating is consistent with the claim that James has not consistently given his best effort this season on the defensive end.

Does the same hold true on offense? No. Offensively LeBron has been as dominant as ever, and his +8.32 ORPM leads the league by a comfortable margin. This means that despite his less-than-stellar defensive contribution, LeBron's total RPM (+8.11) still rates as the NBA's highest this season.

It couldn't be more wrong. It attempts to refute the allegations that LeBron's impact has been less in '13-14 by using a stat that essentially gives him the benefit of the doubt based on prior seasons...and which has a cousin stat that does NOT color the data the same way. Guess how LeBron looks when you eliminate the priors? LIke a guy who is coasting a good bit this season on offense.

In the end, everyone should be using +/- stats to go along with box score stats and of course observations and critical thinking. If this stat helps people do that it's probably for the best, but it's off to a pretty questionable start.

[Leslie Forman]

As long as the benefits take place while Augustin is on the court, then of course it does. Why would you think it doesn't? It would take this into account because this would show up in the +/- ratings. Anything, literally anything that you do that contributes to your team scoring a point, or stopping the opposing team from scoring a point will show up in this stat, because it is based on how the TEAM DOES WHILE YOU ARE ON THE COURT. So anything, literally anything, that you do while you are on the court to help your team will show up in this stat.

What doesn't show up is off the court stuff, locker room presence, how you may help your team in practice or anything that goes on that you may do to help or hinder your team while you are NOT on the court.

That is my point. DJ Augustin is helping everyone on the Bulls, including those who aren't on the court with him.

His +/- says he is having a negative impact, whereas the eye test, production numbers, and, most importantly, the win-loss column all says he's had a positive impact. Hmmmm maybe one of those things is wrong?

I suggest you look at the WAR ranking for a more accurate portrayal of who the top contributer's have been.

You mean the WAR rankings that have DeAndre Jordan in 7th, ahead of every other center in the league?

No thanks.

[MisterHibachi]

Someone clarify this for me.

This stat has Nick Collison 6th at 5.81 and Kevin Love 10th at 5.40.

Does this mean that Collison does better in his role than Love in his? And not that Collison is better than Love in a vacuum. Simply that, Collison is better at doing what his role demands of him (setting screens, rebounding, defense) than Love is at his role (first option, rebounding, defense).

Am I correct in this interpretation?

[jazzfan1971]

1)Is there a single stat out there which you don't disagree with a few outcomes?

I've never seen one that sorts for best player in a way that doesn't have weird results. I think it's possible to get such a sort, we've just not hit upon the methodology yet.

[jinxed]

As long as the benefits take place while Augustin is on the court, then of course it does. Why would you think it doesn't? It would take this into account because this would show up in the +/- ratings. Anything, literally anything that you do that contributes to your team scoring a point, or stopping the opposing team from scoring a point will show up in this stat, because it is based on how the TEAM DOES WHILE YOU ARE ON THE COURT. So anything, literally anything, that you do while you are on the court to help your team will show up in this stat.

What doesn't show up is off the court stuff, locker room presence, how you may help your team in practice or anything that goes on that you may do to help or hinder your team while you are NOT on the court.

That is my point. DJ Augustin is helping everyone on the Bulls, including those who aren't on the court with him.

His +/- says he is having a negative impact, whereas the eye test, production numbers, and, most importantly, the win-loss column all says he's had a positive impact. Hmmmm maybe one of those things is wrong?

I suggest you look at the WAR ranking for a more accurate portrayal of who the top contributer's have been.

You mean the WAR rankings that have DeAndre Jordan in 7th, ahead of every other center in the league?

No thanks.

According to PER (basically the sum of his box score stats - an offensive measure ) Augustin is just slightly above average. RPM has him just barely below average. So not much difference. I think we can agree that Augustin is about an average offensive player. All the data tells us the same thing.

But defensively, his DRPM is atrocious. That's is why his total RPM says he is a bad player, because he stinks defensively. As a team the Bulls give up 5 more points per 100 possessions on the defensive end when Augustin is on the court than Hinrich. Do you have any evidence to suggest that he is not an atrocious defensive player?

As for DeAndre Jordan, which center would you put above him?

Even by other metrics DeAndre Jordan ranks high. Win Shares has him 9th overall and also as the top center.

So maybe it is your perception of DJ that is wrong and not the other way around?

Do you disagree that DJ is a slightly above average offensive center, and a real good, but not spectacular defensive player who plays a lot of minutes and stays healthy? Because that is what RPM says and that is why he is rated highly.

[jinxed]

Someone clarify this for me.

This stat has Nick Collison 6th at 5.81 and Kevin Love 10th at 5.40.

Does this mean that Collison does better in his role than Love in his? And not that Collison is better than Love in a vacuum. Simply that, Collison is better at doing what his role demands of him (setting screens, rebounding, defense) than Love is at his role (first option, rebounding, defense).

Am I correct in this interpretation?

Pretty close. What it is saying is that Collison contributes more to the Thunder winning in his 16 minutes per game than Love contributes to the Wolves winning in a comparable 16 minute stretch.

Now of course, Love plays way more than that, and so when you take into account minutes played KL has added about 13 wins above a replacement player for the season while Collison has only added 6.

You may think this is absurd about Collison, and one could say it was a fluke except that Collison is a rock star in this stat every single year. He is just awesome at what he does. We really need to change our perceptions about him instead of bemoaning the stat because it says he is great.

[Doctor MJ]

Someone clarify this for me.

This stat has Nick Collison 6th at 5.81 and Kevin Love 10th at 5.40.

Does this mean that Collison does better in his role than Love in his? And not that Collison is better than Love in a vacuum. Simply that, Collison is better at doing what his role demands of him (setting screens, rebounding, defense) than Love is at his role (first option, rebounding, defense).

Am I correct in this interpretation?

I think the first thing to have in mind is that a player who plays <20 MPG and does so as a supporting player is not having the gauntlet thrown at him the same way a star is. Guys like Love are out there on the court as much as they can stand and have the entire opposing team geared to stop them, and hence it is a far more significant accomplishment than what Collison is known to be doing.

The second thing I'll note is that when we look at RAPM data throughout the years, we don't typically see role players rack up big numbers year after year. You get fluke years, and then you get a guy like Shane Battier who on a typical year is actually more impactful than a lot of "stars" who aren't actually functioning that well in a team context.

Collison is an interesting case because he's been getting huge RAPM numbers in very limited minutes for the past 5 years. It's a consistent enough trend that it doesn't make sense to simply dismiss it to noise. For whatever reason, the Thunder are really successful when Colliison plays.

This makes one ask the pair of questions:

Is this because the Thunder are using Collison brilliantly?, or Is this because the Thunder are radically underusing Collison?

I'd say the answer is a bit of both.

The Thunder have made it their focus to work on making Ibaka part of a Durant/Westbrook-led Big 3. He's an immense physical talent, he's young, it makes sense. As such they slot in Collison as needed to fill in the gaps. That means he plays disproportionately against weaker opposing players - which is in theory adjusted for by these stats, but if a player is disproportionately effective against weak opponents relative to what others do there's no way to adjust for that individual deviation. It also means that in times where he plays more it's likely because his savvier understanding of the game really comes in handy.

Additionally, while OKC - like most teams - tends to have a forced fit in the starting lineup trying to make it work with huge talents that don't necessarily go together, successful benches are built on great fit. Hence when you see a guy like Collison shine on a successful bench, it's because he works really well with the guys he plays with. This is not to say he doesn't deserve credit, but it's close to a given that if you simply tossed him on another team he would not duplicate this level of success.

However, if the question is whether the OKC would be better if they played Collison more. I think the answer is a resounding yes. Through a process that involves a great deal of luck, they've found a way to great consistently great efficacy with Collison in the lineup for 5 years in a row. At a certain point if you're trying to win a title, not working to capitalize on that more is simply not recognizing what you have in front of you. I understand that if you see it as a choice between Ibaka and Collison it's tough, but good coaching nowadays means breaking the 5 position mold and just figuring out how to exploit everything you've got as much as you can. Sure there will be times you can't play both of those guys together, but that doesn't mean you never do it.

In the end I suppose I feel this is therefore symptomatic of a cowardly approach to coaching from Brooks that ties in also to why they lucked into having James Freaking Harden on their team and they never gave him a shot in a starring role despite the fact the fit between their two established stars has never been as seamless as you'd like. If OKC were a less successful team then I expect Brooks would have tried other things with less fear of failure, but when you're a young coach expecting a blast of success your first time out it takes some serious balls to say "This is good, but I know it can be better".

Spo in Miami has done it, but he also did it in a culture that was "Dynasty or Bust". Had the team failed to win a title the 2nd year, Spo was dead one way or the other so he might as well give it a shot. Brooks on the other hand after the '09-10 season basically guaranteed himself a solid career as an NBA coach simply by being Durant's head cheerleader as long as he didn't do anything that blew the whole thing up, and so he made sure not to take risks.

[20MexicanosIn1Van]

If you don't understand linear regression you can't really comment on the validity of this statistic. It is very obvious to me that many people in this thread don't understand linear regression.

You also cannot disagree with a stat based on some select outcomes. You can, however, disagree with the methods. Unfortunately ESPN isn't going to publish the methods so discussing the validity of this statistic is really kind of pointless. I can imagine the methods they use and if they used these methods they really can isolate the effect of a certain player as they claim.

[Doctor MJ]

If you don't understand linear regression you can't really comment on the validity of this statistic. It is very obvious to me that many people in this thread don't understand linear regression.

You also cannot disagree with a stat based on some select outcomes. You can, however, disagree with the methods. Unfortunately ESPN isn't going to publish the methods so discussing the validity of this statistic is really kind of pointless. I can imagine the methods they use and if they used these methods they really can isolate the effect of a certain player as they claim.

It's really not that great of a mystery. Go over to APBRmetrics, read stuff by the poster J.E. They're just using his stat with a sexier name with perhaps some very slight tweaks.

I would agree with you though that you need to understand regression analysis before you look to criticize or praise.

[HotrodBeaubois]

#1 Power Forward ---One Foot in the Grave Dirk "Diggler" Nowitzki

[og15]

If your new stat says that Channing Frye is better than Blake Griffin, it might not be a very good stat. And Iggy > Durant. And Nick Collison > Tim Duncan. This thing is

Brandan Wright is better than Griffin on PER. PER is trash, too? Every advanced stat is flawed, but having them helps us understanding the game much more.

I used RAPM this years to see which role players are hugely undervalued. Like with PER, it's flawed on some players, but it's still a good stat.

And Iguodala is not having a bad year, he's been outrageously good defensively.

F*** John Hollinger. He has ruined a once very good Memphis Grizzlies team. Outside of the trade of Rudy Gay, all his moves have caused the grizz to take a huge step back, esp the firing of lionell hollins.

Is Hollinger even the Grizzlies GM? Lol

[Knosh]

he's a great coach, but he must be the best coach in the history of the league if the bulls are outperforming their individual WARs by that much. the spurs aren't that far out of whack, and popovich is pretty universally considered an excellent coach

and i don't think WAR is supposed to work like that anyway. it relies solely on what has happened on the court and distributes that value to players

From the ESPN article:
"Drawing on advanced statistical modeling techniques (and the analytical wizardry of RPM developer Jeremias Engelmann, formerly of the Phoenix Suns), the metric isolates the unique plus-minus impact of each NBA player by adjusting for the effects of each teammate, opposing player and coach."

finding the word coach in the analysis does not mean that the coach can be given a WAR rating. which would have to be done to explain how the chicago bulls have such a low sum of individual WARs

the excerpt you just noted backs up what i said: coaching impact is supposed to manifest itself in the final player ratings. put differently, the statistical noise of teammates, coaches, and opponents is supposed to be filtered out, leaving on-court performance to be analyzed numerically. point differentials and team wins are ultimately determined on the court as a result of player actions

As some have already pointed out, this stat isn't new. It is just new on ESPN. I know for a fact that coaching is one of the explanatory variables used in the modell. I gave you the quote, because I thought you would prefer a source instead of a "No, you are wrong". I gave you a link to Engelmann's coaching rating, showing you that Thibs is indeed ranked very highly.

The quote says that the rating is adjusted for the effects of teammates, opponents and coaching. The article doesn't state it clearly, but the method used is a regression (They say "employing techniques similar to those used by scientific researchers when they need to model the effects of numerous variables at the same time").
So from there you might be able to believe me that coaching is one of the explanatory variables, along with teammates and opponents.

[wilt]

#1 Power Forward ---One Foot in the Grave Dirk "Diggler" Nowitzki

good thing he shoots one-footed then

[mysticbb]

but there's no getting around the fact that the bulls have been far more successful with augustin on the roster than they were with deng.

But the Bulls played better basketball with Deng on the court than they are playing with Augustin. Can you imagine that it is mere coincedence that the Bulls won more with Augustin, because they played better players more minutes?

it relies solely on what has happened on the court and distributes that value to players

Coaches are added as 6th player to each lineup in the regression. But Jerry said on APBR that the coaching adjustment is indeed not included, thus the description on ESPN is wrong.

i guess synergy's "eye test" can't be taken seriously either

He is average as an 1on1 defender, that's what Synergy says, but Synergy doesn't tell you anything about him as a team defender, how is help defense is (which is bad), how he rotates (which isn't good) or how he acts in transition or off the ball (which isn't good either). So, that's what my eye test says about it. If you believe, he is actually good at that, well, we have to agree to disagree on that part.

he was indeed a positive for the team, which is just more evidence that losing him and bringing in augustine should not have resulted in the team taking off given the supposed huge negative WAR impact that the move had

No, why should that be the case? Deng's minutes didn't went to Augustin, but to Dunleavy, Butler and Hinrich, al 3 are positive players as well. That the Bulls wouldn't lose much when all of those 3 are available while Deng would be not, was something we could expect. And when you take the minutes of each player into account you will see that it turns out be right were the Bulls played since the Deng trade. So, you might believe that Augustin is actually better, thus which player do you believe is then worse than his rating? Hinrich, with whom the Bulls play about 9 points better than with Augustin and the same 4 temmates (adjusted for the strength of opponents)? In that rating the Bulls have:
Noah +3.9
Mohammed -3.1
Gibson +4.1
Boozer -3.5
Dunleavy +3.2
Butler +1.9
Hinrich +0.3
Snell -5.5
Augustin -4.7

So, the Bulls were in those last 45 games since Deng's trade at +3, the rating measures them out to be at +2.3. If we put all the value "missing" into Augustin we end up with -3.5 for him instead. Sounds that better to you?

and deng started hot for the cavs and fell off hard

Deng for his first 19 games for the Cavs was 14.7/5.5/2.2, 50.7 TS% and 9.2 TOV%, for his last 20 games he was: 14.1/5.5/2.5, 50 TS% and 8.1 TO%. More importantly, the Cavs played about the same with and without Deng for those first games, now they are playing about 2.5 points better with Deng on the court. It needed some time for the Cavs to work with Deng on the court together in order to achieve better team results. But you probably want to refer to his first 6 games, when Deng scored nearly 19 points and had 61 TS%, while rebounded worse, had less assists, turned the ball over more and the Cavs were about -4 with him on the court (per 100 poss)? So, he started "hot" as a scorer while playing worse in all other areas, makes you conclude that he "fell off hard" since then?

[Keller61]

PER is a box score measurement. It tells you which players put up the best box scores and its pretty good at doing that. It does what it claims to do. Not garbage.

it's widely referenced as a single number indicator of a player's value. and by that measure it is severely flawed. it largely ignores defense and overvalues chucking

there ARE no reliable widely referenced single number indicators of player value. i'd say RAPM comes about as close as any, though there is an obvious problem with outliers

Well, people should learn what each stat tells them before using it. It's not the stat's fault that people use it wrongly. PER ignores defense because the box score ignores defense. If you don't trust the box score, you shouldn't trust PER. But the box score does tell you important things. It's important to know who had how many points, rebounds, assists, whatever. Box score ignores context, but that's what the eye test is for.

Obviously no stat is the end-all, be-all, RAPM and PER included. I don't think anyone is arguing that. They're all part of the puzzle, tho.

I don't understand how these stats can be part of a puzzle. What puzzle do they fit into? They are just numbers that tell you no specific information.

[dice]

but there's no getting around the fact that the bulls have been far more successful with augustin on the roster than they were with deng.

But the Bulls played better basketball with Deng on the court than they are playing with Augustin

the bulls were a losing team with deng on the court playing heavy minutes

Can you imagine that it is mere coincedence that the Bulls won more with Augustin, because they played better players more minutes?

deng's minutes were replaced by supposedly roughly equivalent performers. how does the team get much better adding a supposedly negative performer like augustin and playing him significant minutes? doesn't make any sense at all

at the very least augustin has been very good offensively. and you recognize that his man-to-man defense is not bad at least. a negative rating for the guy just has no basis in reality

i promise you that dj augustin has had more of a positive impact on the chicago bulls than mike dunleavy has. yet dunleavy's rating is excellent while augustin's is piss poor? c'mon

[mysticbb]

the bulls were a losing team with deng on the court playing heavy minutes :dontknow:

They played +2.6 with Deng on the court, and +2.1 with Augustin, despite the fact that Deng has a bigger share of his minutes with worse players than Augustin has. The Bulls didn't have a losing record based on Deng's playing level at that time; and they don't have a winning record based on Augustin's now. Augustin is just an improvement afterall over Teague, but that is hardly something incredible good.

deng's minutes were replaced by supposedly roughly equivalent performers. how does the team get much better adding a supposedly negative performer like augustin and playing him significant minutes? doesn't make any sense at all

Yes, it makes sense, because Augustin did not replace a better player afterall, but the Bulls were simply able to give better players more minutes in average after Deng's depature completely unrelated to Deng or Augustin.

at the very least augustin has been very good offensively.

In terms of the production and efficiency combination he is slightly better than average, but he is not as capable of helping his teammates by organizing a good offense as someone like Hinrich is for example. Still, the team overall has a slightly better offense with Augustin instead of Hinrich, but in no way or shape is that what you get from Augustin considered "very good offensively". Such a statement can only be explained by having incredible low expectations about him.

and you recognize that his man-to-man defense is not bad at least. a negative rating for the guy just has no basis in reality

Again, the Bulls with Hinrich instead of Augustin are playing 9! points better per 100 possessions. About 10 on defense better while about 1 worse on offense. Really, that guy is a net negative overall.

i promise you that dj augustin has had more of a positive impact on the chicago bulls than mike dunleavy has. yet dunleavy's rating is excellent while augustin's is piss poor? c'mon

Yeah, for sure ... the data analysis for ALL minutes gives something very, very different, but you are sure that Dunleavy is less helpful :roll:

Why do you think the method comes up with such different values for each player as your perception? Do you think that there is a hidden variable included, which is just biased towards Augustin? Or can you imagine that you are just missing the "little things" happening on the court?

You are just falling for the "little guy who scores looks impressive among those bigger guys", that's all here. A similar story like last season when people tried to paint Nate Robinson as a much better player than he actually was. But that is in part also based on the biased representation of the game, where the focus is mainly on the ball. But well, I doubt we will agree on that matter ...

[dice]

the bulls were a losing team with deng on the court playing heavy minutes

They played +2.6 with Deng on the court, and +2.1 with Augustin, despite the fact that Deng has a bigger share of his minutes with worse players than Augustin has. The Bulls didn't have a losing record based on Deng's playing level at that time; and they don't have a winning record based on Augustin's now. Augustin is just an improvement afterall over Teague, but that is hardly something incredible good.

augustin has been a well above average point guard offensively. that's plain from basic statistics

and this stat is not supposed to be in comparison to other players like regular +/-. it's supposed to filter that out and provide a stand-alone judgement on a player's impact

deng's minutes were replaced by supposedly roughly equivalent performers. how does the team get much better adding a supposedly negative performer like augustin and playing him significant minutes? doesn't make any sense at all

Yes, it makes sense, because Augustin did not replace a better player afterall, but the Bulls were simply able to give better players more minutes in average after Deng's depature completely unrelated to Deng or Augustin.

the players given more minutes are not significantly better than deng (whose minutes they took), though! how are the bulls so much better? augustin got some of deng's minutes too, while supposedly he's way worse than deng

at the very least augustin has been very good offensively.

In terms of the production and efficiency combination he is slightly better than average, but he is not as capable of helping his teammates by organizing a good offense as someone like Hinrich is for example. Still, the team overall has a slightly better offense with Augustin instead of Hinrich, but in no way or shape is that what you get from Augustin considered "very good offensively". Such a statement can only be explained by having incredible low expectations about him.

the only way you can not recognize that augustin has been very good offensively is by blindly relying on some statistic that tells you that

kirk hinrich's offense consists of dribbling around a lot. he's not providing anything noteworthy in terms of "organizing a good offense". meanwhile augustine has been a WAY better scorer as well as providing more assists

Why do you think the method comes up with such different values for each player as your perception? Do you think that there is a hidden variable included, which is just biased towards Augustin? Or can you imagine that you are just missing the "little things" happening on the court?

i think the stat is pretty accurate overall. just not in this particular case. either it's random variation or a huge outlier

You are just falling for the "little guy who scores looks impressive among those bigger guys"

uh, no. no i'm not. not even close

A similar story like last season when people tried to paint Nate Robinson as a much better than he actually was

i didn't. augustin has been far more consistent and it's apparent in the statistics

augustin: 17p 6a 2to on 57.6% ts (excellent)
nate: 19p 5a 2to on 54.0% ts (average)

[jinxed]

Dice

augustin: 17p 6a 2to on 57.6% ts (excellent)

Where are you getting those numbers? For the year (which is what RPM rates, so his time in Toronto is included is ..12.8 ppg and 4 assists. That is not well above average. That is average.

And has been explained to you over and over, it's on DEFENSE that he is terrible.

EDIT: Sorry I see that those are Per 36 numbers while playing for Chicago.

[mysticbb]

augustin has been a well above average point guard offensively.

Yeah ... wait ... no, he isn't. But we are running in circles here. Your statements aren't making any sense to me, while mine likely make not much sense to you. Thus, I don't see the point of such conversation.