#25 Highest Peak of All Time (McGrady '03 wins)

[Doctor MJ]

I'm certainly not willing to put Malone in the Melo zone, but the data is telling me that there's a major disconnect here somewhere.

Of course, were assuming that the data is perfectly unbiased, infallible, and not subject to scrutiny.

Even putting Karl Malone, one of the better offensive PFs of his era who actually could be asked to play some defense, in the same sentence with Carmelo Anthony is incredulous. With all due respect. And I don't even like this guy; I hated watching him during MJ second threepeat and the media fawning over his MVP that should've went to MJ.

This isn't "preconceived notions" talking either.

Do you have any idea how many Melo supporters say the EXACT same things you are? I'm not trying to say that means you personally are incapable of analysis here Westside by any means, but the entire basis of doing legit objective analysis is not let your notions of what is incredulous dismiss data. If you're doing that, then you are just using preconceived notions.

And if you're not doing that, then we're really on the same page. Am I saying Malone was a mediocre player? Not at all, however we now have a pretty significant trend going here which requires evaluation.

[Doctor MJ]

Players who don't play enough minutes don't get assigned ratings.

He has put guys with like 100 minutes before on the list.

Kenyon Martin is on the list you linked and he was in the 2000 draft aka the 2001 season.

Mea culpa as I edited above. I don't understand what's up with that particular set of data.

[An Unbiased Fan]

RAPM is just a number. People are putting way too much stock in a lineup evaluation metric, when it overrides so much other evidence of Malone's impact.

Eh, the problem with this approach is that for the most part it reads to me as saying, "Who knows what it means?"

I understand being cautious about accepting any bold interpretation, and I most certainly understand putting a caveat in the interpretation of the data based on a systematic issue you've observe.

But you aren't doing that, are you? Aren't you just saying, "This is too far away from my preconceived notions, I'm going to ignore what it says about the player in question."?

What we know with a certainty is that for Malone's last 5 years of his "incredible longevity" is that the lineups weren't doing much better with him than without him. Does that mean he wasn't a great player in this time period necessarily? No, that's the part where analysis needs to be done in order to have a conclusion, but how on Earth anyone cannot be incredibly curious about such odd lineup results is beyond me.

Something very unexpected has been noted, this is either a reason to kick your analysis skills into gear or its time to reveal that you really don't have analysis skills, you just have a set of beliefs of sundry origins.

Well to be fair, I have been criticizing RAPM from the start, and it has more to do with lineup makeup, team systems, roster depth, player roles, etc.

If a player has great utility on a team like Manu for the Spurs, then they will have impressive RAPM numbers. That's useful if we're analyzing how the 2005 Spurs would have done without Manu's skillset in that system, but it doesn't tell us how good Manu was.

I always go back to one of the obvious problems with RAPM. Any star putting up significant minutes, will have very unique times when they are off court. Malone was still playing 35+ mpg, so we're taking the 10-13 mpg he's not on court, which is usually 2nd unit vs 2nd unit, and then ascribing the difference to his "impact". What is the real value of those 35 mpg vs 13 mpg? Malone has to anchor the team battling against other starters, and gets diminished because when he's on the bench, Sloan has the 2nd unit guys playing hard against the other benches. Clearly, a guy like Stock will have big numbers considering his role on Utah, and the lack of anyone to take his place. We saw the same dynamic in PHX, where Barbosas was the backup.

I really think "on court" should be what's looked at, not the off court stuff which is over-analyzed.

[MisterWestside]

Do you have any idea how many Melo supporters say the EXACT same things you are? I'm not trying to say that means you personally are incapable of analysis here Westside by any means, but the entire basis of doing legit objective analysis is not let your notions of what is incredulous dismiss data. If you're doing that, then you are just using preconceived notions.

Oh? Name a Melo supporter who thinks that Melo is on Malone's level.

And once again, no; this isn't "preconceived". I've watched the guy play, thank you. And there's plenty of other evidence that suggests Malone > Melo, and not by some close margin.

I used RAPM as much as other metrics/game tape to put together my top 10 from just this past season. So don't claim that I'm "throwing out" the data. I see it with regards to Karl. I'm simply considering other (just as) important sources instead of putting everything under the context of RAPM.

[Doctor MJ]

You aren't alone on your Dwight Howard thoughts.

As for Karl, a few points:

(1) Just because they would be voted back to back doesn't mean the people votING have them back-to-back. I certainly don't.

(2) I thought the RAPM data was for 2001. (This link certainly is. For example, there is no Jeff Hornacek: http://stats-for-the-nba.appspot.com/PBP/2000.html)

(3) That means Karl's RAPM in 01 was 2.7. This is his clearly the end of his prime (last all-nba team, last POM, shooting falls off dramatically as well as rebounding in following year.) Which means if RAPM is "accurate, in 01:"

Shaq
Duncan
Stockton
Marion
Dirk
Kobe
Ray
Robinson
A. Miller
McKie
B. Davis
The Wallaces
Carter
S. Williams
D. Anderson
Outlaw
PJ Brown
Kukoc
Webber
J. O'Neal
Mutombo
Iverson
Frencis
K. Thomas
McGrady
McDyess
Ward

we are all "better." Now I have no idea what the error is on this stat, although it has to be relevant looking at single-season APM error and looking at the movement of the stat based on 15-game PS updates. (eg To see a similar 60% increase in a number that 2012 LBJ saw, after 80 games at 20 ppg, it would take averaging 96 ppg in the next 15g to increase the scoring by 60%.) I assume then a small change bringing 01 Karl to the forefront would impress you instead of bringing you pause. How does the prior/non-prior difference play a role in this as well?

(4) Note that Kobe Bryant showed the exact same RAPM in 2011 (haven't done the STD check) as Karl in 01. Let's say he slowly tapered off after that and that was all the RAPM data we had on him...should we re-think his career/peak bc of such a season?

(5) Stockton. I find any +/- data useful, but I feel for some strange reason people have defaulted heavily to single-season RAPM as some sort of reference point, as if ridge-regression handles most of the many issues. But the big issue is Stockton.

How can I think that there isn't some co-variance that RAPM isn't "noticing" when Stockton is being considered significantly better than Malone at this point in their careers? Without delving into the lineup patterns to unearth something like this, the implication here is that the Jazz would be far better off with a team of:

Polynice
D. Marshall
Russell
Starks
Stockton

than

Polynice
K. Malone
Russell
Starks
Vaughn

This would be one of those WP predictions that is as unlikely as Charlotte winning 40 games. Anything's possible, I suppose. Of course just a quick glance at the minutes distribution of the Jazz team reveals 9-men over 16 mpg...with Karl really the only regular, which obviously opens up the door for some steady lineup pattern/subs that let RAPM "think" he's the one holding the team down. The entire 01 Jazz RAPM:

Stockton 5.6
Karl 2.7
Russell 2.0
D. Marshall 1.6
Manning 0.2
Polynice -0.2
Starks -0.5
Ostertag -0.7
Vaughn -3.4

RAPM "thinks" that Vaughn is pulling down lineup differentials while Stockton is inflating them. My guess is, without looking at opponent strength, that this has to do with Stockton being a part of the 10-best Jazz lineups (+12.9 per 48 total) while Vaughn is part of the 8-worst lineups (-16.9 per 48 total). Karl is on the floor for those 4-worst lineups (and 9 of the top-10), so solving for the Karl Malone variable is affected by this as well.

(6) Then there's the "Ginobili issue" of him playing 28, hand-picked perfect minutes to exert himself because he's 38. If it were Karl playing 28 mpg in this form with a crappy backup and Stockton asked to play 36 mpg with the team while Utah had better guards around him, what do you think the number would look like?

In summary, I'm really surprised to see you re-thinking Karl based on this data point since it didn't give me much pause. Perhaps you could expound...

I'll throw you something since you say you aren't impressed with any of the guys being discussed. The 98 Jazz were ~2 SRS team to start the year without John Stockton. They played a brutal schedule, and posted a +1.7 ORtg and -0.9 DRtg in those first 18 games. But they were also a veteran team, and as such ramped up as the season progressed (Stockton's return is a confound in that equation -- for reference the SRS over the next 20g was 5.2, over the following 20 8.4, and over the last 20+ 7.0 (with ORtg's around +11 for the second half of the year).

So a "non-warmed up" Jazz team of Eisley, Hornacek, Keefe, Karl and Foster looked like a 49-win team. I've been impressed by Barkley for a number of voting slots because I don't think many (any?) guys on the board left could take a Dawkins-Hawkins-Mahorn-Gminski team to ~54 wins...and that's pretty clearly a better team in my mind. Food for thought.

Wanted to respond to this from the previous thread:

1) Point taken. I shouldn't be thinking along those lines.

2) Yeah, something weird there.

3) It's interesting, and yeah when you take out '99-00 from the equation, the general trend is less disconcerting and significantly more worthy of caution.

However, it's well and good to say we don't know what the errors are like here, but a series of independent RAPM years combined showing similar trends is going to have very low uncertainty. I like multi-year APM studies, multi-year RAPM studies are even more reliable, and multiple years taken independently that say similar things is better than that for certainty. Bottom line: This is starting to become a big deal.

Agreed that it's not as big of a deal without '99-00, but a lot of people weren't impressed even with '99-00. To them I just have to ask: At what point do you start paying attention?

Obviously there are some people who simply aren't going to have their opinion swayed for reasons that have nothing to do with objective analysis, but there are others who are reasonably cautious and those are the people I'm talking to here. What will you need to see to start saying there's something here?

4) Remember Kobe, advise caution. Well and good. People should remember though, I'm talking about two separate potential conclusions:

A) That old man Malone is killing his impact by pretending he's still in his prime...which is clearly part of Kobe's issue. So that right there is not in contradiction.

B) That we see a trend going right up to Malone's respectability peak and still not seeing an impact which is scary as hell. We didn't see that with Kobe, so this is new territory that I think it's starting to make sense to really start to worry about. One thing to dismiss one or two years at the end of a guy's career, but when a guy is known for a late peak and we start seeing 5 years of continued meh impact, this is a wow moment...

but of course, as mentioned, there is something weird about the '00 data, so (B) is premature to at least some degree.

5) When we see a bunch of 1-year numbers that are independent and agree with each other, it starts to become pretty dang hard to ignore.

Now "ignore" and saying "I'm not ready to equate his used value with his actual ability" are two different things of course, but it sure looks to me like we're seeing a trend that in retrospect really makes you wonder if the Jazz were running an optimal design those last few Malone-Stockton.

Of course in particular there's the part to note that Kirilenko is a bit of a monster who showed +/- signs of being a monster because Malone & Stockton retired. When a star stands in the way of an emerging star, this can cause some really drastic +/- results.

6) Ginobili-ish issues with old Stockton. Absolutely. We just need to remember that Ginobili never stopped Duncan from looking like a Player of the Decade candidate by +/-.

[Doctor MJ]

Well to be fair, I have been criticizing RAPM from the start, and it has more to do with lineup makeup, team systems, roster depth, player roles, etc.

I'm not accusing you of a pro-Malone bias, I'm literally talking about how you've approached +/- from the start.

To be fair to you though: It is possible to have a general objective perspective but simply be too wary of a particular technique to do anything with it. Reluctance in using a table saw does not mean you are unable to use a hacksaw.

If a player has great utility on a team like Manu for the Spurs, then they will have impressive RAPM numbers. That's useful if we're analyzing how the 2005 Spurs would have done without Manu's skillset in that system, but it doesn't tell us how good Manu was.

I always go back to one of the obvious problems with RAPM. Any star putting up significant minutes, will have very unique times when they are off court. Malone was still playing 35+ mpg, so we're taking the 10-13 mpg he's not on court, which is usually 2nd unit vs 2nd unit, and then ascribing the difference to his "impact". What is the real value of those 35 mpg vs 13 mpg? Malone has to anchor the team battling against other starters, and gets diminished because when he's on the bench, Sloan has the 2nd unit guys playing hard against the other benches. Clearly, a guy like Stock will have big numbers considering his role on Utah, and the lack of anyone to take his place. We saw the same dynamic in PHX, where Barbosas was the backup.

I really think "on court" should be what's looked at, not the off court stuff which is over-analyzed.

The "adjusted" form of +/- doesn't actually use "off court" at all. It's basically looking at what was accomplished by everyone when they were on the court, and distributing credit in the simplest way possible. There's more to a players abilities than this of course, but a player who consistently doesn't do great by this metric simply isn't showing signs that his team is benefiting immensely from his on court presence.

I do understand that in practice, this is only able to be done because a player has off court time, and that that off court context isn't going to be identical to the on court context, but that's why you need nuance in your interpretation. Saying "Ignore what it says about Malone, it's just a lineup tool" isn't a nuance, it's closing your eyes.

At a certain point with superstars, the point is this: These are supposed to be players who are FAR better than the other guys around them. If they are having impact like that, we should see it. One thing to say a lesser player has to little impact-to-noise ratio to say anything about him, but a man among boys should not have that issue.

[Doctor MJ]

Do you have any idea how many Melo supporters say the EXACT same things you are? I'm not trying to say that means you personally are incapable of analysis here Westside by any means, but the entire basis of doing legit objective analysis is not let your notions of what is incredulous dismiss data. If you're doing that, then you are just using preconceived notions.

Oh? Name a Melo supporter who thinks that Melo is on Malone's level.

And once again, no; this isn't "preconceived". I've watched the guy play, thank you. And there's plenty of other evidence that suggests Malone > Melo, and not by some close margin.

I used RAPM as much as other metrics/game tape to put together my top 10 from just this past season. So don't claim that I'm "throwing out" the data. I see it with regards to Karl. I'm simply considering other (just as) important sources instead of putting everything under the context of RAPM.

Not what I meant. I've been talking to Melo freaks for more than a half decade who insist he's just a half step behind LeBron, and as good as the Wades and Pauls of the world. When I tell them that in reality is fraction is a tiny fraction of those other guys, they tune out because its ludicrous to them. Their problem isn't really about Melo, it's about them being unable to analyze contradictory evidence without dismissing the stuff that doesn't fit with their schema as 'ludicrous'.

You saying, "I watched the guy", again, that's what the Melo people say. Now, is it possible to tell a lot about a guy based on your basketball watching ability? Sure is. Am I saying you're not better at it than the Melo people? Not at all. However, when you use the same arguments they use, and you see the absurdity of their opinions, you need to recognize that you have to go beyond that if you want to keep from ever getting absurd.

Re: "I used RAPM when it wasn't ludicrous, so don't claim that I'm throwing data out simply because I'm ignoring the ludicrous stuff."

I'm sure I understand where you're really coming from here. You're just advocating caution and it came out a certain way. I'm just making clear that your wording is exactly the wording I see from the people neither of us thinks are doing great objective analysis. If that wording represents how you think, you have the same problem as them.

If that wording doesn't represent how you think, be more clear in your wording.

[An Unbiased Fan]

The "adjusted" form of +/- doesn't actually use "off court" at all. It's basically looking at what was accomplished by everyone when they were on the court, and distributing credit in the simplest way possible. There's more to a players abilities than this of course, but a player who consistently doesn't do great by this metric simply isn't showing signs that his team is benefiting immensely from his on court presence.

I do understand that in practice, this is only able to be done because a player has off court time, and that that off court context isn't going to be identical to the on court context, but that's why you need nuance in your interpretation. Saying "Ignore what it says about Malone, it's just a lineup tool" isn't a nuance, it's closing your eyes.

At a certain point with superstars, the point is this: These are supposed to be players who are FAR better than the other guys around them. If they are having impact like that, we should see it. One thing to say a lesser player has to little impact-to-noise ratio to say anything about him, but a man among boys should not have that issue.

The adjusted form is an attempt to balance the numbers, but that's really not possible. At the end of the day, you're still essentially comparing on/off court numbers, and trying to extrapolate meaning from the 35 mpg vs 13 mpg. Again, there's a big difference in whom Malone is facing in his 35 mpg "on court", vs others in the 13 mpg "on court". It's just baffling to try and distribute credit based on these numbers.

These numbers don't say anything about what Malone did, but rather how Sloan's system operated. The points he scored, the rebounds he grabbed, etc., his performance in the PnR, his defense, all tell us what we really need to know.

I wish we could go back to the 1950's, and tell stat-keepers to follow everything from synergy type stats, to how many box outs a player had, or how many passes led to FTA, etc. But sadly we just don't have that information. So we still have to relay on observation from watching realtime or video, peer review, and box score estimations. RAPM numbers are not special, or even particularly advanced. People keep equating RAPM to impact, when it's a utility trend estimation.

[ElGee]

What will you need to see to start saying there's something here?

4) Remember Kobe, advise caution. Well and good. People should remember though, I'm talking about two separate potential conclusions:

A) That old man Malone is killing his impact by pretending he's still in his prime...which is clearly part of Kobe's issue. So that right there is not in contradiction.

B) That we see a trend going right up to Malone's respectability peak and still not seeing an impact which is scary as hell. We didn't see that with Kobe, so this is new territory that I think it's starting to make sense to really start to worry about. One thing to dismiss one or two years at the end of a guy's career, but when a guy is known for a late peak and we start seeing 5 years of continued meh impact, this is a wow moment...

I'd need AT LEAST 2 years of Malone's prime. Right now, I have one.

Also, I believe the 2000 link is a different iteration of the 2001 season. I'm prett sure of it (look at the remarkable similarity in the results). So everyone is looking at data that is reinforcing itself from the same year.

Stockton RAPM (Malone parens)
2001 6.1/5.6 (Karl +2.7/+2.1)
2002 3.0 (Karl -0.9)
2003 1.2 (Karl +0.5)

For your B point to be the case here we'd need years from the player's prime. I don't see how one-year of RAPM is now suddenly cutting it for mind-changing data. And that's what we're talking about here -- something with enough puissance to matter.

There's also the practical application of these numbers. Do they go down because Malone's asked to do too much? What's the error in the single-season? Is 2001 all that really matters here? Why are Stockton's defensive numbers so high -- doesn't this suggest a potential problem? (Since when is RAPM not subject to problems?) And perhaps most importantly, why does Stockton have such a huge 2001 season but then drops off in a season that was widely considered a rejuvenation of sorts? Nearly the same box metrics with a mild increase in minutes, and more than double the 20-point games from the previous year.

[therealbig3]

Unless I'm mistaken, RAPM has nothing to do with time spent on the bench, right? It's all about scoring margin while that player is on the court, from what I understand. And it tries to estimate how much of that scoring margin should be credited to an individual player.

Now, I guess theoretically, despite the adjustments, it could still be skewed based on how much of certain lineups a team puts out there around their star. But if a certain player consistently shows up to be unimpressive, or consistently shows up to be very impressive, I think that's valuable information.

[MisterWestside]

Not what I meant. I've been talking to Melo freaks for more than a half decade who insist he's just a half step behind LeBron, and as good as the Wades and Pauls of the world. When I tell them that in reality is fraction is a tiny fraction of those other guys, they tune out because its ludicrous to them. Their problem isn't really about Melo, it's about them being unable to analyze contradictory evidence without dismissing the stuff that doesn't fit with their schema as 'ludicrous'.

Okay, but can any of those Melo freaks (and we'll assume that they're in the casual fanbase that eats up things like awards, titles/team results and such) even point to anything like 1) 2 league MVPs 2) two Finals appearances/three WCF appearances 3) multiple all-D honors? Even by using the flawed "accolade" comparison Malone is way ahead of Melo here (heck, even his own peers at the forward position in the 90s). At least the causal Malone freaks can point to some stuff to support their argument here, and not just make rabid claims of greatness like Melo freaks do.

You saying, "I watched the guy", again, that's what the Melo people say. Now, is it possible to tell a lot about a guy based on your basketball watching ability? Sure is. Am I saying you're not better at it than the Melo people? Not at all. However, when you use the same arguments they use, and you see the absurdity of their opinions, you need to recognize that you have to go beyond that if you want to keep from ever getting absurd.

I only stated that for emphasis, because even in this age of basketball metrics game-watching is still invaluable, and I've observed stark differences between Melo and Malone's game. I'm not using it as the sole source of information, of course.

Re: "I used RAPM when it wasn't ludicrous, so don't claim that I'm throwing data out simply because I'm ignoring the ludicrous stuff."

Right Well again, I'm not throwing out the RAPM data. I'm simply saying "This one data point says that, but we have plenty of other information that says this. The scale weighs in favor of the other information here."

ElGee makes some great points about this with his post on the previous page and his above post ("Since when is RAPM not subject to problems?" ). Also, re: 2001 Malone vs. old man 2011 Kobe, Karl's 2001 usg/ortg numbers adjusted for 2011 is 30/117. That's superstar production (in fact, this put him in elite company http://bkref.com/tiny/YxJfZ), and he was playing better defense as well. Malone didn't play like "I'll chuck up shots like I'm young Malone but brick them instead" until '02.

[An Unbiased Fan]

Unless I'm mistaken, RAPM has nothing to do with time spent on the bench, right? It's all about scoring margin while that player is on the court, from what I understand. And it tries to estimate how much of that scoring margin should be credited to an individual player.

Now, I guess theoretically, despite the adjustments, it could still be skewed based on how much of certain lineups a team puts out there around their star. But if a certain player consistently shows up to be unimpressive, or consistently shows up to be very impressive, I think that's valuable information.

It really has no choice but to factor in off court team performance, even if it's done in an indirect manner. The numbers being evaluated are a result of the 35 mpg vs 13 mpg dynamic, there is simply no way around this.

So when we say a player was "unimpressive" by this metric, we really need to examine what the metric is actually saying. What value is there in the 35 mpg that Malone gave to the jazz against 1st units vs the other 13 mpg against 2nd units, because these are factored into the overall estimation.

[C-izMe]

I think your putting too much value into RAPM Doc. Just imagine you knew 8 people. You wanted to know the fastest way to get somewhere and naturally some people were more reliable than others. 7 of them tell you to go one general direction (with slight varience) but the one you trust most tells you to go a completely different way. Will you competent disregard the 7 to go with the 1?

[tsherkin]

I think your putting too much value into RAPM Doc. Just imagine you knew 8 people. You wanted to know the fastest way to get somewhere and naturally some people were more reliable than others. 7 of them tell you to go one general direction (with slight varience) but the one you trust most tells you to go a completely different way. Will you competent disregard the 7 to go with the 1?

You should, C-iz. You shouldn't necessarily trust it if there isn't anything behind the suggestion, but appealing to the majority if a fallacy. The 1 disagrees, so you look to see if there is substance to the notion. You have to challenge accepted thoughts from time to time, in case you (which I use in the general sense) are actually wrong.

EDIT: I'm not saying that RAPM is correct in this case, necessarily, just that sometimes if you see something that is divergent from what you believe, you should examine WHY that is the case.

[C-izMe]

I think your putting too much value into RAPM Doc. Just imagine you knew 8 people. You wanted to know the fastest way to get somewhere and naturally some people were more reliable than others. 7 of them tell you to go one general direction (with slight varience) but the one you trust most tells you to go a completely different way. Will you competent disregard the 7 to go with the 1?

You should, C-iz. You shouldn't necessarily trust it if there isn't anything behind the suggestion, but appealing to the majority if a fallacy. The 1 disagrees, so you look to see if there is substance to the notion. You have to challenge accepted thoughts from time to time, in case you (which I use in the general sense) are actually wrong.

EDIT: I'm not saying that RAPM is correct in this case, necessarily, just that sometimes if you see something that is divergent from what you believe, you should examine WHY that is the case.

Good point but I think it's wrong to try see what is wrong with the other numbers instead of trying to see what's wrong with the minority. We should be looking more into WHY the RAPM numbers for Malone are low instead of assuming they mean something was wrong with him.

[therealbig3]

For the Malone supporters (mainly ElGee), why 98 over a year like 95?

I read ronnymac's posts in the RPOY thread, and he seems to think that 94-96 is probably when Malone peaked. I'd probably go with 95 out of those years, just by glancing at his production in the RS and PS. I understand Malone's insane physical fitness, but he was still an effective passer and jump shooter in those years, and you'd still assume that he was a tad quicker and faster at 30-32 than 34.

[ElGee]

I think your putting too much value into RAPM Doc. Just imagine you knew 8 people. You wanted to know the fastest way to get somewhere and naturally some people were more reliable than others. 7 of them tell you to go one general direction (with slight varience) but the one you trust most tells you to go a completely different way. Will you competent disregard the 7 to go with the 1?

You should, C-iz. You shouldn't necessarily trust it if there isn't anything behind the suggestion, but appealing to the majority if a fallacy. The 1 disagrees, so you look to see if there is substance to the notion. You have to challenge accepted thoughts from time to time, in case you (which I use in the general sense) are actually wrong.

EDIT: I'm not saying that RAPM is correct in this case, necessarily, just that sometimes if you see something that is divergent from what you believe, you should examine WHY that is the case.

Good point but I think it's wrong to try see what is wrong with the other numbers instead of trying to see what's wrong with the minority. We should be looking more into WHY the RAPM numbers for Malone are low instead of assuming they mean something was wrong with him.

To be clear, I don't quite agree with this line of thinking. In your analogy, the answer is "it depends on how much you trust the one friend." If it's a question of a dynamic traffic calculation, and my 7 friends say take one route and my other friend, Lt. Data, says take another, I'm listening to Data. Every time. If my dissenting friend is Apple Maps, I'm probably thinking "there goes that Apple Maps malfunctioning again." RAPM falls somewhere in between those 2, but in this case we have the issue of its "advice" only providing a single street name (the year 2001 for Malone), not a complete route.

[C-izMe]

Well I guess I trust in/out, PER, WS, and those other numbers more than most. It's just really odd for me that people immediately started doubting Malone on (possibly inaccurate) post prime RAPM numbers. I really like the Kobe analogy here when it comes to how we should look at these numbers.

[tsherkin]

Good point but I think it's wrong to try see what is wrong with the other numbers instead of trying to see what's wrong with the minority. We should be looking more into WHY the RAPM numbers for Malone are low instead of assuming they mean something was wrong with him.

I think we're saying the same thing here; I'm not saying that you should assume the RAPM numbers are correct. I'm saying you should see what proof there is that agrees with what RAPM is saying, as opposed to discarding RAPM because it doesn't agree with everything else, that's all.

[Doctor MJ]

It really has no choice but to factor in off court team performance, even if it's done in an indirect manner. The numbers being evaluated are a result of the 35 mpg vs 13 mpg dynamic, there is simply no way around this.

So when we say a player was "unimpressive" by this metric, we really need to examine what the metric is actually saying. What value is there in the 35 mpg that Malone gave to the jazz against 1st units vs the other 13 mpg against 2nd units, because these are factored into the overall estimation.

Right, so to your mind when you say you want a focus on "on court" as opposed to "off court" you don't mean whether they literally are only looking at on court time, you mean you don't want them to use +/- data at all, which means you want to have no analysis done on the actual scoreboard. "Just give me the arbitary stats that we know the players can manipulate to negative effect please."

Re: actually saying. When I talk about nuance, I'm talking about not trying to compare Karl Malone's +/- to those of some scrub because we know they are doing very different things. If you are a man among boys though, we should be seeing glaring signs that things are better when you're on the court as opposed to eating an ice cream cone.