Questions on PI RPM / Garnett

[McBubbles]

Was finally taking a look at Engelman's PI RAPM. Namely the PI DRAPM and I have a few questions.

1. Do these numbers measure impact full stop, or do they measure impact relative to role? I'm pretty sure I know the answer to this but I'm just checking. I ask because Shane Battiers PI DRAPM peaks at 4.51, whilst Duncan's and Ben Wallace's peak at 5.4 and 4.94 respectively, while orbiting between 4.50-5.00 most other years for the former and 4.00-4.50 for the latter. Does this mean that Battier at his defensive peak had a prime ATG rim protector level impact? Or does it mean that Battier was just as dominant at perimeter defence at his peak as Wallace and Duncan were at interior defence in their prime, but that Battiers impact wasn't similar?

2. Why on earth are KG's DRAPM numbers so ridiculously **** high after he leaves Minnesota? They're a massive outlier relative to other players and a massive outlier relative to his prior self. His best from 02-07 was 4.51. His worst from 08-14 was 5.07 Is it the effect of him utilizing a revolutionary Thibbs defensive scheme? And if so you'd think that effect would have lasted for a shorter period of time than it did, especially not up until 2014.

Is it a result of him being put closer to the interior on defence instead of him being "wasted" further out on the perimeter? Or should it be taken at face value with KG actually just being a mega GOAT defender?

Also why do his DRAPM numbers peak at an absolutely monsterous 6.79 in 2013 when he's 36 years old? So many questions.

[Morb]

It's just another weak defence stat.

Sent from my Nokia 5.1 Plus using RealGM mobile app

[Jaivl]

Was finally taking a look at Engelman's PI RAPM. Namely the PI DRAPM and I have a few questions.

1. Do these numbers measure impact full stop, or do they measure impact relative to role? I'm pretty sure I know the answer to this but I'm just checking. I ask because Shane Battiers PI DRAPM peaks at 4.51, whilst Duncan's and Ben Wallace's peak at 5.4 and 4.94 respectively, while orbiting between 4.50-5.00 most other years for the former and 4.00-4.50 for the latter. Does this mean that Battier at his defensive peak had a prime ATG rim protector level impact? Or does it mean that Battier was just as dominant at perimeter defence at his peak as Wallace and Duncan were at interior defence in their prime, but that Battiers impact wasn't similar?

2. Why on earth are KG's DRAPM numbers so ridiculously **** high after he leaves Minnesota? They're a massive outlier relative to other players and a massive outlier relative to his prior self. His best from 02-07 was 4.51. His worst from 08-14 was 5.07 Is it the effect of him utilizing a revolutionary Thibbs defensive scheme? And if so you'd think that effect would have lasted for a shorter period of time than it did, especially not up until 2014.

Is it a result of him being put closer to the interior on defence instead of him being "wasted" further out on the perimeter? Or should it be taken at face value with KG actually just being a mega GOAT defender?

Also why do his DRAPM numbers peak at an absolutely monsterous 6.79 in 2013 when he's 36 years old? So many questions.

- Battier is an excellent defender, obviously not as good as Duncan or Wallace, but he left a huge defensive statistical imprint in some good Memphis teams (top 5 defense out of thin air in 2005, #2 in 2006 with healthy Gasol and Eddie Jones, LAST when he leaves in 2007 -although Gasol and Jones got injured as well-).

- Also, remember RAPM is per-possesion, and Garnett is still a DPOY caliber defender when he's on the court up to 2013, with huge point differentials to his credit, and as he does not have that big of an offensive role or scoring ability anymore, there's some merit for his defensive results being "better".

But anyway, I would focus on the approximate rank, and not really the number. There's some math stuff in there that makes the number not really 1-to-1 with "impact".

[Doctor MJ]

Good questions:

1. They measure a correlation between a player’s presence in the court and team success within the category being measure. If you’re focusing on DRtg, then all that matters is the team’s opponents scoreboard result (divided by number of possessions).

Anything we do calling this “impact” is connecting that correlation as if it is causation, and this is something where there are always expected to be confounding factors. Nevertheless, the more data we get, the less reasonable it is to ignore what that data says.

Please feel free to ask more questions drilling down here.

2. In Boston, Kevin Garnett focused on defense, and did so within Tom Thibodeau’s (then) cutting edge defense. (Definitely worth googling for some analysis of what that system looked like.)

Specifically what made Garnett really become a transcendent defense force there was the combination of

a - Garnett’s ideal body for a defensive player (based in the Bill Russell long and lithe archetype)

b - Garnett being focused more on defense rather than carrying the offensive burden he did in Minny.

c - Other player’s doing their darnedest to carry out their part in a way that made people use the expression “5 guys on 1 string”. Note: That in earlier eras this could have been called illegal defense…though that has nothing to do with why Minny never played like this.

d - Garnett taking on a “middle linebacker” or “defensive quarterback” role where he would read what the offense was doing from his position in the interior and tell his teammates where to go.

To do this and do this well, you need an ultra-high defensive BBIQ with vocabulary. You need to not just be able to read the offense instantly, but no how each teammate should optimally respond and what language to use to communicate successfully to them so that they can respond before the offense can successfully exploit them with the next move.

Incidentally, since Garnett, we’ve seen two other players who can truly claim to be like this (that I’m aware of): Marc Gasol and Draymond Green.

All 3 would win DPOY, and I would consider Green to be the best defender of the post-KG era.

e - it’s worth noting that the Warriors at this point have spent years expanding this style of defensive communication to the point now where it’s not just Green or (also exceptional) Andre Iguodala that do this, but that the entire defense is looking to communicate to the point of attack defenders all the time with Steph Curry becoming quite proficient at it himself.

This is part of what’s allowing a guy like Payton II to be not just a terror in defense, but a guy who doesn’t get his aggression used against him the way he normally would.


Sent from my iPhone using Tapatalk

[feyki]

2010 Celtics - w/KG in 69G 3,5 SRS, without KG in 12G 0 SRS,
2011 Celtics - w/KG in 71G 4,5 SRS, without KG in 10G 5,5 SRS,
2013 Celtics - w/KG in 68G -0,5 SRS, without KG in 13G -2,5 SRS

. Don't know about 2011 and also toooo limited sample size, but there was 2-3 SRS difference between the w/KG and the without KG teams. Assuming his offensive impact as positive(somewhere around 0 to 1 SRS), don't think his defensive impact on the dpoy level, Wolves KG was way better defensive impact player than any version of Celt Garnett, except 2008.

[Jaivl]

2010 Celtics - w/KG in 69G 3,5 SRS, without KG in 12G 0 SRS,
2011 Celtics - w/KG in 71G 4,5 SRS, without KG in 10G 5,5 SRS,
2013 Celtics - w/KG in 68G -0,5 SRS, without KG in 13G -2,5 SRS

. Don't know about 2011 and also toooo limited sample size, but there was 2-3 SRS difference between the w/KG and the without KG teams. Assuming his offensive impact as positive(somewhere around 0 to 1 SRS), don't think his defensive impact on the dpoy level, Wolves KG was way better defensive impact player than any version of Celt Garnett, except 2008.

Are you being deliberately sloppy? - I have 2013's numbers on hand and you round -0.21 to -0.5 (Garnett's ON sample) but then round -2.78 to -2.5 (Garnett's OFF sample), so Garnett's with/without differential becomes +2 instead of +2.6. That's a 30% deviation lol.

Also, using again 2013 for practicality, considering Boston loses 6.something points offensively on games with KG but gains 9 points on defense, and his offensive +/- is -3, and he doesn't have that many playmaking duties anymore, and he scores under 15 points per game at exactly league average... yeah, I think a +1 offensive impact is a big stretch.

EDIT: just ran 2011 numbers and you shaved another half a point (you say -1 when it's actually -0.5, defense improves nearly 5.5 points and offense worsens the same amount, the half a point difference is due to the strength of schedule, although Boston played at home during most of that KG-less stretch - 8 games out of 11, maybe that balances it out). Can't really rely on your calcs.

[trex_8063]

2010 Celtics - w/KG in 69G 3,5 SRS, without KG in 12G 0 SRS,
2011 Celtics - w/KG in 71G 4,5 SRS, without KG in 10G 5,5 SRS,
2013 Celtics - w/KG in 68G -0,5 SRS, without KG in 13G -2,5 SRS

. Don't know about 2011 and also toooo limited sample size, but there was 2-3 SRS difference between the w/KG and the without KG teams. Assuming his offensive impact as positive(somewhere around 0 to 1 SRS), don't think his defensive impact on the dpoy level, Wolves KG was way better defensive impact player than any version of Celt Garnett, except 2008.

Are you being deliberately sloppy? - I have 2013's numbers on hand and you round -0.21 to -0.5 (Garnett's ON sample) but then round -2.78 to -2.5 (Garnett's OFF sample), so Garnett's with/without differential becomes +2 instead of +2.6. That's a 30% deviation lol.

Also, using again 2013 for practicality, considering Boston loses 6.something points offensively on games with KG but gains 9 points on defense, and his offensive +/- is -3, and he doesn't have that many playmaking duties anymore, and he scores under 15 points per game at exactly league average... yeah, I think a +1 offensive impact is a big stretch.

EDIT: just ran 2011 numbers and you shaved another half a point (you say -1 when it's actually -0.5, defense improves nearly 5.5 points and offense worsens the same amount, the half a point difference is due to the strength of schedule, although Boston played at home during most of that KG-less stretch - 8 games out of 11, maybe that balances it out). Can't really rely on your calcs.

I'd also note that prior poster stated DPOY-level defensive impact when on the court.........he's not "on the court" every minute of the games he played in. He was only averaging ~30-31 mpg in those years, so that's 17-18 mpg where he's NOT on the court in those games included in the "with" sample.
Thus we should note the difference in team DRtg when ON vs when OFF [not just with/without games]:
-4.6 better in '10
-6.3 better in '11
An atypically low -1.5 better in '12
Then a whopping -9.0 better in '13

[kcktiny]

2. In Boston, Kevin Garnett focused on defense, and did so within Tom Thibodeau’s (then) cutting edge defense.

Specifically what made Garnett really become a transcendent defense force there was...

Are you seriously trying to suggest that Kevin Garnett was a better defender in Boston than he was in Minnesota, that he only then focused on defense, just because DRAPM says so?

Garnett did not play in Boston until age 31. He played 12 seasons in Minnesota before ever getting to Boston, was named to the All-Defensive 1st Team 6 straight seasons (1990-00 to 2004-05) as a Timberwolve, was All-Defensive 2nd team twice, and was also 2nd in DPOY voting twice and 3rd another time.

He was a great defender long before he ever got to Boston. Maybe what you should really be asking is why this concocted calculation didn't have him rated higher as a defender when in Minnesota.

[Doctor MJ]

2. In Boston, Kevin Garnett focused on defense, and did so within Tom Thibodeau’s (then) cutting edge defense.

Specifically what made Garnett really become a transcendent defense force there was...

Are you seriously trying to suggest that Kevin Garnett was a better defender in Boston than he was in Minnesota, that he only then focused on defense, just because DRAPM says so?

Garnett did not play in Boston until age 31. He played 12 seasons in Minnesota before ever getting to Boston, was named to the All-Defensive 1st Team 6 straight seasons (1990-00 to 2004-05) as a Timberwolve, was All-Defensive 2nd team twice, and was also 2nd in DPOY voting twice and 3rd another time.

He was a great defender long before he ever got to Boston. Maybe what you should really be asking is why this concocted calculation didn't have him rated higher as a defender when in Minnesota.

Interesting where you ended here. Something you should understand is that many of us here are people who began using stats like RAPM because we understood that the stat was doing, and thus saw the clear validity in the stat.

That doesn’t mean the stat is a perfect measure of goodness or value or impact, but it does mean something.

In this case what it means is that Garnett’s presence in Boston correlated more with team defensive effectiveness after normalizing for major confounding variables, than it did in Minnesota.

If you have any questions about stuff like that, you should ask, but I’d hope that you’d agree that a stat like that has basketball meaning (if implemented well).

You need to also understand what it felt like doing this analysis in real-time. Forget about the stats. Those Celtics shocked the world. People nowadays talk as if their success was expected because they were a “super team”, but this couldn’t be further from the truth.

Those Celtics drastically overperformed expectations, and did so not because their expected strength (offense) was better than than expected, but because their expected weakness (defense) instead dominated the league and quickly set the new standard for state-of-the-art defensive strategy.

And that defense did what it did because of what happened with Garnett on the court, playing the role I’ve explained in my earlier post.

Does this mean Garnett was physically more capable of a defender at his Boston age compared to his Minny age? No.

But did it mean he had greater defensive impact in Boston than Minny? Yes, I believe he did.

Hence the question isn’t so much how a stat could be so wrong, but what Garnett could have achieved if he were better architected around in Minny. And it’s not just the defense - the defense is just the part we got to see him with superior strategy. Offensively he likely could have been a monster 3-point shooter and a monster helio.

And of course he’s not the only one for whom this is true. It’s how such evolution works. As optimization occurs in a field, it allows better utilization of the talent in the field. We know this is true, we just haven’t traditionally been able to point to data when we saw it in action.

But now we can, with tools like this, and it looks the example above.

The discrepant data that made you think the stat was wrong is indeed an indication that something was wrong about basketball - relative to later eras.


Sent from my iPhone using Tapatalk

[kcktiny]

many of us here are people who began using stats like RAPM because we understood that the stat was doing, and thus saw the clear validity in the stat

If there is one thing any derivative of plus/minus is not is clear.

That doesn’t mean the stat is a perfect measure of goodness or value or impact

Correct. It is none of that.

but it does mean something

Like what?

In this case what it means is that Garnett’s presence in Boston correlated more with team defensive effectiveness after normalizing for major confounding variables, than it did in Minnesota.

Could you possibly be more evasive? Confounding variables? What confounding variables?

Those Celtics shocked the world.

The San Antonio Spurs were even better defensively the half decade prior to Garnett going to Boston than Boston was with Garnett. Did they shock the world too?

And that defense did what it did because of what happened with Garnett on the court

Well then what did Garnett do on the court for Minnesota such that he was All-Defensive 1st Team 6 years running and 2nd/3rd in DPOY 3 of those seasons? Evidently the people that saw him play and voted for those awards thought he was one of the very best defenders in the league - for a long time.

playing the role I’ve explained in my earlier post.

What role did he play in Minnesota such that those that saw him play were so impressed by his defense? Can you explain that?

But did it mean he had greater defensive impact in Boston than Minny? Yes

False.

Hence the question isn’t so much how a stat could be so wrong

On the contrary, this is exactly the question.

Right now the ESPN Real Plus/Minus shows Nikola Jokic as the best defender in the league with a DRPM of 6.26. Is this or is this not a derivative of Englemann's RAPM?

Yet if you go to the stats webpages at NBA.com you'll find that as of today of the 14 players in the league to have faced 400+ FGAs from within 10 feet of the basket, Jokic is allowing a 58.0% FG% (344/593), but best in the league is Jarrett Allen allowing just a 47.2% FG% (218/462). That's almost 11% lower than Jokic. Jokic's 58.0% FG% allowed is the 3rd highest/worst among those 14 players.

And - not surprising - right now on defense Denver as a team allows the 2nd highest/worst 2pt FG% at 55.0% among all 30 teams in the league. Due in large part to Jokic, who has faced on defense almost twice as many shots as any other Nuggets player (593, next most faced is 303 by Aaron Gordon).

Jokic is an excellent defensive rebounder, but defensive rebounding is after the fact. Stopping shots from going in is key to a great defense, and the NBA.com numbers certainly show Jokic's FG% allowed is high/poor for a C.

So how does that square with the ESPN RPM? Is this where those major confounding variables come in? How can it rate Jokic as best in DRPM when he allows one of the highest/worst FG%s within 10' of the basket?

And it’s not just the defense

But defense is what we are discussing here.

The discrepant data that made you think the stat was wrong is indeed an indication that something was wrong about basketball - relative to later eras.

Did you even watch Garnett play during his tenure with Minnesota?

[McBubbles]

many of us here are people who began using stats like RAPM because we understood that the stat was doing, and thus saw the clear validity in the stat

If there is one thing any derivative of plus/minus is not is clear.

That doesn’t mean the stat is a perfect measure of goodness or value or impact

Correct. It is none of that.

but it does mean something

Like what?

In this case what it means is that Garnett’s presence in Boston correlated more with team defensive effectiveness after normalizing for major confounding variables, than it did in Minnesota.

Could you possibly be more evasive? Confounding variables? What confounding variables?

Those Celtics shocked the world.

The San Antonio Spurs were even better defensively the half decade prior to Garnett going to Boston than Boston was with Garnett. Did they shock the world too?

And that defense did what it did because of what happened with Garnett on the court

Well then what did Garnett do on the court for Minnesota such that he was All-Defensive 1st Team 6 years running and 2nd/3rd in DPOY 3 of those seasons? Evidently the people that saw him play and voted for those awards thought he was one of the very best defenders in the league - for a long time.

playing the role I’ve explained in my earlier post.

What role did he play in Minnesota such that those that saw him play were so impressed by his defense? Can you explain that?

But did it mean he had greater defensive impact in Boston than Minny? Yes

False.

Hence the question isn’t so much how a stat could be so wrong

On the contrary, this is exactly the question.

Right now the ESPN Real Plus/Minus shows Nikola Jokic as the best defender in the league with a DRPM of 6.26. Is this or is this not a derivative of Englemann's RAPM?

Yet if you go to the stats webpages at NBA.com you'll find that as of today of the 14 players in the league to have faced 400+ FGAs from within 10 feet of the basket, Jokic is allowing a 58.0% FG% (344/593), but best in the league is Jarrett Allen allowing just a 47.2% FG% (218/462). That's almost 11% lower than Jokic. Jokic's 58.0% FG% allowed is the 3rd highest/worst among those 14 players.

And - not surprising - right now on defense Denver as a team allows the 2nd highest/worst 2pt FG% at 55.0% among all 30 teams in the league. Due in large part to Jokic, who has faced on defense almost twice as many shots as any other Nuggets player (593, next most faced is 303 by Aaron Gordon).

Jokic is an excellent defensive rebounder, but defensive rebounding is after the fact. Stopping shots from going in is key to a great defense, and the NBA.com numbers certainly show Jokic's FG% allowed is high/poor for a C.

So how does that square with the ESPN RPM? Is this where those major confounding variables come in? How can it rate Jokic as best in DRPM when he allows one of the highest/worst FG%s within 10' of the basket?

And it’s not just the defense

But defense is what we are discussing here.

The discrepant data that made you think the stat was wrong is indeed an indication that something was wrong about basketball - relative to later eras.

Did you even watch Garnett play during his tenure with Minnesota?

I know very little about these stats (obviously lol, hence my question) but to my knowledge RAPM is very different from PI RAPM in terms of accuracy on account of a single seasons worth of data having a lot of noise. I think it takes a couple of years to get a solid and accurate single year data set.

So I don't think Jokic's DRPM being super high in a single season sample size of about 50 games is enough to discount PI DRAPM.

Also what's your opinion on why KG's DRAPM is so ridiculously high? Is it just crappy stats? Cuz the stat seems to reflect reality with everyone else.

[feyki]

2010 Celtics - w/KG in 69G 3,5 SRS, without KG in 12G 0 SRS,
2011 Celtics - w/KG in 71G 4,5 SRS, without KG in 10G 5,5 SRS,
2013 Celtics - w/KG in 68G -0,5 SRS, without KG in 13G -2,5 SRS

. Don't know about 2011 and also toooo limited sample size, but there was 2-3 SRS difference between the w/KG and the without KG teams. Assuming his offensive impact as positive(somewhere around 0 to 1 SRS), don't think his defensive impact on the dpoy level, Wolves KG was way better defensive impact player than any version of Celt Garnett, except 2008.

Are you being deliberately sloppy? - I have 2013's numbers on hand and you round -0.21 to -0.5 (Garnett's ON sample) but then round -2.78 to -2.5 (Garnett's OFF sample), so Garnett's with/without differential becomes +2 instead of +2.6. That's a 30% deviation lol.

Also, using again 2013 for practicality, considering Boston loses 6.something points offensively on games with KG but gains 9 points on defense, and his offensive +/- is -3, and he doesn't have that many playmaking duties anymore, and he scores under 15 points per game at exactly league average... yeah, I think a +1 offensive impact is a big stretch.

EDIT: just ran 2011 numbers and you shaved another half a point (you say -1 when it's actually -0.5, defense improves nearly 5.5 points and offense worsens the same amount, the half a point difference is due to the strength of schedule, although Boston played at home during most of that KG-less stretch - 8 games out of 11, maybe that balances it out). Can't really rely on your calcs.

I don't much understand of your point that where you want to bring it. Already it is what it is, I didn't calcuate the numbers detailed and rounded it. So, half or so SRS differences are predictable.

I read several times your post, from start to end, really don't get any idea what's your counterpoint with that post. Probably one thing you said that 13 KG was not positive offensive impact player.

[Doctor MJ]

many of us here are people who began using stats like RAPM because we understood that the stat was doing, and thus saw the clear validity in the stat

If there is one thing any derivative of plus/minus is not is clear.

That doesn’t mean the stat is a perfect measure of goodness or value or impact

Correct. It is none of that.

but it does mean something

Like what?

In this case what it means is that Garnett’s presence in Boston correlated more with team defensive effectiveness after normalizing for major confounding variables, than it did in Minnesota.

Could you possibly be more evasive? Confounding variables? What confounding variables?

Those Celtics shocked the world.

The San Antonio Spurs were even better defensively the half decade prior to Garnett going to Boston than Boston was with Garnett. Did they shock the world too?

Sigh. Okay, this is as far as I'm going to go, and then I'm out because you're just being so confrontational and rude for no good reason.

You call me "evasive" when I summarize something for your benefit. Put yourself in my shoes. What am I to expect will help you understand this stuff? And make no mistake, that's what I was trying to do. I don't need to have a conversation with you in order to understand this stuff for my benefit, and you have no authority that I need to persuade in order for my analysis to have respect 'round these parts. I reached out a hand to help, and you slapped it.

Should I just say RAPM is what we get when we put lineup play-by-play data into a linear regression with regularization? I mean, that's all it is, but if you understand that stuff then I doubt we're in a conversation where you're saying that nothing is clear to you.

It's understandable to respond "Okay, what confounding variables are those?", but you weren't doing that. You were attributing cynical explanations for all that was said on your behalf, and that's just going to leave you unable to learn from people who legitimately understand what you don't and are willing (god knows why) to take time away from their day to help you.

Regardless, the important thing is to understand that we level up from team records to raw +/- to more sophisticated metrics to try to normalize for stuff. +/- is more on-point than team record because it doesn't credit the player with stuff that happens when he's off the court. Other stuff is more on-point than +/- because it factors in who else was on the floor at the same time. Regularization helps with something called multicollinearity.

I was going to stop there but the "shocked the world" line is just so concrete I feel like I can get a definite answer that can't help but be understood.

I say "shock the world" specifically in the sense that they drastically overperformed expectations and sent tongues a-wagging. It's not meant as a quantitative or analytical statement. It's about the fact that when Boston came out of the gates in 2007 beating the hell out of everyone, it made people ask questions and try to find explanations, and the conclusions they drew had everything to do with why the mainstream press selected Garnett for his first DPOY that season.

All your points about Garnett's defense in Minnesota seem to be based on the idea that only DRAPM says anything uniquely amazing about Garnett in Boston, but the reality is the data just points in the same direction that everyone was seeing with their eyes that season.

[kcktiny]

Sigh. Okay, this is as far as I'm going to go, and then I'm out because you're just being so confrontational and rude for no good reason.

I've noticed a tendency of those who are adherents of Plus/Minus calculations like RAPM, DRAPM, and the like that when confronted and questioned with having to defend the output of their calculations that - when unable to - decide to take their ball and go home.

I reached out a hand to help, and you slapped it.

Sorry you feel offended, but the fact is you are scripting and crafting a narrative to substantiate the DRAPM results rather than taking into consideration actual player raw numbers and the opinions of those that actually saw the players play and who voted on awards for them.

I am willing to bet that the individuals who calculated these numbers you are taking as factual representations of reality did not even see these players play at the time they in fact played.

Should I just say RAPM is what we get

We all understand the basic concept of Plus/Minus and it's offshoots. It is the posted results that we question, which apparently is something those who defend Plus/Minus often have to make excuses for, by saying things like "...well it's just one season of data...".

It's understandable to respond "Okay, what confounding variables are those?", but you weren't doing that. You were attributing cynical explanations for all that was said on your behalf

You yourself said, quote:

2. In Boston, Kevin Garnett focused on defense
b - Garnett being focused more on defense rather than carrying the offensive burden he did in Minny.

You are crafting your logic here for why his DRAPM was better in Boston, by saying that he was a better defender in Boston vs. in Minnesota, without any proof whatsoever, simply in an attempt to validate the DRAPM results.

d - Garnett taking on a “middle linebacker” or “defensive quarterback” role where he would read what the offense was doing from his position in the interior and tell his teammates where to go.

And then making statements like this. So you are saying it took 12 years in the NBA and multiple All-Defensive 1st Team nominations plus multiple 2nd/3rd placings for DPOY before Garnett ever did this, ever inspired his teammates on the defensive end of the floor? Again you are trying to authenticate DRAPM results as real by making ad hoc statements without any basis in fact.

The fact is Garnett was considered an all-time great defender years before he got to Boston.

Specifically what made Garnett really become a transcendent defense force there
a - Garnett’s ideal body for a defensive player

Was his body somehow different in Boston than those 12 years he played in Minnesota?

and that's just going to leave you unable to learn from people who legitimately understand what you don't

This is typical of people who defend Plus/Minus offshoots. The results are true whether you like it or not, and when you question Plus/Minus results it's not that they are wrong, but that you don't understand what they mean.

All your points about Garnett's defense in Minnesota seem to be based on the idea that only DRAPM says anything uniquely amazing about Garnett in Boston,

False. What I have stated is that he was a great defender in both Minnesota and Boston. Not better in one city or the other.

but the reality is the data just points in the same direction that everyone was seeing with their eyes that season.

"The data"? What "data" might you be referring to?

Boston was the best team in the league defensively in 2007-08 (lowest points per team possession allowed). Also 2nd best in 2008-09, 5th best in 2009-10, and 2nd best in 2010-11.

Also, I notice you did not comment on Nikola Jokic. Again, this is typical of those who defend Plus/Minus and it's offshoots, especially when actual data conflicts with Plus/Minus calculated results.

Like this statement by someone else:

So I don't think Jokic's DRPM being super high in a single season sample size of about 50 games is enough to discount PI DRAPM.

So if one really believes this statement, my question then is this - just what exactly is DRPM, DRAPM, or PI DRAPM telling you about a player?

Same with Duncan Robinson, which ESPN Real Plus/Minus shows has the highest DRPM for SGs.

To do this and do this well, you need an ultra-high defensive BBIQ with vocabulary. You need to not just be able to read the offense instantly, but no how each teammate should optimally respond and what language to use to communicate successfully to them so that they can respond before the offense can successfully exploit them with the next move.

Incidentally, since Garnett, we’ve seen two other players who can truly claim to be like this (that I’m aware of): Marc Gasol and Draymond Green.

All 3 would win DPOY, and I would consider Green to be the best defender of the post-KG era.

Since the 2014-15 season, the best defensive team (lowest points per team possession allowed) in the league has been the Utah Jazz. And during this time Jazz C Rudy Gobert was DPOY 3 times and All-Defensive 1st Team 5 times. Golden State was the 3rd best defensive team during this time, and Draymond Green was DPOY once and All-Defensive 1st Team 4 times.