Jordan and Pippen Plus/Minus Numbers in the 90s

[ShaqAttac]

I'd be hesitant to call the '91 data an RAPM, but I didn't mean to say Rodman was bad or anything, just that he was well off MJ/Pippen and more comparable to Kukoc. Being (probably) the #3 on a GOAT tier team is no shame or anything. KD was still pretty good afterall.

You aren't kidding, are you?

Bro got hit with the fishing lure cheese

it aint bait if its true

[Djoker]

MJ having such a massive net rating with his #2 player on the court, while simultaneously that #2 player has a negative net rating without him (and even when the third best player is on the court with the #2) is incredibly impressive and plainly supportive of him as GOAT. A +15 with your best teammate is very rare and not something most GOAT candidates achieved as far as we know. Of course it’s all caveated by not being full samples, though.

They aren't full samples but they match the overall team W-L record almost perfectly (see posts above) and it is a 157-game sample which is equivalent to almost two full seasons of data. It's not 20 or 30 games.

Both the +13.9 ON and the +19.5 ON-OFF are incredibly impressive. Scratch impressive... Historic is more the word.

Lebron's best two-season sample is at +13.3 ON and +19.0 ON-OFF from 2008-2010 which trails Jordan in both categories.

Garnett's best two-season sample is at +7.9 ON and +22.2 ON-OFF from 2002-2004. KG has a better ON-OFF but a much much worse ON rating.

Duncan's best two-season sample is at +13.9 ON and +12.8 ON-OFF from 2003-2005. He actually matches MJ's ON rating but has a much worse ON-OFF.

Shaq doesn't have any samples that come even close.

The only better two-season samples I found are from Curry who is +17.4 ON and +20.4 ON-OFF from 2014-2016 and +17.6 ON and +19.9 OFF from 2015-2017.

At this point, Jordan has by far the best box score signal (not that close) and GOAT caliber scoreboard signal. The second one wasn't evident till recently but with the playoff numbers released recently and now this, it's clear that he looks like a GOAT in terms of plus-minus as well.

Of course we don't have any plus-minus data for GOAT candidates from prior eras so it's possible that someone is an outlier in these stats like Russell or Kareem and just blows everyone else away. But from the 90's till present, most of the data points in the "Jordan has the GOAT peak" direction.

[lessthanjake]

MJ having such a massive net rating with his #2 player on the court, while simultaneously that #2 player has a negative net rating without him (and even when the third best player is on the court with the #2) is incredibly impressive and plainly supportive of him as GOAT. A +15 with your best teammate is very rare and not something most GOAT candidates achieved as far as we know. Of course it’s all caveated by not being full samples, though.

how do numbers that dont exist for russell or kareem put mj over russell or kareem? cant wait till someone shows a graph or something where bron was doing this in his 30's

Yeah, we don’t have those numbers for Russell or Kareem (besides some numbers for older Lakers Kareem), but I was more just making a point in relation to the person this board placed #1, without saying his name specifically because that’d raise a bat signal for people, which was something I didn’t feel like dealing with.

[ShaqAttac]

MJ having such a massive net rating with his #2 player on the court, while simultaneously that #2 player has a negative net rating without him (and even when the third best player is on the court with the #2) is incredibly impressive and plainly supportive of him as GOAT. A +15 with your best teammate is very rare and not something most GOAT candidates achieved as far as we know. Of course it’s all caveated by not being full samples, though.

how do numbers that dont exist for russell or kareem put mj over russell or kareem? cant wait till someone shows a graph or something where bron was doing this in his 30's

Yeah, we don’t have those numbers for Russell or Kareem (besides some numbers for older Lakers Kareem), but I was more just making a point in relation to the person this board placed #1, without saying his name specifically because that’d raise a bat signal for people, which was something I didn’t feel like dealing with.

you said most goat candidates and said it supports mj as the goat...

im guessin it also doesnt really help vs bron, but idk how ppl find that so i guess ill leave it to the batmans ur tryna avoid

[Jaivl]

Refusing to modify your views in lieu of new information is such an anti-LeBron take, actually.

[70sFan]

im guessin it also doesnt really help vs bron, but idk how ppl find that so i guess ill leave it to the batmans ur tryna avoid

Seriously? Is this how you react to new information? Just ignore it because someone said something to keep you at your comfortable position?

[MrLurker]

These are pretty great numbers as one might expect. I do think the idea this supports him as the greatest player ever seems a touch overzealous.

[eminence]

MJ+Pippen: 8547 possessions, +15.0
MJ, no Pippen: 1410 possessions, +7.4
Pippen, no MJ: 1223 possessions, -5.9
Neither: 1410 possessions: -5.0

From the shared chart.

Steph/Dray from pbpstats.com, RS+PO from '15-'23, minus '20

Steph+Dray: 17278 minutes, +13.4
Steph, no Dray: 5728 minutes, +7.2
Dray, no Steph: 5532 minutes, +3.4
Neither: 9479 minutes, -4.5

Not so dissimilar, the only meaningful difference being the large Dray vs Pippen solo gap, which doesn't bother me too much for Pippen, as we saw him with a much more impressive extended solo stint during MJs first retirement. Overall, I continue to see the duos similarly.

It isn't overwhelming evidence for MJ as the GOAT peak/prime (MJ isn't in my own career GOAT discussion duo to longevity issues), but I'd say it's supports him being a member of that discussion (the other databall era guys we have that show strong +/- evidence for that would probably be KG/Duncan/LeBron/Steph, with maybe a few new guys emerging recently, but I want to see a bit more of to feel solid on that - Giannis/Jokic).

Earlier guys we don't have the info to really compare.

[MrLurker]

MJ+Pippen: 8547 possessions, +15.0
MJ, no Pippen: 1410 possessions, +7.4
Pippen, no MJ: 1223 possessions, -5.9
Neither: 1410 possessions: -5.0

From the shared chart.

Steph/Dray from pbpstats.com, RS+PO from '15-'23, minus '20

Steph+Dray: 17278 minutes, +13.4
Steph, no Dray: 5728 minutes, +7.2
Dray, no Steph: 5532 minutes, +3.4
Neither: 9479 minutes, -4.5

Not so dissimilar, the only meaningful difference being the large Dray vs Pippen solo gap, which doesn't bother me too much for Pippen, as we saw him with a much more impressive extended solo stint during MJs first retirement. Overall, I continue to see the duos similarly.

It isn't overwhelming evidence for MJ as the GOAT peak/prime (MJ isn't in my own career GOAT discussion duo to longevity issues), but I'd say it's supports him being a member of that discussion (the other databall era guys we have that show strong +/- evidence for that would probably be KG/Duncan/LeBron/Steph, with maybe a few new guys emerging recently, but I want to see a bit more of to feel solid on that - Giannis/Jokic).

Earlier guys we don't have the info to really compare.

This seems like a more measured and reasonable conclusion to me. I do think on its own - with no other sort of input - these numbers might support parity between the two - and potentially Garnett and Steph and Duncan and the like. But these aren't the only numbers on the subject - that Jordan's help looks markedly better in years he doesn't play shouldn't be ignored.

Or, perhaps put another way. The recent chorus suggesting Jordan as a match is impossible goes too far. I do think an optimistic view - like simply taking these numbers and running - can get Jordan on even footing with Lebron. But I think a similarly optimistic view can get Lebron well ahead of the pack. And - looking at what each side has offered - I'm not sure an optimistic view of Jordan is as easy to support as an optimistic view of James. And yes - Lebron's sustained excellence and body of work across different leagues and teams helps his peak or prime case greatly for me.

Then we have the matter of most of the players in nba history not having this data - and I can understand why there's pushback when these numbers are provided and is trumpeted of proof - not that Jordan is in the conversation - but that he was the very best ever.

[lessthanjake]

how do numbers that dont exist for russell or kareem put mj over russell or kareem? cant wait till someone shows a graph or something where bron was doing this in his 30's

Yeah, we don’t have those numbers for Russell or Kareem (besides some numbers for older Lakers Kareem), but I was more just making a point in relation to the person this board placed #1, without saying his name specifically because that’d raise a bat signal for people, which was something I didn’t feel like dealing with.

you said most goat candidates and said it supports mj as the goat...

im guessin it also doesnt really help vs bron, but idk how ppl find that so i guess ill leave it to the batmans ur tryna avoid

LeBron James never had a +15 net rating in even one single season when on the court with Wade, with Davis, with Kyrie, with Bosh, or with Kevin Love. Michael Jordan was simply better as a force multiplier with other great players, and this is yet more evidence of it.

And I said it supports MJ’s GOAT candidacy because a +15 net rating with a player’s best teammate is “not something most GOAT candidates achieved as far as we know.” If you want to say MJ, LeBron, Kareem, and Russell are GOAT candidates, then we know LeBron hasn’t achieved it, and we do actually have at least some data for Kareem with Magic and I don’t *think* it gets that high either (though I know it is very good, and someone can correct me if I’m wrong). So I’d say that is supportive of the idea that “most” GOAT candidates haven’t achieved it “as far as we know.”

I’d also note that it’s highly implausible Russell achieved this, given that the teams’ total net rating was never *super* high (highest ever was +7.0) and Russell played incredibly high minutes so he was out there almost all the time. It’s technically *possible* but what would’ve had to happen for this to be the case for Russell is that the Celtics were incredibly good with Russell and his 2nd best teammate on the court while also actually being genuinely bad with Russell on and that 2nd teammate off. For instance, let’s take that year where they had their highest ever net rating—the 1961-1962 season with a +7.0 net rating. Let’s say Cousy was the 2nd best player on the team (debatable at that point, but he did finish the highest of the teammates in MVP voting). Russell played 45.2 minutes a game (out of the team’s total 48.24 minutes a game). Cousy played 28.2 minutes a game. If the Celtics were +15.0 with Russell+Cousy and we assume that essentially all of Cousy’s minutes were with Russell (a pretty reasonable assumption, since Russell basically played all but garbage time), they’d have to have had a -4.26 net rating with Cousy off the court for the team’s overall net rating to come out to +7.0. Even if we assume that the Celtics had something horrible like a -10.0 net rating in the small number of no-Russell-no-Cousy minutes, they’d still have to be -3.23 with Russell on and Cousy off for it to be possible for them to have also been +15.0 with Russell and Cousy both on. And that’s in the Celtics’s highest net rating year. There’s a bit of leeway in the numbers there, since it’s possible that Cousy had a very small number of minutes with Russell off. But basically, given the Celtics’ overall net ratings and the number of minutes Russell played, it’s not really mathematically possible for Russell to have ever had a +15.0 net rating with his 2nd best teammate on the court, unless it was also the case that the Celtics were genuinely bad with Russell on and the 2nd best teammate off. And that is both implausible and also definitely wouldn’t support Russell for GOAT (and, in fact, wouldn’t necessarily even support Russell as that era’s Celtics best player!).

So yeah, I'm pretty comfortable with the assertion I made. The only thing I'll note is that my guess is Kareem + Oscar put up similar net ratings in their best couple years and that it seems plausible to me that Kareem + Magic might’ve done it at some point if we had fuller data.

[eminence]

.

Would anyone with some spare time on their hands combine this with the '97/'98 data?

[DraymondGold]

LeBron James never had a +15 net rating in even one single season when on the court with Wade, with Davis, with Kyrie, with Bosh, or with Kevin Love. Michael Jordan was simply better as a force multiplier with other great players, and this is yet more evidence of it.

And I said it supports MJ’s GOAT candidacy because a +15 net rating with a player’s best teammate is “not something most GOAT candidates achieved as far as we know.” If you want to say MJ, LeBron, Kareem, and Russell are GOAT candidates, then we know LeBron hasn’t achieved it, and we do actually have at least some data for Kareem with Magic and I don’t *think* it gets that high either (though I know it is very good, and someone can correct me if I’m wrong). So I’d say that is supportive of the idea that “most” GOAT candidates haven’t achieved it “as far as we know.”

I’d also note that it’s highly implausible Russell achieved this, given that the teams’ total net rating was never *super* high (highest ever was +7.0) and Russell played incredibly high minutes so he was out there almost all the time. It’s technically *possible* but what would’ve had to happen for this to be the case for Russell is that the Celtics were incredibly good with Russell and his 2nd best teammate on the court while also actually being genuinely quite bad with Russell on and that 2nd teammate off. For instance, let’s take that year where they had their highest ever net rating—the 1961-1962 season with a +7.0 net rating. Let’s say Cousy was the 2nd best player on the team (debatable at that point, but he did finish the highest of the teammates in MVP voting). Russell played 45.2 minutes a game (out of the team’s total 48.24 minutes a game). Cousy played 28.2 minutes a game. If the Celtics were +15.0 with Russell+Cousy and we assume that essentially all of Cousy’s minutes were with Russell (a pretty reasonable assumption, since Russell basically played all but garbage time), they’d have to have had a -4.26 net rating with Cousy off the court for the team’s overall net rating to come out to +7.0. Even if we assume that the Celtics had something horrible like a -10.0 net rating in the small number of no-Russell-no-Cousy minutes, they’d still have to be -3.23 with Russell on and Cousy off for it to be possible for them to have also been +15.0 with Russell and Cousy both on. And that’s in the Celtics’s highest net rating year. There’s a bit of leeway in the numbers there, since it’s possible that Cousy had a very small number of minutes with Russell off. But basically, given the Celtics’ overall net ratings and the number of minutes Russell played, it’s not really mathematically possible for Russell to have ever had a +15.0 net rating with his 2nd best teammate on the court, unless it was also the case that the Celtics were genuinely bad with Russell on and the 2nd best teammate off. And that is both implausible and also definitely wouldn’t support Russell for GOAT (and, in fact, wouldn’t necessarily even support Russell as that era’s Celtics best player!).

So yeah, I'm pretty comfortable with the assertion I made. The only thing I'll note is that my guess is Kareem + Oscar put up similar net ratings in their best couple years and that it seems plausible to me that Kareem + Magic might’ve done it at some point if we had fuller data.

Great post jake! And especially great post given what you were responding to...
*You show number that are objectively more favorable to Jordan over LeBron* Response: "im guessin it also doesnt really help vs bron, but idk how ppl find that".
*you state that don't want to turn this into another Jordan vs LeBron debate.* Response: "so i guess ill leave it to the batmans ur tryna avoid". Wow, lol Not that these numbers are definitive or anything. LeBron absolutely still has a reasonable GOAT case. But like you say, they are supportive of Jordan's GOAT level ceiling raising. Me personally, I tend to say Russell, Kareem, Jordan, and LeBron all have reasonable GOAT cases (and I could maybe justify others like Wilt if you do dramatic criteria changes). I don't care much if you happen to have a different GOAT than me, though I do get more irritated when people start making ridiculous claims like "LeBron has the only reasonable modern GOAT case, and it's not reasonable to think Jordan's the GOAT", or vice versa. Which you'd think wouldn't be too much to ask for. Alas, even on this forum, you get people who are so sure that they're right and others are wrong that you end up getting worse discussion occasionally.

Regardless, I was just thinking of trying something similar when you posted this! Given that we know the team net rating for Kareem/Russell/Wilt and we know how many minutes they played each season, it seems like we should be able to do the math for how bad the team would have needed to be when those stars were off the court for those 3 players to match Jordan in his ON-rating and his ON-OFF. We could then do a bit of a sanity check: is this a reasonable off-rating for a team without a Top 10 player, or is this outlier bad to the extent that it seems quite unlikely.

Of course there are limitations: given how little Plus/minus data we have from the 50s–70s, it's possible that the general scaling of plus minus was different back then. Minutes and rotations were distributed differently after all. If so, someone might still be able to make an 'era-relative' plus minus argument for an older player like Russell/Wilt/Kareem, using e.g. the standard deviation of their plus minus rather than the raw plus minus. If we had the data to do this of course. I'm not sure how compelling an argument like that would be (and again, we don't have the data either way) -- I'm more just exploring the different arguments you could still make make if for Russell/Wilt/Kareem to be GOATs, in the case they end up having worse ON / ON-OFF than Jordan. Obviously if they have better ON / ON-OFF, that could be supportive of their GOAT case too.

Another factor to consider: units! On and On-off are usually per 100 possessions, rather than Per Season or something like that. Since Russell/Kareem/Wilt are playing more minutes than Jordan usually, it's possible they could look worse per 100 possessions but end up better in total season value. The argument there might be they were helping their team more per game and per season, but needed to conserve energy more on a per-possession level since they were playing more.

It's also interesting that giving peak Russell + Cousy the same ON rating as Jordan + Pippen actually mathematically requires Russell look *worse* in On-Off than both Jordan *and Cousy*! Not what I would have guessed, but it makes sense when you write out the math like that. Question: does this assume an equal pace when Russell/Cousy are on/off? We don't have any lineup pace data , so we can't do much better, but that's another source of uncertainty that could be referenced to argue that it's possible for Russell's ON/ON-Off to be comparable to Jordan's.

Edit:
second question: Let's take your methodology for Russell. As I understand it, you assumed 1) that Russell + Cousy's ON rating matches Jordan + Pippen's ON rating, 2) that Cousy's minutes overlapped with Russell's mostly (very reasonable like you say), 3) that the celtics had some "horrible like a -10.0 net rating in the small number of no-Russell-no-Cousy minutes" .... -> to see what would be required for Russell to match Jordan's ON rating (assumption 1).

What if, instead we assume 1) that Russell (or Russell + Cousy)'s ON rating at least matches Jordan's (or Jordan + Pippen's), 2) that Cousy's minutes overlapped with Russell's mostly, and 3) that Russell's ON-OFF at least matches Jordan's? I.e., just change the third assumption, then see what OFF would be required for Russell to at least match both Jordan's ON and his ON-OFF (or do better in either). Would this be unreasonable or even mathematically possible, given the net rating and Russell's minutes?

[lessthanjake]

Another factor to consider: units! On and On-off are usually per 100 possessions, rather than Per Season or something like that. Since Russell/Kareem/Wilt are playing more minutes than Jordan usually, it's possible they could look worse per 100 possessions but end up better in total season value. The argument there might be they were helping their team more per game and per season, but needed to conserve energy more on a per-possession level since they were playing more.

It's also interesting that giving peak Russell + Cousy the same ON rating as Jordan + Pippen actually mathematically requires Russell look *worse* in On-Off than both Jordan *and Cousy*! Not what I would have guessed, but it makes sense when you write out the math like that. Question: does this assume an equal pace when Russell/Cousy are on/off? We don't have any lineup pace data , so we can't do much better, but that's another source of uncertainty that could be referenced to argue that it's possible for Russell's ON/ON-Off to be comparable to Jordan's.

Yeah, I was assuming equal pace, which is another confounding factor. The numbers definitely aren’t exact, because the pace could be a bit different in the different scenarios and also Cousy may have played a small number of minutes with Russell off. But it gives us a general ballpark for roughly where the numbers would have to be for Russell in order for him to have matched Jordan’s +15 number with his 2nd best player. It just seems implausible to me.

I think you’re right that the plus-minus scaling for great players was probably a fair bit different in the past and also playing more minutes is a benefit that mitigates a lower plus-minus, so there’s definitely a potential objection that Russell can’t be expected to have the same kind of plus-minus as more recent players. Those are perfectly fine nuances to raise (and you rightly raised them in your post), but of course that sort of thing would be a contextualization of what I had initially said, rather than a demonstration that what I said was factually incorrect. Not that you’re saying otherwise—I’m saying this more for other people.

Edit:
second question: Let's take your methodology for Russell. As I understand it, you assumed 1) that Russell + Cousy's ON rating matches Jordan + Pippen's ON rating, 2) that Cousy's minutes overlapped with Russell's mostly (very reasonable like you say), 3) that the celtics had some "horrible like a -10.0 net rating in the small number of no-Russell-no-Cousy minutes" .... -> to see what would be required for Russell to match Jordan's ON rating (assumption 1).

What if, instead we assume 1) that Russell (or Russell + Cousy)'s ON rating at least matches Jordan's (or Jordan + Pippen's), 2) that Cousy's minutes overlapped with Russell's mostly, and 3) that Russell's ON-OFF at least matches Jordan's? I.e., just change the third assumption, then see what OFF would be required for Russell to at least match both Jordan's ON and his ON-OFF (or do better in either). Would this be unreasonable or even mathematically possible, given the net rating and Russell's minutes?

Well, let’s take the 1962-1963 season I used (i.e. the season the Celtics had the highest net rating). It is not mathematically possible for Russell’s ON and his ON-OFF to match Jordan. Given the Celtics’ overall net rating, if Russell’s ON was as high as Jordan’s in this data (+13.9), then the OFF would need to be implausibly terrible to get to the +7.0 net rating the Celtics had overall. Specifically, if Russell matched Jordan’s +13.9 overall net rating while on, then the Celtics would’ve had to have a net rating of about -48.55 with Russell off in order to only have a +7.0 net rating overall. They only played like 426 minutes without Russell, but that number seems obviously implausible.

Another way of looking at this is to say what the net rating with Russell would be if we assume some specific OFF rating. Let’s say we assume the Celtics were an awful -10.0 with Russell off the court (almost certainly not the case IMO). And we again use the 1962-1963 season, where they had an overall net rating of +7.0. If the Celtics were -10.0 with Russell off the court, they could only have been +9.11 with Russell on the court (assuming equal pace on and off). Which would be an amazing ON-OFF up there with this Jordan data, but would fall well short in ON. Even if we assumed an absolutely awful -15.0 with Russell off the court, it’s still just lead to a +9.73 with Russell on the court. The real numbers here are surely even lower, since the Celtics were likely not nearly this bad with Russell off. And, again, this is in the Celtics’ best season. There’s basically no way Russell matched Jordan’s ON that we see in this data.

__________________

Full Disclosure: I realize by the way that my calculations in my above post were slightly wrong, since I used their minutes per game to weight the net ratings, rather than their total minutes, and these guys didn’t quite play every game so that matters. It actually changes the numbers a decent bit, but not the directional point. Basically, if Cousy played every minute with Russell and we assume equal pace with various lineups, then, for Russell + Cousy to be +15.0, the Celtics would need to be -2.69 with Cousy off. And even if we assume the Celtics had an awful -10.0 net rating in the few minutes with Russell + Cousy both off (almost certainly not actually the case), then the Celtics would still need to have a negative rating with Russell on + Cousy off. And that’s in the season where the Celtics had the highest net rating they got with Russell—it’d be even worse in other seasons. Of course, again, there’s various assumptions there, so I wouldn’t put much stock in the actual number, but it’s just demonstrative of the overall point that basically it’s not really possible for Russell to match the +15.0 with his second best player, unless the net rating was pretty poor without that second best player. And that seems highly unlikely, and also if it were the case then it wouldn’t really reflect all that well on Russell anyways.

[ShaqAttac]

Yeah, we don’t have those numbers for Russell or Kareem (besides some numbers for older Lakers Kareem), but I was more just making a point in relation to the person this board placed #1, without saying his name specifically because that’d raise a bat signal for people, which was something I didn’t feel like dealing with.

you said most goat candidates and said it supports mj as the goat...

im guessin it also doesnt really help vs bron, but idk how ppl find that so i guess ill leave it to the batmans ur tryna avoid

LeBron James never had a +15 net rating in even one single season when on the court with Wade, with Davis, with Kyrie, with Bosh, or with Kevin Love. Michael Jordan was simply better as a force multiplier with other great players, and this is yet more evidence of it.

so basically ur arguin port again but coz you know no one here takes that seriously you tried to put it with a bunch of other stuff to make it look better than it is

obviously better than most goat candidates was cap, but bron wasnt enough ig

[ShaqAttac]

im guessin it also doesnt really help vs bron, but idk how ppl find that so i guess ill leave it to the batmans ur tryna avoid

Seriously? Is this how you react to new information? Just ignore it because someone said something to keep you at your comfortable position?

why dont you keep this energy for when information that goes against mj or for lebron comes? or maybe youre just comfortable pretending theyre the same...

idk why im getting flack for not taking random numbers tossed with bs attached.

how did most goat candidates turn into bron? and why didnt you check if those numbers were actually better than bron's? all lessthan is doin here is pushin port again. that aint new information, that's the same old same old and you wanna be in the middle so you pretended it was something special

[Clyde Frazier]

im guessin it also doesnt really help vs bron, but idk how ppl find that so i guess ill leave it to the batmans ur tryna avoid

Seriously? Is this how you react to new information? Just ignore it because someone said something to keep you at your comfortable position?

why dont you keep this energy for when information that goes against mj or for lebron comes? or maybe youre just comfortable pretending theyre the same...

idk why im getting flack for not taking random numbers tossed with bs attached.

how did most goat candidates turn into bron? and why didnt you check if those numbers were actually better than bron's? all lessthan is doin here is pushin port again. that aint new information, that's the same old same old and you wanna be in the middle so you pretended it was something special

If a respected poster here shares manual tracking data (which is a huge undertaking to say the least) we absolutely take it seriously. If it supports a player you aren't in favor of that's too bad, but it doesn't automatically make it invalid either.

[MyUniBroDavis]

im guessin it also doesnt really help vs bron, but idk how ppl find that so i guess ill leave it to the batmans ur tryna avoid

Seriously? Is this how you react to new information? Just ignore it because someone said something to keep you at your comfortable position?

why dont you keep this energy for when information that goes against mj or for lebron comes? or maybe youre just comfortable pretending theyre the same...

idk why im getting flack for not taking random numbers tossed with bs attached.

how did most goat candidates turn into bron? and why didnt you check if those numbers were actually better than bron's? all lessthan is doin here is pushin port again. that aint new information, that's the same old same old and you wanna be in the middle so you pretended it was something special

70sfan is the last person I expected to make shaqattack break character

[eminence]

Rodman has good numbers in the 1991 RAPM in a season where Isiah missed half of the season. He was also solid in 1997.

The Jordan/Pippen/Rodman/Kukoc lineups were insanely good. They were +17 in 1996, +29 in 1997, and +20 in 1998.

I'd be hesitant to call the '91 data an RAPM, but I didn't mean to say Rodman was bad or anything, just that he was well off MJ/Pippen and more comparable to Kukoc. Being (probably) the #3 on a GOAT tier team is no shame or anything. KD was still pretty good afterall.

You aren't kidding, are you?

I meant to get back to this.

Not really kidding I guess, though I mostly said KD instead of Dray for dramatic effect. I have Dray/KD pretty close for '17-'19, leaning Dray due to the playoff injuries for KD. In terms of quality when each played I'm not sure which way I'd go off the top (both clearly behind Steph).

[MyUniBroDavis]

Another factor to consider: units! On and On-off are usually per 100 possessions, rather than Per Season or something like that. Since Russell/Kareem/Wilt are playing more minutes than Jordan usually, it's possible they could look worse per 100 possessions but end up better in total season value. The argument there might be they were helping their team more per game and per season, but needed to conserve energy more on a per-possession level since they were playing more.

It's also interesting that giving peak Russell + Cousy the same ON rating as Jordan + Pippen actually mathematically requires Russell look *worse* in On-Off than both Jordan *and Cousy*! Not what I would have guessed, but it makes sense when you write out the math like that. Question: does this assume an equal pace when Russell/Cousy are on/off? We don't have any lineup pace data , so we can't do much better, but that's another source of uncertainty that could be referenced to argue that it's possible for Russell's ON/ON-Off to be comparable to Jordan's.

Yeah, I was assuming equal pace, which is another confounding factor. The numbers definitely aren’t exact, because the pace could be a bit different in the different scenarios and also Cousy may have played a small number of minutes with Russell off. But it gives us a general ballpark for roughly where the numbers would have to be for Russell in order for him to have matched Jordan’s +15 number with his 2nd best player. It just seems implausible to me.

I think you’re right that the plus-minus scaling for great players was probably a fair bit different in the past and also playing more minutes is a benefit that mitigates a lower plus-minus, so there’s definitely a potential objection that Russell can’t be expected to have the same kind of plus-minus as more recent players. Those are perfectly fine nuances to raise (and you rightly raised them in your post), but of course that sort of thing would be a contextualization of what I had initially said, rather than a demonstration that what I said was factually incorrect. Not that you’re saying otherwise—I’m saying this more for other people.

Edit:
second question: Let's take your methodology for Russell. As I understand it, you assumed 1) that Russell + Cousy's ON rating matches Jordan + Pippen's ON rating, 2) that Cousy's minutes overlapped with Russell's mostly (very reasonable like you say), 3) that the celtics had some "horrible like a -10.0 net rating in the small number of no-Russell-no-Cousy minutes" .... -> to see what would be required for Russell to match Jordan's ON rating (assumption 1).

What if, instead we assume 1) that Russell (or Russell + Cousy)'s ON rating at least matches Jordan's (or Jordan + Pippen's), 2) that Cousy's minutes overlapped with Russell's mostly, and 3) that Russell's ON-OFF at least matches Jordan's? I.e., just change the third assumption, then see what OFF would be required for Russell to at least match both Jordan's ON and his ON-OFF (or do better in either). Would this be unreasonable or even mathematically possible, given the net rating and Russell's minutes?

Well, let’s take the 1962-1963 season I used (i.e. the season the Celtics had the highest net rating). It is not mathematically possible for Russell’s ON and his ON-OFF to match Jordan. Given the Celtics’ overall net rating, if Russell’s ON was as high as Jordan’s in this data (+13.9), then the OFF would need to be implausibly terrible to get to the +7.0 net rating the Celtics had overall. Specifically, if Russell matched Jordan’s +13.9 overall net rating while on, then the Celtics would’ve had to have a net rating of about -48.55 with Russell off in order to only have a +7.0 net rating overall. They only played like 426 minutes without Russell, but that number seems obviously implausible.

Another way of looking at this is to say what the net rating with Russell would be if we assume some specific OFF rating. Let’s say we assume the Celtics were an awful -10.0 with Russell off the court (almost certainly not the case IMO). And we again use the 1962-1963 season, where they had an overall net rating of +7.0. If the Celtics were -10.0 with Russell off the court, they could only have been +9.11 with Russell on the court (assuming equal pace on and off). Which would be an amazing ON-OFF up there with this Jordan data, but would fall well short in ON. Even if we assumed an absolutely awful -15.0 with Russell off the court, it’s still just lead to a +9.73 with Russell on the court. The real numbers here are surely even lower, since the Celtics were likely not nearly this bad with Russell off. And, again, this is in the Celtics’ best season. There’s basically no way Russell matched Jordan’s ON that we see in this data.

__________________

Full Disclosure: I realize by the way that my calculations in my above post were slightly wrong, since I used their minutes per game to weight the net ratings, rather than their total minutes, and these guys didn’t quite play every game so that matters. It actually changes the numbers a decent bit, but not the directional point. Basically, if Cousy played every minute with Russell and we assume equal pace with various lineups, then, for Russell + Cousy to be +15.0, the Celtics would need to be -2.69 with Cousy off. And even if we assume the Celtics had an awful -10.0 net rating in the few minutes with Russell + Cousy both off (almost certainly not actually the case), then the Celtics would still need to have a negative rating with Russell on + Cousy off. And that’s in the season where the Celtics had the highest net rating they got with Russell—it’d be even worse in other seasons. Of course, again, there’s various assumptions there, so I wouldn’t put much stock in the actual number, but it’s just demonstrative of the overall point that basically it’s not really possible for Russell to match the +15.0 with his second best player, unless the net rating was pretty poor without that second best player. And that seems highly unlikely, and also if it were the case then it wouldn’t really reflect all that well on Russell anyways.

Y’all Running a virtual pick and roll is hilarious

[ShaqAttac]

Seriously? Is this how you react to new information? Just ignore it because someone said something to keep you at your comfortable position?

why dont you keep this energy for when information that goes against mj or for lebron comes? or maybe youre just comfortable pretending theyre the same...

idk why im getting flack for not taking random numbers tossed with bs attached.

how did most goat candidates turn into bron? and why didnt you check if those numbers were actually better than bron's? all lessthan is doin here is pushin port again. that aint new information, that's the same old same old and you wanna be in the middle so you pretended it was something special

If a respected poster here shares manual tracking data (which is a huge undertaking to say the least) we absolutely take it seriously. If it supports a player you aren't in favor of that's too bad, but it doesn't automatically make it invalid either.

no one said anything about colt

its about lessthan tryna use numbers which dont say mj is the goat to say mj is the goat