What was the ceiling for the 90 and 95 Bulls?

eminence · Post #21 » by **eminence** » Sat Aug 12, 2023 10:14 pm

I don't think the squared data is viable for cross team comparisons, more useful for establishing impact hierarchies on the few big sample teams (Bulls, Lakers, Celtics if memory serves). Even for that, it is very narrow sample (against a few specific matchups), with some worries on which games were recorded for posterity as being potentially biased.

MrVorp · Post #22 » by **MrVorp** » Sat Aug 12, 2023 11:35 pm

lessthanjake wrote:
homecourtloss wrote:I remember when the ‘97 and ‘98 and then the ‘96 play by data became available and people were rightfully high on Jordan and the line of thinking was always “imagine what peak Jordan looked like.” Well, it looks like peak Jordan wasn’t making some mythological untouchable impact by the data we have, i.e., ElGee’s playoff on/off data, squared2020’s partial RAPM data (great impact, but not anywhere near the myths, and not anymore impressive than 1996-1998 Jordan), and Moonbeam’s Regressed WOWY data

The problem is that this isn’t really true. Much of the data you refer to there shows Jordan as having GOAT-level impact.

- Ben Taylor’s playoff on/off data has Jordan at around a +23 playoff on/off per 48 minutes from 1988 to 1990. Which is obviously enormous. In the first-three-peat years, that goes down to about a +7.5 on/off (because the team does better with Jordan off the court than before). What do those numbers look like combined? Well, if we do weighted averages of the figures in those time periods, using the actual number of minutes on and off in each period and eyeballing the values in the charts in Ben’s video (for 1988-1990, I used -20 off and +3 on, while for 1991-1993 I used +1.5 off and +9 on), we get an overall average of about +6.41 on and -6.84 off, in the playoffs from 1988-1993. That’s a +13.25 on/off. Which looks great, but that’s also in per 48 minutes terms, not per 100 possessions. The weighted average of the Bulls’ pace in those playoffs is 90.06. Which means that, using the Thinking Basketball on-off numbers, in the 6-year span from 1988-1993, Michael Jordan had a playoff on/off per 100 possessions of about +14.71! That certainly looks pretty GOAT-like!

- And then, Ben Taylor also provides similar data for the 1996 playoffs, with Jordan having about a +16 on/off per 48 minutes, which would translate to about +18.4 on/off per 100 possessions. And we know it was +23.6 in the 1997 playoffs and +13.1 in the 1998 playoffs. So, you’re right in the sense that the on-off numbers from 1988-1993 are not more impressive than those from 1996-1998, but it’s *all* super impressive. If we combine those 1998-1993 numbers with the 1996-1998 numbers, we get a playoff on/off per 100 possessions of about +16.0! Which is enormous and obviously GOAT-like. Obviously, that’s missing some shorter playoff runs (i.e. his first three years and 1995), where we don’t have this data, but we have the most important years and they look incredibly good.

- I genuinely don’t understand the idea that Squared’s RAPM data suggests anything but great things about Jordan. Squared has RAPM for four different seasons for Jordan: 1984-1985, 1987-1988, 1990-1991, and 1995-1996. Jordan is ranked 1st in the NBA in 1987-1988, 1990-1991, and 1995-1996, and no one is even particularly close to him in any of those years. Meanwhile, 1984-1985 was Jordan’s rookie season, and he’s still ranked 6th in the NBA (and only behind three actual star players). While Squared’s data is obviously just a snippet, this is clearly quite supportive of Jordan as having been a dominant impact guy.

- The Moonbeam stuff is interesting, but it has some unsolvable problems with distinguishing impact when multiple good players didn’t miss many games. Given that Jordan, Pippen, and Grant missed very few games for many years, the model has virtually no way of distinguishing who had what impact between those guys in those years. And it has some pretty obvious issues with overestimating certain players (the BJ Armstrong Problem is one we’ve talked about in that thread, where his impact on the Bulls was clearly overestimated due to the minutes cutoff, which suppressed others’ impact; and there are plenty of other examples). In any event, until Jordan retired the first time (at which point the model quickly recognizes Jordan as super elite), it has basically no data to distinguish Jordan/Pippen/Grant’s impact, except for the 1986 Jordan injury season that factors into those early data points for Jordan. And, of course, even that injury-season signal is pretty iffy when we realize that Jordan had a steadily increasing minutes restriction and the Bulls got destroyed in the games where Jordan played but had super low minutes and those games end up counting against him for purposes of that signal. If the model even just only counted specific games where Jordan played the minutes cutoff (i.e. 18+ minutes), that signal would look way better, and Jordan would end up looking substantially higher in those early years (since that’s really the only signal for him that the model has and it’s clearly distorted). I should also note that a version of Moonbeam’s analysis (Lasso) has Jordan pretty immediately vaulting to the top range once that year is out of the system and then has him staying there for a long time. And even the Ridge version has Jordan hovering right near the top for quite a while in a way that only a handful of other all-time greats have. Meanwhile, the aggregate of multiple WOWYR ratings—which is the same sort of analysis as Moonbeam’s—has Jordan ranked above every single other serious GOAT candidate. Overall, I think it’d be really flimsy analysis to decide that Jordan’s impact is not GOAT-like on the basis specifically of Moonbeam’s Ridge model—which is one piece of data (in which Jordan still looks really great, just not the best) amidst a sea of data going the other way, and it pretty clearly has inherent limitations that make it not all that suited to accurately measure Jordan’s impact in his peak years. And, frankly, I don’t even think Moonbeam would disagree with me.

Yeah, the mental gymnastics that it takes to make Jordan’s peak as anything other than GOAT like is pretty funny. As for the WOWY stuff, a time decayed version (which fixes a lot of the issues of smoothing out large time spans) ran by the creator of DARKO has Jordan peaking as the highest right at his first retirement.

Read on Twitter

?s=46&t=SSGuQgt0b5os6LeB9cXuIQ

OhayoKD · Post #23 » by **OhayoKD** » Sun Aug 13, 2023 9:18 am

Just had a night full of booze, tats, dumplings, and headbutts but let's see if I can put together something semi-coherent before I doze off.

First, some house-cleaning:
Let's start with the thing your jordan posts in the 86 Rockets thread centered around...

Despite the fact that his supporting cast shot so badly in the series they lost that it’s almost impossible to find any series in the history of the NBA where a supporting cast shot so badly and the team still won. That’s just not a serious argument

Uh...

https://www.basketball-reference.com/playoffs/1974-nba-western-conference-finals-bulls-vs-bucks.html
https://www.basketball-reference.com/playoffs/1977-nba-western-conference-semifinals-warriors-vs-lakers.html
https://www.basketball-reference.com/playoffs/1964-nba-western-division-finals-hawks-vs-warriors.html
https://www.basketball-reference.com/playoffs/1962-nba-eastern-division-semifinals-nationals-vs-warriors.html
https://www.basketball-reference.com/playoffs/2013-nba-eastern-conference-finals-pacers-vs-heat.html
https://www.basketball-reference.com/playoffs/2018-nba-eastern-conference-finals-cavaliers-vs-celtics.html
https://www.basketball-reference.com/playoffs/2018-nba-eastern-conference-first-round-pacers-vs-cavaliers.html

Not really?

As enigma covered, it's pretty dumb to ignore roles here. Jordan's pretty much sole job was to score and the defense was outstanding(-5) despite being average in 89 and below average before Pippen/Grant ascended over the course of 1990. And this gets us to...

Lol, a guy has a 31.7 playoff PER and 13.7 playoff BPM (both of which are top 10 all time for a single playoffs)

Yeah, notice how I said "despite subjective box-weightings". PER and BPM can't really account for pesky details like creation quality, ball-handling, team-wide orchestration, defensive attention(the triangle made doubling Jordan very difficult), or most-everything that happens defensively. All components Jordan is disadvantaged at relative to the best of the best. You have of course read something along those lines from different posters at several points and have chosen to pretend all of the above are not disadvantages.

Hakeem has 3 series with a similar scoring-efficiency gap as Jordan's 1990 and had he had the defensive support Jordan had, I imagine those may have ended as wins. The 1994 finals and the 1995 conference finals are also in the vicinity and thanks to Hakeem's much better defense, both ended in wins.

Lebron also put up monstrous per/bpm/whatever for the entirety of 2009. The difference here is that he was also playing better defense than Jordan ever could, was facing more defensive attention as the primary ball-handler and was a more efficient passer. Let's cut the semantics. If Jordan played goat-level ball in 1990, then you should be able to explain to me in basketball terms how Jordan compensated for all the aforementioned disadvantages. As covered "ball-dominance -> worse offense" is not really a thing

Moving on to on/off...

homecourtloss wrote:

As unibro pointed out about a year ago, Shaq's data here is inaccurate. Corrected, Shaq lands somewhere around MJ and Steph.

alongside a playoff on-off that was around +36 per 100 possessions (maybe the highest ever recorded for a player that got to at least the conference finals)

Nope. In 2001 Duncan, scored at +38. Also, as always, single-year playoff is noisy, and, as has been covered before, when we extend the sample, Jordan comes out looking a fair bit worse than Lebron(who is the only other popular goat candidate we have that sort of on/off for):

Spoiler:

He loses pretty decisively to a guy we do have the data for and we do not have this data for other players he might lose to like Kareem, Russell, Hakeem, and Magic so I'm confused how you decided playoff on/off affirms MJ as "goat-level".

Rs on/off doesn't really help(on+on/off for two of his best rs teams during his potential impact peak are middle of the pack among Lebron years...), but whatver. Jordan's impact signals have been discussed to death so I'll just end with some points on Darko(or the latest variant):

MrVorp wrote:Yeah, the mental gymnastics that it takes to make Jordan’s peak as anything other than GOAT like is pretty funny. As for the WOWY stuff, a time decayed version (which fixes a lot of the issues of smoothing out large time spans) ran by the creator of DARKO has Jordan peaking as the highest right at his first retirement.

Honestly would want to know how RAPM(which we don't have pre-97) is being used here(with WOWY vs WOWY it swings strongly in Lebron's favor) but regardless, while Darko's a great metric, it's important to understand how it works:

Spoiler:

OhayoKD wrote:

So it isn't spiking or deflating player performance by checking how players compare relative to expected trajectories?

If I've counted correctly, Jordan trails Magic's trajectory through the end of 1990 and matches a second Magic peak(somehwere between 87-90) as of 1993.

That big spike comes during the second-three-peat where Jordan is aged 32-34(magic was forced to retire at 29 and came back one season at 36).

Now you can go with 96-98 Mj just being >>>>> "prime" MJ(in which case we actually have data-ball for MJ's "peak" which now doesn't look that great), the reason Jordan is spiking in his 30's is because he was unusually good at 30+ and not so unusually good at what would be his typical peak. Actually Magic and Bird also benefit from this(notice how they go up as they age?) though they for different reasons they are not participating during the height of expansion where 30+ players were peaking

You can also see this with Lebron whose trajectory is well ahead of the field during his mid-late 20's is mysteriously upsurped by 30+ Jordan, and then reaches newer heights than everyone else in his 30's. Is this because 30+ Lebron was more valuable than mid-20's Lebron? No. It's because Lebron was a bigger outlier relative to other players.

Additionally because you are relying strictly on WOWY, players are at the mercy of timing. Duncan barely misses games up until 2004 so he doesn't do too well higher. But if we swtich to Darko DPM, Duncan is kicking everyone's ass because Darko now has access to that early stretch where Duncan was already one of the best players ever by the age of 22. The only player who tops Duncan's early trajectory is Lebron but then Lebron has 2011 at an age where Darko expects you to peak and Lebron's trajectory collapses to 2nd best behind the big fundamental.

Darko is a great stat which is very good at what it does. But you have to understand what it does to use it properly. Darko suggests that Jordan was crazy good for a 30+ guy during the second-three peat. It also suggests he was good but not crazy good for a guy in his 20's/mid-20's during the 80's and early 90's.

I know how the stat works, which is why I'm not just looking at which lines are higher on the chart. Lebron at game 1600 was almost certainly not a better player than peak Micheal. And while there's certainly evidence to suggest the second three-peat might have been Jordan's apex in terms of "era-relative situational impact", he was not a waaaay better there than he was in the years some consider the best ever.[/quote]

Jordan skyrocketing during the second threepeat should be a bit of a tell here. While there's a decent chance that would be his impact peak anyway(in which case various metrics do not paint him particularly close to "goat lvl"), what Darko is really assessing here is how MJ compares relative to other players at that point in their trajectory. Jordan isn't looking that hot when he was, in an absolute sense, "better", but expansion offers him aid other players in these graphs do not get(magic retires, bird is injured, Lebron isn't alive), and Jordan still seems to fall short when the same data is used for everyone even with the likes of Kareem, Russell, and Darko king Duncan excluded. HCL covered a bunch of the rest, and I think at this point most regulars here have some awareness of the impact stuff so let's end this with some granular analysis...

OldSchoolNoBull wrote:
OhayoKD wrote:
OldSchoolNoBull wrote:I don't think MJ was the issue.

His 1995 playoff numbers: 31.5ppg/6.5rpg/4.5apg/2.3spg/1.4bpg, 55.7% TS, 8.0BPM, 6.3 PO RAPTOR
His 1996 playoff numbers: 30.7ppg/4.9rpg/4.1apg/1.8spg/0.3bpg, 56.4% TS, 10.7BPM, 9.0 PO RAPTOR

Given that those Bulls played at a 50-win pace before Mj arrived, I'd say throwing out box-numbers to argue he "wasn't the issue" when they lost to a soon-to-be swept finalist is pretty dubious

Well, there's not a whole lot non-box to go by in terms of MJ's individual

[/quote][/quote]
By non-box I really meant defense. But honestly I don't have a strong opinion on the degree of the gap so we can table it I guess. You make good points about the matchup tbf and Grant's arrival coincided with significant improvement for Magic. Rodman was basically his replacement in a way and offensive rebounding proved very important for Chicago:

Spoiler:

Rodman also looks pretty strong if you go by WOWY for those years(similar lift to oscar in 72, kd on the warriors) so I'm certainly not opposed to giving him more credit

Jordan recorded 31 points, 9 assists, and 8 rebounds on 48% FG in that Game 7. Now I realize that his efficiency in that game wasn't quite up to his standards, but here's what his teammates did in that game:

Pippen: 2 points, 2 assists, and 4 rebounds on 1/10 shooting with a migraine
Paxson: Didn't play at all due to an ankle injury
Grant: 3/17 from the field(albeit he did grab 14 boards, so he contributed in the way)
BJ Armstrong: 1/8 FG
Craig Hodges: 3/13 FG

It just seems nuts to me to blame MJ for that when his whole team laid an egg outside of Grant's rebounding.

Okay, but Jordan doesn't need to win game 7 to win the series. Here was the box-score for their closest loss:
https://www.basketball-reference.com/boxscores/199005220DET.html

Also, I mostly covered this at the top, but just because someone seems to be doing "everything possible" in the box-score doe snot mean they are actually doing everything possible. Jordan was not the reason the Bulls held the pistons offense 5 points under their usual performance. Jordan also was not running the team on either end, and I'd guess the main reason his numbers improve between the 89 series and the 90 one is because he is facing way less defensive attention thanks to Pippen operating as the primary ball-handler. Jordan is basically being allowed to shoot a bunch in single coverage, he's being given fantastic defensive support, and I do think it would be fair to argue that some of his teammates ppg/ts/fg% ends up looking better if you replace him with a better passer:

Spoiler:

We can lower the bar of "goat-lvl" to include these types of performances, but that's really besides the point. As is, Jordan's in a bit of a bind, because we know there's a player who can match what he does in the box-score(even facing more defensive attention)...
https://www.basketball-reference.com/playoffs/2009-nba-eastern-conference-finals-magic-vs-cavaliers.html
...while also doing a lot more beyond the slashlines.

Here was Lebron's worst game that playoff-run...
https://www.basketball-reference.com/boxscores/200905300ORL.html

We can blame or not blame but ultimately, if you're going to use the word "goat", I think "greatest" needs to be at least be viable. Even if you want to dismiss the Russells and Magics because of uncertainty and being a different type of player, we have a crystal clear example where a guy did more just as efficiently. I'd say that precludes "goat-lvl" as a description.

MrVorp · Post #24 » by **MrVorp** » Sun Aug 13, 2023 12:33 pm

MrVorp wrote:Yeah, the mental gymnastics that it takes to make Jordan’s peak as anything other than GOAT like is pretty funny. As for the WOWY stuff, a time decayed version (which fixes a lot of the issues of smoothing out large time spans) ran by the creator of DARKO has Jordan peaking as the highest right at his first retirement.

Honestly would want to know how RAPM(which we don't have pre-97) is being used here(with WOWY vs WOWY it swings strongly in Lebron's favor) but regardless, while Darko's a great metric, it's important to understand how it works:

Spoiler:

OhayoKD wrote:

So it isn't spiking or deflating player performance by checking how players compare relative to expected trajectories?

If I've counted correctly, Jordan trails Magic's trajectory through the end of 1990 and matches a second Magic peak(somehwere between 87-90) as of 1993.

That big spike comes during the second-three-peat where Jordan is aged 32-34(magic was forced to retire at 29 and came back one season at 36).

Now you can go with 96-98 Mj just being >>>>> "prime" MJ(in which case we actually have data-ball for MJ's "peak" which now doesn't look that great), the reason Jordan is spiking in his 30's is because he was unusually good at 30+ and not so unusually good at what would be his typical peak. Actually Magic and Bird also benefit from this(notice how they go up as they age?) though they for different reasons they are not participating during the height of expansion where 30+ players were peaking

You can also see this with Lebron whose trajectory is well ahead of the field during his mid-late 20's is mysteriously upsurped by 30+ Jordan, and then reaches newer heights than everyone else in his 30's. Is this because 30+ Lebron was more valuable than mid-20's Lebron? No. It's because Lebron was a bigger outlier relative to other players.

Additionally because you are relying strictly on WOWY, players are at the mercy of timing. Duncan barely misses games up until 2004 so he doesn't do too well higher. But if we swtich to Darko DPM, Duncan is kicking everyone's ass because Darko now has access to that early stretch where Duncan was already one of the best players ever by the age of 22. The only player who tops Duncan's early trajectory is Lebron but then Lebron has 2011 at an age where Darko expects you to peak and Lebron's trajectory collapses to 2nd best behind the big fundamental.

Darko is a great stat which is very good at what it does. But you have to understand what it does to use it properly. Darko suggests that Jordan was crazy good for a 30+ guy during the second-three peat. It also suggests he was good but not crazy good for a guy in his 20's/mid-20's during the 80's and early 90's.

I know how the stat works, which is why I'm not just looking at which lines are higher on the chart. Lebron at game 1600 was almost certainly not a better player than peak Micheal. And while there's certainly evidence to suggest the second three-peat might have been Jordan's apex in terms of "era-relative situational impact", he was not a waaaay better there than he was in the years some consider the best ever.[/quote]

Jordan skyrocketing during the second threepeat should be a bit of a tell here. While there's a decent chance that would be his impact peak anyway(in which case various metrics do not paint him particularly close to "goat lvl"), what Darko is really assessing here is how MJ compares relative to other players at that point in their trajectory. Jordan isn't looking that hot when he was, in an absolute sense, "better", but expansion offers him aid other players in these graphs do not get(magic retires, bird is injured, Lebron isn't alive), and Jordan still seems to fall short when the same data is used for everyone even with the likes of Kareem, Russell, and Darko king Duncan excluded. HCL covered a bunch of the rest, and I think at this point most regulars here have some awareness of the impact stuff so let's end this with some granular analysis...
[quote="OldSchoolNoBull"][quote="OhayoKD"][quote="OldSchoolNoBull"]I don't think MJ was the issue.

Yeah that’s not how the metric I displayed works. Think WOWY RAPM as a ridge regressed version of WOWY data. It’s also time decayed, meaning it weighs older games less. That’s different than the DARKO model itself, which is a machine learning model that combines a bunch of projected statistics, age, on/off data, etc. The reason it likes Duncan so much is because he had a really high time decay RAPM and the model is likely overfit to him. Also I’ve noticed that these WOWY type stats, even the ones that try to account for teammates have only so much granularity that a player’s score is strongly correlated with his team’s SRS. So I think the most likely reason it loves second three peat Jordan is because of the dominant SRS of the Bulls at that time. Jordan’s RAPM at the end of his Bulls career is basically within a standard error of peak Lebron, Duncan, Garnett anyway.

homecourtloss · Post #25 » by **homecourtloss** » Sun Aug 13, 2023 5:11 pm

lessthanjake wrote:
homecourtloss wrote:
I didn’t say that Jordan’s impact data looks anything but great and is in the discussion for possible GOATdom.

I said that the idea that his peak of 1991 and the surrounding years is some type of outlier mythological monster isn’t supported by any of the data at all. Everything you just listed is very nice and good, but it’s type of data that other players have done similarly or in the case of someone like LeBron, even better since we have RAPM numbers.

But that chart is just taking the absolute best +/- periods for other players over relatively short timespans in a low-sample-size measure. Jordan’s best playoff on-off years aren’t all distributed next to each other. When you go to larger sample sizes—which is obviously an important and superior thing to do with playoff on/off data since it’s all low sample size in general and we want to look at as large a sample as we can—we see that in the data we have over 9 playoffs, Jordan has about a +16.0 on-off per 100 possessions overall. For reference, the best 9-playoff span LeBron has in playoff on-off is +13.9, and it was +10.4 in his prime years (defining 2009-2020 as his prime) and +10.2 for his career. Of course, we don’t have *all* the years for Jordan, so we can’t make a full comparison, but it certainly looks very GOAT-like for Jordan when we actually properly look at all the data we have.

Again, Jordan looks good but nothing outlier or anything that seems to show consensus GOAT peak. Also, not sure if you’re trying to apply BKREF’s pace estimates to eyeball estimates of a graph or not here. Pace probably does favor 1990 Jordan compared to 2017 Lebron for single, but we have surety of LeBron’s numbers while it’s fuzzy for Jordan’s. I’m assuming you’re using DraymondGold’s numbers, which uses some average which we discussed but since it’s being used:

Jordan's best 1-year average: +33.5 or 33.75 (1990; +24 in 1989)
Jordan's best 2-year average: +28.7 (1989-1990)
Jordan's best 3-year average: +23.4 (1989-1991)
Jordan's best 4-year average: +20.2 (1988-1991)
Jordan's best 5-year average: +18.5 (1989-1993)
Jordan's best 6-year average: +14.4 (1988-1993)

Then, he discounts 1995:

Jordan's best 7-year average:+16.6 (1989-1997 discounting 1995 which has greater uncertainty; it's +12.3 to +12.9 ish if we include 1995)
Jordan's best 8-year average: +16.0 (1989-1998 discounting 1995 which has greater uncertainty; it's +13.6 to +14.2 ish if we include 1995)
Jordan's best 9-year average: +15.2 (1988-1998 discounting 1995 which has greater uncertainty; it's +13.5 to +13.9 ish if we include 1995)

I’m not sure whether he is counting 1995 in an existing range or instead of a border year on that range. I recall he said there is no 10-year number when literally speaking there should be one, so either he skipped the 7-year average or he skipped the 10 year average.

Lebron’s 2017 postseason on/off per 48 minutes is +33.4
Lebron’s 2016/17 postseason (2-year) on/off per 48 minutes is +24.9, and he also has a smaller sample stretch of +25.8 in 2007/08.
Lebron’s 2007-10 postseason (4-year) on/off per 48 minutes is +21
Lebron’s 2015-21 postseason (6-year) on/off per 48 minutes is +16, and his 2007-12 stretch is +15.4
Lebron’s 2014-21 postseason (7-year) on/off per 48 minutes is +14.9
Lebron’s 2012-20 postseason (8-year) on/off per 48 minutes is +13.1
Lebron’s 2012-21 postseason (9-year) on/off per 48 minutes is +14.4
Lebron’s 2007-21 postseason (14-year) on/off per 48 minutes is +13.8.

Lebron’s 9-average is better than Jordan’s best 6-year average, and his 14-year average is close to that 6-year average. Nothing stands out here really.

lessthanjake wrote:Did Jordan play a ball-dominant heliocentric style? No, not really. All but two of the years we have data for here are in the triangle, so I don’t see your point. The same factor pretty obviously doesn’t apply.

Firstly, LeBron doesn’t play the exaggerated ball dominant game you speak of. Secondly, we have seen that true ball dominance (something like what Luka or Harden play) didn’t have that effect on offense. It’s easy to just basically say “Jordan’s on/off is so high because of how good Jordan was and how poor his teammates played while LeBron’s is so high because of ball domiance.” Where’s the proof?

Here is chart comparing LeBron’s, Luka’s, and Harden’s respective avg. time touching the ball and dribbles per touch. Harden and Luka are routinely in the top 10. James is nowhere to be found atop these lists.

@letskissbro gave a little snippet of some of these other players and their respective teams

letskissbro wrote:Both Luka Doncic and James Harden have actually played under the exaggerated caricature that people have made "LeBron ball" out to be, yet their teams fare significantly better when they're off the court than LeBron's teams and often field positive net ratings with their stars on the bench.

LeBron's average times of possession in that 16-21 stretch from Ben's graph (minus 2019): 5.3, 6.4, 6.7, 7.4, 6.4
Corresponding net ratings for those teams while he sat: -4.3, -8.8, -0.5, -0.9, -2.0

Luka (20-23): 8.9, 8.9, 9.3, 9.1
Off NetRtg: +4.2, +0.3, +3.4, -2.7

Harden (17-23): 9.3, 8.8, 9.3, 8.6, 8.6, 9.2, 8.6
Off NetRtg: +3.7, +5.2, +1.1, -3.4, +2.4, +1.2, +2.8

If LeBron spends approximately six seconds with the ball in his hands, and that leads to over-reliance on him, why is it that Harden and Luka, who hold the ball for 50% longer, dribble the ball significantly more per touch, play under a similar team setup, and have comparable usage rates, see their teams perform well in their absence?

Most teams see their offensive rating plummet when their offensive superstar goes to the bench. People just selectively choose language that really just projects their aesthetic biases through their analysis. When it's a shoot first guard like Steph Curry or Michael Jordan they're celebrated: "Look at how impactful their ceiling raising is!". When it's LeBron James, it's "Well he takes his teammates out of rhythm."

The more plausible, albeit uncomfortable for some, explanation is that LeBron's unique combination of volume scoring + playmaking + all-time wing defense provides a greater lift on both ends of the court compared to Jordan, allowing him to take worse rosters to higher highs. The only caveat being that he couldn't do it for 48 minutes a game.

We have tracking data for ball dominance going back to the 2013-2014 season.

1) First the numbers that discuss how much LeBron and Kyrie dribbled the ball and how long they held on to it:

2) Next, since the beginning of the 2013-2014 season, the LESS ball dominant LeBron was, the GREATER the ON/OFF ORtg became. The argument is always that “he holds the ball and controls everything, so when he goes to the bench players are out of rhythm and this creates the crazy ON-OFF numbers.”

You can see all the numbers here (note how ball dominant LeBron was in 2020 curiously the Off ORtg didn’t plummet:

[3)We have LeBron’s playoff ball dominance data that OhayoKD and others posts in this tweet:

OhayoKD wrote:.

Read on Twitter

Here is the playoffs and ball dominance data (I need to update for 2023 playoffs which will make the trend line about neutral, maybe slightly positive:

Here is the playoffs data table:

Strange this entire “LeBron’s teams struggle without him because he controls everything” only applies to him. Bulls offense went in the toilet in the playoffs without Jordan but that’s because of “ressons.”

When LeBon’s teammates play well and make shots, the argument goes away. One of the biggest reasons the Lakers were so dominant in the playoffs in 2020 was because they played well with LeBron off court in non-garbage time minutes and LeBron in 2020 controlled the ball more as much as he ever has. Why did they do well? Because they made shots, especially AD being fed by Rondo. The offense was really good without LeBron on court in a season in which LeBron controlled the ball as much as ever.

2020 playoffs, Lebron off, AD ON
ORTG 118.5
DRTG 113.0
+5.5

With other minutes, mostly garbage time minutes, the Lakers were a negative with LeBron off, but the argument that the team struggles “because he controls everything” is not factually supported.

Bulls without Jordan in 1997 playoffs [numbers behind BKREF paywall now]:
ORTG with Jordan, 108.3
ORTG without Jordan, 96.6
Overall without Jordan, -14.8

Bulls without Jordan in 1998 playoffs:

ORTG with Jordan, 110.2
ORTG without Jordan, 95.7
Overall without Jordan, -13.1

There are multiple other examples too numerous to write down all here but it’s only James who somehow gets blamed for this. Is Shaq ball dominant who controls everything? Why wasn’t Kobe scoring when Shaq was off court? Was it because “Shaq controlled everything”?

Lakers’ playoffs ORTG without Shaq
1997, -11.9
1998, -9.8
1999, -6.9
2000, -13.9
2001, +5.3
2002, -22.1
2003, -16.0
2004, -22.8

Wade, Kyrie, Love, and his other supposed “super teammates” didn’t do much of anything with him off court even though they were running their own sets they were used to all season long, many of them ISO in which they’re good scorers but failed in the playoffs in limited possessions.

Respective Playoff ORTGs without James or Jordan or Shaq [using BKREF numbers for all for sake of ease]

BTW, Cavs had the drop off because the Cavs offenses had such high ceilings WITH LeBron and were better on offense than the Bulls were with Jordan even relative to league averages during the playoffs. If we broke it down by individual defenses faced, took out the Cavs/Heat/Lakers offenses from the league averages, it’s even more skewed in favor of James for the available data for here it is unadjusted:

1997 Bulls offense with Jordan: 108.3 [+.9 rORtg]
1998 Bulls: 110.2 [+4.6 rORtg]
2006 Cavs: 104.0 [-4.2 rORtg]
2007 Cavs: 103.5 [-1.4 rORtg]
2008 Cavs: 106.6 [-.8 rORtg]
2009 Cavs: 115.1 [+7.4 rORtg]
2010 Cavs: 110.1 [+1.5 rORtg]
2011 Heat: 107.5 [+1.5 rORtg]
2012 Heat: 111.9 [+8.3 rORtg]
2013 Heat: 111.9 [+7.1 rORtg]
2014 Heat: 112.7 [+4.0 rORtg]
2015 Cavs: 107.3 [+2.0 rORtg]
2016 Cavs: 118.2 [+11.5 rORtg]
2017 Cavs: 124.0 [+12.7 rORtg]
2018 Cavs: 111.9 [+3.0 rORtg ]
2020 Lakers: 118.3 [+7.0 rORtg]

Now look at some of these offenses without James, Shaq, and Jordan in the playoffs:

2021 Lakers, -30.5 without James
2004, Lakers -22.8 without Shaq
2008 Cavs, -22.2 without James [limited off court minutes since James basically played entire games]
2002 Lakers, -22.1 without Shaq
2017 Cavs, -19.9 without James
2003 Lakers, -16.0 without Shaq
2012 Heat, -15.0 without James
1998 Bulls, -14.5 without Jordan
2010 Cavs, -14.1 without James [limited off court minutes]
2018 Cavs, -14.1 without James
2000 Lakers, -13.9 without Shaq
2016 Cavs, -12.6 without James
1997 Lakers, -11.9 without Shaq
1997 Bulls, -11.7 without Jordan
2007 Cavs, -10.3 without James [limited off court minutes]
1998 Lakers, -9.8 without Shaq
2013 Heat, -8.7 without James
2009 Cavs, -7.8 without James
1999 Lakers, -6.9 without Shaq
2011 Heat, -5.2 without James
2020 Lakers, -4.6 without James
2015 Cavs, -2.5 without James
2014 Heat, +2.0 without James [collapsed defensively]
2001 Lakers, +5.3 without Shaq [Kobe played great and role players made shots]

lessthanjake wrote:I’m really not seeing how being #1 by a significant margin in every season there’s data on except his rookie year could in any way be described as “not approach[ing] anything close to an outlier mythological peak that is above everyone else’s.” In the Squared data, Jordan has literally got an outlier RAPM that is above everyone else’s in every year except his rookie year!

And that includes Magic Johnson. Saying Magic “looks every bit as good” is just incorrect. They overlapped in three years in the Squared data. Magic is above Jordan in Jordan’s rookie season. But then Jordan is well above Magic in the other two seasons where they’re both actually in their prime. And to the extent you’re saying “peaks higher” by comparing RAPM values in different years, let’s remember that we’ve agreed at your insistence that that type of cross-year comparison of RAPM values is invalid.

Define “significant.” Is the gap bigger than the gap for others in other years? 1991 Jordan didn’t have an outlier gap. 1991 Jordan is a modest plus on defense (a modest negstive in 85 and 88). There’s nothing outlier here when Magic looks just as good.

lessthanjake wrote: The thing is that Jordan doesn’t need to have had *more* impact in his earlier years than in the second three-peat for him to have been GOAT-level, because the impact in the second three-peat is itself GOAT-level overall (with that last season maybe being a shade off that, but the overall three-year span being GOAT-level impact for sure). So basically, I agree we don’t have enough info to really know for sure exactly how his earlier impact shaped up compared to the later impact, but the signals for all of it that we have is really really great and entirely consistent with a GOAT-level player, so I don’t really see the point of parsing through limited data trying to determine which of his time periods were more or less impactful. It looks like it was all incredibly impactful!

If the consensus is that 1991 is his peak and 1990 is peak adjacent and that 1988-1993 is a GOAT 6 year stretch, well, it would be nice to see that that period had an outlier level of impact. But it doesn’t.

lessthanjake wrote: And no, I don’t really care what the output of the data is—there’s just obvious limitations and potential distortions of the data that shouldn’t require any motivation to notice.

Of course you don’t because Jordan does not look to have GOAT impact in those regressions (more below). Other WoWY data you meet with enthusiasm (such as Draymond’s) but from the beginning you have been trying to find ways to caution this, keep in mind, etc., which is true of data but you seem keen on cautioning this data since Moon’s data doesn’t seem to corroborate any outlier GOAT peak or consensus GOAT impact at all compared to others even when we have plenty of OFF data in these early years,

lessthanjake wrote: Again, though, Jordan’s “peak” cannot really be properly measured in Moonbeam’s data, because the three best players on his team (Jordan included) barely missed games during that timeframe, so the model has virtually no way of figuring out who was having what impact between them. It can just try to estimate the role players’ impact based on their missed time (and sometimes do that in a way that ends up being flawed, see, e.g., BJ Armstrong), and then make a rough guess at how to distribute the rest of the impact between Jordan, Pippen, and Grant, using the extremely limited number of missed games that each one had. Once the data actually saw what the Bulls did without Jordan, he quickly catapults to the top, with no player with a substantial playtime throughout the time period being ahead of him for the next 5 time periods starting with 1991-1995 (said based on Ridge version). This isn’t really a picture of a guy who was less impactful in those earlier years. It’s a situation where the model doesn’t have the data to isolate out the impact of the top few guys on the team, and then once the model gets more data to do that with (which also includes more data from Grant leaving Chicago), Jordan catapults to the top of the league.

It doesn’t show that he is an outlier impact king either—in fact, it hints that his impact peak wasn’t during the 1988-1993 period. If we had lineup data we’d have a better idea but we don’t but there are players who played all years in these 5 year segments who come out looking better EVEN though we have big chunks of off data. Also, all the limitations of the data you mention apply to Magic Johnson as well and he looks like an impact god.

Moonbeam’s RWOWY for Jordan (Jordan in the Ridge set, Jordan in the LASSO set)

This was a cursory count, and others can do their own, but I took out every player who only played one year in the data segments unless stated or the player in question, and I also took out every player who didn’t at least play solid rotational minutes for two of the years in the data segments (1,000+ minutes a season in at least two seasons).

1981-1985: 4th, 3rd
1982-1986: 12th, 20th [here you have a large off segment]
1983-1987: 16th, 30th [Again a large of segment]
1984-1988: 30th, 21st [and again]
1985-1989: 40th, 17th [and again]
1986-1990: 65th, 27th [And again, including 1990, close to his consensus peak]
1987-1991: 56th, 52nd [Includes his peak and 1990]
1988-1992: 23rd, 7th
1989-1993: 15th, 5th
1990-1994: 7th, 5th
1991-1995: 2nd, 4th
1992-1996: 1st, 1st
1993-1997: 1st, 2nd
1994-1998: 2nd, 3rd
1995-1999: 1st, 2nd
1996-2000: 9th, not listed
1997-2001: 3rd, 4th
1998-2002: 6th, 18th
1999-2003: not listed, not listed
2000-2004: 86th, not listed [note: did not check for players were only played one year or non-rotational minutes in this segment]
2001-2005: not listed, not listed
2002-2006: not listed, not listed
2003-2007: not listed, not listed

Moonbeam’s RWOWY for Magic (Magic the Ridge set, Magic in the LASSO set)

1976-1980: 7th, 6th [counting one year of Bird]
1977-1981: 3rd, 7th
1978-1982: 2nd, 3rd
1979-1983: 1st, 2nd
1980-1984: 1st, 3rd
1981-1985: 2nd, 3rd
1982-1986: 3rd, 2nd
1983-1987: 4th, 4th
1984-1988: 4th, 2nd
1985-1989: 3rd, 1st
1986-1990: 3rd, 1st
1987-1991: 5th, 1st
1988-1992: 4th, 2nd
1989-1993: 2nd, 1st
1990-1994: 1st, 2nd
1991-1999: Magic has only one year samples (i.e., 1991 and 1996) but he’s basically
top 5 in Ridge in every segment of 1991-1995, 1992-1996, 1993-1997, 1994-1998, 1995-1999

So basically top 5 every year which is beyond ridiculous in a pure, non-prior informed set. :lol:

This is what one would expect Jordan’s peak years to look like given the mythological and unassailable status it has reached.

70sFan · Post #26 » by **70sFan** » Sun Aug 13, 2023 6:01 pm

All this long discussion can be summed up quickly - there is no clear GOAT peak in the NBA history. I know it hurts Jordan or LeBron fans, but we don't have any clear proof that one players reached a level head and shoulders ahead of anyone else ever. Even players like Russell and Mikan have question marks, let alone Jordan.

lessthanjake · Post #27 » by **lessthanjake** » Sun Aug 13, 2023 7:22 pm

homecourtloss wrote:Again, Jordan looks good but nothing outlier or anything that seems to show consensus GOAT peak. Also, not sure if you’re trying to apply BKREF’s pace estimates to eyeball estimates of a graph or not here. Pace probably does favor 1990 Jordan compared to 2017 Lebron for single, but we have surety of LeBron’s numbers while it’s fuzzy for Jordan’s. I’m assuming you’re using DraymondGold’s numbers, which uses some average which we discussed but since it’s being used:

Jordan's best 1-year average: +33.5 or 33.75 (1990; +24 in 1989)
Jordan's best 2-year average: +28.7 (1989-1990)
Jordan's best 3-year average: +23.4 (1989-1991)
Jordan's best 4-year average: +20.2 (1988-1991)
Jordan's best 5-year average: +18.5 (1989-1993)
Jordan's best 6-year average: +14.4 (1988-1993)

Then, he discounts 1995:

Jordan's best 7-year average:+16.6 (1989-1997 discounting 1995 which has greater uncertainty; it's +12.3 to +12.9 ish if we include 1995)
Jordan's best 8-year average: +16.0 (1989-1998 discounting 1995 which has greater uncertainty; it's +13.6 to +14.2 ish if we include 1995)
Jordan's best 9-year average: +15.2 (1988-1998 discounting 1995 which has greater uncertainty; it's +13.5 to +13.9 ish if we include 1995)

I’m not sure whether he is counting 1995 in an existing range or instead of a border year on that range. I recall he said there is no 10-year number when literally speaking there should be one, so either he skipped the 7-year average or he skipped the 10 year average.

Lebron’s 2017 postseason on/off per 48 minutes is +33.4
Lebron’s 2016/17 postseason (2-year) on/off per 48 minutes is +24.9, and he also has a smaller sample stretch of +25.8 in 2007/08.
Lebron’s 2007-10 postseason (4-year) on/off per 48 minutes is +21
Lebron’s 2015-21 postseason (6-year) on/off per 48 minutes is +16, and his 2007-12 stretch is +15.4
Lebron’s 2014-21 postseason (7-year) on/off per 48 minutes is +14.9
Lebron’s 2012-20 postseason (8-year) on/off per 48 minutes is +13.1
Lebron’s 2012-21 postseason (9-year) on/off per 48 minutes is +14.4
Lebron’s 2007-21 postseason (14-year) on/off per 48 minutes is +13.8.

Lebron’s 9-average is better than Jordan’s best 6-year average, and his 14-year average is close to that 6-year average. Nothing stands out here really.

To begin with, I think it’s important to note that I’m not claiming Jordan has impact signals that are complete outliers as compared to other very top-tier all-time players. No one has that. So if your argument is that Jordan isn’t statistically at some mythical status that is just standard deviations above the rest of the pack in everything, then I agree with you. But I also think that that’s an artificial standard.

Anyways, I wasn’t using any information from DraymondGold. I don’t even know what you’re referring to from DraymondGold on this, since he’s not posted in this thread. I assume you’re referring to a past thread I’ve never seen. I did my own analysis. What I used in that analysis is exactly what I said I used—which is (1) the charts presented in the Thinking Basketball video (which can be eyeballed in terms of what the “on” and “off” values are for 1988-1993 & 1996, and I mentioned what I was eyeballing the values to be); (2) the actual minutes Jordan was on and off in each individual playoffs; (3) the Bulls’ playoff pace in those years; and (4) Jordan’s actual known playoff on-off numbers for 1997 and 1998. With that information, one can first easily derive a weighted average for both Jordan’s “on” and “off” values per 48 minutes for 1988-1993 + 1996, then adjust those to per 100 possessions using the Bulls’ playoff pace, and then do a weighted average for Jordan’s “on” and “off” values per 100 possessions for 1988-1993 + 1996-1998, by adding in the known 1997 and 1998 on-off per 100 possession numbers. It’s pretty simple math, with the only real flaws being (1) I’m eyeballing values on the Thinking Basketball charts, so they will be close but not exact; and (2) it’s possible that the pace in the “on” and “off” minutes differed a bit, which might make the pace adjustment from per 48 to per 100 possessions slightly off. These are likely only really minor differences though, so the number is almost certainly very close.

And that number is about +16.0 on-off per 100 possessions in the 9 playoffs we have data on from 1988-1993 + 1996-1998. That is an extremely high number, and is objectively substantially higher than any 9-playoff stretch for LeBron (and by enough of a margin that it’s IMO not plausible that the two very minor sources of error above could cause the difference).

You refer to 1995, but I’m not sure what you’re referring to there, since I don’t think we actually have 1995 playoff on-off data for Jordan. I think Ben Taylor *may* have tracked it but the video doesn’t really refer to it in any meaningful way so I don’t really know what the “on” and “off” values that year were. If we do actually have that data somewhere, then please refer me to it and I’m happy to incorporate it.

Spoiler:
lessthanjake wrote:Did Jordan play a ball-dominant heliocentric style? No, not really. All but two of the years we have data for here are in the triangle, so I don’t see your point. The same factor pretty obviously doesn’t apply.

Firstly, LeBron doesn’t play the exaggerated ball dominant game you speak of. Secondly, we have seen that true ball dominance (something like what Luka or Harden play) didn’t have that effect on offense. It’s easy to just basically say “Jordan’s on/off is so high because of how good Jordan was and how poor his teammates played while LeBron’s is so high because of ball domiance.” Where’s the proof?

Here is chart comparing LeBron’s, Luka’s, and Harden’s respective avg. time touching the ball and dribbles per touch. Harden and Luka are routinely in the top 10. James is now we’re to be found atop these lists.

letskissbro wrote:Both Luka Doncic and James Harden have actually played under the exaggerated caricature that people have made "LeBron ball" out to be, yet their teams fare significantly better when they're off the court than LeBron's teams and often field positive net ratings with their stars on the bench.

LeBron's average times of possession in that 16-21 stretch from Ben's graph (minus 2019): 5.3, 6.4, 6.7, 7.4, 6.4
Corresponding net ratings for those teams while he sat: -4.3, -8.8, -0.5, -0.9, -2.0

Luka (20-23): 8.9, 8.9, 9.3, 9.1
Off NetRtg: +4.2, +0.3, +3.4, -2.7

Harden (17-23): 9.3, 8.8, 9.3, 8.6, 8.6, 9.2, 8.6
Off NetRtg: +3.7, +5.2, +1.1, -3.4, +2.4, +1.2, +2.8

If LeBron spends approximately six seconds with the ball in his hands, and that leads to over-reliance on him, why is it that Harden and Luka, who hold the ball for 50% longer, dribble the ball significantly more per touch, play under a similar team setup, and have comparable usage rates, see their teams perform well in their absence?

Most teams see their offensive rating plummet when their offensive superstar goes to the bench. People just selectively choose language that really just projects their aesthetic biases through their analysis. When it's a shoot first guard like Steph Curry or Michael Jordan they're celebrated: "Look at how impactful their ceiling raising is!". When it's LeBron James, it's "Well he takes his teammates out of rhythm."

The more plausible, albeit uncomfortable for some, explanation is that LeBron's unique combination of volume scoring + playmaking + all-time wing defense provides a greater lift on both ends of the court compared to Jordan, allowing him to take worse rosters to higher highs. The only caveat being that he couldn't do it for 48 minutes a game.

We have tracking data for ball dominance going back to the 2013-2014 season.

1) First the numbers that discuss how much LeBron and Kyrie dribbled the ball and how long they held on to it:

2) Next, since the beginning of the 2013-2014 season, the LESS ball dominant LeBron was, the GREATER the ON/OFF ORtg became. The argument is always that “he holds the ball and controls everything, so when he goes to the bench players are out of rhythm and this creates the crazy ON-OFF numbers.”

You can see all the numbers here:

[3)We have LeBron’s playoff ball dominance data that OhayoKD and others posts in this tweet:

OhayoKD wrote:.

Read on Twitter

Here is the playoffs and ball dominance data (I need to update for 2023 playoffs which will make the trend line about neutral, maybe slightly positive:

Here is the playoffs data table:

Strange this entire “LeBron’s teams struggle without him because he controls everything” only applies to him. Bulls offense went in the toilet in the playoffs without Jordan but that’s because of “ressons.”

I think those charts with a trend line for on-off and ball dominance % are pretty meaningless. There’s literally only a handful of data points, the correlation is really low, and the trendline isn’t steep. And the underlying dependent variable there (i.e. on-off) is subject to so many other variables that we’d just expect it to be super random (especially when we’re talking about different years with different teams, so it’s just completely different players playing in the “on” and “off” minutes in the different data points). So I don’t really think much value can be derived from that at all. It’s pretty obviously just noise.

Of course, I’m sure you’ll want to refer back to the charts I provided on Bill Walton with the 1986 Celtics. And it’d be a fair point. But I’d note that (1) those charts had substantially more data points, (2) I was presenting them to refute a pretty hyperbolic point (i.e. that Walton was the most important part of the 1986 Celtics defense) while you’re presenting these to refute me merely having said something is a factor rather than that it is the biggest factor; (3) the underlying data points there weren’t from different years and with different teams, which makes these data points even sillier to try to draw a trend line between—it’d be like if I did what I did with Walton but included prior years’ games with other teams in the analysis.

As for the thing about Luka and Harden, I’ve addressed that to you and others multiple times now. I’ve demonstrated that those guys’ teams have often had very different lineup strategies, where their teams’ other stars (Chris Paul, Brunson, etc.) have been substantially more staggered, which obviously would have a huge effect on their on-off. No one is suggesting that this is some effect that completely swamps all the other many things that affect on-off numbers.

I’d also repeat that my point is legitimately based on hearing NBA role players specifically say this. You and others keep acting like it’s crazy, but I put more credence in what I’ve heard actual NBA players say than what homecourtloss says.

In any event, this is all plainly off topic and derives from you bringing up subject matter from months ago that may still irk you but that has no actual relevance to this thread, so I don’t think I’ll continue down this irrelevant path moving forward.

Bulls without Jordan in 1997 playoffs [numbers behind BKREF paywall now]:

This is not a substantive point, but just FYI for you and others, you can actually still get to basketball-reference on-off numbers without going through a paywall. If you’re on a player’s page, go to the URL, delete the “.html” part at the end, and add a “/on-off/YEAR”. So, for instance, if I wanted to see Jordan’s 1997 on-off numbers, I can go to: https://www.basketball-reference.com/players/j/jordami01/on-off/1997

lessthanjake wrote:I’m really not seeing how being #1 by a significant margin in every season there’s data on except his rookie year could in any way be described as “not approach[ing] anything close to an outlier mythological peak that is above everyone else’s.” In the Squared data, Jordan has literally got an outlier RAPM that is above everyone else’s in every year except his rookie year!

And that includes Magic Johnson. Saying Magic “looks every bit as good” is just incorrect. They overlapped in three years in the Squared data. Magic is above Jordan in Jordan’s rookie season. But then Jordan is well above Magic in the other two seasons where they’re both actually in their prime. And to the extent you’re saying “peaks higher” by comparing RAPM values in different years, let’s remember that we’ve agreed at your insistence that that type of cross-year comparison of RAPM values is invalid.

Define “significant.” Is the gap bigger than the gap for others in other years? 1991 Jordan didn’t have an outlier gap. 1991 Jordan is a modest plus on defense (a modest negstive in 85 and 88). There’s nothing outlier here when Magic looks just as good.

I’m not understanding what you’re saying here. There’s six years of the Squared RAPM data that I’m aware of: 1969-1970, 1979-1980, 1984-1985, 1987-1988, 1990-1991, and 1995-1996. The first two years have so little data at the moment that the numbers are completely meaningless (they’re literally from tracking of 15-20 total games across the entire league, and you can tell from the possession numbers that it’s like 1-4 games for each player). Meanwhile, Michael Jordan is #1 in the last three of those years. The only year with any real sample size at all that Jordan isn’t #1 in is the 1984-1985 season. There, Magic Johnson is ahead by 0.47 points (or a 5.6% increase) over the #2 player. In 1987-1988, Michael Jordan is ahead by 0.85 points (or a 12.8% increase) over the #2 player. In 1990-1991, Michael Jordan is ahead by 1.45 points (or a 29.3% increase) over the #2 player. In 1995-1996, Michael Jordan is ahead by 0.78 points (or a 12.7% increase) over the #2 player. In other words, Michael Jordan is #1 in Squared RAPM in all but his rookie season, and in every single one of those seasons the gap he has over #2 is larger than the gap Magic has in the one year Magic is on top. There is no argument that “Magic looks just as good” in terms of having a “gap” over others. And Jordan is ahead in 3 of the 4 years there’s any real data on. Michael Jordan is the clear standout in Squared’s RAPM data.

Of course, one could just say that Squared’s data is too low sample size to draw much of a concrete conclusion from. And you could object to the idea that comparing the size of a gap across years is invalid (though you’re the one that suggested looking at that). That would all be valid. But what is not valid is to try to argue that Squared’s RAPM data actively goes against Jordan’s GOAT case. We can debate how much (if any) value to put on Squared’s data, but it does *not* go against Jordan’s GOAT case.

lessthanjake wrote: The thing is that Jordan doesn’t need to have had *more* impact in his earlier years than in the second three-peat for him to have been GOAT-level, because the impact in the second three-peat is itself GOAT-level overall (with that last season maybe being a shade off that, but the overall three-year span being GOAT-level impact for sure). So basically, I agree we don’t have enough info to really know for sure exactly how his earlier impact shaped up compared to the later impact, but the signals for all of it that we have is really really great and entirely consistent with a GOAT-level player, so I don’t really see the point of parsing through limited data trying to determine which of his time periods were more or less impactful. It looks like it was all incredibly impactful!

If the consensus is that 1991 is his peak and 1990 is peak adjacent and that 1988-1993 is a GOAT 6 year stretch, well, it would be nice to see that that period had an outlier level of impact. But it doesn’t.

Outlier compared to what? We know he had about a +14.7 playoff on-off in the six years from 1988-1993 (see my prior posts with that analysis). That’s GOAT-like. For reference, LeBron does not have any 6-year span where he actually made the playoffs every year and had as high a playoff on-off as that (his best such span is +13.9). LeBron does have a 6-playoff span from 2015-2021 (6 playoffs over 7 years, since he missed the playoffs one of those years) that’s slightly above that (it’s +15.5). So it’s basically a wash overall, where peak Jordan and peak LeBron look very comparable in this regard. I’m failing to see how peak Jordan’s playoff impact looking like peak LeBron’s playoff impact somehow suggests peak Jordan’s impact wasn’t GOAT-like. (And of course this isn’t even factoring in that Jordan’s playoff on-off in the second three-peat is even higher).

lessthanjake wrote: And no, I don’t really care what the output of the data is—there’s just obvious limitations and potential distortions of the data that shouldn’t require any motivation to notice.

Of course you don’t because Jordan does not look to have GOAT impact in those regressions (more below). Other WoWY data you meet with enthusiasm (such as Draymond’s) but from the beginning you have been trying to find ways to caution this, keep in mind, etc., which is true of data but you seem keen on cautioning this data since Moon’s data doesn’t seem to corroborate any outlier GOAT peak or consensus GOAT impact at all compared to others even when we have plenty of OFF data in these early years,

No, I’m cautioning the data because there’s obvious limitations that need to be acknowledged, rather than looking at the data and not thinking critically about what it means and what issues there are. It’s just bad analysis to not think about these things. I’m very consistent about this stuff—which is why others have criticized me before for constantly saying that we should assume that every metric we have is substantially flawed (which I did even while arguing for a guy like Steph, who would probably *benefit* from blind acceptance of what impact metrics say). There’s serious methodological flaws/issues/limitations with everything, and I don’t think Moonbeam himself would claim that that doesn’t apply to his model.

lessthanjake wrote: Again, though, Jordan’s “peak” cannot really be properly measured in Moonbeam’s data, because the three best players on his team (Jordan included) barely missed games during that timeframe, so the model has virtually no way of figuring out who was having what impact between them. It can just try to estimate the role players’ impact based on their missed time (and sometimes do that in a way that ends up being flawed, see, e.g., BJ Armstrong), and then make a rough guess at how to distribute the rest of the impact between Jordan, Pippen, and Grant, using the extremely limited number of missed games that each one had. Once the data actually saw what the Bulls did without Jordan, he quickly catapults to the top, with no player with a substantial playtime throughout the time period being ahead of him for the next 5 time periods starting with 1991-1995 (said based on Ridge version). This isn’t really a picture of a guy who was less impactful in those earlier years. It’s a situation where the model doesn’t have the data to isolate out the impact of the top few guys on the team, and then once the model gets more data to do that with (which also includes more data from Grant leaving Chicago), Jordan catapults to the top of the league.

It doesn’t show that he is an outlier impact king either—in fact, it hints that his impact peak wasn’t during the 1988-1993 period. If we had lineup data we’d have a better idea but we don’t but there are players who played all years in these 5 year segments who come out looking better EVEN though we have big chunks of off data. Also, all the limitations of the data you mention apply to Magic Johnson as well and he looks like an impact god.

Moonbeam’s RWOWY for Jordan (Jordan in the Ridge set, Jordan in the LASSO set)

This was a cursory count, and others can do their own, but I took out every player who only played one year in the data segments unless stated or the player in question, and I also took out every player who didn’t at least play solid rotational minutes for two of the years in the data segments (1,000+ minutes a season in at least two seasons).

1981-1985: 4th, 3rd
1982-1986: 12th, 20th [here you have a large off segment]
1983-1987: 16th, 30th [Again a large of segment]
1984-1988: 30th, 21st [and again]
1985-1989: 40th, 17th [and again]
1986-1990: 65th, 27th [And again, including 1990, close to his consensus peak]
1987-1991: 56th, 52nd [Includes his peak and 1990]
1988-1992: 23rd, 7th
1989-1993: 15th, 5th
1990-1994: 7th, 5th
1991-1995: 2nd, 4th
1992-1996: 1st, 1st
1993-1997: 1st, 2nd
1994-1998: 2nd, 3rd
1995-1999: 1st, 2nd
1996-2000: 9th, not listed
1997-2001: 3rd, 4th
1998-2002: 6th, 18th
1999-2003: not listed, not listed
2000-2004: 86th, not listed [note: did not check for players were only played one year or non-rotational minutes in this segment]
2001-2005: not listed, not listed
2002-2006: not listed, not listed
2003-2007: not listed, not listed

Moonbeam’s RWOWY for Magic (Magic the Ridge set, Magic in the LASSO set)

1976-1980: 7th, 6th [counting one year of Bird]
1977-1981: 3rd, 7th
1978-1982: 2nd, 3rd
1979-1983: 1st, 2nd
1980-1984: 1st, 3rd
1981-1985: 2nd, 3rd
1982-1986: 3rd, 2nd
1983-1987: 4th, 4th
1984-1988: 4th, 2nd
1985-1989: 3rd, 1st
1986-1990: 3rd, 1st
1987-1991: 5th, 1st
1988-1992: 4th, 2nd
1989-1993: 2nd, 1st
1990-1994: 1st, 2nd
1991-1999: Magic has only one year samples (i.e., 1991 and 1996) but he’s basically
top 5 in Ridge in every segment of 1991-1995, 1992-1996, 1993-1997, 1994-1998, 1995-1999

So basically top 5 every year which is beyond ridiculous in a pure, non-prior informed set. This is what one would expect Jordan’s peak years to look like given the mythological and unassailable status it has reached.

Again, you’re failing to acknowledge that the model has virtually no way of figuring out who was having what impact between a group of players when they all barely missed games during that timeframe. Given that Jordan barely missed games in his peak years and was playing with Pippen and Grant who also barely missed games in that timeframe, the model can’t really properly assess Jordan’s impact in those years. And once it got more “off” info on those players (i.e. from Jordan retiring and Grant leaving Chicago), Jordan quickly skyrocketed to the top of the league and stayed there for quite a while. And that's not even getting into things like the BJ Armstrong Problem (which pretty clearly unduly hurt the numbers for the actually highly impactful Bulls players like Jordan)

If you want to look at that and actually conclude that Jordan’s impact in his peak years1 wasn’t very high, then you can go ahead and do that I suppose. But I’m just pointing out that you’d be basing that conclusion off of results spit out when the model had plainly inadequate information with which to parse out Jordan’s individual impact and that once that same model got a bit more info with which to do so, the result looked great for Jordan.

homecourtloss · Post #28 » by **homecourtloss** » Sun Aug 13, 2023 7:27 pm

lessthanjake wrote:
homecourtloss wrote:Again, Jordan looks good but nothing outlier or anything that seems to show consensus GOAT peak. Also, not sure if you’re trying to apply BKREF’s pace estimates to eyeball estimates of a graph or not here. Pace probably does favor 1990 Jordan compared to 2017 Lebron for single, but we have surety of LeBron’s numbers while it’s fuzzy for Jordan’s. I’m assuming you’re using DraymondGold’s numbers, which uses some average which we discussed but since it’s being used:

Jordan's best 1-year average: +33.5 or 33.75 (1990; +24 in 1989)
Jordan's best 2-year average: +28.7 (1989-1990)
Jordan's best 3-year average: +23.4 (1989-1991)
Jordan's best 4-year average: +20.2 (1988-1991)
Jordan's best 5-year average: +18.5 (1989-1993)
Jordan's best 6-year average: +14.4 (1988-1993)

Then, he discounts 1995:

Jordan's best 7-year average:+16.6 (1989-1997 discounting 1995 which has greater uncertainty; it's +12.3 to +12.9 ish if we include 1995)
Jordan's best 8-year average: +16.0 (1989-1998 discounting 1995 which has greater uncertainty; it's +13.6 to +14.2 ish if we include 1995)
Jordan's best 9-year average: +15.2 (1988-1998 discounting 1995 which has greater uncertainty; it's +13.5 to +13.9 ish if we include 1995)

I’m not sure whether he is counting 1995 in an existing range or instead of a border year on that range. I recall he said there is no 10-year number when literally speaking there should be one, so either he skipped the 7-year average or he skipped the 10 year average.

Lebron’s 2017 postseason on/off per 48 minutes is +33.4
Lebron’s 2016/17 postseason (2-year) on/off per 48 minutes is +24.9, and he also has a smaller sample stretch of +25.8 in 2007/08.
Lebron’s 2007-10 postseason (4-year) on/off per 48 minutes is +21
Lebron’s 2015-21 postseason (6-year) on/off per 48 minutes is +16, and his 2007-12 stretch is +15.4
Lebron’s 2014-21 postseason (7-year) on/off per 48 minutes is +14.9
Lebron’s 2012-20 postseason (8-year) on/off per 48 minutes is +13.1
Lebron’s 2012-21 postseason (9-year) on/off per 48 minutes is +14.4
Lebron’s 2007-21 postseason (14-year) on/off per 48 minutes is +13.8.

Lebron’s 9-average is better than Jordan’s best 6-year average, and his 14-year average is close to that 6-year average. Nothing stands out here really.

To begin with, I think it’s important to note that I’m not claiming Jordan has impact signals that are complete outliers as compared to other very top-tier all-time players. No one has that. So if your argument is that Jordan isn’t statistically at some mythical status that is just standard deviations above the rest of the pack in everything, then I agree with you. But I also think that that’s an artificial standard.

Anyways, I wasn’t using any information from DraymondGold. I don’t even know what you’re referring to from DraymondGold on this, since he’s not posted in this thread. I assume you’re referring to a past thread I’ve never seen. I did my own analysis. What I used in that analysis is exactly what I said I used—which is (1) the charts presented in the Thinking Basketball video (which can be eyeballed in terms of what the “on” and “off” values are for 1988-1993 & 1996, and I mentioned what I was eyeballing the values to be); (2) the actual minutes Jordan was on and off in each individual playoffs; (3) the Bulls’ playoff pace in those years; and (4) Jordan’s actual known playoff on-off numbers for 1997 and 1998. With that information, one can first easily derive a weighted average for both Jordan’s “on” and “off” values per 48 minutes for 1988-1993 + 1996, then adjust those to per 100 possessions using the Bulls’ playoff pace, and then do a weighted average for Jordan’s “on” and “off” values per 100 possessions for 1988-1993 + 1996-1998, by adding in the known 1997 and 1998 on-off per 100 possession numbers. It’s pretty simple math, with the only real flaws being (1) I’m eyeballing values on the Thinking Basketball charts, so they will be close but not exact; and (2) it’s possible that the pace in the “on” and “off” minutes differed a bit, which might make the pace adjustment from per 48 to per 100 possessions slightly off. These are likely only really minor differences though, so the number is almost certainly very close.

And that number is about +16.0 on-off per 100 possessions in the 9 playoffs we have data on from 1988-1993 + 1996-1998. That is an extremely high number, and is objectively substantially higher than any 9-playoff stretch for LeBron (and by enough of a margin that it’s IMO not plausible that the two very minor sources of error above could cause the difference).

You refer to 1995, but I’m not sure what you’re referring to there, since I don’t think we actually have 1995 playoff on-off data for Jordan. I think Ben Taylor *may* have tracked it but the video doesn’t really refer to it in any meaningful way so I don’t really know what the “on” and “off” values that year were. If we do actually have that data somewhere, then please refer me to it and I’m happy to incorporate it.

Spoiler:
lessthanjake wrote:Did Jordan play a ball-dominant heliocentric style? No, not really. All but two of the years we have data for here are in the triangle, so I don’t see your point. The same factor pretty obviously doesn’t apply.

Firstly, LeBron doesn’t play the exaggerated ball dominant game you speak of. Secondly, we have seen that true ball dominance (something like what Luka or Harden play) didn’t have that effect on offense. It’s easy to just basically say “Jordan’s on/off is so high because of how good Jordan was and how poor his teammates played while LeBron’s is so high because of ball domiance.” Where’s the proof?

Here is chart comparing LeBron’s, Luka’s, and Harden’s respective avg. time touching the ball and dribbles per touch. Harden and Luka are routinely in the top 10. James is now we’re to be found atop these lists.

letskissbro wrote:Both Luka Doncic and James Harden have actually played under the exaggerated caricature that people have made "LeBron ball" out to be, yet their teams fare significantly better when they're off the court than LeBron's teams and often field positive net ratings with their stars on the bench.

LeBron's average times of possession in that 16-21 stretch from Ben's graph (minus 2019): 5.3, 6.4, 6.7, 7.4, 6.4
Corresponding net ratings for those teams while he sat: -4.3, -8.8, -0.5, -0.9, -2.0

Luka (20-23): 8.9, 8.9, 9.3, 9.1
Off NetRtg: +4.2, +0.3, +3.4, -2.7

Harden (17-23): 9.3, 8.8, 9.3, 8.6, 8.6, 9.2, 8.6
Off NetRtg: +3.7, +5.2, +1.1, -3.4, +2.4, +1.2, +2.8

If LeBron spends approximately six seconds with the ball in his hands, and that leads to over-reliance on him, why is it that Harden and Luka, who hold the ball for 50% longer, dribble the ball significantly more per touch, play under a similar team setup, and have comparable usage rates, see their teams perform well in their absence?

Most teams see their offensive rating plummet when their offensive superstar goes to the bench. People just selectively choose language that really just projects their aesthetic biases through their analysis. When it's a shoot first guard like Steph Curry or Michael Jordan they're celebrated: "Look at how impactful their ceiling raising is!". When it's LeBron James, it's "Well he takes his teammates out of rhythm."

The more plausible, albeit uncomfortable for some, explanation is that LeBron's unique combination of volume scoring + playmaking + all-time wing defense provides a greater lift on both ends of the court compared to Jordan, allowing him to take worse rosters to higher highs. The only caveat being that he couldn't do it for 48 minutes a game.

We have tracking data for ball dominance going back to the 2013-2014 season.

1) First the numbers that discuss how much LeBron and Kyrie dribbled the ball and how long they held on to it:

2) Next, since the beginning of the 2013-2014 season, the LESS ball dominant LeBron was, the GREATER the ON/OFF ORtg became. The argument is always that “he holds the ball and controls everything, so when he goes to the bench players are out of rhythm and this creates the crazy ON-OFF numbers.”

You can see all the numbers here:

[3)We have LeBron’s playoff ball dominance data that OhayoKD and others posts in this tweet:

OhayoKD wrote:.

Read on Twitter

Here is the playoffs and ball dominance data (I need to update for 2023 playoffs which will make the trend line about neutral, maybe slightly positive:

Here is the playoffs data table:

Strange this entire “LeBron’s teams struggle without him because he controls everything” only applies to him. Bulls offense went in the toilet in the playoffs without Jordan but that’s because of “ressons.”

I think those charts with a trend line for on-off and ball dominance % are pretty meaningless. There’s literally only a handful of data points, the correlation is really low, and the trendline isn’t steep. And the underlying dependent variable there (i.e. on-off) is subject to so many other variables that we’d just expect it to be super random (especially when we’re talking about different years with different teams, so it’s just completely different players playing in the “on” and “off” minutes in the different data points). So I don’t really think much value can be derived from that at all. It’s pretty obviously just noise.

Of course, I’m sure you’ll want to refer back to the charts I provided on Bill Walton with the 1986 Celtics. And it’d be a fair point. Except that (1) those charts had substantially more data points, (2) I was presenting them to refute a pretty hyperbolic point (i.e. that Walton was the most important part of the 1986 Celtics defense) while you’re presenting these to refute me merely having said something is a factor; (3) the underlying data points there weren’t from different years and with different teams, which makes these data points even sillier to try to draw a trend line between—it’d be like if I did what I did with Walton but included prior years’ games with other teams in the analysis.

As for the thing about Luka and Harden, I’ve addressed that to you and others multiple times now. I’ve pretty conclusively demonstrated that those guys’ teams have had very different lineup strategies, where their teams’ other stars (Chris Paul, Brunson, etc.) have been substantially more staggered, which obviously would have a huge effect on their on-off. No one is suggesting that is some effect that completely swamps all the other many things that affect on-off numbers.

I’d also repeat that my point is legitimately based on hearing NBA role players specifically say this. You and others keep acting like it’s crazy, but I put more credence in what I’ve heard actual NBA players say than what homecourtloss says.

In any event, this is all plainly off topic and derives from you bringing up subject matter from months ago that apparently still irks you but that has no actual relevance to this thread, so I don’t think I’ll continue down this irrelevant path moving forward.

Bulls without Jordan in 1997 playoffs [numbers behind BKREF paywall now]:

This is not a substantive point, but just FYI for you and others, you can actually still get to basketball-reference on-off numbers without going through a paywall. If you’re on a player’s page, go to the URL, delete the “.html” part at the end, and add a “/on-off/YEAR”. So, for instance, if I wanted to see Jordan’s 1997 on-off numbers, I can go to: https://www.basketball-reference.com/players/j/jordami01/on-off/1997

lessthanjake wrote:I’m really not seeing how being #1 by a significant margin in every season there’s data on except his rookie year could in any way be described as “not approach[ing] anything close to an outlier mythological peak that is above everyone else’s.” In the Squared data, Jordan has literally got an outlier RAPM that is above everyone else’s in every year except his rookie year!

And that includes Magic Johnson. Saying Magic “looks every bit as good” is just incorrect. They overlapped in three years in the Squared data. Magic is above Jordan in Jordan’s rookie season. But then Jordan is well above Magic in the other two seasons where they’re both actually in their prime. And to the extent you’re saying “peaks higher” by comparing RAPM values in different years, let’s remember that we’ve agreed at your insistence that that type of cross-year comparison of RAPM values is invalid.

Define “significant.” Is the gap bigger than the gap for others in other years? 1991 Jordan didn’t have an outlier gap. 1991 Jordan is a modest plus on defense (a modest negstive in 85 and 88). There’s nothing outlier here when Magic looks just as good.

I’m not understanding what you’re saying here. There’s six years of the Squared RAPM data that I’m aware of: 1969-1970, 1979-1980, 1984-1985, 1987-1988, 1990-1991, and 1995-1996. The first two years have so little data at the moment that the numbers are completely meaningless (they’re literally from tracking of 15-20 total games across the entire league, and you can tell from the possession numbers that it’s like 1-4 games for each player). Meanwhile, Michael Jordan is #1 in the last three of those years. The only year with any real sample size at all that Jordan isn’t #1 in is the 1984-1985 season. There, Magic Johnson is ahead by 0.47 points (or a 5.6% increase) over the #2 player. In 1987-1988, Michael Jordan is ahead by 0.85 points (or a 12.8% increase) over the #2 player. In 1990-1991, Michael Jordan is ahead by 1.45 points (or a 29.3% increase) over the #2 player. In 1995-1996, Michael Jordan is ahead by 0.78 points (or a 12.7% increase) over the #2 player. In other words, Michael Jordan is #1 in Squared RAPM in all but his rookie season, and in every single one of those seasons the gap he has over #2 is larger than the gap Magic has in the one year Magic is on top. There is no argument that “Magic looks just as good” in terms of having a “gap” over others. Michael Jordan is the clear standout in Squared’s RAPM data.

Of course, one could just say that Squared’s data is too low sample size to draw much of a concrete conclusion from. And you could object to the idea that the comparing the size of a gap across years is invalid (though you’re the one that suggested looking at that). That would all be valid. But what is not valid is to try to argue that Squared’s RAPM data actively goes against Jordan’s GOAT case. We can debate how much (if any) value to put on Squared’s data, but it does *not* go against Jordan’s GOAT case.

lessthanjake wrote: The thing is that Jordan doesn’t need to have had *more* impact in his earlier years than in the second three-peat for him to have been GOAT-level, because the impact in the second three-peat is itself GOAT-level overall (with that last season maybe being a shade off that, but the overall three-year span being GOAT-level impact for sure). So basically, I agree we don’t have enough info to really know for sure exactly how his earlier impact shaped up compared to the later impact, but the signals for all of it that we have is really really great and entirely consistent with a GOAT-level player, so I don’t really see the point of parsing through limited data trying to determine which of his time periods were more or less impactful. It looks like it was all incredibly impactful!

If the consensus is that 1991 is his peak and 1990 is peak adjacent and that 1988-1993 is a GOAT 6 year stretch, well, it would be nice to see that that period had an outlier level of impact. But it doesn’t.

Outlier compared to what? We know he had about a +14.7 playoff on-off in the six years from 1988-1993 (see my prior posts with that analysis). That’s GOAT-like. For reference, LeBron does not have any 6-year span where he actually made the playoffs every year and had as high a playoff on-off as that (his best such span is +13.9). LeBron does have a 6-playoff span from 2015-2021 (6 playoffs over 7 years, since he missed the playoffs one of those years) that’s slightly above that (it’s +15.5). So it’s basically a wash overall, where peak Jordan and peak LeBron look very comparable in this regard. I’m failing to see how peak Jordan’s playoff impact looking like peak LeBron’s playoff impact somehow suggests peak Jordan’s impact wasn’t GOAT-like. (And of course this isn’t even factoring in that Jordan’s playoff on-off in the second three-peat is even higher).

lessthanjake wrote: And no, I don’t really care what the output of the data is—there’s just obvious limitations and potential distortions of the data that shouldn’t require any motivation to notice.

Of course you don’t because Jordan does not look to have GOAT impact in those regressions (more below). Other WoWY data you meet with enthusiasm (such as Draymond’s) but from the beginning you have been trying to find ways to caution this, keep in mind, etc., which is true of data but you seem keen on cautioning this data since Moon’s data doesn’t seem to corroborate any outlier GOAT peak or consensus GOAT impact at all compared to others even when we have plenty of OFF data in these early years,

No, I’m cautioning the data because there’s obvious limitations that need to be acknowledged, rather than looking at the data and not thinking critically about what it means and what issues there are. It’s just bad analysis to not think about these things. I’m very consistent about this stuff—which is why others have criticized me before for constantly saying that we should assume that every metric we have is substantially flawed (which I did even while arguing for a guy like Steph, who would probably *benefit* from blind acceptance of what impact metrics say). There’s serious methodological flaws/issues/limitations with everything, and I don’t think Moonbeam himself would claim that that doesn’t apply to his model.

lessthanjake wrote: Again, though, Jordan’s “peak” cannot really be properly measured in Moonbeam’s data, because the three best players on his team (Jordan included) barely missed games during that timeframe, so the model has virtually no way of figuring out who was having what impact between them. It can just try to estimate the role players’ impact based on their missed time (and sometimes do that in a way that ends up being flawed, see, e.g., BJ Armstrong), and then make a rough guess at how to distribute the rest of the impact between Jordan, Pippen, and Grant, using the extremely limited number of missed games that each one had. Once the data actually saw what the Bulls did without Jordan, he quickly catapults to the top, with no player with a substantial playtime throughout the time period being ahead of him for the next 5 time periods starting with 1991-1995 (said based on Ridge version). This isn’t really a picture of a guy who was less impactful in those earlier years. It’s a situation where the model doesn’t have the data to isolate out the impact of the top few guys on the team, and then once the model gets more data to do that with (which also includes more data from Grant leaving Chicago), Jordan catapults to the top of the league.

It doesn’t show that he is an outlier impact king either—in fact, it hints that his impact peak wasn’t during the 1988-1993 period. If we had lineup data we’d have a better idea but we don’t but there are players who played all years in these 5 year segments who come out looking better EVEN though we have big chunks of off data. Also, all the limitations of the data you mention apply to Magic Johnson as well and he looks like an impact god.

Moonbeam’s RWOWY for Jordan (Jordan in the Ridge set, Jordan in the LASSO set)

This was a cursory count, and others can do their own, but I took out every player who only played one year in the data segments unless stated or the player in question, and I also took out every player who didn’t at least play solid rotational minutes for two of the years in the data segments (1,000+ minutes a season in at least two seasons).

1981-1985: 4th, 3rd
1982-1986: 12th, 20th [here you have a large off segment]
1983-1987: 16th, 30th [Again a large of segment]
1984-1988: 30th, 21st [and again]
1985-1989: 40th, 17th [and again]
1986-1990: 65th, 27th [And again, including 1990, close to his consensus peak]
1987-1991: 56th, 52nd [Includes his peak and 1990]
1988-1992: 23rd, 7th
1989-1993: 15th, 5th
1990-1994: 7th, 5th
1991-1995: 2nd, 4th
1992-1996: 1st, 1st
1993-1997: 1st, 2nd
1994-1998: 2nd, 3rd
1995-1999: 1st, 2nd
1996-2000: 9th, not listed
1997-2001: 3rd, 4th
1998-2002: 6th, 18th
1999-2003: not listed, not listed
2000-2004: 86th, not listed [note: did not check for players were only played one year or non-rotational minutes in this segment]
2001-2005: not listed, not listed
2002-2006: not listed, not listed
2003-2007: not listed, not listed

Moonbeam’s RWOWY for Magic (Magic the Ridge set, Magic in the LASSO set)

1976-1980: 7th, 6th [counting one year of Bird]
1977-1981: 3rd, 7th
1978-1982: 2nd, 3rd
1979-1983: 1st, 2nd
1980-1984: 1st, 3rd
1981-1985: 2nd, 3rd
1982-1986: 3rd, 2nd
1983-1987: 4th, 4th
1984-1988: 4th, 2nd
1985-1989: 3rd, 1st
1986-1990: 3rd, 1st
1987-1991: 5th, 1st
1988-1992: 4th, 2nd
1989-1993: 2nd, 1st
1990-1994: 1st, 2nd
1991-1999: Magic has only one year samples (i.e., 1991 and 1996) but he’s basically
top 5 in Ridge in every segment of 1991-1995, 1992-1996, 1993-1997, 1994-1998, 1995-1999

So basically top 5 every year which is beyond ridiculous in a pure, non-prior informed set. This is what one would expect Jordan’s peak years to look like given the mythological and unassailable status it has reached.

Again, you’re failing to acknowledge that the model has virtually no way of figuring out who was having what impact between a group of players when they all barely missed games during that timeframe. Given that Jordan barely missed games in his peak years and was playing with Pippen and Grant who also barely missed games in that timeframe, the model can’t really properly assess Jordan’s impact in those years. And once it got more “off” info on those players (i.e. from Jordan retiring and Grant leaving Chicago), Jordan quickly skyrocketed to the top of the league and stayed there for quite a while. And that's not even getting into things like the BJ Armstrong Problem (which pretty clearly unduly hurt the numbers for the actually highly impactful Bulls players like Jordan)

If you want to look at that and actually conclude that Jordan’s impact in his peak years1 wasn’t very high, then you can go ahead and do that I suppose. But I’m just pointing out that you’d be basing that conclusion off of results spit out when the model had plainly inadequate information with which to parse out Jordan’s individual impact and that once that same model got a bit more info with which to do so, the result looked great for Jordan.

A long post basically disregarding data or anything that deviates from a unanimous consensus goat peak other preconceived notions. Sounds about right.

homecourtloss · Post #29 » by **homecourtloss** » Sun Aug 13, 2023 7:28 pm

70sFan wrote:All this long discussion can be summed up quickly - there is no clear GOAT peak in the NBA history. I know it hurts Jordan or LeBron fans, but we don't have any clear proof that one players reached a level head and shoulders ahead of anyone else ever. Even players like Russell and Mikan have question marks, let alone Jordan.

I tend to agree with this— they are just so many things to consider in so many players we can make a legitimate argument for.

OhayoKD · Post #30 » by **OhayoKD** » Sun Aug 13, 2023 8:22 pm

MrVorp wrote:Yeah that’s not how the metric I displayed works. Think WOWY RAPM as a ridge regressed version of WOWY data. It’s also time decayed, meaning it weighs older games less.

Do you have a link or something to the methodology? I haven't seen anything from the tweets. Everyone seems to be doing better as they age including Lebron(whose teams posted worse srs from that point) in both variants, which is why I was thinking they were still doing "relative to expected trajectory" like Darko does. The first tweet from a few days earlier called the stat Darko WOWY and that regression seems like a progression from that.

Also I’ve noticed that these WOWY type stats, even the ones that try to account for teammates have only so much granularity that a player’s score is strongly correlated with his team’s SRS.

Well, yeah, that's why I much prefer granular year-to-year analysis and concentrated samples over the psuedo-rapm attempts. With WOWY, over an extended stretch, it's still a tiny per-season sample and adding a bunch of tinier-sampled adjustments(including stuff several years removed) could easily just distort things. Even with actual RAPM, there's value to looking at the actual gaps, because ultimately rapm(and it's copycats) are still artificially scaled approximations of the real thing.

So I think the most likely reason it loves second three peat Jordan is because of the dominant SRS of the Bulls at that time.

But then why does it love 30+ Lebron whose teams posted worse SRS? To be clear, it's perfectly possible that MJ's Actual impact peak is during the second three-peat(moonbeam's rapm approximation also sees Mj go from not standing out that much to becoming a big outlier in the mid 90's), but the gap shouldn't be that large. Moonbeam mentioned his regression also has a similar issue but neither the raw or regressed stuff he is using has Jordan jump --that-- much.

Jordan’s RAPM at the end of his Bulls career is basically within a standard error of peak Lebron, Duncan, Garnett anyway.

[/quote][/quote]
Maybe for Duncan and Garnett, but Lebron is pretty far removed from either whether you go by cheema or je's stuff. Of course in actual Darko, Duncan is the outlier jumping Lebron's trajectory after a poorly timed down-year in 2011...

lessthanjake · Post #31 » by **lessthanjake** » Sun Aug 13, 2023 8:42 pm

OhayoKD wrote:Just had a night full of booze, tats, dumplings, and headbutts but let's see if I can put together something semi-coherent before I doze off.

First, some house-cleaning:
Let's start with the thing your jordan posts in the 86 Rockets thread centered around...

Despite the fact that his supporting cast shot so badly in the series they lost that it’s almost impossible to find any series in the history of the NBA where a supporting cast shot so badly and the team still won. That’s just not a serious argument

Uh...

https://www.basketball-reference.com/playoffs/1974-nba-western-conference-finals-bulls-vs-bucks.html
https://www.basketball-reference.com/playoffs/1977-nba-western-conference-semifinals-warriors-vs-lakers.html
https://www.basketball-reference.com/playoffs/1964-nba-western-division-finals-hawks-vs-warriors.html
https://www.basketball-reference.com/playoffs/1962-nba-eastern-division-semifinals-nationals-vs-warriors.html
https://www.basketball-reference.com/playoffs/2013-nba-eastern-conference-finals-pacers-vs-heat.html
https://www.basketball-reference.com/playoffs/2018-nba-eastern-conference-finals-cavaliers-vs-celtics.html
https://www.basketball-reference.com/playoffs/2018-nba-eastern-conference-first-round-pacers-vs-cavaliers.html

Not really?

I’m not sure what the examples you’ve given are supposed to be showing. I opened the first link you provided and it’s a series where the Bucks won the series with the supporting cast having a 55.1% TS%, against an opponent that gave up an average of a 50.2% TS% in a league where the average TS% was 50.3%. So the supporting cast actually shot great! This is *obviously* not an example of a team winning while their supporting cast shoots worse than the 1990 Bulls did against the Pistons. Quite the opposite actually. I then looked at the second link you provided. There, at least, the winning team’s supporting cast did actually shoot badly. But the 1990 Bulls scoring efficiency was worse in absolute terms, worse in league-relative terms, and worse in opponent-relative terms. So that is also not an example at all. I stopped looking at that point, since it’s clear that you must have a misunderstanding of what we’re talking about (or maybe the booze clouded your response haha).

More generally, you seem to have tried to scour NBA history and not come up with any example that occurred in the 1980s, 1990s, or 2000s (and, as described above, the examples you provided in the 1970s were objectively wrong). That should tell you something. Honestly, I’d have expected there to be *something* in that span (and perhaps there actually is, I don’t know for sure, since I didn’t look at every single series in all those decades).

As enigma covered, it's pretty dumb to ignore roles here. Jordan's pretty much sole job was to score and the defense was outstanding(-5) despite being average in 89 and below average before Pippen/Grant ascended over the course of 1990. And this gets us to...

The defense was good, but teams fairly often have good defenses and still lose series when the supporting cast shoots *that* badly. I think you’ll find that it’s much more common to lose with a -5 rDRTG than it is to win with a supporting cast that shoots that badly.

Lol, a guy has a 31.7 playoff PER and 13.7 playoff BPM (both of which are top 10 all time for a single playoffs)

Yeah, notice how I said "despite subjective box-weightings". PER and BPM can't really account for pesky details like creation quality, ball-handling, team-wide orchestration, defensive attention(the triangle made doubling Jordan very difficult), or most-everything that happens defensively. All components Jordan is disadvantaged at relative to the best of the best. You have of course read something along those lines from different posters at several points and have chosen to pretend all of the above are not disadvantages.

Hakeem has 3 series with a similar scoring-efficiency gap as Jordan's 1990 and had he had the defensive support Jordan had, I imagine those may have ended as wins. The 1994 finals and the 1995 conference finals are also in the vicinity and thanks to Hakeem's much better defense, both ended in wins.

“In the vicinity” is doing quite a lot of work there. In reality, the Rockets supporting cast did not at all shoot as badly in the 1994 finals or 1995 conference finals as the Bulls supporting cast did in the 1990 conference finals.

For instance, in the 1994 finals, the Rockets supporting cast had a 48.93% TS%. That was a -3.87 rTS% in league-relative terms (compared to -7.5 for the 1990 Bulls), and -1.22 rTS% in opponent-relative terms (compared to -5.1 for the 1990 Bulls). And that was a very close 7-game series where the Rockets actually got outscored but won in 7 games. It seems extremely likely that the Rockets would’ve lost that series if the supporting cast shot as badly as the 1990 Bulls supporting cast did.

Meanwhile, I have no idea why you’re even mentioning the 1995 conference finals at all. In the 1995 conference finals, the Rockets supporting cast had a 52.32% TS%. That was a -2.07 rTS% in league-relative terms, and a +0.04 rTS% in opponent-relative terms. So it’s not even remotely close to the 1990 Bulls. And, of course, while we don’t know for sure what would’ve happened, given that the Rockets barely outscored the Spurs in that series, it’s probably a very good guess that the Rockets would’ve lost the series if their supporting cast shot as badly as the 1990 Bulls did against the Pistons.

Lebron also put up monstrous per/bpm/whatever for the entirety of 2009. The difference here is that he was also playing better defense than Jordan ever could, was facing more defensive attention as the primary ball-handler and was a more efficient passer. Let's cut the semantics. If Jordan played goat-level ball in 1990, then you should be able to explain to me in basketball terms how Jordan compensated for all the aforementioned disadvantages.

The fact that the Bulls took the Pistons (a team that only lost 2 playoff games to any team without Jordan in a two-year stretch) to the brink while the supporting cast shot awfully and did awfully in the Jordan “off” minutes (unless you think that roughly -28 “off” net rating was somehow driven by earlier series that the Bulls easily won) is pretty strong evidence that Jordan was playing at an extremely high level. As are the box score stats. You provide nothing to refute that besides just vague criticisms of his game. What he was doing from a basketball perspective was the same sorts of things he always did from a basketball perspective—things that made him a 5-time MVP, consensus GOAT, 6-time title winner, etc. Which includes dominant scoring, great on-ball creation, great off-ball creation (albeit in an era with less space), great defense, incredible ball security, fantastic rebounding for his position, etc. If you can’t identify what made Jordan great, then that’s on you, and perhaps you should go back and watch more film.

Moving on to on/off...
homecourtloss wrote:

As unibro pointed out about a year ago, Shaq's data here is inaccurate. Corrected, Shaq lands somewhere around MJ and Steph.

Again, as I pointed out to homecourtloss, this chart is just based on artificially short time periods in a low-sample-size stat, and Jordan’s best playoff on-off years aren’t all clustered together. It’s obviously better to look at larger sample sizes in playoff on-off, rather than artificially smaller snippets. And if we do that and actually look at all the data we have (i.e. from 1988-1993 & 1996-1998), we see Jordan had a roughly +16 on-off per 100 possessions in those 9 playoffs (see my prior posts detailing that). That’s incredibly high, and, for reference, LeBron has no 9-playoff span that is as high as that. Playoff on-off is noisy even in larger sample sizes, but this is clearly a GOAT-level signal from Jordan. It’s not the full data set for Jordan (it doesn’t have 1985-1987 or 1995) so we can’t say anything super definitive, but what we have is the most important years and it looks clearly superior to any comparable span from LeBron.

I don’t know what analysis you’re referring to later in your post with spoilered stuff, but the analysis I did is pretty simple and transparent, and so you should respond to that, since I am not going to vet someone else’s method when I did my own analysis.

alongside a playoff on-off that was around +36 per 100 possessions (maybe the highest ever recorded for a player that got to at least the conference finals)

Nope. In 2001 Duncan, scored at +38.

Lol, okay, there’s one example of someone having a higher playoff on-off. Really knocks Jordan’s performance off of the GOAT-level!

Okay, but Jordan doesn't need to win game 7 to win the series. Here was the box-score for their closest loss:
https://www.basketball-reference.com/boxscores/199005220DET.html

First of all, the degree of difficulty on a series is quite high if a player’s team will lose if they have even a single down shooting night. Second, are you aware that Jordan played that game injured (and also actually played essentially the entirety of game 1—which the Bulls also lost—injured)? He had a bad fall in the first quarter of game 1 that gave him a sprained wrist and a deep bruise in his hip. Maybe if he’d not been injured, the Bulls could’ve won the series. But, as it was, he objectively played great in the series, even though he did have one down night in terms of shooting (while injured).

LukaTheGOAT · Post #32 » by **LukaTheGOAT** » Sun Aug 13, 2023 10:24 pm

70sFan wrote:All this long discussion can be summed up quickly - there is no clear GOAT peak in the NBA history. I know it hurts Jordan or LeBron fans, but we don't have any clear proof that one players reached a level head and shoulders ahead of anyone else ever. Even players like Russell and Mikan have question marks, let alone Jordan.

THE GOAT peak is 2009 Lebron.

OhayoKD · Post #33 » by **OhayoKD** » Sun Aug 13, 2023 11:12 pm

lessthanjake wrote:
homecourtloss wrote:.

Yeah, I'm not sure why you keep going off on ball-dominance:

Spoiler:

"Ball-dominance -> bad bench" is not a thing.

Helio led offense also hits higher highs than non-helio led offense and is more replicable across different contexts:

Spoiler:

I really do not get your insistence telling other people they do not have your "independent knowledge"...

nd also whether you should be confidently asserting what people who watched basketball thought in a time period that you did not watch basketball and other posters did).

...when you keep whiffing on the basics and are repeatedly caught pushing theories that are outright contradicted by what we actually have.

Speaking of...

I’m not sure what the examples you’ve given are supposed to be showing.

A bunch of series where the gap between the lead-scorer's efficiency and the rest of the cast was similar to what it was for Jordan and his help in Chicago.

As has been pointed out to you, even going by absolute effeciency...

Your other claim was that the only exception (now other than the 2004 Pistons/Pacers series) was the 2021 Bucks/Nets. Also wrong: I found a couple (and probably could find a bunch from the 1960s), and they just so happen to come from the perimetre player ahead of Jordan on this forum’s player rankings.

2018 first round
2018 Pacers defence: 55.78%
Cavaliers without Lebron: 49.93%
(Lebron himself scored at 65.5% efficiency. )

2007 conference semifinals
2007 Nets defence: 53.97%
Cavaliers without Lebron: 47.67%

Anyway, I am not really blaming Jordan for the loss to any particular degree, because even if the defence was good, there are scales here, and obviously he was not playing for the 2004 Pistons. While I am skeptical that playing to their usual standards changes the result of Game 7, they definitely let him down in that game specifically. However, the team also gave him chances to end the series before Game 7, and while in a literal sense that single game decided their season, perhaps other players could have made use of the support provided by the starters and closed it out before that point.

...2021 Bucks/Nets was not the sole example. From what I linked there is also Wilt.

More generally, you seem to have tried to scour NBA history

Uh...no? Nba history was not scoured. These were just some candidates that came to mind. If we were to brainstorm a bit longer...
https://www.basketball-reference.com/playoffs/1976-aba-finals-nets-vs-nuggets.html
https://www.basketball-reference.com/playoffs/1964-nba-western-division-finals-hawks-vs-warriors.html
https://www.basketball-reference.com/playoffs/2004-nba-eastern-conference-finals-pistons-vs-pacers.html
The last series I linked, was also brought to your attention, and in response you politely read "lead scorer" as "best player"

The fact all these examples are lying around(and apparently were not caught by your search) really makes me wonder just how much you actually looked. You might note that multiple popular goat candidates, whose value is not nearly as tied to scoring as Jordan's is, are included here.

The defense was good, but teams fairly often have good defenses and still lose series when the supporting cast shoots *that* badly. I think you’ll find that it’s much more common to lose with a -5 rDRTG than it is to win with a supporting cast that shoots that badly.

Possibly, but I think most of us know better than to take your word for it. As is, the Bulls built their team to make scoring as easy as possible for a player whose greatness heavily relies on his claim as the league's greatest volume scorer. A great volume scorer largely being allowed to operate 1 on 1(or wide-open) also typically does not result in a loss.

You have insisted it is "not serious" to argue Jordan did not play "goat" level basketball when he lost here. Telling me that Jordan was great does not actually support that.

The fact that the Bulls took the Pistons (a team that only lost 2 playoff games to any team without Jordan in a two-year stretch) to the brink while the supporting cast shot awfully and did awfully in the Jordan “off” minutes (unless you think that roughly -28 “off” net rating was somehow driven by earlier series that the Bulls easily won) is pretty strong evidence that Jordan was playing at an extremely high level.

"Extremely high" =/ "goat-lvl" and pivoting from granulars to broad-strokes does not help you here.

Kareem played better teams closer or as close in 72 and 74, Lebron actually beat a better team by rsrs, psrs, or standard deviation in 2013 in what is considered one of the worse series of his prime before taking the 2015 Warriors to 6(smoked everyone else like Detroit) on the back of an elite defense built around him and defenders who were negative/neutral on their previous teams.

There is of course Hakeem's 86 run, Wilt taking the Celtics to 7 multiple times, Garnett pushing the Lakers...and no, that list also is not exhaustive.

Otherwise, I am very confused on what you think has not been addressed here. You listed PER and BPM and on/off(which doesn't actually support Jordan as a goat-lvl player) and all of those were covered(and have been covered multiple times over by several different posters.

You are just playing with words to lower the bar. If all you can say for Jordan being "goat-lvl" is "he played great", then "goat-lvl" he is not.

First of all, the degree of difficulty on a series is quite high if a player’s team will lose if they have even a single down shooting night.

Uhuh. And I and HCL, and Enigma, and Shaq, and Cieling and everyone else have definitely been saying Jordan wasn't great and the pistons series was a cake-walk. Gain the bar is "goat", the first letter in that acronym is "greatest", and even when we just restrict it to perimiter players, Jordan very clearly did not top the scale.

Again, explain to me how Jordan's performance was comparable to Lebron's in 2009 in basketball terms. What did he do to off-set all these advantages reflected, not only in impact and offensive ratings, and his ability to replicate his results in different situations, but also in his record-breaking box-stuff and his sophisticated box. Lebron is not far beyond any peak ever(if anything he would be well behind Russell), but he is a player who is relatively easy to Jordan, and you have not been able to make a "serious" case against him via granulars, via holistics, or whatever else. "Ball-dominance makes bench bad" would not still be getting airtime if you could.

You want to dismiss it as a matter of injury? Fine, then show me what series you think Jordan arguably hit the apex of basketball. 1990 is probably his best, by box, playoff-run if you account for opposing defense, and, perhaps your independent knowledge did not inform you of this, but Jordan had injury issues in 1991 too(this is probably a factor in his defensive involvement permanently decreasing post 88). Just like Lebron in 2010 and 2015, and Curry in 2016 and 2018, and Duncan basically forever after his first championship, and Russell in the only loss of his career he was available for more than half of the series.

You bring up perception(never mind Jordan's MVP wins were less dominant and he was winning less frequently until expansion), you bring up team success(5 mvp winner and 11-championshp winner Russell irrelevant I guess), and you bring up qualities which are not unique to Jordan(38 ppg is not dominant scoring I guess?), and then you insist that you're the one whose watched the films and the games when you have yet to acknowledge the basics of the triangle(Jordan was put off-ball to avoid extra defensive attention, not summon it).

By your standards your habit is "rude", but as importantly it's empty. You literally cherrypicked an aspect of basketball and then called everyone else "not serious" for looking at things holistically as if analysis of a component somehow refutes analysis of the whole that component is a part of.

Then you typed in all-caps "STOP" when a poster rightfully pointed out you are the one whose analysis is filled to the brim with holes despite kicking everything off with an assertion they didn't watch basketball during the time-period in question. Do you think it's just coincidence that 'healthy" 89 Jordan was not putting up the same production against the same team "injured" jordan posted maybe his best series(by box) against when his defensive attention was significantly lesser?

Jordan did not lose 20 pounds and gain athleticism after 89. His scoring improvement came with a playmaking trade-off(passer-rating and box-creation go down). His defensive activity lessened. Yet apparently he is a tier up from when he carried a near-30 win cast to(at best) 50 win regular seasons and playoff runs worse than the 2015 Cavs and the 86 Rockets(led by players I imagine you would say it is "not serious" to argue were playing "Goat-lvl"). This is just vibes. Consequently, you keep tossing whatever might stick for a few seconds before throwing something else when what you tossed first inevitably falls.

LukaTheGOAT wrote:
70sFan wrote:All this long discussion can be summed up quickly - there is no clear GOAT peak in the NBA history. I know it hurts Jordan or LeBron fans, but we don't have any clear proof that one players reached a level head and shoulders ahead of anyone else ever. Even players like Russell and Mikan have question marks, let alone Jordan.

THE GOAT peak is 2009 Lebron.

Absolute, value over replacement? Probably. But I don't think we can ignore how shooting variance rendered it moot in a contest to a not all-time team. Granted the Wallace injury was a factor, but I don't really see how you can argue for Lebron era-relative vs Russell when he was beating much better teams with weak casts as a player-coach.

Nonetheless, I don't think anyone here has asserted Lebron was a peak far beyond every other peak. But that is a completely different question to how he compares to another perimeter player in Jordan. From what I can tell the "advantage" that counters everything Lebron has is...secondary creation as an off-ball player(an adjustment that was literally designed to get him away from extra defenders?)

That seems like a stretch and a half.

lessthanjake · Post #34 » by **lessthanjake** » Mon Aug 14, 2023 4:24 am

OhayoKD wrote:
lessthanjake wrote:
homecourtloss wrote:.

Yeah, I'm not sure why you keep going off on ball-dominance:

Actually, homecourtloss brought the topic up in this thread completely randomly, with no actual bearing on the subject at hand. So the idea that I “keep going off on” it is a bit odd in the context of this thread.

"Ball-dominance -> bad bench" is not a thing.

Again, you can get as exasperated about this as you want, but I’ve heard actual NBA role players talk about this, and I’m quite certain that actual NBA players know more about what is “a thing” in NBA basketball than OhayoKD does. But also, this subject is irrelevant to this thread and is literally only being talked about because homecourtloss apparently couldn’t help but go off about some completely irrelevant month-old discussion we’d had because I guess he (and maybe you) are sitting there stewing about it still. Let’s try to get back on topic.

Helio led offense also hits higher highs than non-helio led offense and is more replicable across different contexts:

This is also off topic, but I just want to flag that it’s not really true, since you casually make this claim a lot. If you look at the top 15 regular season rORTG teams in history, it’s basically just Steve Nash teams, the peak Magic year (1987), and a bunch of non-heliocentric teams. Steve Nash personally makes up a lot of the top 15, but it’s still 8 very much non-helio teams in the top 15 (2016 Warriors; 1997 Bulls; 1998 Jazz; 1996 Bulls; 2004 Kings; 1982 Nuggets; 1988 Celtics; and 1992 Bulls). Similarly, if you look at the top 15 playoff rORTG teams in history, it’s also mostly non-helio teams (this time 9 out of 15, with the other six being four Nash teams and a couple LeBron teams). So yeah, this claim you keep making isn’t really right at all. Though I guess if your claim was more specifically that Steve Nash-led offenses led to higher highs than any other offense, then perhaps I might be able to get on board, since that actually is arguably true (with the 2005 Suns being the best offense ever IMO). I actually will probably be pushing Nash fairly soon in the top 100 project—he was easily my favorite player of the 2000s, and perhaps my favorite player ever.

I really do not get your insistence telling other people they do not have your "independent knowledge"...
nd also whether you should be confidently asserting what people who watched basketball thought in a time period that you did not watch basketball and other posters did).

...when you keep whiffing on the basics and are repeatedly caught pushing theories that are outright contradicted by what we actually have.

Speaking of...
I’m not sure what the examples you’ve given are supposed to be showing.

A bunch of series where the gap between the lead-scorer's efficiency and the rest of the cast was similar to what it was for Jordan and his help in Chicago.

But that’s not in any way on point, and I think you know that. The discussion was about the supporting cast shooting really badly, and you’re just talking about something only tangentially related.

As has been pointed out to you, even going by absolute effeciency...
Your other claim was that the only exception (now other than the 2004 Pistons/Pacers series) was the 2021 Bucks/Nets. Also wrong: I found a couple (and probably could find a bunch from the 1960s), and they just so happen to come from the perimetre player ahead of Jordan on this forum’s player rankings.

2018 first round
2018 Pacers defence: 55.78%
Cavaliers without Lebron: 49.93%
(Lebron himself scored at 65.5% efficiency. )

2007 conference semifinals
2007 Nets defence: 53.97%
Cavaliers without Lebron: 47.67%

Anyway, I am not really blaming Jordan for the loss to any particular degree, because even if the defence was good, there are scales here, and obviously he was not playing for the 2004 Pistons. While I am skeptical that playing to their usual standards changes the result of Game 7, they definitely let him down in that game specifically. However, the team also gave him chances to end the series before Game 7, and while in a literal sense that single game decided their season, perhaps other players could have made use of the support provided by the starters and closed it out before that point.

...2021 Bucks/Nets was not the sole example. From what I linked there is also Wilt.
More generally, you seem to have tried to scour NBA history

Uh...no? Nba history was not scoured. These were just some candidates that came to mind. If we were to brainstorm a bit longer...
https://www.basketball-reference.com/playoffs/1976-aba-finals-nets-vs-nuggets.html
https://www.basketball-reference.com/playoffs/1964-nba-western-division-finals-hawks-vs-warriors.html
https://www.basketball-reference.com/playoffs/2004-nba-eastern-conference-finals-pistons-vs-pacers.html

You still seem to not understand what we’re talking about here. That 1964 series you link to does not work at all. The Warriors won while Wilt’s supporting cast shot a 45.1% TS%. That seems low. But league-average TS% was 48.5% that year. We don’t know what TS% their opponent gave up, but they were a pretty average defense (just 0.6 lower DRTG than league average), so they probably gave up a pretty average TS%. The shooting display of the 1990 Bulls supporting cast was substantially worse in league-relative terms and almost certainly worse in opponent-relative terms as well. Similarly, the supporting cast in that ABA series (lol!) that you pointed to shot better in absolute terms, league-relative terms, and opponent-relative terms, so it also clearly does not work.

The difficulty you have in finding series that work here is telling. Between multiple of us, we’ve actually only found three examples (and only one league-relative example), and none in the first like 50 years of the league’s history. That is telling.

The last series I linked, was also brought to your attention, and in response you politely read "lead scorer" as "best player"

No, this is completely off-base. I was the one that first brought up that series, so it was in no way “brought to [my] attention.” And I didn’t misread anything, because the argument being made about it was just completely off-point. The entire point here is to identify when the “help” shot terribly in a series. So what is relevant here is who the best player on the team is because that’s the player that has a supporting cast. The rest of the Pistons were not the “help” for Rip Hamilton. The idea of that is basically laughable, as he was the 4th best player on the team. Whether he was the lead scorer or not, he was part of the “help,” similar to how Byron Scott was part of the “help” on the 1988 Lakers. If you want to argue that the rest of the 2004 Pistons were a supporting cast for Rip Hamilton, then we’re just not going to agree, and I don’t think any reasonable person would agree with you.

The fact all these examples are lying around(and apparently were not caught by your search) really makes me wonder just how much you actually looked.

The "examples" you've found literally have been repeatedly wrong (which is itself telling about how difficult these are to find). And I've said exactly what I looked at in my search. I looked at playoff series on almost a decade either side of that 1990 series, and then also looked up some additional series that popped into my head as possible options (which included, among others, that 2021 Bucks/Nets series and the 2004 Pistons/Pacers series). I've been extremely transparent about what I looked at and been very explicit in saying that I did not look at every series in NBA history. I looked at enough to see that what we're talking about is a really rare occurrence, and there's been absolutely nothing pointed out since then that has suggested otherwise. Between multiple people looking for these, we've found 3 examples in NBA history, only 1 of which was actually worse in league-relative *and* opponent-relative terms (the 2021 Bucks example). I imagine there's probably other examples hiding somewhere, but it's quite obvious that this is extremely rare, which was my entire point. This whole exercise has just proven it further.

The defense was good, but teams fairly often have good defenses and still lose series when the supporting cast shoots *that* badly. I think you’ll find that it’s much more common to lose with a -5 rDRTG than it is to win with a supporting cast that shoots that badly.

Possibly, but I think most of us know better than to take your word for it. As is, the Bulls built their team to make scoring as easy as possible for a player whose greatness heavily relies on his claim as the league's greatest volume scorer. A great volume scorer largely being allowed to operate 1 on 1(or wide-open) also typically does not result in a loss.

You have insisted it is "not serious" to argue Jordan did not play "goat" level basketball when he lost here. Telling me that Jordan was great does not actually support that.

The fact that you think Michael Jordan was "largely…allowed to operate 1 on 1(or wide open" against the Detroit Pistons is just wild. In any event, you have no basis for your claim that Jordan was *not* playing at a GOAT-level, when both his box stats and his on-off in those playoffs was unbelievably high, and in the series his team lost, his supporting cast shot awfully and Jordan's team still managed to win more games in that one series than their opponent lost combined in every playoff series they played against any other team during a two-year span. Your basis for this is literally just that you have decided in a conclusory manner that the way that Michael Jordan played basketball did not provide GOAT-level value. The fact that you're demanding further support from me here is quaint. You have no support for your position except your own personal opinion (which, I'll note, is not even informed by watching the player contemporaneously).

The fact that the Bulls took the Pistons (a team that only lost 2 playoff games to any team without Jordan in a two-year stretch) to the brink while the supporting cast shot awfully and did awfully in the Jordan “off” minutes (unless you think that roughly -28 “off” net rating was somehow driven by earlier series that the Bulls easily won) is pretty strong evidence that Jordan was playing at an extremely high level.

"Extremely high" =/ "goat-lvl" and pivoting from granulars to broad-strokes does not help you here.

Kareem played better teams closer or as close in 72 and 74, Lebron actually beat a better team by rsrs, psrs, or standard deviation in 2013 in what is considered one of the worse series of his prime before taking the 2015 Warriors to 6(smoked everyone else like Detroit) on the back of an elite defense built around him and defenders who were negative/neutral on their previous teams.

There is of course Hakeem's 86 run, Wilt taking the Celtics to 7 multiple times, Garnett pushing the Lakers...and no, that list also is not exhaustive.

Otherwise, I am very confused on what you think has not been addressed here. You listed PER and BPM and on/off(which doesn't actually support Jordan as a goat-lvl player) and all of those were covered(and have been covered multiple times over by several different posters.

Okay, so please find me an extended playoff run where anyone has matched or exceeded Jordan's box-score metrics while also matching or exceeding his on-off. It literally doesn’t exist. For instance, the only playoff runs where someone else has had a higher BPM than Jordan in 1990 while actually at least making the conference finals are LeBron in 2009 and Kawhi in 2017. Kawhi basically didn’t play in the conference finals so that hardly counts, but either way we know that the playoff on-off in both of those runs were not close to Jordan’s on-off in 1990. The only playoff run you’ve found where someone made the conference finals and had a higher on-off than Jordan in 1990 was Duncan in a year where his BPM was barely half that of Jordan’s in 1990. If no one else has ever matched the combination of box stats and raw impact data that Jordan had in that playoff run, then I think it’s pretty obvious he was playing at a GOAT-level, and your own personal vibes are not remotely persuasive otherwise.

Anyways, to clarify, I’m not invested in the idea that it was literally the best playoff run anyone has ever had. I meant “GOAT-level” in the sense of a player playing at the level you’d expect from a GOAT player, not that it was literally the GOAT playoff run—I don’t think even think it’s Jordan’s best playoff run. So it’s possible we’re arguing past each other, but I interpret you as arguing that Jordan was not playing like a GOAT-level player (feel free to correct me if I’m wrong). And that is just a completely vibes-based argument from you that is pretty well contradicted by box and impact data we have and your argument to the contrary is basically just “Yeah, but I don’t like Michael Jordan.”

You want to dismiss it as a matter of injury? Fine, then show me what series you think Jordan arguably hit the apex of basketball. 1990 is probably his best, by box, playoff-run if you account for opposing defense, and, perhaps your independent knowledge did not inform you of this, but Jordan had injury issues in 1991 too(this is probably a factor in his defensive involvement permanently decreasing post 88). Just like Lebron in 2010 and 2015, and Curry in 2016 and 2018, and Duncan basically forever after his first championship, and Russell in the only loss of his career he was available for more than half of the series.

This is getting off-topic regarding the 1990 Bulls specifically, but I’ll note that being the best player in a sport is not really about hitting “the apex” for an extremely small sample size. It’s about consistency. Like, I could point to Jordan in the 1991 Finals or something as being “the apex,” but I’m not actually super concerned with an argument about whether LeBron James played as well or better in the 2009 ECF or something (or a number of other incredible series from other players, a good number of whom have hit similar levels over small samples). Jordan’s playoff greatness is about having essentially always played really well in every series (something that LeBron very clearly cannot make claim to).

You bring up perception(never mind Jordan's MVP wins were less dominant and he was winning less frequently until expansion), you bring up team success(5 mvp winner and 11-championshp winner Russell irrelevant I guess), and you bring up qualities which are not unique to Jordan(38 ppg is not dominant scoring I guess?), and then you insist that you're the one whose watched the films and the games when you have yet to acknowledge the basics of the triangle(Jordan was put off-ball to avoid extra defensive attention, not summon it).

I promise you that I am more aware of what the triangle offense is than you, since there’s obviously just a really wide gap here in how much we’ve watched of this. I know you bristle at that assertion, but I mention it because it is true and very relevant to discussing with someone who crusades across RealGM against a player they did not actually contemporaneously watch. Anyways, your knowledge of the triangle offense seems to be wholly borne out of trying to figure out how to twist its existence in order to denigrate Michael Jordan. At its most basic level, it was about maintaining spacing, while using motion and ball movement. This is good basketball (primarily because it maximizes the threat posed by everyone on the floor), and it’s definitely not a negative to be able to work well within such a system. You try to twist things to denigrate Jordan, by saying it was designed to “avoid extra defensive attention” on Jordan, but you’re ignoring why it might do that—which is by increasing the threat posed by the rest of the Bulls players on the floor. Again, that’s obviously good basketball. The main reason not to use a system like that is because maximizing the threat posed by the rest of the players on the floor might not be able to be achieved without also substantially minimizing the value provided by your star player. So there’s a tradeoff. And not every star player can play in a system like that without their value being so minimized that it’s not worth it for the team. If a star player can get their full (or at least almost full) value in a system where the rest of the players are able to threaten the defense and extract as much value as possible, then that is great! This is fundamentally a positive, not a negative! Jordan’s ability to be a dominant scorer within the flow of that offense—basically being a source of very frequent quick and efficient buckets when the ball movement found him at any position within the triangle’s spacing—was incredibly valuable. And then, of course, you layer on great playmaking (both on and off ball), absolutely elite ball security, great defense, fantastic rebounding for a guard, elite transition scoring, etc. Jordan’s value is obvious. And it shows up in the data too, of course—indeed, that’s usually my main focus. You just have some vibes-based arguments for why you don’t like players like Jordan (though I suspect that those arguments are less a matter of general principle and more constructed in the first place specifically to argue against Jordan).

lessthanjake · Post #35 » by **lessthanjake** » Mon Aug 14, 2023 5:17 am

One tiny thing I’d add is that I realize we actually have plus-minus data for another playoffs in Jordan’s career: his rookie year. This is because the Squared historical plus-minus data for Jordan actually includes that playoff series (https://squared2020.com/2022/07/31/some-michael-jordan-plus-minus-numbers/).

And, in that series, the Bulls were +10 with Jordan on the court, and -22 overall, which means they were -32 with Jordan off the court. Jordan played 171 minutes in the series, and so that means he was off the court for 21 minutes. Therefore, we know that, in the playoffs in his rookie year, the Bulls were +5.85 per 48 minutes with Jordan on the court, and -73.14 per 48 minutes with Jordan off the court. The series was played at a 96.8 pace, so in per-100-possession terms that was probably more like +6.04 with Jordan on and -75.56 with Jordan off.

Obviously, this is very low sample size (only a 21 minute “off” sample) but it’s a data point we can add to the mix. And that’d take Jordan’s overall playoff on-off in the data we have even higher, to more like +17.6.

70sFan · Post #36 » by **70sFan** » Mon Aug 14, 2023 7:19 am

LukaTheGOAT wrote:
70sFan wrote:All this long discussion can be summed up quickly - there is no clear GOAT peak in the NBA history. I know it hurts Jordan or LeBron fans, but we don't have any clear proof that one players reached a level head and shoulders ahead of anyone else ever. Even players like Russell and Mikan have question marks, let alone Jordan.

THE GOAT peak is 2009 Lebron.

It's a reasonable opinion, but not a fact.

OhayoKD · Post #37 » by **OhayoKD** » Tue Aug 15, 2023 2:33 am

NBA Role player and widely respected basketball analyst JJ Reddick has repeatedly noted Jordan was not facing many double-teams during his title-runs. In Golden State, role-players, starters, and a certain back to back MVP recognized Durant as the Warrior's best player. Phil Jackson, the winningest head coach in nba history, claimed that the Bulls had not only the best, but the two best players in the league. And yes, there are role-players who have argued a guy doing everything galvanized the team.

If we're done with the authority appeals...

And, in that series, the Bulls were +10 with Jordan on the court, and -22 overall, which means they were -32 with Jordan off the court. Jordan played 171 minutes in the series, and so that means he was off the court for 21 minutes. Therefore, we know that, in the playoffs in his rookie year, the Bulls were +5.85 +2.81 per 48 minutes with Jordan on the court, and -73.14 per 48 minutes with Jordan off the court. The series was played at a 96.8 pace, so in per-100-possession terms that was probably more like +6.04 with Jordan on and -75.56 with Jordan off.

Obviously, this is very low sample size (only a 21 minute “off” sample) but it’s a data point we can add to the mix. And that’d take Jordan’s overall playoff on-off in the data we have even higher, to more like +17.6.

You're on-calc was off by 3, okay whatever. Is there a reason you're tossing in 85 while ignoring what Ben provided for 95? Could it be that the Bulls had an off of +14.7 giving Jordan an on/off of -14.7? In fact, If Ben's graph isn't including 1994, it gets down to -16.7. Either way, Jordan comes out looking significantly worse. From 2007-2021, Lebron posts an on/off of 13.8 roughly matching an 11-year mark of 13.75 for Jordan. If were to include 1994, Jordan would drop to +11.05. I am not seeing a strong "goat-lvl" claim here. Especially when we remember it's Lebron whose playing more minutes, and Lebron whose staggering more with star-teammates.

What I do see is you pretending 1995 was not discernible. And that is rather bad form.

Also interesting how much stock you're putting in these small samples when you have called larger regular-season ones "meaningless" on the basis of...being too small of a sample.

But I digress, most of nba-history is excluded and Jordan does not come out top. Even if we restrict this to 1-year. Duncan tops 1990 with his own conference final run and Lebron matches with a finals trip...

Helio led offense also hits higher highs than non-helio led offense and is more replicable across different contexts:

This is also off topic, but I just want to flag that it’s not really true, since you casually make this claim a lot. If you look at the top 15 regular season rORTG teams in history, it’s basically just Steve Nash teams, the peak Magic year (1987), and a bunch of non-heliocentric teams.

And then we do the playoffs and it's Lebron, Nash, Magic, a first-round exit from Bird, Steph with KD, Dray, and Klay 1 year, Shaq with Kobe turned MJ, and....Westbrook with KD and...Adams?

Naturally you opted to ignore that second clause.

I’m not sure what the examples you’ve given are supposed to be showing.

A bunch of series where the gap between the lead-scorer's efficiency and the rest of the cast was similar to what it was for Jordan and his help in Chicago.

But that’s not in any way on point, and I think you know that. The discussion was about the supporting cast shooting really badly, and you’re just talking about something only tangentially related.
[/quote]
I'd say it's pretty relevant when assessing how well Jordan played, but you were specifically talking teammate shooting so fair enough.

You've been flipping between different targets(absolute, league-relative, opponent relative), but regardless of approach, take Jordan out of the picture, and only some of the series mentioned qualify. If we do things properly(opponent-relative), we get 3 series including two from a perimeter player with better on/off whose value is tied less to scoring. The same can be said for Billups and the Pacers, a series you've tried to dismiss on the basis that the best scorer is not always the best player...

When we insert Jordan into the picture and look at the relative gaps, more examples emerge. It is a rare occurrence, but it is not "virtually impossible" to find or a "unique historical event". "Scour history" I did not. Brainstorm some obvious candidates and then link a bunch while under the influence I did. That list is not exhaustive. And even if it was...

we’ve actually only found three examples (and only one league-relative example), and none in the first like 50 years of the league’s history. That is telling.

...It still does not logically lead to your conclusion. Your claim was not that the cast shot woat-lvl bad. Your claim was not even that the series showcased goat-lvl scoring. It was that it showcased goat-lvl overall play. The first two has not been challenged by anyone. And we can just point to how you dismissed series #3 to why proving the first two would not prove the third...

Whether he was the lead scorer or not, he was part of the “help,” similar to how Byron Scott was part of the “help” on the 1988 Lakers.

No. This is just you playing word games. The "best player" is also part of the "help" for a teammate. When you are assessing "scoring" and not "overall play", you do not get to fall back on "but he was not the best player". Moreover, if it is "unreasonable" to think Billups was better than Wallace despite Wallace being a mediocre scorer, then I have no idea you can argue it is "not serious" to argue a player with "goat-lvl scoring" was not playing the best basketball ever.

Do you know what helps teammates shoot better? Playmaking. Feel free to check how Lebron and Jordan stack up in box-creation/passer-rating(even the 'inflated" era-adjusted score) if you think they were equals. Why does Ben Wallace get to be the best player on the Pistons despite **** scoring? Defense

You know what also helps teammates shoot better? Drawing defensive attention. Of course there is also something to be said for keeping turnovers down, as a primary ball-handler, running the team on both ends and a whole bunch of things you continue to dismiss as vague before you come back with...

And not every star player can play in a system like that without their value being so minimized that it’s not worth it for the team.

True. Jordan is the only player who can retain value while scaling-down to make use of better teammates...

...oh right. Lebron literally saw their ball-dominance plummet before...checks notes...taking a team that was bad without him to 60+ wins with and an all-time great playoff run? Woah. Craaaazy. Gee, no way they posted jordan-pippen level lineup-ratings when paired with an extremely similar player, or look better in playoff on/off despite staggering with co-stars, or went for scoring titles while posting great impact in lineups with westbrook...

Yeah let me fix this for you

Anyways, your knowledge of the Warrior's motion-offense seems to be wholly borne out of trying to figure out how to twist its existence in order to denigrate Michael Jordan Kevin Durant. At its most basic level, it was about maintaining spacing, while using motion and ball movement. This is good basketball (primarily because it maximizes the threat posed by everyone on the floor), and it’s definitely not a negative to be able to work well within such a system.

Or actually...

Anyways, your knowledge of the triangle seems to be wholly borne out of trying to figure out how to twist its existence in order to denigrate Michael Jordan Dennis Rodman. At its most basic level, it was about maintaining spacing, while using motion and ball movement. This is good basketball (primarily because it maximizes the threat posed by everyone on the floor), and it’s definitely not a negative to be able to work well within such a system.

Don't forget Rodman also led the league in IBM's computer formula while letting everyone else score and pass for him. This is not a negative. It is fundamentally a positive. I am definitely not throwing whatever might possibly stick.

The main reason not to use a system like that is because maximizing the threat posed by the rest of the players on the floor might not be able to be achieved without also substantially minimizing the value provided by your star player.

Or....maybe you don't do it because it

A. Requires specific personel
B. Requires strong coaching
C. Does not actually lead to better results?

Very persuasive my guy.

This is getting off-topic regarding the 1990 Bulls specifically, but I’ll note that being the best player in a sport is not really about hitting “the apex” for an extremely small sample size.

Fascinating defense of a player who joined a 27-win team, saw that team get way better despite, by your own admission, scaling down, and then was apparently playing goat-lvl basketball as they approached(kind of?) the highs of the

-> 86 Rockets
-> 2015 Cavs

Yup definitely goat-lvl.

but I mention it because it is true and very relevant to discussing with someone who crusades across RealGM against a player they did not actually contemporaneously watch.

You are aware this "crusade" involves various posters who were alive for the 90's, 80's, and even the 70's? But sure, you definitely watched more basketball than all of us. That's why you are repeatedly caught having no idea what you're talking about by posters you accuse of "Not having watched the games"(cough standard deviation cough).

And then, of course, you layer on great playmaking (both on and off ball), absolutely elite ball security, great defense, fantastic rebounding for a guard, elite transition scoring, etc. Jordan’s value is obvious.

Amazing how half of these skills are things the guy with better playoff on/off is better at. Again, definitely not "vibes".

And it shows up in the data too, of course—indeed, that’s usually my main focus.

Is it? Because a couple months ago your focus was where Sam Jones ranked on the last top 100. And, not yet realizing you could "twist" on/off, your application of "data" involved pretending it wasn't relevant when compiling "all the impact metrics" which you also were very clear "didn't mean much".

It would seem to me your focus is pushing whatever might favor the guys you like. And then of course projecting this focus onto others. But hey, if you are under the impression people find that convincing, feel free to proceed.

VanWest82 · Post #38 » by **VanWest82** » Tue Aug 15, 2023 2:43 am

1990 had the higher ceiling in large part because it had the much better Mike. I will go to my grave believing they win the title if Rodman doesn't undercut Jordan end of 1st q game one. MJ dominated that first quarter but was hobbling up and down the court for the rest of game one and two. It wasn't until they got the extra day off that Jordan started to look like himself. No way Bulls lose both if he doesn't get hurt, and Pippen probably doesn't have a "migraine" if there's no game 7.

lessthanjake · Post #39 » by **lessthanjake** » Tue Aug 15, 2023 3:38 am

OhayoKD wrote:You're on-calc was off by 3, okay whatever. Is there a reason you're tossing in 85 while ignoring what Ben provided for 95? Could it be that the Bulls had an off of +14.7 giving Jordan an on/off of -14.7? In fact, If Ben's graph isn't including 1994, it gets down to -16.7. Either way, Jordan comes out looking significantly worse. From 2007-2021, Lebron posts an on/off of 13.8 roughly matching an 11-year mark of 13.75 for Jordan. If were to include 1994, Jordan would drop to +11.05. I am not seeing a strong "goat-lvl" claim here. Especially when we remember it's Lebron whose playing more minutes, and Lebron whose staggering more with star-teammates.

What I do see is you pretending 1995 was not discernible. And that is rather bad form.

I’m not going to bother responding to the vast majority of this, since it’s just going around in circles and I’ll end up banging my head against a wall repeating myself to someone who is not going to listen, and I don’t think there’s much of any danger that your post will actually persuade anyone of pretty much anything.

What I will note is that I posted a thread compiling Jordan’s on-off data here: viewtopic.php?f=64&t=2314587

I tightened up the methodology I’d been using (read: took more thorough but time-consuming routes of analysis) and corrected some mistakes (including, for example, on the 1985 playoffs). You can see the data there. The bottom line is that Jordan’s playoff on-off in 1985 + 1988-1993 + 1996-1998 is about +17.21 per 100 possessions. You’ll find I also made a follow-up post trying to incorporate 1995 playoff data from the Thinking Basketball video (by trying to roughly derive it using the 1993-1995 chart—which is a bit rough but I’m not aware of any better way to do it). If we incorporate that in, the total playoff on-off comes down to +15.03 per 100 possessions. Which is enormously high, and well clear of any comparable span from LeBron (note: to correct your post, LeBron does not have a +13.8 playoff on-off from 2007-2021. Basketball Reference has it at +12.9).

What was the ceiling for the 90 and 95 Bulls?

Re: What was the ceiling for the 90 and 95 Bulls?

Re: What was the ceiling for the 90 and 95 Bulls?

Re: What was the ceiling for the 90 and 95 Bulls?

Re: What was the ceiling for the 90 and 95 Bulls?

Re: What was the ceiling for the 90 and 95 Bulls?

Re: What was the ceiling for the 90 and 95 Bulls?

Re: What was the ceiling for the 90 and 95 Bulls?

Re: What was the ceiling for the 90 and 95 Bulls?

Re: What was the ceiling for the 90 and 95 Bulls?

Re: What was the ceiling for the 90 and 95 Bulls?

Re: What was the ceiling for the 90 and 95 Bulls?

Re: What was the ceiling for the 90 and 95 Bulls?

Re: What was the ceiling for the 90 and 95 Bulls?

Re: What was the ceiling for the 90 and 95 Bulls?

Re: What was the ceiling for the 90 and 95 Bulls?

Re: What was the ceiling for the 90 and 95 Bulls?

Re: What was the ceiling for the 90 and 95 Bulls?

Re: What was the ceiling for the 90 and 95 Bulls?

Re: What was the ceiling for the 90 and 95 Bulls?

Re: What was the ceiling for the 90 and 95 Bulls?

Re: What was the ceiling for the 90 and 95 Bulls?

Re: What was the ceiling for the 90 and 95 Bulls?

Re: What was the ceiling for the 90 and 95 Bulls?

Re: What was the ceiling for the 90 and 95 Bulls?

Re: What was the ceiling for the 90 and 95 Bulls?

Re: What was the ceiling for the 90 and 95 Bulls?

Re: What was the ceiling for the 90 and 95 Bulls?

Re: What was the ceiling for the 90 and 95 Bulls?

Re: What was the ceiling for the 90 and 95 Bulls?

Re: What was the ceiling for the 90 and 95 Bulls?

Re: What was the ceiling for the 90 and 95 Bulls?

Re: What was the ceiling for the 90 and 95 Bulls?

Re: What was the ceiling for the 90 and 95 Bulls?

Re: What was the ceiling for the 90 and 95 Bulls?

Re: What was the ceiling for the 90 and 95 Bulls?

Re: What was the ceiling for the 90 and 95 Bulls?

Re: What was the ceiling for the 90 and 95 Bulls?

Re: What was the ceiling for the 90 and 95 Bulls?