What impact metrics show MJ as a GOAT candidate?

[AEnigma]

In general agreement with you on all of that.

Hi AEnigma,
Is LeBron the greatest player of all time, in your opinion? Is he well ahead of Jordan?

(If you rank players between eras), who are your top 10?

Was just genuinely curious. Thanks for your time!

Will DM you so as to prevent this from being derailed by squabbles over one specific individual’s ranking; thanks for the interest!

I have said in other threads though that I have a top three tier of Lebron, Kareem, and Russell, depending on the specific era/positional/achievement framing used.

[AEnigma]

In general agreement with you on all of that. Personally, I tend to see less of a strict aging curve among the all-time greats than most — perhaps because I do not wallow as much in the box score values lol. It actually makes it a bit frustrating when talking Jordan because his older years are too often assumed to be this major step down in overall impact from his peak, when I think he maintained his individual impact pretty impressively. And that even extends to his Wizards seasons, which a lot of people like to pretend never happened but which I think reflect very well for him on an aging curve (albeit likely not quite as well as for Lebron or a couple of others).

Interesting read on the trouble with aging curves, lifted from the DARKO site:
http://blog.philbirnbaum.com/2019/11/why-you-cant-calculate-aging.html?m=1

Four notes on Jordan:

(1) I actually prefer his offensive game to LeBron's, because he was able to turn low percentage shots (long 2s and post ups) into game theoretical anomalies. That said, so was Adrian Dantley. So my belief is, LeBron is the clearly better floor raiser, Jordan is (probably) the better ceiling raiser.

(2) From Ben's research using Squared2020's data, there isn't really a next level for MJ in the playoffs. Jordan is a known quantity who will give you consistent performance (with minor caveats for 93 and 98, though I think both of those were due to injuries from the rest of the Bulls' supporting casts).

(3) There is this belief that 96 and 97 were big steps down from 88-93, but I don't think that's necessarily the case at all. He was still extremely athletic, put on more muscle, had a similar motor. There isn't reason to believe that first threepeat Jordan was on another level from the second threepeat, and it's a bit of a stretch from Jordan proponents given they know that the data doesn't fully exist.

(4) I think it's very, very likely that if we do eventually have more data from the 80s/early 90s, we'll see that Jordan's biggest rival/peer isn't a ghost from another era, but Magic Johnson. Squared2020's data (again, very limited sample) portrays Johnson in a very strong light. Jordan and Johnson are very close offensively, and not far off defensively. I would be fascinated in data on Magic's +/- in playoffs in his career, as I think he compares very favorably to Jordan.

Feel very similarly to you on 2, 3, and 4.

For 1, I think that can make some sense as an era relative assessment for specifically offensive ceiling raising. In the 1990s, that type of midrange scoring resilience was worth a lot and enabled easier offensive fit. In the absolute I do not see that as equally true today, with improved defensive coverages on pure scorers and increased league efficiency outpacing Jordan’s midrange shot profile. Lebron has some of that Shaq-esque ceiling raising value now with how potently he can score near the basket in nearly any situation. Jordan had a strong post game too, but Lebron is a clear historical outlier for a wing there.

Great thoughts all-around.

[LukaTheGOAT]

Four notes on Jordan:

(1) I actually prefer his offensive game to LeBron's, because he was able to turn low percentage shots (long 2s and post ups) into game theoretical anomalies. That said, so was Adrian Dantley. So my belief is, LeBron is the clearly better floor raiser, Jordan is (probably) the better ceiling raiser.

(2) From Ben's research using Squared2020's data, there isn't really a next level for MJ in the playoffs. Jordan is a known quantity who will give you consistent performance (with minor caveats for 93 and 98, though I think both of those were due to injuries from the rest of the Bulls' supporting casts).

(3) There is this belief that 96 and 97 were big steps down from 88-93, but I don't think that's necessarily the case at all. He was still extremely athletic, put on more muscle, had a similar motor. There isn't reason to believe that first threepeat Jordan was on another level from the second threepeat, and it's a bit of a stretch from Jordan proponents given they know that the data doesn't fully exist.

(4) I think it's very, very likely that if we do eventually have more data from the 80s/early 90s, we'll see that Jordan's biggest rival/peer isn't a ghost from another era, but Magic Johnson. Squared2020's data (again, very limited sample) portrays Johnson in a very strong light. Jordan and Johnson are very close offensively, and not far off defensively. I would be fascinated in data on Magic's +/- in playoffs in his career, as I think he compares very favorably to Jordan.[/quote]

I mean from plenty of people's evaluations, they see 96 and 97 Jordan as lesser. For one, I don't agree 96 and 97 Jordan's motor was the same as 88 or 89 Jordan. This is something Ben Taylor in film study has noted. Ben Taylor has also mentioned he feels as if Jordan's defense while less him running-around and getting off the ground as consistently in the 2nd 3 peat years in Q And A/Podcasts than his earlier years. He still thinks of Jordan as a strong defender, however, I don't think he would consider Jordan to have the same motor from 96-98 as later years.

We don't have complete plus-minus data of Jordan from earlier years, however metrics such as PIPM, RAPTOR, Backpicks BPM, are regressed against box-score numbers (similar idea to the again-curve idea you proposed, where you try to find correlations between variables) to attempt to get an estimate of his true impact.

From 1988-1993 in the PS,

Jordan averaged an Inflation Adjusted 34.5 pts per 75 on rTS% of 5.9% ahead of the league average.

From 1995-1997 in the PS,

Jordan averaged an Inflation Adjusted 32.8 pts per 75 on rTS% of 2.1% ahead of the league average. If you take out the 95 year because you believe he was rusty, the numbers are still basically the same at IA 33.1 pts per 75 on rTS% of 1.9% ahead of the league average.

Jordan dropped multiple tiers as a scorer, which would be a pretty big hit to his impact considering that is his strongest skill. Jordan's estimated shots created goes from about 12.4 shots per 100 possessions to 8.1 or 8.2 depending on the stretch you use. His impact metrics such as RAPTOR, PIPM, and Backpicks BPM all take a hefty cut as well and he doesn't look close to a prime worthy of being an Offensive GOAT candidate.

The 88-93 sample wasn't small either, as it was 4,211 minutes which is almost 2 seasons worth of play in today's league. The 1,957 minutes played from 95-97, for instance would be more minutes than any NBA player other than Anthony Edwards, Mikal Bridges, and Julius Randle have played this season (just for a frame of reference).

[ceiling raiser]

For 1, I think that can make some sense as an era relative assessment for specifically offensive ceiling raising. In the 1990s, that type of midrange scoring resilience was worth a lot and enabled easier offensive fit. In the absolute I do not see that as equally true today, with improved defensive coverages on pure scorers and increased league efficiency outpacing Jordan’s midrange shot profile. Lebron has some of that Shaq-esque ceiling raising value now with how potently he can score near the basket in nearly any situation. Jordan had a strong post game too, but Lebron is a clear historical outlier for a wing there.

Great thoughts all-around.

I agree to some extent that it was era specific (especially given the restrictions on zone defense), however fundamentally I think being able to make shots at a high efficiency that the defense is trained to give up (caveat being that these were largely in single coverage, as Jordan in the 90s was very much an off-ball player) is a huge advantage. I do want to point out that Jordan's shooting advantage wasn't just midrange proper, but long 2s as well. Certainly though at that point, the question of whether that shot is valuable or if you should take a step back and shoot a three at a slightly lower percentage at a higher EV makes sense.

From a game theoretical perspective, I think the biggest current arb is the deep three still: https://fansided.com/2018/07/31/nylon-calculus-game-theory-deep-3/ https://blogs.cornell.edu/info2040/2019/10/28/the-game-theory-behind-the-deep-3-in-the-nba/ . There is a big reason why the DeRozan/Aldridge Spurs were incredibly limited. Much of it has to do with smarter, more athletic defenses and legalized zone vs the 80s/90s.

[ceiling raiser]

I mean from plenty of people's evaluations, they see 96 and 97 Jordan as lesser. For one, I don't agree 96 and 97 Jordan's motor was the same as 88 or 89 Jordan. This is something Ben Taylor in film study has noted. Ben Taylor has also mentioned he feels as if Jordan's defense while less him running-around and getting off the ground as consistently in the 2nd 3 peat years in Q And A/Podcasts than his earlier years. He still thinks of Jordan as a strong defender, however, I don't think he would consider Jordan to have the same motor from 96-98 as later years.

I don't see a massive difference in motor. Jordan isn't in transition as often, but some of that was a macro playstyle consideration given the rest of the roster age. Jordan's off-ball activity offensively increased, and in the half court he was basically Reggie Miller. The only age-based decline was likely the dropoff in lateral quickness (his first step was still there) in the second threepeat, but that might've been due to putting on additional muscle.

We don't have complete plus-minus data of Jordan from earlier years, however metrics such as PIPM, RAPTOR, Backpicks BPM, are regressed against box-score numbers (similar idea to the again-curve idea you proposed, where you try to find correlations between variables) to attempt to get an estimate of his true impact.

Like I said above, my issue with any box score based argument is that all a box score can tell you is what a player's role was. KG has similar box score stats in 04 and 05, but was a clearly worse player in 05. A player's role is a lagging indicator vis a vis a player's impact, so that continuity isn't meaningful to me.

[falcolombardi]

In general agreement with you on all of that. Personally, I tend to see less of a strict aging curve among the all-time greats than most — perhaps because I do not wallow as much in the box score values lol. It actually makes it a bit frustrating when talking Jordan because his older years are too often assumed to be this major step down in overall impact from his peak, when I think he maintained his individual impact pretty impressively. And that even extends to his Wizards seasons, which a lot of people like to pretend never happened but which I think reflect very well for him on an aging curve (albeit likely not quite as well as for Lebron or a couple of others).

Interesting read on the trouble with aging curves, lifted from the DARKO site:
http://blog.philbirnbaum.com/2019/11/why-you-cant-calculate-aging.html?m=1

Four notes on Jordan:

(1) I actually prefer his offensive game to LeBron's, because he was able to turn low percentage shots (long 2s and post ups) into game theoretical anomalies. That said, so was Adrian Dantley. So my belief is, LeBron is the clearly better floor raiser, Jordan is (probably) the better ceiling raiser.

(2) From Ben's research using Squared2020's data, there isn't really a next level for MJ in the playoffs. Jordan is a known quantity who will give you consistent performance (with minor caveats for 93 and 98, though I think both of those were due to injuries from the rest of the Bulls' supporting casts).

(3) There is this belief that 96 and 97 were big steps down from 88-93, but I don't think that's necessarily the case at all. He was still extremely athletic, put on more muscle, had a similar motor. There isn't reason to believe that first threepeat Jordan was on another level from the second threepeat, and it's a bit of a stretch from Jordan proponents given they know that the data doesn't fully exist.

(4) I think it's very, very likely that if we do eventually have more data from the 80s/early 90s, we'll see that Jordan's biggest rival/peer isn't a ghost from another era, but Magic Johnson. Squared2020's data (again, very limited sample) portrays Johnson in a very strong light. Jordan and Johnson are very close offensively, and not far off defensively. I would be fascinated in data on Magic's +/- in playoffs in his career, as I think he compares very favorably to Jordan.

I understand the theory behind point 1 but couldnt it be the opposite? Specially in the more modern game

A reliable moderately high value output from midrange scoring being floor raising, while lebron ability to dominate inside ans unlock higher value/volume creation for teammates a higher ceiling? Midrange game is a floor to a team.

Having a high floor by being able to score well from "bad" spots simultaneously raises a team ceiling (because good teams will eventually be able to force you to take those) i agree with that. But i believe better creation ability and inside scoring in the more important area of the court more than matches that

Lebron teams actually have done better offensively in the playoffs even if in theory his game doesnt fit the "ceiling raiser" aesthetic. So he is the one who has actually shown his game to scale to a higher ceiling

Curry:
2015 +4 (RS) +4.1(PS)
2016 +7.9(RS)+5.7(PS)
2017 +6.8(RS)+11.6 (PS)
2018 + 5.0(RS)+6.5(PS)
2019 + 5.5(RS)+5.4 (PS)
average: 5.85 (RS) 6.6(PS)
combined average: +6.2

Lebron
2013 +6.4 (RS) +7.2 (PS)
2014 +4.2 (RS) +10.6 (PS)
2015 +5.5(RS) +5.5 (PS)
2016 +4.5(RS) +12.5 (PS)
2017 +4.8 (RS) +13.7 (PS)
Average +5.1(RS) +9.9 (PS)
combined average: +7.5

jordan* (i had to use his first 5 championship seasons)
1991 +6.7(RS) +11.7 (PS)
1992 +7.3(RS) +6.5 (PS)
1993 +4.9 (RS) +9.8 (PS)
1996 +7.6 (RS) +8.6 (PS)
1997 +7.7(RS) +6.5(PS)
average +6.85 (RS) +8.6(PS)
combined average:+7.7

nash

2005 suns. +8.4(RS) +17 (PS)
2006 suns +5.3(RS) +9.5 (PS)
2007 suns +7.4(RS)+7.6 (PS)
2008 suns. +5.8(RS) + 3.1 (PS)
2010 suns +7.7(RS) +13.4 (PS)
Average +6.9(RS) + 10.1 (PS)
combined average: +8.5

shaq

1998 +6.9(RS), +10.1(PS)
1999 +5.4(RS), +4.7(PS)
2000 +3.2(RS), +9.3(PS)
2001 +5.4 (RS) +13.6(PS)
2002 +4.9(RS), +6.4 (PS)
Average +5.2(RS) +8.8(PS)
combined average: +7

bird

1984 +3.3 (RS) +6.4 (PS)
1985 +4.9 (RS) +3.9 (PS)
1986 +4.6 (RS) + 8.3 (PS)
1987 +5.2 (RS) + 8.7 (PS)
1988 +7.4 (RS) +4.2 (PS)
average +5.1(RS) +6.3(PS)
combined average: +5.7

magic

1986 +6.1(RS) +6.7
1987 +7.6 (RS) +10.7
1988 +5.1(RS) +8.3
1989 +6 (RS) +9.3
1990 +5.9(RS) +8.4
Average +6.1(RS), + 8.7 (PS)
combined average: +7.4


Notice how the theorically more portable and ceiling raising archetype players (curry, bird, jordan) trend lower in team offenses result/ceiling than the "floor raiser" archetype ones (nash, lebron, magic)

I think absolute creation ability on-ball is underated whrn it comes to offense ceiling raiser whereas portability/off ball game are overstated

[ShaqAttac]

For 1, I think that can make some sense as an era relative assessment for specifically offensive ceiling raising. In the 1990s, that type of midrange scoring resilience was worth a lot and enabled easier offensive fit. In the absolute I do not see that as equally true today, with improved defensive coverages on pure scorers and increased league efficiency outpacing Jordan’s midrange shot profile. Lebron has some of that Shaq-esque ceiling raising value now with how potently he can score near the basket in nearly any situation. Jordan had a strong post game too, but Lebron is a clear historical outlier for a wing there.

Great thoughts all-around.

I agree to some extent that it was era specific (especially given the restrictions on zone defense), however fundamentally I think being able to make shots at a high efficiency that the defense is trained to give up (caveat being that these were largely in single coverage, as Jordan in the 90s was very much an off-ball player) is a huge advantage. I do want to point out that Jordan's shooting advantage wasn't just midrange proper, but long 2s as well. Certainly though at that point, the question of whether that shot is valuable or if you should take a step back and shoot a three at a slightly lower percentage at a higher EV makes sense.

From a game theoretical perspective, I think the biggest current arb is the deep three still: https://fansided.com/2018/07/31/nylon-calculus-game-theory-deep-3/ https://blogs.cornell.edu/info2040/2019/10/28/the-game-theory-behind-the-deep-3-in-the-nba/ . There is a big reason why the DeRozan/Aldridge Spurs were incredibly limited. Much of it has to do with smarter, more athletic defenses and legalized zone vs the 80s/90s.

how about their d. isn't bron's big "impact" edge on d?

[LukaTheGOAT]

I mean from plenty of people's evaluations, they see 96 and 97 Jordan as lesser. For one, I don't agree 96 and 97 Jordan's motor was the same as 88 or 89 Jordan. This is something Ben Taylor in film study has noted. Ben Taylor has also mentioned he feels as if Jordan's defense while less him running-around and getting off the ground as consistently in the 2nd 3 peat years in Q And A/Podcasts than his earlier years. He still thinks of Jordan as a strong defender, however, I don't think he would consider Jordan to have the same motor from 96-98 as later years.

I don't see a massive difference in motor. Jordan isn't in transition as often, but some of that was a macro playstyle consideration given the rest of the roster age. Jordan's off-ball activity offensively increased, and in the half court he was basically Reggie Miller. The only age-based decline was likely the dropoff in lateral quickness (his first step was still there) in the second threepeat, but that might've been due to putting on additional muscle.

We don't have complete plus-minus data of Jordan from earlier years, however metrics such as PIPM, RAPTOR, Backpicks BPM, are regressed against box-score numbers (similar idea to the again-curve idea you proposed, where you try to find correlations between variables) to attempt to get an estimate of his true impact.

Like I said above, my issue with any box score based argument is that all a box score can tell you is what a player's role was. KG has similar box score stats in 04 and 05, but was a clearly worse player in 05. A player's role is a lagging indicator vis a vis a player's impact, so that continuity isn't meaningful to me.

Kevin Garnett in 04
VORP-10
WS/48-0.272
TWPR-77.46
Backpicks BPM-6.5

Kevin Garnett in 05
BPM-9.1
WS/48-0.248
TWPR-57.77 (5th best season per this metric)

04 and 05 KG don't look that similar but to each their own I guess.

[AEnigma]

I mean from plenty of people's evaluations, they see 96 and 97 Jordan as lesser.

The dispute was not whether they were lesser, it was by how much.

For one, I don't agree 96 and 97 Jordan's motor was the same as 88 or 89 Jordan. This is something Ben Taylor in film study has noted. Ben Taylor has also mentioned he feels as if Jordan's defense while less him running-around and getting off the ground as consistently in the 2nd 3 peat years in Q And A/Podcasts than his earlier years. He still thinks of Jordan as a strong defender, however, I don't think he would consider Jordan to have the same motor from 96-98 as later years.

You can have a lower “motor” while still being a similar value defender.

His defensive plus-minus values from ’97 and ’98 were both strong for a wing — around 35th among scaled four-year peaks, next to Bruce Bowen — but given his post-retirement defensive prowess, I consider it unlikely that his earlier years were much more impactful.

We don't have complete plus-minus data of Jordan from earlier years, however metrics such as PIPM, RAPTOR, Backpicks BPM, are regressed against box-score numbers (similar idea to the again-curve idea you proposed, where you try to find correlations between variables) to attempt to get an estimate of his true impact.

And those box score regressions still contain all the flaws inherent in any box score analysis — they can only give one value to a box score input. More steals better than fewer steals. More blocks better than fewer blocks. More rebounds better than fewer rebounds. More assists better than fewer assists. Correlative as a rule, but not a substitute for analysis in itself.

From 1988-1993 in the PS,

Jordan averaged an Inflation Adjusted 34.5 pts per 75 on rTS% of 5.9% ahead of the league average.

From 1995-1997 in the PS,

Jordan averaged an Inflation Adjusted 32.8 pts per 75 on rTS% of 2.1% ahead of the league average. If you take out the 95 year because you believe he was rusty, the numbers are still basically the same at IA 33.1 pts per 75 on rTS% of 1.9% ahead of the league average.

Jordan dropped multiple tiers as a scorer, which would be a pretty big hit to his impact considering that is his strongest skill.

And how are you adjusting for opponents? How are you adjusting for team circumstance?

Jordan was a less capable driver in his second threepeat — but he also had Ron Harper and Dennis Rodman instead of BJ Armstrong and Horace Grant. He was a worse scorer in an absolute sense, yes; the question is whether that box score drop is 1:1 with a drop in his real impact.

Just because Lebron’s best three-year boxscores were in 2008-10 (or 2009/10 for two-year, or 2009 for one-year) does not mean that is the period we are all obligated or even likely to mark as his peak.

Jordan's estimated shots created goes from about 12.4 shots per 100 possessions to 8.1 or 8.2 depending on the stretch you use. His impact metrics such as RAPTOR, PIPM, and Backpicks BPM all take a hefty cut as well and he doesn't look close to a prime worthy of being an Offensive GOAT candidate.

I guess my question would be whether you feel he became a less capable passer during the second threepeat. I tend to assume most players maintain or improve on that front, and while I agree the reduced driving acumen limits his creation potential, I would again question how much that would be true on other teams. You can even see some of that pop up in his better stretches with the Wizards.

Box scores and box metrics are convenient shorthands, but at this level we should be putting forth the extra effort and analysing everything we can: individual skillset, opponent quality, team situation, etc.

[ceiling raiser]

Lebron teams actually have done better offensively in the playoffs even if in theory his game doesnt fit the "ceiling raiser" aesthetic. So he is the one who has actually shown his game to scale to a higher ceiling

[...]

Notice how the theorically more portable and ceiling raising archetype players (curry, bird, jordan) trend lower in team offenses result/ceiling than the "floor raiser" archetype ones (nash, lebron, magic)

I think absolute creation ability on-ball is underated whrn it comes to offense ceiling raiser whereas portability/off ball game are overstated

Quick question, are these raw offensive numbers, or are they adjusted for opponent DRtg? Nash is interesting, however LeBron and Magic played in the clearly lesser conference for the bulk of their careers.

Just want to make sure we're hitting all bases here.

[falcolombardi]

Lebron teams actually have done better offensively in the playoffs even if in theory his game doesnt fit the "ceiling raiser" aesthetic. So he is the one who has actually shown his game to scale to a higher ceiling

[...]

Notice how the theorically more portable and ceiling raising archetype players (curry, bird, jordan) trend lower in team offenses result/ceiling than the "floor raiser" archetype ones (nash, lebron, magic)

I think absolute creation ability on-ball is underated whrn it comes to offense ceiling raiser whereas portability/off ball game are overstated

Quick question, are these raw offensive numbers, or are they adjusted for opponent DRtg? Nash is interesting, however LeBron and Magic played in the clearly lesser conference for the bulk of their careers.

Just want to make sure we're hitting all bases here.

They were adjusted for rival defense yeah

Also, lebron eastern conference rivals were better defenses than jordan rivals by a sizable margin at that. Which is why his team offense numbers look better when adjusted to rival

Lebron from 2007-2016 played one of the toughest defensive gauntlets ever. I think only wilt had it clearly tougher over his prime with dealing with monster defense regularly

(magic western rivals were **** defenses usually but that is accounted for in relative offense)

[canada_dry]

All of em.

Sent from my SM-G960W using RealGM mobile app

[ceiling raiser]

They were adjusted for rival defense yeah

Also, lebron eastern conference rivals were better defenses than jordan rivals by a sizable margin at that. Which is why his team offense numbers look better when adjusted to rival

Lebron from 2007-2016 played one of the toughest defensive gauntlets ever. I think only wilt had it clearly tougher over his prime with dealing with monster defense regularly

(magic western rivals were **** defenses usually but that is accounted for in relative offense)

So this is very interesting. Can you show me your calculations? Not that I don't believe you, just that there are a number of ways to calculate this. is this:

On Offense - (Opponent DRtg - League Average DRtg)
(On Offense - Off Offense) - (Opponent DRtg - League Average DRtg)
Something else maybe?

I am a bit curious, what do the defensive numbers look like for those guys?

[falcolombardi]

They were adjusted for rival defense yeah

Also, lebron eastern conference rivals were better defenses than jordan rivals by a sizable margin at that. Which is why his team offense numbers look better when adjusted to rival

Lebron from 2007-2016 played one of the toughest defensive gauntlets ever. I think only wilt had it clearly tougher over his prime with dealing with monster defense regularly

(magic western rivals were **** defenses usually but that is accounted for in relative offense)

So this is very interesting. Can you show me your calculations? Not that I don't believe you, just that there are a number of ways to calculate this. is this:

On Offense - (Opponent DRtg - League Average DRtg)
(On Offense - Off Offense) - (Opponent DRtg - League Average DRtg)
Something else maybe?

I am a bit curious, what do the defensive numbers look like for those guys?

I got most of the data from this data base and manually calculated the few ones the database was not updated for (is from ben taylor if i am not mistaken)

viewtopic.php?t=1703897&start=20

https://docs.google.com/spreadsheets/d/1lMHVWmmq6lEy9O9XqLk0Ji-xawtX8gPRtHHwbvV9634/edit#gid=999526014

Only issue is that is team level rather than with the player ON (for obvious reasons as there is no +/- tracked league wide before 97)

so there is a slight margin of uncertainty in how the player ON differs from the team average

[ceiling raiser]

They were adjusted for rival defense yeah

Also, lebron eastern conference rivals were better defenses than jordan rivals by a sizable margin at that. Which is why his team offense numbers look better when adjusted to rival

Lebron from 2007-2016 played one of the toughest defensive gauntlets ever. I think only wilt had it clearly tougher over his prime with dealing with monster defense regularly

(magic western rivals were **** defenses usually but that is accounted for in relative offense)

So this is very interesting. Can you show me your calculations? Not that I don't believe you, just that there are a number of ways to calculate this. is this:

On Offense - (Opponent DRtg - League Average DRtg)
(On Offense - Off Offense) - (Opponent DRtg - League Average DRtg)
Something else maybe?

I am a bit curious, what do the defensive numbers look like for those guys?

It was series offensive rating - rival regular season defensive rating (scoring a 110 ORTG vs a 100 DRTG regular season team is a +10) then i averaged the series in each playoff run

I got most of the data from this data base and manually calculated the few ones the database was not updated for (is from ben taylor if i am not mistaken)

viewtopic.php?t=1703897&start=20

https://docs.google.com/spreadsheets/d/1lMHVWmmq6lEy9O9XqLk0Ji-xawtX8gPRtHHwbvV9634/edit#gid=999526014

Only issue is that is team level rather than with the player ON (for obvious reasons as there is no +/- tracked league wide before 97)

so there is a slight margin of uncertainty in how the player ON differs from the team average

Thanks for the clarification. I think that is somewhat flawed. So the issue is:

ORtg_team1 - DRtg_team2 = ORtg_team1 + ((DRtg_team2 - DRtg_league) + DRtg_league) = ORtg_team1 - (rDRtg_team2 + DRtg_league) = adj_ORtg - DRtg_league

If instead, we use the following:

ORtg_team1 - (DRtg_team2 - DRtg_league) = adj_ORtg

there's no additional term buoyed by league average.

[falcolombardi]

So this is very interesting. Can you show me your calculations? Not that I don't believe you, just that there are a number of ways to calculate this. is this:

On Offense - (Opponent DRtg - League Average DRtg)
(On Offense - Off Offense) - (Opponent DRtg - League Average DRtg)
Something else maybe?

I am a bit curious, what do the defensive numbers look like for those guys?

It was series offensive rating - rival regular season defensive rating (scoring a 110 ORTG vs a 100 DRTG regular season team is a +10) then i averaged the series in each playoff run

I got most of the data from this data base and manually calculated the few ones the database was not updated for (is from ben taylor if i am not mistaken)

viewtopic.php?t=1703897&start=20

https://docs.google.com/spreadsheets/d/1lMHVWmmq6lEy9O9XqLk0Ji-xawtX8gPRtHHwbvV9634/edit#gid=999526014

Only issue is that is team level rather than with the player ON (for obvious reasons as there is no +/- tracked league wide before 97)

so there is a slight margin of uncertainty in how the player ON differs from the team average

Thanks for the clarification. I think that is somewhat flawed. So the issue is:

ORtg_team1 - DRtg_team2 = ORtg_team1 + ((DRtg_team2 - DRtg_league) + DRtg_league) = ORtg_team1 - (rDRtg_team2 + DRtg_league) = adj_ORtg - DRtg_league

If instead, we use the following:

ORtg_team1 - (DRtg_team2 - DRtg_league) = adj_ORtg

there's no additional term buoyed by league average.

It definetely coukd have adjustments, ON-OFF particularly would be particularly useful if we ever get it for older seasons

But my main point here remains that those players who often get the semi backhanded "floorraiser" label for being on-ball players (magic, nash, lebron) often lead better offenses that the off ball "ceiling raising" players which are assumed based on playstyle to lead to higjer offensive ceilings (bird, curry, jordan)

I have no actual results based reason to believe nash or magic or lebron have ceiling raising limitstions as offensive players when they literally have the top 3 ever playoff offense runs

[LukaTheGOAT]

I mean from plenty of people's evaluations, they see 96 and 97 Jordan as lesser.

The dispute was not whether they were lesser, it was by how much.

For one, I don't agree 96 and 97 Jordan's motor was the same as 88 or 89 Jordan. This is something Ben Taylor in film study has noted. Ben Taylor has also mentioned he feels as if Jordan's defense while less him running-around and getting off the ground as consistently in the 2nd 3 peat years in Q And A/Podcasts than his earlier years. He still thinks of Jordan as a strong defender, however, I don't think he would consider Jordan to have the same motor from 96-98 as later years.

You can have a lower “motor” while still being a similar value defender.

His defensive plus-minus values from ’97 and ’98 were both strong for a wing — around 35th among scaled four-year peaks, next to Bruce Bowen — but given his post-retirement defensive prowess, I consider it unlikely that his earlier years were much more impactful.

We don't have complete plus-minus data of Jordan from earlier years, however metrics such as PIPM, RAPTOR, Backpicks BPM, are regressed against box-score numbers (similar idea to the again-curve idea you proposed, where you try to find correlations between variables) to attempt to get an estimate of his true impact.

And those box score regressions still contain all the flaws inherent in any box score analysis — they can only give one value to a box score input. More steals better than fewer steals. More blocks better than fewer blocks. More rebounds better than fewer rebounds. More assists better than fewer assists. Correlative as a rule, but not a substitute for analysis in itself.

From 1988-1993 in the PS,

Jordan averaged an Inflation Adjusted 34.5 pts per 75 on rTS% of 5.9% ahead of the league average.

From 1995-1997 in the PS,

Jordan averaged an Inflation Adjusted 32.8 pts per 75 on rTS% of 2.1% ahead of the league average. If you take out the 95 year because you believe he was rusty, the numbers are still basically the same at IA 33.1 pts per 75 on rTS% of 1.9% ahead of the league average.

Jordan dropped multiple tiers as a scorer, which would be a pretty big hit to his impact considering that is his strongest skill.

And how are you adjusting for opponents? How are you adjusting for team circumstance?

Jordan was a less capable driver in his second threepeat — but he also had Ron Harper and Dennis Rodman instead of BJ Armstrong and Horace Grant. He was a worse scorer in an absolute sense, yes; the question is whether that box score drop is 1:1 with a drop in his real impact.

Just because Lebron’s best three-year boxscores were in 2008-10 (or 2009/10 for two-year, or 2009 for one-year) does not mean that is the period we are all obligated or even likely to mark as his peak.

Jordan's estimated shots created goes from about 12.4 shots per 100 possessions to 8.1 or 8.2 depending on the stretch you use. His impact metrics such as RAPTOR, PIPM, and Backpicks BPM all take a hefty cut as well and he doesn't look close to a prime worthy of being an Offensive GOAT candidate.

I guess my question would be whether you feel he became a less capable passer during the second threepeat. I tend to assume most players maintain or improve on that front, and while I agree the reduced driving acumen limits his creation potential, I would again question how much that would be true on other teams. You can even see some of that pop up in his better stretches with the Wizards.

Box scores and box metrics are convenient shorthands, but at this level we should be putting forth the extra effort and analysing everything we can: individual skillset, opponent quality, team situation, etc.

He specifically uses the words "motor in his post," hegnce I argued against the idea.

The rTS% is adjusted for opponent (His TS%-average opposing team allowed).

And players typically don't peak as creators later in their career, because being able to pressure the defense with scoring matters. This is why you see declines in shots created for all greats, such as Magic, Nash, Bird, etc.

Per Ben Taylor's words, Jordan peaked ad a playmaker before his first retirement.

[McBubbles]

Just wanna say, this an excellent and wonderfully enlightening thread. Nice to see something with MJ's name in it not turn into a cesspool.

[DraymondGold]

Focus: Team Performance as an Argument for Jordan > LeBron

Part 1: Flaws in the usual argument for LeBron's teams performance > Jordan's
I’ve seen people use team performance to argue LeBron > Jordan. The argument usually goes something like this:
1. LeBron’s best offensive teams have a higher offensive rating and relative offensive rating than Jordan’s teams. Since LeBron’s the best offensive player on them, that suggests LeBron’s close to or greater than Jordan on offense.
2. LeBron’s the better defender over Jordan (based on rim protection / BBIQ / versatility arguments, concerns for Jordan’s defensive gambling, some available statistics).
3. Therefore LeBron’s the better player.

There are some (potential) flaws with this line of thinking. The reasoning assumes:
A. That LeBron’s better offensive rating comes from his offensive equality/superiority to Jordan, and not from different teammate or team contexts.
B. Lebron is a better defender (point #2) at the same time he’s having a better offensive rating.
C. That LeBron was capable of having a better offensive rating at the time he was a better defender

Assumption A may be faulty. LeBron’s better offensive ratings came during his 2nd Cleveland stint. The Cavs often prioritized offense at the expense of defense (especially in 17 and 18, when LeBron’s offense was at its best). Their best offensive ratings came when LeBron and Kevin Love were the two bigs out there, but this came at the cost of the defense. Love’s a better shooting big man than almost anybody else in the league, but he certainly has defensive liabilities — if LeBron had a big that was better defensively, it’s hard to imagine he would get a big that fit as well as a spacer and pick-and-pop threat as Kevin Love. LeBron is a great non-big defender, but if he’s your Power Forward defender without a strong rim protector, your defense may not be at the top of the league (at least as LeBron got older). Plus, many of their supporting cast and bench were offense-first players… Kyrie Irving, Kyle Korver, J.R. Smith all prioritize the team offense over the defense.

In fact, if we look at every player who played more than 15 minutes in the playoffs in 2016 and 2017, only 1/7 had a better defensive RAPM than offensive RAPM! The other 6 players were more valuable on offense over defense by a significant margin. In 2018, 8/9 of the Cavs top rotation players were significantly better on offense over defense.

Meanwhile, the Bulls were often defensive-first teams. Jordan’s second-stint Bulls especially focused on defense, relative to LeBron’s equivalent second-stint Cavs who focused on offense. In 1997 (our first season with full RAPM data, and the parallel team to LeBron’s 2017 cavs), 4/8 of the top rotation players were more valuable on defense over offense according to RAPM. And these 4 don’t include Pippen or Rodman, who are also more defense-focused compared to LeBron’s costars in Kyrie and Love. Phil Jackson as a coach was also more willing to focus on defense over offense when planning rotations ore lineups compared to Tyronn Lue.

If Jordan's teammates were more defense-focused while LeBron's were more offense-focused, just looking at offensive rating (rather than net rating) as a judge for both players is likely to systematically overrate LeBron / underrate Jordan.

Assumption B/C might also be faulty. LeBron was clearly a better defender in 2009–2012, when his teams did not reach a higher relative offense than the Bulls. LeBron had stretches of similar defense in 13 and 16, and LeBron’s teams had stretches of better offense then compared to 09-12, but he doesn’t really gain any separation over Jordan’s teams offensively in those years. Likewise, in the years when LeBron’s teams produced his best offense (2017, but also 2014 and 2018 if you curve for the supporting cast), LeBron had down-years defensively and his teams did not have better defenses than the Bulls. In short: even if LeBron did lend teams with better offensive ratings than Jordan, it’s not clear-cut that LeBron could combine his better individual defense with his teams’ better offense at the same time.

Part 2: An argument for Jordan’s ceiling-raising over LeBron's:
Most analytically-minded people who argue for peak/prime Jordan > LeBron say that Jordan’s a better ceiling raiser than LeBron. And while you can argue LeBron’s ceiling raising is higher based on his superior offensive rating, I think it’s worth noting there’s other evidence that Jordan’s ceiling raising is higher.

For example: 96 Bulls >> any LeBron team.
In basically every single standard team stat that we have — whether it’s regular season-only (Regular Season Record, Regular Season SRS), postseason-only (Postseason Record, Postseason SRS, relative Net Rating, common-opponent Net Rating), or full-season (Fivethirtyeight’s ELO, overall SRS) — the 96 Bulls are universally better than any single LeBron team.

There’s not one stat that takes even one LeBron team over the 96 Bulls, whether it’s regular season, playoffs-only, or both. And Jordan was clearly the best player on the 96 Bulls. In the actual RAPM data we have for 96, Jordan was more valuable on both offense and defense than Pippen or Rodman. In 96, Jordan was still one of the best defensive guards ever, and had matured in his defensive decision making and gambling. On offense, his all-time scoring ability and above-average efficiency was incredibly valuable, being one of the primary sources of playmaking on the Bulls. His scoring volume enabled Rodman to focus more on his best offensive skill (rebounding), and his scoring efficiency made those rebounds more valuable. His historic off-ball ability enabled other Bulls to focus on some of their better offensive skills (e.g. Pippen/Kukoc focusing on playmaking). And his all-time low turnover economy helped sustain the Bulls’ immense possession advantage.

There’s also a strong argument that 91 Bulls > Any LeBron team.
The 91 Bulls are better than any single LeBron team in every postseason-only stat (Postseason Record, Postseason SRS, relative Net Rating, common-opponent Net Rating) and every full-season stat (Fivethirtyeight’s ELO, overall SRS).

They’re also better than the vast majority of LeBron’s regular season teams. The only LeBron teams with a better regular season record are the 13 Heat and 09 Cavs, while the only LeBron team with a better regular season SRS are the 09 Cavs. However, the 90-91 Bulls are a well-known for having a steady improvement throughout their regular seasons. By the second half of the season, the 91 Bulls had a better regular season SRS than any LeBron team ever.

Notably, 1991 is commonly considered Jordan’s best individual year. If Jordan’s best individual year coincided with a team that’s clearly better than any LeBron team, that’s definitely a point in favor of the argument that Jordan’s the better ceiling raiser.

There’s also arguments for other Jordan teams being better than any LeBron team. Our two best standard full-season team stats are Overall SRS and ELO. In overall SRS, both the 92 Bulls and the 97 Bulls are better than any LeBron team ever. In ELO, the 92 Bulls, 97 Bulls, and 98 Bulls are better than any LeBron team ever.

Summary
I pointed out 3 flaws in the argument for LeBron’s team > Jordan’s teams. I pointed out that basically every stat we have favors the 96 Bulls and 91 Bulls (during Jordan’s peak) over any LeBron team ever. Both standard full-season team stats we have favor the 92 Bulls and 97 Bulls over any LeBron team ever. These performances favor the idea that Jordan is a better ceiling raiser than LeBron.

Now this is by no means a conclusive argument. You could make counter-arguments to defend the LeBron argument, and I’m sure you can make counters against my Jordan argument. With any team-performance-based argument for a player, there’s greater uncertainty as it’s difficult to fully isolate for an individual player’s contribution vs their teammates/coach.

However, there’s still some value in looking at team stats as an order-of-magnitude estimate or proof-of-concept estimate for a great player. For example, an all-time player with a good supporting cast should be able to lead an all-time team. And while you might favor LeBron’s team argument, there is evidence to the contrary: Jordan’s best teams were likely better overall than LeBron’s best teams.

[falcolombardi]

Focus: Team Performance as an Argument for Jordan > LeBron

Part 1: Flaws in the usual argument for LeBron's teams performance > Jordan's
I’ve seen people use team performance to argue LeBron > Jordan. The argument usually goes something like this:
1. LeBron’s best offensive teams have a higher offensive rating and relative offensive rating than Jordan’s teams. Since LeBron’s the best offensive player on them, that suggests LeBron’s close to or greater than Jordan on offense.
2. LeBron’s the better defender over Jordan (based on rim protection / BBIQ / versatility arguments, concerns for Jordan’s defensive gambling, some available statistics).
3. Therefore LeBron’s the better player.

There are some (potential) flaws with this line of thinking. The reasoning assumes:
A. That LeBron’s better offensive rating comes from his offensive equality/superiority to Jordan, and not from different teammate or team contexts.
B. Lebron is a better defender (point #2) at the same time he’s having a better offensive rating.
C. That LeBron was capable of having a better offensive rating at the time he was a better defender

Assumption A may be faulty. LeBron’s better offensive ratings came during his 2nd Cleveland stint. The Cavs often prioritized offense at the expense of defense (especially in 17 and 18, when LeBron’s offense was at its best). Their best offensive ratings came when LeBron and Kevin Love were the two bigs out there, but this came at the cost of the defense. Love’s a better shooting big man than almost anybody else in the league, but he certainly has defensive liabilities — if LeBron had a big that was better defensively, it’s hard to imagine he would get a big that fit as well as a spacer and pick-and-pop threat as Kevin Love. LeBron is a great non-big defender, but if he’s your Power Forward defender without a strong rim protector, your defense may not be at the top of the league (at least as LeBron got older). Plus, many of their supporting cast and bench were offense-first players… Kyrie Irving, Kyle Korver, J.R. Smith all prioritize the team offense over the defense.

In fact, if we look at every player who played more than 15 minutes in the playoffs in 2016 and 2017, only 1/7 had a better defensive RAPM than offensive RAPM! The other 6 players were more valuable on offense over defense by a significant margin. In 2018, 8/9 of the Cavs top rotation players were significantly better on offense over defense.

Meanwhile, the Bulls were often defensive-first teams. Jordan’s second-stint Bulls especially focused on defense, relative to LeBron’s equivalent second-stint Cavs who focused on offense. In 1997 (our first season with full RAPM data, and the parallel team to LeBron’s 2017 cavs), 4/8 of the top rotation players were more valuable on defense over offense according to RAPM. And these 4 don’t include Pippen or Rodman, who are also more defense-focused compared to LeBron’s costars in Kyrie and Love. Phil Jackson as a coach was also more willing to focus on defense over offense when planning rotations ore lineups compared to Tyronn Lue.

If Jordan's teammates were more defense-focused while LeBron's were more offense-focused, just looking at offensive rating (rather than net rating) as a judge for both players is likely to systematically overrate LeBron / underrate Jordan.

Assumption B/C might also be faulty. LeBron was clearly a better defender in 2009–2012, when his teams did not reach a higher relative offense than the Bulls. LeBron had stretches of similar defense in 13 and 16, and LeBron’s teams had stretches of better offense then compared to 09-12, but he doesn’t really gain any separation over Jordan’s teams offensively in those years. Likewise, in the years when LeBron’s teams produced his best offense (2017, but also 2014 and 2018 if you curve for the supporting cast), LeBron had down-years defensively and his teams did not have better defenses than the Bulls. In short: even if LeBron did lend teams with better offensive ratings than Jordan, it’s not clear-cut that LeBron could combine his better individual defense with his teams’ better offense at the same time.

Part 2: An argument for Jordan’s ceiling-raising over LeBron's:
Most analytically-minded people who argue for peak/prime Jordan > LeBron say that Jordan’s a better ceiling raiser than LeBron. And while you can argue LeBron’s ceiling raising is higher based on his superior offensive rating, I think it’s worth noting there’s other evidence that Jordan’s ceiling raising is higher.

For example: 96 Bulls >> any LeBron team.
In basically every single standard team stat that we have — whether it’s regular season-only (Regular Season Record, Regular Season SRS), postseason-only (Postseason Record, Postseason SRS, relative Net Rating, common-opponent Net Rating), or full-season (Fivethirtyeight’s ELO, overall SRS) — the 96 Bulls are universally better than any single LeBron team.

There’s not one stat that takes even one LeBron team over the 96 Bulls, whether it’s regular season, playoffs-only, or both. And Jordan was clearly the best player on the 96 Bulls. In the actual RAPM data we have for 96, Jordan was more valuable on both offense and defense than Pippen or Rodman. In 96, Jordan was still one of the best defensive guards ever, and had matured in his defensive decision making and gambling. On offense, his all-time scoring ability and above-average efficiency was incredibly valuable, being one of the primary sources of playmaking on the Bulls. His scoring volume enabled Rodman to focus more on his best offensive skill (rebounding), and his scoring efficiency made those rebounds more valuable. His historic off-ball ability enabled other Bulls to focus on some of their better offensive skills (e.g. Pippen/Kukoc focusing on playmaking). And his all-time low turnover economy helped sustain the Bulls’ immense possession advantage.

There’s also a strong argument that 91 Bulls > Any LeBron team.
The 91 Bulls are better than any single LeBron team in every postseason-only stat (Postseason Record, Postseason SRS, relative Net Rating, common-opponent Net Rating) and every full-season stat (Fivethirtyeight’s ELO, overall SRS).

They’re also better than the vast majority of LeBron’s regular season teams. The only LeBron teams with a better regular season record are the 13 Heat and 09 Cavs, while the only LeBron team with a better regular season SRS are the 09 Cavs. However, the 90-91 Bulls are a well-known for having a steady improvement throughout their regular seasons. By the second half of the season, the 91 Bulls had a better regular season SRS than any LeBron team ever.

Notably, 1991 is commonly considered Jordan’s best individual year. If Jordan’s best individual year coincided with a team that’s clearly better than any LeBron team, that’s definitely a point in favor of the argument that Jordan’s the better ceiling raiser.

There’s also arguments for other Jordan teams being better than any LeBron team. Our two best standard full-season team stats are Overall SRS and ELO. In overall SRS, both the 92 Bulls and the 97 Bulls are better than any LeBron team ever. In ELO, the 92 Bulls, 97 Bulls, and 98 Bulls are better than any LeBron team ever.

Summary
I pointed out 3 flaws in the argument for LeBron’s team > Jordan’s teams. I pointed out that basically every stat we have favors the 96 Bulls and 91 Bulls (during Jordan’s peak) over any LeBron team ever. Both standard full-season team stats we have favor the 92 Bulls and 97 Bulls over any LeBron team ever. These performances favor the idea that Jordan is a better ceiling raiser than LeBron.

Now this is by no means a conclusive argument. You could make counter-arguments to defend the LeBron argument, and I’m sure you can make counters against my Jordan argument. With any team-performance-based argument for a player, there’s greater uncertainty as it’s difficult to fully isolate for an individual player’s contribution vs their teammates/coach.

However, there’s still some value in looking at team stats as an order-of-magnitude estimate or proof-of-concept estimate for a great player. For example, an all-time player with a good supporting cast should be able to lead an all-time team. And while you might favor LeBron’s team argument, there is evidence to the contrary: Jordan’s best teams were likely better overall than LeBron’s best teams.

Everythingh except for how good they were with jordan on the court

Like this was a very in depth argument but the point was just "jordan team did better hence he was better"