5 year plus-minus of various all-time nba peaks(Feat. Shaq, Lebron, Jordan, Robinson and Curry)

[OhayoKD]

(got it from bucksin6 on reddit)
From Ben's mosr recent video

https://youtu.be/RqGDLV-do9c?t=842

Left to Right is how good a team is during the postseason. Down to Up is a player's plus/minus during the postseason.

6 5 year streches are in this video:

98-01 David Robinson: +25 (note: in limited minuites)

00-04 Shaq: +22

16-21 Lebron: +18

88-93 Jordan: +15

15-19 Curry: +14

91-96 Jordan: +6

Note that these are just 6 stretches that Ben singled out, not necceasrily the 6 most impressive ones.

Also note that these only date back to the 80's so legends like Kareem and Bill Russell(R.I.P) wouldn't be on this.

I also don't have access to the full data set but maybe one of you do.

[Proxy]

Man having this data available for most of Jordan's prime now is so awesome. Shoutout to Ben's team, Squared2020, and everyone that provided the film for real, great stuff.

[MyUniBroDavis]

(got it from bucksin6 on reddit)
From Ben's mosr recent video

https://youtu.be/RqGDLV-do9c?t=842

Left to Right is how good a team is during the postseason. Down to Up is a player's plus/minus during the postseason.

6 5 year streches are in this video:

98-01 David Robinson: +25 (note: in limited minuites)

00-04 Shaq: +22

16-21 Lebron: +18

88-93 Jordan: +15

15-19 Curry: +14

91-96 Jordan: +6

Note that these are just 6 stretches that Ben singled out, not necceasrily the 6 most impressive ones.

Also note that these only date back to the 80's so legends like Kareem and Bill Russell(R.I.P) wouldn't be on this.

I also don't have access to the full data set but maybe one of you do.

How is Shaq +22?

From pbp

2001 +1.4
2002 +21.3
2003 +9.2
2004 +25

I don’t know what his +/- was in 2000 but it would be about the same as his 02 and 04 runs I think from other stuff

Wouldn’t that add up to +18 if you take the 5 years?

[eminence]

Misunderstanding above - the x-axis is plus/minus (per 48 minutes, so a bit different than pbp/100), the y-axis is (plus/minus on-court) - (plus/minus off-court).

Eg:

Team +7 with star on-court, -4 with star off. +7 - -4 = +11

I would also note - MJ is the only mid 90's earlier star, everyone else is from '97 onwards. So no Hakeem, Magic, Bird, etc.

Edit: If you're around Ben, did you do any of the other main Bulls guys? At least Scottie?

[RCM88x]

That 1992 team is certainly underrated in a historical sense I think, Scottie and Grant really were at their best that year it seems.

The data is interesting, unsurprisingly no crazy outlier years, pretty impossible to have that for a title winning team. But also pretty consistent in each of the two runs which sort of points to a pretty stable roster and relative performance/lineups. Probably not something we'll see for such a stretch in todays NBA, unless something drastic changes.

[eminence]

That 1992 team is certainly underrated in a historical sense I think, Scottie and Grant really were at their best that year it seems.

The data is interesting, unsurprisingly no crazy outlier years, pretty impossible to have that for a title winning team. But also pretty consistent in each of the two runs which sort of points to a pretty stable roster and relative performance/lineups. Probably not something we'll see for such a stretch in todays NBA, unless something drastic changes.

With the runs sharing 3/5 years I don't think it's too different than the modern game. Would have to see multiple runs filled in for the other top guys to see how it compares. LeBron in particular I'm certain has a few of the other top dots.

[RCM88x]

That 1992 team is certainly underrated in a historical sense I think, Scottie and Grant really were at their best that year it seems.

The data is interesting, unsurprisingly no crazy outlier years, pretty impossible to have that for a title winning team. But also pretty consistent in each of the two runs which sort of points to a pretty stable roster and relative performance/lineups. Probably not something we'll see for such a stretch in todays NBA, unless something drastic changes.

With the runs sharing 3/5 years I don't think it's too different than the modern game. Would have to see multiple runs filled in for the other top guys to see how it compares. LeBron in particular I'm certain has a few of the other top dots.

I guess my comment was more about the single year plots he had in the video, or even the 3 year plots. There was a good amount of consistency year to year in each of the runs.

You don't really see than with LeBron or Curry's teams, they seem to jump around a lot more than Jordans teams did year to year.

[eminence]

That 1992 team is certainly underrated in a historical sense I think, Scottie and Grant really were at their best that year it seems.

The data is interesting, unsurprisingly no crazy outlier years, pretty impossible to have that for a title winning team. But also pretty consistent in each of the two runs which sort of points to a pretty stable roster and relative performance/lineups. Probably not something we'll see for such a stretch in todays NBA, unless something drastic changes.

With the runs sharing 3/5 years I don't think it's too different than the modern game. Would have to see multiple runs filled in for the other top guys to see how it compares. LeBron in particular I'm certain has a few of the other top dots.

I guess my comment was more about the single year plots he had in the video, or even the 3 year plots. There was a good amount of consistency year to year in each of the runs.

You don't really see than with LeBron or Curry's teams, they seem to jump around a lot more than Jordans teams did year to year.

Looking at Curry 3 year stretches to compare.

MJ first, estimating from image at 12:25
'88-'90: +3 on, -20 off
'89-'91: +8 on, -13 off
'90-'92: +9 on, -4 off
'91-'93: +9 on, +1 off
'92-'94: +7 on, +3 off
'93-'95: +5 on, +3 off

Curry from pbp, converted to /48 minutes from /100 possessions
'13-'15: +6.6 on, -12.5 off
'14-'16: +6.0 on, +1.4 off
'15-'17: +10.8 on, +2.1 off
'16-'18: +11.9 on, +4.2 off
'17-'19: +12.3 on, +0.1 off
'18-'22: +8.6 on, +0.3 off

MJ: 2.2 Oncourt rating σ, 8.0 On/Off σ
Curry: 2.5 Oncourt rating σ, 4.6 On/Off σ

Single Year

MJ, estimating from 11:32
'88: -4 on, -8 off
'89: +5 on, -20 off
'90: +7 on, ??? off (not sure what's going on there, is it off the charts bad? same spot? forgot to plot?)
'91: +13 on, +1 off
'92: +7 on, +5 off
'93: +7 on, -3 off

Curry from pbp converted to /48 minutes (skipping his single series run as MJs were as well)
'13: +5.0 on, -25.5 off
'15: +9.6 on, -0.5 off
'16: +3.7 on, +5.1 off
'17: +19.6 on, -3.3 off
'18: +12.7 on, +6.6 off
'19: +6.8 on, -10.0 off
'22: +7.6 on, -1.6 off

MJ: 5.0 Oncourt rating σ, 8.1 On/Off σ (no result entered for '90)
Curry: 5.0 Oncourt rating σ, 10.0 On/Off σ

It seems pretty close one way or the other.

[OhayoKD]

That 1992 team is certainly underrated in a historical sense I think, Scottie and Grant really were at their best that year it seems.

The data is interesting, unsurprisingly no crazy outlier years, pretty impossible to have that for a title winning team. But also pretty consistent in each of the two runs which sort of points to a pretty stable roster and relative performance/lineups. Probably not something we'll see for such a stretch in todays NBA, unless something drastic changes.

kinda suprised lebron and shaq's playoff teams had comparable on-differential with the players on. I would have expected the playoff bulls to be significantly ahead of everyone but the warriors

[falcolombardi]

Man now i wonder if we are going to have bill russel and chamberlain numbers one day

If there is a older guy i wouldnt bet against "breaking the scale" of impact metrics it is russel

[DraymondGold]

That 1992 team is certainly underrated in a historical sense I think, Scottie and Grant really were at their best that year it seems.

The data is interesting, unsurprisingly no crazy outlier years, pretty impossible to have that for a title winning team. But also pretty consistent in each of the two runs which sort of points to a pretty stable roster and relative performance/lineups. Probably not something we'll see for such a stretch in todays NBA, unless something drastic changes.

kinda suprised lebron and shaq's playoff teams had comparable on-differential with the players on. I would have expected the playoff bulls to be significantly ahead of everyone but the warriors

That's exactly what I was thinking! It's not shocking to me at all that LeBron/Shaq have a higher on/off. They're absolutely top Tier in on/off. What is surprising is that their team's performance with them on was as good as the GOAT-level Jordan Bulls, or that Jordan doesn't get more separation when looking at his on-court differential. He doesn't get this, inspire of his teams being clearly the best of the non-Curry bunch over a 5-year average... like the video suggests, the team's success does indeed seem more tied with teammates and less with Jordan than his Legend would suggest.

By my count, Jordan has one 5-year stretch that's top 15 in on/off ever and both stretches are top 15 ever in on-court differential. Considering we're using unadjusted plus-minus, one of the noisiest stats around, Thinking Basketball's point that this doesn't preclude Jordan from being the GOAT is a good one.

The fact that he doesn't dominate in this first ever comprehensive plus-minus study does raise some questions.

First, if we're saying the noise of unadjusted plus minus shouldn't make us that lower (yet) for Jordan, should this data make us higher for peak Shaq/LeBron/Robinson/Curry?
-While Shaq isn't as statistically as dominant as people usually suggest (e.g. in RAPM/PIPM), his on-off is otherworldly.
-While people often detract from LeBron because he needs to play LeBron ball, having his team play LeBron ball produces Jordan-Bulls-level success.
-While people detract from Robinson's supposed lack of resilience, he seems GOAT level when he has a good team around him.
-Similarly, while people detract Curry for a lack of resilience / well fitting teammates, he also seems GOAT level.

Second, should we expect changes in the relative rankings when adding adjustments to this on/off data? For example, it would be hard to estimate Jordan's teammate impact without just flat out calculating APM for everyone. But we could estimate Jordan's relative opponent difficulty (vs Shaq/LeBron/Robinson/Curry) by looking at opponent SRS. Might Jordan be facing more difficult opponents, on average? Even with a small boost in relative opponent difficulty, Jordan might pop into Tier 1 All Time in (semi-)Adjusted Plus Minus.

[capfan33]

The data definitely lends credence to the idea that MJ actually peaked from 89-91, his team was just straight cheeks the 1st 2 years.

[falcolombardi]

(got it from bucksin6 on reddit)
From Ben's mosr recent video

https://youtu.be/RqGDLV-do9c?t=842

Left to Right is how good a team is during the postseason. Down to Up is a player's plus/minus during the postseason.

6 5 year streches are in this video:

98-01 David Robinson: +25 (note: in limited minuites)

00-04 Shaq: +22

16-21 Lebron: +18

88-93 Jordan: +15

15-19 Curry: +14

91-96 Jordan: +6

Note that these are just 6 stretches that Ben singled out, not necceasrily the 6 most impressive ones.

Also note that these only date back to the 80's so legends like Kareem and Bill Russell(R.I.P) wouldn't be on this.

I also don't have access to the full data set but maybe one of you do.

Do you have the reddit thread link? Wonder what the more "mainstream" (for lack of a better word) fans in reddit think about this data compares to here in rela gm

[DraymondGold]

That 1992 team is certainly underrated in a historical sense I think, Scottie and Grant really were at their best that year it seems.

The data is interesting, unsurprisingly no crazy outlier years, pretty impossible to have that for a title winning team. But also pretty consistent in each of the two runs which sort of points to a pretty stable roster and relative performance/lineups. Probably not something we'll see for such a stretch in todays NBA, unless something drastic changes.

kinda suprised lebron and shaq's playoff teams had comparable on-differential with the players on. I would have expected the playoff bulls to be significantly ahead of everyone but the warriors

That's exactly what I was thinking! It's not shocking to me at all that LeBron/Shaq have a higher on/off. They're absolutely top Tier in on/off. What is surprising is that their team's performance with them on was as good as the GOAT-level Jordan Bulls, or that Jordan doesn't get more separation when looking at his on-court differential. He doesn't get this, inspire of his teams being clearly the best of the non-Curry bunch over a 5-year average... like the video suggests, the team's success does indeed seem more tied with teammates and less with Jordan than his Legend would suggest.

By my count, Jordan has one 5-year stretch that's top 15 in on/off ever and both stretches are top 15 ever in on-court differential. Considering we're using unadjusted plus-minus, one of the noisiest stats around, Thinking Basketball's point that this doesn't preclude Jordan from being the GOAT is a good one.

The fact that he doesn't dominate in this first ever comprehensive plus-minus study does raise some questions.

First, if we're saying the noise of unadjusted plus minus shouldn't make us that lower (yet) for Jordan, should this data make us higher for peak Shaq/LeBron/Robinson/Curry?
-While Shaq isn't as statistically as dominant as people usually suggest (e.g. in RAPM/PIPM), his on-off is otherworldly.
-While people often detract from LeBron because he needs to play LeBron ball, having his team play LeBron ball produces Jordan-Bulls-level success.
-While people detract from Robinson's supposed lack of resilience, he seems GOAT level when he has a good team around him.
-Similarly, while people detract Curry for a lack of resilience / well fitting teammates, he also seems GOAT level.

Second, should we expect changes in the relative rankings when adding adjustments to this on/off data? For example, it would be hard to estimate Jordan's teammate impact without just flat out calculating APM for everyone. But we could estimate Jordan's relative opponent difficulty (vs Shaq/LeBron/Robinson/Curry) by looking at opponent SRS. Might Jordan be facing more difficult opponents, on average? Even with a small boost in relative opponent difficulty, Jordan might pop into Tier 1 All Time in (semi-)Adjusted Plus Minus.

Comparing Average Opponents for LeBron vs Jordan

Cool, so I figured I'd do the approximate opponent comparison. We can't do an exact adjustment (turning this on/off to APM) based on the actual rotations of teammates/opponents, but we can look at the average opponent over the course of 48 minutes.

In short:

On/Off per 48:
1989-1993 Jordan: +15
2016-2021 LeBron: +18 (+3 better)

Average Opponent SRS per 48:
1989-1993 Jordan: +4.17
2016-2021 LeBron: +3.98

Approximate "Semi"-Adjusted Plus Minus per 48:
1989-1993 Jordan: +19.17
2016-2021 LeBron: +21.98 (+2.8 better)
(adjusting for average opponent quality, but not specific opponent rotations or teammate quality)

SO: estimating for Jordan's higher average opponent difficulty, he slightly closes the gap with LeBron but doesn't fully close it. Of course, he might close the gap further if we did the full RAPM calculation, but we can't know for sure without the full data.

[eminence]

.

https://www.reddit.com/r/nba/comments/wcsi3w/thinking_basketball_i_tracked_every_michael/

[falcolombardi]

.

https://www.reddit.com/r/nba/comments/wcsi3w/thinking_basketball_i_tracked_every_michael/

"Floaters for algernon" is the greatest username in the history of usernames

Flowers for algerbon being one of my favorite books ever also helps

[RCM88x]

kinda suprised lebron and shaq's playoff teams had comparable on-differential with the players on. I would have expected the playoff bulls to be significantly ahead of everyone but the warriors

That's exactly what I was thinking! It's not shocking to me at all that LeBron/Shaq have a higher on/off. They're absolutely top Tier in on/off. What is surprising is that their team's performance with them on was as good as the GOAT-level Jordan Bulls, or that Jordan doesn't get more separation when looking at his on-court differential. He doesn't get this, inspire of his teams being clearly the best of the non-Curry bunch over a 5-year average... like the video suggests, the team's success does indeed seem more tied with teammates and less with Jordan than his Legend would suggest.

By my count, Jordan has one 5-year stretch that's top 15 in on/off ever and both stretches are top 15 ever in on-court differential. Considering we're using unadjusted plus-minus, one of the noisiest stats around, Thinking Basketball's point that this doesn't preclude Jordan from being the GOAT is a good one.

The fact that he doesn't dominate in this first ever comprehensive plus-minus study does raise some questions.

First, if we're saying the noise of unadjusted plus minus shouldn't make us that lower (yet) for Jordan, should this data make us higher for peak Shaq/LeBron/Robinson/Curry?
-While Shaq isn't as statistically as dominant as people usually suggest (e.g. in RAPM/PIPM), his on-off is otherworldly.
-While people often detract from LeBron because he needs to play LeBron ball, having his team play LeBron ball produces Jordan-Bulls-level success.
-While people detract from Robinson's supposed lack of resilience, he seems GOAT level when he has a good team around him.
-Similarly, while people detract Curry for a lack of resilience / well fitting teammates, he also seems GOAT level.

Second, should we expect changes in the relative rankings when adding adjustments to this on/off data? For example, it would be hard to estimate Jordan's teammate impact without just flat out calculating APM for everyone. But we could estimate Jordan's relative opponent difficulty (vs Shaq/LeBron/Robinson/Curry) by looking at opponent SRS. Might Jordan be facing more difficult opponents, on average? Even with a small boost in relative opponent difficulty, Jordan might pop into Tier 1 All Time in (semi-)Adjusted Plus Minus.

Comparing Average Opponents for LeBron vs Jordan

Cool, so I figured I'd do the approximate opponent comparison. We can't do an exact adjustment (turning this on/off to APM) based on the actual rotations of teammates/opponents, but we can look at the average opponent over the course of 48 minutes.

In short:

On/Off per 48:
1989-1993 Jordan: +15
2016-2021 LeBron: +18 (+3 better)

Average Opponent SRS per 48:
1989-1993 Jordan: +4.17
2016-2021 LeBron: +3.98

Approximate "Semi"-Adjusted Plus Minus per 48:
1989-1993 Jordan: +19.17
2016-2021 LeBron: +21.98 (+2.8 better)
(adjusting for average opponent quality, but not specific opponent rotations or teammate quality)

SO: estimating for Jordan's higher average opponent difficulty, he slightly closes the gap with LeBron but doesn't fully close it. Of course, he might close the gap further if we did the full RAPM calculation, but we can't know for sure without the full data.

The Bulls were just far too dominant to be completely reliant on Jordan, I wouldn't have personally expected to see any outlier +/- numbers.

I think if +/- tells us anything, that the difference between the teams of truly ATG peak players is almost always how those teams perform when their stars are on the bench.

With LeBron on the court the 2017 or 2016 Cavs are absolutely on the level as Jordans best teams were with him on the court. The difference is how those teams performed with their stars on the bench, not really that surprising.

While those ATG players definitely can play a role in that. It do think that ultimately most of that is out of their control and up to the team/coach/FO along with just the state of the league. Boiling it down to players is more fun so we usually do that, but in the grand scheme of things much is usually outside of their control.

[eminence]

Man I wish all the dots were filled in, I want to know if the 3 to the right of Robinson after Curry '15-'19 are all Curry/Dray runs.

[homecourtloss]

That 1992 team is certainly underrated in a historical sense I think, Scottie and Grant really were at their best that year it seems.

The data is interesting, unsurprisingly no crazy outlier years, pretty impossible to have that for a title winning team. But also pretty consistent in each of the two runs which sort of points to a pretty stable roster and relative performance/lineups. Probably not something we'll see for such a stretch in todays NBA, unless something drastic changes.

kinda suprised lebron and shaq's playoff teams had comparable on-differential with the players on. I would have expected the playoff bulls to be significantly ahead of everyone but the warriors

That's exactly what I was thinking! It's not shocking to me at all that LeBron/Shaq have a higher on/off. They're absolutely top Tier in on/off. What is surprising is that their team's performance with them on was as good as the GOAT-level Jordan Bulls, or that Jordan doesn't get more separation when looking at his on-court differential. He doesn't get this, inspire of his teams being clearly the best of the non-Curry bunch over a 5-year average... like the video suggests, the team's success does indeed seem more tied with teammates and less with Jordan than his Legend would suggest.

By my count, Jordan has one 5-year stretch that's top 15 in on/off ever and both stretches are top 15 ever in on-court differential. Considering we're using unadjusted plus-minus, one of the noisiest stats around, Thinking Basketball's point that this doesn't preclude Jordan from being the GOAT is a good one.

The fact that he doesn't dominate in this first ever comprehensive plus-minus study does raise some questions.

First, if we're saying the noise of unadjusted plus minus shouldn't make us that lower (yet) for Jordan, should this data make us higher for peak Shaq/LeBron/Robinson/Curry?
-While Shaq isn't as statistically as dominant as people usually suggest (e.g. in RAPM/PIPM), his on-off is otherworldly.
-While people often detract from LeBron because he needs to play LeBron ball, having his team play LeBron ball produces Jordan-Bulls-level success.
-While people detract from Robinson's supposed lack of resilience, he seems GOAT level when he has a good team around him.
-Similarly, while people detract Curry for a lack of resilience / well fitting teammates, he also seems GOAT level.

Second, should we expect changes in the relative rankings when adding adjustments to this on/off data? For example, it would be hard to estimate Jordan's teammate impact without just flat out calculating APM for everyone. But we could estimate Jordan's relative opponent difficulty (vs Shaq/LeBron/Robinson/Curry) by looking at opponent SRS. Might Jordan be facing more difficult opponents, on average? Even with a small boost in relative opponent difficulty, Jordan might pop into Tier 1 All Time in (semi-)Adjusted Plus Minus.

It has always been interesting speculated about with Jordans plus minus numbers look like and before we really didn’t have anything. Then we got the 97 and 98 data and then the 94 to 96 data, and now some data from the other years.

I always thought there would be one of three outcomes to knowing this data:

1. +/- numbers are so staggering that the GOAT talk increases and his position amongst the majority is strengthened.

2. +/- numbers are about what we thought so he stays where he is for most

3. +/- numbers are underwhelming and some serious cracks open up.

From what we have so far, #1 hasn’t happened and we’re really in the #2 range.

[homecourtloss]

Man I wish all the dots were filled in, I want to know if the 3 to the right of Robinson after Curry '15-'19 are all Curry/Dray runs.

Also those bottom left ugly bubbles.