Compilation of Michael Jordan’s on-off data with the Bulls

[trex_8063]

Locking this topic until a PC board moderator steps in.

That exchange where you talk to each other without quoting each other is quite silly. You can make an effort to be better... since you should be aware other adults will have to spend their time reading that and dealing with it applying their "forum justice."

(Just a temporary lock, the topic has great discussion and will be unlocked by a PC mod)

Re-opening, and asking posters to simply ignore perceived baits, as the perpetrator is in time-out anyway.

[Djoker]

Ya know, I'm not fully up on what these numbers would look like /48 for other notables (Shaq/Duncan/KG/LeBron/Steph maybe?).

Anybody have any of those sitting around?

Here you go!

[OhayoKD]

Ya know, I'm not fully up on what these numbers would look like /48 for other notables (Shaq/Duncan/KG/LeBron/Steph maybe?).

Anybody have any of those sitting around?

Lebron is +11 for his career and also near +15 over a similar stretch(though more minutes) from 2006-2021. No idea about the rest.

[eminence]

Ya know, I'm not fully up on what these numbers would look like /48 for other notables (Shaq/Duncan/KG/LeBron/Steph maybe?).

Anybody have any of those sitting around?

Here you go!

Would you mind sharing a RS one as well?

[lessthanjake]

Ya know, I'm not fully up on what these numbers would look like /48 for other notables (Shaq/Duncan/KG/LeBron/Steph maybe?).

Anybody have any of those sitting around?

Lebron is +11 for his career and also near +15 over a similar stretch(though more minutes) from 2006-2021. No idea about the rest.

That is factually inaccurate. Using plus-minus data from Statsmuse and minutes and team MOV data from basketball-reference, we can see that, from 2006-2021, LeBron James’s playoff on-off per 48 minutes was +12.95 (it was +1310 in 11,035 minutes on, for a +5.70 ON, and -274 in 1,813 minutes off, for a -7.25 OFF). That is not “near +15.” Of course, perhaps we can give you the benefit of the doubt and assume you meant to say 2007-2021, since the 2006 playoffs bring down LeBron’s numbers in this regard and it is actually 2007-2021 that is LeBron’s best span in playoff on-off. But even then, from 2007-2021, LeBron’s on-off per 48 minutes was +13.84. I would say it is materially misleading to call that “near +15.” Jordan’s career playoff on-off could perhaps accurately be described as “near +15,” but even picking out his best span LeBron’s could not. Not sure if the misleading statement was intentional, but either way it is worth correcting here.

[DraymondGold]

That is factually inaccurate. Using plus-minus data from Statsmuse and minutes and team MOV data from basketball-reference, we can see that, from 2006-2021, LeBron James’s playoff on-off per 48 minutes was +12.95 (it was +1310 in 11,035 minutes on, for a +5.70 ON, and -274 in 1,813 minutes off, for a -7.25 OFF). That is not “near +15.” Of course, perhaps we can give you the benefit of the doubt and assume you meant to say 2007-2021, since the 2006 playoffs bring down LeBron’s numbers in this regard and it is actually 2007-2021 that is LeBron’s best span in playoff on-off. But even then, from 2007-2021, LeBron’s on-off per 48 minutes was +13.84. I would say it is materially misleading to call that “near +15.” Jordan’s career playoff on-off could perhaps accurately be described as “near +15,” but even picking out his best span LeBron’s could not. Not sure if the misleading statement was intentional, but either way it is worth correcting here.

Hey jake! Agreed it's not +15

Just for another timeframe, we could also compare over 14-year spans. Since Jordan played over a 14 year span with 1 missed year (1985–1998 missing 1995), we might look at LeBron over a 14 year span with 1 missed year (2007–2020 missing 2019 or 2008–2021 missing 2019).

Per Djoker's chart, Jordan's +14.6.
2007–2020 LeBron: +13.02
2008–2021 LeBron: +12.625
Details:

Spoiler:

source: pbpstats, 'wowy' page.

Calculation details:
2007–2020
On:
Cle: +765 in 5817 mins
Mia: +424 in 3629 mins
Lal: 165 in 762 mins
=1354 total
on: (765 + 424 + 165)*48/(5817 + 3629 + 762)
= +6.367 per 48

Off:
Cle: -195 in 895 mins
Mia: -22 in 577 mins
Lal: -21 in 246 mins
=-238 total
off: -6.650 per 48

total on/off: +13.02 per 48

2008–2021
On:
Cle: +694 in 4924 mins
Mia: +424 in 3629 mins
Lal: +175 in 985 mins
on: (694 + 424 + 175)*48/(4924 + 3629 + 985)
= +6.507 per 48

off:
Cle: -167 in 818 mins
Mia: -22 in 577 mins
Lal: -70 in 637 mins
off: (-167 + -22 + -70)*48/(818 + 577 + 637)
= - 6.118 per 48

on-off: +12.625

. One small detail: looks like we get very tiny disagreements between pbpstats and statsmuse, such that pbpstats gives LeBron a +5 better total on score in this span. In the Miami span for example, statues says his team was +420 total when he was on (https://www.statmuse.com/nba/ask?q=what+was+lebron%27s+total+plus+minus+in+the+playoffs+from+2010-11+to+2013-14) while pbpstats says he was +424 (https://www.pbpstats.com/wowy/nba?0Exactly1OnFloor=2544&TeamId=1610612748&Season=2010-11,2011-12,2012-13,2013-14&SeasonType=Playoffs&Type=Team). Really quite small stuff so won't make a big difference, but one wonders where the disagreement comes from. Pbpstats is the more respected website so I would have a slight preference to trust it first, but without knowing what the disagreement comes from, it's hard to tell for sure, and there's definitely a 'ease of usability' benefit to the quicker search from statmuse.

Can we take much from this? Not so sure. Me personally, I have peak/prime Jordan slightly over LeBron with LeBron having better longevity, and this on/off data definitely supports that. Still, they're close, and there's plenty of limits and uncertainties with on/off, so it's far from the end-all and be-all.

Hoping this doesn't just turn into another LBJ MJ thread, so wrapping in some other players...

Ya know, I'm not fully up on what these numbers would look like /48 for other notables (Shaq/Duncan/KG/LeBron/Steph maybe?).

Anybody have any of those sitting around?

Here you go!

Would you mind sharing a RS one as well?

Great stuff Djoker! Agreed RS would be interesting. Some sort of more restrictive timespan (e.g. best 8, 10, 12, or 15 years, or whatever) to decrease the unequal effects of differing number of non-prime years would also be cool, but I'm sure these plots take a lot of work, so no worries if not Djoker!

In the plus-minus era, LeBron, Garnett, and Curry seem to dominate the on/off data in the regular season, so it's interesting that they also dominate in the playoffs. Curry and Garnett's playoff reputation vs the other top 15 candidates (Kobe/Jokic/Shaq) would give you the impression that Curry and Garnett would drop off more, but they're still clearly a level up.

Curry's two worst playoff on/off's are in his injured years (2016 and 2018), which would have otherwise been right in his 4-year peak stretch... I wonder how much higher his would have been wth just two fewer injuries. Likewise Garnett was on pretty putrid teams right between two of his best seasons in 04 and 08, and missed the playoffs in all 3 years. Alas.

Shaq and Duncan look close in on/off in career regular season (slight advantage Shaq in available data... but missing some prime and pre-prime seasons for Shaq), but Shaq seems to jump a level above Duncan in the playoffs. Indeed over 10 year primes among available top ~25 candidates, Shaq is one of the 2 biggest improvers in playoff on/off. Not sure how much this is due to the resilience of Shaq's playstyle, vs having him taking days off in the regular season (e.g. due to poorer conditioning/effort) and fully committing in the playoffs. But it's interesting nonetheless.

[Djoker]

Ya know, I'm not fully up on what these numbers would look like /48 for other notables (Shaq/Duncan/KG/LeBron/Steph maybe?).

Anybody have any of those sitting around?

Here you go!

Would you mind sharing a RS one as well?

I never made a RS one because the MJ sample is partial and not evenly sampled either.

[OhayoKD]

Ya know, I'm not fully up on what these numbers would look like /48 for other notables (Shaq/Duncan/KG/LeBron/Steph maybe?).

Anybody have any of those sitting around?

Lebron is +11 for his career and also near +15 over a similar stretch(though more minutes) from 2006-2021. No idea about the rest.

That is factually inaccurate. Using plus-minus data from Statsmuse and minutes and team MOV data from basketball-reference, we can see that, from 2006-2021, LeBron James’s playoff on-off per 48 minutes was +12.95 (it was +1310 in 11,035 minutes on, for a +5.70 ON, and -274 in 1,813 minutes off, for a -7.25 OFF). That is not “near +15.” Of course, perhaps we can give you the benefit of the doubt and assume you meant to say 2007-2021, since the 2006 playoffs bring down LeBron’s numbers in this regard and it is actually 2007-2021 that is LeBron’s best span in playoff on-off. But even then, from 2007-2021, LeBron’s on-off per 48 minutes was +13.84. I would say it is materially misleading to call that “near +15.” Jordan’s career playoff on-off could perhaps accurately be described as “near +15,” but even picking out his best span LeBron’s could not. Not sure if the misleading statement was intentional, but either way it is worth correcting here.

your correction is correct.

[lessthanjake]

...

Alright so I've been able to do some serious data clean-up. In this post I will summarize the updated data and point out a few small inconsistencies I see in the OP.

Regular Season

I looked at the Raw Plus Minus spreadsheet that I think Eminence also posted in this thread and realized that there is MJ regular season data for 1994-95 on there for the 17 games he played: +91 ON 668 ON Mins +26 OFF 153 OFF Mins. 1994-95 regular season can be added to the OP.

Can you point me to where you got this 1994-1995 regular season data? I think you’re referring to the thread Eminence linked to (https://forums.realgm.com/boards/viewtopic.php?f=344&t=1343246), but the links in the OP of that thread don’t work anymore, and I can’t seem to find this data in the the other posts in the thread (but definitely may be missing it!).

I noticed that the OP has different ON and OFF values for the 1987-88 season. Like mine, I think his data only includes the 42 full games and excludes the one game fragment on 01-09-88 so not sure why it's different?

I looked at it again and I too get +205 for the Jordan plus-minus. But the team MOV I get after counting three times kept being +104 (which would lead to -101 off value). You got +100 for the team MOV. Not sure what causes the discrepancy there, since I too am not counting the partial game with the 20-16 score.

I also have a different OFF for the 1992-93 season. For the 63 sampled games, the margin is +414 so since the ON is +489 the OFF value should be -75. Please confirm.

I’ve checked and you are definitely right. Not sure how I was that far off on the OFF value, but it should definitely be -75. Will change that in the OP.

Playoffs

First off I noticed a typo in the OP. The 1985 playoffs should have an OFF value of -32 and not -22.

You are absolutely correct. I must’ve accidentally used the total MOV as the OFF value. It should be -32. Will change that.

I logged the Knicks and Pistons series for 1989 because the Thinking Basketball (TB) numbers didn't match the total score margins.

Knicks: +45 ON -20 OFF (was +65 ON -30 OFF)
Pistons: -16 ON -9 OFF (was -19 OFF +12 OFF)

1989 Total: +51 ON -47 OFF (was +63 ON -41 OFF)

Thanks! I’ll change that, and update the methodology section to note where the numbers come from.

I also logged the Sixers series for 1990 because the TB numbers again didn't match the total score margins.

Sixers: +45 ON -7 OFF (was +45 ON -9 OFF)

1990 Total: +100 ON -45 OFF (was +100 ON -47 OFF)

Thanks for this too! Will update for this too.

In the 1995 playoffs, I found an arithmetic error in Game 1 of the Hornets series that I logged. Instead of +11 ON -3 OFF it's actually +6 ON +2 OFF.

1995 Total: -14 ON +24 OFF (was -9 ON +19 OFF)

Ditto for this as well!

[Djoker]

...

Alright so I've been able to do some serious data clean-up. In this post I will summarize the updated data and point out a few small inconsistencies I see in the OP.

Regular Season

I looked at the Raw Plus Minus spreadsheet that I think Eminence also posted in this thread and realized that there is MJ regular season data for 1994-95 on there for the 17 games he played: +91 ON 668 ON Mins +26 OFF 153 OFF Mins. 1994-95 regular season can be added to the OP.

Can you point me to where you got this 1994-1995 regular season data? I think you’re referring to the thread Eminence linked to (https://forums.realgm.com/boards/viewtopic.php?f=344&t=1343246), but the links in the OP of that thread don’t work anymore, and I can’t seem to find this data in the the other posts in the thread (but definitely may be missing it!).

I noticed that the OP has different ON and OFF values for the 1987-88 season. Like mine, I think his data only includes the 42 full games and excludes the one game fragment on 01-09-88 so not sure why it's different?

I looked at it again and I too get +205 for the Jordan plus-minus. But the team MOV I get after counting three times kept being +104 (which would lead to -101 off value). You got +100 for the team MOV. Not sure what causes the discrepancy there, since I too am not counting the partial game with the 20-16 score.

I also have a different OFF for the 1992-93 season. For the 63 sampled games, the margin is +414 so since the ON is +489 the OFF value should be -75. Please confirm.

I’ve checked and you are definitely right. Not sure how I was that far off on the OFF value, but it should definitely be -75. Will change that in the OP.

Playoffs

First off I noticed a typo in the OP. The 1985 playoffs should have an OFF value of -32 and not -22.

You are absolutely correct. I must’ve accidentally used the total MOV as the OFF value. It should be -32. Will change that.

I logged the Knicks and Pistons series for 1989 because the Thinking Basketball (TB) numbers didn't match the total score margins.

Knicks: +45 ON -20 OFF (was +65 ON -30 OFF)
Pistons: -16 ON -9 OFF (was -19 OFF +12 OFF)

1989 Total: +51 ON -47 OFF (was +63 ON -41 OFF)

Thanks! I’ll change that, and update the methodology section to note where the numbers come from.

I also logged the Sixers series for 1990 because the TB numbers again didn't match the total score margins.

Sixers: +45 ON -7 OFF (was +45 ON -9 OFF)

1990 Total: +100 ON -45 OFF (was +100 ON -47 OFF)

Thanks for this too! Will update for this too.

In the 1995 playoffs, I found an arithmetic error in Game 1 of the Hornets series that I logged. Instead of +11 ON -3 OFF it's actually +6 ON +2 OFF.

1995 Total: -14 ON +24 OFF (was -9 ON +19 OFF)

Ditto for this as well!

The link to the spreadsheet that includes the 1994-95 data is here:

https://docs.google.com/spreadsheets/d/1Z-0DQCSQr5vMLUp1zY4DbsmzBM4UFghqG4dc1eHfBh4/edit#gid=1672058251

I think only the 1987-88 regular season is a point of difference between us. I checked yet again and still get MOV of +100. It's +104 with the 20-16 fragment but we are not counting the fragment.

[lessthanjake]

Alright so I've been able to do some serious data clean-up. In this post I will summarize the updated data and point out a few small inconsistencies I see in the OP.

Regular Season

I looked at the Raw Plus Minus spreadsheet that I think Eminence also posted in this thread and realized that there is MJ regular season data for 1994-95 on there for the 17 games he played: +91 ON 668 ON Mins +26 OFF 153 OFF Mins. 1994-95 regular season can be added to the OP.

Can you point me to where you got this 1994-1995 regular season data? I think you’re referring to the thread Eminence linked to (https://forums.realgm.com/boards/viewtopic.php?f=344&t=1343246), but the links in the OP of that thread don’t work anymore, and I can’t seem to find this data in the the other posts in the thread (but definitely may be missing it!).

I noticed that the OP has different ON and OFF values for the 1987-88 season. Like mine, I think his data only includes the 42 full games and excludes the one game fragment on 01-09-88 so not sure why it's different?

I looked at it again and I too get +205 for the Jordan plus-minus. But the team MOV I get after counting three times kept being +104 (which would lead to -101 off value). You got +100 for the team MOV. Not sure what causes the discrepancy there, since I too am not counting the partial game with the 20-16 score.

I also have a different OFF for the 1992-93 season. For the 63 sampled games, the margin is +414 so since the ON is +489 the OFF value should be -75. Please confirm.

I’ve checked and you are definitely right. Not sure how I was that far off on the OFF value, but it should definitely be -75. Will change that in the OP.

Playoffs

First off I noticed a typo in the OP. The 1985 playoffs should have an OFF value of -32 and not -22.

You are absolutely correct. I must’ve accidentally used the total MOV as the OFF value. It should be -32. Will change that.

I logged the Knicks and Pistons series for 1989 because the Thinking Basketball (TB) numbers didn't match the total score margins.

Knicks: +45 ON -20 OFF (was +65 ON -30 OFF)
Pistons: -16 ON -9 OFF (was -19 OFF +12 OFF)

1989 Total: +51 ON -47 OFF (was +63 ON -41 OFF)

Thanks! I’ll change that, and update the methodology section to note where the numbers come from.

I also logged the Sixers series for 1990 because the TB numbers again didn't match the total score margins.

Sixers: +45 ON -7 OFF (was +45 ON -9 OFF)

1990 Total: +100 ON -45 OFF (was +100 ON -47 OFF)

Thanks for this too! Will update for this too.

In the 1995 playoffs, I found an arithmetic error in Game 1 of the Hornets series that I logged. Instead of +11 ON -3 OFF it's actually +6 ON +2 OFF.

1995 Total: -14 ON +24 OFF (was -9 ON +19 OFF)

Ditto for this as well!

The link to the spreadsheet that includes the 1994-95 data is here:

https://docs.google.com/spreadsheets/d/1Z-0DQCSQr5vMLUp1zY4DbsmzBM4UFghqG4dc1eHfBh4/edit#gid=1672058251

I think only the 1987-88 regular season is a point of difference between us. I checked yet again and still get MOV of +100. It's +104 with the 20-16 fragment but we are not counting the fragment.

Thanks! I’ve made all the changes in the OP now. That includes the 1987-1988 regular season changes. I recounted it and got a +100 MOV. I’m pretty sure I’d previously been consistently counting a 94-106 score as a -8

[NoParticular]

I think there’s a pretty good argument that the four-year period from 1987-1988 to 1990-1991 is Michael Jordan’s peak individually. After all, it’s the period of time where he peaks in terms of box stats like PER, win shares per 48, and BPM in both regular season and playoffs. Of course, there’s arguments for other time periods for Jordan, especially since the 1987-1988 to 1990-1991 time period wasn’t his most successful in terms of team success. But let’s assume for now that his box-score-metric peak was his peak.

It occurs to me that because the two biggest Squared samples are from that time period, we actually have a lot of on-off data for Jordan for that time period. Indeed, between regular season and playoffs, we have on-off data for 41% of Jordan’s games (158 out of 387 games) in these peak years, and we can combine all that data together. So does the on-off data back up that Jordan was incredible in that time period? Well, in short, yes:

Peak Jordan On-Off per 100 Possessions in Regular Season + Playoffs (1987-1988 to 1990-1991) (sample: 158 out of 387 games)

- On: +7.63
- Off: -13.35
- On-Off: +20.98

In other words, we have data for almost half of peak Jordan’s games, and he’s basically got a +21 on-off per 100 possessions in those games. It’s not all the games, of course, so the actual overall number might differ a bit from that, but this is suggestive of peak Jordan perhaps being the most impactful player ever, as I’m not aware of anyone with that kind of on-off over a four-year span (the closest I’m aware of is Steph from 2014-2017, who had about a +17 RS+PS on-off, and with a higher “on” value). And of course, we can map this onto the fact that this is also probably the highest four-year peak in terms of advanced box measures. Certainly seems supportive of it being the highest peak ever.

With all the updated data we have, is this still accurate?

[lessthanjake]

I think there’s a pretty good argument that the four-year period from 1987-1988 to 1990-1991 is Michael Jordan’s peak individually. After all, it’s the period of time where he peaks in terms of box stats like PER, win shares per 48, and BPM in both regular season and playoffs. Of course, there’s arguments for other time periods for Jordan, especially since the 1987-1988 to 1990-1991 time period wasn’t his most successful in terms of team success. But let’s assume for now that his box-score-metric peak was his peak.

It occurs to me that because the two biggest Squared samples are from that time period, we actually have a lot of on-off data for Jordan for that time period. Indeed, between regular season and playoffs, we have on-off data for 41% of Jordan’s games (158 out of 387 games) in these peak years, and we can combine all that data together. So does the on-off data back up that Jordan was incredible in that time period? Well, in short, yes:

Peak Jordan On-Off per 100 Possessions in Regular Season + Playoffs (1987-1988 to 1990-1991) (sample: 158 out of 387 games)

- On: +7.63
- Off: -13.35
- On-Off: +20.98

In other words, we have data for almost half of peak Jordan’s games, and he’s basically got a +21 on-off per 100 possessions in those games. It’s not all the games, of course, so the actual overall number might differ a bit from that, but this is suggestive of peak Jordan perhaps being the most impactful player ever, as I’m not aware of anyone with that kind of on-off over a four-year span (the closest I’m aware of is Steph from 2014-2017, who had about a +17 RS+PS on-off, and with a higher “on” value). And of course, we can map this onto the fact that this is also probably the highest four-year peak in terms of advanced box measures. Certainly seems supportive of it being the highest peak ever.

With all the updated data we have, is this still accurate?

So, the answer to that question is largely yes, but not exactly. We do not actually have data on any additional games than we did when I posted that. It’s still the same 158 games. However, since that post was made, there’s a few relevant things that have changed in how I’m looking at the data: (1) we got the exact playoff data from Thinking Basketball, rather than estimating based on the YouTube video; (2) I’ve shifted to using Djoker’s tracking for two of the playoffs, because the Thinking Basketball data didn’t add up to the right total margin of victory; and (3) I’ve shifted to putting all the data in per-48-minute terms, rather than making certain assumptions in order to put it into per-100-possession terms. The result of all those changes is that, in those same 158 games in his peak years, we have an on-off per 48 minutes of +18.09 for Jordan (+7.21 on; -10.88 off). That raw number is a little lower than what I’d posted earlier in the post you quote, but remember that we’d generally expect per-48-minute numbers to be a bit lower than per-100-possession numbers, since pace is usually below 100. So basically, while the numbers are slightly different and reported out in a different format (i.e. per-48 as opposed to per-100-possessions), what we have now is still quite similar to the above, and definitely still indicative of unbelievably high impact that I think is supportive of it being the GOAT peak.

[VanWest82]

I think 88-91 is the four year regular season peak. I'd say 89-92 is the post season peak. I know that isn't necessarily supported in the net on-off data but I'm pretty sure it's right. 88 Jordan was like 09 Lebron: unbelievably dominant player but there were some weaknesses if we're going to nitpick.

[NoParticular]

I think there’s a pretty good argument that the four-year period from 1987-1988 to 1990-1991 is Michael Jordan’s peak individually. After all, it’s the period of time where he peaks in terms of box stats like PER, win shares per 48, and BPM in both regular season and playoffs. Of course, there’s arguments for other time periods for Jordan, especially since the 1987-1988 to 1990-1991 time period wasn’t his most successful in terms of team success. But let’s assume for now that his box-score-metric peak was his peak.

It occurs to me that because the two biggest Squared samples are from that time period, we actually have a lot of on-off data for Jordan for that time period. Indeed, between regular season and playoffs, we have on-off data for 41% of Jordan’s games (158 out of 387 games) in these peak years, and we can combine all that data together. So does the on-off data back up that Jordan was incredible in that time period? Well, in short, yes:

Peak Jordan On-Off per 100 Possessions in Regular Season + Playoffs (1987-1988 to 1990-1991) (sample: 158 out of 387 games)

- On: +7.63
- Off: -13.35
- On-Off: +20.98

In other words, we have data for almost half of peak Jordan’s games, and he’s basically got a +21 on-off per 100 possessions in those games. It’s not all the games, of course, so the actual overall number might differ a bit from that, but this is suggestive of peak Jordan perhaps being the most impactful player ever, as I’m not aware of anyone with that kind of on-off over a four-year span (the closest I’m aware of is Steph from 2014-2017, who had about a +17 RS+PS on-off, and with a higher “on” value). And of course, we can map this onto the fact that this is also probably the highest four-year peak in terms of advanced box measures. Certainly seems supportive of it being the highest peak ever.

With all the updated data we have, is this still accurate?

So, the answer to that question is largely yes, but not exactly. We do not actually have data on any additional games than we did when I posted that. It’s still the same 158 games. However, since that post was made, there’s a few relevant things that have changed in how I’m looking at the data: (1) we got the exact playoff data from Thinking Basketball, rather than estimating based on the YouTube video; (2) I’ve shifted to using Djoker’s tracking for two of the playoffs, because the Thinking Basketball data didn’t add up to the right total margin of victory; and (3) I’ve shifted to putting all the data in per-48-minute terms, rather than making certain assumptions in order to put it into per-100-possession terms. The result of all those changes is that, in those same 158 games in his peak years, we have an on-off per 48 minutes of +18.09 for Jordan (+7.21 on; -10.88 off). That raw number is a little lower than what I’d posted earlier in the post you quote, but remember that we’d generally expect per-48-minute numbers to be a bit lower than per-100-possession numbers, since pace is usually below 100. So basically, while the numbers are slightly different and reported out in a different format (i.e. per-48 as opposed to per-100-possessions), what we have now is still quite similar to the above, and definitely still indicative of unbelievably high impact that I think is supportive of it being the GOAT peak.

Thanks a ton for all your work! This is great data. With that said, I think giving date for some players 3 year peaks and their primes as comparison for MJ's data could be useful to properly contextualise these stats.

[lessthanjake]

With all the updated data we have, is this still accurate?

So, the answer to that question is largely yes, but not exactly. We do not actually have data on any additional games than we did when I posted that. It’s still the same 158 games. However, since that post was made, there’s a few relevant things that have changed in how I’m looking at the data: (1) we got the exact playoff data from Thinking Basketball, rather than estimating based on the YouTube video; (2) I’ve shifted to using Djoker’s tracking for two of the playoffs, because the Thinking Basketball data didn’t add up to the right total margin of victory; and (3) I’ve shifted to putting all the data in per-48-minute terms, rather than making certain assumptions in order to put it into per-100-possession terms. The result of all those changes is that, in those same 158 games in his peak years, we have an on-off per 48 minutes of +18.09 for Jordan (+7.21 on; -10.88 off). That raw number is a little lower than what I’d posted earlier in the post you quote, but remember that we’d generally expect per-48-minute numbers to be a bit lower than per-100-possession numbers, since pace is usually below 100. So basically, while the numbers are slightly different and reported out in a different format (i.e. per-48 as opposed to per-100-possessions), what we have now is still quite similar to the above, and definitely still indicative of unbelievably high impact that I think is supportive of it being the GOAT peak.

Thanks a ton for all your work! This is great data. With that said, I think giving date for some players 3 year peaks and their primes as comparison for MJ's data could be useful to properly contextualise these stats.

Okay, so here’s some data on other four-year impact peaks of players in the play-by-play era (using four-year since that 1988-1991 timeframe for Jordan is four years, not three). I used StatsMuse to get the players’ plus-minus data, and then used basketball-reference to get the team MOVs and the minutes data:

- LeBron’s 2008-2009 to 2011-2012 time period is both LeBron’s box score peak and his four-year impact peak as well. In that four-year span, he had a +14.52 on-off per 48 minutes in regular season + playoffs (+10.05 on; -4.47 off). (Note: His 2013-2014 to 2016-2017 time period was not his box score peak, but in terms of on-off he was only slightly below the 2009-2012 time period, with a +14.36 on-off per 48 minutes in that timespan).

- Steph’s peak four years were 2014-2015 to 2017-2018. In that four-year span, he had a +17.08 on-off per 48 minutes (+15.94 on; -1.14 off). I should note that if we instead took 2013-2014 to 2016-2017, Steph’s on-off per 48 minutes goes up to +17.99. I don’t think that actually represents his peak four years as a player, but it is his peak four years in terms of on-off.

- For Kevin Garnett, his four-year impact peak was 2000-2001 to 2003-2004 (I’d say his actual peak as a player may have been more like 2001-2002 to 2004-2005, but the 2004-2005 on-off was not good at all, so I’ll use the better on-off period for these purposes). In those four years, Garnett had a +15.47 on-off per 48 minutes (+5.26 on; -10.21 off).

- In the last four years (including this year), Jokic has an on-off per 48 minutes of +16.85 (+9.18 on; -7.67 off)

I don’t think there’s any player that we have on-off data for that has a better on-off than the above guys. So basically, the data we have on Jordan’s four-year peak has a higher on-off per 48 minutes than any player has gotten in a four-year span in the play-by-play era. So I think we can certainly say that this info is supportive of Jordan having had the GOAT peak. That said, there’s two important caveats. First, we do not have the full data for Jordan in those years, but rather a sample that makes up about 41% of his games in those years. So we do not know exactly what his full on-off was. Second, the “on” rating is important context here, since it gets harder to push your team’s net rating up the higher that net rating goes. So, while Jordan’s +18 on-off might be higher than Steph’s +17 on-off, I think it’d actually be reasonable to be more impressed by Steph’s +17, because it comes with a significantly higher “on” rating. So overall, it’s definitely far from conclusive data (and hopefully we can eventually fill it out with data from the entire set of games), but it is supportive info for Jordan.

[Superjohnstarks]

Yeah, not sure about those 92 marks:

There is no way that +45 on/off net rating is remotely accurate.

1. Dipper had similarly high on/off estimates for Barkley (+36.3) and Hakeem (+34.5) but those estimates were disproven with Harvey Pollack's full season data.

Barkley net on/off
1987: +7.9
1988: +2.7
1989: +11.0
1990: +8.3
1991: +8.8
1992: +6.0
1993: ?
1994: +6.8
1995: +6.8
1996: +7.8

Hakeem net on/off
1994: +14.5
1995: +9.0
1996: +10.3

In none of the 12 full seasons did they come anywhere near the +35 on/off estimate over the same period.

2. There has only been a handful of +20 on/off net seasons in the last 25 years, none above +25. A +35 or + 45 on/off was never realistic in the first place, especially not by 3 players at the same time in just the few seasons prior. This was most likely just a calculation error somewhere in Dipper's spreadsheet which was repeated when he plugged in new numbers for different players.

Taking it at face value anyway...

Excellent work, man, thanks!

So, MJs career playoffs on-off per 100 poss is 14,7 with 7.55 ON. Maaan... LeBrons is 10,2 with 5,9 ON. Its not close actually the more we use analytical approach and impact data... In regular season I suppose the margin would be even wider because of "coasting".

Make that 2006-2021 Lebron and the playoff stuff is comparable with a 1994 Bull-sized elephant missing and the regular-season stuff being comparable without 94 or the bulk of 95.

That is also despite Lebron playing a much larger minute sample and staggering significantly more with co-stars.

IOW, nice try

I'm a bit confused, is the 1995 playoffs included in the sample??u said lebron has a comparable value in the playoffs from 2006 to 2021 which is essentially jordan entire playoff range. A source will be nice on that including the minutes disparity of their various co stars

[Superjohnstarks]

So, the answer to that question is largely yes, but not exactly. We do not actually have data on any additional games than we did when I posted that. It’s still the same 158 games. However, since that post was made, there’s a few relevant things that have changed in how I’m looking at the data: (1) we got the exact playoff data from Thinking Basketball, rather than estimating based on the YouTube video; (2) I’ve shifted to using Djoker’s tracking for two of the playoffs, because the Thinking Basketball data didn’t add up to the right total margin of victory; and (3) I’ve shifted to putting all the data in per-48-minute terms, rather than making certain assumptions in order to put it into per-100-possession terms. The result of all those changes is that, in those same 158 games in his peak years, we have an on-off per 48 minutes of +18.09 for Jordan (+7.21 on; -10.88 off). That raw number is a little lower than what I’d posted earlier in the post you quote, but remember that we’d generally expect per-48-minute numbers to be a bit lower than per-100-possession numbers, since pace is usually below 100. So basically, while the numbers are slightly different and reported out in a different format (i.e. per-48 as opposed to per-100-possessions), what we have now is still quite similar to the above, and definitely still indicative of unbelievably high impact that I think is supportive of it being the GOAT peak.

Thanks a ton for all your work! This is great data. With that said, I think giving date for some players 3 year peaks and their primes as comparison for MJ's data could be useful to properly contextualise these stats.

Okay, so here’s some data on other four-year impact peaks of players in the play-by-play era (using four-year since that 1988-1991 timeframe for Jordan is four years, not three). I used StatsMuse to get the players’ plus-minus data, and then used basketball-reference to get the team MOVs and the minutes data:

- LeBron’s 2008-2009 to 2011-2012 time period is both LeBron’s box score peak and his four-year impact peak as well. In that four-year span, he had a +14.52 on-off per 48 minutes in regular season + playoffs (+10.05 on; -4.47 off). (Note: His 2013-2014 to 2016-2017 time period was not his box score peak, but in terms of on-off he was only slightly below the 2009-2012 time period, with a +14.36 on-off per 48 minutes in that timespan).

- Steph’s peak four years were 2014-2015 to 2017-2018. In that four-year span, he had a +17.08 on-off per 48 minutes (+15.94 on; -1.14 off). I should note that if we instead took 2013-2014 to 2016-2017, Steph’s on-off per 48 minutes goes up to +17.99. I don’t think that actually represents his peak four years as a player, but it is his peak four years in terms of on-off.

- For Kevin Garnett, his four-year impact peak was 2000-2001 to 2003-2004 (I’d say his actual peak as a player may have been more like 2001-2002 to 2004-2005, but the 2004-2005 on-off was not good at all, so I’ll use the better on-off period for these purposes). In those four years, Garnett had a +15.47 on-off per 48 minutes (+5.26 on; -10.21 off).

- In the last four years (including this year), Jokic has an on-off per 48 minutes of +16.85 (+9.18 on; -7.67 off)

I don’t think there’s any player that we have on-off data for that has a better on-off than the above guys. So basically, the data we have on Jordan’s four-year peak has a higher on-off per 48 minutes than any player has gotten in a four-year span in the play-by-play era. So I think we can certainly say that this info is supportive of Jordan having had the GOAT peak. That said, there’s two important caveats. First, we do not have the full data for Jordan in those years, but rather a sample that makes up about 41% of his games in those years. So we do not know exactly what his full on-off was. Second, the “on” rating is important context here, since it gets harder to push your team’s net rating up the higher that net rating goes. So, while Jordan’s +18 on-off might be higher than Steph’s +17 on-off, I think it’d actually be reasonable to be more impressed by Steph’s +17, because it comes with a significantly higher “on” rating. So overall, it’s definitely far from conclusive data (and hopefully we can eventually fill it out with data from the entire set of games), but it is supportive info for Jordan.

Hi there,how do u calculate the on/off per 48 minuted and why not per 75 possesion which nen taylor outlines is the standard possesion star players play in post merger nba

[lessthanjake]

Thanks a ton for all your work! This is great data. With that said, I think giving date for some players 3 year peaks and their primes as comparison for MJ's data could be useful to properly contextualise these stats.

Okay, so here’s some data on other four-year impact peaks of players in the play-by-play era (using four-year since that 1988-1991 timeframe for Jordan is four years, not three). I used StatsMuse to get the players’ plus-minus data, and then used basketball-reference to get the team MOVs and the minutes data:

- LeBron’s 2008-2009 to 2011-2012 time period is both LeBron’s box score peak and his four-year impact peak as well. In that four-year span, he had a +14.52 on-off per 48 minutes in regular season + playoffs (+10.05 on; -4.47 off). (Note: His 2013-2014 to 2016-2017 time period was not his box score peak, but in terms of on-off he was only slightly below the 2009-2012 time period, with a +14.36 on-off per 48 minutes in that timespan).

- Steph’s peak four years were 2014-2015 to 2017-2018. In that four-year span, he had a +17.08 on-off per 48 minutes (+15.94 on; -1.14 off). I should note that if we instead took 2013-2014 to 2016-2017, Steph’s on-off per 48 minutes goes up to +17.99. I don’t think that actually represents his peak four years as a player, but it is his peak four years in terms of on-off.

- For Kevin Garnett, his four-year impact peak was 2000-2001 to 2003-2004 (I’d say his actual peak as a player may have been more like 2001-2002 to 2004-2005, but the 2004-2005 on-off was not good at all, so I’ll use the better on-off period for these purposes). In those four years, Garnett had a +15.47 on-off per 48 minutes (+5.26 on; -10.21 off).

- In the last four years (including this year), Jokic has an on-off per 48 minutes of +16.85 (+9.18 on; -7.67 off)

I don’t think there’s any player that we have on-off data for that has a better on-off than the above guys. So basically, the data we have on Jordan’s four-year peak has a higher on-off per 48 minutes than any player has gotten in a four-year span in the play-by-play era. So I think we can certainly say that this info is supportive of Jordan having had the GOAT peak. That said, there’s two important caveats. First, we do not have the full data for Jordan in those years, but rather a sample that makes up about 41% of his games in those years. So we do not know exactly what his full on-off was. Second, the “on” rating is important context here, since it gets harder to push your team’s net rating up the higher that net rating goes. So, while Jordan’s +18 on-off might be higher than Steph’s +17 on-off, I think it’d actually be reasonable to be more impressed by Steph’s +17, because it comes with a significantly higher “on” rating. So overall, it’s definitely far from conclusive data (and hopefully we can eventually fill it out with data from the entire set of games), but it is supportive info for Jordan.

Hi there,how do u calculate the on/off per 48 minuted and why not per 75 possesion which nen taylor outlines is the standard possesion star players play in post merger nba

The reason to do it on a per-48-minute basis is that a lot of this data we have for Jordan is raw plus-minus data that does not tell us how many possessions there were with Jordan on and off the court. So the only way to put it in per 100 possession terms (note: per-100-possessions is the standard way to measure on-off) would be to try to estimate the number of possessions on and off the court using the team’s overall pace and assuming equal pace with Jordan on and off the court. That’s what I originally did in the initial version of the OP, but people objected that the process of estimating per-100-possession data involved assumptions and uses of estimations that likely led to inaccuracy. I saw the point and didn’t want to have argument about that continue to take over the discussion on this, so I eventually converted the data to per-48-minute data—which does not involve any estimations or assumptions to calculate.

In terms of how per-48-minute data is calculated, it’s very simple. To get the “on” value, you take the plus-minus with the player on the court, divide by the number of minutes the player played, and multiply by 48. To get the “off” value, you take the plus-minus with the player off the court, divide by the number of minutes the player was off the court, and multiply by 48. And, of course, the “on-off” is just the “on” value minus the “off” value.

[Djoker]

Okay, so here’s some data on other four-year impact peaks of players in the play-by-play era (using four-year since that 1988-1991 timeframe for Jordan is four years, not three). I used StatsMuse to get the players’ plus-minus data, and then used basketball-reference to get the team MOVs and the minutes data:

- LeBron’s 2008-2009 to 2011-2012 time period is both LeBron’s box score peak and his four-year impact peak as well. In that four-year span, he had a +14.52 on-off per 48 minutes in regular season + playoffs (+10.05 on; -4.47 off). (Note: His 2013-2014 to 2016-2017 time period was not his box score peak, but in terms of on-off he was only slightly below the 2009-2012 time period, with a +14.36 on-off per 48 minutes in that timespan).

- Steph’s peak four years were 2014-2015 to 2017-2018. In that four-year span, he had a +17.08 on-off per 48 minutes (+15.94 on; -1.14 off). I should note that if we instead took 2013-2014 to 2016-2017, Steph’s on-off per 48 minutes goes up to +17.99. I don’t think that actually represents his peak four years as a player, but it is his peak four years in terms of on-off.

- For Kevin Garnett, his four-year impact peak was 2000-2001 to 2003-2004 (I’d say his actual peak as a player may have been more like 2001-2002 to 2004-2005, but the 2004-2005 on-off was not good at all, so I’ll use the better on-off period for these purposes). In those four years, Garnett had a +15.47 on-off per 48 minutes (+5.26 on; -10.21 off).

- In the last four years (including this year), Jokic has an on-off per 48 minutes of +16.85 (+9.18 on; -7.67 off)

I don’t think there’s any player that we have on-off data for that has a better on-off than the above guys. So basically, the data we have on Jordan’s four-year peak has a higher on-off per 48 minutes than any player has gotten in a four-year span in the play-by-play era. So I think we can certainly say that this info is supportive of Jordan having had the GOAT peak. That said, there’s two important caveats. First, we do not have the full data for Jordan in those years, but rather a sample that makes up about 41% of his games in those years. So we do not know exactly what his full on-off was. Second, the “on” rating is important context here, since it gets harder to push your team’s net rating up the higher that net rating goes. So, while Jordan’s +18 on-off might be higher than Steph’s +17 on-off, I think it’d actually be reasonable to be more impressed by Steph’s +17, because it comes with a significantly higher “on” rating. So overall, it’s definitely far from conclusive data (and hopefully we can eventually fill it out with data from the entire set of games), but it is supportive info for Jordan.

Hi there,how do u calculate the on/off per 48 minuted and why not per 75 possesion which nen taylor outlines is the standard possesion star players play in post merger nba

The reason to do it on a per-48-minute basis is that a lot of this data we have for Jordan is raw plus-minus data that does not tell us how many possessions there were with Jordan on and off the court. So the only way to put it in per 100 possession terms (note: per-100-possessions is the standard way to measure on-off) would be to try to estimate the number of possessions on and off the court using the team’s overall pace and assuming equal pace with Jordan on and off the court. That’s what I originally did in the initial version of the OP, but people objected that the process of estimating per-100-possession data involved assumptions and uses of estimations that likely led to inaccuracy. I saw the point and didn’t want to have argument about that continue to take over the discussion on this, so I eventually converted the data to per-48-minute data—which does not involve any estimations or assumptions to calculate.

In terms of how per-48-minute data is calculated, it’s very simple. To get the “on” value, you take the plus-minus with the player on the court, divide by the number of minutes the player played, and multiply by 48. To get the “off” value, you take the plus-minus with the player off the court, divide by the number of minutes the player was off the court, and multiply by 48. And, of course, the “on-off” is just the “on” value minus the “off” value.

Good explanation.

Also an issue is that different sources use different methodologies for counting possessions. Thus even in the play-by-play era which is since 1996-97, we get some discrepancies in the per 100 numbers. For instance compare pbpstats, basketball-reference and NBA stats. The ON-OFF numbers are different for all three and honestly I have not been able to reproduce any of their net ratings using raw plus minus and total possession count which I should be able to. It leads to many disagreements. Usually the discrepancies aren't huge but it is often 1-2 points per 100 which is kind of significant.

Thus better to give all numbers as per 48 minutes. In case you didn't see the comparison with other players, I'm posting it here.