Orlando Shaq Playoff Plus Minus and Impact Assessment

Moderators: Clyde Frazier, Doctor MJ, trex_8063, penbeast0, PaulieWal

Djoker
Starter
Posts: 2,102
And1: 1,799
Joined: Sep 12, 2015
 

Orlando Shaq Playoff Plus Minus and Impact Assessment 

Post#1 » by Djoker » Tue Sep 10, 2024 6:41 pm

PLUS MINUS

I tracked every one of Shaq's playoff games from his Magic tenure (36 in all) and here is the data.

Image

Image

Like with Jordan previously, I also compiled his career totals. He is now the second player before the PBP era for whom we have the entire playoff career plus minus.

Image

IMPACT ASSESSMENT

Offense

Unlike many great bigs in NBA history that had a mix of solid offensive and solid defensive impact (aka two-way bigs), Orlando Shaq was an offensive monster and the data support it in a big way. Across the eight series played, the Magic offenses were astonishingly dominant with Shaq on the court.

rORtg with Shaq ON/Shaq OFF

1994 IND: +3.3/-12.0
1995 BOS: +0.4/+11.9
1995 CHI: +10.4/+2.7
1995 IND: +17.0/+8.1
1995 HOU: +4.4/-18.1
1996 DET: +19.5/+16.1
1996 ATL: +20.8/-3.4
1996 CHI: +0.8/-29.5

Average: +9.6/-3.0

He was a load to handle down low and the defensive work began before he ever got the ball. The defense had a major job to do to prevent him from getting deep post position. If Shaq set up within 8 feet of the basket, it was basically over. He would get the basket and/or get fouled if the foolish defender tried to stop him once he went up. His touch, footwork, and the nasty spin moves and up and unders were just insane for a man that big. And you had to box him out because he could just destroy you on the offensive glass. Most teams doubled him when he caught the ball but he was a very good passer that could hit shooters for open threes but was even adept at finding cutters. His best passes were one handed bullets to cutters than ended up as layups or free throws for teammates. I consider him a very good passer in 1995 and 1996.

Shaq as a Laker improved even more as a passer and was able to make reads faster and punish double teams a bit more ruthlessly. This would help against elite defenses like the 1996 Bulls which were able to stymie Orlando's offense (Orlando only had a only +0.8 rORtg with Shaq on the court) by very quickly double-teaming and recovering after he passed the ball. The quickness of Harper/Jordan/Pippen is something most teams don't have and these guys were forcing Shaq into quite a few turnovers and broke up Orlando's plays afterwards as shooters didn't get their normally open looks. Slightly quicker reads could make a lot of difference.

Image

Defense

While ability to captain GOAT-level offense is pretty evident from the data, pinpointing his defensive impact carries much more uncertainty. The Magic defense was inconsistent and mostly below league average for the majority of the eight series. Negative numbers are good and positive numbers are bad.

rDRtg with Shaq ON/Shaq OFF

1994 IND: +2.4/+15.1
1995 BOS: -11.3/-0.6
1995 CHI: +2.9/+2.9
1995 IND: +7.3/+18.1
1995 HOU: +5.5/+27.2
1996 DET: +2.7/-5.7
1996 ATL: -0.1/+11.9
1996 CHI: -1.1/+42.7

Average: +1.0/+14.0

On one hand the defenses with Shaq being below league average by a point is strong indication that he is not an elite defensive anchor. On the other hand, the teams completely collapsing on defense without him suggest that he did likely provide at least some lift on that end.

Shaq did offer plenty of intimidation. When settled under the rim, penetrators would rarely challenge him so there was some deterrence at play. When he didn't have to move laterally, his athleticism allowed him to contest shots effectively at the rim. His verticality was excellent. In terms of individual post defense, he was fantastic and this is in my opinion his biggest strength defensively. In the 1995 Finals in particular, he did a terrific job defending Hakeem Olajuwon in the post. Hakeem's methodical attack rarely worked on Shaq who rarely jumped and simply used his superior vertical reach to bother Hakeem's shots. When trying to face him up further from the basket, Hakeem had more success as he blew by Shaq a few times (300 lbs --> slow feet) but that consumed more energy and was only an occasional strategy.

In terms of weaknesses, the main problem was excessive fouling. He gave up a ton of points by committing at times stupid hip check fouls (hoping the ref wouldn't notice?) and at the same time removing himself from the game. Shaq had 4 or more fouls in 21/36 (58%) of playoff games between 1994-1996 resulting in reduced minutes on the court which ate away at his impact. In latter years in LA, this is one era he cleaned up. During the threepeat years from 2000-2002, Shaq only had 4 or more fouls in 22/58 (38%) of playoff games. In Orlando, he also displayed poor awareness often, either failing to rotate or deciding to leave the paint and defend someone on the perimeter where he would get promptly beat off the dribble and the other team getting an uncontested rim finish. In the regular season, Orlando defenses actually fared better but in the postseason, these weaknesses of Shaq were exploited more aggressively by opposing coaches.

The sideburns confused or distracted opponents in 1994 though their actual effectiveness is disputed.

Image
User avatar
AEnigma
Assistant Coach
Posts: 4,086
And1: 5,921
Joined: Jul 24, 2022
 

Re: Orlando Shaq Playoff Plus Minus and Impact Assessment 

Post#2 » by AEnigma » Tue Sep 10, 2024 7:13 pm

Yeah you probably should not be doing pace adjustments like that. Right away I can see your 2001 value is off.

Nice work logging the 1994-96 games though.
User avatar
MacGill
Veteran
Posts: 2,766
And1: 565
Joined: May 29, 2010
Location: From Parts Unknown...
     

Re: Orlando Shaq Playoff Plus Minus and Impact Assessment 

Post#3 » by MacGill » Tue Sep 10, 2024 9:40 pm

After 9 pages of posters arguing the right way to interpret the numbers and being coached on what sets to use, we'll start to see the usual arguments against him.

Nice work Djoker with the post! :)
Image
DraymondGold
Senior
Posts: 590
And1: 763
Joined: May 19, 2022

Re: Orlando Shaq Playoff Plus Minus and Impact Assessment 

Post#4 » by DraymondGold » Wed Sep 11, 2024 1:06 am

Djoker wrote:PLUS MINUS
...
Image
Brilliant stuff as usual Djoker! This must have taken a massive amount of work. But as a reward... I think we now have the raw plus minus data for Shaq's entire career :D (edit: except for his rookie year regular season).

It looks like his dominant peak sample doesn't get any better. His best 3-year PS On/off rank is still ~4th all-time (min 200 off mins, which gets rid of e.g. 99-01 Ray Allen; one sample per player... Robinson and Garnett have 2 each). His 5-year PS On/off still ranks ~1st all-time (min 300 off-minutes... 86-90 Jordan becomes first if we lower the off-minute threshold any more).

But it does look like his phenomenal postseason prime run gets extended early, as expected based on the film. He was pretty great even in his sophomore season. I haven't checked rigorously, but my bet would be Shaq might be first all time in say ~8 year or 10 year PS on/off. He also might be in contention for the biggest postseason improver according to prime on/off among top 20–30 players. It's a noisy stat and the change might be noisier, but it definitely backs up his reputation as pretty resilient in the postseason (and lazy in the regular season).

Hopefully I'll get the chance to look more in depth at the details later this week, but just wanted to say thanks for doing this!
One_and_Done
General Manager
Posts: 8,625
And1: 5,429
Joined: Jun 03, 2023

Re: Orlando Shaq Playoff Plus Minus and Impact Assessment 

Post#5 » by One_and_Done » Wed Sep 11, 2024 2:32 am

As usual, take these fan made stats well salted.
Warspite wrote:Billups was a horrible scorer who could only score with an open corner 3 or a FT.
Ol Roy
Junior
Posts: 446
And1: 541
Joined: Dec 03, 2023

Re: Orlando Shaq Playoff Plus Minus and Impact Assessment 

Post#6 » by Ol Roy » Wed Sep 11, 2024 3:04 am

One_and_Done wrote:As usual, take these fan made stats well salted.


Posters who spend hours working to compile new data for the benefit of others, in good faith and for the love of the sport, don't deserve to be crapped on with your baseless, passive-aggressive insinuations.
One_and_Done
General Manager
Posts: 8,625
And1: 5,429
Joined: Jun 03, 2023

Re: Orlando Shaq Playoff Plus Minus and Impact Assessment 

Post#7 » by One_and_Done » Wed Sep 11, 2024 3:36 am

Ol Roy wrote:
One_and_Done wrote:As usual, take these fan made stats well salted.


Posters who spend hours working to compile new data for the benefit of others, in good faith and for the love of the sport, don't deserve to be crapped on with your baseless, passive-aggressive insinuations.

It's not about good faith, it's about the reliability of one fan trying to calculate complex volume stats in a way that isn't auditable and is subject to massive amounts of human error.
Warspite wrote:Billups was a horrible scorer who could only score with an open corner 3 or a FT.
User avatar
eminence
RealGM
Posts: 16,839
And1: 11,682
Joined: Mar 07, 2015

Re: Orlando Shaq Playoff Plus Minus and Impact Assessment 

Post#8 » by eminence » Wed Sep 11, 2024 3:57 am

While there's some small differences in how +/- can be tracked (mostly with FT substitutions), there's really nothing complex about raw +/- tracking. Error is of course possible and even guaranteed at some rate, but if we take the poster as baseline competent and well intentioned than I expect the data is fairly accurate.

Anywho, thanks Djoker.
I bought a boat.
70sFan
RealGM
Posts: 29,637
And1: 24,953
Joined: Aug 11, 2015
 

Re: Orlando Shaq Playoff Plus Minus and Impact Assessment 

Post#9 » by 70sFan » Wed Sep 11, 2024 6:48 am

Huge thanks for your work Djoker, fantastic post! I hope I will find time to discuss the differences between Orlando Shaq and Lakers Shaq, as I have analyzed over 40 2000-01 Shaq games (mostly in the PS).

Remarkable job!
70sFan
RealGM
Posts: 29,637
And1: 24,953
Joined: Aug 11, 2015
 

Re: Orlando Shaq Playoff Plus Minus and Impact Assessment 

Post#10 » by 70sFan » Wed Sep 11, 2024 6:49 am

One_and_Done wrote:
Ol Roy wrote:
One_and_Done wrote:As usual, take these fan made stats well salted.


Posters who spend hours working to compile new data for the benefit of others, in good faith and for the love of the sport, don't deserve to be crapped on with your baseless, passive-aggressive insinuations.

It's not about good faith, it's about the reliability of one fan trying to calculate complex volume stats in a way that isn't auditable and is subject to massive amounts of human error.

It's a lot of effort, but there is nothing complex about tracking raw +/-, it's not more complex than trakcing anything else really.
Owly
Lead Assistant
Posts: 5,614
And1: 3,132
Joined: Mar 12, 2010

Re: Orlando Shaq Playoff Plus Minus and Impact Assessment 

Post#11 » by Owly » Wed Sep 11, 2024 3:56 pm

70sFan wrote:
One_and_Done wrote:
Ol Roy wrote:
Posters who spend hours working to compile new data for the benefit of others, in good faith and for the love of the sport, don't deserve to be crapped on with your baseless, passive-aggressive insinuations.

It's not about good faith, it's about the reliability of one fan trying to calculate complex volume stats in a way that isn't auditable and is subject to massive amounts of human error.

It's a lot of effort, but there is nothing complex about tracking raw +/-, it's not more complex than trakcing anything else really.

Caveat... I haven't done this.
Maybe not "complex" but in terms of possibilities (and going slightly beyond +/- to on-off stuff)

- single individual's attention
- time limitations
- quality of footage to distinguish players.
- Accuracy of clock tracking (and there at the margin I wonder if multiple generations of broadcast to vhs to digitized to possibly re-encoded onto a host's platform could make it harder if time is marginally distorted)
- Degree of commitment/intent to accuracy in clock tracking - e.g. above all given in whole minutes. Were stints to the second and then rounded at the end of the game? Or are stints themselves rounded?
- occasional possibility of footage being incomplete

I remember "AScreaming"'s RAPM didn't have Dale and Antonio Davis for at least one year because presumably at least in some places the data was muddy.

Not all of these will apply to the case above (e.g. Shaq's fairly distinctive and by 90s footage we're less likely to be relying on 240 resolution off a grainy VHS). More generally the big picture stuff would probably be the more official it is the more hours, ability to check and willingness to check there may be and the further back one goes the hard it is to independently check. And whilst the poster says there's no insinuation with regard to intent here there is always a non-zero possibility ... some people do like to just lie on the internet. I'm not saying that's the case here but there's ... less cost than official sourced numbers put out by employees of the league.

I appreciate anyone willing to put the time in to this type of stuff. I wouldn't phrase it others have: "fan made stats" (or taking them "well salted). I can see some possibility of error ... well an easier path for more errors or imprecision and especially errors being less likely to be caught (officially generated stuff having better footage, more checks, more incentives to check). I think - to my mind - some greater level of caution is fine, perhaps sensible (though moreso ... it's a noisy measure from small samples anyhow so bear that in mind) ... but as stated I wouldn't have phrased it as above.
Djoker
Starter
Posts: 2,102
And1: 1,799
Joined: Sep 12, 2015
 

Re: Orlando Shaq Playoff Plus Minus and Impact Assessment 

Post#12 » by Djoker » Wed Sep 11, 2024 6:49 pm

AEnigma wrote:Yeah you probably should not be doing pace adjustments like that. Right away I can see your 2001 value is off.

Nice work logging the 1994-96 games though.


You reminded me of something that I've been wanting to post for a while. Per 100 estimates like ORtg/DRtg/Net Rtg vary quite a lot from different sources even though they all use the same raw possession data?!? My first thought is that it may be due to rounding but the relatively large differences make me think it may be the methodology that varies.

For instance, let's look at 2001 Shaq Net Rating in the PS.

B-Ref: +13.8 ON, +14.1 OFF
PBPStats: +13.98 ON, +12.58 OFF
NBA.com: +14.0 ON, +9.2 OFF (NBA.com gives only ON prior to 2008 but can easily estimate OFF from total team Net Rtg)
My Estimate: +14.4 ON, +9.7 OFF

ON ratings are quite stable but OFF ratings are quite different. Then again, OFF ratings for a single PS are so noisy anyways (97 mins in this example) that I wouldn't have too much confidence in them to begin with. Even for an entire career, Shaq has just 2772 OFF minutes so I'd still expect a lot of noise there. OFF samples have large errors.
User avatar
AEnigma
Assistant Coach
Posts: 4,086
And1: 5,921
Joined: Jul 24, 2022
 

Re: Orlando Shaq Playoff Plus Minus and Impact Assessment 

Post#13 » by AEnigma » Wed Sep 11, 2024 7:42 pm

I have never really understood the variance on NBA.com. BBRef is just estimating possession counts, which on average is okay but for the sake of this exercise does not work because you have no idea how the pace (or rather, pace estimate) changes when Shaq is in or out even if we were to take the series estimates as relatively accurate.
Djoker
Starter
Posts: 2,102
And1: 1,799
Joined: Sep 12, 2015
 

Re: Orlando Shaq Playoff Plus Minus and Impact Assessment 

Post#14 » by Djoker » Wed Sep 11, 2024 10:08 pm

Hmm I thought B-Ref only estimates possessions counts prior to PBP which is 1996-97. Why would they estimate when there are exact possession counts...

Anyways the pace ON and pace OFF are hardly different in most cases if you look at the official data we have so I doubt the assumption of equal pace is leading to any major inaccuracies in the 90's either.

But yea.. all per 100 metrics are estimates. ON ratings I would be very confident in and OFF ratings less so.
DraymondGold
Senior
Posts: 590
And1: 763
Joined: May 19, 2022

Re: Orlando Shaq Playoff Plus Minus and Impact Assessment 

Post#15 » by DraymondGold » Thu Sep 12, 2024 10:13 pm

Djoker wrote:
AEnigma wrote:Yeah you probably should not be doing pace adjustments like that. Right away I can see your 2001 value is off.

Nice work logging the 1994-96 games though.


You reminded me of something that I've been wanting to post for a while. Per 100 estimates like ORtg/DRtg/Net Rtg vary quite a lot from different sources even though they all use the same raw possession data?!? My first thought is that it may be due to rounding but the relatively large differences make me think it may be the methodology that varies.

For instance, let's look at 2001 Shaq Net Rating in the PS.

B-Ref: +13.8 ON, +14.1 OFF
PBPStats: +13.98 ON, +12.58 OFF
NBA.com: +14.0 ON, +9.2 OFF (NBA.com gives only ON prior to 2008 but can easily estimate OFF from total team Net Rtg)
My Estimate: +14.4 ON, +9.7 OFF

ON ratings are quite stable but OFF ratings are quite different. Then again, OFF ratings for a single PS are so noisy anyways (97 mins in this example) that I wouldn't have too much confidence in them to begin with. Even for an entire career, Shaq has just 2772 OFF minutes so I'd still expect a lot of noise there. OFF samples have large errors.
Thanks for this!

This kind of thing is pretty interesting, as it gives a sense of the uncertainty in possession estimates. It seems like with 676 on minutes (~1296 possessions), all 4 different methods get within 0.6 in ON per 100. In 97 off minutes (191 possessions), all 4 different methods get within 4.9 in Off per 100.

If you discount BBR which I believe assumes the same pace for On and Off and go with the actual possession estimates, the max differences are 0.42 in On and 3.38 in Off. It would be interesting to see how this scales (~ possessions^-2 ?) to get some sense for the error bars just from possession estimates as you get to say ~3-5 year samples.

I think I've shared this before, but pbpstats has https://darrylblackport.com/posts/2019-04-03-why-pbpstats-possession-counts-lower/ to discuss the different definition for possessions vs Nba.com. No discussion of bbref though.

We might also use something similar to get the error rate in a good hand tracker. E.g. colt18 hand tracked the 1995 finals and got the same value is you (https://forums.realgm.com/boards/viewtopic.php?t=2370986), which is definitely encouraging. You and I got slightly different results from one Jordan game we tracked (?), and I suppose we could do similar comparative studies with other times people have both hand tracked the same game.

Overall, it seems like we all tend to get similar answers when we do it, which should absolutely be encouraging, but having a quantitative error bar might help assuage the odd number-doubters out there. Of course hand tracking might not be perfect, but really it's just a question of whether there's some systematic bias and what the range/error bars are, and it seems like you're generally doing pretty good!

Getting back to Shaq, It's interesting to look at his playoff runs in order ranked by on/off:
Spoiler:
2000
2004
1994
2002
1996
1999
2008
1997
2003
1995
1998
2005
2001
2011
2010
2006
2007

And Ranked by the average in On and On/off (2x On/off weighting, like in the formula for AuPM, which is fit to RAPM, although also with the addition of box stats and some smart statistical techniques for smaller playoff samples)
Spoiler:
2000
2004
2002
1996
1994
2008
1999
1997
2003
1995
1998
2001
2005
2011
2006
2010
2007
As I've reminded people before, on/off is noisy enough that we absolutely shouldn't expect the best seasons to always come out on top. Still, 2000 being his best regular season and having the highest impact metrics here gives credence to it being his single-year peak.

2004 and 2002 are also generally high, which makes sense considering they're right in the middle of his prime, during his multi-year peak (at least 02 is), and on a team with good fit.

2001 is oddly low in on-off (with the tiny off sample we mentioned earlier), but the overall team was really excellent in those playoffs. Seems like there was good 3 point shooting both when he was on and off the court. Taken as part of the multi-year peak, 2001's nothing concerning given the noise of on/off and how good his multi-year on/off is. If you want to argue something causal, 3 point shooting variability's definitely it. I suppose the most favorable argument you could make for 01 Shaq in this context is that his gravity when he was on-court helped the shooters get hot, and they just stayed hot when he was off the court, which would help the overall team but diminish his on/off -- not sure how good of an argument, but it's an interesting thought.

It looks like 96/94 generally come next after 04/02. It's perhaps a slightly surprising that they're above 95, though nothing so crazy that noise couldn't explain it. 94 only has 18 off min autes, so the noise there is going to dominate. Regardless, Shaq gets a lot of his longevity value / career value from starting his prime earlier than a lot of players in his ~sophomore year, and the data seems to support it.

That said, his post-prime years weren't as good as other all-time players. People generally think he dropped off out of his prime in 05 and certainly by 06 -- which is a tad young for a big man, and the drop off seemed to be larger than other all-time players -- and the data seems to support that too. Some of it is of course how much Shaq used his athleticism rather than skill/IQ, which might age better. I wonder how much of it also comes from Shaq focusing on (athletic) offense over defense -- defensive focused bigs still retain some size and IQ as they get older, even if they lose mobility, which might be a model for how some bigs seem to have good late-career longevity.
Djoker
Starter
Posts: 2,102
And1: 1,799
Joined: Sep 12, 2015
 

Re: Orlando Shaq Playoff Plus Minus and Impact Assessment 

Post#16 » by Djoker » Fri Sep 13, 2024 3:28 am

DraymondGold wrote:
Djoker wrote:
AEnigma wrote:Yeah you probably should not be doing pace adjustments like that. Right away I can see your 2001 value is off.

Nice work logging the 1994-96 games though.


You reminded me of something that I've been wanting to post for a while. Per 100 estimates like ORtg/DRtg/Net Rtg vary quite a lot from different sources even though they all use the same raw possession data?!? My first thought is that it may be due to rounding but the relatively large differences make me think it may be the methodology that varies.

For instance, let's look at 2001 Shaq Net Rating in the PS.

B-Ref: +13.8 ON, +14.1 OFF
PBPStats: +13.98 ON, +12.58 OFF
NBA.com: +14.0 ON, +9.2 OFF (NBA.com gives only ON prior to 2008 but can easily estimate OFF from total team Net Rtg)
My Estimate: +14.4 ON, +9.7 OFF

ON ratings are quite stable but OFF ratings are quite different. Then again, OFF ratings for a single PS are so noisy anyways (97 mins in this example) that I wouldn't have too much confidence in them to begin with. Even for an entire career, Shaq has just 2772 OFF minutes so I'd still expect a lot of noise there. OFF samples have large errors.
Thanks for this!

This kind of thing is pretty interesting, as it gives a sense of the uncertainty in possession estimates. It seems like with 676 on minutes (~1296 possessions), all 4 different methods get within 0.6 in ON per 100. In 97 off minutes (191 possessions), all 4 different methods get within 4.9 in Off per 100.

If you discount BBR which I believe assumes the same pace for On and Off and go with the actual possession estimates, the max differences are 0.42 in On and 3.38 in Off. It would be interesting to see how this scales (~ possessions^-2 ?) to get some sense for the error bars just from possession estimates as you get to say ~3-5 year samples.

I think I've shared this before, but pbpstats has https://darrylblackport.com/posts/2019-04-03-why-pbpstats-possession-counts-lower/ to discuss the different definition for possessions vs Nba.com. No discussion of bbref though.

We might also use something similar to get the error rate in a good hand tracker. E.g. colt18 hand tracked the 1995 finals and got the same value is you (https://forums.realgm.com/boards/viewtopic.php?t=2370986), which is definitely encouraging. You and I got slightly different results from one Jordan game we tracked (?), and I suppose we could do similar comparative studies with other times people have both hand tracked the same game.

Overall, it seems like we all tend to get similar answers when we do it, which should absolutely be encouraging, but having a quantitative error bar might help assuage the odd number-doubters out there. Of course hand tracking might not be perfect, but really it's just a question of whether there's some systematic bias and what the range/error bars are, and it seems like you're generally doing pretty good!

Getting back to Shaq, It's interesting to look at his playoff runs in order ranked by on/off:
Spoiler:
2000
2004
1994
2002
1996
1999
2008
1997
2003
1995
1998
2005
2001
2011
2010
2006
2007

And Ranked by the average in On and On/off (2x On/off weighting, like in the formula for AuPM, which is fit to RAPM, although also with the addition of box stats and some smart statistical techniques for smaller playoff samples)
Spoiler:
2000
2004
2002
1996
1994
2008
1999
1997
2003
1995
1998
2001
2005
2011
2006
2010
2007
As I've reminded people before, on/off is noisy enough that we absolutely shouldn't expect the best seasons to always come out on top. Still, 2000 being his best regular season and having the highest impact metrics here gives credence to it being his single-year peak.

2004 and 2002 are also generally high, which makes sense considering they're right in the middle of his prime, during his multi-year peak (at least 02 is), and on a team with good fit.

2001 is oddly low in on-off (with the tiny off sample we mentioned earlier), but the overall team was really excellent in those playoffs. Seems like there was good 3 point shooting both when he was on and off the court. Taken as part of the multi-year peak, 2001's nothing concerning given the noise of on/off and how good his multi-year on/off is. If you want to argue something causal, 3 point shooting variability's definitely it. I suppose the most favorable argument you could make for 01 Shaq in this context is that his gravity when he was on-court helped the shooters get hot, and they just stayed hot when he was off the court, which would help the overall team but diminish his on/off -- not sure how good of an argument, but it's an interesting thought.

It looks like 96/94 generally come next after 04/02. It's perhaps a slightly surprising that they're above 95, though nothing so crazy that noise couldn't explain it. 94 only has 18 off min autes, so the noise there is going to dominate. Regardless, Shaq gets a lot of his longevity value / career value from starting his prime earlier than a lot of players in his ~sophomore year, and the data seems to support it.

That said, his post-prime years weren't as good as other all-time players. People generally think he dropped off out of his prime in 05 and certainly by 06 -- which is a tad young for a big man, and the drop off seemed to be larger than other all-time players -- and the data seems to support that too. Some of it is of course how much Shaq used his athleticism rather than skill/IQ, which might age better. I wonder how much of it also comes from Shaq focusing on (athletic) offense over defense -- defensive focused bigs still retain some size and IQ as they get older, even if they lose mobility, which might be a model for how some bigs seem to have good late-career longevity.


Nice post. Thanks for the link to the Darryl Blackport post. The methodology is indeed quite different between PBP and NBA.com as they count possessions differently. And for what it's worth I think I like the PBP method even though the cutoff for 2 seconds at the end of the quarter being a possession is completely arbitrary. Why not 1.5 seconds or 3 seconds or something else... but it's not like there is a perfect choice for a cutoff.

As for Shaq, yes his post-prime years are quite poor and bring down his averages. His 1994-2004 stretch which is 11 postseasons is just fantastic. +6.0 ON -14.0 OFF +20.0 ON-OFF.
User avatar
MacGill
Veteran
Posts: 2,766
And1: 565
Joined: May 29, 2010
Location: From Parts Unknown...
     

Re: Orlando Shaq Playoff Plus Minus and Impact Assessment 

Post#17 » by MacGill » Sun Sep 15, 2024 12:59 pm

Djoker wrote:As for Shaq, yes his post-prime years are quite poor and bring down his averages. His 1994-2004 stretch which is 11 postseasons is just fantastic. +6.0 ON -14.0 OFF +20.0 ON-OFF.


This is always why I never understand why some would get down on Shaq with longevity as a center, and I would always say that I'd put up his 92-06 years against almost anyone.
Image

Return to Player Comparisons