RealGM 2023 Top 100 Project - #7 (Wilt Chamberlain)

[ijspeelman]

I think these nominees can basically be divided into two different groups. Group A: Two-way players who excel on both sides of the ball but probably aren't good enough to carry a really good offense (Garnett, Chamberlain). Group B: Offensive centerpieces who can carry a good offense but are not super impactful on defense (Steph, Magic, Shaq).

On reasons of general principle, I consider the second group more valuable, because I think offensive centerpieces are rare and because I think they're more valuable in basketball. I tend to think that basketball offense is more of a strongest-link thing, and basketball defense is more of a weakest-link thing. All other things being equal, I think great offensive centerpieces carry more value for their teams. And I don't think the winning imprints of Garnett and Chamberlain really change that.

I think its funny because I agree with everything you've said besides that I find the first group more valuable to winning. Comparing each groups historical winning would tend to support your favoring, but if you include some past entries (Hakeem, Russell, Duncan) then I think there is an argument that the two way guys have a lot more impact even if their offense doesn't reach their insane defense.

I wouldn't necessarily put Hakeem or Duncan into either one of those categories. I don't think these are categories that all NBA players can be sorted into - it's just how I happen to feel about this particular group of 5 players.

Hakeem and Duncan both have scoring and offensive games that I think really highly of in addition to being super-elite defenders. Both of them proved that they can be centerpieces of playoff offenses on deep playoff runs, so I wouldn't compare Garnett or Wilt to them. And then with Russell, I'm a lot lower than most other people here on Russell. That might be wrong, but to me the same arguments I'm making here would apply to Russell. On general principles, for a playoff team I'd rather have the scoring and offense if I had to choose between these players. As a result I'm lower on Garnett and Wilt.

When I look at their careers in more detail, I don't really see anything that makes me change my mind. For instance, it's not like you can say that Wilt and Garnett had demonstrably greater impact in terms of leading their teams on deep playoff runs. Obviously, playoff success is heavily dependent on luck and overall team quality; I don't think Garnett is necessarily to blame for his lack of playoff success, for instance. But we can look and see that Steph, Shaq and Magic demonstrably had consistently led their teams on deep playoff runs over and over and over again during their primes when healthy. So at the very least, there's nothing that indicates that Garnett and Wilt were demonstrably *more* impactful for playoff winning than the other 3 players based on team outcome. When I look at their playoff offensive output, I don't see anything that makes me think that I'm underrating the value of their offense significantly. And it's definitely possible that there's an argument for thinking that their defensive contributions are so valuable as to outweigh the gap on offense between Steph / Shaq / Magic. But right now I just don't see it.

Yes, Hakeem and Duncan had potent offensive games, but they didn't really lead fantastic offenses. It was more about what they brought to the defensive end that gave them their biggest boost in impact. Obviously, Russell is a bit of a cherry picked example as he won 11 titles and that was through mostly defense, but there is a lot of context there.

Its arguable that Garnett had a better offensive game than Duncan, but was used in the wrong role. No he could not shoulder the same back to the basket game, but on similar volume and similar efficiency (about 1% on average lower TS%), Garnett played an elite face-up and spacing game. He would have made a perfect guy to play off of another play creator, but instead he was stuck generating a lot of offense for himself due to his roster.

Kevin Garnett's teams offenses also have similar offensive ratings to those of Duncan's with a worse supporting cast.

[OhayoKD]

10. Draymond, +6.6, lower bound +4.5, upper bound 8.6
11. Curry, +6.4, lower bound +4.7, upper bound 8.2
17. Shaq, +5.8, lower bound +4.3, upper bound 7.4

If someone wants to make the case that Draymond was as impactful as Curry and Shaq, I would be very interested in hearing that argument! And it would definitely make me a lot lower on their overall contributions if there's a solid case there. I'm a little skeptical because it definitely doesn't align with how I've felt watching the Warriors but it's an intriguing idea.

I'm not saying as impactful but

I do sort of make a case he was
-> very impactful
and
-> probably juices curry's impact
here:
https://forums.realgm.com/boards/viewtopic.php?p=107743950#p107743950
(bottom half of post)

[Owly]

I don’t think Kobe should be punished for those last few years. He basically gave the Lakers exactly what they needed at that time period: a marketable face to carry the brand while also being a tank commander. They actually drafted some incredible talent over that time period that could have been massive for them going forward if they didn’t throw it all into massively overpaying for Anthony Davis when no one else was going to make an offer anywhere close.

My nominees going to be David Robinson though. He honestly has a case right there with Hakeem as the second best defender of all-time and he was a more efficient score than Kobe on not that much less volume.

So as I say I don't think there's a simple easy answer ...

Some possible quibbles/counterpoints/ ...
a) He was still paid an awful lot.
b) The Lakers brand is/was itself the draw
c) Should a tank commander be rewarded or acknowledged in general
d) Should a tank commander not knowing that they're tank commanding be rewarded (and how does that speak to basketball iq more generally etc)
e) Did the Lakers ever "need" a tank commander. LA and the Laker brand have allowed them to get established stars through free agency or players "forcing" (or whatever, don't like the tone, but can't find something better) their way their via trade..
f) Should a player be rewarded for the strength of their marketability
g) Are Randle, Russell, Ingram (the Kobe tank era resultant picks) "incredible talent". Do any of them show a career trend of significant positive impact?
h) Is a superstar still gunning at high volume whilst ineffective sending positive developmental messages for potential young stars.
i) Assuming Bryant's brand was valuable to LA ... should that matter to a more holistic player evaluation where most players won't play in that particular situation.

I think, to me, the best case for a low weighting is most teams don't win anyway, thus the cost of (immediate-term) harm is low.

I'm not saying we should actually rank Kobe higher on the all-time list for his skills as a tank commander, but we certainly shouldn't punish him for those years. He helped them reach their organizational objectives. I think there would have been a lot of fan backlash to tanking 3 years in a row without Kobe front and center and I don't know if the front office would have had the stomach for it without Kobe scoring points and re-assuring everyone that the Lakers were still a show.

I disagree with certainly.
I don't know whether LA paid him that much to tank. I don't know that he thought he was tanking, rather suspect he didn't (cf above how that might actually negatively affect his entire career). I don't know when and to what degree LA purposefully tanked. If it did happen to help cf above re how much this should be "he 'helped' so it's ... fine?/good?/not harmful?/not bad basketball?".

And you're seriously asking if Randle and Ingram show any significant positive impact?

I am, per the quote, quibbling with " "incredible talent".
And inquiring "Do any of them show a career trend of significant positive impact?

And whilst it was a question rather than statement of the 3 only one has seen his teams trend better with him on the court for his career (Randle -2.9; Russell -1.6; Ingram +1).

Randle was deservedly all-NBA last year for the second time in 3 years.

Is an opinion.

He was 17th in the NBA last year by EPM expected wins.

At the surface level his team was worse with him on the floor last year. Cleaning the Glass raises that to +0.1.
I'm not clear on what the best state-of-the art public stuff is. My impression was DARKO DPM was liked slightly better. He ranks quite a bit lower than your EPM figure by that (cf: https://apanalytics.shinyapps.io/DARKO/). My guess might be more box-side influence measures might use his passing to guess he's an okay defender because on average that court awareness correlates somewhat with good D and they just miss here. I don't know. I also assume him playing 77 games, 2737 minutes allows the cumulative model to get him above players it thinks are better but less available (by your chosen metric, in your chosen year, he's ranked rate wise in the 30s amongst this pack
Fred VanVleet
Derrick White
Steven Adams
Trae Young
Julius Randle
O.G. Anunoby
Brook Lopez
Nic Claxton
Jarrett Allen
Franz Wagner
Ensemble pieces mostly, more than an "incredible talent" "deservedly all-NBA".
As I say I don't know the cutting edge stuff well but if he's actually an okay defender he's had awfully bad luck in his career with lineups.

Ingram had an up and down season last year dealing with injuries, but he was stil 51st in EPM despite not being quite as good as he was the year before when he was the clearcut best player on a Pels team that really pushed the Suns hard in the first round averaging 27/6/6 on .584 TS% in the postseason.

And given I did not say Ingram has had not one single good season nor playoff series this is fine but even assumed to fully accurate and the full story, doesn't counter my point.
Fwiw ...
I note, and I get it - its not wild or anything, the shift to the rate version for the guy where that is more favorable. If Randle 's minutes do really matter to you then one has to note BI's good year (and more generally) is less of a signal and less valuable that year and in general. Assuming clear cut best player to be true (and the aforementioned minutes issue would seem to come up as a negative if seeking internal consistency) this was a negative SRS team (9th record in conference).

As such I would stand by ...
I am, per the quote, quibbling with "incredible talent".
And inquiring "Do any of them show a career trend of significant positive impact?

[DraymondGold]

Looks like Wilt’s out to a pretty big early lead so far. My count right now has:

Wilt 5
Shaq 2
Steph 1

I get why he’s up there but I think if you really try to tease out his impact, it’s not that strong. He was traded twice for junk and both times it’s not clear if the team who he joined got better or worse when he arrived. He did have one glorious season in 1967, but aside from that one year, I don’t see him as on the same level as any of the other currently nominated candidates.

Just following up on this, because I used to be pretty concerned about Wilt's poorer WOWY. I actually think a closer look is a bit more favorable to Wilt.

The biggest multi-year lineup-change samples you're talking about are...
-Wilt's Rookie Year
-1965 Trade, leaving Warriors
-1965 Trade, joining 76ers
-1968–69 Trade, leaving 76ers
-1968–69 Trade, joining Lakers
-1970 Injury
-Wilt's Retirement

The changes in team results are obviously highly dependent on changes in the other players too (e.g. other teammates getting better, getting worse, other roster changes), changes to coaching, the context, etc.. But if we just look at changes to Wilt's Team's performance (SRS for full-year data, MoV for mid-year injuries / trades), we get:

Wilt's (un-health corrected) Multi-Year large WOWY samples:
1959–60 Warriors: 2.27 SRS with, -2.29 SRS without. Total change: +5.06 [Rookie year]
1965 Warriors: -4.97 MoV with, -7.26 MoV without. Total change: +2.29. [trade, leaving Warriors]
1965 76ers: 0.29 MoV with, -0.49 MoV without. Total change: +0.78. [trade, joining 76ers]
1968–69 76ers: 7.96 SRS with, 4.79 SRS without. Total change: +3.17 [trade, leaving 76ers]
1968–69 Lakers: 3.84 SRS with, 4.99 SRS without. Total change: -1.15. [trade, joining Lakers]
1969-70 Lakers: 3.64 MoV with, 1.94 MoV without. Total change: +1.7 [Injury year. Need to combine 1969–70 to have a usable 'With' sample. You get 1.2 if you use 1970–71 instead]
1973–74 Lakers: 8.16 SRS with, 0.85 SRS without. Total change: +7.31. [Retirement]

But Zeppelin noted that Wilt was injured in 1965! (see first links below). And these are two of Wilt's worse samples! We can help correct for this by wrapping in the neighboring healthy years. For example, rather than just looking at 1964 Warriors with and without (injured) Wilt, we might look at how the 1964–65 Warriors look with and without Wilt, and how 1965–66 76ers look with and without Wilt. The results....

Wilt's (healthy) Multi-Year large WOWY samples:
1959–60 Warriors: 2.27 SRS with, -2.29 SRS without. Total change: +5.06 [Rookie year]
1964–65 Warriors: +1.44 MoV with, -7.26 MoV without. Total change: +8.7 [trade, leaving Warriors]
1965–66 76ers: +3.0 MoV with, -0.49 MoV without. Total change: +3.49 [trade, joining 76ers]
1968–69 76ers: 7.96 SRS with, 4.79 SRS without. Total change: +3.17 [trade, leaving 76ers]
1968–69 Lakers: 3.84 SRS with, 4.99 SRS without. Total change: -1.15. [trade, joining Lakers.]
Note: +2.0 comparing 1968-1969 playoffs
1969-70 Lakers: 3.64 MoV with, 1.94 MoV without. Total change: +1.7 [Injury year. Need to combine 1969–70 to have a usable 'With' sample. You get 1.2 if you use 1970–71 instead]
1973–74 Lakers: 8.16 SRS with, 0.85 SRS without. Total change: +7.31. [Retirement]

For context, here's Hakeem's:
1984–85 Rockets: 1.38 with, -3.12 without. Total change: 4.5 [Rookie year]
1991–92 Rockets: 1.7 with, -1.79 without. Total change: 3.49 [injury years]
1998 Rockets: 0 with, -1.77 without. Total change: 1.77 [injury year]
2000 Rockets: -3.7 with, 2.42 without. Total change: -6.12 [injury year]
2001–02 Rockets: 2.71 with, -4.31 without. Total change: 7.02 [trade, leaving Rockets]
2001–02 Raptors: -0.7 with, 1.69 without. Total change: -2.39 [trade, joining Raptors]
2002–03 Raptors: -0.7 with, -6.1 without. Total change: 5.4 [Retirement]

Wilt's Career Average: +2.74
Wilt's healthy Career Average: +4.04
Hakeem's Career Average: +1.95 (but most samples are from post-prime)
Wilt's 10-year prime: +2.03 (10-year 60–69 prime). +2.35 (best 10-year stretch 1965-1974)
Wilt's 10-year healthy prime: +3.83 (10-year 60-69 prime). +3.87 (best 10-year stretch 1965-1974)
Hakeem's 10-year prime: +3.49 (10-year 86-95 prime, but this only includes one sample. 14-year average: +3.25 from 1985–1998)

One starts to wonder why people have been saying raw WOWY is a point for Hakeem over Wilt... But I digress
To tie this back to the players that are up for voting, by their 10-year time, the players rank like this:
10-year Prime in Multi-year WOWY samples:
Curry: +8.39 in 2014–2023, with 3 large-sample WOWY in 2018 injury, 2020–21 injury, 2022–23 injury
Magic: +7.68 in 1982–1991, with 1 sample in 1991. +6.06 in 2 samples in 11-years from 1981-1991. +4.85 in 3 samples in 12 years from 1980–91.
Garnett: +6.88 in 2000–2009, with 3 samples in 2007-08 trade from Timberwolves, 2007-08 trade to Celtics, 2008-2009 injury. Obviously the Celtics' roster was entirely revamped, so it's +3.83 if we discount that trade to the Celtics sample, just below Wilt.
Shaq: +3.86 in 1995–2004 in 3 samples, 1996-97 trade from Magic, 1996-97 trade to Lakers, 2004-05 trade from Lakers.
Wilt: +3.83 in 1960-1969 in 5 samples, see above. It would be + 4.48 if we switched the 1968–69 regular season 76ers number for the playoffs number.
Hakeem: +3.49 (see above, just 1 sample in 91-92).

So if we correct for Wilt's health, Wilt certainly doesn't seem clearly below the others in this tier. He's tied with Shaq for his 10-year prime. You can argue Wilt's 10-year prime over Shaq if you trust Wilt's larger number of samples here, or if you use Wilt's playoff change for the 1969 Lakers. You could argue Wilt's 10-year prime over Garnett's if you disregard the 2008 Celtics sample entirely since some of that change was from the rest of the roster revamping (I wouldn't throw it out entirely, but I also wouldn't take it at face-value).

Curry and Magic seem the best, but they also have the worst longevity whether you're counting total number of games (for Garnett / Shaq) or longevity relative to era (for Wilt, since players in the 2000s had on average about 70% more career games than players in the 60s)

Note that this poor health can also be used to explain Wilt's low raw WOWY numbers in Thinking Basketball's "10-year Prime WOWY" score that you see on the old Top 40 pages. It also biases the WOWYR somewhat, although WOWYR is more positive on Wilt in general because it sees Wilt correlates much better with his team's success than his teammates, even if he doesn't correlate as much in 1965 as we think he might when healthy.

...

More details on Wilt's down years here: https://forums.realgm.com/boards/viewtopic.php?p=107709117#p107709117
More info on Wilt's WOWY / team changes without him here: https://forums.realgm.com/boards/viewtopic.php?p=107709162#p107709162

[Owly]

As a general note, I find JE’s RAPM to be pretty confusing, to be honest. As far as I’m aware, it’s not supposed to have a box component. But there’s outcomes that don’t really make sense as a pure RAPM outcome. Like, for instance, Kevin Garnett ranked #2 in the NBA (and #14 across the entire set) in JE’s RAPM in 2004-2005, in a year where his raw on-off was only +0.7 (and he had no playoff numbers so that didn’t change it). Now, I get that raw on-off and RAPM aren’t the same, but they’re very closely related (RAPM basically just takes raw on-off and makes some adjustments based on who’s on the floor), and it’s quite difficult for me to understand how a player could come out as #2 in the NBA in RAPM in a season where their raw on-off is +0.7. I think the “prior-informed” part of it must be doing quite a lot of work in that model. But I don’t really know what his “priors” are, and whether they’re box-score-based priors, or past years’ on-off, or something else (JE has apparently even used height as a prior in RAPM before). It seems obvious to me that it is very far from just raw RAPM. Does anyone know what priors he’s using? My guess is that it’s at least in part past years’ on-off—which would explain why 2004-2005 Garnett is ranked really highly, since Garnett had really high on-off the previous couple years—but then if there’s a really strong prior for past years’ on-off (which there’d have to be to get a +0.7 raw on-off to #2 in the NBA and #14 in the last 25 years in RAPM), then the model is going to fail to account for quick changes in a player’s ability/impact (i.e. for instance, Steph Curry getting incredibly good really fast, after having had merely pretty good on-off his first few seasons while not really being used to his full potential).

Very much not an expert but I would guess/venture...

Having the two previous years at circa +20 on-off might give great confidence in pushing priors pretty high.
I think at first glance Minny's 6-8 in minutes look relatively (and absolutely) better than 2-5. If the priors reflect this and if the rest of the top 5 are the guys playing the highest proportion of minutes with Garnett then that probably is a factor.

You do make a case for future as well as past priors though that would never be able to be available immediately.

I don't know what you mean by "raw RAPM". I guess NPI. Googlesites (I think that was JE) had that much lower +0.7 (101st). It is a purer measure, but might in some cases think things like Garnett is benefiting more from playing with Hoiberg than vice-versa. It's a trade-off. Whilst I really like1 number metrics I guess it's just a reminder than there is no one holy grail and to look at these things in context. Fwiw that site's PI version had Garnett 8th that year.

[eminence]

I find it more likely Dray is draining Steph than the reverse (or that someone else on the squad is missing out, namely Klay as the other constant).

'14-'15 to '22-'23 (minus '20) RS/PO combined, P/M and possessions from pbpstats.com, Net calculated and rounded to the nearest tenth
Both: 17278 minutes, +13.9 Net
Steph: 5728 minutes, +8.1 Net
Dray: 5532 minutes, +1.6 Net
Neither: 9479 minutes, -5.2 Net

[lessthanjake]

10. Draymond, +6.6, lower bound +4.5, upper bound 8.6
11. Curry, +6.4, lower bound +4.7, upper bound 8.2
17. Shaq, +5.8, lower bound +4.3, upper bound 7.4

If someone wants to make the case that Draymond was as impactful as Curry and Shaq, I would be very interested in hearing that argument! And it would definitely make me a lot lower on their overall contributions if there's a solid case there. I'm a little skeptical because it definitely doesn't align with how I've felt watching the Warriors but it's an intriguing idea.

I'm not saying as impactful but

I do sort of make a case he was
-> very impactful
and
-> probably juices curry's impact
here:
https://forums.realgm.com/boards/viewtopic.php?p=107743950#p107743950
(bottom half of post)

That’s what happens when a superstar fits quite well with another star. The other star has room to be very impactful. And Steph Curry is very easy to fit well with—which is a good thing!

The “juices Curry’s impact” thing simply isn’t actually true though. If we look at pbpstats data for the last decade in regular season + playoffs (excluding 2019-2020), we find that when Curry + Draymond are both on the floor, the Warriors have been +13.31 in net rating and when Curry is off but Draymond is on, they’ve gone +2.10. Therefore, when filtered to just minutes that Draymond is on, Steph being on has had a +11.21 effect on net rating. Meanwhile, when Curry is on the floor without Draymond, the Warriors have had a +6.92 net rating, and then when both are off the floor, the Warriors’ net rating has been -4.58. So, when filtered to just minutes that Draymond is off, Steph being on has had a +11.50 effect on net rating. In other words, Curry’s impact is basically exactly the same regardless of whether Draymond is on or off the court, and indeed is actually very slightly higher with Draymond off.

The only sense in which Draymond “juices Curry’s impact” is that Curry’s effect on the raw on-off is able to stay the same with Draymond on despite the net ratings being much better (improving net rating is harder to do the higher it is)—which in part reflects Curry being a ceiling raiser who’s able to extract his full value without stepping on other great teammates’ toes much. Which is of course a good thing! I’d also note the obvious that, like the effect of mostly playing with your teams’ better players in general, Draymond helps Curry’s raw on-off because their minutes overlap a lot. But that’s just a basic fact of how raw on-off works for everyone, and isn’t the point you were making. The fact is that when you control for whether Draymond is on or off, Curry’s impact over the last decade has been the same (and indeed very very slightly better when Draymond is off as opposed to on).

[iggymcfrack]

Looks like Wilt’s out to a pretty big early lead so far. My count right now has:

Wilt 5
Shaq 2
Steph 1

I get why he’s up there but I think if you really try to tease out his impact, it’s not that strong. He was traded twice for junk and both times it’s not clear if the team who he joined got better or worse when he arrived. He did have one glorious season in 1967, but aside from that one year, I don’t see him as on the same level as any of the other currently nominated candidates.

Just following up on this, because I used to be pretty concerned about Wilt's poorer WOWY. I actually think a closer look is a bit more favorable to Wilt.

The biggest multi-year lineup-change samples you're talking about are...
-Wilt's Rookie Year
-1965 Trade, leaving Warriors
-1965 Trade, joining 76ers
-1968–69 Trade, leaving 76ers
-1968–69 Trade, joining Lakers
-1970 Injury
-Wilt's Retirement

The changes in team results are obviously highly dependent on changes in the other players too (e.g. other teammates getting better, getting worse, other roster changes), changes to coaching, the context, etc.. But if we just look at changes to Wilt's Team's performance (SRS for full-year data, MoV for mid-year injuries / trades), we get:

Wilt's (un-health corrected) Multi-Year large WOWY samples:
1959–60 Warriors: 2.27 SRS with, -2.29 SRS without. Total change: +5.06 [Rookie year]
1965 Warriors: -4.97 MoV with, -7.26 MoV without. Total change: +2.29. [trade, leaving Warriors]
1965 76ers: 0.29 MoV with, -0.49 MoV without. Total change: +0.78. [trade, joining 76ers]
1968–69 76ers: 7.96 SRS with, 4.79 SRS without. Total change: +3.17 [trade, leaving 76ers]
1968–69 Lakers: 3.84 SRS with, 4.99 SRS without. Total change: -1.15. [trade, joining Lakers]
1969-70 Lakers: 3.64 MoV with, 1.94 MoV without. Total change: +1.7 [Injury year. Need to combine 1969–70 to have a usable 'With' sample. You get 1.2 if you use 1970–71 instead]
1973–74 Lakers: 8.16 SRS with, 0.85 SRS without. Total change: +7.31. [Retirement]

But Zeppelin noted that Wilt was injured in 1965! (see first links below). And these are two of Wilt's worse samples! We can help correct for this by wrapping in the neighboring healthy years. For example, rather than just looking at 1964 Warriors with and without (injured) Wilt, we might look at how the 1964–65 Warriors look with and without Wilt, and how 1965–66 76ers look with and without Wilt. The results....

Wilt's (healthy) Multi-Year large WOWY samples:
1959–60 Warriors: 2.27 SRS with, -2.29 SRS without. Total change: +5.06 [Rookie year]
1964–65 Warriors: +1.44 MoV with, -7.26 MoV without. Total change: +8.7 [trade, leaving Warriors]
1965–66 76ers: +3.0 MoV with, -0.49 MoV without. Total change: +3.49 [trade, joining 76ers]
1968–69 76ers: 7.96 SRS with, 4.79 SRS without. Total change: +3.17 [trade, leaving 76ers]
1968–69 Lakers: 3.84 SRS with, 4.99 SRS without. Total change: -1.15. [trade, joining Lakers.]
Note: +2.0 comparing 1968-1969 playoffs
1969-70 Lakers: 3.64 MoV with, 1.94 MoV without. Total change: +1.7 [Injury year. Need to combine 1969–70 to have a usable 'With' sample. You get 1.2 if you use 1970–71 instead]
1973–74 Lakers: 8.16 SRS with, 0.85 SRS without. Total change: +7.31. [Retirement]

For context, here's Hakeem's:
1984–85 Rockets: 1.38 with, -3.12 without. Total change: 4.5 [Rookie year]
1991–92 Rockets: 1.7 with, -1.79 without. Total change: 3.49 [injury years]
1998 Rockets: 0 with, -1.77 without. Total change: 1.77 [injury year]
2000 Rockets: -3.7 with, 2.42 without. Total change: -6.12 [injury year]
2001–02 Rockets: 2.71 with, -4.31 without. Total change: 7.02 [trade, leaving Rockets]
2001–02 Raptors: -0.7 with, 1.69 without. Total change: -2.39 [trade, joining Raptors]
2002–03 Raptors: -0.7 with, -6.1 without. Total change: 5.4 [Retirement]

Wilt's Career Average: +2.74
Wilt's healthy Career Average: +4.04
Hakeem's Career Average: +1.95 (but most samples are from post-prime)
Wilt's 10-year prime: +2.03 (10-year 60–69 prime). +2.35 (best 10-year stretch 1965-1974)
Wilt's 10-year healthy prime: +3.83 (10-year 60-69 prime). +3.87 (best 10-year stretch 1965-1974)
Hakeem's 10-year prime: +3.49 (10-year 86-95 prime, but this only includes one sample. 14-year average: +3.25 from 1985–1998)

One starts to wonder why people have been saying raw WOWY is a point for Hakeem over Wilt... But I digress
To tie this back to the players that are up for voting, by their 10-year time, the players rank like this:
10-year Prime in Multi-year WOWY samples:
Curry: +8.39 in 2014–2023, with 3 large-sample WOWY in 2018 injury, 2020–21 injury, 2022–23 injury
Magic: +7.68 in 1982–1991, with 1 sample in 1991. +6.06 in 2 samples in 11-years from 1981-1991. +4.85 in 3 samples in 12 years from 1980–91.
Garnett: +6.88 in 2000–2009, with 3 samples in 2007-08 trade from Timberwolves, 2007-08 trade to Celtics, 2008-2009 injury. Obviously the Celtics' roster was entirely revamped, so it's +3.83 if we discount that trade to the Celtics sample, just below Wilt.
Shaq: +3.86 in 1995–2004 in 3 samples, 1996-97 trade from Magic, 1996-97 trade to Lakers, 2004-05 trade from Lakers.
Wilt: +3.83 in 1960-1969 in 5 samples, see above. It would be + 4.48 if we switched the 1968–69 regular season 76ers number for the playoffs number.
Hakeem: +3.49 (see above, just 1 sample in 91-92).

So if we correct for Wilt's health, Wilt certainly doesn't seem clearly below the others in this tier. He's tied with Shaq for his 10-year prime. You can argue Wilt's 10-year prime over Shaq if you trust Wilt's larger number of samples here, or if you use Wilt's playoff change for the 1969 Lakers. You could argue Wilt's 10-year prime over Garnett's if you disregard the 2008 Celtics sample entirely since some of that change was from the rest of the roster revamping (I wouldn't throw it out entirely, but I also wouldn't take it at face-value).

Curry and Magic seem the best, but they also have the worst longevity whether you're counting total number of games (for Garnett / Shaq) or longevity relative to era (for Wilt, since players in the 2000s had on average about 70% more career games than players in the 60s)

Note that this poor health can also be used to explain Wilt's low raw WOWY numbers in Thinking Basketball's "10-year Prime WOWY" score that you see on the old Top 40 pages. It also biases the WOWYR somewhat, although WOWYR is more positive on Wilt in general because it sees Wilt correlates much better with his team's success than his teammates, even if he doesn't correlate as much in 1965 as we think he might when healthy.

...

More details on Wilt's down years here: https://forums.realgm.com/boards/viewtopic.php?p=107709117#p107709117
More info on Wilt's WOWY / team changes without him here: https://forums.realgm.com/boards/viewtopic.php?p=107709162#p107709162

OK, but Wilt got traded twice in his prime. With Hakeem most of those changes you're looking at are his impact from age 35-39 when he was well past his prime. It makes sense that Hakeem from age 35-39 shouldn't show the same level of change as peak Wilt. If you look at the overall WOWY number for Wilt and Hakeem you get:

Wilt: +6.0 WOWYR (38th), 1.1 prime WOWY (114th)
Hakeem: +5.7 WOWYR (49th), 5.2 prime WOWY (10th)

Furthermore, if you compare Wilt to Shaq who also moved teams twice in his prime, here's what you get:

1992-93 Magic: 1.35 SRS with, -6.52 SRS without. Total change: +7.87 [Rookie year]
1995-95 Magic: 5.40 SRS with, -0.07 SRS without. Total change: +5.47 [trade, leaving Magic]
1996-97 Lakers: 4.21 SRS with, +.66 without. Total change: -0.55 [trade, joining Lakers]
2003-04 Lakers : 4.35 SRS with, -2.32 SRS without. Total change: +6.67 [trade, leaving Lakers]
2004-05 Heat: 5.77 SRS with, -0.13 SRS without. Total change: +5.90 [trade, joining Heat]

So, looking at the rookie year and prime moves, you get a net average change of +5.07 for Shaq whereas for Wilt's moves during his prime you get an average of +2.03.

[One_and_Done]

Vote: Shaquille O'Neal

It's time for Shaq. In comparing the candidates, let's start by looking at the WOWY numbers:

Prime WOWY
KG 5.7
Shaq 5.5
Magic 4.7
Wilt 1.2

WOWYR
Magic 10.1
Shaq 6.7
KG 6.2
Wilt 6.0

OK, so all we can really determine from this is that Wilt's at the back of the pack. The other three are close enough that how you order them depends on your methodology. These numbers kinda fit what we've seen looking at team level results year-by-year in that Wilt just wasn't much of a winner and he didn't contribute to team results anywhere near as much as his reputation would have you believe. The fact that Magic scores higher in WOWYR, but lower in prime WOWY makes me think that he might be getting an artificual boost in the second category from skipping the decline phase to his career.

Next, let's look at longevity.
Shaq played 19 seasons, 1207 games, 41,918 minutes, and made all-NBA 14 times.
KG played 21 seasons, 1462 games, 50,418 minutes, and made all-NBA 9 times.
Magic played 13 seasons, 906 games, 33,245 minutes, and made all-NBA 10 times.

So KG has the edge in overall longevity, but Shaq likely has the edge in longevity at a high-level with so many more NBA teams and a 12 year stretch where his teams won 50 games (or equivalent pace in a lockout shortened season) and made it into at least the second round every single year. With KG, Shaq, and Magic all having similar all-time peaks, I think we can eliminate Magic next on the longevity question.

So now it's down to Shaq vs. KG. Shaq won more, while KG had better impact stats. In that sort of situation, I'd normally tend to go with the numbers except Shaq was a playoff riser. Especially after he won his first ring in 2000, he tended to coast a lot more through the regular season and then kick it into full gear in the postseason. I don't know if calling KG a postseason "faller" would be accurate since his defense always held up extremely well, but his offense did have some question marks against locked in defenses. Let's look at some simple numbers:

Regular season Shaq: 26.4 PER, 5.1 BPM, +8.5 on/off
Regular season KG: 22.7 PER, 5.6 BPM, +11.3 on/off

Postseason Shaq: 26.1 PER, 5.6 BPM, +11.7 on/off
Postseason KG: 21.1 PER, 5.1 BPM, +14.5 on/off

Notice that the leader by BPM switches when you go between regular season and postseason stats. KG does still have the edge in on/off even if he's starting from a lower base. IDK, you really can't go wrong with either guy here. I think they're incredibly close and there might not be a meaningful differentiator. Ultimately, I'm going to trust what Shaq ultimately accomplished over what KG could have done in theory since he had shown himself to rise to a new level in the playoffs above and beyond what KG did. It's a hair's breadth difference though and I certainly wouldn't argue with someone who prefers Garnett.

Alternate: Kevin Garnett

Nominate David Robinson

One of the very best defensive players of all-time. He has a very compelling case against Hakeem as the 2nd best defensive player in league history. He joined a league average defense as a rookie and immediately made them the 3rd best defense in the league in his very first season. They would go on to rank 1st, 1st, 10th, 9th, 5th, and 3rd the following seasons before he got injured and missed a season where they promptly fell to last in the league. Upon his return (and the addition of Tim Duncan), they followed that up with 4 top 2 defenses in a row. His offense was very good too as in his prime years before the initial injury, he averaged 26 PPG on .592 TS%, far above the league average which hovered between .528 and .543 over those seasons.

Iggy as someone who is going to rate D.Rob higher than most I am sympthetic to your nomination. At this stage though you're throwing your vote away. To get traction on D.Rob there is going to need to be an absolute mountain of good analysis done, not a small para about it. It's also still a little soon. I say that as someone who will be turning my eye to D.Rob and Giannis quite soon. For now though you're better advocatng for a more plausible candidate like Dirk or K.Malone.

[iggymcfrack]

Vote: Shaquille O'Neal

It's time for Shaq. In comparing the candidates, let's start by looking at the WOWY numbers:

Prime WOWY
KG 5.7
Shaq 5.5
Magic 4.7
Wilt 1.2

WOWYR
Magic 10.1
Shaq 6.7
KG 6.2
Wilt 6.0

OK, so all we can really determine from this is that Wilt's at the back of the pack. The other three are close enough that how you order them depends on your methodology. These numbers kinda fit what we've seen looking at team level results year-by-year in that Wilt just wasn't much of a winner and he didn't contribute to team results anywhere near as much as his reputation would have you believe. The fact that Magic scores higher in WOWYR, but lower in prime WOWY makes me think that he might be getting an artificual boost in the second category from skipping the decline phase to his career.

Next, let's look at longevity.
Shaq played 19 seasons, 1207 games, 41,918 minutes, and made all-NBA 14 times.
KG played 21 seasons, 1462 games, 50,418 minutes, and made all-NBA 9 times.
Magic played 13 seasons, 906 games, 33,245 minutes, and made all-NBA 10 times.

So KG has the edge in overall longevity, but Shaq likely has the edge in longevity at a high-level with so many more NBA teams and a 12 year stretch where his teams won 50 games (or equivalent pace in a lockout shortened season) and made it into at least the second round every single year. With KG, Shaq, and Magic all having similar all-time peaks, I think we can eliminate Magic next on the longevity question.

So now it's down to Shaq vs. KG. Shaq won more, while KG had better impact stats. In that sort of situation, I'd normally tend to go with the numbers except Shaq was a playoff riser. Especially after he won his first ring in 2000, he tended to coast a lot more through the regular season and then kick it into full gear in the postseason. I don't know if calling KG a postseason "faller" would be accurate since his defense always held up extremely well, but his offense did have some question marks against locked in defenses. Let's look at some simple numbers:

Regular season Shaq: 26.4 PER, 5.1 BPM, +8.5 on/off
Regular season KG: 22.7 PER, 5.6 BPM, +11.3 on/off

Postseason Shaq: 26.1 PER, 5.6 BPM, +11.7 on/off
Postseason KG: 21.1 PER, 5.1 BPM, +14.5 on/off

Notice that the leader by BPM switches when you go between regular season and postseason stats. KG does still have the edge in on/off even if he's starting from a lower base. IDK, you really can't go wrong with either guy here. I think they're incredibly close and there might not be a meaningful differentiator. Ultimately, I'm going to trust what Shaq ultimately accomplished over what KG could have done in theory since he had shown himself to rise to a new level in the playoffs above and beyond what KG did. It's a hair's breadth difference though and I certainly wouldn't argue with someone who prefers Garnett.

Alternate: Kevin Garnett

Nominate David Robinson

One of the very best defensive players of all-time. He has a very compelling case against Hakeem as the 2nd best defensive player in league history. He joined a league average defense as a rookie and immediately made them the 3rd best defense in the league in his very first season. They would go on to rank 1st, 1st, 10th, 9th, 5th, and 3rd the following seasons before he got injured and missed a season where they promptly fell to last in the league. Upon his return (and the addition of Tim Duncan), they followed that up with 4 top 2 defenses in a row. His offense was very good too as in his prime years before the initial injury, he averaged 26 PPG on .592 TS%, far above the league average which hovered between .528 and .543 over those seasons.

Iggy as someone who is going to rate D.Rob higher than most I am sympthetic to your nomination. At this stage though you're throwing your vote away. To get traction on D.Rob there is going to need to be an absolute mountain of good analysis done, not a small para about it. It's also still a little soon. I say that as someone who will be turning my eye to D.Rob and Giannis quite soon. For now though you're better advocatng for a more plausible candidate like Dirk or K.Malone.

My top 4 candidates to nominate would be D-Rob, Giannis, CP3, and Jokic. I don’t see any point in trying to nominate someone who would be 15th or 16th on my overall list just because they have a better chance of getting nominated when they likely aren’t going ahead of Bird or Kobe anyway. I’ll vote strategically with my main alternate vote if the race seems to be between two people who aren’t my first choice, but I don’t see any point in doing it on my nomination vote.

[Dooley]

Yes, Hakeem and Duncan had potent offensive games, but they didn't really lead fantastic offenses. It was more about what they brought to the defensive end that gave them their biggest boost in impact.

OK this is definitely a fair point - it would definitely be more accurate to say that Hakeem and Duncan lead playoff offenses that were good *enough* to win titles, not that they lead fantastic playoff offenses. And not necessarily better than Garnett's best team playoff offenses in terms of rORTG.

Its arguable that Garnett had a better offensive game than Duncan, but was used in the wrong role. No he could not shoulder the same back to the basket game, but on similar volume and similar efficiency (about 1% on average lower TS%), Garnett played an elite face-up and spacing game. He would have made a perfect guy to play off of another play creator, but instead he was stuck generating a lot of offense for himself due to his roster.

This, I definitely don't agree with. Not just bc Duncan's box stats were a little better - but also bc IMO Duncan's resilient scoring stuff does actually provide a lot more value in the playoffs. Having a player who is able to get points consistently in the playoffs is so valuable to winning playoff games. And that's really what I value about Hakeem's and Duncan's offensive games, and where Garnett is weaker. I do think that the fact that Garnett was less good generating offense himself is a big deal in the playoffs (maybe even in a way that isn't born out by oRTG?). And, IDK, whenever I go back and watch Garnett it always feels like he's settling for long-range jumpers on offense, and I hate that, and I think it genuinely provides a lot less value.

So, as I look at it, Duncan and Hakeem were providing a lot more individual offensive value in the playoffs with consistent bucket-getting *in addition* to elite defense, and that's why I see them in a different category than Wilt or Garnett. And, again, I don't see what the dimension is for Wilt and Garnett that makes up for that.

[DraymondGold]

I find it more likely Dray is draining Steph than the reverse (or that someone else on the squad is missing out, namely Klay as the other constant).

'14-'15 to '22-'23 (minus '20) RS/PO combined, P/M and possessions from pbpstats.com, Net calculated and rounded to the nearest tenth
Both: 17278 minutes, +13.9 Net
Steph: 5728 minutes, +8.1 Net
Dray: 5532 minutes, +1.6 Net
Neither: 9479 minutes, -5.2 Net

For those who prefer raw WOWY (@OhayoKD, since I think the debate between you and jake were what started this discussion?), it's worth noting that WOWY also clearly favors Steph over Dray:

'14-'15 to '22-'23 (minus '20) RS/PO combined, raw WOWY (changes in margin of victory per game), data from statmuse.com and rounded to the nearest hundredths:

with Steph: 9.86 (451 games 2014-2019), 4.35 (218 games 2021-2023). Average: +8.06
without Steph: 0.3 (64 games 2014-2019), -3.04 (53 games 2020-2023). Average: -1.21
with Draymond: 8.88 (476 games 2014-2019), 4.23 (216 games 2020-2023). Average: +7.43.
without Draymond: 6.13 (39 games 2014-2019), -2.31 (55 games 2020-2023). Average: +1.19

Steph's raw WOWY in this stretch: +9.27 (49% better than Draymond)
Draymond's raw WOWY in this stretch: +6.24

[OhayoKD]

I find it more likely Dray is draining Steph than the reverse (or that someone else on the squad is missing out, namely Klay as the other constant).

'14-'15 to '22-'23 (minus '20) RS/PO combined, P/M and possessions from pbpstats.com, Net calculated and rounded to the nearest tenth
Both: 17278 minutes, +13.9 Net
Steph: 5728 minutes, +8.1 Net
Dray: 5532 minutes, +1.6 Net
Neither: 9479 minutes, -5.2 Net

Well I think if you're trying to iisolate for their influence on each other, WOWY is probably preferable(though I do not know how one would filter out 2020 net nor do i ran out of statmuse searches before i could check net). And by record Draymond does not look like a drain at all.

Nonetheless going by lineup-stuff

For that frame, in the regular-season, PBP has steph no dray at +7.97 and Dray no Steph at +2.22. Both +14, neither at -4.14

For full-season dray no steph is +7.3 and steph no dray is +11.6. Both is +18.30, neither is -7.12

For the 15-17 regular seasons(the years they really good) steph no dray is +11.6, Dray no steph is +7.34, both is +18.3, neither is -7.12
For full season it's +8.6 for dray no steph and +9.6 for steph no dray, without -7.29, both +17.09

If you go just 15-16 things actually swing to Draymond both rs and playoffs.

Doesn't look like a drain to me

[OhayoKD]

I find it more likely Dray is draining Steph than the reverse (or that someone else on the squad is missing out, namely Klay as the other constant).

'14-'15 to '22-'23 (minus '20) RS/PO combined, P/M and possessions from pbpstats.com, Net calculated and rounded to the nearest tenth
Both: 17278 minutes, +13.9 Net
Steph: 5728 minutes, +8.1 Net
Dray: 5532 minutes, +1.6 Net
Neither: 9479 minutes, -5.2 Net

For those who prefer raw WOWY (@OhayoKD, since I think the debate between you and jake were what started this discussion?), it's worth noting that WOWY also clearly favors Steph over Dray:

'14-'15 to '22-'23 (minus '20) RS/PO combined, raw WOWY (changes in margin of victory per game), data from statmuse.com and rounded to the nearest hundredths:

with Steph: 9.86 (451 games 2014-2019), 4.35 (218 games 2021-2023). Average: +8.06
without Steph: 0.3 (64 games 2014-2019), -3.04 (53 games 2020-2023). Average: -1.21
with Draymond: 8.88 (476 games 2014-2019), 4.23 (216 games 2020-2023). Average: +7.43.
without Draymond: 6.13 (39 games 2014-2019), -2.31 (55 games 2020-2023). Average: +1.19

Steph's raw WOWY in this stretch: +9.27 (49% better than Draymond)
Draymond's raw WOWY in this stretch: +6.24

Thanks!

Steph looks better as I'd expect, But Draymond looks like a superstar in his own right(was expecting it to be substantially less impressive past 17 actually).

[DraymondGold]

Looks like Wilt’s out to a pretty big early lead so far. My count right now has:

Wilt 5
Shaq 2
Steph 1

I get why he’s up there but I think if you really try to tease out his impact, it’s not that strong. He was traded twice for junk and both times it’s not clear if the team who he joined got better or worse when he arrived. He did have one glorious season in 1967, but aside from that one year, I don’t see him as on the same level as any of the other currently nominated candidates.

Just following up on this, because I used to be pretty concerned about Wilt's poorer WOWY. I actually think a closer look is a bit more favorable to Wilt.

The biggest multi-year lineup-change samples you're talking about are...
-Wilt's Rookie Year
-1965 Trade, leaving Warriors
-1965 Trade, joining 76ers
-1968–69 Trade, leaving 76ers
-1968–69 Trade, joining Lakers
-1970 Injury
-Wilt's Retirement

The changes in team results are obviously highly dependent on changes in the other players too (e.g. other teammates getting better, getting worse, other roster changes), changes to coaching, the context, etc.. But if we just look at changes to Wilt's Team's performance (SRS for full-year data, MoV for mid-year injuries / trades), we get:

Wilt's (un-health corrected) Multi-Year large WOWY samples:
1959–60 Warriors: 2.27 SRS with, -2.29 SRS without. Total change: +5.06 [Rookie year]
1965 Warriors: -4.97 MoV with, -7.26 MoV without. Total change: +2.29. [trade, leaving Warriors]
1965 76ers: 0.29 MoV with, -0.49 MoV without. Total change: +0.78. [trade, joining 76ers]
1968–69 76ers: 7.96 SRS with, 4.79 SRS without. Total change: +3.17 [trade, leaving 76ers]
1968–69 Lakers: 3.84 SRS with, 4.99 SRS without. Total change: -1.15. [trade, joining Lakers]
1969-70 Lakers: 3.64 MoV with, 1.94 MoV without. Total change: +1.7 [Injury year. Need to combine 1969–70 to have a usable 'With' sample. You get 1.2 if you use 1970–71 instead]
1973–74 Lakers: 8.16 SRS with, 0.85 SRS without. Total change: +7.31. [Retirement]

But Zeppelin noted that Wilt was injured in 1965! (see first links below). And these are two of Wilt's worse samples! We can help correct for this by wrapping in the neighboring healthy years. For example, rather than just looking at 1964 Warriors with and without (injured) Wilt, we might look at how the 1964–65 Warriors look with and without Wilt, and how 1965–66 76ers look with and without Wilt. The results....

Wilt's (healthy) Multi-Year large WOWY samples:
1959–60 Warriors: 2.27 SRS with, -2.29 SRS without. Total change: +5.06 [Rookie year]
1964–65 Warriors: +1.44 MoV with, -7.26 MoV without. Total change: +8.7 [trade, leaving Warriors]
1965–66 76ers: +3.0 MoV with, -0.49 MoV without. Total change: +3.49 [trade, joining 76ers]
1968–69 76ers: 7.96 SRS with, 4.79 SRS without. Total change: +3.17 [trade, leaving 76ers]
1968–69 Lakers: 3.84 SRS with, 4.99 SRS without. Total change: -1.15. [trade, joining Lakers.]
Note: +2.0 comparing 1968-1969 playoffs
1969-70 Lakers: 3.64 MoV with, 1.94 MoV without. Total change: +1.7 [Injury year. Need to combine 1969–70 to have a usable 'With' sample. You get 1.2 if you use 1970–71 instead]
1973–74 Lakers: 8.16 SRS with, 0.85 SRS without. Total change: +7.31. [Retirement]

For context, here's Hakeem's:
1984–85 Rockets: 1.38 with, -3.12 without. Total change: 4.5 [Rookie year]
1991–92 Rockets: 1.7 with, -1.79 without. Total change: 3.49 [injury years]
1998 Rockets: 0 with, -1.77 without. Total change: 1.77 [injury year]
2000 Rockets: -3.7 with, 2.42 without. Total change: -6.12 [injury year]
2001–02 Rockets: 2.71 with, -4.31 without. Total change: 7.02 [trade, leaving Rockets]
2001–02 Raptors: -0.7 with, 1.69 without. Total change: -2.39 [trade, joining Raptors]
2002–03 Raptors: -0.7 with, -6.1 without. Total change: 5.4 [Retirement]

Wilt's Career Average: +2.74
Wilt's healthy Career Average: +4.04
Hakeem's Career Average: +1.95 (but most samples are from post-prime)
Wilt's 10-year prime: +2.03 (10-year 60–69 prime). +2.35 (best 10-year stretch 1965-1974)
Wilt's 10-year healthy prime: +3.83 (10-year 60-69 prime). +3.87 (best 10-year stretch 1965-1974)
Hakeem's 10-year prime: +3.49 (10-year 86-95 prime, but this only includes one sample. 14-year average: +3.25 from 1985–1998)

One starts to wonder why people have been saying raw WOWY is a point for Hakeem over Wilt... But I digress
To tie this back to the players that are up for voting, by their 10-year time, the players rank like this:
10-year Prime in Multi-year WOWY samples:
Curry: +8.39 in 2014–2023, with 3 large-sample WOWY in 2018 injury, 2020–21 injury, 2022–23 injury
Magic: +7.68 in 1982–1991, with 1 sample in 1991. +6.06 in 2 samples in 11-years from 1981-1991. +4.85 in 3 samples in 12 years from 1980–91.
Garnett: +6.88 in 2000–2009, with 3 samples in 2007-08 trade from Timberwolves, 2007-08 trade to Celtics, 2008-2009 injury. Obviously the Celtics' roster was entirely revamped, so it's +3.83 if we discount that trade to the Celtics sample, just below Wilt.
Shaq: +3.86 in 1995–2004 in 3 samples, 1996-97 trade from Magic, 1996-97 trade to Lakers, 2004-05 trade from Lakers.
Wilt: +3.83 in 1960-1969 in 5 samples, see above. It would be + 4.48 if we switched the 1968–69 regular season 76ers number for the playoffs number.
Hakeem: +3.49 (see above, just 1 sample in 91-92).

So if we correct for Wilt's health, Wilt certainly doesn't seem clearly below the others in this tier. He's tied with Shaq for his 10-year prime. You can argue Wilt's 10-year prime over Shaq if you trust Wilt's larger number of samples here, or if you use Wilt's playoff change for the 1969 Lakers. You could argue Wilt's 10-year prime over Garnett's if you disregard the 2008 Celtics sample entirely since some of that change was from the rest of the roster revamping (I wouldn't throw it out entirely, but I also wouldn't take it at face-value).

Curry and Magic seem the best, but they also have the worst longevity whether you're counting total number of games (for Garnett / Shaq) or longevity relative to era (for Wilt, since players in the 2000s had on average about 70% more career games than players in the 60s)

Note that this poor health can also be used to explain Wilt's low raw WOWY numbers in Thinking Basketball's "10-year Prime WOWY" score that you see on the old Top 40 pages. It also biases the WOWYR somewhat, although WOWYR is more positive on Wilt in general because it sees Wilt correlates much better with his team's success than his teammates, even if he doesn't correlate as much in 1965 as we think he might when healthy.

...

More details on Wilt's down years here: https://forums.realgm.com/boards/viewtopic.php?p=107709117#p107709117
More info on Wilt's WOWY / team changes without him here: https://forums.realgm.com/boards/viewtopic.php?p=107709162#p107709162

OK, but Wilt got traded twice in his prime. With Hakeem most of those changes you're looking at are his impact from age 35-39 when he was well past his prime. It makes sense that Hakeem from age 35-39 shouldn't show the same level of change as peak Wilt.

I mean... I said this exact point in my post already

I mentioned that Hakeem's samples are dominated by the late-career stuff directly next to his career average, so that's why I went to their 10-year primes (which favor Wilt, but not by a big margin). But yes agreed, that's a good reason not to take much faith in the plain 'ol career average I gave when just using these limited multi-year samples, since some peoples' samples come predominantly from their non-prime years.

If you look at the overall WOWY number for Wilt and Hakeem you get:

Wilt: +6.0 WOWYR (38th), 1.1 prime WOWY (114th)
Hakeem: +5.7 WOWYR (49th), 5.2 prime WOWY (10th)

Yep, like I said at the end of the post, wilt's "+1.1 prime WOWY (114th)" score from Thinking Basketball is dominated by the injured year. The Thinking Basketball career prime scores does not use multi-year WOWY (such as trades that occur between seasons). It only looks at WOWY samples within one season (such as trades that occur half way through a season, mid-season injuries/absences, etc.). It does this to avoid being biased by other roster changes that occur between years, at the cost of a potentially smaller sample size.

For Wilt's sample, by my count (might be slightly off?), it uses: 43 missed games from 1965 76ers (vs compared to when he was on the court that year *but playing injured*), 42 missed games from 1965 Warriors (*compared to when he was on the court that year *but playing injured*), 1 missed game from 1966, 1 missed game from 1969, and... depending on when the prime years are set (can't check, the old website's down currently), either 3 missed games from 1960, or 70 missed games in 1970.

So... that Thinking Basketball raw Prime WOWY score is based off a sample that's either 54% or 94% Injured Wilt only. So that's why I mentioned that we can go to multi-year WOWY samples for Wilt, to correct for the health concern.

Or we can go to WOWYR. This adjusts for teammates (just like we adjust for teammates in RAPM), but it still uses the same sample for Wilt. And just by adjusting for teammates (with the majority of Wilt's WOWY sample in this database coming from when he was playing injured)... Wilt still ends up ahead of Hakeem.

To sum:
-When we don't adjust for Wilt's health or for teammates, prime Hakeem >> prime Wilt
-When we adjust for Wilt's health but not for teammate, prime Wilt > prime Hakeem
-When we adjust for Wilt's teammates but not for Wilt's health, prime Wilt > prime Hakeem
-When we adjust for both Wilt's teammates and for Wilt's health... we would expect prime Wilt >> prime Hakeem.

Furthermore, if you compare Wilt to Shaq who also moved teams twice in his prime, here's what you get:

1992-93 Magic: 1.35 SRS with, -6.52 SRS without. Total change: +7.87 [Rookie year]
1995-95 Magic: 5.40 SRS with, -0.07 SRS without. Total change: +5.47 [trade, leaving Magic]
1996-97 Lakers: 4.21 SRS with, +.66 without. Total change: -0.55 [trade, joining Lakers]
2003-04 Lakers : 4.35 SRS with, -2.32 SRS without. Total change: +6.67 [trade, leaving Lakers]
2004-05 Heat: 5.77 SRS with, -0.13 SRS without. Total change: +5.90 [trade, joining Heat]

So, looking at the rookie year and prime moves, you get a net average change of +5.07 for Shaq whereas for Wilt's moves during his prime you get an average of +2.03.

Yep, no qualms here. There's an argument for Shaq over Wilt in raw WOWY, absolutely! I just wanted to mention that a counter argument *was possible*, even if you still end up going for Shaq.

[Dooley]

10. Draymond, +6.6, lower bound +4.5, upper bound 8.6
11. Curry, +6.4, lower bound +4.7, upper bound 8.2
17. Shaq, +5.8, lower bound +4.3, upper bound 7.4

If someone wants to make the case that Draymond was as impactful as Curry and Shaq, I would be very interested in hearing that argument! And it would definitely make me a lot lower on their overall contributions if there's a solid case there. I'm a little skeptical because it definitely doesn't align with how I've felt watching the Warriors but it's an intriguing idea.

I'm not saying as impactful but

I do sort of make a case he was
-> very impactful
and
-> probably juices curry's impact
here:
https://forums.realgm.com/boards/viewtopic.php?p=107743950#p107743950
(bottom half of post)

I definitely think it's difficult to tease out the impact of Dray and Steph and that's one of the big challenges with evaluating Steph. But intuitively it feels like most of the credit should go to Steph on offense and (based on discussion in the last few posts in the thread) it seems that the various on/off numbers bear out that Steph is more valuable. I'm also not sure that you can rely too much on the fact that Steph's meteoric rise happened at the same time Dray became a starter, because there were several other changes at the same time - Steph was getting better as he entered his prime, and it's also obviously coincident with a big coaching change and the implementation of a new offensive system (which was basically built around Steph's unique strengths).

However - I'm also curious if there's really an argument that Dray was on the same impact tier as Shaq and Curry, because it's quite counterintuitive (and because it seems like that would have big implications for thinking about KG).

[eminence]

I find it more likely Dray is draining Steph than the reverse (or that someone else on the squad is missing out, namely Klay as the other constant).

'14-'15 to '22-'23 (minus '20) RS/PO combined, P/M and possessions from pbpstats.com, Net calculated and rounded to the nearest tenth
Both: 17278 minutes, +13.9 Net
Steph: 5728 minutes, +8.1 Net
Dray: 5532 minutes, +1.6 Net
Neither: 9479 minutes, -5.2 Net

Well I think if you're trying to iisolate for their influence on each other, WOWY is probably preferable(though I do not know how one would filter out 2020 net nor do i ran out of statmuse searches before i could check net). And by record Draymond does not look like a drain at all.

Nonetheless going by lineup-stuff

For that frame, in the regular-season, PBP has steph no dray at +7.97 and Dray no Steph at +2.22. Both +14, neither at -4.14

For full-season dray no steph is +7.3 and steph no dray is +11.6. Both is +18.30, neither is -7.12

For the 15-17 regular seasons(the years they really good) steph no dray is +11.6, Dray no steph is +7.34, both is +18.3, neither is -7.12
For full season it's +8.6 for dray no steph and +9.6 for steph no dray, without -7.29, both +17.09

If you go just 15-16 things actually swing to Draymond both rs and playoffs.

Doesn't look like a drain to me

Neither is ‘draining’ the other in terms of team success(you need multiple great players playing well to hit the Warriors sustained heights). But when you’re talking about Dray ‘juicing’ Stephs APM type numbers, I’d say those numbers suggest the opposite. Steph juicing Dray at his own expense.

[OhayoKD]

If someone wants to make the case that Draymond was as impactful as Curry and Shaq, I would be very interested in hearing that argument! And it would definitely make me a lot lower on their overall contributions if there's a solid case there. I'm a little skeptical because it definitely doesn't align with how I've felt watching the Warriors but it's an intriguing idea.

I'm not saying as impactful but

I do sort of make a case he was
-> very impactful
and
-> probably juices curry's impact
here:
https://forums.realgm.com/boards/viewtopic.php?p=107743950#p107743950
(bottom half of post)

I definitely think it's difficult to tease out the impact of Dray and Steph and that's one of the big challenges with evaluating Steph. But intuitively it feels like most of the credit should go to Steph on offense and (based on discussion in the last few posts in the thread) it seems that the various on/off numbers bear out that Steph is more valuable. I'm also not sure that you can rely too much on the fact that Steph's meteoric rise happened at the same time Dray became a starter, because there were several other changes at the same time - Steph was getting better as he entered his prime, and it's also obviously coincident with a big coaching change and the implementation of a new offensive system (which was basically built around Steph's unique strengths).

However - I'm also curious if there's really an argument that Dray was on the same impact tier as Shaq and Curry, because it's quite counterintuitive (and because it seems like that would have big implications for thinking about KG).

That would be a one or two-year "peak" argument centered around 15/16(and yes that can be made emperically). Over those two years steph no dray ~ dray no steph in the rs and full season it favors dray. Do I think he would be more valuable in a variety of situations? No. But situationally Dray does have favorable snippets here and there(especially if you lean into the playoffs). Steph probably should receive more credit, but in a comparison with the field, Steph is advantaged situationally I think having had is minutes tied very closely(79% of his minutes from 2015-2019 are with Draymond), a guy who looks very impactful without Steph in the lineup and in entire games where Steph isn't a facotr:

I find it more likely Dray is draining Steph than the reverse (or that someone else on the squad is missing out, namely Klay as the other constant).

'14-'15 to '22-'23 (minus '20) RS/PO combined, P/M and possessions from pbpstats.com, Net calculated and rounded to the nearest tenth
Both: 17278 minutes, +13.9 Net
Steph: 5728 minutes, +8.1 Net
Dray: 5532 minutes, +1.6 Net
Neither: 9479 minutes, -5.2 Net

Well I think if you're trying to iisolate for their influence on each other, WOWY is probably preferable(though I do not know how one would filter out 2020 net nor do i ran out of statmuse searches before i could check net). And by record Draymond does not look like a drain at all.

Nonetheless going by lineup-stuff

For that frame, in the regular-season, PBP has steph no dray at +7.97 and Dray no Steph at +2.22. Both +14, neither at -4.14

For full-season dray no steph is +7.3 and steph no dray is +11.6. Both is +18.30, neither is -7.12

For the 15-17 regular seasons(the years they really good) steph no dray is +11.6, Dray no steph is +7.34, both is +18.3, neither is -7.12
For full season it's +8.6 for dray no steph and +9.6 for steph no dray, without -7.29, both +17.09

If you go just 15-16 things actually swing to Draymond both rs and playoffs.

Doesn't look like a drain to me

Neither is ‘draining’ the other in terms of team success(you need multiple great players playing well to hit the Warriors sustained heights). But when you’re talking about Dray ‘juicing’ Stephs APM type numbers, I’d say those numbers suggest the opposite. Steph juicing Dray at his own expense.

Then we are interpreting those numbers very differently. How do you arrive at "steph's apm is being drained by dray" from any of this? Steph plays almost all his minutes with someone who looks great without him(impact-wise at least). Both should benefit from playing with each other

[ZeppelinPage]

OK, but Wilt got traded twice in his prime. With Hakeem most of those changes you're looking at are his impact from age 35-39 when he was well past his prime. It makes sense that Hakeem from age 35-39 shouldn't show the same level of change as peak Wilt.

For Wilt's sample, by my count (might be slightly off?), it uses: 43 missed games from 1965 76ers (vs compared to when he was on the court that year *but playing injured*), 42 missed games from 1965 Warriors (*compared to when he was on the court that year *but playing injured*), 1 missed game from 1966, 1 missed game from 1969, and... depending on when the prime years are set (can't check, the old website's down currently), either 3 missed games from 1960, or 70 missed games in 1970.

So... that Thinking Basketball raw Prime WOWY score is based off a sample that's either 54% or 94% Injured Wilt only. So that's why I mentioned that we can go to multi-year WOWY samples for Wilt, to correct for the health concern.

Admittedly not big into WOWY but wouldn't the 76ers having injuries to Greer, Costello, and Jackson after starting 9-2 impact Wilt's score?

As well as the Warriors replacing Wilt after the trade with a Hall of Fame center in Nate Thurmond?

I genuinely don't know how much WOWY accounts for things like that. But, along with the small sample size and him coming off a possible heart attack, it definitely seems like that would impact his WOWY. That "Prime WOWY" score just doesn't seem right to me considering how close Wilt had injured teams like the '62 Warriors and '65 76ers to beating the Celtics. The way you accounted for Wilt's injury definitely seems closer to what one would expect for Wilt's impact.

[LukaTheGOAT]

As a general note, I find JE’s RAPM to be pretty confusing, to be honest. As far as I’m aware, it’s not supposed to have a box component. But there’s outcomes that don’t really make sense as a pure RAPM outcome. Like, for instance, Kevin Garnett ranked #2 in the NBA (and #14 across the entire set) in JE’s RAPM in 2004-2005, in a year where his raw on-off was only +0.7 (and he had no playoff numbers so that didn’t change it). Now, I get that raw on-off and RAPM aren’t the same, but they’re very closely related (RAPM basically just takes raw on-off and makes some adjustments based on who’s on the floor), and it’s quite difficult for me to understand how a player could come out as #2 in the NBA in RAPM in a season where their raw on-off is +0.7. I think the “prior-informed” part of it must be doing quite a lot of work in that model. But I don’t really know what his “priors” are, and whether they’re box-score-based priors, or past years’ on-off, or something else (JE has apparently even used height as a prior in RAPM before). It seems obvious to me that it is very far from just raw RAPM. Does anyone know what priors he’s using? My guess is that it’s at least in part past years’ on-off—which would explain why 2004-2005 Garnett is ranked really highly, since Garnett had really high on-off the previous couple years—but then if there’s a really strong prior for past years’ on-off (which there’d have to be to get a +0.7 raw on-off to #2 in the NBA and #14 in the last 25 years in RAPM), then the model is going to fail to account for quick changes in a player’s ability/impact (i.e. for instance, Steph Curry getting incredibly good really fast, after having had merely pretty good on-off his first few seasons while not really being used to his full potential).

I'm pretty sure each year takes other years in account as a prior. So KG's 05 year, is influenced by his 03 and 04 years which are some of the most bonkers stuff you will see.