RealGM

I know I have LeBron at 1 but I am vascillating between Jordan and Shaq for 2.

Can someone make the case for 1991 Jordan over 2000 Shaq?

homecourtloss wrote:

falcolombardi wrote:

JordansBulls wrote:1. 1991 MJ - All around great season, all nba first team and defense, led in scoring beat the 2x defending champion, beat the guy who finished 2nd in MVP voting on an all time great team for a franchise that never won before and drafted him and was the only allstar on the squad.

2. 1971 Kareem - 2nd year in the league won league mvp, finals mvp for a franchise that never won before and drafted him.

3. 1994 Hakeem - MVP, DPOY the same year, won title for franchise that never won before and that drafted him and didn't switch teams to win.

sorry but what the f**k?

scottie pippen was a star, the "only star on his team" argument applies to 94 hakeem, not to 91 jordan

jordan is great enough there is no need to make up this stuff to prop him up lol, jordan doesnt need pippen diminished to be argued as the goat

At minimum, Pippen was a top 10 impact player in 1991 and probably higher. And, since the poster who posted this likes anecdotal comments, Magic Johnson, though no oracle of analysis, in 1992 while calling the NBA Finals mentioned that if you have Pippen and Jordan on the floor, you’re going to win. Jordan had at minimum top 10 player playing with him every season from 1991 to 1998.

Pippen production in '91 is not Top 10. He didn't make the AS game that season so technically Jordan is the only Bulls star.

'91 top 10 in no order who are better than Pip

Jordan
Hakeem
D-Rob
Barkley
Malone
Ewing
Clyde
Magic
Bird
Bernard

Pip has to compete with the like of Stockton, Issiah Thomas, Dominque to name a few..

2klegend wrote:

homecourtloss wrote:

falcolombardi wrote:
sorry but what the f**k?

scottie pippen was a star, the "only star on his team" argument applies to 94 hakeem, not to 91 jordan

jordan is great enough there is no need to make up this stuff to prop him up lol, jordan doesnt need pippen diminished to be argued as the goat

At minimum, Pippen was a top 10 impact player in 1991 and probably higher. And, since the poster who posted this likes anecdotal comments, Magic Johnson, though no oracle of analysis, in 1992 while calling the NBA Finals mentioned that if you have Pippen and Jordan on the floor, you’re going to win. Jordan had at minimum top 10 player playing with him every season from 1991 to 1998.

Pippen production in '91 is not Top 10. He didn't make the AS game that season so technically Jordan is the only Bulls star.

'91 top 10 in no order who are better than Pip

Jordan
Hakeem
D-Rob
Barkley
Malone
Ewing
Clyde
Magic
Bird
Bernard

Pip has to compete with the like of Stockton, Issiah Thomas, Dominque to name a few..

that pippen didnt make the all star doesnt mean he was not an all star level player at minimum

i would never take 2022 wiggins over 1991 pippen eithet and guess who was an all star

ceiling raiser wrote:Question for people choosing MJ -

If we had data demonstrating that MJ at his best had a comparable defensive impact (in terms of DRAPM, or defense on/off the court, or other data) to Kobe, how would it impact your selection of his early 90s seasons as GOAT peak?

A great question to be asked.

So first, broadly: You better believe I take this data seriously when I get access to it. Getting access to the 76er dataset (thank to you I believe!) was eye-opening about Dr. J, and it has lowered his standing on my lists.

I will zag back in the other direction to emphasize that such data doesn't remove the end-accomplishments of the players involved. If the team just keeps winning when you're out there running the show all the way to the chip, well, that's your job. Beating the other team by as many points as possible, by contrast, is not the goal

But lastly:

I'm a bit surprised people (because someone else mentioned this possibility recently) are talking as if we have no data along these lines. It's not a question, for example, whether Jordan ever had a regular season +/- north of Kobe's best, because we have data from Jordan's last 2 years in Chicago, and that already settles the matter.

We also have access to regular season +/- data back through '93-94, and that gives us enough so that we already know that Jordan led his team in +/- at least as many times as Kobe did - Kobe did it 3 times in his career, Jordan did it in '95-96, '96-97 & '97-98.

So yeah, while it's within the realm of possibility that Kobe-like tendencies will result in +/- data underrating Jordan's total contribution, there's nothing like the "Whoa, Kobe really doesn't look that impactful by these metrics?" thing I detect for Jordan.

giordunk wrote:I still always think it's literally impossible for 2009 LeBron to be the best LeBron.

So, I feel compelled to ask:

What is it that makes you feel it's impossible besides the fact his team didn't win the title?

ftr, I like the arguments being made for '15-16 LeBron, and I think '12-13 LeBron has a good case too, so I'm not necessarily looking to advocate for '08-09 LeBron, but I wouldn't dismiss it lightly.

falcolombardi wrote:

2klegend wrote:

homecourtloss wrote:
At minimum, Pippen was a top 10 impact player in 1991 and probably higher. And, since the poster who posted this likes anecdotal comments, Magic Johnson, though no oracle of analysis, in 1992 while calling the NBA Finals mentioned that if you have Pippen and Jordan on the floor, you’re going to win. Jordan had at minimum top 10 player playing with him every season from 1991 to 1998.

Pippen production in '91 is not Top 10. He didn't make the AS game that season so technically Jordan is the only Bulls star.

'91 top 10 in no order who are better than Pip

Jordan
Hakeem
D-Rob
Barkley
Malone
Ewing
Clyde
Magic
Bird
Bernard

Pip has to compete with the like of Stockton, Issiah Thomas, Dominque to name a few..

that pippen didnt make the all star doesnt mean he was not an all star level player at minimum

i would never take 2022 wiggins over 1991 pippen eithet and guess who was an all star

In hindsight, it is an easy choice. As always, in '91 it wasn't that clear cut. Pip was still an up-and-coming raw talent. No doubt he exceeds people expectations largely due to MJ pushing him. But he wasn't anything spectacular. You can't run an offense through him. He is the best Robin-type role. He is not a franchise-level player.

ardee wrote:I know I have LeBron at 1 but I am vascillating between Jordan and Shaq for 2.

Can someone make the case for 1991 Jordan over 2000 Shaq?

I might start by taking a look at this vid if you haven't already (Shaq segment around the 18 minute mark), and then watching until the end. It also shines some positive light on what Miami Lebron's offense, and shows why some might think more highly of that version than his 2nd Cleveland stint.

;t=297s

Doctor MJ wrote:I'm a bit surprised people (because someone else mentioned this possibility recently) are talking as if we have no data along these lines. It's not a question, for example, whether Jordan ever had a regular season +/- north of Kobe's best, because we have data from Jordan's last 2 years in Chicago, and that already settles the matter.

We also have access to regular season +/- data back through '93-94, and that gives us enough so that we already know that Jordan led his team in +/- at least as many times as Kobe did - Kobe did it 3 times in his career, Jordan did it in '95-96, '96-97 & '97-98.

Great point. This is data I take pretty seriously. I think there are some issues with the Pollack numbers, however they are probably in the ballpark.

So yeah, while it's within the realm of possibility that Kobe-like tendencies will result in +/- data underrating Jordan's total contribution, there's nothing like the "Whoa, Kobe really doesn't look that impactful by these metrics?" thing I detect for Jordan.

So I might be making a mountain out of a molehill with the limited numbers Squared2020 produced. Two things came to mind (taking the data at face value, which is dangerous given the error bounds and sample size!):

(1) It did indeed remind me of the Dr. J data showing very limited defensive impact. Counterpoint: Jordan's on-off in 96, 97, 98 looks very strong.

(2) Looking like a peer of Magic's in 85, 88, 91 in terms of overall impact as well as similar in off/def split has caused me to adjust my priors a bit, and elevate Magic while dropping Jordan. In actuality, such limited data shouldn't cause me to do both (and even one is questionable). But I guess my larger point is that if Jordan was similar to Kobe defensively, that he'd be comparable to Magic in terms of peak, and the two should be a lot closer than conventional wisdom suggests. Counterpoint: I believe that in 97 and 98 NPI RAPM Jordan looks strong defensively. Counter-counterpoint: Kobe looked pretty strong in DRAPM in 09 and 10 iirc.

Doctor MJ wrote:

giordunk wrote:I still always think it's literally impossible for 2009 LeBron to be the best LeBron.

So, I feel compelled to ask:

What is it that makes you feel it's impossible besides the fact his team didn't win the title?

ftr, I like the arguments being made for '15-16 LeBron, and I think '12-13 LeBron has a good case too, so I'm not necessarily looking to advocate for '08-09 LeBron, but I wouldn't dismiss it lightly.

there is a in-built bias towards title winning seasons, a natural dogmatic belief that a player is better in his late 20's or maybe early 30's than before,

than a somewhat older version of a player that is more skilled and experienced is better for winning than a younger version that is more dependant on superior athletism

a plauer greater stats in his mid 20's get diminisjed and weaker stats in their late 20's or early 30's handwaved away, usally because the latter are winners and the former are not, when often it may be simply that the team was weaker in that player mid 20's

88-89 jordan teams vs 91-93 jordan teams are a good example where player improvement may get the credit that should go instead to supporting cast improvement

mid 20's kareem winning in 71 but losing in 77 is a gold example of the opposite, 71 was not necesarrily better but it had much better teammstes)

this is specially important cause teams that get superstars usually do throufg the draft amd are bad teams, making title winning teams around your franchise player is hard and if done at all is likely to take a while so playera are mpre likely to win older than younger regardless of their level of impact

when the opposite happens you usually see players win more when younger (again think kareem in his early 20's vs late 20"s to very early 30's)

all of these are sometimes treated as a dogma

BTW, only ~6% of the season, but 95-96 RAPM so far from Squared2020:



Splits aren't meaningful, but yet another data point showing MJ as an absolute impact monster.

1) LeBron James (2016)
B) LeBron James (2017)
2) Michael Jordan (1991)
Michael Jordan (1990)
3) Tim Duncan (2003)

I have LeBron James at #1 as I believe he reached a level in the post-season no other player has reached. He was impacting more possessions (Offense and Defense combined) in games no other player in NBA history could have impacted and did so for multiple games per series.

This is my quote from Odinn's Top 12 seasons of Player X and Y thread, which I still standby to this day.

The first tier are the 4 seasons where these players had the combination of physical imprint combined with a clear understanding of the system and could bring their games to levels, when needed, that no other perimeter players in NBA history have achieved. I value Lebron's offensive heliocentrism here and his team defense more in a vacuum than Jordan's scoring and turnover economy with his risk-defined style on the defensive end.

Focusing more on Jordan and LeBron versus the rest of the perimeter players, I am not seriously considering another perimeter player until #10, which shows just how incredible Jordan and LeBron were as two-way players, high volume scorers and offensive savant's in offensive systems they understood inside-and-out.

Jumping into some statistics maybe not covered in the 10-series novel posted earlier in this thread (which I haven't read in-full yet)...

LeBron James vs Detroit: 4 Games, 22.8/9.0/6.8, 1.8 steals, 21.1 3P%, 3.3 TO

Essentially a mediocre to below average prime LeBron series by the box-score, but the most important part of every player at this stage in the Top Peaks (and really all 40 of the top peaks) is how much impact a player has beyond the box-score. Kyrie Irving, for example, had a spectacularly efficient series: 27.5 Points, 6.5 TOV%, 58.5 TS%. The point emphasis and focus though should be this: The Cavaliers played at a net +16.8 Points/100 with James on the court and +3.0 Points/100 with Kyrie on the court. The Cavaliers offense exceeded 120 Ortg with James on the court, a level reserved for the GOAT offensive players.

LeBron James vs Atlanta: 24.3/8.5/7.8, 3.0 steals, 4.25 TO, 57.3 TS%

Box-Score looks a bit better, but let's back up. Atlanta is a 101.4 Defense Rating (2nd in NBA) and 3.49 SRS team. This team is significantly better than either opponent Jordan faced in round 1 or 2 in 1991. The Hawks are coming off a series played in 2016 holding their opponents to a 94.5 Ortg. That's incredible, and the team was Boston who was respectable (2.83 SRS, 8th in NBA, 10th rated Offense and 4th Rated Defense).

Now, look at what LeBron and the Cavaliers did to dismantle this defense--the 2nd best in the league and one looking like an all-time great defense through the regular season and 1st round of the playoffs...The Cavaliers played at a net +19,1 Points/100 with James on the court and +18.4 Points/100 with Kyrie on the court and +23.3 with Kevin Love. If anyone think LeBron is difficult to play with or diminishing his teammates efficiency...
LeBron James On-Court Ortg: 122.2
Kevin Love On-Court Ortg: 123.7
Kyrie Irving On-Court Ortg: 124.6

These numbers are against a 101.4 Rated Defense, that's producing an offense over +20 points per 100 better than what this defense gave up in the regular season and nearly +30 points per 100 better than what they gave up against the 10th best offense in round 1 over 6 games.

Next-up, LeBronto--the team that wins the championship in its first year after LeBron leaves the conference.

LeBron James vs Toronto: 26.0/8.5/6.7, 2.0 steals, 66.5 TS%

There is the ultra-efficient LeBron scoring series but perhaps not even the best part of the series from LeBron. While he posted an all-time great scoring series, and led the Cavaliers to a 121.3 Ortg when on the court resulting in a +19.2 Net Rating per 100 when on the court, it was his defense against the 5th best offense which leaves me the most impressed. The Cavaliers best defenders by on-court rating were point-of-attack defenders (We saw this in 2020 with Caruso as well where LeBron and a POA defender were all-time great levels defensively) and this is something when combined with LeBron James makes his defense unfair. LeBron is so cerebral and intelligent defensively, having someone be able to disrupt and force the offense into making a decision on your clock is playing right into the hands of a defense and IQ Savant like James. This skill is also something which made Jordan so amazing at getting steals and being disruptive when his Point-of-attack point guards like Armstrong and Harper where running around hounding the ball.

NBA Finals--It has been talked about extensively. Best Finals ever and best 3-game stretch in NBA History and not just because LeBron was hot shooting.

I will focus more on Jordan and Duncan in the next thread (likely LeBron again since Jordan will deservedly win this thread by the looks of it)

#2 Jordan: Simply the best with LeBron as a perimeter player, good enough defensively to have enough impact to match the all-time great two-way big men (Hakeem, Kareem, Wilt, Russell (more defensively and less two-way), Walton, Duncan, Garnett, Robinson off the top of my head).

#3 Duncan: I voted him #1 in the last peaks project. Simply able to appreciate Jordan and LeBron as defenders a bit more as well as their volume/resiliency as scorers carrying more weight for me than it has in the past. I like Duncan the best as a two-way guy out of all the big-men which makes him #3.

1. 1991 Jordan
(Honorable mention years: 1b. 1990 Jordan, 1c. 1989 Jordan)
2. 2013 LeBron
(Honorable mention years: 2b. 2012 LeBron, 2c. 2016 LeBron, 2009 2d. LeBron).
3. 1967 Wilt
(Honorable mention opponents: 2017 Curry, 2000 Shaq).

To me, Jordan and LeBron have some separation as the clear greatest peaks of all time.

Jordan's year choice: Jordan's best years were 89-91, with the earlier years having greater athleticism (so better driving on offense and motor on defense), while the later years having more greater skill, control, and scalability (so better midrange shooting / touch, better passing, better off-ball ability to pair with teammates). In 92 and on, his lost athleticism and motor pulls down his value. Of 89-91, I prefer the older one of the three.
Why 91 > 90, 89? Qualitatively, I tend to be higher on creation and scalability with teammates over raw defensive motor. Quantitatively, you can see the impact of Jordan's greater scalability / teammate fit on the performance of the Bulls. The Bulls clearly improved from 89 to 91 (https://projects.fivethirtyeight.com/complete-history-of-the-nba/#bulls). Some of this is attributable to Jordan's improved teammates of course, but I credit some of it to Jordan's increased familiarity and effectiveness within the Triangle system. It's also noteworthy that 91 is likely his best playoffs of the three -- I credit his increased experience as the cause for his resilience here.

LeBron's year choice: LeBron is probably the hardest player to choose a single peak year for. There's at least 3 clear local peaks: in 2009, in 2012 playoffs / 2013 regular season, and in 2016 (with smaller secondary peaks in 2017 or 2018 playoffs if people are very low on the regular season and on defense). I favor 2012/2013 > 2009 and 2016?
Why 2013 > 2009? I can see arguments for each. 2009 may be his best floor-raising year, and the regular season impact metrics may peak here with his all-time motor. However, I have concerns about his lack of scalability and resilience. I think 2009 would have had the same playoff resilience struggles that 2010/2011 LeBron facing the more drop/zone centric Celtics/Mavs/Spurs defensive schemes. This came from a lack of offensive versatility, particularly with his lack of shooting, post game, off-ball ability, and without reaching his peak yet with passing, BBIQ, or controlling the pace of the game. That same lack of versatility also hampers his value in a ceiling-raising ability alongside high-volume all star teammates (again, see 2011 LeBron's fit with Wade vs 2012/13's fit with Wade).
Why 2013 ? 2016? I could see arguments for either, and there certainly have been some compelling arguments for 2016 in this thread already. The 2013 regular season certainly surpasses the 2016 regular season (e.g. in AuPM, Backpacks Box Plus Minus, and Corp Plus Minus). 2016 does have some superior metrics in the playoffs (AuPM, Backpicks BPM, Goldstein RAPM/PIPM), and visibly improved on defense vs regular season 2016 LeBron, and improved in his passing / control of the game vs 2013 LeBron. But 2016 also had some favorable context in the playoffs: he coasted in the regular season and so had more energy, he had healthy teammates (vs Wade who was injured), and he had injured/suspended opponents. If 2013 LeBron had this kind of favorable context, perhaps we'd be talking about 2013 as his best playoffs. The fact that he had one of his best playoffs immediately before in 2012 gives credence to this idea. When I also consider that 2013 LeBron had a higher motor for more consistent defensive effort, and had a better off-ball game and post-game, I end up siding with 2013 over 2016 (though it's understandable if you disagree).

Jordan vs LeBron:

Let's consider impact stats. I'm using the closest years I can find for Jordan, and showing their rank among the top ~20 peaks of all time in parentheses (only 1 year per player).
Ai. Regular Season Augmented Plus Minus per game: +6.6 2013 LeBron (4th all time) > +6.5 1991-1993/1996 Jordan (5th all time). [+7.8 2009 LeBron (would be 1st all time), +5.7 2016 LeBron (would be 7th all time)]
Aii. Postseason Augmented Plus Minus per game: +6.1 1997 Jordan (4th all time) > +5.8 2013 LeBron (6th all time)
[+6.9 2012 LeBron (would be 3rd all time), +8.7 2009 LeBron (would be 1st all time), +7.3 2016 Lebron (would be 3rd all time)].
Bi. Regular Season Backpicks Box Plus Minus: +9 1991 Jordan (1st all time) > +8.8 2013 LeBron (3rd all time) [+8.5 2009 LeBron (would be 5th all time), +6 2016 LeBron (not top 20)]
Bii. Postseason Season Backpicks Box Plus Minus: +12.1 1991 Jordan (1st all time) > +8.3 2013 LeBron (8th all time) [+12 2009 LeBron (would be 2nd all time), +9.5 2016 LeBron (would be 2nd all time)]
C. Goldstein RAPM / Historical RAPM: +6.82 1991 Jordan (12th al time) > +6.4 2013 LeBron (13th all time) [+7.47 1988 Jordan (would be 7th all time), +9.5 2012 LeBron (would be 2nd all time), +8.84 2009 LeBron (would be 3rd all time), +8.62 2016 LeBron (would be 3rd all time)]
D. Backpicks Championship Over Replacement Player percentage: 31.1% 1991 Jordan (1st all time) > 28.7% 2013 LeBron (3rd all time) [31.2% 1989 Jordan (would be 1st all time), 31.1% 1990 Jordan (would be 1st all time), 27.1% 2009 LeBron (would be 3rd all time), 26.5% 2016 LeBron (4th all time)]

Jordan and LeBron both have numerous years that we could say are their peaks, and whichever we pick, advanced metrics would have them near the top. Ultimately, I side with Jordan. With many metrics, we only have small samples or off-peak years for Jordan, yet he still compares favorably.
A. For Regular season AuPM, his years from 1991-1993/1996 are 1 rank below 2013 LeBron; for postseason RAPM, 1997 Jordan is 1 rank below LeBron. I think it's fair to say in his peak, we would expect 1991 Jordan to improve over 2013 LeBron in AuPM.
B. For both regular season and post season Backpicks BPM, 1991 ranks first all time over 2013 LeBron.
C. In our (very limited) RAPM sample, 1991 Jordan is just over LeBron. The Bulls actually underperformed vs their season average in this sample, so it's possible Jordan would rank higher for a full season RAPM, though of course small-sample noise is an issue.
D. In Backpicks CORP%, 1991 Jordan ranks first all time over 2013 LeBron.

Of course, you could take another year for LeBron. 2016 LeBron performs worse in regular season AuPM, better in postseason AuPM, much worse in regular season BPM, better in postseason BPM, better in RAPM (though worse vs 2012's RAPM), and worse in CORP%. If we were to combine the better qualities of LeBron's various peaks (e.g. 2013's regular season motor and defense with 2016's postseason performance), it would absolutely beat 1991 Jordan. But... that's not what happened. This is what ultimately puts Jordan's peak for me. Jordan put together all his best qualities into a single season. LeBron's best qualities are spread across numerous seasons 7+ years apart, giving him clearly better longevity, but a worse one-year peak.

*** 2 Down, 1 to go

After a small gap, we reach the next peak of players. To me, some possible candidates are: Shaq, Wilt, Hakeem, Curry, Duncan, Kareem.

Here are their peak years ranked by the advanced stats I discussed above.
First, the modern stats:
Ai. Regular Season AuPM: [2016 Curry...], 2017 Curry, 2000 Shaq, 2001 Shaq, 2003 Duncan... 1994 Hakeem (missing: Wilt, Kareem -- too old)
Aii. Postseason AuPM: 2017 Curry, 2003 Duncan, 2001 Shaq, 2000 Shaq... (missing: Hakeem, Wilt, Kareem)
C. GoldsteinRAPM: ... 2003 Duncan... 2000 Shaq, 2017 Curry... 2001 Shaq, 91 Hakeem (missing: Wilt, Kareem, peak Hakeem)

Next, the historical stats:
Bi. Regular Season Backpicks BPM: [2016 Curry]... 1967 Wilt... [1972 Kareem]... 00 Shaq... 2017 Curry... 93 Hakeem... 03 Duncan... 01 Shaq... 94 Hakeem... 77 Kareem.
Bii. Postseason Backpicks BPM: ... 2017 Curry/2003 Duncan... 1977 Kareem... 1967 Wilt, 1994 Hakeem... 2001 Shaq... 1993 Hakeem... 2000 Shaq
D. CORP%: ... 2000 Shaq... 1993 Hakeem /[ 2016 Healthy Curry], 2001 Shaq, 1977 Kareem... 1967 Wilt, 2017 Curry, 2003 Duncan, ... 1994 Hakeem

Looking at each player's actual plus minus value (e.g. taking the average), we find a clear top 3 of: 1967 Wilt, 2000 Shaq, and 2017 Curry. We find the same top 3 if we look at each player's rank. All three gain small separation over Duncan and clearer separation over Hakeem and Kareem.
Does this make sense? I think so. 2001 Shaq might have been slightly better than 2000 Shaq in the postseason, but he suffers from a noticeably worse regular season (and this coasting may have enabled his improved performance in the post season). 1977 Kareem also suffers from a poorer regular season relative to competition. Like LeBron, if we combined their better regular seasons with their better postseason, we might get someone clearly better than the competition. Without this, they gain longevity but fall back in 1 year peaks. Hakeem also falls back. Duncan has the best argument, and I wouldn't be upset if other data supported Duncan. Basketball reference's BPM favors him, and WS/48 says he's in the mix, though WOWY, on/off, ESPN's RPM, and scalability seem to favor the current top 3. Like Kareem, it seems Duncan gains when we consider longevity, and falls slightly for 1 year peak.

So who do we take between 1967 Wilt, 2000 Shaq, and 2017 Curry?
Curry vs Shaq: AuPM certainly favors 2017 Curry > 2000 Shaq in both the regular season and post season (and it favors healthy 2016 Curry by even more). Box Plus Minus also favors Curry > 2000 Shaq (by a lot in the 2017 postseason in 2016 regular season, though 2017 Curry falls just behind in the regular season). 2000 Shaq does beat 2017 Curry in RAPM by a touch and in CORP% (though healthy 2016 Curry wins in RAPM and closes some of the gap in CORP%). TBH I'm surprised ... I thought Shaq would be the clear better peak.
Maybe more stats would help Shaq come out? Curry still comes out on top in ESPN's RPM and raw on/off (by a lot, though I've never been a fan of raw on/off). My back-of-the-napkin calculation of WOWY has them similar, while WS/48 and BR's BPM aren't conclusive either (Shaq takes the regular season, while Curry takes the postseason).
... I knew I had a Curry bias, but I was sure I'd take Shaq over Curry. Despite all that... I think I'm convincing myself of 2017 Curry over 2000 Shaq. Certainly the data seems to support it.
The main counters I can think of are: 1. if you're lower on health risks (I'm not for 1 year peaks, though I am for longevity), 2. if you think all the data is underrating Shaq's defense (... eh), 3. if you think the resilience disparity is wayyy in Shaq's favor (I don't think it's enough if we give Curry health), 4. if you think Curry's fit improves his data more than Shaq's fit (maybe?), or 5. if you use a backward Time Machine argument (while ignoring a forward time machine argument).
You know what? Screw it. I'm doing it. I'm saying it. 2017 Curry > 2000 Shaq. Come at me haters (seriously, please do... I'd love to hear if there's data to prove this wrong).

Wilt vs Curry: This is much harder comparison, given our lack of data. Backpicks Box Plus Minus is mixed: 1967 Wilt beats 2017 Curry in the regular season, though he loses in the playoffs (and loses to 2016 healthy Curry). There's a similar pattern with CORP: 1967 Wilt beats 2017 Curry, though he loses to healthy 2016 Curry. For what it's worth, WS/48 and and WOWY favor Wilt. We now get to the same questions Curry faced with Shaq:
1. Health risks (which favor Wilt > Curry)
2. measuring defense with limited data (the data is certainly sparser and noisier in the 60s, but there's no doubt Wilt > Curry on defense)
3. resilience (both were seen by media as playoff strugglers. Wilt suffered from GOAT-level defensive opponents, while Curry suffered from health. With favorable context, I see them equally resilient)
4. Fit (both had favorable fit in their peak year)
5. Time machine argument (the 60s certainly favor Wilt over Curry. In future eras, Wilt gains value with data / coaching / strategy / medicine, while losing value with the 3 point line and increased perimeter focus).
Certainly, 1967 Wilt has greater uncertainty vs 2017 Curry. But it seems to me like he's slightly favored in the data, and remains favored when we consider the 5 contextual factors. Phew, I was worried everyone would see me as a biased Curry fan! 1967 Wilt > 2017 Curry.

Sources:
Backpicks source: https://backpicks.com/metrics/player-seasons/ and https://backpicks.com/metrics/historical/corp-valuations/.
AuPM source: https://docs.google.com/spreadsheets/d/1A-vid8kuilwdyjKNLDKu6Qvc-6pARwQGjAXMf-2ku4k/edit#gid=0
RAPM source: https://docs.google.com/spreadsheets/d/1eK0i6L0q2Brih5nKOKZLGHVofY0JWKOlnnEaSMu1LTM/edit#gid=0)
Historical RAPM pre-1997: https://squared2020.com/2021/09/11/1990-1991-nba-rapm/

ceiling raiser wrote:

Doctor MJ wrote:I'm a bit surprised people (because someone else mentioned this possibility recently) are talking as if we have no data along these lines. It's not a question, for example, whether Jordan ever had a regular season +/- north of Kobe's best, because we have data from Jordan's last 2 years in Chicago, and that already settles the matter.

We also have access to regular season +/- data back through '93-94, and that gives us enough so that we already know that Jordan led his team in +/- at least as many times as Kobe did - Kobe did it 3 times in his career, Jordan did it in '95-96, '96-97 & '97-98.

Great point. This is data I take pretty seriously. I think there are some issues with the Pollack numbers, however they are probably in the ballpark.

So yeah, while it's within the realm of possibility that Kobe-like tendencies will result in +/- data underrating Jordan's total contribution, there's nothing like the "Whoa, Kobe really doesn't look that impactful by these metrics?" thing I detect for Jordan.

So I might be making a mountain out of a molehill with the limited numbers Squared2020 produced. Two things came to mind (taking the data at face value, which is dangerous given the error bounds and sample size!):

(1) It did indeed remind me of the Dr. J data showing very limited defensive impact. Counterpoint: Jordan's on-off in 96, 97, 98 looks very strong.

(2) Looking like a peer of Magic's in 85, 88, 91 in terms of overall impact as well as similar in off/def split has caused me to adjust my priors a bit, and elevate Magic while dropping Jordan. In actuality, such limited data shouldn't cause me to do both (and even one is questionable). But I guess my larger point is that if Jordan was similar to Kobe defensively, that he'd be comparable to Magic in terms of peak, and the two should be a lot closer than conventional wisdom suggests. Counterpoint: I believe that in 97 and 98 NPI RAPM Jordan looks strong defensively. Counter-counterpoint: Kobe looked pretty strong in DRAPM in 09 and 10 iirc.

All good thoughts.

I'm just not ready to draw conclusions from Squared2020's data yet. Big fan of what he's doing, but as an analyst, the fact that we know that there is a bias to which games have been most readily available really concerns me, and it makes me unsure of how much of a sample he needs to get to before I start attempting to make sense of what the results tell me.

falcolombardi wrote:

Doctor MJ wrote:

giordunk wrote:I still always think it's literally impossible for 2009 LeBron to be the best LeBron.

So, I feel compelled to ask:

What is it that makes you feel it's impossible besides the fact his team didn't win the title?

ftr, I like the arguments being made for '15-16 LeBron, and I think '12-13 LeBron has a good case too, so I'm not necessarily looking to advocate for '08-09 LeBron, but I wouldn't dismiss it lightly.

there is a in-built bias towards title winning seasons, a natural dogmatic belief that a player is better in his late 20's or maybe early 30's than before,

than a somewhat older version of a player that is more skilled and experienced is better for winning than a younger version that is more dependant on superior athletism

a plauer greater stats in his mid 20's get diminisjed and weaker stats in their late 20's or early 30's handwaved away, usally because the latter are winners and the former are not, when often it may be simply that the team was weaker in that player mid 20's

88-89 jordan teams vs 91-93 jordan teams are a good example where player improvement may get the credit that should go instead to supporting cast improvement

mid 20's kareem winning in 71 but losing in 77 is a gold example of the opposite, 71 was not necesarrily better but it had much better teammstes)

this is specially important cause teams that get superstars usually do throufg the draft amd are bad teams, making title winning teams around your franchise player is hard and if done at all is likely to take a while so playera are mpre likely to win older than younger regardless of their level of impact

when the opposite happens you usually see players win more when younger (again think kareem in his early 20's vs late 20"s to very early 30's)

all of these are sometimes treated as a dogma

So, I think you make some great points, but feel like I need to push back on some points.

Stating up front that chip bias is going to be huge in this project to all of us, no matter how aware of the concept of 'winning bias' we are, and that you gave examples where there absolutely is a component of that:

- I think how one views pre-Phil Jordan with Phil-Jordan matters a lot here. The fact of the matter is that Jordan had to learn to adapt to play the style Phil wanted him to play, and to allow Pippen to come into his own. I value the MJ who has learned this.

- I think it's hard to know when Kareem peaked. I'll acknowledge that I believe his scoring game got more resilient after '71. My question for you would be: How do you see Kareem's loss of youth explosion playing in to his arc of defensive impact over his career? When does it become clear to you he can't possibly be at his peak any more?

Doctor MJ wrote:

ceiling raiser wrote:
So first, broadly: You better believe I take this data seriously when I get access to it. Getting access to the 76er dataset (thank to you I believe!) was eye-opening about Dr. J, and it has lowered his standing on my lists.

This isn't completely relevant *yet*(as noone has mentioned Dr. J for contention for GOAT peak), but has anyone tracked lineup/rotation data for the 70s-80s 76ers? I know the data for him looks relatively unimpressive for someone considered an all-timer(very inconsistent as well) - but those numbers could also be heavily influenced by that stuff and I would find it significantly more convincing if we had that to look at as well.

falcolombardi wrote:

Texas Chuck wrote:I'm not qualified for this project. But it is interesting to me that there has been zero mention of Russell. Which makes me worried that peak is once again being connected mostly to offensive numbers. But Russell absolutely belongs in any discussions of greatest peak. His impact data if we had it, would almost certainly be the best we'd ever seen at least based on how far ahead the Celtics defense was of the league as a whole.

I hope he and other elite defenders start to get some traction sooner than later and this doesn't just become about the offensive stars.

i actually dont feel completely qualified either but i am gonna be voting the picks i feel more confident on and the pkayers i feel more knowledgeable and sure about

Wanted to continue this thread of conversation.

I don't want to claim to know what all you feel about this project, but I'll share some of my thoughts on the matter:

It's easier to judge basketball careers, and even basketball seasons, than it is the quality of actual basketball play - particularly when we do so cross-era back to eras none of us were able to witness to the level we could claim to be experts.

When we do Peak conversations like we do, we're doing an obviously imperfect job and putting out limited understandings of basketball's subtle things potentially on full display.

It can really inhibit us - particularly if people are being confrontational.

But as with other projects we do on this board, I believe the process of thinking through something like this to the best of our ability is worthwhile for us and those who may read our thoughts and use them to develop their own ideas.

Additionally, I think we have to reckon with the fact that when it comes to fervent historical basketball analysis, the people we really want to defer to are all either a) dead or b) temporarily not dead. While we need to be able to laugh at ourselves in recognition of how much we know we'll always be missing, I think we also need to cut ourselves some slack as we recognize that there are very few communities out there that have learned as much about the history of the game as we have, and while it can be argued that we're silly for our ranking-based approach, if it means that stars of yesteryear gets discussed in ways they otherwise would not, I think there's something poignant about that.

This to say: I don't actually think being too knowledge-less should be a reason to DQ someone from the voter pool. The important thing is just that they are sincere in trying (researching and thinking), and trying to be sincere (positive, curious, etc).

Doctor MJ wrote:

ceiling raiser wrote:Question for people choosing MJ -

If we had data demonstrating that MJ at his best had a comparable defensive impact (in terms of DRAPM, or defense on/off the court, or other data) to Kobe, how would it impact your selection of his early 90s seasons as GOAT peak?

A great question to be asked.

So first, broadly: You better believe I take this data seriously when I get access to it. Getting access to the 76er dataset (thank to you I believe!) was eye-opening about Dr. J, and it has lowered his standing on my lists.

I will zag back in the other direction to emphasize that such data doesn't remove the end-accomplishments of the players involved. If the team just keeps winning when you're out there running the show all the way to the chip, well, that's your job. Beating the other team by as many points as possible, by contrast, is not the goal

But lastly:

I'm a bit surprised people (because someone else mentioned this possibility recently) are talking as if we have no data along these lines. It's not a question, for example, whether Jordan ever had a regular season +/- north of Kobe's best, because we have data from Jordan's last 2 years in Chicago, and that already settles the matter.

We also have access to regular season +/- data back through '93-94, and that gives us enough so that we already know that Jordan led his team in +/- at least as many times as Kobe did - Kobe did it 3 times in his career, Jordan did it in '95-96, '96-97 & '97-98.

So yeah, while it's within the realm of possibility that Kobe-like tendencies will result in +/- data underrating Jordan's total contribution, there's nothing like the "Whoa, Kobe really doesn't look that impactful by these metrics?" thing I detect for Jordan.

Great points! Using CORP +/- (not perfect but an easily accessible and stable metric with offensive/defensive splits), we find: peak defense MJ had +1.25 on defense, overall peak MJ had +1 on defense, peak defense Kobe had +1–0.5 on defense, and overall peak Kobe had +0.25 on defense. Dropping peak MJ from +1 on defense to +0.5 or +0.25 would be a loss of 0.5–0.75 points of value, which would take him beneath LeBron, Shaq, and possibly Hakeem into the next tier of players in that metric.

Also, re: the discussion on early RAPM estimates pre-97, it's certainly possible that there's systematic bias in the games saved and sampled, but at the bottom of each RAPM article they do include how the team performed in that sample relative to the season, which gives us at least a little more context.

For example, in the 1991 RAPM estimates (https://squared2020.com/2021/09/11/1990-1991-nba-rapm/), we have 38 regular season games where the Bulls were on pace to underperform their full-season results by ~3 games. That's a small difference, though it's hard to know how MJ's individual performance in these games compares to his full season performance. Perhaps someone could try finding some other estimate (BPM, win shares, whatever we have) for these games to compare to the full season? FWIW, the Celtics in 85 greatly underperformed in the games sampled for RAPM (they were on pace for 10 more loses in the sampled games vs their true regular season record), the Lakers in 85 also slightly underperformed, and the samples measured for superstars in 1970 so far are absolutely too small to take anything away from them.

DraymondGold wrote:

Doctor MJ wrote:

ceiling raiser wrote:Question for people choosing MJ -

If we had data demonstrating that MJ at his best had a comparable defensive impact (in terms of DRAPM, or defense on/off the court, or other data) to Kobe, how would it impact your selection of his early 90s seasons as GOAT peak?

A great question to be asked.

So first, broadly: You better believe I take this data seriously when I get access to it. Getting access to the 76er dataset (thank to you I believe!) was eye-opening about Dr. J, and it has lowered his standing on my lists.

I will zag back in the other direction to emphasize that such data doesn't remove the end-accomplishments of the players involved. If the team just keeps winning when you're out there running the show all the way to the chip, well, that's your job. Beating the other team by as many points as possible, by contrast, is not the goal

But lastly:

I'm a bit surprised people (because someone else mentioned this possibility recently) are talking as if we have no data along these lines. It's not a question, for example, whether Jordan ever had a regular season +/- north of Kobe's best, because we have data from Jordan's last 2 years in Chicago, and that already settles the matter.

We also have access to regular season +/- data back through '93-94, and that gives us enough so that we already know that Jordan led his team in +/- at least as many times as Kobe did - Kobe did it 3 times in his career, Jordan did it in '95-96, '96-97 & '97-98.

So yeah, while it's within the realm of possibility that Kobe-like tendencies will result in +/- data underrating Jordan's total contribution, there's nothing like the "Whoa, Kobe really doesn't look that impactful by these metrics?" thing I detect for Jordan.

Great points! Using CORP +/- (not perfect but an easily accessible and stable metric with offensive/defensive splits), we find: peak defense MJ had +1.25 on defense, overall peak MJ had +1 on defense, peak defense Kobe had +1–0.5 on defense, and overall peak Kobe had +0.25 on defense. Dropping peak MJ from +1 on defense to +0.5 or +0.25 would be a loss of 0.5–0.75 points of value, which would take him beneath LeBron, Shaq, and possibly Hakeem into the next tier of players in that metric.

Also, re: the discussion on early RAPM estimates pre-97, it's certainly possible that there's systematic bias in the games saved and sampled, but at the bottom of each RAPM article they do include how the team performed in that sample relative to the season, which gives us at least a little more context.

For example, in the 1991 RAPM estimates (https://squared2020.com/2021/09/11/1990-1991-nba-rapm/), we have 38 regular season games where the Bulls were on pace to underperform their full-season results by ~3 games. That's a small difference, though it's hard to know how MJ's individual performance in these games compares to his full season performance. Perhaps someone could try finding some other estimate (BPM, win shares, whatever we have) for these games to compare to the full season? FWIW, the Celtics in 85 greatly underperformed in the games sampled for RAPM (they were on pace for 10 more loses in the sampled games vs their true regular season record), the Lakers in 85 also slightly underperformed, and the samples measured for superstars in 1970 so far are absolutely too small to take anything away from them.

I think that guy tweeted out an updated 1991 RAPM estimate

Eddy_JukeZ wrote:

DraymondGold wrote:

Doctor MJ wrote:
A great question to be asked.

So first, broadly: You better believe I take this data seriously when I get access to it. Getting access to the 76er dataset (thank to you I believe!) was eye-opening about Dr. J, and it has lowered his standing on my lists.

I will zag back in the other direction to emphasize that such data doesn't remove the end-accomplishments of the players involved. If the team just keeps winning when you're out there running the show all the way to the chip, well, that's your job. Beating the other team by as many points as possible, by contrast, is not the goal

But lastly:

I'm a bit surprised people (because someone else mentioned this possibility recently) are talking as if we have no data along these lines. It's not a question, for example, whether Jordan ever had a regular season +/- north of Kobe's best, because we have data from Jordan's last 2 years in Chicago, and that already settles the matter.

We also have access to regular season +/- data back through '93-94, and that gives us enough so that we already know that Jordan led his team in +/- at least as many times as Kobe did - Kobe did it 3 times in his career, Jordan did it in '95-96, '96-97 & '97-98.

So yeah, while it's within the realm of possibility that Kobe-like tendencies will result in +/- data underrating Jordan's total contribution, there's nothing like the "Whoa, Kobe really doesn't look that impactful by these metrics?" thing I detect for Jordan.

Great points! Using CORP +/- (not perfect but an easily accessible and stable metric with offensive/defensive splits), we find: peak defense MJ had +1.25 on defense, overall peak MJ had +1 on defense, peak defense Kobe had +1–0.5 on defense, and overall peak Kobe had +0.25 on defense. Dropping peak MJ from +1 on defense to +0.5 or +0.25 would be a loss of 0.5–0.75 points of value, which would take him beneath LeBron, Shaq, and possibly Hakeem into the next tier of players in that metric.

Also, re: the discussion on early RAPM estimates pre-97, it's certainly possible that there's systematic bias in the games saved and sampled, but at the bottom of each RAPM article they do include how the team performed in that sample relative to the season, which gives us at least a little more context.

For example, in the 1991 RAPM estimates (https://squared2020.com/2021/09/11/1990-1991-nba-rapm/), we have 38 regular season games where the Bulls were on pace to underperform their full-season results by ~3 games. That's a small difference, though it's hard to know how MJ's individual performance in these games compares to his full season performance. Perhaps someone could try finding some other estimate (BPM, win shares, whatever we have) for these games to compare to the full season? FWIW, the Celtics in 85 greatly underperformed in the games sampled for RAPM (they were on pace for 10 more loses in the sampled games vs their true regular season record), the Lakers in 85 also slightly underperformed, and the samples measured for superstars in 1970 so far are absolutely too small to take anything away from them.

I think that guy tweeted out an updated 1991 RAPM estimate

Read on Twitter

Does he have the second sheet? I’d want to do the standard deviation thing with it

From a cursory look vs 2016 single season, since there isn’t a press sheet for it, that 2-50 are further from 0 than 2-50 in that other dataset, makes me thing that the std dev is probably a bit higher or comparable, so I’d wager that MJs RAPM in that regard here is about as impressive as the elite seasons we have on record but doesn’t stand out in that regard

Intuitively it’s about where I thought of him, GOAT tier offensively but more elite defender than DPOY type defensively

Doctor MJ wrote:

falcolombardi wrote:

Texas Chuck wrote:I'm not qualified for this project. But it is interesting to me that there has been zero mention of Russell. Which makes me worried that peak is once again being connected mostly to offensive numbers. But Russell absolutely belongs in any discussions of greatest peak. His impact data if we had it, would almost certainly be the best we'd ever seen at least based on how far ahead the Celtics defense was of the league as a whole.

I hope he and other elite defenders start to get some traction sooner than later and this doesn't just become about the offensive stars.

i actually dont feel completely qualified either but i am gonna be voting the picks i feel more confident on and the pkayers i feel more knowledgeable and sure about

Wanted to continue this thread of conversation.

I don't want to claim to know what all you feel about this project, but I'll share some of my thoughts on the matter:

It's easier to judge basketball careers, and even basketball seasons, than it is the quality of actual basketball play - particularly when we do so cross-era back to eras none of us were able to witness to the level we could claim to be experts.

When we do Peak conversations like we do, we're doing an obviously imperfect job and putting out limited understandings of basketball's subtle things potentially on full display.

It can really inhibit us - particularly if people are being confrontational.

But as with other projects we do on this board, I believe the process of thinking through something like this to the best of our ability is worthwhile for us and those who may read our thoughts and use them to develop their own ideas.

Additionally, I think we have to reckon with the fact that when it comes to fervent historical basketball analysis, the people we really want to defer to are all either a) dead or b) temporarily not dead. While we need to be able to laugh at ourselves in recognition of how much we know we'll always be missing, I think we also need to cut ourselves some slack as we recognize that there are very few communities out there that have learned as much about the history of the game as we have, and while it can be argued that we're silly for our ranking-based approach, if it means that stars of yesteryear gets discussed in ways they otherwise would not, I think there's something poignant about that.

This to say: I don't actually think being too knowledge-less should be a reason to DQ someone from the voter pool. The important thing is just that they are sincere in trying (researching and thinking), and trying to be sincere (positive, curious, etc).

I didn’t realize it was a qualification thing lol

I feel there’s been less discussion than like the year I joined (2015 I think lol) vs more kind of place a ballot and dip kind of thing now, but maybe I’m just tripping

It feels like most of it early on was which lebron year is his peak lol

On a side note, I do think there’s a bit of a kind of, general consensuses on the site that I think are a result of just the general discussion that has been had, which of course is normal but a large deviation feels weird to say in projects like these sometimes (I think people would barf on my thoughts on Kawhi lol)