Updating my top 50

[70sFan]

...

After thinking about it a bit longer, I think I will change my ranking for 1969 Russell to MVP level. I also think about 1971-72 Kareem and 2016-17 James, but I think for now I'll keep them at all-time tier.

[krii]

I said about to approximate, there is no exact numbers on any of these, but you need some approximate numbers to do math.
All-Star might be 23 or 29, but calling it an all-star (or all-nba) gives you ranges.

I still disagree with the above. In the league of the late 50s/early 60s there were basically league-average [or scarcely better] players being selected as All-Stars, simply because they needed 20 or so guys to fill the All-Star rosters......but they ran out of guys truly deserving of such distinction after about 10-12 or so.

That guy who's 23rd-best in 1960 [even in era-relative terms] is a comically lesser player than the 29th-best player of 2023.

Consequently, if/when I get far enough down my CORP-list that I run into guys who were 20th-24th in the league circa-1960, I'm not giving them credit for an "All-Star level" season for those years simply because they were top 24.

You're free to do so, of course. But to me, rating a clearly worse player better [than the other player he's worse than] is kinda bonkers.

It's worthwhile looking at the degree of parity in the league in given years (i.e. don't take raw metrics at face value, as though the numbers mean exactly the same thing from decade to decade), sure. But I'm just not going to take that as far as settling on a number of persons (give or take a few) in the league, when the league is nearly 4x larger today.

My point is that they are still at a somewhat non scarce level. There are 60-90 or so Thompsons, Armstrongs etc.in the league. Virtually every team has a couple.

So?
Again, I think you're too hung up on proportion of the league. Being relatively common in the league does not = unvaluable.

And fwiw, some CORP methodologies have slightly different values for the various tiers, if that makes you feel better about it. For example, I read one where a "GOAT-level" season was given 40% CORP, while "Average [Role] player" seasons only 1.5% (instead of the 33% and 2% cited by 70sFan).

Personally, I'm going with this range (which is in between): 35% [GOAT]/29% [All-Time Great]/21% [MVP]/16% [weak MVP]/10% [All-NBA]/6.5% [All-Star]/4% [Sub All-Star]/1.5% [Average player]

Yes, Tristan Thompson and B.J. Armstrong help a team win a title......

......Being as good as B.J. doesnt mean you have a better chance to win a title than being as good as Jud Buechler. The key to winning NBA titles is really to have guys who are better than that.

This seems like, "Yes, they help.....except that they don't help."

OBVIOUSLY you need guys [multiple guys] better than "Average" on your team to have any legit shot at a title. Precisely NO ONE is suggesting otherwise. The principle of CORP% analysis is absolutely NOT suggesting otherwise.

It does, however, recognize that even GOAT-tier players need help; look at '89 and '90 Jordan, '77 Kareem, '09 LeBron, '64 Wilt. All of those are at least "All-Time Great" seasons, if not even "GOAT-level", yet NONE of them won the title ['64 Wilt is the ONLY one among them who even made it to the Finals, losing soundly].

Having Armstrong [or Harper] instead of Hudson, Smith instead of Rivers, Thompson instead of Jerebko, or Love instead of West, etc: either you agree these things make a difference, maybe even enough to turn those particular series's in a different direction........or you don't.

If you DO agree they do, then that is acknowledging the core of the CORP analysis view of average players. Because again: MOST of the league [in ANY year] is replacement level or worse. So even though there may be, as you say, 60-90 "Average" players in the league, they're still more scarce than the guys who are WORSE (if you insist on keeping the focus on scarcity).

Just out of curiosity, as I'm not sure if a similar system has already been proposed, wouldn't this problem be solved by estimating the level/power of the league's specific season in a numerical value, e.g. on a scale of 0.00-1.00? Each season would have to be given an appropriate valorisation relative to the other seasons. Many factors would have to be taken into account that is usually used for player evaluation and can be quantified in some way - what was the offensive/defensive level of the season, what model of play prevailed and what type of players this system favoured and benefited*, what was the average pace, what was the relative scale of talent (e.g. volume of players with corresponding levels per season - e.g. how many players have MVP/subpar MVP levels in a given season, compared to other seasons – think, how many great all-time runs you had in 2015 or 1978). Then the seasons would have to be compared to each other to be sure that, for example, the 1996/97 season compared to the 2011/12 season had such and such value (e.g. 0.76-0.68).

Once the strength of the season had been worked out, the player's rating would not only relate to the season's rating in the context of the entire history of the league (i.e. as the rating now takes into account) but also in comparison to the specific 'strength' rating of the season. The season 'strength' indicator would not necessarily have the same weight as the overall-history rating but would be relevant in comparing the achievements of players from different eras.

* - that could be used for yet another comparison in which players playing in different eras might be compared more thoroughly. If we have a player like Stockton, who is proving on low volume that he can be an effective three-point shooter, but focuses on pure playmaking rather than throwing - is this a result of the dominant model of play in the league, and if so, does this affect the evaluation of his play in space usage, and volume scoring; compared to, say, the 2010s and 2020s, when the premise of the playmaker is basically a dominant 3p shooting and perimeter space-passing?

@70sFan – have you considered this?

[70sFan]

I still disagree with the above. In the league of the late 50s/early 60s there were basically league-average [or scarcely better] players being selected as All-Stars, simply because they needed 20 or so guys to fill the All-Star rosters......but they ran out of guys truly deserving of such distinction after about 10-12 or so.

That guy who's 23rd-best in 1960 [even in era-relative terms] is a comically lesser player than the 29th-best player of 2023.

Consequently, if/when I get far enough down my CORP-list that I run into guys who were 20th-24th in the league circa-1960, I'm not giving them credit for an "All-Star level" season for those years simply because they were top 24.

You're free to do so, of course. But to me, rating a clearly worse player better [than the other player he's worse than] is kinda bonkers.

It's worthwhile looking at the degree of parity in the league in given years (i.e. don't take raw metrics at face value, as though the numbers mean exactly the same thing from decade to decade), sure. But I'm just not going to take that as far as settling on a number of persons (give or take a few) in the league, when the league is nearly 4x larger today.




So?
Again, I think you're too hung up on proportion of the league. Being relatively common in the league does not = unvaluable.

And fwiw, some CORP methodologies have slightly different values for the various tiers, if that makes you feel better about it. For example, I read one where a "GOAT-level" season was given 40% CORP, while "Average [Role] player" seasons only 1.5% (instead of the 33% and 2% cited by 70sFan).

Personally, I'm going with this range (which is in between): 35% [GOAT]/29% [All-Time Great]/21% [MVP]/16% [weak MVP]/10% [All-NBA]/6.5% [All-Star]/4% [Sub All-Star]/1.5% [Average player]




This seems like, "Yes, they help.....except that they don't help."

OBVIOUSLY you need guys [multiple guys] better than "Average" on your team to have any legit shot at a title. Precisely NO ONE is suggesting otherwise. The principle of CORP% analysis is absolutely NOT suggesting otherwise.

It does, however, recognize that even GOAT-tier players need help; look at '89 and '90 Jordan, '77 Kareem, '09 LeBron, '64 Wilt. All of those are at least "All-Time Great" seasons, if not even "GOAT-level", yet NONE of them won the title ['64 Wilt is the ONLY one among them who even made it to the Finals, losing soundly].

Having Armstrong [or Harper] instead of Hudson, Smith instead of Rivers, Thompson instead of Jerebko, or Love instead of West, etc: either you agree these things make a difference, maybe even enough to turn those particular series's in a different direction........or you don't.

If you DO agree they do, then that is acknowledging the core of the CORP analysis view of average players. Because again: MOST of the league [in ANY year] is replacement level or worse. So even though there may be, as you say, 60-90 "Average" players in the league, they're still more scarce than the guys who are WORSE (if you insist on keeping the focus on scarcity).

Just out of curiosity, as I'm not sure if a similar system has already been proposed, wouldn't this problem be solved by estimating the level/power of the league's specific season in a numerical value, e.g. on a scale of 0.00-1.00? Each season would have to be given an appropriate valorisation relative to the other seasons. Many factors would have to be taken into account that is usually used for player evaluation and can be quantified in some way - what was the offensive/defensive level of the season, what model of play prevailed and what type of players this system favoured and benefited*, what was the average pace, what was the relative scale of talent (e.g. volume of players with corresponding levels per season - e.g. how many players have MVP/subpar MVP levels in a given season, compared to other seasons – think, how many great all-time runs you had in 2015 or 1978). Then the seasons would have to be compared to each other to be sure that, for example, the 1996/97 season compared to the 2011/12 season had such and such value (e.g. 0.76-0.68).

Once the strength of the season had been worked out, the player's rating would not only relate to the season's rating in the context of the entire history of the league (i.e. as the rating now takes into account) but also in comparison to the specific 'strength' rating of the season. The season 'strength' indicator would not necessarily have the same weight as the overall-history rating but would be relevant in comparing the achievements of players from different eras.

* - that could be used for yet another comparison in which players playing in different eras might be compared more thoroughly. If we have a player like Stockton, who is proving on low volume that he can be an effective three-point shooter, but focuses on pure playmaking rather than throwing - is this a result of the dominant model of play in the league, and if so, does this affect the evaluation of his play in space usage, and volume scoring; compared to, say, the 2010s and 2020s, when the premise of the playmaker is basically a dominant 3p shooting and perimeter space-passing?

@70sFan – have you considered this?

If I had the infinite time for basketball analysis, I'd do that but unfortunately it's impossible to me to do such a huge project.

[krii]

Just out of curiosity, as I'm not sure if a similar system has already been proposed, wouldn't this problem be solved by estimating the level/power of the league's specific season in a numerical value, e.g. on a scale of 0.00-1.00? Each season would have to be given an appropriate valorisation relative to the other seasons. Many factors would have to be taken into account that is usually used for player evaluation and can be quantified in some way - what was the offensive/defensive level of the season, what model of play prevailed and what type of players this system favoured and benefited*, what was the average pace, what was the relative scale of talent (e.g. volume of players with corresponding levels per season - e.g. how many players have MVP/subpar MVP levels in a given season, compared to other seasons – think, how many great all-time runs you had in 2015 or 1978). Then the seasons would have to be compared to each other to be sure that, for example, the 1996/97 season compared to the 2011/12 season had such and such value (e.g. 0.76-0.68).

Once the strength of the season had been worked out, the player's rating would not only relate to the season's rating in the context of the entire history of the league (i.e. as the rating now takes into account) but also in comparison to the specific 'strength' rating of the season. The season 'strength' indicator would not necessarily have the same weight as the overall-history rating but would be relevant in comparing the achievements of players from different eras.

* - that could be used for yet another comparison in which players playing in different eras might be compared more thoroughly. If we have a player like Stockton, who is proving on low volume that he can be an effective three-point shooter, but focuses on pure playmaking rather than throwing - is this a result of the dominant model of play in the league, and if so, does this affect the evaluation of his play in space usage, and volume scoring; compared to, say, the 2010s and 2020s, when the premise of the playmaker is basically a dominant 3p shooting and perimeter space-passing?

@70sFan – have you considered this?

If I had the infinite time for basketball analysis, I'd do that but unfortunately it's impossible to me to do such a huge project.

I believe I could prompt the AI to do most of the job. Bearing in mind the variables and how they affect the season rating, the AI should adjust the rating of individual years (e.g. bearing in mind the changing nature of the game in specific years, caused by new rules). With such a list, this could be run through the forum, testing whether the community's rating is aligning with the final prompt, and then running it. After doing so, it can be applied to your list at the level of a simple addition to your own formula.

Correct me if you feel it's too much

[OhayoKD]

Thread with Russell's defensive footage sounds good to me

If you're following the evidence, then you shouldn't worry about overrating who you're evaluating. The Lakers went 8-1(+10 and +17) against non-kareem opposition in round 2 and the finals of the playoffs. Taking away "inflation" with san's standard deviation,the 72 Lakers and the 71 and 72 Bucks still score higher than the 2000 Lakers. Kareem outscored the first team I just listed and led the two other teams, notably doing the first bit with oscar banged up, and having the lakers playing like a champion in games oscar missed entirely. Can you construct good positive arguments against 72 for all of the non-kareem "peaks" you currently have a tier higher? Can you construct one from surrounding years from those peaks?

I think I can, but you're giving me solid arguments for reconsidering 1972 case for GOAT-level peak. This one, along with 2016 and 2017 James, are the ones that gave me the most trouble to choose. These cutoffs are always a bit arbitrary and the change wouldn't make any difference in the final ranking, for what it's worth.

Would be curious to see what a positive case against 72 looks like. Honestly haven't seen much of one against 71 for players people rank higher or the same(shaq, jordan, ect) and we can scale 72 directly off since he's basically just a better version.

Also against 2016 Lebron or 2015-2017 as a three-year stretch. Per lineup-adjustment(believe you've asked about this before) 2015-2017 Lebron lines right up with rs Steph per nbashotchart's 3-year RAPM. Using Ahmed's non-box informed 5-year set, 2013-2017 Lebron actually stands clear of any non-lebron stretch per-possession and is an even bigger-outlier accounting for #of possessions:
https://public.tableau.com/app/profile/ahmed.cheema8618/viz/FiveYearRAPMPeaks1997-2021/FiveYearRAPMPeaks1997-2021. Using Ben's scaled APM, all-three of 2015, 2016, and 2017 score at the very top and as I'm sure you're aware, whether you go by lineup-splits, WOWY, indirect, or extended WOWY, raw-stuff paints those three-regular seasons as top-tier(notably this remains true if you go with kyrie/love+lebron or no kyrie/love + lebron).

The Cavs then get substantially better in the playoffs(15-17), lebron-specific lineups improve(with and without his co-stars), and the cavs do better in the later rounds than the early ones and knock off an opponent that posts a +10.8 PSRS despite Curry missing a big chunk of that postseason. By the numbers, those rs's are at or above various seasons you have in your goat-tier and then there is big playoff-improvement. Even 2015 looks up there holistically if you can get past the scoring numbers.

And then there's replication where Lebron and Kareem are just completely unrivalled and therefore probably deserving of "benefit of the doubt" in a way other "goat-tier" peaks probably aren't(you might note Lebron owns all 3 of the dots standing further out from the rest in the graph I linked above).

The other guy who has never really failed on that front(though he's had limited opportunity to show) is Russell.

(I'll address "skillset" below)

Kareem obviously still has a big defensive advantage against people like peak shaq and such, and obviously has a big offensive advantage over someone like peak wilt. How does Kareem being limited offensively relative to himself matter for how he compares to a Shaq, a Wilt, a Jordan, or a Lebron?

Internal-scaling inofitself establishes nothing about how a player compares to everyone else. We currently have no real way to derive an upper-limit to player goodness or an upper-limit to what the gaps between great players can be situationally let alone what the gaps are across a variety of situations. 72 Kareem<74 Kareem does not establish anything about 72 Kareem and 2000 Shaq. Just like 2016 Lebron<2013 Lebron does not establish anything about 2016 Lebron and 1989 MJ.

If you're not establishing comparisons between different players directly, what good does looking at internal-scaling achieve? In most cases players have certain advantages and certain disadvantages in terms of skill-set. Player A protects the rim, player b protects the rim less but helps more. Player A scores and creates worse but player b puts opposing defenses in foul-trouble. Player A scores more and creates as much, Player B creates more efficiently and is a handful of players who can act as an on-court coach.

In any of these cases, player A can maintain an overall advantage over player b even if they've regressed from being an outright match for player b in a specific skill or attribute or set of attributes/skills. If a player still looks as or more valuable when they have yet to acquire certain skills, or have lost another skill(worth noting the skills people tend to fixate-on are scoring-related), the fact they had yet to acquire or have lost something may well just be outweighed by their pre-existing advantages(88,92/93 hakeem, 2015 Lebron, 69 Russell, and 05 Duncan come to mind)

Kareem can be less resilient offensively than he was in 74 and still be better than numerous goat-tier peaks because he can make defenses 4-points better/league-best and Shaq, Jordan, Magic, Bird have never come close(2001 is the only postseason of any of these guys where you can argue they anchored a good defense).

Why do you think that I don't do comparisons across different players? I responded to DQuinn1575's post which clearly was about the comparison of younger vs older Kareem. It doesn't mean I don't do comparisons between different players directly, it's impossible to create such list without it.

Because I don't see it much. At least not between the highest all-time peaks/primes. I haven't seen positive comparisons made between Kareem and Lebron, Kareem and Jordan, Kareem and Shaq, Lebron and Jordan, Lebron and Shaq, Hakeem and Jordan, Shaq and Duncan, Shaq and Hakeem, ect. I do see you compare players to themselves or make claims about x gap being bigger than y gap, or x player being more consistent than y player, without much reasoning.

This is not a novel point, but I have not seen it addressed once: Unless you are not valuing era-relativity highly, putting 69 Russell at "weak mvp" makes negative sense. Even using RK's +5 assessment(derived by using noisy playoff d-rating and ignoring the larger sample of regular-season d-rating or the Celtic's overall shift in 70(which would place Russ at +7), Being worth 5-points of SRS in 1969 =/ Weak MVP.

It's another tough season to rate, but when I evaluate given year, I always look at the combination of impact, skillset and RS+PS performance. In Russell's case, it was clear that he had an immense impact on his team. The thing is that impact can be situational, in some cases players show impact signals that underrate their value (see 2011 James for example), sometimes it's the opposite.

Sure.

I think that Russell's defensive pressence was enormous. I have got quite a lot of footage from 1968/69 season in recent years and he looks as good as ever on defensive end there (if you'd like to, we can create a new thread where I will upload videos). At the same time though, I also see how his offensive game deteriorated even from 1967. At this point he was a legit negative on offense, though to be fair he doesn't look nearly as bad when he didn't have to face Wilt on the other end (he looks quite decent against Reed and the Knicks). With that in mind, I struggle to put him higher than MVP tier, because of that weakness. Maybe I should put him at the MVP level, but anything beyond that is tough to sell for me, considering the games I have seen.

But why does he need to be a positive or neutral offensive player to be "anything beyond" mvp in 1969? The Celtics were the most dominant team ever on the strength of their defense which you say for Russell was still "as good as ever". They were not a great offense in 69(-1.7, -6.4 in the rs, +2, -5 in the playoffs). Skill-set analysis is fine, but when you're comparing different-types of players in different-types of situations/contexts, we still need some sort of basis to weigh one attribute to another. The Celtics post a league-best srs based entirely on their defense. They beat a gauntlet pre-dominantly on their defense. And we still aren't factoring in what Russell offers as a coach who is telling his teammates where to go and what to do.

I've made this point with Kareem but it applies to the others you're evaluating too.

Where do you register that Russell's combination of defense and iq, which was always the vast majority of the most successful formula in American Sports, is not enough to be the best now when he still gets the job done against unusually difficult opposition?

Where do you register that Lebron's myriad non-box advantages over virtually any non-big ever(paint-deterrence, inside-scoring, gravity, rebounding) or virtually any player ever(passing, Creation-quality, defensive play-calling, offensive play-calling, pace-control, help), don't make up for lesser rs box-score production vs "goat-tiers" when the data would suggest it does(in the RS)?

Skill-set analysis is probably off if you're putting significant weight on offense for russell or comparing Lebron to "goat-tier" players simply via raw scoring/creation.

If anyone feels differently, I would invite them to list 5 examples of players they think have ever, at any point, been worth more on their own, than the gap between average and the highest-srs mark in the league.

I mean, you can find such examples. 1982 Moses Malone for example - the Rockets went from -0.4 SRS to -11.1 SRS with his absence (and they got a solid replacement in Caldwell Jones, unlike Russell). I think you are against putting Moses anywhere near all-time tier, right?

Well first, keep in mind we're not actually using the full-drop off from 1969-1970(7-points), but just the 5-point guess Rk made(undercutting Russ's signal by 2-points of srs). Russell was replaced by a bad center in 1970. Is that worth 5-10 wins over Jordan replacing a bad but not quite as bad shooting guard in 1984?

Secondly, you're using a 1-game sample 70's. One that is contradicted by all the larger ones:

Malone missed 23 games that year, and without him the “healthy” Rockets played at a 21-win pace (-7.3 SRS), but with him only a 29-win pace (-4.1 SRS).6 Although he was just 22 and rapidly developing, that’s a less-than-desirable result for any star-level player.

When Moses arrived in Philadelphia in 1983, he joined an upper-crust club that had reached the Finals two of the three prior seasons. The 76ers crushed the league that year, finishing with a defense 3.8 points better than league average and clocking along at a 64-win pace at full-strength (8.8 SRS). But that excellence evaporated in 1984, despite no notable roster changes and a core with five players between 26 and 28 years old. There were grumblings of disappointment from ownership about a lack of effort and the Sixers sputtered to a 52-win pace (3.7 SRS).8

In ’85, Philly bounced back, playing at a 58-win pace (6.0 SRS) with the addition of rookie Charles Barkley. But again, the team regressed in ’86 (50-win pace, or 3.1 SRS) under new coach Matt Guokas as Moses and Erving aged and scoring dynamo Andrew Toney missed most of the season with stress fractures in his feet. Malone was traded to Washington for the ’87 season, where an overhauled Bullets squad played at a pace nearly identical to their ’86 team (38-win pace or -1.0 SRS). Meanwhile, back in Philadelphia, with Moses and Bobby Jones gone and Erving in his final year, the 76ers treaded water as an average team (0.8 SRS when healthy).

We also have..

Fortunately, we have Harvey Pollack’s plus-minus data for Moses’ four seasons in Philadelphia to help evaluate his impact. His AuPM oscillates between strong (but not transcendent) and pedestrian in those four seasons, with ’83 and ’85 looking like typical top-20 seasons and ’84 and ’86 lacking impact. Similarly, Moses’ regressed game-level data tells us that he made a difference, but that his impact was far short of a Grade-A superstar’s.

In comparison, a 7-point evaluation would be coming off the largest possible sample of off. It also is corraborated by the 7-game sample in 1969, Russell's career WOWY, and the Celtics mostly staying league-best when non-Russell players missed time. The only "contradiction" comes if we take his rookie-year when Russell had a superteam with a bunch of key teammates that Russell would win 5 rings without after they'd all left or fallen off. But if we're going that far, we can also reference him winning 2 rings with a team that didn't make the ncaa without him, or him setting the record for olympic point-differential. And crucially, there's the bit where he never lost when healthy. Why would we take 1969 Russell's impact to be a matter of noise?. Like I said, we're not using the 7-point drop that actually occurred, so who Russell was replaced by shouldn't really matter.

We've both agreed that replication/corroboration matters. Russell has never lost when healthy, and all the signals with relevance noisy they may be support what the largest possible sample says. We can acknowledge uncertainty but uncertainty itself is not a good reason for claiming a player is better or worse than another. Ditto with Lebron and Kareem. These are 3 players who from what we have were posting outlier signals relative to the field from teen-aged to their 30's and beyond. They also happen to have a whole collection "goat-lvl" or "outlier" looking prime/peak stuff, even in down-years(2015, 1975). Why would we take their signals to be noise, when it's exactly what we should expect given where they started and ended(or in Lebron's case ending) and they've posted "goat-level" or better impact again and again?

[70sFan]

Thread with Russell's defensive footage sounds good to me

I will start the thread in a moment.

Would be curious to see what a positive case against 72 looks like. Honestly haven't seen much of one against 71 for players people rank higher or the same(shaq, jordan, ect) and we can scale 72 directly off since he's basically just a better version.

I have some reservations about his ability to anchor offense at the top level, but to be honest 1972 Kareem is probably the strongest all-time level season I can imagine. I think it's reasonable to put him at the top tier. It's just that the Bucks offense underperformed a little bit in the playoffs in 1971-73 period that gives me some reservations about taking him at the same level as the best peaks ever.

I just realized that I ranked 2001 Shaq as the GOAT tier season (probably too high), but I am confident with taking 1971 and 1972 Kareem over 2001 Shaq, so it seems that I should swap these seasons.

Also against 2016 Lebron or 2015-2017 as a three-year stretch. Per lineup-adjustment(believe you've asked about this before) 2015-2017 Lebron lines right up with rs Steph per nbashotchart's 3-year RAPM. Using Ahmed's non-box informed 5-year set, 2013-2017 Lebron actually stands clear of any non-lebron stretch per-possession and is an even bigger-outlier accounting for #of possessions:
https://public.tableau.com/app/profile/ahmed.cheema8618/viz/FiveYearRAPMPeaks1997-2021/FiveYearRAPMPeaks1997-2021. Using Ben's scaled APM, all-three of 2015, 2016, and 2017 score at the very top and as I'm sure you're aware, whether you go by lineup-splits, WOWY, indirect, or extended WOWY, raw-stuff paints those three-regular seasons as top-tier(notably this remains true if you go with kyrie/love+lebron or no kyrie/love + lebron).

The Cavs then get substantially better in the playoffs(15-17), lebron-specific lineups improve(with and without his co-stars), and the cavs do better in the later rounds than the early ones and knock off an opponent that posts a +10.8 PSRS despite Curry missing a big chunk of that postseason. By the numbers, those rs's are at or above various seasons you have in your goat-tier and then there is big playoff-improvement. Even 2015 looks up there holistically if you can get past the scoring numbers.

And then there's replication where Lebron and Kareem are just completely unrivalled and therefore probably deserving of "benefit of the doubt" in a way other "goat-tier" peaks probably aren't(you might note Lebron owns all 3 of the dots standing further out from the rest in the graph I linked above).

The other guy who has never really failed on that front(though he's had limited opportunity to show) is Russell.

I have my thoughts about 2015 LeBron, but I don't want to start another James debate. I just want to mention that I made a few minor changes in the top 10:

- 2016 James up to "GOAT tier",
- 1971 and 1972 Kareem up to "GOAT tier",
- 1969 Russell up to "MVP tier",
- 1987 Jordan up to "MVP tier",
- 2001 Shaq down to "all-time tier".

It didn't change the order, but it changed the overall career value estimations a little bit.

Because I don't see it much. At least not between the highest all-time peaks/primes. I haven't seen positive comparisons made between Kareem and Lebron, Kareem and Jordan, Kareem and Shaq, Lebron and Jordan, Lebron and Shaq, Hakeem and Jordan, Shaq and Duncan, Shaq and Hakeem, ect. I do see you compare players to themselves or make claims about x gap being bigger than y gap, or x player being more consistent than y player, without much reasoning.

Well, I can't compare everyone to everyone. It's time consuming to create such list and to keep everything in mind, I don't have (and won't have) the time to compare every season between all top 10 players. If you want to search for my reasonings, you can read this project for a start:

viewtopic.php?f=64&t=2120549

But why does he need to be a positive or neutral offensive player to be "anything beyond" mvp in 1969?

Because I don't evaluate players strictly relative to one season, but to the overall era they played in. 1969 Russell wasn't a better player than younger Russell or peak Wilt.

The Celtics were the most dominant team ever on the strength of their defense which you say for Russell was still "as good as ever".

Yeah, but his offense hurt Celtics and Celtics defense didn't do nearly as well in the playoffs (though maybe it's a noise of estimations).

They were not a great offense in 69(-1.7, -6.4 in the rs, +2, -5 in the playoffs).

Ben Taylor's estimates paints a different picture:

1969 RS: +0.1 rORtg, -4.6 rDRtg
1969 PS: +6.4 rORtg, -0.6 rDRtg

Do I trust these estimations? I'm not really sure, I think I'd have to make some calculations with Ben's database because sometimes I find out that the results for the early years don't add up.

Skill-set analysis is fine, but when you're comparing different-types of players in different-types of situations/contexts, we still need some sort of basis to weigh one attribute to another. The Celtics post a league-best srs based entirely on their defense.

2nd best to be precise.

They beat a gauntlet pre-dominantly on their defense.

That's my intuition, but Ben's data disagrees.

And we still aren't factoring in what Russell offers as a coach who is telling his teammates where to go and what to do.

Yes, but I try not to give any additional credit for that in this project.

Well first, keep in mind we're not actually using the full-drop off from 1969-1970(7-points), but just the 5-point guess Rk made(undercutting Russ's signal by 2-points of srs). Russell was replaced by a bad center in 1970. Is that worth 5-10 wins over Jordan replacing a bad but not quite as bad shooting guard in 1984?

I don't think I mentioned rookie Jordan, rookie Jordan is definitely worse player than 1969 Russell in my evaluation, he's now two tiers below 1969 Russell.

Secondly, you're using a 1-game sample 70's. One that is contradicted by all the larger ones:

No one year, but two years sample identical to 1969-70 Russell.

Malone missed 23 games that year, and without him the “healthy” Rockets played at a 21-win pace (-7.3 SRS), but with him only a 29-win pace (-4.1 SRS).6 Although he was just 22 and rapidly developing, that’s a less-than-desirable result for any star-level player.

That's about 1978 Moses who was 22 years old. I don't think we can use 1957 sample for 1969 Russell, right?

When Moses arrived in Philadelphia in 1983, he joined an upper-crust club that had reached the Finals two of the three prior seasons. The 76ers crushed the league that year, finishing with a defense 3.8 points better than league average and clocking along at a 64-win pace at full-strength (8.8 SRS). But that excellence evaporated in 1984, despite no notable roster changes and a core with five players between 26 and 28 years old. There were grumblings of disappointment from ownership about a lack of effort and the Sixers sputtered to a 52-win pace (3.7 SRS).8

What does it have to do with 1982 Rockets? Malone dealt with injuries in 1983/84 season.

Not to mention that Moses missed 10 games in that season and the Sixers went 5-6 without him, far from an all-time great supporting cast.

In ’85, Philly bounced back, playing at a 58-win pace (6.0 SRS) with the addition of rookie Charles Barkley. But again, the team regressed in ’86 (50-win pace, or 3.1 SRS) under new coach Matt Guokas as Moses and Erving aged and scoring dynamo Andrew Toney missed most of the season with stress fractures in his feet. Malone was traded to Washington for the ’87 season, where an overhauled Bullets squad played at a pace nearly identical to their ’86 team (38-win pace or -1.0 SRS). Meanwhile, back in Philadelphia, with Moses and Bobby Jones gone and Erving in his final year, the 76ers treaded water as an average team (0.8 SRS when healthy).

Again, no idea why we're talking about the yeat 5 years after the situation I am describing, but fair.
I have no idea why Ben made such a misleading quote about the Bullets situation:

1. Malone missed 10 games in 1986/87 season and the Bullets went 4-5 without him (36 wins team) vs 38-35 with him (43 wins team).
2. That alone doesn't explain anything, but Ben also doesn't adjust for the fact that the Bullets roster was completely different in 1986 and 1987:

1986 Bullets top 5 players in minutes:

Jeff Malone
Cliff Robinson
Dan Roundfield
Gus Williams
Manute Bol

1987 Bullets top 5 players in minutes:

Jeff Malone
Moses Malone
Terry Catlege
Ennis Whatley
John Williams
(Bol was still on the team)

Outside of Jeff Malone and Manute Bol, the Bullets looks completely different. They got Moses in trade, but they also lost Cliff Robinson, Dan Roundfield and old Gus Williams, replacing them with scrubs.

Fortunately, we have Harvey Pollack’s plus-minus data for Moses’ four seasons in Philadelphia to help evaluate his impact. His AuPM oscillates between strong (but not transcendent) and pedestrian in those four seasons, with ’83 and ’85 looking like typical top-20 seasons and ’84 and ’86 lacking impact. Similarly, Moses’ regressed game-level data tells us that he made a difference, but that his impact was far short of a Grade-A superstar’s.

Moses plus-minus data looks phenomenal in 1983 and 1985, I have no idea why Ben uses such language. I am afraid Ben wanted to show in his article why he's so low on 3 times MVP and it came out a bit too negative in my opinion.

In comparison, a 7-point evaluation would be coming off the largest possible sample of off.

Just like with 1982->1983 Moses situation.

It also is corraborated by the 7-game sample in 1969

Moses impact is also visible in 1983/84 season.

Russell's career WOWY, and the Celtics mostly staying league-best when non-Russell players missed time.

You don't need to convince me that Russell is a better player than Moses, my point was that one sample across two seasons can show various results and Moses 1982->1983 sample is at least comparable to Russell's 1969->1970.

The only "contradiction" comes if we take his rookie-year when Russell had a superteam with a bunch of key teammates that Russell would win 5 rings without after they'd all left or fallen off. But if we're going that far, we can also reference him winning 2 rings with a team that didn't make the ncaa without him, or him setting the record for olympic point-differential. And crucially, there's the bit where he never lost when healthy. Why would we take 1969 Russell's impact to be a matter of noise?. Like I said, we're not using the 7-point drop that actually occurred, so who Russell was replaced by shouldn't really matter.

I have never said that Russell's impact was a noise. Again, that's a baseless assumption.
I only stated that the Celtics were very reliant on his defense and in 1970 they didn't replace him with even average center, but they played with no reliable center all season long. It means that Russell's impact was absurdly high for them, but it also means that it could be a bit lower in different situation.

We've both agreed that replication/corroboration matters. Russell has never lost when healthy, and all the signals with relevance noisy they may be support what the largest possible sample says. We can acknowledge uncertainty but uncertainty itself is not a good reason for claiming a player is better or worse than another. Ditto with Lebron and Kareem. These are 3 players who from what we have were posting outlier signals relative to the field from teen-aged to their 30's and beyond. They also happen to have a whole collection "goat-lvl" or "outlier" looking prime/peak stuff, even in down-years(2015, 1975). Why would we take their signals to be noise, when it's exactly what we should expect given where they started and ended(or in Lebron's case ending) and they've posted "goat-level" or better impact again and again?

You almost talk about it like these two would be out of my top 10, while they are the clear 3 outliers at the top, with Jordan/Duncan/Hakeem being far behind Russell...

[AEnigma]

They were not a great offense in 69(-1.7, -6.4 in the rs, +2, -5 in the playoffs).

Ben Taylor's estimates paints a different picture:

1969 RS: +0.1 rORtg, -4.6 rDRtg
1969 PS: +6.4 rORtg, -0.6 rDRtg

Do I trust these estimations? I'm not really sure, I think I'd have to make some calculations with Ben's database because sometimes I find out that the results for the early years don't add up.

They beat a gauntlet pre-dominantly on their defense.

That's my intuition, but Ben's data disagrees.

Think we may have talked about this recently (do not care enough to search it out), but just on its face those ratings seem untenable. Until some legitimate justification is shown, I would trust common sense over what is posted on a stat-site.

Keeping in mind that the Celtics held teams to 105.4 points per game in the regular season…
76ers: down from 118.9 points per game in the regular season to 106 points per game against the Celtics
Knicks: down from 110.8 (112.4 with Debusschere) points per game in the regular season to 101.8 points per game against the Celtics
Lakers: down from 112.2 points per game in the regular season to 106.3 points per game against the Celtics

Ohayo is quoting one of Ben’s older PSRS databases for that +2 ortg and -5 drtg number, which seems far more realistic.

[70sFan]

They were not a great offense in 69(-1.7, -6.4 in the rs, +2, -5 in the playoffs).

Ben Taylor's estimates paints a different picture:

1969 RS: +0.1 rORtg, -4.6 rDRtg
1969 PS: +6.4 rORtg, -0.6 rDRtg

Do I trust these estimations? I'm not really sure, I think I'd have to make some calculations with Ben's database because sometimes I find out that the results for the early years don't add up.

They beat a gauntlet pre-dominantly on their defense.

That's my intuition, but Ben's data disagrees.

Think we may have talked about this recently (do not care enough to search it out), but just on its face those ratings seem untenable. Until some legitimate justification is shown, I would trust common sense over what is posted on a stat-site.

Keeping in mind that the Celtics held teams to 105.4 points per game in the regular season…
76ers: down from 118.9 points per game in the regular season to 106 points per game against the Celtics
Knicks: down from 110.8 (112.4 with Debusschere) points per game in the regular season to 101.8 points per game against the Celtics
Lakers: down from 112.2 points per game in the regular season to 106.3 points per game against the Celtics

One of his older PSRS databases has that team as +2 ortg and -5 drtg in the postseason, which seems far more realistic.

Yeah, that's my intuition in this case as well. I find it hard to believe that the Celtics and their opponents simply played at much, much lower pace, especially when we look at their opponents efficiency shift:

LAL in RS: 46.9% FG, 51.1% TS
LAL vs BOS: 44.2% FG, 48.9% TS

NYK in RS: 45.9% FG, 50.7% TS
NYK vs BOS: 42.1% FG, 47.5% TS

PHI in RS: 45.4% FG, 50.6% TS
PHI vs BOS: 42.4% FG, 47.6% TS

Again, with the available evidences I can't trust Ben's estimations in this case. Maybe he has a mistake in his database?

[DQuinn1575]

Ben Taylor's estimates paints a different picture:

1969 RS: +0.1 rORtg, -4.6 rDRtg
1969 PS: +6.4 rORtg, -0.6 rDRtg

Do I trust these estimations? I'm not really sure, I think I'd have to make some calculations with Ben's database because sometimes I find out that the results for the early years don't add up.


That's my intuition, but Ben's data disagrees.

Think we may have talked about this recently (do not care enough to search it out), but just on its face those ratings seem untenable. Until some legitimate justification is shown, I would trust common sense over what is posted on a stat-site.

Keeping in mind that the Celtics held teams to 105.4 points per game in the regular season…
76ers: down from 118.9 points per game in the regular season to 106 points per game against the Celtics
Knicks: down from 110.8 (112.4 with Debusschere) points per game in the regular season to 101.8 points per game against the Celtics
Lakers: down from 112.2 points per game in the regular season to 106.3 points per game against the Celtics

One of his older PSRS databases has that team as +2 ortg and -5 drtg in the postseason, which seems far more realistic.

Yeah, that's my intuition in this case as well. I find it hard to believe that the Celtics and their opponents simply played at much, much lower pace, especially when we look at their opponents efficiency shift:

LAL in RS: 46.9% FG, 51.1% TS
LAL vs BOS: 44.2% FG, 48.9% TS

Celtics went from about 101 fga in RS vs 95 in playoffs, FTA were about 32 in both. So 6 fga, adjust for TO & ORB, but difference in pace is probably about 5ish.
B-Ref has 117.5/111.6 - the differential of about 6 seems right.
But the numbers themselves are subject to some estimates, and I dont like reading too much into them - you're estimating ORB, TO, and the impact of different FT rules.

NYK in RS: 45.9% FG, 50.7% TS
NYK vs BOS: 42.1% FG, 47.5% TS

PHI in RS: 45.4% FG, 50.6% TS
PHI vs BOS: 42.4% FG, 47.6% TS

Again, with the available evidences I can't trust Ben's estimations in this case. Maybe he has a mistake in his database?

[OhayoKD]

I will start the thread in a moment.

Cool.

Would be curious to see what a positive case against 72 looks like. Honestly haven't seen much of one against 71 for players people rank higher or the same(shaq, jordan, ect) and we can scale 72 directly off since he's basically just a better version.

I have some reservations about his ability to anchor offense at the top level, but to be honest 1972 Kareem is probably the strongest all-time level season I can imagine. I think it's reasonable to put him at the top tier. It's just that the Bucks offense underperformed a little bit in the playoffs in 1971-73 period that gives me some reservations about taking him at the same level as the best peaks ever.

Sure. That still gets us back to weighing Kareem's defensive advantage vs the other folks offensive advantage. You're not really making a comparative case for or against here. From the broadstrokes 71 looks to me like 91 and 67 with less steps in between and I know Kareem plays better the next postseason which also looks like a better version of 90. What reason is there to think the offense disadvantage was outweighed by the defensive advantage?

I have my thoughts about 2015 LeBron, but I don't want to start another James debate. I just want to mention that I made a few minor changes in the top 10:

- 2016 James up to "GOAT tier",
- 1971 and 1972 Kareem up to "GOAT tier",
- 1969 Russell up to "MVP tier",
- 1987 Jordan up to "MVP tier",
- 2001 Shaq down to "all-time tier".

It didn't change the order, but it changed the overall career value estimations a little bit.

Yeah, I still don't see much of a case for alot of the years you have ranked higher but whatever.

Because I don't see it much. At least not between the highest all-time peaks/primes. I haven't seen positive comparisons made between Kareem and Lebron, Kareem and Jordan, Kareem and Shaq, Lebron and Jordan, Lebron and Shaq, Hakeem and Jordan, Shaq and Duncan, Shaq and Hakeem, ect. I do see you compare players to themselves or make claims about x gap being bigger than y gap, or x player being more consistent than y player, without much reasoning.

Well, I can't compare everyone to everyone. It's time consuming to create such list and to keep everything in mind, I don't have (and won't have) the time to compare every season between all top 10 players. If you want to search for my reasonings, you can read this project for a start:

viewtopic.php?f=64&t=2120549

Alright. Going through your reasoning with the aforementioned examples in chronological order(excepting the ones that haven't had a thread-created)

Kareem vs Shaq

1. 1977 Kareem Abdul-Jabbar - can't imagine other choice here. Jabbar at his offensive peak wasn't any less intimidating or effective than Shaq. On top of that, his defensive effort and tendencies were clearly more impressive. Prime Shaq never played with as bad team as 1977 Lakers, so we shouldn't put lack of ring against Kareem.

2. 1974 Kareem Abdul-Jabbar - I am a bit more impressed by Kareem's skillset in 1977, but he was basically a finished product by 1974. Good RS backed up by dominant postseason run. His performance against 1st Bulls defense is spectacular - you should check it out.

3. 2000 Shaquille O'Neal - by far his best RS, strong defensive effort for his standards and very impressive postseason run. He didn't face any team close to 1972 Lakers, but I give him benefit of doubt.

4. 1972 Kareem Abdul-Jabbar - I decided to go with more matured version over 1971. It's the best Kareem RS performance (and one of the best ever period), but he did struggle against Thurmond and Wilt. To his credit, Bucks were injured and Jabbar had to carry his team heavily against GOAT-level team.

77 features a positive comparison, "same offense+better defense". No comparison is really made between the next 3-years beyond "faced easier comp, won". More time is spent comparing Kareem and Shaq with themselves than with each other.

Kareem vs Lebron

1. 1977 Kareem Abdul-Jabbar - you all know that Kareem is my GOAT and it's not only related to his longevity. 1977 is one of the greatest carry-jobs in NBA history and his postseason run is on par with any of James. Why 1977 over 2009? I think that Kareem's playoff performance was less related to variance ("luck" in shooting) and was more repetitve. Is it fair? I don't know, but I do prefer Kareem's skillset and ceilling raising ability over 2009 James.

2. 2012 LeBron James - unreal postseason run, James dominated his competiton through sheer physicality. I've never seen James being so dominant and focused before or since.

3. 2009 LeBron James - very, very tough choice between 2009 vs 1974. As high as I am on this season, it's possible that I may underrate it still.

You basically just say Lebron's 2009 playoff shooting was a hot-streak without addressing the b2b 40-win lift regular-seasons, and don't really compare in terms of skillset or how their strengths/weaknesses contribute to winning that-year or in general. Granted I think it's probably defensible from an era-relative impact perspective ifyou account for lower-srs tresholds post-merger(top-teams were +4/+5 during that period), but nothing here justifies favoring one player over the other, and no attempt is made at all with the year you have at #2. The 2012 bit only compares Lebron with himself.
Hakeem vs Jordan

- 1993/1994 Hakeem vs 1991/1990/1989 Jordan is a very close debate, I'm close to picking Hakeem here but I have concerns about his offense,
- 1995 Hakeem vs 1992/1993 Jordan is also interesting comparison, though I'm afraid that Hakeem's mediocre RS isn't enough to put him ahead of 1993 Jordan (not sure about 1992),
- 1989/1990 Hakeem vs 1988 Jordan is another tier of comparison, I also think that 1996 should be there,
- 1997 Jordan will probably finish my list, I don't think that 1987 Hakeem was good enough to compete here.

There's basically no comparative case here. Just an allusion to "offense vs defense" and an implication that 93 RS>95.

Hakeem v Shaq

1. 1993 Hakeem Olajuwon - probably the most surprising pick in this thread, I'll be probably the only one who has Hakeem ahead of Shaq peak-wise. I'm not super comfortable with my decision, but I've seen 35 games of peak Hakeem and around 20 games of peak Shaq within last 4 months and Hakeem's defensive edge is so massive that I don't believe it can be neutralized by Shaq's superior offense. Don't get me wrong, Shaq was clearly better offensively but unlike Kareem, his defense was tiers below Hakeem.
2. 1994 Hakeem Olajuwon - very close to 1993, with slightly weaker defensive motor.

3. 2000 Shaquille O'Neal - probably the best offensive season in this comparison, monstrous postseason and great effort throughout the whole season.
4. 2001 Shaquille O'Neal - just as good in the playoffs as 2000 version, but worse RS.

We get more of a comparison here, "hakeem defensive adv>shaq offensive adv", but no justification is offered. 2000, 2001 and 1994 are compared to themselves.

Lebron vs Jordan.

1. 1991 Michael Jordan - as I said, I think that Jordan's effort was more consistent throughout the season. It's close and I do think that James had higher highs (especially on defense), but consistency matters and you won't find many more consistent players than Jordan.
2. 1990 Michael Jordan - similar to 1991, only with slightly lesser experience on playing in traingle offense.
3. 2012 LeBron James - the debate was mostly focused on 2009 James, but I think that people forget how ridiculous he was in 2012. His Miami years are very underrated as a whole.
4. 2009 LeBron James - I'd compare that season to 1989 Jordan and I'm simply more impressed with James carryjob. Probably the most productive season ever on offensive end, though not necessarily the best.
5. 2013 LeBron James - another monstrous season, could be higher but I'm hestitant with his relatively weaker postseason run.

You do make a positive comparison here. The only issue is that said claim("2009 Lebron is the most productive offensive season ever") doesn't support where you end up ranking the players. You say "Jordan is more consistent" even though we know Lebron consistently torches Jordan, as well as several other "goat-tier" peak-holders(Shaq, Duncan, KG) over full-season, multi-season, and career-wide samples via impact. Lebron wins 66-games in the year in question(which is also the biggest modern rs outlier signal in terms of impact), 61 games the next(second biggest modern rs outlier in terms of impact), and puts up the "most productive offensive season ever" carrying a larger offensive load, a larger defensive load, and facing significantly more defensive attention than MJ in the two years you rank above(this was one of the big benefits of the triangle/ceding primacy to Pippen). Lebron also manages similar box-offensive efficiency in the playoffs, while also being the play-caller for his team on both ends. Yet Jordan is ahead because of..."consistency throughout the season" even though he's literally leading teams that are better without him(and well-suited for Mike by 1990) to worse outcomes until he gets himself one of the most loaded-decks in history.

You purportedly value defense, yet you ignore what is basically a more clear-cut version of the comparison you made between 77 Kareem and 2000 Shaq. You say his Miami years were "underrated as a whole", yet scaling Miami over the "most productive offensive season in history", the biggest modern outlier in terms of impact, and one of the best non-big defensive years ever(anchoring a -5.5 defense with the Cavs 2nd best defender a non-factor for half of the season) doesn't actually get him to the top. I'm not sure how that works unless you're using the approach of "claim this year is peak and curve all the signals that aren't this year down below said peak".

Unfortunately there is no link for Lebron vs Shaq or Shaq vs Duncan, so I guess I'll stop there. "Consistency" as you use it does read pretty similar to "dominance" with Shaq and "unstoppable" with Durant.

But why does he need to be a positive or neutral offensive player to be "anything beyond" mvp in 1969?

Because I don't evaluate players strictly relative to one season, but to the overall era they played in. 1969 Russell wasn't a better player than younger Russell or peak Wilt.

Sure. That doesn't really matter in a comparison with modern players though. Wilt also only won when his team was super-good so it's hard to be confident in him as a peak.

The Celtics were the most dominant team ever on the strength of their defense which you say for Russell was still "as good as ever".

Yeah, but his offense hurt Celtics and Celtics defense didn't do nearly as well in the playoffs (though maybe it's a noise of estimations).

I think Ben's site is just bugging, the -5, +2 split came from his database. All the Boston playoff-splits swapped bizarrely iirc

Skill-set analysis is fine, but when you're comparing different-types of players in different-types of situations/contexts, we still need some sort of basis to weigh one attribute to another. The Celtics post a league-best srs based entirely on their defense.

2nd best to be precise.

By a few tenths of percentage, but sure. "League-best lvl" seems fair, though if you want to say marginally behind the best in the league, that still is a lot better than I can say for "all-time" carry-jobs like 1988 or "goat" carry jobs like 1989 while multiple teams are doubling the team in queston by srs. This is also probably a good-point to clarify that 84 is not being used for rookie Jordan, it's being used to establish an upper-bound for prime Jordan with the generous assumption that Jordan was the only reason the Bulls srs improved by 8-points from 1984 to 1988. That is a year you have "all-time" but with me giving Jordan all the credit for the best full-strength srs he posted pre-triangle(alongside his best RAPM score), it really is a "peak" signal. And with that in mind...

I don't think I mentioned rookie Jordan, rookie Jordan is definitely worse player than 1969 Russell in my evaluation, he's now two tiers below 1969 Russell.

The 1988 Bulls were a +3.8 team at full-strength in a season where the best team(iow the best chance at stopping you from a championship) was +6.15. In 1989 the Bulls were +2.3 in a league where the best team was near +8, the 2nd best team was near +7 and 4 teams were higher than +6. Forget that Boston actually won, Is being worth 7-points on a team that is right behind league-best worse than being worth 8 points on a team that is half as good as the best competition?

That 8-points is an artificially juiced signal, and I don't think, specifically looking at championship likelihood, it's actually better. Both were replaced by or replacements for bad players. You can say there's still uncertainty, but that comes with the territory when you're rating players you don't have all the data for and putting them higher or lower than others.

Secondly, you're using a 1-game sample 70's. One that is contradicted by all the larger ones:

No one year, but two years sample identical to 1969-70 Russell.

Fair enough. TBH, I don't have a strong opinion on Moses, but er...

mmmm...The Rockets also lost Robert Reid and Mike Dunleavy, and Elvin Hayes had age catch up to him. And Houston made no real attempt to shore up their offense when Malone was traded; the "plan" was to have Allen Leavell shoot more. He wasn't a volume scorer; his FG% dropped precipitously. I'm pretty sure Philadelphia traded a first round pick they had acquired from Cleveland, who was in the gutter at that time. At the time, I remember that everyone figured the Rockets were setting themselves up for the top pick in 1983; they wanted Ralph Sampson and another great player. they ended up getting Sampson at #1, and picking up Rodney McCray with #3. McCray was a good player, but didn't turn out to be worth the #3 pick. At the time, I remember Houston sportswriters and fans wondering why the Rockets didn't go for the local guy who played in two consecutive Final Fours with the University of Houston...the guy named Clyde Drexler.

If you're thinking, "Hey, there's no way the Rockets purposefully let their team be lousy to be sure they'd get the top picks" ... think again. They did the same thing the next year. The Rockets were 18-26 at the mid point of the season...and went 11-27 the rest of the way, ending the season with a 3-14 run. Bill Fitch, who disliked Elvin Hayes pretty intensely, mysteriously let Elvin played extended minutes in the final two months, when the Big E was totally shot as a player. The story was that Fitch wanted Elvin to get to 50,000 minutes. Right.

Not saying all of that made up anything like 32 games worth of difference...but those things add up. Sure, Moses leaving was a big part of the drop. So was losing Reid (who was good) and Dunleavy (who was a very good bench player--comparable to Steve Blake today) and having Hayes slow down (he played 25% less) and not making any real moves to improve the team.

That seems more like a 1999 Bulls or 2015 Miami situation than a 2011 Cavs, 1970 Celtics, or 94 Bulls one. Not going to challenge the rest though. If you think he has the impact-profile of an all-timer, it's possible he deserves to be considered as such, at least without a compelling counter-case based on his skillset or whatever. The point of bringing other years would be to see if there's replication(the more something happens, the less likely it's situational or noise). Rookie Shaq and Bird have crazy looking rookie-lift but then they never replicate it even with substantial off-samples later on and the team regresses. Kareem has crazy looking rookie-lift but then the team gets better the next two years and was considered one of the bitw before he entered the league. So even if the raw-numbers say shaq/bird = Kareem, I would rate rookie Kareem higher. Now if Ben was disingenuous with his samples...

2. That alone doesn't explain anything, but Ben also doesn't adjust for the fact that the Bullets roster was completely different in 1986 and 1987:

...then maybe we rate him too low.

Fortunately, we have Harvey Pollack’s plus-minus data for Moses’ four seasons in Philadelphia to help evaluate his impact. His AuPM oscillates between strong (but not transcendent) and pedestrian in those four seasons, with ’83 and ’85 looking like typical top-20 seasons and ’84 and ’86 lacking impact. Similarly, Moses’ regressed game-level data tells us that he made a difference, but that his impact was far short of a Grade-A superstar’s.

Moses plus-minus data looks phenomenal in 1983 and 1985, I have no idea why Ben uses such language. I am afraid Ben wanted to show in his article why he's so low on 3 times MVP and it came out a bit too negative in my opinion.

Ben does seem better at accumulating data than he is at logically applying it...

We've both agreed that replication/corroboration matters. Russell has never lost when healthy, and all the signals with relevance noisy they may be support what the largest possible sample says. We can acknowledge uncertainty but uncertainty itself is not a good reason for claiming a player is better or worse than another. Ditto with Lebron and Kareem. These are 3 players who from what we have were posting outlier signals relative to the field from teen-aged to their 30's and beyond. They also happen to have a whole collection "goat-lvl" or "outlier" looking prime/peak stuff, even in down-years(2015, 1975). Why would we take their signals to be noise, when it's exactly what we should expect given where they started and ended(or in Lebron's case ending) and they've posted "goat-level" or better impact again and again?

You almost talk about it like these two would be out of my top 10, while they are the clear 3 outliers at the top, with Jordan/Duncan/Hakeem being far behind Russell...

[/quote]
I mean, those 3 are outliers based on how long they played, right? You don't have them gaining significant separation at their best(if at all), and besides Russell, it doesn't seem you think their primes are stand-out beyond how long they were. And that's despite them having better peak signals/data and/or better replication than a bunch of the guys you seem to see as peers longetvity aside(Shaq, Jordan, Wilt, Hakeem, Duncan), ect.

If having the best data, replicating it across a variety of contexts, having major advantages over virtually anyone at your position or overall for basically the entirety of your prime/career, and demonstrating you can impact winning at the highest-level(or even higher) over and over again isn't enough to gain separation. What is?

[70sFan]

Sure. That still gets us back to weighing Kareem's defensive advantage vs the other folks offensive advantage. You're not really making a comparative case for or against here. From the broadstrokes 71 looks to me like 91 and 67 with less steps in between and I know Kareem plays better the next postseason which also looks like a better version of 90. What reason is there to think the offense disadvantage was outweighed by the defensive advantage?

Well, for once I don't think Kareem has defensive advantage over 1967 Wilt. If I am a bit lower on early Kareem offense, that could lead me to pick 1967 Wilt over him, considering that I am also quite high on his offense.

As the time went on, I think it's harder and harder to compare perimeter players and centers, especially across different eras. What makes you confident that 1971 equals to 1991? Similar team success?


Alright. Going through your reasoning with the aforementioned examples in chronological order(excepting the ones that haven't had a thread-created)

Kareem vs Shaq

77 features a positive comparison, "same offense+better defense". No comparison is really made between the next 3-years beyond "faced easier comp, won". More time is spent comparing Kareem and Shaq with themselves than with each other.

I have made a few long posts comparing Kareem and Shaq in the top 100 project (or was it peak project?) if you want to know my opinion about this comparison.

Kareem vs Lebron

You basically just say Lebron's 2009 playoff shooting was a hot-streak without addressing the b2b 40-win lift regular-seasons,

I didn't, because I don't 100% trust this signal. 2009 and 2011 Cavs team were not the same and the Cavs tanked heavily in that season. I didn't mention 1969->1970 lift either, because I am less focused on such WOWY samples than you.

and don't really compare in terms of skillset or how their strengths/weaknesses contribute to winning that-year or in general.

I think it's obvious for posters here what advantages one has over the other. It's not really that groundbreaking to mention that Kareem was a better defender or LeBron was a better playmaker. There are some interesting things there (I think Kareem's isolation scoring was clearly more valuable than LeBron's relative to the era, I talked about it a few times on this forum) but overall it's a matter of what you value higher. I think that Kareem defensive advantage proved to be a little more valuable and led to better results, but again - different systems, different rosters, different opponents...

Hakeem vs Jordan
There's basically no comparative case here. Just an allusion to "offense vs defense" and an implication that 93 RS>95.

If you have any questions, feel free to ask.

Hakeem v Shaq

We get more of a comparison here, "hakeem defensive adv>shaq offensive adv", but no justification is offered. 2000, 2001 and 1994 are compared to themselves.

Sorry that I not everyone has enough time to write 10 full articles per day, but people have real lifes, you know?

You say "Jordan is more consistent" even though we know Lebron consistently torches Jordan, as well as several other "goat-tier" peak-holders(Shaq, Duncan, KG) over full-season, multi-season, and career-wide samples via impact.

Except that we don't know that... We don't have comparable stats for Jordan and even if we had, they are not always conclusive. Again, not everybody focuses almost solely on WOWY analysis.

Lebron wins 66-games in the year in question(which is also the biggest modern rs outlier signal in terms of impact), 61 games the next(second biggest modern rs outlier in terms of impact), and puts up the "most productive offensive season ever" carrying a larger offensive load, a larger defensive load, and facing significantly more defensive attention than MJ in the two years you rank above(this was one of the big benefits of the triangle/ceding primacy to Pippen). Lebron also manages similar box-offensive efficiency in the playoffs, while also being the play-caller for his team on both ends. Yet Jordan is ahead because of..."consistency throughout the season" even though he's literally leading teams that are better without him(and well-suited for Mike by 1990) to worse outcomes until he gets himself one of the most loaded-decks in history.

What argument do you have for 1990 and 1991 Bulls playing better without him? What the hell are you talking about?

You purportedly value defense, yet you ignore what is basically a more clear-cut version of the comparison you made between 77 Kareem and 2000 Shaq.

Maybe because I am not nearly as high on James defense as you are (while having it higher than Jordan).

By the way, I'd probably change the order if I participate in the same project now. At this point, I think that James peaked higher than Jordan.

You say his Miami years were "underrated as a whole", yet scaling Miami over the "most productive offensive season in history", the biggest modern outlier in terms of impact, and one of the best non-big defensive years ever(anchoring a -5.5 defense with the Cavs 2nd best defender a non-factor for half of the season) doesn't actually get him to the top. I'm not sure how that works unless you're using the approach of "claim this year is peak and curve all the signals that aren't this year down below said peak".

Yes, I scaled Miami LeBron (2012 specifically) higher than 2009 because I think he was capable of doing the same thing 2009 LeBron did, while being more impressed with his offensive development. I also don't think there is any defensive advantage either way between 2009 vs 2012.

"Consistency" as you use it does read pretty similar to "dominance" with Shaq and "unstoppable" with Durant.

Cool, at least I don't say things like:

"we know Lebron consistently torches Jordan, as well as several other "goat-tier" peak-holders(Shaq, Duncan, KG) over full-season, multi-season, and career-wide samples via impact."

with deciding what impact is and what isn't, while using completely arbitrary and absurd WOWY samples like 1984 to 1990 vs 2009 to 2011.

Sorry, but if you think your way of thinking is the only way, then we won't have a conversation here.

By a few tenths of percentage, but sure. "League-best lvl" seems fair, though if you want to say marginally behind the best in the league, that still is a lot better than I can say for "all-time" carry-jobs like 1988 or "goat" carry jobs like 1989 while multiple teams are doubling the team in queston by srs.

We don't know how good 1988 team would be without Jordan, we don't have any significant sample. Based on roster construction, I think it's fair they'd be significantly worse than 1970 Celtics.

This is also probably a good-point to clarify that 84 is not being used for rookie Jordan, it's being used to establish an upper-bound for prime Jordan with the generous assumption that Jordan was the only reason the Bulls srs improved by 8-points from 1984 to 1988.

Yeah, except it doesn't make any sense. 1984 team has no linking points with 1988 team. This is almost as smart as taking 2017 Spurs and 2023 Spurs and concluding that Kawhi is a GOAT level player.

That is a year you have "all-time" but with me giving Jordan all the credit for the best full-strength srs he posted pre-triangle(alongside his best RAPM score), it really is a "peak" signal. And with that in mind...

The 1988 Bulls were a +3.8 team at full-strength in a season where the best team(iow the best chance at stopping you from a championship) was +6.15. In 1989 the Bulls were +2.3 in a league where the best team was near +8, the 2nd best team was near +7 and 4 teams were higher than +6. Forget that Boston actually won, Is being worth 7-points on a team that is right behind league-best worse than being worth 8 points on a team that is half as good as the best competition?

That 8-points is an artificially juiced signal, and I don't think, specifically looking at championship likelihood, it's actually better. Both were replaced by or replacements for bad players. You can say there's still uncertainty, but that comes with the territory when you're rating players you don't have all the data for and putting them higher or lower than others.

Yes, but again - not everybody rank players based on RS lifts. Basketball isn't limited to that...

Fair enough. TBH, I don't have a strong opinion on Moses, but er...

mmmm...The Rockets also lost Robert Reid and Mike Dunleavy, and Elvin Hayes had age catch up to him. And Houston made no real attempt to shore up their offense when Malone was traded; the "plan" was to have Allen Leavell shoot more. He wasn't a volume scorer; his FG% dropped precipitously. I'm pretty sure Philadelphia traded a first round pick they had acquired from Cleveland, who was in the gutter at that time. At the time, I remember that everyone figured the Rockets were setting themselves up for the top pick in 1983; they wanted Ralph Sampson and another great player. they ended up getting Sampson at #1, and picking up Rodney McCray with #3. McCray was a good player, but didn't turn out to be worth the #3 pick. At the time, I remember Houston sportswriters and fans wondering why the Rockets didn't go for the local guy who played in two consecutive Final Fours with the University of Houston...the guy named Clyde Drexler.

If you're thinking, "Hey, there's no way the Rockets purposefully let their team be lousy to be sure they'd get the top picks" ... think again. They did the same thing the next year. The Rockets were 18-26 at the mid point of the season...and went 11-27 the rest of the way, ending the season with a 3-14 run. Bill Fitch, who disliked Elvin Hayes pretty intensely, mysteriously let Elvin played extended minutes in the final two months, when the Big E was totally shot as a player. The story was that Fitch wanted Elvin to get to 50,000 minutes. Right.

Not saying all of that made up anything like 32 games worth of difference...but those things add up. Sure, Moses leaving was a big part of the drop. So was losing Reid (who was good) and Dunleavy (who was a very good bench player--comparable to Steve Blake today) and having Hayes slow down (he played 25% less) and not making any real moves to improve the team.

That's a good post applying context and showing that raw lifts are usually not representative for actual impact - which was my point. At no point I tried to suggest that Moses was as good (or better) as Russell.

Not going to challenge the rest though. If you think he has the impact-profile of an all-timer, it's possible he deserves to be considered as such, at least without a compelling counter-case based on his skillset or whatever.

I didn't rank any Moses season in "all-time" tier to be precise, I had his best seasons in "MVP level" tier.

The point of bringing other years would be to see if there's replication(the more something happens, the less likely it's situational or noise).

Yes, but we don't have any reliable replication signals for Russell.

Again - I don't criticize Russell as the GOAT candidate (I have him there, quite clearly), only the methodology that focuses too much on WOWY signals.

I mean, those 3 are outliers based on how long they played, right?

Not really:

- Russell played less seasons than Jordan, yet he has a significant career value advantage (he would have a slim advantage even without the last season),
- Kareem would have an advantage even if you count 1970-81 only,
- LeBron would have an advantage even if you only count 2004-17 (or 2010-23 only).
- Russell/Kareem/LeBron all have 4 GOAT-level seasons, Jordan has 3,
- Russell/Kareem/LeBron have 11 MVP+ seasons, Jordan has 10.

You don't have them gaining significant separation at their best(if at all),

Well yes, to do that I'd have to create another tier to differentiate the best seasons ever, but that wasn't my intention in this thread. I do have Jordan's peak below these 3, but I think that at the highest level the ranking becomes too subjective to rely on it for such a project.

and besides Russell, it doesn't seem you think their primes are stand-out beyond how long they were.

That's not really true either, they (especially LeBron) have more All-time+ seasons than anyone else (besides Jordan), they have more MVP+ seasons than anyone else as well. I mean, here is the list for top 10:

All-time+:

LeBron: 9
Kareem: 6
Russell: 7
Jordan: 7
Duncan: 4
Wilt: 4
Hakeem: 5
Shaq: 3
Garnett: 2
Magic: 3

MVP+:

LeBron: 11
Kareem: 11
Russell: 11
Jordan: 10
Duncan: 6
Wilt: 8
Hakeem: 7
Shaq: 8
Garnett: 7
Magic: 7

And that's despite them having better peak signals/data and/or better replication than a bunch of the guys you seem to see as peers longetvity aside(Shaq, Jordan, Wilt, Hakeem, Duncan), ect.

I do think there are way more GOAT-level seasons than you think, yes. I don't think the data you provide are the proof, it only gives us some evidence.

[DQuinn1575]

GOAT-level - 33%
All-time - 27%
MVP level - 20%
Weak MVP level - 15%
All-nba level - 10%
All-star level - 7%
Sub all-star - 4%
Role player - 2%

Thanks, this is a great thread, I really do value your opinions, and your willingness to get input from others, even someone as dense as me sometimes.


Any chance you can list out all your player seasons for at least the top 2 or 3 levels? Like all your GOAT and all-time (and if not a big pain MVP) in one post? It would be a great reference point.


Either way, thanks

[70sFan]

GOAT-level - 33%
All-time - 27%
MVP level - 20%
Weak MVP level - 15%
All-nba level - 10%
All-star level - 7%
Sub all-star - 4%
Role player - 2%

Thanks, this is a great thread, I really do value your opinions, and your willingness to get input from others, even someone as dense as me sometimes.


Any chance you can list out all your player seasons for at least the top 2 or 3 levels? Like all your GOAT and all-time (and if not a big pain MVP) in one post? It would be a great reference point.


Either way, thanks

GOAT-level seasons:

LeBron James: 4 (2009, 2012, 2013, 2016)
Kareem Abdul-Jabbar: 4 (1971, 1972, 1974, 1977)
Bill Russell: 4 (1962-65)
Michael Jordan: 3 (1989-91)
Tim Duncan: 2 (2002, 2003)
Wilt Chamberlain: 2 (1964, 1967)
Hakeem Olajuwon: 2 (1993, 1994)
Shaquille O'Neal: 1 (2000)
Nikola Jokic: 1 (2023)
George Mikan: 1 (1950)

Total: 24 seasons

All-time seasons:

LeBron James: 5 (2010, 2014, 2017, 2018, 2020)
Kareem Abdul-Jabbar: 2 (1979, 1980)
Bill Russell: 3 (1960, 1961, 1966)
Michael Jordan: 4 (1988, 1992, 1993, 1996)
Tim Duncan: 2 (2004, 2007)
Wilt Chamberlain: 2 (1962, 1968)
Hakeem Olajuwon: 2 (1989, 1990, 1995)
Shaquille O'Neal: 2 (2001, 2002)
Magic Johnson: 3 (1987, 1988, 1990)
Kevin Garnett: 2 (2003, 2004)
Oscar Robertson: 2 (1963, 1964)
Larry Bird: 3 (1984, 1986, 1987)
Jerry West: 2 (1965, 1966)
Stephen Curry: 3 (2015-17)
Julius Erving: 1 (1976)
David Robinson: 2 (1995, 1996)
Giannis Antetokumpo: 2 (2021, 2022)
Nikola Jokic: 1 (2022)
George Mikan: 3 (1951, 1953, 1954)
Bill Walton: 1 (1977)

Total: 47 seasons

[DQuinn1575]

GOAT-level - 33%
All-time - 27%
MVP level - 20%
Weak MVP level - 15%
All-nba level - 10%
All-star level - 7%
Sub all-star - 4%
Role player - 2%

Thanks, this is a great thread, I really do value your opinions, and your willingness to get input from others, even someone as dense as me sometimes.


Any chance you can list out all your player seasons for at least the top 2 or 3 levels? Like all your GOAT and all-time (and if not a big pain MVP) in one post? It would be a great reference point.


Either way, thanks

GOAT-level seasons:

LeBron James: 4 (2009, 2012, 2013, 2016)
Kareem Abdul-Jabbar: 4 (1971, 1972, 1974, 1977)
Bill Russell: 4 (1962-65)
Michael Jordan: 3 (1989-91)
Tim Duncan: 2 (2002, 2003)
Wilt Chamberlain: 2 (1964, 1967)
Hakeem Olajuwon: 2 (1993, 1994)
Shaquille O'Neal: 1 (2000)
Nikola Jokic: 1 (2023)
George Mikan: 1 (1950)

Total: 24 seasons

All-time seasons:

LeBron James: 5 (2010, 2014, 2017, 2018, 2020)
Kareem Abdul-Jabbar: 2 (1979, 1980)
Bill Russell: 3 (1960, 1961, 1966)
Michael Jordan: 4 (1988, 1992, 1993, 1996)
Tim Duncan: 2 (2004, 2007)
Wilt Chamberlain: 2 (1962, 1968)
Hakeem Olajuwon: 2 (1989, 1990, 1995)
Shaquille O'Neal: 2 (2001, 2002)
Magic Johnson: 3 (1987, 1988, 1990)
Kevin Garnett: 2 (2003, 2004)
Oscar Robertson: 2 (1963, 1964)
Larry Bird: 3 (1984, 1986, 1987)
Jerry West: 2 (1965, 1966)
Stephen Curry: 3 (2015-17)
Julius Erving: 1 (1976)
David Robinson: 2 (1995, 1996)
Giannis Antetokumpo: 2 (2021, 2022)
Nikola Jokic: 1 (2022)
George Mikan: 3 (1951, 1953, 1954)
Bill Walton: 1 (1977)

Total: 47 seasons

Thanks,
when they do all-time posts/threads on this board, this is getting my vote.

[trex_8063]

So I'm in process of doing something similar (the CORP-thing); not too far along.

I'm going with the following valuations:
GOAT - 35%
All-Time Great - 29%
MVP - 21%
Weak MVP - 16%
All-NBA - 10%
All-Star - 6.5%
Sub All-Star - 4%
Average/Role Player - 1.5%

I'm assessing how they were relative to their contemporary competition, and that gives me their total cumulative CORP (we'll call that Version 1.0).


Version 2.0 addresses the potential need of some manner of longevity calibration (this is a counter or correction made to the whole "average player for 100 years" argument). But I didn't make the calibration too extreme.

For instance, what if you were looking at Player A and Player B......
Player A had a 20-year career and accumulated 3.000 CORP shares (i.e. 300%).
Player B had a 12-year career and accumulated 2.900 CORP shares (290%).

Who did better? Off the cuff, I would want to say maybe Player B (an additional 10% or 0.1 is not a lot to show for 8 additional years).
However, what if on closer inspection I discovered that Player A had accumulated 2.9 by the end of his 12th season......then had 1 Sub All-Star season, FOUR "average/role player" seasons, followed by three twilight-year seasons of no consequence??

There it would seem Player A had more or less matched Player B year-for-year for his 12-year career........then went on to do a tiny bit more. So he should be considered the greater (EDIT: though things like salary being drawn during those last eight seasons could be considered).
Things playing out like that is probably a bit unlikely; it would be more likely that Player A was simply not quite as dominant peak vs peak, or avg prime year vs avg prime year.......but it could have played out like that. For example, the last several years of Vince Carter's career sort of went like that.

Anyway, so I made the adjustment such that these two players would have basically the same longevity-calibrated CORP score [with Player B just negligibly higher]. One practical example of how it affected things: based on how I valuated their respective seasons, Dirk Nowitzki came out 1% [0.01] higher than Kobe Bryant in raw CORP......but he played 21 seasons [to Kobe's 20]. My calibration makes their calibrated figure exactly equal.
So again, just a small adjustment, but it's there.


Version 3.0 takes that longevity-calibrated [V2.0] number, and makes an adjustment against my subjective assessment of strength of era/environment (I have year-by-year figures, which I may adjust at some point).


Anyway.....all of that was just a preface to asking you [70sFan] how you placed Stephen Curry's seasons. It seems I peg him notably lower than you (you have him 14th).
For me, he finishes just 24th in raw CORP shares (v2.0 number remains at 24th). In v3.0 [with era accounted] he bounces up to 22nd (it's Pettit and Mikan who fall behind him here).

[OhayoKD]

As I've said before, I'm not a fan of pleasantries, but since you're special 70's, I'll take a second to clarify that I don't really have any issues with you. My dryness is just that, dryness. It's not meant as some indication of anger which you seem to take it as at points. When I'm typing I read what I'm typing as monotone in my head(though my actual voice is uhhhhh the opposite ), including the question marks.

That out of the way...

Sure. That still gets us back to weighing Kareem's defensive advantage vs the other folks offensive advantage. You're not really making a comparative case for or against here. From the broadstrokes 71 looks to me like 91 and 67 with less steps in between and I know Kareem plays better the next postseason which also looks like a better version of 90. What reason is there to think the offense disadvantage was outweighed by the defensive advantage?

Well, for once I don't think Kareem has defensive advantage over 1967 Wilt. If I am a bit lower on early Kareem offense, that could lead me to pick 1967 Wilt over him, considering that I am also quite high on his offense.

As the time went on, I think it's harder and harder to compare perimeter players and centers, especially across different eras. What makes you confident that 1971 equals to 1991? Similar team success?

Never said anything about "confident", but I don't really have an issue making probabilistic judgments(uncertainity is fine). Main thing about 1971 vs 1991 is
-> there is no triangle equivalent(Bulls offense goes from +2.3 to +6.5 between the first and 2nd half of 1990),
-> there's no equivalent to the defensive jump(Bulls go from below average at the start of the season to a -3 defense by the 90 playoffs(-5 in the last 2 rounds)).

The Bucks are closer by 1970 with rookie Dandridge than the Bulls get pre-triangle despite Kareem joining a similar team. The Bucks are also still great In 72 in the games Oscar completely misses and unlike 91 where there's no real discernible improvement(Mj's on/off, rapm, defensive tape all looks worse actually)despite facing significantly weaker competition(pistons are way worse defensively and overall in the first two rounds of 1991 compared to 1990).

Kareem's production jumps between 70, 71, and 72 despite worse help(oscar hobbled) and much tougher comp(west+wilt) in 72 vs 71(west hurt). Their full-strength srs also improves iirc. It's also obvious the Bulls were historically loaded when we look at the full-lineup performances in 94 and 95(58-win without Jordan, 52-win without Jordan and Grant), and there's nothing that indicates the same for the Bucks

With Wilt, the issue is the team doesn't really fall off that much without him and then he forms what should have been a dyansty and goes 0 for 2.

I don't have any confidence putting 71 Kareem ahead, but i see more reasons to put it ahead than behind. I am more confident in 72 because it's a very obvious improvement from a season which looks as good to me if not better than 67 and 91 and also looks like a better version of MJ's 90(does better against a better opponent with less help after a more impressive rs), and that's not getting into replication where Kareem only looks better and better compared to Wilt or MJ the more surrounding years you allow into the eval.

Hence I have 72 Goat-tier and everything else chucked down to all-time. I actually didn't have 74 GOAT-tier, not because I think Kareem is necessarily worse(74 might be his 2nd best playoff run from a granualar perspective), but because I can't really come up with a favorable comparison the way I can for 77 or 72.

To be confident in one player over the other I want to see either
A. Player did more with less or as much help
B. Player did as much or more with less help

Skillset-analysis is fine, but if it can't be tied to the one of the above or established historical trends(helio-offenses are more resilient than non-helio offenses, paint-protection >>>> everything else on d), saying one skillset is better than the other is a crapshoot.

I guess i should be fair and note that WOWYR says Wilt is actually ~ Russell, but that's a very small sample(2.2 gms/szn for russell) and the actual implication of that would be that the team Bill won 11 rings with was on average, average.

This is also probably a good-point to clarify that 84 is not being used for rookie Jordan, it's being used to establish an upper-bound for prime Jordan with the generous assumption that Jordan was the only reason the Bulls srs improved by 8-points from 1984 to 1988.

Yeah, except it doesn't make any sense. 1984 team has no linking points with 1988 team. This is almost as smart as taking 2017 Spurs and 2023 Spurs and concluding that Kawhi is a GOAT level player.[/quote]
The linking point is that Jordan's cast was better in 88 than it was in 84(oakley, dropped a bunch of defensive negatives, drafted pippen and grant, coaching upgrade, average defense becomes good), just like in the case for Kawhi it would be that the 2023 Spurs were worse than the 2017 Spurs. The Bulls made various improvements and I'm pretending the only improvement that occurred was Jordan getting better. Hence it establishes an upper-bound. Jordan is probably not as valuable as 88-84, but he's certainly not more valuable.

I also do the opposite with Kareem. As you pointed out, the 30-win team he joined was gutted before they added him. So we can set a lower-bound by subtracting 75 srs from 76 or 77 where there were no major additions to counteract that roster being gutted.

We can logically conclude then that Kareem had less help than Jordan when he led a 55-win full-strength team to Jordan's 52 or 53-win full-strength team in a league with lower srs-tresholds. There's also an abundance of corraboration because Kareem has a whole set of samples which get close to the aforementioned upper-bound(70, 76, 75 ect). As a result I have 77 at "goat-tier" and all of Jordan's years at "all-time". I juiced up his value and he still wasn't a match.

If the gap between the 2017 Spurs and the 2023 Spurs was somehow lower than the gap between the 2018 Rockets and the 2019 Rockets having traded chris-paul for steph curry, it would be fair to say Steph Curry was probably more valuable right?

77 features a positive comparison, "same offense+better defense". No comparison is really made between the next 3-years beyond "faced easier comp, won". More time is spent comparing Kareem and Shaq with themselves than with each other.

I have made a few long posts comparing Kareem and Shaq in the top 100 project (or was it peak project?) if you want to know my opinion about this comparison.

[/quote]
Okay, will look later.

You basically just say Lebron's 2009 playoff shooting was a hot-streak without addressing the b2b 40-win lift regular-seasons,

I didn't, because I don't 100% trust this signal. 2009 and 2011 Cavs team were not the same and the Cavs tanked heavily in that season. I didn't mention 1969->1970 lift either, because I am less focused on such WOWY samples than you.

Well we actually have a pretty good sample(21 games/szn) just with that same rotation-lebron where the Cavs played like an 18-win team per srs. That's also corroborated by the three-year off from 08-10 and you have similarly crazy lift in his second cleveland stint(the 30ish win mark i throw around comes from taking lineup-ratings with kyrie/love and no lebron which view the Cavs as a 30ish win team as opposed to WOWY which views them as a 20ish win team). There's also the aforementioned RAPM(a stat that tends to suppress outliers) dominance. Another set we can look at is a 20-year one from 2000-2021. Applying a 100,000 possession filter, there's a bigger gap between Lebron and #2 KG then KG and #5 Tim Duncan per-possession despite Lebron playing 70, 000 more possessions.

I don't "100%" trust the signal either, but that signal still adds to his case probabilistically, especially when that sort of craziness is hardly a one-off. No one else post-merger puts up those sorts of signals repeatedly. No one else in data-ball touches him in gap-suppressing regularized samples. And this is a player who has the best playoff winning% in history among MVPs as an srs-underdog. He's already a statistical outlier in the regular season, per-possession, for his career and 2009/10 stuff are the biggest(empirically) outliers of all those outliers. That's an important factor a proper counter-case should address(though for Kareem and Russell, the Cavs losing early in playoffs thanks to 3-point variance helps).

and don't really compare in terms of skillset or how their strengths/weaknesses contribute to winning that-year or in general.

I think it's obvious for posters here what advantages one has over the other. It's not really that groundbreaking to mention that Kareem was a better defender or LeBron was a better playmaker. There are some interesting things there (I think Kareem's isolation scoring was clearly more valuable than LeBron's relative to the era, I talked about it a few times on this forum) but overall it's a matter of what you value higher. I think that Kareem defensive advantage proved to be a little more valuable and led to better results, but again - different systems, different rosters, different opponents...

Can you link me to the isolation scoring post? I've generally given Kareem a scoring edge and been pretty firm about Kareem's defense generally being better even if there are situations where Lebron can come close(small-ball opponents help). But I think Lebron has more advantages than just playmaking. Ball-handling is a pretty big deal, not just in terms of playmaking, but also in terms of being able to get to spots to score or create from. And you have the mental side of the game where Lebron is clearly advantaged as a floor-general/playcaller on both ends.

Hakeem vs Jordan
There's basically no comparative case here. Just an allusion to "offense vs defense" and an implication that 93 RS>95.

If you have any questions, feel free to ask.

[/quote]
How are you comparing MJ's 93 and Hakeem's 95 as regular seasons?

I guess another question is how valuable Drexler was in 95.

Hakeem v Shaq

We get more of a comparison here, "hakeem defensive adv>shaq offensive adv", but no justification is offered. 2000, 2001 and 1994 are compared to themselves.

Sorry that I not everyone has enough time to write 10 full articles per day, but people have real lifes, you know?

I'm just sorting through what you linked me to. Don't conflate me talking in a dry manner with anger. I just prefer being direct.

You say "Jordan is more consistent" even though we know Lebron consistently torches Jordan, as well as several other "goat-tier" peak-holders(Shaq, Duncan, KG) over full-season, multi-season, and career-wide samples via impact.

Except that we don't know that... We don't have comparable stats for Jordan and even if we had, they are not always conclusive. Again, not everybody focuses almost solely on WOWY analysis.

Well we also now have some prime rapm data(can be compared directly looking at "the gaps" in sets), playoff on/off, and on+on for Jordan's 2nd and 5th best regular season teams. I never said the data was conclusive. But it is consistently favoring one guy. Maybe it's not "proof", but what's the counter-case?

Wouldn't you expect the bigger, smarter and more versatile player to be better at generating wins? I think NBA history would suggest yes.

We also do have complete data for players like duncan, shaq, and kg. Do you think MJ was a different calibre? Raw impact-stuff would say no as does your own seasonal-tier list.

None of that's definitive, but if we're assessing things probabilistically...

You purportedly value defense, yet you ignore what is basically a more clear-cut version of the comparison you made between 77 Kareem and 2000 Shaq.

Maybe because I am not nearly as high on James defense as you are (while having it higher than Jordan).

By the way, I'd probably change the order if I participate in the same project now. At this point, I think that James peaked higher than Jordan.

Cool, I'm just working off what you linked me. We've gone over the defensive stuff in-depth(was specifically a second-cavs lebron-brook comparison iirc), and I'm sure you've seen the various granular-break downs from hcl, hikari, colts, ect. so no reason to continue there.

You say his Miami years were "underrated as a whole", yet scaling Miami over the "most productive offensive season in history", the biggest modern outlier in terms of impact, and one of the best non-big defensive years ever(anchoring a -5.5 defense with the Cavs 2nd best defender a non-factor for half of the season) doesn't actually get him to the top. I'm not sure how that works unless you're using the approach of "claim this year is peak and curve all the signals that aren't this year down below said peak".

Yes, I scaled Miami LeBron (2012 specifically) higher than 2009 because I think he was capable of doing the same thing 2009 LeBron did, while being more impressed with his offensive development. I also don't think there is any defensive advantage either way between 2009 vs 2012.

I think 09 has a very clear advantage in terms of man-d/shot-contests but sure. There are a bunch of explanations for the impact-drop(playing pf instead of sf, overlap with a similar player, staggered lineups with co-stars), but an important part of scaling is looking the evidence for the year you're scaling a player above, no? There are three-ways to scale internally with an assumed peak:
-> Assume all evidence outside said peak is noise and curve everything down below aforementioned peak(this is what Ben does)
-> Assume the peak is noise and curve it up above everything else with the better evidence
-> Assume both can be noise and have your assessments meet somewhere in between

Which approach are you using here?

"Consistency" as you use it does read pretty similar to "dominance" with Shaq and "unstoppable" with Durant.

Cool, at least I don't say things like:

"we know Lebron consistently torches Jordan, as well as several other "goat-tier" peak-holders(Shaq, Duncan, KG) over full-season, multi-season, and career-wide samples via impact."

I mean, that is a claim that can be scrutinized referencing data everyone has access to with a fully transparent approach/methodology anyone here can replicate. I also don't think any of the decisions being made here are "arbitrary" but you're welcome to specify any issues you have. The alternative I'm apparently being unfairly dismissive of is an uh...adjective.

with deciding what impact is and what isn't

Uh, I've looked at and used all the metrics someone who had Jordan #1 and Shaq #3 agreed were valid as impact with the exception of WOWYR. You were also there when a bunch of different posters, myself included, offered justification for not taking WOWYR seriously. If you have a gripe with said justification, then just say it.

I don't think the issue is there being only one way of thinking. The issue is that no one has really been able to offer a coherent alternative that logically leads to the reverse conclusion. In lieu of that alternative, why should anyone see the counter-stance(on championship likelihood specifically) as reasonable?

I also guess you forgot it's you, not me, that has said another person's approach on evaluating player-goodness is "objectively wrong"

That is a year you have "all-time" but with me giving Jordan all the credit for the best full-strength srs he posted pre-triangle(alongside his best RAPM score), it really is a "peak" signal. And with that in mind...

The 1988 Bulls were a +3.8 team at full-strength in a season where the best team(iow the best chance at stopping you from a championship) was +6.15. In 1989 the Bulls were +2.3 in a league where the best team was near +8, the 2nd best team was near +7 and 4 teams were higher than +6. Forget that Boston actually won, Is being worth 7-points on a team that is right behind league-best worse than being worth 8 points on a team that is half as good as the best competition?

That 8-points is an artificially juiced signal, and I don't think, specifically looking at championship likelihood, it's actually better. Both were replaced by or replacements for bad players. You can say there's still uncertainty, but that comes with the territory when you're rating players you don't have all the data for and putting them higher or lower than others.

Yes, but again - not everybody rank players based on RS lifts. Basketball isn't limited to that...

Do the playoffs hurt Russell here? Celtics obviously get way better. Hondo steps up, but he also improves the next year anyway. I don't know what Russell looks like throughout the postseason but him being in foul trouble nearly causes a crazy comeback in game 7. You also say he does better offensively in specific playoff matchups.

I only weigh raw lifts to that extent when we dont have much else to go off. But I think if someone not named Russell faced similar competition and won, there'd be a lot of people clamoring that it was the best year ever. You can say Russell was clearly worse than he was earlier(though anenigma offers a decent counter-case in the other thread), but that only helps anyone else if you establish a limit for earlier Bill, and it's hard to establish an upper-limit in terms of winning championships for someone who has never lost when healthy.

Uncertainty is one thing, but why place 69 a tier below seasons it compares favorably to in terms of winning because it's possibly misrepresentative?

Fair enough. TBH, I don't have a strong opinion on Moses, but er...

That's a good post applying context and showing that raw lifts are usually not representative for actual impact - which was my point. At no point I tried to suggest that Moses was as good (or better) as Russell.

I mean, that depends on how the roster changes no? If the roster is basically the same or it changes in one direction, why wouldn't it be representative outside of replacement quality(and in this case, neither Jordan or Bill replaced good players)

The point of bringing other years would be to see if there's replication(the more something happens, the less likely it's situational or noise).

Yes, but we don't have any reliable replication signals for Russell.

"Reliable", arguably not, but that's what comes with the territory with someone from the 60's who rarely missed substantial time. Those unreliable signals do tend to strengthen Russell's case though, with that rookie stretch being the only real counterpoint. Also, while it's not "off", Russell did win 5 championships in 6 with a completely different core. If Moses had notched a win without Erving we'd probably view him very differently.

Again - I don't criticize Russell as the GOAT candidate (I have him there, quite clearly), only the methodology that focuses too much on WOWY signals.

What other methodology should we apply? He always won on the strength of his defense on teams that gained separation almost entirely on defense and in the year in question, you still say he was as good defensively.

We can use box-production or tape to compare him with himself but where does get you against anyone else? The internal-scaling only works once you establish the external scaling. You can assume Bill Russell's peak on the same level as other players but there's no real reason to assume that sort of ceiling on someone who always won. And if him declining still has him leading the worst cast of his Celtics tenure to one of their best runs(arguably best considering competition), did that aforementioned decline really change anything about how he compares with everyone else?

SRS, point-differential, "production" and "lift" aren't actually what we're looking for with championships. It's just that we need proxies to compare individuals because winning a championship is an inherently unlikely proposition for anyone...except Bill Russell. "Proving" a limit requires failure. Russell basically never failed. We don't know what circumstance he couldn't win in besides "first-year player-coach carrying an injury" or "literally not on the floor".

I mean, those 3 are outliers based on how long they played, right?

Not really:

- Russell played less seasons than Jordan, yet he has a significant career value advantage (he would have a slim advantage even without the last season),
- Kareem would have an advantage even if you count 1970-81 only,
- LeBron would have an advantage even if you only count 2004-17 (or 2010-23 only).
- Russell/Kareem/LeBron all have 4 GOAT-level seasons, Jordan has 3,
- Russell/Kareem/LeBron have 11 MVP+ seasons, Jordan has 10.

You don't have them gaining significant separation at their best(if at all),

Well yes, to do that I'd have to create another tier to differentiate the best seasons ever, but that wasn't my intention in this thread. I do have Jordan's peak below these 3, but I think that at the highest level the ranking becomes too subjective to rely on it for such a project.

Interesting. It would seem we find a way to similar conclusions despite different approaches. I guess I'll point out for posterity, that all else being equal, from a predictive lens, a player who has more "goat-lvl" seasons would be more likely to have "the" goat season based on sheer probability. Part of why I think it's better to start by comparing players "in general" looking for positive comparisons to be made across their careers and then applying whatever internal-scaling there is to be applied.

I am seemingly more picky on what can be "goat-lvl" though. Players have to "earn" consideration beyond 1-year plausibility, especially in comparison with generally better ones.

I also seem to view the gaps between players as larger. I have Russell's prime in a tier of his own above everyone and then Lebron and Kareem's arguable "peaks" in a second one.

That is all era-relative though. Things change considerably if I use a more modernist or "across time" approach.

[70sFan]

Anyway.....all of that was just a preface to asking you [70sFan] how you placed Stephen Curry's seasons. It seems I peg him notably lower than you (you have him 14th).
For me, he finishes just 24th in raw CORP shares (v2.0 number remains at 24th). In v3.0 [with era accounted] he bounces up to 22nd (it's Pettit and Mikan who fall behind him here).

Stephen Curry

GOAT-level: 0
All-time: 3 (2015-17)
MVP: 3 (2019, 2021, 2022)
Weak MVP: 2 (2014, 2018)
All-nba: 2 (2013, 2023)
All-star: 1 (2011)
Sub all-star: 1 (2010)
Role player: 0

I can see 2014 going lower one tier and 2010+2011 going lower one tier as well, which would put him way down to 21st spot, so I understand why you have him clearly lower.

[70sFan]

As I've said before, I'm not a fan of pleasantries, but since you're special 70's, I'll take a second to clarify that I don't really have any issues with you. My dryness is just that, dryness. It's not meant as some indication of anger which you seem to take it as at points. When I'm typing I read what I'm typing as monotone in my head(though my actual voice is uhhhhh the opposite ), including the question marks.

Don't worry, I am fine with your responses. I don't have enough time to respond to your post this week, but I will read it closely and keep your notes in mind.

[trex_8063]

Anyway.....all of that was just a preface to asking you [70sFan] how you placed Stephen Curry's seasons. It seems I peg him notably lower than you (you have him 14th).
For me, he finishes just 24th in raw CORP shares (v2.0 number remains at 24th). In v3.0 [with era accounted] he bounces up to 22nd (it's Pettit and Mikan who fall behind him here).

Stephen Curry

GOAT-level: 0
All-time: 3 (2015-17)
MVP: 3 (2019, 2021, 2022)
Weak MVP: 2 (2014, 2018)
All-nba: 2 (2013, 2023)
All-star: 1 (2011)
Sub all-star: 1 (2010)
Role player: 0

I can see 2014 going lower one tier and 2010+2011 going lower one tier as well, which would put him way down to 21st spot, so I understand why you have him clearly lower.

That makes sense. It's probably those differences [I've probably done most of the ones you've thought about] that have pegged him down on my listing. I knew that you had listed all of '15-'17 as "All-Time Great"; I actually did not.

NOTE:
One thing I'm allowing myself to do is to dole out half-measures, as well as pooling years. That is: if I can't decide if a season was merely "MVP" or "All-Time Great" [i.e. it seemed halfway between], or if any other season is roughly halfway between two other tiers, I'll list it as such: 0.5 in one tier, 0.5 in the other.

And what I mean by "pooling" years is: suppose there's a guy who had three seasons that seemed mostly like "weak MVP", though maybe 2-3 of them seemed a little better, but not enough to give even a half-measure toward "MVP" to any individual season......maybe I'll credit that three-year span as 0.5 to 1 "MVP" seasons and 2 to 2.5 "weak MVP" seasons.

That may seem dodgey, but I want to credit these guys with what they deserve; I don't want to be too rigid with a protocol that says I need to consider each season on an island by itself, and pick the ONE tier that's closest........because some players are going to get overrated in doing so [because they're consistently marginal or toward the lower end of that nearest tier], while others will be underrated [because they're consistently on the higher end of the tier].


Anyway, for Curry '15-'17, I've credited him with 1.5 "All-Time" seasons and 1.5 "MVP" seasons; I could see going as far as 2 and 1, respectively, but that's it. And I'll try to explain my reasoning.....

'15 Curry I see as on the cusp of All-Time Great in the rs (could even see going all the way there). But his individual brilliance comes down to Earth somewhat in the post-season (down to "merely" MVP level, perhaps), and the Warrior offense, fwiw, also falters in the ps (admittedly against mostly very good defenses). But it was largely their defense that pushed them them thru (some key injuries to Cleveland's roster didn't hurt either). So I'm not sold on giving the year an "All-Time Great" ranking, personally (though admittedly, we have a definite semantic ambiguity about what that even means).

In '16, Curry easily has an "All-Time Great" rs. Hell, you can even consider throwing it in the "GOAT" tier. God, it seemed like he was hitting everything [from everywhere] that year. Teams were attacking him as soon as he crossed half-court, miscue double-teams OFF-ball were happening, teams were so panicked about him. It was nuts.
But then he misses the first 6 games of the playoffs [fortunate to be in a circumstance where that doesn't matter because his team can breeze thru the first round without him], and then is [on average] playing at roughly a "weak MVP" level through the rest of their playoff run: scoring is down [from his arguably GOAT-tier rs], assisting is down, shooting efficiency is way down, turnovers are pretty substantially up.......via a number of different ways, this post-season is multiple tiers removed from his rs excellence; AND he missed those six games.
So here again it's at least a little hard for me to give him full credit for an "All-Time Great" season, even though I'm one who is typically putting more emphasis on the rs than other people. I still view this year as at least halfway between "MVP" and "All-Time", fwiw.

In '17, his efficiency dips a bit relative to '16 [or even slightly vs '15], in what [theoretically] should be an easier circumstance (attention has been diverted from him to some degree [again: theoretically]). But his 3pt% in '17 is comfortably the worst of the three years [worst of his career to that point, actually], his FT% is also lowest of the three, with 2pt% also being comfortably less than in '16.
His numbers [volume] dip a bit too, though I view that mostly as "making room" for Durant to get his shots. So it's without a doubt a weaker rs than '16 [what isn't?], and arguably marginally weaker than '15. His gravity is still tremendous (and there's no question [to me] that he was far superior to Durant in this), but I'm a little hesitant to give it "All-Time" status. It still feels more like "MVP"-tier to me in the rs.
But then he has his BEST playoff run, going for 28/6/7 on just silly shooting efficiency en route to a title, while his gravity paved the way [from time to time] for Durant. This kinda lifts the season up a bit. All the way to "All-Time"? Not sure about that.


When I look at all of that collectively, I have to give at least half-credit toward that higher category for the three years as a whole. Maybe even as far as two-thirds credit. But [for me] not full credit for each. None of them is quite so bullet-proof.

[dygaction]

GOAT-level - 33%
All-time - 27%
MVP level - 20%
Weak MVP level - 15%
All-nba level - 10%
All-star level - 7%
Sub all-star - 4%
Role player - 2%

Thanks, this is a great thread, I really do value your opinions, and your willingness to get input from others, even someone as dense as me sometimes.


Any chance you can list out all your player seasons for at least the top 2 or 3 levels? Like all your GOAT and all-time (and if not a big pain MVP) in one post? It would be a great reference point.


Either way, thanks

GOAT-level seasons:

LeBron James: 4 (2009, 2012, 2013, 2016)
Kareem Abdul-Jabbar: 4 (1971, 1972, 1974, 1977)
Bill Russell: 4 (1962-65)
Michael Jordan: 3 (1989-91)
Tim Duncan: 2 (2002, 2003)
Wilt Chamberlain: 2 (1964, 1967)
Hakeem Olajuwon: 2 (1993, 1994)
Shaquille O'Neal: 1 (2000)
Nikola Jokic: 1 (2023)
George Mikan: 1 (1950)

Total: 24 seasons

All-time seasons:

LeBron James: 5 (2010, 2014, 2017, 2018, 2020)
Kareem Abdul-Jabbar: 2 (1979, 1980)
Bill Russell: 3 (1960, 1961, 1966)
Michael Jordan: 4 (1988, 1992, 1993, 1996)
Tim Duncan: 2 (2004, 2007)
Wilt Chamberlain: 2 (1962, 1968)
Hakeem Olajuwon: 2 (1989, 1990, 1995)
Shaquille O'Neal: 2 (2001, 2002)
Magic Johnson: 3 (1987, 1988, 1990)
Kevin Garnett: 2 (2003, 2004)
Oscar Robertson: 2 (1963, 1964)
Larry Bird: 3 (1984, 1986, 1987)
Jerry West: 2 (1965, 1966)
Stephen Curry: 3 (2015-17)
Julius Erving: 1 (1976)
David Robinson: 2 (1995, 1996)
Giannis Antetokumpo: 2 (2021, 2022)
Nikola Jokic: 1 (2022)
George Mikan: 3 (1951, 1953, 1954)
Bill Walton: 1 (1977)

Total: 47 seasons

Would you think it makes sense to separate playoffs each year with a lower weight, taking into account performance and depth? Curry's 06 would be GOAT regular and discounted playoffs. Garnett's 2005 is arguably better than 2003 from all advanced stats, so maybe another all time season with no playoffs.

Also curious how did you rate Luka's, as he gets on your list already. Did not expect that.