Greatest Peaks Project (2022): #12 - 2003-04 Kevin Garnett

[ceoofkobefans]

10. 2008 Kobe Bryant

I know this is probably going to be controversial on this forum since he’s usually fringe t15 on peak lists around here (due to what seems like RS impact metrics). Most Impact metrics generally do have him around the fringe t15 range (like 13-17ish) but Kobe is one of the biggest PO risers ever. Here’s 08-10 Kobe from the RS to PO (biggest peak PO sample we have without 2 first round exits skewing results).

(Box numbers are IA/75)
RS
28.3 PTS
5.3 AST
5.8 TRB
3 TOV
+1.8 rTS
+5.3 BBR BPM (+3.9/g)
+5 BP BPM/g (+6.7/100)
+4.1 AuPM/g (+5.5/100)
+5.96 RAPTOR (+4.4/g)

PO
30.5 PTS
5.6 AST
5.8 TRB
3 TOV/75
+3.9 rTS
+7.8 BPM (+6/g)
+6.3 BP BPM/g (+8.2/100)
+4.7 AuPM/g (+6.1/100)
+8.07 RAPTOR (+6.2/g)

Here’s just 2008 since that’s his best season

RS
28.1 PTS
5.3 AST
6.1 TRB
3 TOV
+3.6 rTS
+5.8 BBR BPM (+4.5/g)
+6.1 BP BPM/g (+7.9/100)
+4.2 AuPM/g (+5.4/100)
+7.09 RAPTOR (+5.5/g)

PO
30.5 PTS
5.6 AST
5.6 TRB
3.2 TOV
+4.9 rTS
+7.4 BBR BPM (+5.9/g)
+6.7 BP BPM/g (+8.4/100)
+2.1 3yr AuPM/g (+2.6/100 this is obviously skewed by the 2 previous years)
+7.63 RAPTOR (+6.06/g)

31 IA PTS/75 on +5 rTS is absolutely insane when you consider that he’s playing in 2 center lineups with his best spacer being him and facing more gravity than anyone in nba history that’s name doesn’t start with an S. Him being able to pretty much maintain that in the PO over a 3yr stretch of finals runs against GOAT tier PO comp (same points on +4 rTS) is pretty damn good evidence for him being not only an all time PO riser but this scoring production being real for him. Him being an all time PO riser makes since because he’s arguably the best tough shot maker of all time and is a clear all time self creator which is the number 1 way for your scoring to be resilient in the PO against tougher defenses and more defensive attention. His defense was also pretty solid in 2008. He did still have a bit of a motor issue in the RS but it consistently would shoot up in the PO and this was no different in 08. He was a very good on ball defender but was also a good off ball defender (really good trapper, was the lakers’ primary communicator, and I thought his off ball awareness was improved from his past few years, although his closeout D wasn’t great which hurt his overall off ball D)

Overall i think it’s pretty fair to put Kobe in that top 10 range although I could see him at like 14ish(?) depending how high you are on others/low on him

12. 2004 Kevin Garnett

I was surprised he didn’t get voted into the top 10 with how seemingly reliant this server is on impact metrics specifically in the RS. How I can see the argument to have KG at 8 but I can also see him under my number 13 (Bill Russell who was already voted. I may do him a profile for him and then add an extra vote along with him so y’all at least get to see my thoughts on him). Now 2004 KG is usually a GOAT tier season by impact metrics but this is because his offense is very boosted (for example in Ben Taylor’s Scaled APM/g KG is a +9.4 overall and a +5.6 offensively which would put him in the same tier as guys like Kobe Bryant Steve Nash and Dwyane Wade) and his offense drops in the PO. The reason he isn’t as good offensively is because he isn’t a good enough self creator to rise as a scorer (especially as a number 1 option) and isn’t good enough on the ball to maintain the same quality of play when receiving more defensive attention (which he did face in the PO which is why his TOV rates went up in 2004). Now KG is still a very good offensive player and I would consider him an Offensive star but I would say he’s worse than a Tim Duncan Offensively because of his PO resilience. We don’t have very good PO samples for prime KG outside of 2004 and 2008 but he does fall across the board in the PO if you look at his PO stats in 04 and 08

(IA/75)

RS
24.9 PTS
4.9 AST
13.4 TRB
2.5 TOV
+3.4 rTS
+9.4 BBR BPM
+6.6 BP BPM/g
+9.05 RAPTOR


PO (44 GP)
23.8 PTS
4.4 AST
13.1 TRB
3 TOV
+.2 rTS
+6.5 BBR BPM
+5.7 BP BPM
+6.26 RAPTOR

KG I think is a really good representation of why the O2 D1 archetype isn’t amazing. It is rare that it will be a maximized archetype and they aren’t good enough on the ball to maintain value in the PO offensively in situations they’re more likely to be in which hurts their overall value. I could see him higher if you’re more confident in his offense in a PO environment but if you’re lower on his defense (which I didn’t really talk about but I don’t think there’s much to discuss ATG perimeter D big goat tier versatility goat tier help defender elite but not all time rim protector fringe t5 defender ever) I could see him behind my number 13

13. 1965 Bill Russell

Now I know bill Russell was already picked but he is my 13th best peak I wanted to talk about why he’s so low for me. Now bill Russell is the best defender of all time but how big the gap much he is than Hakeem or Tim Duncan is something I do question a lot. I feel like a +6 is the highest I can BR as a defender and that might be gassing it. The 60s Celtics were had the greatest defensive coach ever and idk if there was a single - defender on the entire team. They also had Peak KC Jones who’s one of the best defenders ever and prime Tom sanders. Bill Russell was 100% the most valuable piece of that defense I just question how much value he has. It is hard to determine which year is truly Bill Russell’s best with how little film and data we have of him but 65 is when he was being used more as a playmaker and his scoring likely didn’t change much. Bill Russell isn’t a very good offensive player but he definitely isn’t a - in that aspect. He is an ok scorer and an all time Playmaking center and likely a goat tier offensive rebounder. I like other players two way ability more and have more confidence in their value which is just natural when you have as little information on someone as we do for Bill. I could see him at 10 but could also potentially see him lower 13 is the Lowest I go generally

[Doctor MJ]

Looking at the voring trends I might just be voting for Dr J and Moses for the next 10-15 spots or so but that's completely fine. The top 10 peaks are our top 10 careers in a slightly shuffled order but now it's going to get really interesting. When opinions are this far apart you're going to get some of the most interesting discussions after all, although most of the difference in approach seems to be based on the definition of a peak season.

I've briefly touched upon Doctor MJ's point before a while ago but it's probably a good thing to bring it up again. No disrespect to the people who voted in the earliest peaks and top 100 projects but it's clear to me the process has evolved since then. There were no explanations required, which makes it hard for anyone to convince others and makes people default to their initial assumptions. The voting system itself has also been greatly improved to give more people a say instead of just counting first place votes. The most important factor for seemingly random changes in rankings of retired players is because we're not a monolith. Just because a majority of people 20 years ago held a certain opinion or even if that opinion was present in the previous project that doesn't mean we should all hold that opinion now. What is kind of overlooked in this is voter turnover. Some people who participated in older projects no longer do so, while there are also new people with different perspectives joining (this is my first peaks project as well). That's why I think while it's interesting to compare results to previous projects, this 2022 ranking should be seen first and foremost as it's own thing. Unless we'd have a large voter pool that remained consistent throughout the years, there will always be fluctuations of rankings due to some players having more champions or detractors participate than in other years.

I'm glad you brought it up again because I don't recall seeing your first post.

To try to boil down your words, you seem to be saying:

1. It makes sense that voting will be different when there's voting turnover.
2. We have reason to think we know better than the people from before.

(1) is a definite thing, and (2) may well be true.

But keep in mind that I'm not asserting the opposite of (2) - I'm not saying there's something definitively wrong with this list because of its divergence from previous list. I'm just pointing out things that seem like trends to me, and suggesting that people should meditate on what might cause those trends.

As with any apparent trend, the noise of small sample size is something we need to be concerned with. But there's a difference between running away with a conclusion based on too small a sample, and considering what might cause the trend to be something more than randomness, just as there's a difference between being cautious to come to conclusions based on the apparent trend, and refusing to consider the possibility that there are meaningful causal explanations for some of what we're seeing.

[Doctor MJ]

I definitely understand the support for KG as his statistical profile is elite, not only for his peak season but consistenly so. While I get that some people come to the conclusion that KG not being voted in yet is ring bias, I don't agree with it. A purely statistical comparison is always going to be skewed more towards the regular season. Some stats don't even have play-off ratings and the ones that do have full season ratings don't weigh play-off games any heavier than regular season games, which makes these stats 80-90% reliant on the regular season.

You could say post-season success is highly dependent on teammates and the regular season gives a more even sample size for everyone and you'd be right. However, play-off basketball is where every team is locked in, where defense suddenly matters again and where every game is must win. In 2004 KG had an all-time regular season but I don't think he had an all-time play-off run. This isn't even to do with that the Timberwovles didn't end up winning a ring or even making the finals, this is just about individual performance. He never really had really bad games but about half of his outings in the 2004 post-season were decent at best. Lots of turnovers as well didn't help.

Let me note 2 thing:

1. The +/- data for Garnett in those playoffs continues to make him look extremely impactful. This data, as I'd hope goes without saying but I know doesn't, doesn't make every other piece of information irrelevant, but I doubt you'll find any star getting to the finals when his teammates put up a -24.2 per 100 possessions in the minutes without him.

2. If we're looking at box score to have a sense of underperformance, I'd note that the Timberwolves performed better offensively against that Laker team than any of the Lakers other opponents, and did so despite the fact that Cassell was injured and was only the #10 man on the team in that series. So if there's any sense that Garnett should have been more Duncan-like offensively, it really doesn't make a lot of sense when you compare the Wolves to what the Laker D did to the Spurs in the previous round.

Now, I know as I point to things like "their offense did better than anyone else's", it makes others say "we're talking all-time here so that doesn't cut it!". But when we can compare with the year favorable to in-prime players that are already voted in, I think we need to be careful lest we fall prey to a subtle wrinkle of winning bias.

If a star player receives a favorable set of matchups one year, then he's more likely to have his most effective games, which will tend to lead to more impressive stats for him and a better chance for his team winning. If this results in him leading his team the whole way, then that season will look extremely impressive naturally and that's just fine. But if the year's surrounding that year seem to tell a very different tale, then one needs to ask whether the difference was really that the player in question was at a unique apex in his golden year, or whether he was largely the same player for multiple years, and that one year represents how he looks when the gauntlet before him proves ideal to his strengths.

This to say, for anyone who voted Duncan in a long time ago, and who is still on the fence about Garnett, I think they need to ask themselves if they think Duncan was the more deserving MVP/POY in '03-04. If not, then I think it needs to be considered whether these players were actually close enough to each other in their prime that who looked better was dependent on context, and that perhaps what is seen as a superior peak from Duncan is simply him having a set of circumstances that year more fortunate than Garnett at the height of his powers ever did.

If this is in fact the case, I don't think it's wrong for someone to err on the side of picking the year that managed to be more ideal, but I do think it should make them hesitant to see clear tier differences between the two.

[SickMother]

01 Erving 75-76: 28.7 PER | .569 TS% | 110 TS+ | 17.7 WS | .262 WS/48
01 Erving 75-76 Playoffs?!?: 32.0 PER | .610 TS% | 3.7 WS | .321 WS/48
[a peak so high the NBA absorbed a whole other league to get this guy under their banner. Doctor turned in a top tier do it all regular season, then followed it up with one of thee largest postseason efficiency increases of all time.]

02 Hawkins 67-68: 28.8 PER | .597 TS% | 124 TS+ | 17.5 WS | .273 WS/48
02 Hawkins 67-68 Playoffs?!?: 30.0 PER | .651 TS% | 4.0 WS | .310 WS/48
[for what amounts to spot #13 on the list I'll go with maybe thee unluckiest player in basketball history, robbed of his collegiate & early NBA career by completely spurious gambling allegations, then derailed by a knee injury which occurred amidst the ABA Championship in this very season.

but there was never any doubt Connie could ball from the outset as one of the original NYC greats at Rucker Park & being named the best high school player in the country in 1960. As for his 67-68 peak, the Hawk simply did everything. Topped the brand new ABA in PPG on monster efficiency, 2nd in the league in RPG, 3rd in the league in APG. Then took his game to a whole other level in the playoffs, hurting his knee during the Championship series & returning to lead his team to the top in heroic fashion even before Willis Reed (or Giannis) did it.

I get it, the 67-68 ABA is probably the 2nd weakest competition level to receive a vote so far besides 49-50 Mikan, but Hawkins had no alternative, it was the best league available to him due to his illegal blacklisting from the NBA & he thoroughly dominated it across the board. Also think that Connie's peak game works in any era with his mix of size, athleticism and all around skillset.]

***I have a tier break here which goes from #14 to #25 with the following seasons under consideration (in chrono order) 63-64 Oscar, 65-66 West, 82-83 Moses, 94-95 Admiral, 03-04 Garnett, 05-06 Wade, 05-06 Nowitzki, 08-09 Kobe, 16-17 Kawhi, 16-17 Durant, 20-21 Giannis and 21-22 Jokic.***

03 Kawhi 16-17: 27.6 PER | .610 TS% | 111 TS+ | 13.6 WS | .264 WS/48
03 Kawhi 16-17 Playoffs?!?: 31.5 PER | .672 TS% | 2.8 WS | .314 WS/48
[ultimately going with Kawhi for two main reasons, his elite defense as a wing defender is unique among the remaining contenders, and that postseason was shaping up as a best ever candidate with Leonard posting absolutely insane efficiency before Zaza stepped in.]

[iggymcfrack]

1. 2003-04 Garnett: With KG on the bench, the Wolves were -10.9. That’s over 3.5 points worse than the worst team in the league. And yet KG was so dominant on both ends that he led Minnesota to the 1 seed over peak Tim Duncan playing with Parker and Ginobili and a super team of Shaq, Kobe, Malone, and Payton. He led the league in the box score composites while also putting up the best defensive impact season of the modern era. Sam Cassell was literally the only other above average player on the entire team and he got hurt early in the conference finals causing the Wolves to lose 4-2. If he’d just had one decent teammate to help him, KG could have led Minnesota to one of the most impressive individual rings of all-time.

2. 2021-22 Nikola Jokic: Best regular season ever by PER and BPM, 5th best postseason ever by PER with impact numbers that dominate the competition much more than the box score metrics. Incredible playmaking as Jokic combines the best passing from a big man ever with a surprisingly low amount of time holding the ball for such an offensive hub.

3. 2020-21 Giannis Antetokounmpo: Some of the biggest performances ever when it mattered most. 40/13/5 on .663 TS% in a Game 7 to close out KD and the Nets, then 3 more 40/10 games in the Finals including a ridiculous 50/14 game with 5 blocks on .749 TS% to close out the Suns.

Switched 2 and 3 after DraymondGold convinced me that Giannis's playoffs in 2021 were not spectacular enough to warrant him passing Jokic.

[falcolombardi]

I definitely understand the support for KG as his statistical profile is elite, not only for his peak season but consistenly so. While I get that some people come to the conclusion that KG not being voted in yet is ring bias, I don't agree with it. A purely statistical comparison is always going to be skewed more towards the regular season. Some stats don't even have play-off ratings and the ones that do have full season ratings don't weigh play-off games any heavier than regular season games, which makes these stats 80-90% reliant on the regular season.

You could say post-season success is highly dependent on teammates and the regular season gives a more even sample size for everyone and you'd be right. However, play-off basketball is where every team is locked in, where defense suddenly matters again and where every game is must win. In 2004 KG had an all-time regular season but I don't think he had an all-time play-off run. This isn't even to do with that the Timberwovles didn't end up winning a ring or even making the finals, this is just about individual performance. He never really had really bad games but about half of his outings in the 2004 post-season were decent at best. Lots of turnovers as well didn't help.

Let me note 2 thing:

1. The +/- data for Garnett in those playoffs continues to make him look extremely impactful. This data, as I'd hope goes without saying but I know doesn't, doesn't make every other piece of information irrelevant, but I doubt you'll find any star getting to the finals when his teammates put up a -24.2 per 100 possessions in the minutes without him.

2. If we're looking at box score to have a sense of underperformance, I'd note that the Timberwolves performed better offensively against that Laker team than any of the Lakers other opponents, and did so despite the fact that Cassell was injured and was only the #10 man on the team in that series. So if there's any sense that Garnett should have been more Duncan-like offensively, it really doesn't make a lot of sense when you compare the Wolves to what the Laker D did to the Spurs in the previous round.

Now, I know as I point to things like "their offense did better than anyone else's", it makes others say "we're talking all-time here so that doesn't cut it!". But when we can compare with the year favorable to in-prime players that are already voted in, I think we need to be careful lest we fall prey to a subtle wrinkle of winning bias.

If a star player receives a favorable set of matchups one year, then he's more likely to have his most effective games, which will tend to lead to more impressive stats for him and a better chance for his team winning. If this results in him leading his team the whole way, then that season will look extremely impressive naturally and that's just fine. But if the year's surrounding that year seem to tell a very different tale, then one needs to ask whether the difference was really that the player in question was at a unique apex in his golden year, or whether he was largely the same player for multiple years, and that one year represents how he looks when the gauntlet before him proves ideal to his strengths.

This to say, for anyone who voted Duncan in a long time ago, and who is still on the fence about Garnett, I think they need to ask themselves if they think Duncan was the more deserving MVP/POY in '03-04. If not, then I think it needs to be considered whether these players were actually close enough to each other in their prime that who looked better was dependent on context, and that perhaps what is seen as a superior peak from Duncan is simply him having a set of circumstances that year more fortunate than Garnett at the height of his powers ever did.

If this is in fact the case, I don't think it's wrong for someone to err on the side of picking the year that managed to be more ideal, but I do think it should make them hesitant to see clear tier differences between the two.

I know we have debated this before....but is not this reminiscent of paul vs curry?

Paul vs curry, like garnett vs duncan before, matches steph in impact metrics as well as team offense results with curry having the ring success

But i would guess you likely have like a 15-20 spots gap between them ( since you had curry 3rs overall and paul is still not in your rankings)

Not saying you are wrong to have curry so high over paul, just pointing out is not a thingh unique to "pro duncan" people in comparisions with garnett

[Doctor MJ]

I definitely understand the support for KG as his statistical profile is elite, not only for his peak season but consistenly so. While I get that some people come to the conclusion that KG not being voted in yet is ring bias, I don't agree with it. A purely statistical comparison is always going to be skewed more towards the regular season. Some stats don't even have play-off ratings and the ones that do have full season ratings don't weigh play-off games any heavier than regular season games, which makes these stats 80-90% reliant on the regular season.

You could say post-season success is highly dependent on teammates and the regular season gives a more even sample size for everyone and you'd be right. However, play-off basketball is where every team is locked in, where defense suddenly matters again and where every game is must win. In 2004 KG had an all-time regular season but I don't think he had an all-time play-off run. This isn't even to do with that the Timberwovles didn't end up winning a ring or even making the finals, this is just about individual performance. He never really had really bad games but about half of his outings in the 2004 post-season were decent at best. Lots of turnovers as well didn't help.

Let me note 2 thing:

1. The +/- data for Garnett in those playoffs continues to make him look extremely impactful. This data, as I'd hope goes without saying but I know doesn't, doesn't make every other piece of information irrelevant, but I doubt you'll find any star getting to the finals when his teammates put up a -24.2 per 100 possessions in the minutes without him.

2. If we're looking at box score to have a sense of underperformance, I'd note that the Timberwolves performed better offensively against that Laker team than any of the Lakers other opponents, and did so despite the fact that Cassell was injured and was only the #10 man on the team in that series. So if there's any sense that Garnett should have been more Duncan-like offensively, it really doesn't make a lot of sense when you compare the Wolves to what the Laker D did to the Spurs in the previous round.

Now, I know as I point to things like "their offense did better than anyone else's", it makes others say "we're talking all-time here so that doesn't cut it!". But when we can compare with the year favorable to in-prime players that are already voted in, I think we need to be careful lest we fall prey to a subtle wrinkle of winning bias.

If a star player receives a favorable set of matchups one year, then he's more likely to have his most effective games, which will tend to lead to more impressive stats for him and a better chance for his team winning. If this results in him leading his team the whole way, then that season will look extremely impressive naturally and that's just fine. But if the year's surrounding that year seem to tell a very different tale, then one needs to ask whether the difference was really that the player in question was at a unique apex in his golden year, or whether he was largely the same player for multiple years, and that one year represents how he looks when the gauntlet before him proves ideal to his strengths.

This to say, for anyone who voted Duncan in a long time ago, and who is still on the fence about Garnett, I think they need to ask themselves if they think Duncan was the more deserving MVP/POY in '03-04. If not, then I think it needs to be considered whether these players were actually close enough to each other in their prime that who looked better was dependent on context, and that perhaps what is seen as a superior peak from Duncan is simply him having a set of circumstances that year more fortunate than Garnett at the height of his powers ever did.

If this is in fact the case, I don't think it's wrong for someone to err on the side of picking the year that managed to be more ideal, but I do think it should make them hesitant to see clear tier differences between the two.

I know we have debated this before....but is not this reminiscent of paul vs curry?

Paul vs curry, like garnett vs duncan before, matches steph in impact metrics as well as team offense results with curry having the ring success

But i would guess you likely have like a 15-20 spots gap between them ( since you had curry 3rs overall and paul is still not in your rankings)

Not saying you are wrong to have curry so high over paul, just pointing out is not a thingh unique to "pro duncan" people in comparisions with garnett

I understand the similarities that make you draw the analogy, but I believe I've elaborate a good deal on differences though.

In a nutshell, I see Paul & Curry as too very different players whose games shouldn't be expected to scale all that similarly. People draw the connection between them because of their size, but I'd be more likely to compare Paul to Harden due to their reliance of on-ball dominance. It's not just that Curry has played for more successful teams, but that I think it's more clear how the way he plays helps achieve this compared to Paul or Harden.

Of course one can respond that they feel the same way about Garnett & Duncan - that Duncan just helps teams reach top tier status better than Garnett - and if they do, then naturally they'll side with Duncan and possibly side with him with tier separation.

I don't personally believe this to be the case however. I don't think an offense built around Duncan's post scoring necessarily scaled better than what Garnett was doing with what was around him back in the day, and certainly I'm more comfortable with Garnett's long-distance shooting, passing, and agility, all of which would give him an edge in a more sophisticated strategic era.

Defensively, I think Duncan did have more impact at the time, and I think that's something that can quite reasonably be used in the debate pertaining to this project...but I think that if you use prime Garnett as he was used in Boston, I think he has defensive impact at least comparable to peak Duncan did in the day, and once again his agility - along with his anticipation - only allow him to scale better to strategically superior competition.

[Dutchball97]

I definitely understand the support for KG as his statistical profile is elite, not only for his peak season but consistenly so. While I get that some people come to the conclusion that KG not being voted in yet is ring bias, I don't agree with it. A purely statistical comparison is always going to be skewed more towards the regular season. Some stats don't even have play-off ratings and the ones that do have full season ratings don't weigh play-off games any heavier than regular season games, which makes these stats 80-90% reliant on the regular season.

You could say post-season success is highly dependent on teammates and the regular season gives a more even sample size for everyone and you'd be right. However, play-off basketball is where every team is locked in, where defense suddenly matters again and where every game is must win. In 2004 KG had an all-time regular season but I don't think he had an all-time play-off run. This isn't even to do with that the Timberwovles didn't end up winning a ring or even making the finals, this is just about individual performance. He never really had really bad games but about half of his outings in the 2004 post-season were decent at best. Lots of turnovers as well didn't help.

Let me note 2 thing:

1. The +/- data for Garnett in those playoffs continues to make him look extremely impactful. This data, as I'd hope goes without saying but I know doesn't, doesn't make every other piece of information irrelevant, but I doubt you'll find any star getting to the finals when his teammates put up a -24.2 per 100 possessions in the minutes without him.

2. If we're looking at box score to have a sense of underperformance, I'd note that the Timberwolves performed better offensively against that Laker team than any of the Lakers other opponents, and did so despite the fact that Cassell was injured and was only the #10 man on the team in that series. So if there's any sense that Garnett should have been more Duncan-like offensively, it really doesn't make a lot of sense when you compare the Wolves to what the Laker D did to the Spurs in the previous round.

Now, I know as I point to things like "their offense did better than anyone else's", it makes others say "we're talking all-time here so that doesn't cut it!". But when we can compare with the year favorable to in-prime players that are already voted in, I think we need to be careful lest we fall prey to a subtle wrinkle of winning bias.

If a star player receives a favorable set of matchups one year, then he's more likely to have his most effective games, which will tend to lead to more impressive stats for him and a better chance for his team winning. If this results in him leading his team the whole way, then that season will look extremely impressive naturally and that's just fine. But if the year's surrounding that year seem to tell a very different tale, then one needs to ask whether the difference was really that the player in question was at a unique apex in his golden year, or whether he was largely the same player for multiple years, and that one year represents how he looks when the gauntlet before him proves ideal to his strengths.

This to say, for anyone who voted Duncan in a long time ago, and who is still on the fence about Garnett, I think they need to ask themselves if they think Duncan was the more deserving MVP/POY in '03-04. If not, then I think it needs to be considered whether these players were actually close enough to each other in their prime that who looked better was dependent on context, and that perhaps what is seen as a superior peak from Duncan is simply him having a set of circumstances that year more fortunate than Garnett at the height of his powers ever did.

If this is in fact the case, I don't think it's wrong for someone to err on the side of picking the year that managed to be more ideal, but I do think it should make them hesitant to see clear tier differences between the two.

The last paragraph is interesting because that is exactly what causes me to have KG a bit lower than the average. To me 04 KG looks amazing in the regular season and good to great in the post-season (exact same scenario for Robinson btw). I don't just pick title winners as I wouldn't have someone like Elvin Hayes or Billups in their title years ahead of KG as they were not the same caliber players but when it comes to the likes of Dr J and Moses they were both the best players in the league in their respective peak years. When it is reasonably close and especially when comparing 1 or more players that played in an era where we have limited stats I gravitate towards the most succesful seasons.

[Doctor MJ]

I definitely understand the support for KG as his statistical profile is elite, not only for his peak season but consistenly so. While I get that some people come to the conclusion that KG not being voted in yet is ring bias, I don't agree with it. A purely statistical comparison is always going to be skewed more towards the regular season. Some stats don't even have play-off ratings and the ones that do have full season ratings don't weigh play-off games any heavier than regular season games, which makes these stats 80-90% reliant on the regular season.

You could say post-season success is highly dependent on teammates and the regular season gives a more even sample size for everyone and you'd be right. However, play-off basketball is where every team is locked in, where defense suddenly matters again and where every game is must win. In 2004 KG had an all-time regular season but I don't think he had an all-time play-off run. This isn't even to do with that the Timberwovles didn't end up winning a ring or even making the finals, this is just about individual performance. He never really had really bad games but about half of his outings in the 2004 post-season were decent at best. Lots of turnovers as well didn't help.

Let me note 2 thing:

1. The +/- data for Garnett in those playoffs continues to make him look extremely impactful. This data, as I'd hope goes without saying but I know doesn't, doesn't make every other piece of information irrelevant, but I doubt you'll find any star getting to the finals when his teammates put up a -24.2 per 100 possessions in the minutes without him.

2. If we're looking at box score to have a sense of underperformance, I'd note that the Timberwolves performed better offensively against that Laker team than any of the Lakers other opponents, and did so despite the fact that Cassell was injured and was only the #10 man on the team in that series. So if there's any sense that Garnett should have been more Duncan-like offensively, it really doesn't make a lot of sense when you compare the Wolves to what the Laker D did to the Spurs in the previous round.

Now, I know as I point to things like "their offense did better than anyone else's", it makes others say "we're talking all-time here so that doesn't cut it!". But when we can compare with the year favorable to in-prime players that are already voted in, I think we need to be careful lest we fall prey to a subtle wrinkle of winning bias.

If a star player receives a favorable set of matchups one year, then he's more likely to have his most effective games, which will tend to lead to more impressive stats for him and a better chance for his team winning. If this results in him leading his team the whole way, then that season will look extremely impressive naturally and that's just fine. But if the year's surrounding that year seem to tell a very different tale, then one needs to ask whether the difference was really that the player in question was at a unique apex in his golden year, or whether he was largely the same player for multiple years, and that one year represents how he looks when the gauntlet before him proves ideal to his strengths.

This to say, for anyone who voted Duncan in a long time ago, and who is still on the fence about Garnett, I think they need to ask themselves if they think Duncan was the more deserving MVP/POY in '03-04. If not, then I think it needs to be considered whether these players were actually close enough to each other in their prime that who looked better was dependent on context, and that perhaps what is seen as a superior peak from Duncan is simply him having a set of circumstances that year more fortunate than Garnett at the height of his powers ever did.

If this is in fact the case, I don't think it's wrong for someone to err on the side of picking the year that managed to be more ideal, but I do think it should make them hesitant to see clear tier differences between the two.

The last paragraph is interesting because that is exactly what causes me to have KG a bit lower than the average. To me 04 KG looks amazing in the regular season and good to great in the post-season (exact same scenario for Robinson btw). I don't just pick title winners as I wouldn't have someone like Elvin Hayes or Billups in their title years ahead of KG as they were not the same caliber players but when it comes to the likes of Dr J and Moses they were both the best players in the league in their respective peak years. When it is reasonably close and especially when comparing 1 or more players that played in an era where we have limited stats I gravitate towards the most succesful seasons.

Does this mean you have many players between Duncan & Garnett, or no? Can't tell from what you say.

If it ends up being simply a tiebreak between contemporaries, that's one thing, if you end up putting one guy over another because he had a golden year in his era and the other did not, that's quite another. At that point it ceases to become a tiebreak, and starts to risk taking effect for the cause early in your analysis.

[Dutchball97]

Let me note 2 thing:

1. The +/- data for Garnett in those playoffs continues to make him look extremely impactful. This data, as I'd hope goes without saying but I know doesn't, doesn't make every other piece of information irrelevant, but I doubt you'll find any star getting to the finals when his teammates put up a -24.2 per 100 possessions in the minutes without him.

2. If we're looking at box score to have a sense of underperformance, I'd note that the Timberwolves performed better offensively against that Laker team than any of the Lakers other opponents, and did so despite the fact that Cassell was injured and was only the #10 man on the team in that series. So if there's any sense that Garnett should have been more Duncan-like offensively, it really doesn't make a lot of sense when you compare the Wolves to what the Laker D did to the Spurs in the previous round.

Now, I know as I point to things like "their offense did better than anyone else's", it makes others say "we're talking all-time here so that doesn't cut it!". But when we can compare with the year favorable to in-prime players that are already voted in, I think we need to be careful lest we fall prey to a subtle wrinkle of winning bias.

If a star player receives a favorable set of matchups one year, then he's more likely to have his most effective games, which will tend to lead to more impressive stats for him and a better chance for his team winning. If this results in him leading his team the whole way, then that season will look extremely impressive naturally and that's just fine. But if the year's surrounding that year seem to tell a very different tale, then one needs to ask whether the difference was really that the player in question was at a unique apex in his golden year, or whether he was largely the same player for multiple years, and that one year represents how he looks when the gauntlet before him proves ideal to his strengths.

This to say, for anyone who voted Duncan in a long time ago, and who is still on the fence about Garnett, I think they need to ask themselves if they think Duncan was the more deserving MVP/POY in '03-04. If not, then I think it needs to be considered whether these players were actually close enough to each other in their prime that who looked better was dependent on context, and that perhaps what is seen as a superior peak from Duncan is simply him having a set of circumstances that year more fortunate than Garnett at the height of his powers ever did.

If this is in fact the case, I don't think it's wrong for someone to err on the side of picking the year that managed to be more ideal, but I do think it should make them hesitant to see clear tier differences between the two.

The last paragraph is interesting because that is exactly what causes me to have KG a bit lower than the average. To me 04 KG looks amazing in the regular season and good to great in the post-season (exact same scenario for Robinson btw). I don't just pick title winners as I wouldn't have someone like Elvin Hayes or Billups in their title years ahead of KG as they were not the same caliber players but when it comes to the likes of Dr J and Moses they were both the best players in the league in their respective peak years. When it is reasonably close and especially when comparing 1 or more players that played in an era where we have limited stats I gravitate towards the most succesful seasons.

Does this mean you have many players between Duncan & Garnett, or no? Can't tell from what you say.

If it ends up being simply a tiebreak between contemporaries, that's one thing, if you end up putting one guy over another because he had a golden year in his era and the other did not, that's quite another. At that point it ceases to become a tiebreak, and starts to risk taking effect for the cause early in your analysis.

I have quite a few players between Duncan and KG. I voted Duncan when he got in and KG is still just outside my ballot. This circles back to what I said in the beginning that most differences come down to a difference in approach. I'm mainly looking for the best seasons by a player instead of trying to find the best player in a vacuum. I have Duncan (and a few others) ahead of KG because they had more "complete" peak seasons. I don't view them as being on another tier. I think it is plausible that KG could've done as well or even better than Moses and Dr J if he was in a similar situation but we can't say for sure that he could.

[iggymcfrack]

I definitely understand the support for KG as his statistical profile is elite, not only for his peak season but consistenly so. While I get that some people come to the conclusion that KG not being voted in yet is ring bias, I don't agree with it. A purely statistical comparison is always going to be skewed more towards the regular season. Some stats don't even have play-off ratings and the ones that do have full season ratings don't weigh play-off games any heavier than regular season games, which makes these stats 80-90% reliant on the regular season.

You could say post-season success is highly dependent on teammates and the regular season gives a more even sample size for everyone and you'd be right. However, play-off basketball is where every team is locked in, where defense suddenly matters again and where every game is must win. In 2004 KG had an all-time regular season but I don't think he had an all-time play-off run. This isn't even to do with that the Timberwovles didn't end up winning a ring or even making the finals, this is just about individual performance. He never really had really bad games but about half of his outings in the 2004 post-season were decent at best. Lots of turnovers as well didn't help.

Let me note 2 thing:

1. The +/- data for Garnett in those playoffs continues to make him look extremely impactful. This data, as I'd hope goes without saying but I know doesn't, doesn't make every other piece of information irrelevant, but I doubt you'll find any star getting to the finals when his teammates put up a -24.2 per 100 possessions in the minutes without him.

2. If we're looking at box score to have a sense of underperformance, I'd note that the Timberwolves performed better offensively against that Laker team than any of the Lakers other opponents, and did so despite the fact that Cassell was injured and was only the #10 man on the team in that series. So if there's any sense that Garnett should have been more Duncan-like offensively, it really doesn't make a lot of sense when you compare the Wolves to what the Laker D did to the Spurs in the previous round.

Now, I know as I point to things like "their offense did better than anyone else's", it makes others say "we're talking all-time here so that doesn't cut it!". But when we can compare with the year favorable to in-prime players that are already voted in, I think we need to be careful lest we fall prey to a subtle wrinkle of winning bias.

If a star player receives a favorable set of matchups one year, then he's more likely to have his most effective games, which will tend to lead to more impressive stats for him and a better chance for his team winning. If this results in him leading his team the whole way, then that season will look extremely impressive naturally and that's just fine. But if the year's surrounding that year seem to tell a very different tale, then one needs to ask whether the difference was really that the player in question was at a unique apex in his golden year, or whether he was largely the same player for multiple years, and that one year represents how he looks when the gauntlet before him proves ideal to his strengths.

This to say, for anyone who voted Duncan in a long time ago, and who is still on the fence about Garnett, I think they need to ask themselves if they think Duncan was the more deserving MVP/POY in '03-04. If not, then I think it needs to be considered whether these players were actually close enough to each other in their prime that who looked better was dependent on context, and that perhaps what is seen as a superior peak from Duncan is simply him having a set of circumstances that year more fortunate than Garnett at the height of his powers ever did.

If this is in fact the case, I don't think it's wrong for someone to err on the side of picking the year that managed to be more ideal, but I do think it should make them hesitant to see clear tier differences between the two.

This is an excellent post. KG and Duncan are really the ultimate what-if in that they had remarkably similar skills and talent and were born less than a month apart, but one was drafted into the best possible situation in the NBA while the other was drafted into the worst. You can't help but wonder, if the situation were reversed, would KG be just as successful in San Antonio as Duncan was if not moreso. Yeah, KG slipped in the mid-2000s for a while when the Wolves were completely stripped of any other remotely talented basketball players, but can you really blame him? Would Duncan have given the same effort next to Ricky Davis, Mark Blount, Trenton Hassell, and Mike James in the starting lineup as he did next to Manu and Tony? I think Duncan deserves a slight edge, but it's very slight and I'd consider KG much closer to Timmy and Hakeem than to the people who have been drafted since.

[iggymcfrack]

The last paragraph is interesting because that is exactly what causes me to have KG a bit lower than the average. To me 04 KG looks amazing in the regular season and good to great in the post-season (exact same scenario for Robinson btw). I don't just pick title winners as I wouldn't have someone like Elvin Hayes or Billups in their title years ahead of KG as they were not the same caliber players but when it comes to the likes of Dr J and Moses they were both the best players in the league in their respective peak years. When it is reasonably close and especially when comparing 1 or more players that played in an era where we have limited stats I gravitate towards the most succesful seasons.

Does this mean you have many players between Duncan & Garnett, or no? Can't tell from what you say.

If it ends up being simply a tiebreak between contemporaries, that's one thing, if you end up putting one guy over another because he had a golden year in his era and the other did not, that's quite another. At that point it ceases to become a tiebreak, and starts to risk taking effect for the cause early in your analysis.

I have quite a few players between Duncan and KG. I voted Duncan when he got in and KG is still just outside my ballot. This circles back to what I said in the beginning that most differences come down to a difference in approach. I'm mainly looking for the best seasons by a player instead of trying to find the best player in a vacuum. I have Duncan (and a few others) ahead of KG because they had more "complete" peak seasons. I don't view them as being on another tier. I think it is plausible that KG could've done as well or even better than Moses and Dr J if he was in a similar situation but we can't say for sure that he could.

This was years ago so it wouldn't encompass all the modern seasons, but at one point Haralabob (top sports bettor ever) said that 2004 Kevin Garnett was the best season he had for anyone in his entire database. His impact stats and simple on/off are absolutely historic for that one season. Probably better than anyone else we have since those numbers have been available. I actually feel like grouping more seasons together makes it easier for KG to slip while looking at that one season in a vacuum makes KG all the more impressive.

In the Western Conference Finals that season, the teammates who played the most minutes for Shaq were Kobe, Karl Malone, and Gary Payton. KG's teammates who played the most were Latrell Sprewell (.493 TS% for the season, would be out of the league in 2 years), Wally Szczerbiak (only 2 more years in the league as a starter), and Fred Hoiberg (out of the league 2 years later at age 33). I think the fact that KG accomplished as much as he did that season is incredible and the only non-Finals season that would be better or even comparable is LeBron in '09.

[MyUniBroDavis]

Is there a way to sort team on-off by round for pre 2008 seasons?

I’m curious to see KGs round by round, I think a lot of the time round by round analysis on plus minus is noisy especially in that era where the minutes with players off the court is so minuscule, but I’d be curious to see it round by round if it was constant or not (constantly ATG)

[eminence]

Is there a way to sort team on-off by round for pre 2008 seasons?

I’m curious to see KGs round by round, I think a lot of the time round by round analysis on plus minus is noisy especially in that era where the minutes with players off the court is so minuscule, but I’d be curious to see it round by round if it was constant or not (constantly ATG)

KG only made it out of the first round once pre 2008 ('04), I really don't think it'd be worth the time.

[MyUniBroDavis]

Is there a way to sort team on-off by round for pre 2008 seasons?

I’m curious to see KGs round by round, I think a lot of the time round by round analysis on plus minus is noisy especially in that era where the minutes with players off the court is so minuscule, but I’d be curious to see it round by round if it was constant or not (constantly ATG)

KG only made it out of the first round once pre 2008 ('04), I really don't think it'd be worth the time.

I just meant for 04 specifically lol

[Proxy]

Is there a way to sort team on-off by round for pre 2008 seasons?

I’m curious to see KGs round by round, I think a lot of the time round by round analysis on plus minus is noisy especially in that era where the minutes with players off the court is so minuscule, but I’d be curious to see it round by round if it was constant or not (constantly ATG)

I feel like you could probably find this with an opponent filter on pbpstats.com but if u try and can't find it, lmk, and I could try looking for an alternative

[eminence]

Is there a way to sort team on-off by round for pre 2008 seasons?

I’m curious to see KGs round by round, I think a lot of the time round by round analysis on plus minus is noisy especially in that era where the minutes with players off the court is so minuscule, but I’d be curious to see it round by round if it was constant or not (constantly ATG)

KG only made it out of the first round once pre 2008 ('04), I really don't think it'd be worth the time.

I just meant for 04 specifically lol

I did per game by hand, apologies if I mixed anything up. Rounded to nearest minute for minutes played.

Round 1 vs Nuggets
41 mpg
+11.6 +/- per game on court
-7.6 +/- per game off court

Round 2 vs Kings
45 mpg (1 overtime game)
-1.3 +/- per game on court
+0.1 +/0 per game off court

WCF vs Lakers
44 mpg
-1.5 +/- per game on court
-0.7 +/- per game off court

4 mpg offcourt in the 2nd/3rd rounds, peaking in game 4 vs the Lakers where he played all but 31 seconds.

So on a per possession basis they were unbelievably better with KG oncourt in round 1 (in the realm of +55 per 100 if pace was similar on vs off).

Very slightly better with KG off vs the Kings (-2 range)

Around +6 vs the Lakers.

[MyUniBroDavis]

Is there a way to sort team on-off by round for pre 2008 seasons?

I’m curious to see KGs round by round, I think a lot of the time round by round analysis on plus minus is noisy especially in that era where the minutes with players off the court is so minuscule, but I’d be curious to see it round by round if it was constant or not (constantly ATG)

I feel like you could probably find this with an opponent filter on pbpstats.com but if u try and can't find it, lmk, and I could try looking for an alternative

I went through every tab on pbp and am so confused how this works,‘might be because I’m on the onos

[jalengreen]

I tried to use pbpstats.com game logs to run the calculations. Fair warning that it's quite possible that I messed something up, some of these numbers are weird lol



We do have the overall postseason on/off from pbpstats which I included in the Total (Actual) row. I included the overall postseason values you'd get using the same game log technique in the Total (Calculated) row to show that there's a slight margin of error for the defensive rating figures, but nothing massive.

Should note that Garnett played a lot of minutes, so these aren't huge samples we're talking about at all. But as for whether the on/off is constant throughout the postseason ... doesn't seem like it.

[eminence]

I tried to use pbpstats.com game logs to run the calculations. Fair warning that it's quite possible that I messed something up, some of these numbers are weird lol



We do have the overall postseason on/off from pbpstats which I included in the Total (Actual) row. I included the overall postseason values you'd get using the same game log technique in the Total (Calculated) row to show that there's a slight margin of error for the defensive rating figures, but nothing massive.

Should note that Garnett played a lot of minutes, so these aren't huge samples we're talking about at all. But as for whether the on/off is constant throughout the postseason ... doesn't seem like it.

This looks solidly more sound than my rough version, good work.

Would you mind doing any of the deeper runs by Wade/Kobe/AD/Kawhi/Giannis (modern guys I've seen with a little traction) for a baseline comparison to how variable those types of numbers can be?

I understand if it takes too much time to be worth it as well

On the actual KG numbers -

The on/off variance there looks to be mostly (understatement) in the wildly variable off period, not too surprising. Wolves with KG ran the Nuggets over and then played the Kings/Lakers pretty even.