4year Peaks Study using nbarapm.com

[trelos6]

I pulled just the basic RAPM data for 34 players. Here's the list of top 4 year increments.

Player	end szn	off	def	rapm
Kevin Garnett	2009	3.9	6.8	10.6
Lebron James	2012	6.6	4	10.5
Lebron James	2010	6.8	3.3	10.2
Nikola Jokic	2025	8.4	1.7	10.1
Lebron James	2011	6.4	3.4	9.8
Kevin Garnett	2010	3.6	6.2	9.8
Lebron James	2017	7.8	1.9	9.7
Lebron James	2016	7.3	2.4	9.7
Stephen Curry	2018	8.6	0.5	9.2
Lebron James	2015	7.1	2	9.1
Chris Paul	2018	6.4	2.7	9.1
Kevin Garnett	2006	5.5	3.4	9
Stephen Curry	2020	8.3	0.7	9
Lebron James	2013	6.2	2.7	8.9
Lebron James	2009	6.3	2.6	8.9
Kawhi Leonard	2023	6.7	2.1	8.9
Kevin Garnett	2005	5.2	3.7	8.9
Stephen Curry	2019	8.3	0.6	8.9
Tim Duncan	2005	3.8	5.1	8.9
Chris Paul	2017	5.2	3.6	8.8
Steve Nash	2008	9.3	-0.6	8.8
Chris Paul	2016	5.7	2.9	8.6
Kevin Garnett	2004	5	3.7	8.6
Lebron James	2020	6.7	1.7	8.4
Kevin Garnett	2011	1.7	6.7	8.4
Lebron James	2018	7.7	0.5	8.3
Kevin Garnett	2008	3.5	4.7	8.2
Tim Duncan	2003	3.9	4.3	8.2
Steve Nash	2010	8.9	-0.8	8.1
Manu Ginobili	2008	5.3	2.6	7.9
Nikola Jokic	2024	6.8	1.2	7.9
Stephen Curry	2017	8	-0.1	7.9
Tim Duncan	2006	3.6	4.3	7.9
Tim Duncan	2004	3.3	4.7	7.9
Kawhi Leonard	2018	4.8	3	7.8
Paul George	2022	3.5	4.3	7.8
Shaquille O'Neal	2004	7.2	0.6	7.8
Joel Embiid	2024	4.7	3	7.7
Lebron James	2019	6.9	0.7	7.6
Chris Paul	2019	4.4	3.2	7.6
Dirk Nowitzki	2004	6.4	1.2	7.6
Steve Nash	2009	8.6	-0.9	7.6
Tim Duncan	2007	3.8	3.9	7.6
Manu Ginobili	2014	5.2	2.3	7.5
Shaquille O'Neal	2005	6.1	1.4	7.5
Stephen Curry	2022	5.7	1.8	7.5
Dirk Nowitzki	2014	5.7	1.7	7.4
Manu Ginobili	2007	5.6	1.8	7.4
Shaquille O'Neal	2000	5.4	2	7.4
Giannis Antetokounmpo	2022	2.6	4.7	7.3
Paul George	2024	4.4	2.9	7.3
Shaquille O'Neal	2003	7.6	-0.2	7.3
Tim Duncan	2008	4.3	3	7.3
Kevin Garnett	2012	1.4	5.8	7.2
Lebron James	2021	5.4	1.7	7.1
Chris Paul	2020	4.3	2.7	7
Joel Embiid	2019	3.1	3.8	7
Kevin Garnett	2003	4.4	2.6	7
Shaquille O'Neal	2001	6.7	0.3	7
Manu Ginobili	2006	4.8	2	6.9
Kawhi Leonard	2017	4.1	2.8	6.9
Steve Nash	2011	7.2	-0.4	6.9
Paul George	2021	3.5	3.3	6.9
Stephen Curry	2015	6.3	0.6	6.9
Lebron James	2022	4.8	2	6.8
Chris Paul	2015	4.8	2	6.8
Dirk Nowitzki	2011	5.1	1.7	6.8
Stephen Curry	2016	6.7	0.1	6.8
Jayson Tatum	2023	4.9	1.9	6.8
Dwyane Wade	2009	6.1	0.7	6.8
Chris Paul	2014	4.9	1.7	6.7
Dirk Nowitzki	2006	5.5	1.2	6.7
Giannis Antetokounmpo	2023	2.9	3.9	6.7
Kawhi Leonard	2024	5.1	1.5	6.7
Lebron James	2008	4.9	1.6	6.6
Chris Paul	2021	3.6	3	6.6
Dirk Nowitzki	2013	4.7	1.8	6.6
Kevin Garnett	2007	3.8	2.8	6.6
Chris Paul	2011	5.5	1	6.5
Joel Embiid	2021	2.3	4.2	6.5
Joel Embiid	2020	1.8	4.6	6.5
Nikola Jokic	2023	5.8	0.7	6.5
Nikola Jokic	2018	5.1	1.4	6.5
Kevin Durant	2022	5	1.5	6.5
Kevin Durant	2016	5.1	1.4	6.5
Tim Duncan	2002	2.6	3.9	6.5
Tim Duncan	2001	3.1	3.4	6.5
Lebron James	2023	4.9	1.5	6.4
James Harden	2018	6.8	-0.4	6.4
Kevin Garnett	2013	0.2	6.2	6.4
Dirk Nowitzki	2005	5.4	0.9	6.3
Joel Embiid	2022	2.9	3.5	6.3
Giannis Antetokounmpo	2021	1.4	5	6.3
Giannis Antetokounmpo	2020	1.8	4.5	6.3
Nikola Jokic	2019	4.5	1.8	6.3
Steve Nash	2007	7.7	-1.4	6.3
Shaquille O'Neal	2002	6.5	-0.2	6.3
Stephen Curry	2021	6.3	0.1	6.3
Jayson Tatum	2022	3.7	2.7	6.3
Dwyane Wade	2010	6.3	0	6.3
Lebron James	2014	5.3	0.9	6.2
Manu Ginobili	2009	4.2	1.9	6.2
Manu Ginobili	2005	3.9	2.3	6.2
James Harden	2017	6.4	-0.2	6.2
Jimmy Butler	2020	3.1	3	6.1
Dirk Nowitzki	2007	5.4	0.7	6.1
Giannis Antetokounmpo	2025	3.8	2.3	6.1
Kawhi Leonard	2022	5.3	0.8	6.1
Dirk Nowitzki	2012	4.6	1.4	6
Giannis Antetokounmpo	2024	3.1	2.9	6
Kevin Durant	2021	5.6	0.4	6
Paul George	2020	3	3	6
Shaquille O'Neal	2006	5.2	0.7	6
Dwyane Wade	2012	6	0	6
Manu Ginobili	2011	4	1.9	5.9
Kawhi Leonard	2019	4.3	1.6	5.9
Kevin Durant	2017	4.9	1	5.9
Jason Kidd	2005	4.3	1.5	5.9
Chris Paul	2012	4.4	1.3	5.8
Manu Ginobili	2015	3.8	2	5.8
Rudy Gobert	2022	1.1	4.7	5.8
Nikola Jokic	2020	5.2	0.6	5.8
Kawhi Leonard	2020	4.7	1.1	5.8
Steve Nash	2012	6.1	-0.3	5.8
Jimmy Butler	2021	3.4	2.4	5.7
Dirk Nowitzki	2003	4.9	0.8	5.7
Manu Ginobili	2016	2.5	3.2	5.7
Manu Ginobili	2013	4.7	1	5.7
James Harden	2012	4.8	0.8	5.7
Nikola Jokic	2017	4.4	1.3	5.7
Kawhi Leonard	2025	4	1.7	5.7
Kobe Bryant	2008	6.2	-0.5	5.7
Paul Pierce	2008	5.1	0.7	5.7
Stephen Curry	2023	5	0.7	5.7
Dirk Nowitzki	2008	5.3	0.3	5.6
Manu Ginobili	2017	2.9	2.7	5.6
Kobe Bryant	2009	6.1	-0.4	5.6
Paul George	2023	2.7	2.9	5.6
Paul Pierce	2005	4.5	1.2	5.6
Jayson Tatum	2021	2.9	2.6	5.6
Lebron James	2024	4.3	1.2	5.5
Chris Paul	2022	2	3.5	5.5
Chris Paul	2013	4.7	0.8	5.5
Manu Ginobili	2010	3.4	2.1	5.5
Kevin Durant	2023	3.9	1.6	5.5
Russell Westbrook	2017	6.1	-0.5	5.5
Joel Embiid	2023	2.7	2.7	5.4
Rudy Gobert	2021	0.9	4.5	5.4
Kawhi Leonard	2021	4.1	1.3	5.4
Kobe Bryant	2010	5.1	0.3	5.4
Jimmy Butler	2019	1.9	3.4	5.3
Dwight Howard	2011	2.7	2.6	5.3
Kawhi Leonard	2016	1.4	3.8	5.3
Kevin Garnett	2000	3	2.3	5.3
Russell Westbrook	2016	5.4	-0.1	5.3
Jimmy Butler	2018	2.4	2.8	5.2
Dirk Nowitzki	2009	5.2	0	5.2
Manu Ginobili	2012	4.8	0.4	5.2
Rudy Gobert	2019	1.1	4.2	5.2
Nikola Jokic	2022	4.8	0.4	5.2
Jayson Tatum	2024	4.4	0.8	5.2
James Harden	2019	5.5	-0.4	5.1
Jason Kidd	2007	4.2	0.9	5.1
Kobe Bryant	2007	5.6	-0.5	5.1
Tim Duncan	2010	2.4	2.7	5.1
Dwyane Wade	2011	5.6	-0.4	5.1
Lebron James	2007	3.6	1.3	5
Tim Duncan	2009	3.1	1.9	5
Ray Allen	2004	5.4	-0.5	4.9
Rudy Gobert	2020	0.9	4	4.9
Kevin Durant	2024	3.8	1.2	4.9
Kevin Durant	2019	5	-0.1	4.9
Kevin Durant	2013	3.3	1.5	4.9
Kevin Garnett	2014	-0.5	5.5	4.9
Shaquille O'Neal	2007	4.1	0.9	4.9
Jayson Tatum	2020	2.2	2.7	4.9
Russell Westbrook	2015	5	-0.1	4.9
Rudy Gobert	2024	0.2	4.6	4.8
James Harden	2020	5.5	-0.7	4.8
Paul George	2019	2.1	2.6	4.8
Dwyane Wade	2008	5.3	-0.4	4.8
Russell Westbrook	2018	5.5	-0.7	4.8
Lebron James	2006	4.2	0.5	4.7
Ben Wallace	2001	0.9	3.8	4.7
Jimmy Butler	2022	3.1	1.6	4.7
Chris Paul	2023	2	2.7	4.7
Chris Paul	2010	5	-0.3	4.7
Dirk Nowitzki	2010	4.2	0.5	4.7
Dirk Nowitzki	2002	4.6	0.1	4.7
Joel Embiid	2025	2	2.7	4.7
Kevin Durant	2025	3.7	1	4.7
Jason Kidd	2004	1.9	2.8	4.7
Jason Kidd	2002	2.7	2	4.7
Stephen Curry	2024	4.6	0.1	4.7
Stephen Curry	2014	4.9	-0.2	4.7
James Harden	2021	5.6	-1	4.6
James Harden	2011	3.7	0.9	4.6
Jason Kidd	2006	3.7	0.9	4.6
Tim Duncan	2000	1.2	3.4	4.6
Ray Allen	2006	5.9	-1.4	4.5
Kevin Durant	2018	4.6	-0.2	4.5
Steve Nash	2006	6.7	-2.3	4.5
Stephen Curry	2025	5.3	-0.8	4.5
Jayson Tatum	2018	2.2	2.3	4.5
Dwyane Wade	2007	4.7	-0.2	4.5
Ben Wallace	2004	0.1	4.2	4.4
Jimmy Butler	2024	3.1	1.3	4.4
Dwight Howard	2012	0.5	3.9	4.4
Gary Payton	2000	3.7	0.8	4.4
Paul George	2017	2.1	2.3	4.4
Paul Pierce	2007	4	0.4	4.4
Paul Pierce	2002	3.7	0.7	4.4
Jayson Tatum	2025	3.1	1.2	4.4
Ray Allen	2003	4.7	-0.4	4.3
Dirk Nowitzki	2015	4.1	0.2	4.3
Dirk Nowitzki	2001	3.3	1	4.3
Rudy Gobert	2023	-0.4	4.7	4.3
Rudy Gobert	2018	-0.2	4.6	4.3
Kevin Durant	2020	4.9	-0.6	4.3
Kobe Bryant	2006	5.5	-1.2	4.3
Paul Pierce	2003	3.7	0.6	4.3
Jayson Tatum	2019	2.3	1.9	4.3
Dwyane Wade	2013	4.8	-0.5	4.3
Tim Duncan	2013	0.7	3.5	4.2
Tracy McGrady	2004	5.4	-1.2	4.2
Ray Allen	2002	4.4	-0.4	4.1
Jimmy Butler	2017	1.7	2.4	4.1
Kevin Garnett	2002	2.3	1.7	4.1
Paul George	2016	0.8	3.4	4.1
Jason Kidd	2008	3.4	0.6	4
Jason Kidd	2003	2	2	4
Jason Kidd	2000	2.8	1.2	4
Paul Pierce	2004	3.8	0.2	4
Tim Duncan	2014	1.3	2.7	4

[lessthanjake]

One interesting thing about this is that Garnett’s multi-year RAPM peak comes from his later years, rather than the earlier years where he seemed to have been a better player. This is different than the other guys near the very top, who pretty much all have their RAPM peak in a timeframe where I think we’d generally think of as their peak years (with CP3 being another possible exception to that, if we consider his peak to have been on the Hornets).

I think there’s both a charitable and uncharitable way to interpret this. One might say that Garnett having such great RAPM when he was a bit past his best but was actually on a good team suggests that peak Garnett could’ve put up even higher RAPM numbers than that if he’d been on a good team back then. But another interpretation would be that Garnett surely wasn’t actually at his most impactful in those later years and therefore that that data is perhaps just a product of noise, and we should key in more on the data from his actual best years. I’m not sure where I stand on that (and the truth may be somewhere in the middle), but it feels interesting to me since we don’t see much of this with other top players. I suppose another interpretation is that people are wrong and he actually was a better player in those later years than he’d been earlier, but I’m very skeptical of that.

[ReggiesKnicks]

One interesting thing about this is that Garnett’s multi-year RAPM peak comes from his later years, rather than the earlier years where he seemed to have been a better player. This is different than the other guys near the very top, who pretty much all have their RAPM peak in a timeframe where I think we’d generally think of as their peak years (with CP3 being another possible exception to that, if we consider his peak to have been on the Hornets).

I think there’s both a charitable and uncharitable way to interpret this. One might say that Garnett having such great RAPM when he was a bit past his best but was actually on a good team suggests that peak Garnett could’ve put up even higher RAPM numbers than that if he’d been on a good team back then. But another interpretation would be that Garnett surely wasn’t actually at his impactful in those later years and therefore that that data is perhaps just a product of noise, and we should key in more on the data from his actual best years. I’m not sure where I stand on that (and the truth may be somewhere in the middle), but it feels interesting to me since we don’t see much of this with other top players. I suppose another interpretation is that people are wrong and he actually was a better player in those later years than he’d been earlier, but I’m very skeptical of that.

I think it is fairly easy to interpret and we don't need to overthink things.

Garnett was tasked with monumental responsibilities both offensively and defensive when he was on the Timberwolves.

When you have a team with Curry who can burden the offense and Draymond who can burden the defense, their impacts are easily captured. But, when you have someone who has primacy over both offense and defense to the degree Garnett did, his RAPM isn't going to be as high as when he is on a more balanced roster where his impact can be maximized in a more limited scope.

Another way to think of it is what would someone like Curry's RAPM be if he literally played in Draymond's role defensively? It would crater, right? After all the defense of the Warriors would be laughably bad if Curry was playing rim protector and switching everything.

Few players in the history of the NBA had to endure Garnett's offensive and defensive burdens like he did from 2000-2004 (let alone 2005-2007 when the roster fell off a cliff from bad to all-time bad).

Keying into the data from his best years is fine, but make sure you are applying proper context to the situation.

For me, Garnett's incredible RAPM data when he was on competently constructed teams is a "I told you so" moment about what could have been from 1998-2007. Given Garnett's intangibles and unique and diversified skill-set, there is nothing left to look at and ask "Could Garnett have done in his earlier years what he did in 2008 with better rosters?" Of course he could have.

[lessthanjake]

One interesting thing about this is that Garnett’s multi-year RAPM peak comes from his later years, rather than the earlier years where he seemed to have been a better player. This is different than the other guys near the very top, who pretty much all have their RAPM peak in a timeframe where I think we’d generally think of as their peak years (with CP3 being another possible exception to that, if we consider his peak to have been on the Hornets).

I think there’s both a charitable and uncharitable way to interpret this. One might say that Garnett having such great RAPM when he was a bit past his best but was actually on a good team suggests that peak Garnett could’ve put up even higher RAPM numbers than that if he’d been on a good team back then. But another interpretation would be that Garnett surely wasn’t actually at his impactful in those later years and therefore that that data is perhaps just a product of noise, and we should key in more on the data from his actual best years. I’m not sure where I stand on that (and the truth may be somewhere in the middle), but it feels interesting to me since we don’t see much of this with other top players. I suppose another interpretation is that people are wrong and he actually was a better player in those later years than he’d been earlier, but I’m very skeptical of that.

I think it is fairly easy to interpret and we don't need to overthink things.

Garnett was tasked with monumental responsibilities both offensively and defensive when he was on the Timberwolves.

When you have a team with Curry who can burden the offense and Draymond who can burden the defense, their impacts are easily captured. But, when you have someone who has primacy over both offense and defense to the degree Garnett did, his RAPM isn't going to be as high as when he is on a more balanced roster where his impact can be maximized in a more limited scope.

Another way to think of it is what would someone like Curry's RAPM be if he literally played in Draymond's role defensively? It would crater, right? After all the defense of the Warriors would be laughably bad if Curry was playing rim protector and switching everything.

Few players in the history of the NBA had to endure Garnett's offensive and defensive burdens like he did from 2000-2004 (let alone 2005-2007 when the roster fell off a cliff from bad to all-time bad).

Keying into the data from his best years is fine, but make sure you are applying proper context to the situation.

For me, Garnett's incredible RAPM data when he was on competently constructed teams is a "I told you so" moment about what could have been from 1998-2007. Given Garnett's intangibles and unique and diversified skill-set, there is nothing left to look at and ask "Could Garnett have done in his earlier years what he did in 2008 with better rosters?" Of course he could have.

I’m not sure that that conclusion is all that obvious. At least not to me. In general, I think that a team relying on someone more is a context in which we’d expect their impact to be higher, not lower. Of course, that can come into tension with the limits of human energy, where a guy who is relied on for everything probably can’t be as good at any single thing as they would be if they were more specialized in what the team was asking them to do. But it’s also not at all obvious to me that having multiple responsibilities that you do well will result in less impact than having one responsibility you do even better. Intuitively, one would think that the player who is focusing on one responsibility will have more impact on that area but less impact on other areas, and it would likely even out (or probably be worse than evening out, since there may be diminishing returns on energy in one particular area). And, under the circumstances in which the responsibility-for-one-thing timeframe is when the guy is older, we would probably expect less impact overall, because players don’t have as much energy in later years, so basically defense is getting a bigger portion of the pie but the pie is also a bit smaller.

[lessthanjake]

And, under the circumstances in which the responsibility-for-one-thing timeframe is when the guy is older, we would probably expect less impact overall, because players don’t have as much energy in later years, so basically defense is getting a bigger portion of the pie but the pie is also a bit smaller.

I would say we have such a clear change in roster construction around Garnett that it behooves one to say this.

I would agree if a player stayed in the same circumstance and simply aged. Curry is a great example of this. Same coach, system and 2nd wing (Green) for much of the prime. Curry had his best and more impactful season by RAPM in 2017-2019, his age 27-29 seasons, then started to see a decline.

Garnett saw his heights around ages 31-33.
Duncan saw his heights around ages 26-28.

The thing about all 3 of these players in this range is telling.

For Curry, he had an uber-elite second option in Durant and Draymond at the peak of his powers while in a system for 3-5 years. The role Curry played was clearly defined and laid out, with Durant there Curry's impact only increased even though he had less usage and less roles.

Do you see that? Curry had less offensive primacy and less of a usage and burden compared to his 2015 and 2016 seasons, and his impact per RAPM actually went up.

Now shift a dozen years earlier to Duncan. Duncan's overall impact is increasing he recalls parts of his offensive primacy, allowing the Spurs to foster the growth and blooming of Tony Parker and Manu Ginobili. The offensive burden for the Spurs is no longer solely on the shoulders of Duncan, yet his impact metrics continue to grow.

Why can't this same logic, a pattern we have seen in the best players ever, not apply to Garnett?

I think you’ve replied to this within the wrong thread, so I’ll respond back here again to take it back to the correct thread:

1. I think you’ve raised some reasons why it’s possible a player’s impact could go up in a more specialized role, but I also think it’s very possible for a player’s impact to go up as they are given more responsibility because it means the team is relying on them more (which can hurt how well the team does with a guy off the court). It can go either way, which makes it not at all obvious that that’s the explanation for Garnett’s impact numbers being higher in his later years. It’s certainly possible that in Minnesota he was simply having his offensive responsibility diluting his defensive impact. But it’s also possible that having the team rely on him on both ends was actually helping his impact, and the fact that his RAPM on the Celtics was even higher is a result of something else (noise being one option, but not the only one).

2. You draw analogies to Steph and Duncan, but I think they’re not necessarily factually right. I’ll start with Steph. You say Curry’s RAPM was at its best in 2017-2019. But by most measures, Steph’s best three years in terms of impact were 2015-2017. There’s a lot of RAPM measures and hybrid-RAPM measures out there, so there’s going to be some differences in what they all say, but if you look across the panoply of impact data we have, that’s the general picture you’ll find. I certainly don’t think it’d be valid to try to fashion an argument around an analogy that presupposes that impact data says Steph’s 2017-2019 looks materially more impactful than his 2015 and 2016 years.

3. As for the Duncan analogy, I don’t think that’s factually right either, but for a different reason. I do think it’s right that 2002-2005 looks like Duncan’s most impactful time period. But I definitely don’t see that as a time period that coincides with him having a more specialized role. It actually includes years where he had to take on the most offensive primacy of his career!

4. Notably, related to #2 above, I suppose I should do some due diligence to determine whether the full panoply of impact data we have says Garnett was more impactful in his later years, or whether it just happens to be the case in that one RAPM measure (I’m more familiar with Steph’s impact numbers off the top of my head than Garnett’s). If it’s the latter, then the explanation is probably just that measure randomly being higher on his later years. A quick perusal of xRAPM tells me that it has Garnett 2003-2006 looking a bit more impactful than any similar timespan later in his career. The RAPM on TheBasketballDatabase has Garnett’s RAPM peak being in his Timberwolves years, though it’s all pretty close. EPM has 2002-2005 looking like his best four-year span by a little bit. So yeah, maybe the answer to my inquiry is that Garnett’s impact numbers just aren’t actually better in his later years, despite one particular measure indicating otherwise.

[trelos6]

3 year RAPM peaks

Player	end szn	off	def	RAPM
Kevin Garnett	2009	4.2	6.4	10.6
Lebron James	2011	6.2	3.8	10
Lebron James	2017	7.6	2.2	9.9
Lebron James	2010	7	2.3	9.3
Lebron James	2009	6.1	3.2	9.3
Nikola Jokic	2025	8.3	1.1	9.3
Nikola Jokic	2024	7	2.3	9.3
Lebron James	2012	5.9	3.3	9.2
Kevin Garnett	2004	5.4	3.9	9.2
Kevin Garnett	2005	5.5	3.6	9
Kevin Garnett	2008	3.4	5.5	8.9
Stephen Curry	2018	8.5	0.4	8.9
Stephen Curry	2019	8	0.9	8.8
Lebron James	2016	6.4	2.3	8.7
Tim Duncan	2005	3.9	4.7	8.6
Lebron James	2015	6.8	1.6	8.4
Kawhi Leonard	2022	6.3	2.1	8.4
Chris Paul	2018	5.5	2.8	8.3
Chris Paul	2016	5.2	3.1	8.3
Tim Duncan	2003	4.1	4.2	8.2
Stephen Curry	2017	7.7	0.4	8.1
Lebron James	2018	8	-0.1	8
Steve Nash	2008	8.3	-0.4	8
Manu Ginobili	2007	5.8	2.1	8
Kevin Garnett	2010	1.9	5.9	7.9
Tim Duncan	2004	3.6	4.4	7.9
Kawhi Leonard	2018	4.9	2.9	7.8
Shaquille O'Neal	2004	6.3	1.3	7.7
Paul George	2021	4	3.7	7.7
Dirk Nowitzki	2013	5.4	2.2	7.7
Chris Paul	2017	5.2	2.4	7.6
Tim Duncan	2007	4.4	3.1	7.5
Steve Nash	2009	8.2	-0.7	7.5
Kawhi Leonard	2017	5.2	2.2	7.4
Steve Nash	2007	8.2	-0.7	7.4
Kyle Lowry	2019	5.2	2.2	7.4
Lebron James	2021	4	3.3	7.3
Kawhi Leonard	2023	6.3	0.9	7.2
Dirk Nowitzki	2012	5.3	1.9	7.2
Dirk Nowitzki	2004	6.9	0.3	7.2
Dwyane Wade	2010	7	0.2	7.2
Joel Embiid	2023	4.4	2.7	7.1
Joel Embiid	2019	3	4.1	7.1
Stephen Curry	2016	6.4	0.6	7
Shaquille O'Neal	2005	5.9	1	7
Shaquille O'Neal	2003	7.3	-0.3	7
Lebron James	2008	5	1.9	6.9
Chris Paul	2015	4.4	2.4	6.9
Manu Ginobili	2008	4.5	2.4	6.9
Jayson Tatum	2022	4.3	2.5	6.9
Kevin Garnett	2012	2	4.9	6.8
Nikola Jokic	2023	6.1	0.8	6.8
Stephen Curry	2020	6.3	0.6	6.8
Tim Duncan	2006	2.9	4	6.8
Shaquille O'Neal	2000	5.6	1.2	6.8
Kyle Lowry	2018	4.7	2.1	6.8
Lebron James	2020	4.9	1.8	6.7
Lebron James	2014	6	0.7	6.7
Kevin Garnett	2003	4.3	2.4	6.7
Chris Paul	2019	3.7	3	6.7
Giannis Antetokounmpo	2022	2.4	4.4	6.7
Lebron James	2022	4.9	1.7	6.6
Manu Ginobili	2006	5.6	1	6.6
Joel Embiid	2024	3.6	3	6.6
Dirk Nowitzki	2003	5.5	1.1	6.6
Kevin Durant	2021	5.2	1.4	6.6
Kevin Garnett	2013	0.3	6.2	6.5
Shaquille O'Neal	2002	6.2	0.4	6.5
Giannis Antetokounmpo	2020	1.7	4.8	6.5
Lebron James	2019	6.4	-0.1	6.4
Steve Nash	2010	7.3	-0.8	6.4
Manu Ginobili	2005	4	2.4	6.4
Paul George	2023	3.2	3.2	6.4
Joel Embiid	2020	1.7	4.7	6.4
Rudy Gobert	2022	1.8	4.5	6.4
Paul George	2024	3.8	2.4	6.3
Nikola Jokic	2018	4.8	1.4	6.2
Stephen Curry	2022	5.1	1.1	6.2
Tim Duncan	2002	2.1	4.1	6.2
Manu Ginobili	2016	3	3.2	6.2
Shaquille O'Neal	1999	4.7	1.5	6.2
Giannis Antetokounmpo	2021	1.5	4.7	6.2
Nikola Jokic	2019	4.9	1.1	6.1
Stephen Curry	2015	6.1	0	6.1
Manu Ginobili	2012	5.5	0.7	6.1
Jimmy Butler	2020	3.7	2.4	6.1
Kevin Garnett	2011	0.5	5.5	6
Chris Paul	2014	4.3	1.8	6
Manu Ginobili	2013	4.6	1.5	6
Shaquille O'Neal	2001	6.9	-0.9	6
Paul George	2020	2.9	3.1	6
Kobe Bryant	2008	6.1	-0.1	6
Lebron James	2013	4.5	1.4	5.9
Chris Paul	2013	5.2	0.7	5.9
Kawhi Leonard	2016	1.6	4.3	5.9
Tim Duncan	2001	2.6	3.4	5.9
Manu Ginobili	2014	4.3	1.6	5.9
Dirk Nowitzki	2005	4.6	1.3	5.9
Giannis Antetokounmpo	2024	3.7	2.2	5.9
Kevin Durant	2016	4.6	1.3	5.9
Kyle Lowry	2017	3.8	2	5.9
Chris Paul	2020	3.5	2.3	5.8
Dirk Nowitzki	2006	5	0.8	5.8
Jimmy Butler	2019	2.5	3.3	5.8
Kevin Garnett	2006	3.6	2.1	5.7
Stephen Curry	2021	5.1	0.6	5.7
Chris Paul	2010	5	0.7	5.7
Jayson Tatum	2023	4.7	1	5.7
Rudy Gobert	2021	1.3	4.4	5.7
Chris Paul	2011	4.1	1.5	5.6
Paul George	2022	2.3	3.2	5.6
Dwyane Wade	2011	5.7	-0.1	5.6
Dwyane Wade	2007	5.4	0.3	5.6
Kevin Durant	2017	4.6	0.9	5.6
Nikola Jokic	2017	4.3	1.2	5.5
Joel Embiid	2021	2.3	3.2	5.5
Dirk Nowitzki	2008	5.4	0.1	5.5
Dirk Nowitzki	2007	5.2	0.3	5.5
Giannis Antetokounmpo	2025	3.6	1.9	5.5
Kevin Durant	2023	3.8	1.8	5.5
James Harden	2017	5.9	-0.4	5.5
James Harden	2016	5.8	-0.3	5.5
Rudy Gobert	2018	1	4.5	5.5
Stephen Curry	2023	4.5	0.8	5.4
Chris Paul	2021	2.5	3	5.4
Tim Duncan	2008	3.6	1.8	5.4
Manu Ginobili	2015	2.6	2.8	5.4
Manu Ginobili	2009	2.9	2.5	5.4
Dirk Nowitzki	2011	4.1	1.2	5.4
Giannis Antetokounmpo	2023	2.5	2.9	5.4
Jayson Tatum	2020	2.3	3.1	5.4
Kevin Durant	2022	3.8	1.5	5.4
Jason Kidd	2005	3.9	1.6	5.4
Lebron James	2007	4	1.4	5.3
Kevin Garnett	2007	3.1	2.2	5.3
Kawhi Leonard	2025	3.3	2.1	5.3
Shaquille O'Neal	2006	4.3	1	5.3
Dwyane Wade	2008	5.3	-0.1	5.3
Stephen Curry	2014	5.3	0	5.2
Kawhi Leonard	2024	3.1	2.1	5.2
Kawhi Leonard	2021	4.3	0.8	5.2
Steve Nash	2006	7.1	-1.8	5.2
Jayson Tatum	2021	3.1	2.2	5.2
Chris Paul	2012	4	1.2	5.1
Steve Nash	2011	5.5	-0.4	5.1
Joel Embiid	2022	2.2	2.8	5.1
Dwyane Wade	2012	5.6	-0.5	5.1
Russell Westbrook	2017	6.1	-1	5.1
Chris Paul	2022	2.1	2.9	5
Tim Duncan	2009	3.2	1.8	5
Giannis Antetokounmpo	2019	1.4	3.5	5
Jayson Tatum	2024	3.9	1.2	5
Jason Kidd	2004	3.3	1.7	5
Ray Allen	2003	4.8	0.1	5

[LeoClark]

One interesting thing about this is that Garnett’s multi-year RAPM peak comes from his later years, rather than the earlier years where he seemed to have been a better player. This is different than the other guys near the very top, who pretty much all have their RAPM peak in a timeframe where I think we’d generally think of as their peak years (with CP3 being another possible exception to that, if we consider his peak to have been on the Hornets).

I think there’s both a charitable and uncharitable way to interpret this. One might say that Garnett having such great RAPM when he was a bit past his best but was actually on a good team suggests that peak Garnett could’ve put up even higher RAPM numbers than that if he’d been on a good team back then. But another interpretation would be that Garnett surely wasn’t actually at his most impactful in those later years and therefore that that data is perhaps just a product of noise, and we should key in more on the data from his actual best years. I’m not sure where I stand on that (and the truth may be somewhere in the middle), but it feels interesting to me since we don’t see much of this with other top players. I suppose another interpretation is that people are wrong and he actually was a better player in those later years than he’d been earlier, but I’m very skeptical of that.

Strange comment considering in another thread you said Kevin Garnett on/off is less impressive cause he was in a bad team, now when his rapm happens to peak on greats teams the data is noise.

[lessthanjake]

One interesting thing about this is that Garnett’s multi-year RAPM peak comes from his later years, rather than the earlier years where he seemed to have been a better player. This is different than the other guys near the very top, who pretty much all have their RAPM peak in a timeframe where I think we’d generally think of as their peak years (with CP3 being another possible exception to that, if we consider his peak to have been on the Hornets).

I think there’s both a charitable and uncharitable way to interpret this. One might say that Garnett having such great RAPM when he was a bit past his best but was actually on a good team suggests that peak Garnett could’ve put up even higher RAPM numbers than that if he’d been on a good team back then. But another interpretation would be that Garnett surely wasn’t actually at his most impactful in those later years and therefore that that data is perhaps just a product of noise, and we should key in more on the data from his actual best years. I’m not sure where I stand on that (and the truth may be somewhere in the middle), but it feels interesting to me since we don’t see much of this with other top players. I suppose another interpretation is that people are wrong and he actually was a better player in those later years than he’d been earlier, but I’m very skeptical of that.

Strange comment considering in another thread you said Kevin Garnett on/off is less impressive cause he was in a bad team, now when his rapm happens to peak on greats teams the data is noise.

I will say I find it very curious that someone with 9 posts—most recently including voting in the corrupted peaks project, with a ballot that certainly was in line with the corrupters’ goals—is seemingly referring back to discussion I had way before they joined these forums, particularly since those past discussions were ones taken issue with by the very individuals who corrupted the peaks project using alt accounts and sock puppets.

To the extent this post actually warrants a substantive response, I will note that on its face the post you quote provides both a “charitable” and “uncharitable” interpretation and I said “I’m not sure where I stand on that.” So saying my post affirmatively says “the data is noise” is actually clearly misleading.

[homecourtloss]

One interesting thing about this is that Garnett’s multi-year RAPM peak comes from his later years, rather than the earlier years where he seemed to have been a better player. This is different than the other guys near the very top, who pretty much all have their RAPM peak in a timeframe where I think we’d generally think of as their peak years (with CP3 being another possible exception to that, if we consider his peak to have been on the Hornets).

I think there’s both a charitable and uncharitable way to interpret this. One might say that Garnett having such great RAPM when he was a bit past his best but was actually on a good team suggests that peak Garnett could’ve put up even higher RAPM numbers than that if he’d been on a good team back then. But another interpretation would be that Garnett surely wasn’t actually at his most impactful in those later years and therefore that that data is perhaps just a product of noise, and we should key in more on the data from his actual best years. I’m not sure where I stand on that (and the truth may be somewhere in the middle), but it feels interesting to me since we don’t see much of this with other top players. I suppose another interpretation is that people are wrong and he actually was a better player in those later years than he’d been earlier, but I’m very skeptical of that.

Strange comment considering in another thread you said Kevin Garnett on/off is less impressive cause he was in a bad team, now when his rapm happens to peak on greats teams the data is noise.

Double post

[homecourtloss]

One interesting thing about this is that Garnett’s multi-year RAPM peak comes from his later years, rather than the earlier years where he seemed to have been a better player. This is different than the other guys near the very top, who pretty much all have their RAPM peak in a timeframe where I think we’d generally think of as their peak years (with CP3 being another possible exception to that, if we consider his peak to have been on the Hornets).

I think there’s both a charitable and uncharitable way to interpret this. One might say that Garnett having such great RAPM when he was a bit past his best but was actually on a good team suggests that peak Garnett could’ve put up even higher RAPM numbers than that if he’d been on a good team back then. But another interpretation would be that Garnett surely wasn’t actually at his most impactful in those later years and therefore that that data is perhaps just a product of noise, and we should key in more on the data from his actual best years. I’m not sure where I stand on that (and the truth may be somewhere in the middle), but it feels interesting to me since we don’t see much of this with other top players. I suppose another interpretation is that people are wrong and he actually was a better player in those later years than he’d been earlier, but I’m very skeptical of that.

Strange comment considering in another thread you said Kevin Garnett on/off is less impressive cause he was in a bad team, now when his rapm happens to peak on greats teams the data is noise.

Same thing with "the on value being more valuable" and then when some research was done and numbers were calculated and certain players had some nice on-off numbers, then the on-off was very important.

[lessthanjake]

One interesting thing about this is that Garnett’s multi-year RAPM peak comes from his later years, rather than the earlier years where he seemed to have been a better player. This is different than the other guys near the very top, who pretty much all have their RAPM peak in a timeframe where I think we’d generally think of as their peak years (with CP3 being another possible exception to that, if we consider his peak to have been on the Hornets).

I think there’s both a charitable and uncharitable way to interpret this. One might say that Garnett having such great RAPM when he was a bit past his best but was actually on a good team suggests that peak Garnett could’ve put up even higher RAPM numbers than that if he’d been on a good team back then. But another interpretation would be that Garnett surely wasn’t actually at his most impactful in those later years and therefore that that data is perhaps just a product of noise, and we should key in more on the data from his actual best years. I’m not sure where I stand on that (and the truth may be somewhere in the middle), but it feels interesting to me since we don’t see much of this with other top players. I suppose another interpretation is that people are wrong and he actually was a better player in those later years than he’d been earlier, but I’m very skeptical of that.

Strange comment considering in another thread you said Kevin Garnett on/off is less impressive cause he was in a bad team, now when his rapm happens to peak on greats teams the data is noise.

Same thing with "the on value being more valuable" and then when some research was done and numbers were calculated and certain players had some nice on-off numbers, then the on-off was very important.

I don’t know how many times I have to tell you that they’re both important and assessing how good a given set of numbers is requires looking at both and balancing them. This is really not a difficult concept (and, indeed, is the standard way of assessing such numbers), so it is a bit frustrating that you seek to constantly misrepresent the completely consistent viewpoint I have taken on this, because I know perfectly well that you understand it. I also think it is very telling what percent of your posts are aimed at making digs at me. It really feels like the lashing out of a frustrated child—and your frustration has obviously crescendoed further since the banning of your toxic, alt-account-spamming buddies (which you seem to blame me and a few others for, despite them absolutely doing it to themselves).

[homecourtloss]

Strange comment considering in another thread you said Kevin Garnett on/off is less impressive cause he was in a bad team, now when his rapm happens to peak on greats teams the data is noise.

Same thing with "the on value being more valuable" and then when some research was done and numbers were calculated and certain players had some nice on-off numbers, then the on-off was very important.

I don’t know how many times I have to tell you that they’re both important and assessing how good a given set of numbers is requires looking at both and balancing them. This is really not a difficult concept (and, indeed, is the standard way of assessing such numbers), so it is a bit frustrating that you seek to constantly misrepresent the completely consistent viewpoint I have taken on this, because I know perfectly well that you understand it. I also think it is very telling what percent of your posts are aimed at making digs at me. It really feels like the lashing out of a frustrated child—and your frustration has obviously crescendoed further since the banning of your toxic, alt-account-spamming buddies (which you seem to blame me and a few others for, despite them absolutely doing it to themselves).

these are your own words. you are saying that Robinson's on/off is more impressive because the +10.3 on compared to Garnet's lesser on court number.

I also don't appreciate the personal attacks here.

[lessthanjake]

Same thing with "the on value being more valuable" and then when some research was done and numbers were calculated and certain players had some nice on-off numbers, then the on-off was very important.

I don’t know how many times I have to tell you that they’re both important and assessing how good a given set of numbers is requires looking at both and balancing them. This is really not a difficult concept (and, indeed, is the standard way of assessing such numbers), so it is a bit frustrating that you seek to constantly misrepresent the completely consistent viewpoint I have taken on this, because I know perfectly well that you understand it. I also think it is very telling what percent of your posts are aimed at making digs at me. It really feels like the lashing out of a frustrated child—and your frustration has obviously crescendoed further since the banning of your toxic, alt-account-spamming buddies (which you seem to blame me and a few others for, despite them absolutely doing it to themselves).

these are your own words. you are saying that Robinson's on/off is more impressive because the +10.3 on compared to Garnet's lesser on court number.

I also don't appreciate the personal attacks here.

…I genuinely don’t see how you could think that that is remotely inconsistent with what I just said. I actually almost quoted those exact posts myself, in order to show how you were mischaracterizing my position! In the first post you pulled there, I was describing two players who actually had essentially identical on-off. In fact, Robinson’s on-off was very slightly higher. Meanwhile, Robinson’s “on” value was noticeably higher. If two players have essentially equal on-off and the “on” value for one of them is higher, then I am more impressed by the number of the guy with the noticeably higher “on” value. And that’s precisely because of the “diminishing marginal returns” that I describe in the second post you quote. That is not even remotely inconsistent with the idea that both the “on” value and the on-off are important and assessing how good a given set of numbers are requires looking at both and balancing them. Indeed, it’s exactly the straightforward conclusion that that approach would naturally lead someone to. Of course, the issue gets harder if one guy has meaningfully higher on-off and the other guy had a meaningfully higher “on” value. That’s where the actual balancing comes in, and depending on the exact numbers in question one might be more impressive than the other. That wasn’t the case there though. In that post, I was describing a situation where the on-off part of the equation was essentially equal and the “on” part wasn’t, so the conclusion of which number is better is easy. I feel like there has to be some willful reading comprehension issues going on here if you’re concluding that that post categorically suggests that only the “on” value matters. (And it’d have to be willful, because I have actually explained this in the past in response to this same dumb attack).

Meanwhile, you saying you “don’t appreciate the personal attacks here” is really rich, since it seems like almost every time I log onto RealGM these days I find you having made posts that are clearly aimed to jab at me and suggest my posts aren’t made in good faith. You do it subtly without naming names—likely to avoid issues with mods—but it’s typically obvious from the context and knowing my own posting history that it’s at least partially referring to me. I have actually asked you repeatedly to stop doing it. Me negatively characterizing your inability to stop doing so after repeatedly asking you to stop is not the issue. And there’s actually an easy fix here—which is for you to stop making rude posts directed towards me. If you think my post wasn’t accurately assessing your behavior, then not making rude posts referring to me should be easy enough for you to do. Let’s see if you can manage it. I’m not going to hold my breath, but maybe you’ll surprise me.

More generally, I also do want to note that the above-discussed posts from me are from over two years ago, so I still do find it quite curious that “LeoClark” with 11 posts and a 2025 join date was referring back to them. Always find it interesting when a new poster is somehow aware of years-old discussions that banned posters have previously harped on with me—particularly when those banned posters were banned in part for their proclivity with alt accounts. Maybe it’s just a coincidence, and “LeoClark” has been so enthralled by my posts that he’s gone back and read all of them and gotten hung up on the same misreading that those alt-account spammers did!

[penbeast0]

Warning issues for personal attacks. If you have a problem with an action of a poster, notify the mods. If you have a problem with the poster consistently, put them on ignore. You still might have them show up when other people quote them but you will see their comments less and therefore enjoy your interactions more. Seriously, try it.