Part 1
Part 2
The with/without data was really the most time-consuming part. After that, comparing on/off and RAPM over the years is pretty easy.
KG vs Kobe, on/off over the years
Spoiler:
We have on/off data for both of these players going back to 01...luckily, that was the first year of Kobe's prime. Their primes overlapped from 01-08. Over that time, we can see who was making a bigger difference for their teams.
01 KG: +8.8
01 Kobe: +8.8
02 KG: +10.1
02 Kobe: -2.0
03 KG: +23.6
03 Kobe: +10.0
04 KG: +20.7
04 Kobe: +6.0
05 KG: +0.7
05 Kobe: +2.4
06 KG: +10.8
06 Kobe: +12.6
07 KG: +14.8
07 Kobe: +6.0
08 KG: +11.8
08 Kobe: +7.0
On/off isn't anywhere close to perfect, but I think over 8 years, if Garnett smashes Kobe in on/off 5 times, and is only slightly bested by Kobe or tied by Kobe the other 3 times, I think it clearly points to a pattern of either Garnett being asked to do a lot more for his teams than Kobe was, or that Garnett was simply a more impactful player than Kobe. And we see when Kobe is in a situation similar to what Garnett had to face for most of his career, he actually doesn't really come close to what Garnett managed to do (2 seasons with >+20 on/off). Kobe only hit +10 on/off twice over this span...KG hit +10 all BUT twice in this span.
I think it's at least fair to say that KG's teams were living and dying with him to a far greater degree than Kobe's team were with him. And this certainly doesn't hurt KG's argument, as this is another non-box score metric that supports his candidacy over Kobe, as well as with/without.
01 KG: +8.8
01 Kobe: +8.8
02 KG: +10.1
02 Kobe: -2.0
03 KG: +23.6
03 Kobe: +10.0
04 KG: +20.7
04 Kobe: +6.0
05 KG: +0.7
05 Kobe: +2.4
06 KG: +10.8
06 Kobe: +12.6
07 KG: +14.8
07 Kobe: +6.0
08 KG: +11.8
08 Kobe: +7.0
On/off isn't anywhere close to perfect, but I think over 8 years, if Garnett smashes Kobe in on/off 5 times, and is only slightly bested by Kobe or tied by Kobe the other 3 times, I think it clearly points to a pattern of either Garnett being asked to do a lot more for his teams than Kobe was, or that Garnett was simply a more impactful player than Kobe. And we see when Kobe is in a situation similar to what Garnett had to face for most of his career, he actually doesn't really come close to what Garnett managed to do (2 seasons with >+20 on/off). Kobe only hit +10 on/off twice over this span...KG hit +10 all BUT twice in this span.
I think it's at least fair to say that KG's teams were living and dying with him to a far greater degree than Kobe's team were with him. And this certainly doesn't hurt KG's argument, as this is another non-box score metric that supports his candidacy over Kobe, as well as with/without.
KG vs Kobe, RAPM over the years
Spoiler:
And now we get to the most controversial and polarizing stat currently used on this board: RAPM. I don't expect to convince the people that are vehemently against this stat at this point, but it's just another non-box score metric that can be used to see how players are affecting their teams. As you can see, I didn't only use RAPM. I'm trying to use a lot of different methods that give an idea of how much a player is helping his team.
I know acrossthecourt has some RAPM going back to the late 90s, but the data I'm most familiar with is Engelmann's RAPM (before he did some wonky stuff to it), which is available from colts18's sig. With regards to KG and Kobe, we can compare how they stack up to each other from 02-08 (overlapping primes). I agree with a lot of the criticisms of RAPM that mainly center on the fact that it only tells you how much a player helps a team in that specific situation, and doesn't really tell you whether someone is BETTER than someone else or not...however, I find it hard NOT to believe someone is probably a better player if they are CONSISTENTLY beating out someone else, regardless of team environment. KG and Kobe both played on bad teams at the same time (05-07), they both played as complementary offensive options (02-04 for Kobe, 08 for Garnett), and they both played as primary options (05-08 for Kobe, 02-07 for Garnett). And it seems to me that regardless of what kind of situation they find themselves in...Garnett consistently trumps Kobe. It's the consistency that really wins me over.
02 KG: +1.9 (16th)
02 Kobe: +1.9 (16th)
03 KG: +5.4 (1st)
03 Kobe: +2.4 (18th)
04 KG: +8.6 (1st)
04 Kobe: +1.6 (50th)
05 KG: +4.4 (8th)
05 Kobe: +0.7 (117th)
06 KG: +4.4 (11th)
06 Kobe: +4.8 (6th)
07 KG: +7.0 (4th)
07 Kobe: +5.5 (8th)
08 KG: +8.1 (1st)
08 Kobe: +6.1 (6th)
They're on the same level in 02, but then KG destroys Kobe from 03-05. In fact, KG enjoys a very clear advantage every year from 03-08 with the exception of 06...and there's barely a difference between them that year. Again, a player having a higher RAPM in a given season does not necessarily convince me that he's better...but I can't really overlook KG trumping Kobe year after year after year.
And of course, the great thing about Garnett is obviously how well he's aged. Even after 08, Garnett trumps Kobe every year with the exception of 2010.
I know acrossthecourt has some RAPM going back to the late 90s, but the data I'm most familiar with is Engelmann's RAPM (before he did some wonky stuff to it), which is available from colts18's sig. With regards to KG and Kobe, we can compare how they stack up to each other from 02-08 (overlapping primes). I agree with a lot of the criticisms of RAPM that mainly center on the fact that it only tells you how much a player helps a team in that specific situation, and doesn't really tell you whether someone is BETTER than someone else or not...however, I find it hard NOT to believe someone is probably a better player if they are CONSISTENTLY beating out someone else, regardless of team environment. KG and Kobe both played on bad teams at the same time (05-07), they both played as complementary offensive options (02-04 for Kobe, 08 for Garnett), and they both played as primary options (05-08 for Kobe, 02-07 for Garnett). And it seems to me that regardless of what kind of situation they find themselves in...Garnett consistently trumps Kobe. It's the consistency that really wins me over.
02 KG: +1.9 (16th)
02 Kobe: +1.9 (16th)
03 KG: +5.4 (1st)
03 Kobe: +2.4 (18th)
04 KG: +8.6 (1st)
04 Kobe: +1.6 (50th)
05 KG: +4.4 (8th)
05 Kobe: +0.7 (117th)
06 KG: +4.4 (11th)
06 Kobe: +4.8 (6th)
07 KG: +7.0 (4th)
07 Kobe: +5.5 (8th)
08 KG: +8.1 (1st)
08 Kobe: +6.1 (6th)
They're on the same level in 02, but then KG destroys Kobe from 03-05. In fact, KG enjoys a very clear advantage every year from 03-08 with the exception of 06...and there's barely a difference between them that year. Again, a player having a higher RAPM in a given season does not necessarily convince me that he's better...but I can't really overlook KG trumping Kobe year after year after year.
And of course, the great thing about Garnett is obviously how well he's aged. Even after 08, Garnett trumps Kobe every year with the exception of 2010.
Non-box score conclusions: KG vs Kobe
Spoiler:
There's really no debate when it comes to anything outside of the box score: KG trumps Kobe in terms of elevating a team from pretty much any perspective you come from. With/without, on/off, RAPM...it doesn't matter. Those are the three most powerful techniques for determining "goodness" that isn't captured by the points/rebounds/assists/TS% stat line, and they all agree that Garnett CLEARLY surpasses Kobe.
How much stock you're going to put into that is up to you, obviously. I'm going to address the box score differences as well, especially in the playoffs, which seems to be the big point of debate here, but I feel like everyone agrees that the box score isn't perfect. It actually doesn't really come close to encompassing all of what's going on in a basketball game, and ESPECIALLY for someone like Kevin Garnett, whose strengths revolve around everything BUT individual ISO offense, he's going to get underrated badly by the box score. The +/- family of stats actually paints a picture about just how special Garnett is in a way that the box score would never be able to do.
It's kind of the same concept with Bird, and why people tend to be so high on him. He's a guy who would still be a super-elite player even if he had a sub-50% TS...because he does a lot of things at an elite level that help a team BIG TIME, which basic counting stats aren't going to capture.
How much stock you're going to put into that is up to you, obviously. I'm going to address the box score differences as well, especially in the playoffs, which seems to be the big point of debate here, but I feel like everyone agrees that the box score isn't perfect. It actually doesn't really come close to encompassing all of what's going on in a basketball game, and ESPECIALLY for someone like Kevin Garnett, whose strengths revolve around everything BUT individual ISO offense, he's going to get underrated badly by the box score. The +/- family of stats actually paints a picture about just how special Garnett is in a way that the box score would never be able to do.
It's kind of the same concept with Bird, and why people tend to be so high on him. He's a guy who would still be a super-elite player even if he had a sub-50% TS...because he does a lot of things at an elite level that help a team BIG TIME, which basic counting stats aren't going to capture.
