What do impact metrics say about Kobe vs Shaq?

[ceiling raiser]

Thinking about year-by-year plus/minus, net rating, NPI RAPM, PI RAPM, and whatever else here.

At his peak Shaq looks like what you'd expect. In other years however, Shaq seems like a top 5-10 guy. Not best in the league.

Is there an affirmative argument for Kobe having a better career from an impact perspective?

[Owly]

Is there an affirmative argument for Kobe having a better career from an impact perspective?

This source
https://sites.google.com/site/rapmstats/97-14-rapm-2
https://sites.google.com/site/rapmstats/97-14-rapm
to my reading suggests it's a tough sell. Note first that this excludes some valuable Shaq years and a couple of badly negative Bryant years.
It has Shaq clearly better as a rate player.
Then without curving for the higher value of super elite seasons it has Shaq higher in wins above average and tied or very slight Bryant win in wins above replacement.
Then go back to what's missing (4 positive Shaq seasons, 2 negative Bryant ones) and factor in a something like an exponential curve for goodness to championship probability and would be hard for me to see a way to a case for Bryant from that data.

Other datasets may differ, I'm not aware of the best sources, other readings may be possible etc.

[JordansBulls]

An argument can be made for both, but prime is Shaq

[ceiling raiser]

I am increasingly entertaining the thought that Shaq might've been average or worse defensively for a lot of his career. Maybe not justified, though.

[Bad Gatorade]

Kobe often appears surprisingly underwhelming in RAPM data - still very good, mind you, but lower than his reputation. This seems to be largely thanks to defence - he tends to peak as a small positive in certain small clusters, but his defence averages out to be a negative through RAPM (and this seems to hold in the playoffs, too, from what I've seen).

This is Engelmann's 15 year RAPM from 2001-2015, age adjusted (so Kobe's teenage years shouldn't really hold him back) -

https://docs.google.com/spreadsheets/d/1-R9RXLp6eYuRcptQIQVTBIkLrxvrTCfLh_WB2P-DBwE/edit?usp=sharing

What is apparent, though, is that his DRAPM does seem to fluctuate quite a bit - after a good showing around 2010, it absolutely plummets from 2011-2013, and then Kobe gets injured (ergo, his post-2010 years may be somewhat overrated if RAPM is telling the truth). I do believe it was decently positive around 2000-2001 and 2008-2010, but craters at certain other times. This inconsistency could be influencing the results, or Kobe might just not be that good at defence on average, lol.

For a brief example as to how his 5 year clusters might look, there's a searchable table here -

https://www.thespax.com/nba/quantifying-the-nbas-greatest-five-year-peaks-since-1997/

This doesn't split between offence and defence, but it's a good, quick source that contains lots of 5-year stints, and it tells a similar story to what I've seen from Kobe in many other RAPM samples.

Shaq looks really good through RAPM - up until around 2005/2006, that is, at which point it drops off a lot. But prior to then, he's excellent, topping 1998's prior informed RAPM by quite a bit, and being the general 3rd guy to Duncan/Garnett in RAPM after that (and given that Shaq improved in the playoffs, it says positives about Shaq). The RAPM used to construct BPM 2.0 has 5 year splits, and Shaq rates excellently in 1997-01 (+7.7 - I believe he was the top ranked player to play all 5 seasons, but 4th overall), with his defence being only a slight positive (+0.6). 2002-06, his offence wasn't quite as gargantuan, although still great (+5.0), but his defence was a bit better (+1.0). His effort wasn't always there, but simply being a behemoth helps

Shaq looks clearly better from what I've generally seen through the RAPM scores, and it seems far more predicated on offence rather than defence.