ElGee wrote:drza wrote:Spoiler:
I don't agree with this part. Nash was among many offensive quality players in Dallas. RAPM doesn't "know" this or have any way of adjusting for interactions based on skill types.His actual His 3-year offensive average there is +2.2. Overall from 2002-2004:
Dirk +5.8
Finley +3.1
Nash +0.5
In 2004 Antawn Jamison was +1.6. Josh Howard was +1.4. Najera +2.1. Nash was -0.8. I look at those numbers and sort of expect some kind of entanglement where someone is going to come out looking weaker due to diminishing returns on offense. Do you really look at that and think Nash was the 6th most valuable Mav? Or put another way, in 2004 do you think Dirk, Finley, Howard and Jamison were big-minute players helping Dallas and Nash and Antoine Walker were big-minute players hurting them?
And, for a counter data point, consider that from 02-04 Dirk missed 16 controlled games and the Mavs were a +4.6 team without him and only a +5.4 team with him. (Wowy Score +0.7.) Does that diminish Dirk as a player in those years, especially in 03 and 04?
Regarding Kidd, I'm wondering how you rate him when you put a number on it. I have his peak offense at +2 and his peak defense at +2 for a one-year peak of +4 and a number of +3.5 surrounding seasons (99-05). Compare to Nash, I have his peak offense at +7 and defense at -1 for most of his prime for a peak season of +6 with four more years at least +5 (05-10). If you perceive a notable difference there, I'm wondering why (i.e what makes you think Kidd moves offenses in the +3-4 range if you think that, or what makes you think Nash is lower if you think that).
Doctor MJ wrote:Spoiler:
Yeah, +/- is a funny thing. Obviously I use it a lot, but I feel much more comfortable looking at it when judging extreme tiers than I do at judging the guys in the middle. In Dallas, Nash was putting up all-star numbers on an extremely effective offense. Is it really at all reasonable to look at RAPM, see a neutral number, and conclude he wasn't much of a player? I mean, doesn't team redundancy have to be a huge part of it?
I say this not to advocate he be credited with Phoenix-level impact for his time in Dallas, but more to advocate moderation. When I look at Nash, I see a good long time in Phoenix with clear superstar-level impact, and then I see several more years in Phoenix where he was clearly showing enough that you can't look at him and say "he was a completely different player! a scrub!", which to me rounds out his impact, and reminds that Nash is 3rd all-time in assists with almost all of those assists coming in star-level seasons.
From another perspective: Phoenix was only able to get Nash because they went HARD after him. They desperately wanted Nash, and then he came and immediately won an MVP. Obviously, they saw something in Nash that made them think he could come in and make a splash for them, and if you watch some Nash highlights from Dallas, it's not hard to see what that was. To look at those Dallas years as somehow damning then is to take the precise stuff that made D'Antoni & co lust after Nash, and turn your nose up at it.
I chose to answer these quotes separately instead of couched, because there is a lot of similarity in the response and I'd want to address both of you equally on it. The underlined sentiment from both of you seems to be, paraphrased, that the Mavs had a lot of offensive talent and thus that Nash's meh +/- results from that team don't mean he's a bad player.
If that's the sentiment, I completely agree with the bolded.
However, I think it's imperative that we maintain consistency with the way we use/assess +/- data, especially in the current environment. As we've all mentioned before, +/- data is exactly that...data. Nash's meh RAPM scores in Dallas are data points. They happened. It's not true that the +/- results say that he was a bad player...what they say is that he wasn't having much impact on those teams. His presence didn't correlate with any movement in the scoring margin for those Mavs. And not just in the 2004 season with the Antoines...Nash started every game he played in in 1999, about half the ones he played in in 2000, and of course was All Star caliber by 2002 & 03 before ending his Mavs tenure in 2004. His scaled normalized RAPM for those 5 seasons are -1.1, +0.9, +0.6, +1.8, and -0.8. He just flat wasn't having much impact on Dallas's scoring margins year after year after year.
So. If those are data points, then it's up to us to interpret it, and as mentioned above it doesn't directly follow that those results mean that Nash was a bad player. I would testify to just the opposite, actually. As someone who was pulling for the Timberwolves in the 2002 playoffs match-up, I can absolutely attest that Nash was terrifying and that his offense being such a mismatch vs Chauncey's defense was massive in that series. So no, I don't think that Nash wasn't good.
However, I do think that if Nash could be an excellent player but have consistently meh impact, that is of note. We know that when Nash made his transition from the Mavs to the Suns in '04, lots of things changed. He was given the keys to the car to a larger degree. His offensive teammates all thrived as finishers (whereas Dirk and some of the other Mavs through the years like NVE or Walker also did more creating on their own). There's the handcheck rule. Texas Chuck has mentioned anecdotally that there was a potential motivation and/or work ethic change that may have occurred. Now I can't pinpoint exactly how much of each factor (or perhaps other un-named factors) may have contributed to Nash going from a +0.5 to a +8 type impact in RAPM studies. However, for one thing, I don't know how much to credit Nash for his Dallas years when SO much of his credit for the Suns years is tied into the very impact that he lacked in Dallas.
But further, this speaks to me about just how global Nash's peak impact was. I don't know if portable is the right term, but if not then it's portability's cousin. Because if all it takes to render Nash's impact moot is to have a team with other creators on it, then that seriously limits the type of offensiv that can be formed around him. Stoudemire and Marion were great players, but they were designed to be finishers (especially at the 5 and 4). Said another way, couldn't we postulate the exact opposite of El Gee's underline above about Nash's Phoenix days...was his impact so high purely because the team didn't have any other creators besides him? If Nash's impact lacks robustness in the face of different situations, then to me that opens up these lines of questions.
Kidd's presence, on the other hand, always moved the scoring margin needle upward to a significant degree on wildly different types of teams and in wildly different roles. Playing next to Penny on "backcourt 2000", Kidd had strong positive impact. Playing in New Jersey on a defensive squad with athletic finishers he had just-under-Nash-in-Phoenix impact. Playing in Dallas as a starter but no longer the main offensive engine, he still had a very positive impact on the scoring margin. His defense was portable and very strong for a guard. His offense provided value in roles ranging from unipolar offense creator to basically spot-up corner shooter. Kidd's impact was robust, to a very demonstrable degree.
ElGee, I have never really tried to do the type of individual SRS estimations that you do. But if anything, I tend to use the RAPM scale as an estiamte of what a person can do. As such, I don't think the +2 offense and -2 defense for Kidd are representative at all for me, if on that same scale Nash is a +6. Because I haven't tried your SRS approach, those numbers are almost arbitrary for me. However, if I go back to the normalized RAPM numbers that I reported, those resonate for me. I could see Kidd as essentially a +7 peak player with a +4 floor...while Nash might be a +8 ceiling with a +0 floor. So if we look at the evidence of their careers, is a player that can give you +8 impact under specialized circumstances but not nearly that impact in others worth more than a guy that gives you a similar period of +7 impact with an extra 7 - 8 years of +4? I could see an argument either way, but I think player with the more robust impact has a case.
