ElGee wrote:For your B point to be the case here we'd need years from the player's prime. I don't see how one-year of RAPM is now suddenly cutting it for mind-changing data. And that's what we're talking about here -- something with enough puissance to matter.
We're not talking about 1-year vs mulit-year here, we're talking about multi-1-year vs single-multi-year. If a guy gets a rating of +X in several independent, consecutive studies do you really not believe that this is more powerful than one longer +X study?
While I understand that there's some degree of noise period, noise is very much related to variance, and the independent 1-year studies speak to small variance.
ElGee wrote:There's also the practical application of these numbers. Do they go down because Malone's asked to do too much? What's the error in the single-season? Is 2001 all that really matters here? Why are Stockton's defensive numbers so high -- doesn't this suggest a potential problem? (Since when is RAPM not subject to problems?) And perhaps most importantly, why does Stockton have such a huge 2001 season but then drops off in a season that was widely considered a rejuvenation of sorts? Nearly the same box metrics with a mild increase in minutes, and more than double the 20-point games from the previous year.
"because Malone's asked to do too much?", this is a good question to ask, but no answer you respond with makes the concern irrelevant. If Malone putting up the numbers that impressed everybody is him doing too much, this is an eye opener.
"Stockton's defensive numbers a problem?", yup. I'm not focusing one 1-year study here, the importance is that there's more than one.
(And to be clear since in your answer you seemed to rebut things I already agreed you were right on: The fact that the '00 study is wonky obviously has an effect on my original point.)