lorak & Doc - Both of you are touching a larger concern of mine...at what point do we draw a distinction between a player's overall goodness how well a player is playing in a given role?
So, RAPM looks to isolate each player's ability to affect scoring margin. I think we can say that it does a pretty good job of that, but when you're isolating a player from his teammates and opponents, what you're left with is something that represents the player. But he's still playing the game a very specific way, and generally with superstars, it's not the only way he can play (and in the case of high portability players, it's not the only way he can thrive at a high level).
When we have a guy who dramatically shifts his playstyle mid-career, I really don't know if a direct comparison between those seasons is appropriate. RAPM (and WOWY, team ORtg/DRtg, and whatever other similar metrics we're using) aren't going to tell us whether a player was better in year n vs year (n-1) if that change took place. Better (or more poorly, since guys can sometimes be put in suboptimal situations later in their careers, either out of necessity or because of poor coaching/management) utilized, maybe, but since portability is such a big part of the discussion, either we're going to have to make a ton of assumptions or do some very involved analysis to try to determine how a guy would affect scoring margin in year n if he kept his year (n-1) role, or what impact he'd have in year (n-1) if the change was made earlier.
At some point, I think we have to decide what exactly we're trying to accomplish in comparing players. Are we trying to decide how well players would compare in their respective roles, or are we trying to make a comparison after considering the entire spectrum of possible roles they can play, and how well they'd play each of them? I think we're trying to do a bit of both, since it's tough to determine what a guy's optimal role is if we've never seen him physically play in that role, and succeed in it, and seasons in suboptimal roles provide datapoints that tell us both how much a guy triumphs over adversary, and how far removed those suboptimal roles are from his preferred roles.
I think ElGee touched on this a bit on the previous page, and his good post has me wondering about a few things:
ElGee wrote:Well I've said earlier in this project that people need to realize what RAPM is actually "saying." It does not say Collison and Johnson are among the best players in the world. It also does not say they are imparting the most impact. It says that they are very impactful in their given role...which is for a short period of time, i.e. is specialized.
This really stands out to me when it comes to players like old Stockton, old Robinson, old Garnett, etc. It's not that I don't think they were good, it's just that even if the stat is "accurate" it still only says what these players were able to do in the right spots. This not only includes minute allocation for their rest/energy but also the lineup-roles they have. RAPM tries to detect when you have good/bad teammates, but it doesn't know if you only get put out there in lineups that cater to your strengths.
This really isn't an issue at all when a player is forced to play with a diversity of lineups, i.e. play big minutes. But when you get into Stockton's 27-29 mpg territory, I see the results as much more specialized. An indicator of value, no doubt, but of overall goodness? I don't see how the metric is measuring that given the circumstances. I'd say the same thing about Robinson to a degree.
How important are minutes in determining a given player's role? If a guy is specializing, and playing in particular lineups, certainly minutes and lineups (though even for most role players, they're going to be playing in a variety of situations...if he's so linked to a given lineup, maybe we should consider looking at it as a consequence of him playing his role, as opposed to the other way around) are something extremely important to consider. But if we have a guy, who is playing very much the same brand of basketball, while slowly reducing his minutes, and there is a dramatic shift in RAPM or other impact-based metrics, should we consider this an effect of a difference in the role he was playing, or the actual performance in his roles?
It might sound like I'm worrying a bit much about semantics here, but I do think even if there isn't a dichotomy, it's something we have to try to answer. Particularly, when we have aging superstars, whether they find themselves in drastically different situations, or if their role change is the primary difference.
And I've heard people suggest confusion as to why Nash should be heralded for situational volume scoring while others aren't (necessarily). It's simple -- there's nothing situational about the PRESSURE Nash applies on the defense -- it's constant, he just balances his own shots with his teammates. He COULD be a volume scorer if he wanted to, the way some lead guards play by default (from the Francis/Marbury type to guys like Wade and even LeBron), but Nash senses higher efficiency elsewhere and perceives passing as the best option. Hard to argue with the GOAT-level offensive results. When the passing is choked off and the option sub-optimal -- due to defensive adjustment, teammate changes, or simply Nash just hedging his bets and simplifying the game by calling his own number more -- he is still capable of volume scoring well because he's awlays been capable of scoring well. Which is exactly what makes him so different from John Stockton. Similar, but steroids make a huge different.
Full circle to Malone -- Karl was the guy who played more of the Nash role I just described. When the going got tough, Utah went to Malone a lot. Sometimes Karl passed, but he wasn't the creator Nash was. He was a phenomenal passer, and an excellent scorer. There are probably only a handful of players in history who could have performed better against that kind of stress/load he faced annually in the playoffs. Meanwhile, Stockton passed and passed and essentially wilted against better defenses at the most inopportune times. And I still think Stockton's good, but there's a sizable gap between him and Malone.
Great, great point about Nash knowing when to score vs. pass. I would've liked if there was more discussion about Oscar and Magic in that context. I think it was established in earlier threads that they had the ability to score more often, but there wasn't as much discussion about how well they achieved that balance. Their offensive results speak for themselves obviously (as does Nash's), but this is one of those areas that always has a place in conversation, since I think it's very difficult to discern what about these guys in particular gives them the ability to balance the two, and what other offensive anchors with the ability to both create for others and score lack.
Regarding Malone, do we then feel his role was suboptimal throughout his career? If so, when was the closest he came to an optimal role, and what similar players do we feel were placed in what we could consider Malone's optimal role? If he was placed in an optimal role, do we think he handled the load and stress as well as he should've? If not, how much is due to context, and is there a possibility that a subopimal role for Malone would've better served his teams given his particular situation? I wondered aloud above about seasons in suboptimal roles telling us a bit about "how much a guy triumphs over adversary and how far removed those suboptimal roles are from his preferred roles". How valuable is the ability to succeed in those suboptimal situations, and at what point is it worth it for a player to assume those roles if their so far removed from his optimal role (especially if it makes team-building more difficult? Not saying this is the case with Malone, just a general concern?
Apologies for ranting, but the exchange between the three of you (which has been a very good discussion so far, I hope it continues into tomorrow) really got me thinking about this philosophically.
Now that's the difference between first and last place.
) did it to Lucas, because players defended by Robinson (Spencer and Chambers) usually stayed behind 3P line, so DRob had to be close to them and that limited his help defense. I think that's the main reason why Spurs defense was so bad in that series and I think we can't blame Robinson here as it's more coach fault, who was unable to make adjustments (just like Brown and Sloan vs Nelson).
