disenfranchised wrote:Kobe most ever first team ALL-NBA selections.
Kobe most ever first team ALL-DEFENSE selections.
Kobe most consecutive all-star games ever.
31,000+ points and counting.
5 rings.
Not top ten? The incompetence in that notion is bewildering. There's no question he's top 10 all time, but the real question is, is he top 5? I say YES.
I haven't voted for #11 yet, and I very well may end up voting for kobe.
That said, there's no reason to look at all star selections as a viable criteria when discussing players at this level. The average fan who votes for starters doesn't care about ability or production, they just vote for their favorite players. Guys like carter, iverson and kobe would get voted in when they missed huge chunks of games. Made no difference.
As for his all defensive selections, those in the later part of his career were especially suspect. At that point he was getting selected by reputation as opposed to merit. I don't think they hold much importance.
Voted on by coaches, not the media or fans.
So basically, all-star votes don't count because of the fans. All-defense selections don't count with only Kobe because you say so.
Owly wrote:1) You basically answer your own question "I find it strange that peer-review is OK for guys like West & Oscar, but not Kobe in terms of defense", "what will you use to measure Oscar's defense". The former asks why peer review etc is okay, the latter essentially acknowledges that there are no exclusively defensive based measures of impact for that time frame(I guess you could split WoWY numbers in into offensive and defensive). And people have used peer review stuff on Kobe, just not exclusively the all-NBA stuff which has known flaws (pro-stars, often lazy based on stats or reputation, limited to position comparisons where there may be a relatively small range of impact). And I'm not saying Kobe's a bad defender, that I fully buy in to +/- variants or that his all D things are garbage. But you know why what's being used is being used. I don't know that many have voted West or Oscar doing so based on expressed high levels of confidence in high levels of (specifically) defensive impact (though they might assume that some WoWY "impact" was on the defensive end). And it's not like people haven't tried to look at net impact.
The point still stands though. I posted a quote from Doc Rivers about Kobe's defense, and it was dismissed. Obviously All-D teams have no traction here, but when you have NBA players, coaches, GMs, analysts consitently calling him a great defender, one has to wonder why peopel are so quick to overlook such peer review.
What point? That Oscar and West should be measured by things that we don't have data for? Assuming that isn't what the point is what is it? That if we don't have data for West and Robertson we shouldn't use it for anyone (that takes out steals, blocks and for most of their careers All-D)? If it's "what people have said should count for something" then, fine post it, those who are open to such points will be persuaded, those who aren't will try to rebutt, hopefully we learn something. I don't know about specific posts, but I don't know why you're convinced that your points are right yet won't be persuasive.
And as before I don't see many putting massive weights on what people said.
An Unbiased Fan wrote:
In terms of "The whole thing is weird how criteria seems to shift when Kobe's name is mentioned" you might want to address this at specific individuals or it's going to be hard to respond to.
Well, I did reply to MacGill's post, specifically his assertions about Kobe's "minimal" defensive impact, while at the same time bringing up Oscar who was on the worst defense teams of his era. Doesn't seem consistent if he values 2-way players.
In fact, the whole "Kobe wasn't a great defender' meme seems off. It's like when I'm having to argue Kobe vs Bird on defense when Bird himself would say Kobe was clearly better. There's a lot of this getting in the way of deeper discussions.
Okay it was at MacGill. That specific bit didn't read to me as clearly at anyone within the context of the post. I felt it was worth clarify so you got your response. Now you have.
An Unbiased Fan wrote:
2) Those pace adjusted numbers haven't accounted for the meanness in scorekeeping with regard to assists. You also notably ignore the efg% calculated or made any attempt to calculated ts% versus league norms.
pick an equivalent Kobe span if you go 04-05 to 12-13 (Kobe as first option years, but feel free to try a different range) http://bkref.com/tiny/1Z8ZC http://bkref.com/tiny/LbGQq Kobe at 122. A much bigger league sure but still that's another area with a substantial Robertson edge. Kobe is very big volume, above average efficiency. Oscar is big volume (2nd in points over that span), tremendous (best in the league) efficiency.
So basically the numbers you've posted are very misleading and by ignoring relative ts% and artificially diminishing the assist numbers you're hiding the two areas where Oscar evidently had a large impact on offense.
Don't see how the numbers are misleading considering I sourced them. The point of pace-adjusting is the same thing we did for Wilt/Shaq. One wouldn't compare Wilt's raw numbers...would they?
If you want to discuss things like TS5, then good. i would argue that Kobe faced far superior perimeter defense, while Oscar was in a very weak era for perimeter defense. Both their 2p% shooting was around 48-49%, so even if you put Kobe in that era with no 3pt line, he's just as efficient. Oscar feasted on the FT line back in the 60's. His athleticism was at a much higher deviation back then than Kobe by the time the 00's came. There are a lot of things to compare across eras.
Kobe was the more prolific scorer, Oscar was the better playmaker. If you feel Oscar had a bigger offensive imapct, I would need more specifics. I have both fairly even offensively. Kobe however, did it longer. Defensively, I would be interested in how Oscar's the better 2-way player.
You deliberately left part out (including efg that was already there) and easily (ts%) available and ignored something that I suspect is common knowledge amongst voters here (50s and 60s scorekeepers were stingier with assists). Saying you sourced it doesn't mean it's right. You posted it, you made it part of your argument which means it's your responsibility to put it into context. It's not a matter of raw numbers versus pace adjusted it's matter of the full picture. You omitted ts%, you put assists out of context and now you make a straw man about comparing Wilt with raw numbers.
I've made my case for Oscar already there, it's in the numbers, the WoWY, how well he blended with elite talent (at ASG and with Kareem) and, at the margins, to help confirm, what people said about him etc. If you want to bring in time machine stuff to make it seem like the hyperefficient guy either isn't making a larger impact by his efficiency or that that doesn't matter because you think it was a weak era ... whatever. FWiW people will go either way on this e.g. "terrible era, put him in a time machine he's bad", or "weak era poor coaching, nutrition etc and he was dominant imagine what he could do with modern training/resources, allowed palm etc" depending on whether they like the player in question.
Besides which it is wholly inconsistent to adjust one league norm (pace) and not the other "Both their 2p% shooting was around 48-49%" (here it might also be mentioned that it has been suggested that then convention of the time was for superstars then to have lower usage than today, I'm not sure whether I buy into factoring that in, but it is out there).
Which brings me back to, those numbers are incomplete, lacking in context and misleading.
G35 wrote:If this was a thread with the premise of educating posters on the circumstances/history of each player that would be different. This is a ranking thread....and please don't act naive that people will not reference this thread as a point of proof when making future comparisons.
"Look at where ________ was ranked in the recent top 100 list....see! I'm right!"
Fair enough, but can it not be both? Is that so impossible (or even improbable)?
G35 wrote:You can learn a lot about anything...does not mean it changes your mind. In fact education is not an indication of compliance. The public is constantly bombarded with the benefits of eating healthier and exercising but we still have a lot of fat, out of shape people walking around. But the bright side is it provides job security to doctors, nutritionists and organic farmers.......
This is a tenuous analogy at best, imo. What we're doing here in this thread is---let's face it---pretty trivial. This has basically no bearing on our life, lifestyle, reshaping core values or how we see the world. This is merely pastime for nearly everyone here. Changing one's opinion wrt a trivial player comparison hardly compares with altering your lifestyle, say with healthier diet---which can involve breaking bad habits, resisting favorite snacks, planning meals to a higher degree, altering how you shop for groceries, and cooking more (which may require some degree of "study")......in other words potentially a MAJOR adjustment----or with exercising more, or breaking addiction, etc.
Speaking for myself, I HAVE altered my opinion and rank on multiple players since starting this project. You could say I've just been brainwashed by someone else's propaganda. Well, maybe so. But I guess I'll counter by saying it was very convincing propaganda. And it's not like there isn't always propaganda for the other side of the coin; if I was convinced by one side vs. the other it's because that side that had (to me) the most sound reasoning. Further, the assumption that changing one's mind means I've just swallowed enough "spouted rhetoric" to make it my new "truth" is predicated on the assumption that my opinion was superior to everyone else's. I have my opinions and my reasons for having them, but I'm certainly not confident enough in their infallibility that I'm willing to disregard the opinions of dozens of other knowledgeable and studious fans.
"The fact that a proposition is absurd has never hindered those who wish to believe it." -Edward Rutherfurd "Those who can make you believe absurdities, can make you commit atrocities." - Voltaire
disenfranchised wrote:Kobe most ever first team ALL-NBA selections.
Kobe most ever first team ALL-DEFENSE selections.
Kobe most consecutive all-star games ever.
31,000+ points and counting.
5 rings.
Not top ten? The incompetence in that notion is bewildering. There's no question he's top 10 all time, but the real question is, is he top 5? I say YES.
I haven't voted for #11 yet, and I very well may end up voting for kobe.
That said, there's no reason to look at all star selections as a viable criteria when discussing players at this level. The average fan who votes for starters doesn't care about ability or production, they just vote for their favorite players. Guys like carter, iverson and kobe would get voted in when they missed huge chunks of games. Made no difference.
As for his all defensive selections, those in the later part of his career were especially suspect. At that point he was getting selected by reputation as opposed to merit. I don't think they hold much importance.
Voted on by coaches, not the media or fans.
So basically, all-star votes don't count because of the fans. All-defense selections don't count with only Kobe because you say so.
Gotcha.
You need to re-read what I wrote instead of making generalizations, although i guess i'm not surprised given your original post.
When discussing players at this level, all star selections aren't worth looking at considering everyone has multiple all NBA selections, which is a higher accolade. As you get into comparing players who weren't as accomplished, yes, you'd look at the all star selections and see how they performed in those seasons.
And if we're being conservative, since 2011, kobe has declined physically as a defender, yet made 3 straight all defensive first teams. If you think he actually deserved them, then fine, you're entitled to that opinion. I think it's clear he was getting voted in based on reputation, though.
1. If Bryant makes first or second-team All-Defense again this season, I’m officially ignoring this honor indefinitely when it comes to assessing a player’s status and career accomplishments. There will be no more “Player X has seven first-team All-Defense appearances” in this space as a way of justifying a man’s place in the NBA hierarchy. Bryant isn’t the first player coasting to undeserved All-Defense appearances based on reputation, but if he makes it again this season, it will be his most egregious appearance ever, and one of the most egregious in league history. It’s probably already time to start ignoring these honors, since Carlos Boozer received what I hope was an accidental vote last season, but Bryant’s selection would make it official.
Again, he’s a good defender when he wants to be, and he can still be a huge pain in the ass on the ball against top scorers. But playing top-notch on-ball defense on a few possessions per game does not qualify someone for an All-Defense honor, when all those other possessions of hideous off-ball defense exist.
Owly wrote:What point? That Oscar and West should be measured by things that we don't have data for? Assuming that isn't what the point is what is it? That if we don't have data for West and Robertson we shouldn't use it for anyone (that takes out steals, blocks and for most of their careers All-D)? If it's "what people have said should count for something" then, fine post it, those who are open to such points will be persuaded, those who aren't will try to rebutt, hopefully we learn something.
Huh? I didn't say that. I said if peer review is fine for Oscar/West, why not for Kobe. Hence the All-D reference, or the Rivers quote, etc. Not sure what your issue even is with what I said.
You deliberately left part out (including efg that was already there) and easily (ts%) available and ignored something that I suspect is common knowledge amongst voters here (50s and 60s scorekeepers were stingier with assists). Saying you sourced it doesn't mean it's right. You posted it, you made it part of your argument which means it's your responsibility to put it into context. It's not a matter of raw numbers versus pace adjusted it's matter of the full picture. You omitted ts%, you put assists out of context and now you make a straw man about comparing Wilt with raw numbers.
I've made my case for Oscar already there, it's in the numbers, the WoWY, how well he blended with elite talent (at ASG and with Kareem) and, at the margins, to help confirm, what people said about him etc. If you want to bring in time machine stuff to make it seem like the hyperefficient guy either isn't making a larger impact by his efficiency or that that doesn't matter because you think it was a weak era ... whatever. FWiW people will go either way on this e.g. "terrible era, put him in a time machine he's bad", or "weak era poor coaching, nutrition etc and he was dominant imagine what he could do with modern training/resources, allowed palm etc" depending on whether they like the player in question.
Besides which it is wholly inconsistent to adjust one league norm (pace) and not the other "Both their 2p% shooting was around 48-49%" (here it might also be mentioned that it has been suggested that then convention of the time was for superstars then to have lower usage than today, I'm not sure whether I buy into factoring that in, but it is out there).
Which brings me back to, those numbers are incomplete, lacking in context and misleading.
Why would I post a link to the whole stat sheet is I was leaving things out? There's a lot of info on the sheet, and I didn't see the point of listing it all when it was linked.
I was going to respond to the rest of your post, but the "those numbers are incomplete, lacking in context and misleading" kinda tell me its pointless.
7-time RealGM MVPoster 2009-2016 Inducted into RealGM HOF 1st ballot in 2017
I'm voting for oscar Robertson. When people translate stats for him they neglect the fact less baskets were awarded assists back then. He played in an era with 22 assists a game or so, similar to today.
So if we translate his numbers to 2000, as asked in the initial post we get about 25 ppg on a ts% 6-8 # over league average with 9-10 assists on one of the best offenses ever.
His defense was always called good, so for me I don't upgrade or downgrade him. He statistically is a better version of magic.
An Unbiased Fan wrote:Well, I did reply to MacGill's post, specifically his assertions about Kobe's "minimal" defensive impact, while at the same time bringing up Oscar who was on the worst defense teams of his era. Doesn't seem consistent if he values 2-way players.
Kobe's a 2 guard. Tell me how he is a better defender than KG? Or how he can impact the game at a higher level than KG? I never said Oscar was better than Kobe defensively, however with the information I am reading of Oscar I am leaning to him from an overall offensive standpoint. I also said that I was unsure of Oscar's impact defensively, so if you're going to quote me, at least be accurate.
So yes, if I am leaning towards KG that the 2-way impact as my later post indicates. Kobe is being discussed and I am being open of unsure about Oscar defense versus him.
I never said Kobe had bigger defensive impact than KG at all, just that Kobe was a great defender too. That's the thing. Most concede that Kobe was better on offense, but that KG was good too. Both were good 2-way players.
So in terms of comparisons, we have to ask, what had the bigger impact, Kobe's offense vs KG's defense. Kobe as an offensive anchor has consistently led good offenses, even when surrounded by bad talent like in the Smush days. KG's defense however, didn't anchor his Minny teams in a comparable fashion. Further, when you look at what Thibs did in Chicago, one has to wonder how much to attribute the Celtic's gaudy DRtg's to KG too alone. Clearly he was utilized better, but just the year before Minny has dismal numbers.
In the playoffs again, Kobe's offense was consistently more dominant than KG's defense. In 2001 for example, SA has no answer to stop Kobe, that's the power of individual offense. With no great perimeter defender to check him, Kobe ran wild. KG's defense however could only do so much. TD shot only 51% TS, which showed he was having an effect, but still....there's only so much individual defense can do.
Obviously there are limitations. Kobe couldn't get LA past the Suns in 06 or 07, nor could MJ will Chicago past the like of the 86 Celtiics. it's all relative. But looking at their careers, and spans like 06/07 where Kobe led the Smush teams to #8 & #7 offenses and the playoffs, while KG missed them, we can see what is happening.
Well I appreciate the response but I have read this from you before. My point was that 1) Kobe's defensive impact was less than KG's, hence why I said more minimal (he's a guard and an O anchor, this should be fairly obvious). 2) I brought up Oscar acknowledging I was unsure of his defensive prowess.
However, the arguments presented to me show that KG was much more balanced year after year on offense and defense. Kobe is better offensively, but........I am not convinced that his offense is greater than KG's total impact on a daily game basis. And as I mentioned before, Kobe isn't in my top tier offensive players of all-time. He is great but not at the top. KG to me is a much more portable player, he can be the focal point offensively or defensively. And because not all teams, players and eras are created equally, I prefer to analysis their skill set as individual players versus what Kobe did on the Lakers versus KG on the Wolves etc. This is where I am at with Oscar, while he scored at a good clip (and yes I know about pace) his rebounding and assists show me a huge impact on the court. So this is why I am trying to find out more about his defense. Regardless of what is posted, it is almost impossible to be an o & d anchor at the same time and to me, Hakeem and KG are the closest we have seen in that regard. KG may not have netted the highest team results, but I am not one to take away from his ability because of the other roster members.
ardee wrote: It's been great so far but if this continues it'll be down the toilet from here on out.
And sorry if I'm being rude, but this is also a general statement to other people who've said the same thing. Stop trying to be elitist in saying that you're only here for the conversation.
I can't speak for everyone else...I specifically chose not to vote because I wanted to focus on the conversations. If the rankings are what matters the most then everyone could just put together their lists, assign values to the rankings, and just be done with it.
But since we're all being honest and blunt, I would also recommend that you (and others) stop trying to be elitist in saying that people using RAPM to support their conclusions as ruining the project or leading it to go into the toilet. Otherwise we should just absolve all statistics that don't fit people's criterias or conclusions.
So when is this plane going down? I'll ride it til' it hits the ground!
An Unbiased Fan wrote:I came into this project hoping for good discussions, but it's kinda amazing how dismissive things have gotten the last few threads. Quite a few posters have put together some good posts, but they are basically skipped over, or not even addressed by most.
1) I find it strange that peer-review is OK for guys like West & Oscar, but not Kobe in terms of defense. I'm not sure why you say Kobe's D impact was minimal. KG didn't have great defenses in Minnesota. When both have decent support, they had Top 7 or so defenses. KG is definitely the better defender, but from 08-11 LA was #5, #6, #4, #6, for example. The whole thing is weird how criteria seems to shift when Kobe's name is mentioned.
As for Oscar...well the Royals were arguably the worse defensive squad of their era. It's not until he joins KAJ, that he's on a team with a decent defense. Seems strange to question Kobe's impact, and mention Oscar in this context. Better yet, what will you use to measure Oscar's defense, since peer review & videos are not apparently good guides.
The thing is that with Oscar, we have so little to go by with regards to his defense. We don't have steals and blocks numbers, we don't have +/- or RAPM, and we don't have a lot of game footage. We have to either accept what we hear and adjust for presumed hyperbole, take into account the limited footage of him available, or simply say that we have no idea about his defense at all.
61-69 Oscar: 25.7 ppg, 6.9 apg, 5.7 rpg ^ Very good, but don't see how that's better than prime Kobe.
Oscar's efficiency is more impressive relative to his era than Kobe's is, and I think that you need to take into account how assists were kept at the time and increase them again- you might as well keep the assist numbers what they actually were, it's probably a relatively accurate picture of what he'd get in today's game.
[quote]1. KG is a better defender, but the lead is smaller than people think. KG was an all-time defender in Boston but his style of play on that end was not conducive to his team's success in Minny. His 'middle linebacker' (I am sick of that now by the way) role where he was practically a perimeter player left the paint open season for 'Sota. In three straight seasons, 2002-04, the teams that knocked them out absolutely roasted them in the paint on ridiculous volume (65%+ in there every season, in the toughest defensive era of modern NBA history). Why defend mid-range jumpers and come out on PGs when you're leaving the paint wide open for the most efficient shot in basketball? Sounds like irresponsible defensive play (Kobe gets CRUSHED for this). This is why Minny had terrible team defenses even though he had guys like Hassell, Cassell, Rasho, etc. on his team through his career.[quote]
Wait, so you're holding KG responsible because he wasn't able to guard the middle (in the West btw during that time period) given his physical make-up? Was KG the starting center for Minny? You don't have your best player guarding the biggest strongest guys when you're giving up a lot of weight and you're counted on to score on the other end. Would you expect Kobe to guard Derrick Rose for an entire game?
[quote]2. Kobe's offensive lead is bigger. Kobe is an offensive constant, meaning you can rely on him 100% of the time or close to that to make your offense run. Whether it's him creating a shot for himself or someone else, it doesn't matter, it just keeps things moving. KG is not. His entire Minny career in the Playoffs (except for a few games in 2004) is punctuated by him looking lost a good part of the time and unable to do anything other than take mid-range jumpers. Sometimes, they go in. Cool, he'll have good games like he did against the 2003 Lakers. Otherwise, in general, it's just hurting your offense.[quote]
Given the information that has been presented here I think this is a very unfair comment and shouldn't really be taken into context. Your basically saying, what did his teams do under KG, when I am solely looking at how good KG was at basketball as an individual. KG was much more unselfish then Kobe, sometimes to a fault, but he was still a very good offensive player. Playoff failures (and the self awarded heroics) aren't just on him. Although, Kobe takes the last shot and it sometimes goes in, without his team, he isn't in the picture...at all.
Clyde Frazier wrote:And if we're being conservative, since 2011, kobe has declined physically as a defender, yet made 3 straight all defensive first teams. If you think he actually deserved them, then fine, you're entitled to that opinion. I think it's clear he was getting voted in based on reputation, though.
I can't believe people are arguing about Kobe's all defensive team selections as some sort of legit argument because it was voted on by the coaches. Hell I am not sure if LeBron deserved his 2nd team this year if we are just talking about the RS. There is no way Kobe should have been selected over Wade in 11 and 12.
JordansBulls wrote:The Warriors are basically a good college team until they meet a team with bigs in the NBA.
I really don't think "RAPM people" are ruining the project by only voting for those players. I think the critics are setting up some pinata strawmen to beat them senseless.
People who use RAPM are the list to list caveats and limitations.
Like if you see Amir Johnson's huge RAPM, you don't go, here's the next all-star! You look at what he's doing for his team, who he's playing with, how he's playing, et. Amir as a young guy had some of the best per minute block/steal/rebound stats and was an athletic, tall wild force. As he's grown up, he's learned to harness some of that energy on the court and become a great defender. On offense, he's one of the best screen-setters (if you look at big men who do better on RAPM than PER/WS, most of them are great at screens/picks), and he takes a lot of jump shots around 20 feet at the top of the key. He also played a lot next to Bargnani, so there's a possibility for some interference.... In a way he's like Garnett-lite, just without the passing and scoring.
And it seems his biggest issue is with the language in talking about RAPM: "impact." It is just a word.... So a player correlates with positive scoring margins for his teams versus opponents adjusting for teammates and HCA. That's a ... great thing.
edit: RAPM does one thing that most other stats do not do, and it's something stat-haters talk about a lot: adjust for who you're playing against/strength of schedule. I know people love to critique bench guys like Harden and Asik and how they'll flounder in bigger roles because they "play against second units" and run up the scores (never mind how units are almost always mixed and a lot of bench guys play in key situations like end of games.)
Excuse me since I'm not really going to be going into depth into the arguments that people have laid out for certain guys, I'm sure lots of research has been done so don't take it as an attack or anything. And to note, I definitely have a bias against older players. But I just can't fathom how - with the lacking resources we have even still to accurately assess a player- Oscar and West (to a lesser degree it seems) can be raked higher than Kobe. I disagree with notion that KG is a greater player as well, but that's clearly a different story. It make zero sense to me since based on the amount of evidence to judge players of today is fairly substantial and yet for the players from yesteryear we have quite a bit less, we are then judging on anecdotal statements from peers, which is often derided. There's an inconsistency here; if people really want to do these rankings each poster participating IMO should rough an objective standard they are going to follow. Not stringent enough where it doesn't allow for some fluid discussion, but enough so people are not people lost into narrative.
ElGee wrote:Thanks for the reply. It's left me at sort of a loss though because it's not clear to me why you would use these methods here, and in my past (limited) exposure to, specifically, facial recognition machine learning algorithms, it was not used (although I can surmise an explanation). But just know that this really doesn't address in any way your concerns related to RAPM/lineups. If I'm the student on this, I still have no idea why you would use RR in those situations but not with lineup data. The first discussion I had with someone in this area I consider a subject matter expert (quite a smart Math PhD) who is also a basketball and sports fan, his response was essentially "cool!" not "you can't do that."
Maybe were just not clearly communicating here so let me simplify: Is it OK to use OLS regression on PBP data? And if not, why?
You "could" use OLS sure, but ridge would still be a better predictor for what you're going after with RAPM. Again, I'm not sure why people asking me about regularization methods as if I have a problem with them. My issue, is the interpretation of what RAPM results represent. This isn't aimed at you Elgee, because you've been more than cordial in discussions, but the constant inference that because I don't agree with RAPm, that I "simply don't understand", or "don't like ridge regression" gets tiresome.
I made it a point when discussing RAPM, to point out examples of why I think it has problems, and the causes of those issues(like I did with Divac near the beginning of the project). Even still, I have no problem with posters using RAPM for arguments. I just have a problem with RAPM taking over the whole discussion, and other criteria getting dismissed as inferior or biased. I compared KG/Kobe by age, and it was clear that Kobe was regarded higher in his early 20's, and had a longer longevity into their 30's. But....apparently RAPM trumps that, and even KG at low production levels, is rated higher in pre/post prime seasons than prime Kobe. The 98/99 KG to 06/07 Kobe comparison is just mind-boggling. Again, the use of RAPM would be fine if it was reasonably used. But at this point nothing else seems to matter in discussions. And I've spent half this project discussing RAPM, so lets move on.
Oh, and I'm using a more dynamic approach for facial recognition, specifically for smartphones/tablets. The learning datasets approach is limited, and I'm using some of my game programming experience to employ 3D type characterizations. I'm basically trying to reverse engineer faces, and its working surprisingly well.
OK so I think we're aligned now -- that wasn't clear to me based on what you've said before. I'll use this as a jumping off point to re-address RAPM and what it's measuring (Chuck also had a very similar, excellent post on TS%):
MisterHibachi wrote:From what I understand of RAPM, it only measures how well the player did in the role that he played in. So eccentric results such as Amir Johnson being better than Shaq or KG better than Wade in 09 only says that Amir Johnson does a better job at his role of hustle big than Shaq did at his role of first option or that KG was better at his limited role of defensive quarterback than Wade was at his huge role.
It doesn't mean that certain players (like KG over Wade) are better than others in a vacuum. RAPM depends a lot on what role you play. It should only be used to compare players with similar responsibilities, such as Wade and LeBron in 09 or Shaq and Duncan and any other such similar roles.
Now this could be completely wrong because I'm not an expert in RAPM. So someone more informed please correct me if this is all wrong.
This is essentially correct, although we can be even more precise. It gives us the most likely probability of a player's value in a given role. This is based on one thing: the actual scoreboard results. There are confounders present, which is why the stat is not a perfect measurement. There are confounders present with every basketball stat I can think of.
The "in a given role" qualifier shouldn't automatically conjure up examples of if "David Robinson is backed up by David Robinson..." because they may take the court together quite a bit. The issue there is the classic Gortat example, when the minutes delta is very, very small (i.e. they never play together). There are sample size issues, as others have mentioned, based on the lineups (e.g. playing 98% of the minutes versus playing 72% of the minutes). There are sample size issues in the variation of lineups -- more equally weighted combinations is better. These things lead to perhaps the biggest issue with teasing out individual variables called multicollinearity.
There's also the issue of coaches only playing player's in optimal situation because of a specific skill set they have that fits with certain lineups/matchups. This is perhaps the biggest thing I see overlooked in RAPM citation -- role players being perceived to be better than stars based on their per-minute results. THE STAT IS NOT SAYING THE ROLE PLAYER IS BETTER. The classic examples are Amir Johnson and Nick Collison. I think less obvious examples are Ginobili and late or post-prime players like Stockton, Robinson and Garnett.
First, you look to see how efficient the role player is. That's where Collison does great. If you were a team looking for a scrappy defender who was very active in PnR situations (or all the other stuff Collison does) playing next to [type of big Collison is paired with] then you would absolutely look at that and go "Collison's our guy!" This is what the Mavericks do essentially, and they are one of the most successful teams in implementing analytics in the last decade. After that, if you were still trying to come up with some single metric that told you about "situational value" to the team, then you would implement the function Doc MJ alluded to -- multiple RAPM/100 * possessions played. When you do this, you'll see a metric that maps to almost every common sense "rankings" test you could give it. Perhaps more importantly, RAPM predicts trades quite well (not randomly), especially when there are more moving parts IIRC.
This is the opposite of garbage. I can't jump into someone's mind and make them understand something, especially if you have some deep-seated emotional desire to not want to use the metric. But if someone actually wants to understand why it has value, is used by billion-dollar teams, people get jobs from it, why it was considered a great and creative solution to parsing lineup data, and why many people around here consider it a valuable tool, there it is. It is NOT a player ranker. Just like FG% is NOT a shooting ranker. And ppg (or ORtg!) is NOT an offensive player ranker.
Just as loudly, I will again echo Chuck's post -- all of the concerns people have about RAPM confounders still exist in spades with every box score stat you have. Do I throw out the box score? Goodness no -- I incorporate it in my own composite metric. SPR6 has a good beat on this -- counting stats are valuable. They are also more abstract and somewhat orthogonal to "value" (or goodness) and thus when people try and distill these stats in a vacuum into PER or WS you have waaaaay more validity issues than you do with RAPM.
So what do you do? Do you use PPG/TS% to measure offense? Goodness no. You use PPG/TS% to gauge someone's performance within a context, and the more contexts the better. Do you use steals as a measure of defense? Hopefully that's obvious -- it measures very little. Do you use RAPM to rank players? Nooooo...you use it to try and understand the most likely impact a player is having in a given context. From there, JUST LIKE WITH EVERY OTHER STAT, you need to interpret the degree of confounders (i.e. is the sample size or MC an issue?) and the context in which the data came from. If you have a bunch of superstars, and they all play a lot (so there's lineup selection for them), and there's no other player in sight for collinearity, and the samples are large, then yes, that's some pretty compelling data your getting back.
Does it tell you that LeBron is 5% better than Garnett as a basketball player. NO! Just like 3p% doesn't tell me Curry is 4.7% better as a shooter than Miller. But if I look at 3p% and see Miller and Allen and Curry in their own pocket, and I look at some APM studies and see Shaq and LeBron and Garnett in their own pocket, I'm gonna be thinking those guys did some major ass-kicking on the basketball court. THEN, I could say could look at the situations for additional confounders and say something like "Is there something about situational value that can be soooo extremely different depending on circumstances for 40 mpg players? Maybe KG played on horrible teams and it was easier to improve them..."
Put bluntly: the people misinterpreting ppg and ts% are also the people more erroneously misinterpreting RAPM data. It's not that pro-RAPM posters don't misinterpret the data, they do, just not at the near universal rate I see when so-called "counter examples" are cited by its naysayers as a proof of condemnation. In general, if you're someone who doesn't want to understand the +/- family, I suggest you completely leave RAPM alone instead of trying to defame it and shift to thinking about how basketball players score points and how that changes the team's overall efficacy.
Clyde Frazier wrote:And if we're being conservative, since 2011, kobe has declined physically as a defender, yet made 3 straight all defensive first teams. If you think he actually deserved them, then fine, you're entitled to that opinion. I think it's clear he was getting voted in based on reputation, though.
I can't believe people are arguing about Kobe's all defensive team selections as some sort of legit argument because it was voted on by the coaches. Hell I am not sure if LeBron deserved his 2nd team this year if we are just talking about the RS. There is no way Kobe should have been selected over Wade in 11 and 12.
I don't think that's what people are saying. That's just a representation of what peers, journalists seem to think. Some people want to toss that aside. Okay, but then people use the same peer review for older players when making their decisions. I'm seeing post about how a poster had read up more on this player and is now convinced of XYZ. If we're going to be critical of peer review, be critical of it. I understand that peer review is perhaps a way of addressing the lack of resources, but in the end of the day, especially with past players, there's going to be a blind spot that no one can correct.
Clyde Frazier wrote:And if we're being conservative, since 2011, kobe has declined physically as a defender, yet made 3 straight all defensive first teams. If you think he actually deserved them, then fine, you're entitled to that opinion. I think it's clear he was getting voted in based on reputation, though.
I can't believe people are arguing about Kobe's all defensive team selections as some sort of legit argument because it was voted on by the coaches. Hell I am not sure if LeBron deserved his 2nd team this year if we are just talking about the RS. There is no way Kobe should have been selected over Wade in 11 and 12.
I don't think that's what people are saying. That's just a representation of what peers, journalists seem to think. Some people want to toss that aside. Okay, but then people use the same peer review for older players when making their decisions. I'm seeing post about how a poster had read up more on this player and is now convinced of XYZ. If we're going to be critical of peer review, be critical of it. I understand that peer review is perhaps a way of addressing the lack of resources, but in the end of the day, especially with past players, there's going to be a blind spot that no one can correct.
That's a very fair point but as it pertains to Kobe's all defensive selections after 2010, anyone who has seen him play knows they are a joke. He's been a one-way player for a while now except for very few short stretches.
JordansBulls wrote:The Warriors are basically a good college team until they meet a team with bigs in the NBA.
acrossthecourt wrote: And it seems his biggest issue is with the language in talking about RAPM: "impact." It is just a word.... So a player correlates with positive scoring margins for his teams versus opponents adjusting for teammates and HCA. That's a ... great thing.
The biggest issue is how RAPM is being used. I've tried to stay away from it but when its used to say pre-prime 98/99 KG was playing at a superstar level and higher than 06/07 Kobe, it just boggles the mind.
If people want tor reference RAPM, fine. But to pump up KG's longevity by using RAPM from the 90's and the 11-13 time, as if he was playing at a superstar level....just leaves me a bit blank for words.
It gets to the point that we can't argue because production, peer-review is off the table. Playoff performances...nope, can't use them either. What's left
7-time RealGM MVPoster 2009-2016 Inducted into RealGM HOF 1st ballot in 2017
acrossthecourt wrote: And it seems his biggest issue is with the language in talking about RAPM: "impact." It is just a word.... So a player correlates with positive scoring margins for his teams versus opponents adjusting for teammates and HCA. That's a ... great thing.
The biggest issue is how RAPM is being used. I've tried to stay away from it but when its used to say pre-prime 98/99 KG was playing at a superstar level and higher than 06/07 Kobe, it just boggles the mind.
If people want tor reference RAPM, fine. But to pump up KG's longevity by using RAPM from the 90's and the 11-13 time, as if he was playing at a superstar level....just leaves me a bit blank for words.
It gets to the point that we can't argue because production, peer-review is off the table. Playoff performances...nope, can't use them either. What's left
Yeah but timeout -- are you saying that a 19-10-3 +0% TS player can't be better than a 34-6-5 +3% TS player?
PaulieWal wrote:That's a very fair point but as it pertains to Kobe's all defensive selections after 2010, anyone who has seen him play knows they are a joke. He's been a one-way player for a while now except for very few short stretches.
That's true, but the problem is for me I'm not seeing that consistent analysis throughout these discussions for other players. What's more, there seems to be an extra bit of vitriol thrown Kobe's way for what his peers and journalists have said about him. It's interesting really.
acrossthecourt wrote: And it seems his biggest issue is with the language in talking about RAPM: "impact." It is just a word.... So a player correlates with positive scoring margins for his teams versus opponents adjusting for teammates and HCA. That's a ... great thing.
The biggest issue is how RAPM is being used. I've tried to stay away from it but when its used to say pre-prime 98/99 KG was playing at a superstar level and higher than 06/07 Kobe, it just boggles the mind.
If people want tor reference RAPM, fine. But to pump up KG's longevity by using RAPM from the 90's and the 11-13 time, as if he was playing at a superstar level....just leaves me a bit blank for words.
It gets to the point that we can't argue because production, peer-review is off the table. Playoff performances...nope, can't use them either. What's left
Yeah but timeout -- are you saying that a 19-10-3 +0% TS player can't be better than a 34-6-5 +3% TS player?
On its face, that would be extraordinarily unusual. But i did look further, specifically to things like peer review that regarded 06/07 Kobe as a Top 5 MVP/All-NBA 1st player. KG wasn't regarded as elite in 98 or 99.
If someone wants to explain how 98/99 KG was superstar level or better than 06/07 Kobe, I'm all ears.
7-time RealGM MVPoster 2009-2016 Inducted into RealGM HOF 1st ballot in 2017
