Race to the MVP pt. III

[Ozark]

Lol, Durant is Dantley?

Dantley didn't have a 3 pt shot.
Dantley was a piss poor defender.
Dantley was a ball dominant player who burned up the shot clock dribbling around trying to get his look, then passed the ball away at the last second for a crappy look if he didn't get it.

Not even close.

[Doctor MJ]

How can Durant be having the GOAT scoring season when Adrian Dantley had several better ones in the 80's? Durant is a 6'11 Dantley imo.

Funny that we've arrived here. I'm not shocked really but it's really something.

Here's the thing:

The whole reason why you use this is because you know Dantley is nothing like the GOAT scorer. However, by basic numbers, Dantley has a GREAT case for being the greatest scorer of all-time. Hence, you could actually use this argument against literally any player in history. Not even Jordan is immune.

It would be ridiculous to use it against Jordan of course, everyone agrees, but the reasons it would be ridiculous to use against Jordan are the exact same reasons it's ridiculous to use it against Durant. Dantley was nothing remotely in the ballpark of a GOAT scorer because of how his game contorted his team's offense as he ate up the shot clock. Jordan & Durant don't play at all the same way.

[Doctor MJ]

Lebron is the best player in the league, but is Durant more valuable?

I think the reality is this:

LeBron chose to put himself in a situation where there's often clear redundancy in the talent of his lineup. That makes him less able to lift his team than he'd be able to do in other scenarios.

This was the essence of why LeBron lost the MVP in '10-11, and while people may regret that now, I don't think they were unreasonable at all. (Forget about Rose, think about what Howard was doing in Orlando.) This will basically always be an issue now.

That doesn't mean LeBron can't deserve the MVP. He clearly did last year. However we have a situation where LeBron's basically given everyone else a chance.

Durant's not as valuable as LeBron was in Cleveland, but yes, he is more valuable than LeBron is right now in Miami. If that seems just "wrong" to you because you think the big award should always go to the guy you think is inherently best, then I think you're placing unreasonable expectations on the award. The goal of the award should not be to shout the name of the most legendary player as many times as possible any more than we should knock a legendary player for not winning the award enough.

[MisterWestside]

Like I said before DocMJ, if we define MVP that way then we should just give it to the best player on the weakest overall team or "best fit" roster. We don't even need to use stats for that, either.

That's all dandy if people decide to define value that way, but I don't buy it.

[Doctor MJ]

Just replied to a couple of your posts, DocMJ. PM me if you wish, however; currently busy and can't respond to every post in detail here.

No worries mate.

Lets make this crystal-clear before I proceed: throughout his exchange (or elsewhere, for that matter), I have not considered you to be an idiot.

Noted, and right back at you. I consider you to be a very solid poster.

Now, with regards to the priors: you're right, the introduction of the priors muddies the validity of the metric. If I JUST want to see what the player in quesiton has done, then yes; non-prior data is more useful and I'll concede your point. But in actuality, stats that are great at telling you what happened don't exactly have any significant, practical purpose. (Per Alex's from Sports Skeptic great rundown on the various metrics available for public consumption, RAPM isn't even the most useful metric here). It's far better for a model to be able to predict things instead. We're halfway through the season, and Durant's "clear improvement" in 35 games or so is dwarfed by his track record from '10-'12. I cannot simply assume his much smaller sample size from this season is the new "norm". I want to know what he'll do going forward, and therefore I want those priors.

Note that priors are literally just more of "what happened". They improve prediction because they in general improve reliability, and as you conceded, there are scenarios where there are important validity issues with them. Where validity can be pinpointed to as clearly flawed you can very quickly reach the point where any reliability argument makes no sense.

If we're seeing something fundamentally new from Durant this year then there's no reason to look at previous seasons as something we expect him to regress to. One can use those seasons as a baseline for certainty of what his minimal levels habe to be, but not to use them against him.

Additionally, Skeptics' analysis is based on the idea of seeing which stat is closest to being The Lord of the Stats to rule them all. If I want one number to predict X, which will I use?

That's cool and all, but no analysis worth a salt should limit himself to one number, so it doesn't really damn RAPM to "lose" this one. I like +/- stats because of how well they work with other factors not because I can ignore other factors because of them.

Now as I say this, you're probably thinking, "Find Doc, but you talk about +/- an awful lot." Point taken, but note that I'm very careful with how I apply it. I use it prominently when I see examples that show trends to significant to ignore, but I certainly don't jump into every player comparison and side with a guy simply by throwing down a +/- stat.

We're specifically talking about a player who we both agree has changed significantly from the previous year.

Can't agree with that. His production is certainly superior (so far) but in essence his role hasn't changed significantly; he's simply refined the facets of his game at the SF position.

Oh okay. Well that would be a reason for disagreement then. Not that I really care to debate what constitutes a "role change", but when we see dramatic changes to areas of both a guy's box score and +/- stats, and this is happening for the good after his team lost a major talent from the previous year, I can't really fathom saying there hasn't been some significant change relating to that player.

I understand that sample size can be an issue here, but the season really isn't that young any more. Not that young for +/- considering how much of an outlier the numbers are, and really not that young for box score stats which tend to require far less sample size in order to achieve solid reliability.

RE: box score priors. These aren't a big issue with regards to Durant, but you brought them up relating to Nash who is tremendously underrated by box score metrics. That's the problem. If box score metrics were better then they could be used as an excellent prior...of course if box score metrics were better, why would we even bother with +/- data in the first place?

Since the goal is to make better predictions using this single stat (xRAPM), it's better to use the box score and +/- together. People do it all the time anyway, since the box score and +/- metrics present their own unique limitations in analysis. But this is a segue into your next post...

Letting you segue...

Well, by mashing everything into a black box, what he's done is create a flawed stat which can't really be used in an informed manner in conjunction with anything else with a great deal of confidence. Whereas, if you keep the the box score and the +/- separate, you've got two orthogonal indicators than an experienced analysis can factor based on their knowledge of how the two tend to work along with a variety of other factors.

Fair point. On that note, it would be interesting to see if the xRAPM blend would lead to better predictions in blends with other metrics than pure +/- and other metrics, and whether it's increased predictive power over RAPM makes up for the possible overfitting.

Then again, I use everything in analysis anyway (and this is not restricted to the metrics), so it's a bit of a "meh" issue from my point of view. And I still like xRAPM more than RAPM as a predictor despite your reservations.

I suppose what I think the key thing here is that talk of predictive power pretty much always focuses on what does better on average as if we're going to pick only one of these stats to be THE official stat to use all of the time. The reality for me is that I know where the issues are likely to be and tailor my analysis of a given player accordingly. For some comparisons, to the extent I was using a RAPM-based stat at all, the box score component wouldn't bother me in the slightest and I'd welcome the increased reliability. For others it makes the stat utterly worthless.

The fact that I can intelligently recognize when these issues come to the forefront and pick the best-of-breed at any given time makes it basically a given that I'm going to do better using my multi-tooled approach than what the Lord of the Stats winner-take-all approach could give.

On that note, you've almost certainly sensed my antipathy toward xRAPM. The issue isn't that I'm against xRAPM as a new tool added to the mix, but that it came on the seen as a replacement for RAPM, and the old data was literally deleted from the internet for no reason other than the statistician in question quite literally didn't understand why anyone would want the old data now...which said quite a lot about how different his perspective was from an experienced analyst.

If Player A has a weaker supporting cast than Player B, but Player A's team does better than Player B, then Player A has "lifted" his team more than Player B.

This statement doesn't consider the interactive effects and the phenomenon of diminishing returns in basketball. It's easier to "lift" a weaker supporting cast; harder to lift a stronger one. (It's also harder to lift a supporting cast that doesn't "fit".) One must account for these things before making any player comparison.

Where did I say I make player comparisons only on lift? In fact you've already seen me talk about lift being distinct from impact which was distinct from value which was distinct from goodness. This specialized vocabulary might seem absurd, but people use these terms to mean similar but slightly different things. What I'm doing is nailing them down not holding any of them up as the whole picture (not even "goodness" since the term implies there's a ideal numerical value that represents a player which is obviously oversimplistic).

I feel like the crux of the point is this:

You say we don't really know anything about LeBron's limitations because the fit in Miami is so terrible.

If you're talking about "fit" with Wade and Bosh, Bosh is playing his best offensive season since '10 in Toronto; and since Barkley infamously called out Wade in the Knicks blowout win vs. the Heat, Wade's skill curve has been near his performance during the '10 season. This is despite Wade being limited by with knee injuries since LeBron joined the Heat, LeBron and Wade sharing similar skill sets, and all three players changing their roles to some degree in a new system under Spoelstra. For the entire '13 season, the Big 3 together are close to what they did individually on offense by some metrics. It isn't all about RAPM (which you poster earlier), since that stat IS LIMITED. When you consider all the metrics, gametape, and diminishing returns in basketball, they're playing quite well.

No, DocMJ; what I'm referring to is your criticism of the Heat's ortg since LeBron joined the team. The Big 3 have played well, but you grossly overrate the rest of the supporting cast (players who have significant input into any team's ortg), which I will show later on.

Okay.

I say that I still expected better, and that I'm cautious about making assumptions about what ultra-outlier offensive performances LeBron could have given we haven't seen it, and that the issue of talent redundancy that we see in Miami would seem to be something that exists with any team capable of reaching those unheard of levels.

You then say "But you don't really know for sure", and I say "You're right".

I don't think you're just merely being skeptical. It's one thing to wonder how a player would hypothetically perform with a good offensive team with the ideal ingredients in place, but from what I've been reading in your posts you seem to go as far as to assume that that player isn't capable of this (LeBron) compared to another player who is playing with these ideal ingredients (Durant). And no, don't keep insisting that Durant/Westbrook is a bad "fit". Even with Westbrook's flaws, those players have more distinct, complementary games than James/Wade. This IS NOT a fair comparison, and you know it.

Put Durant with any of the other 3 players mentioned and you have a better fit than the other two guys left, and that's not a coincidence. What I'm alleging is that the advantage that you see as a coincidental unfair advantage in the comparison is pretty much what we'd expect in general.

When you have a player like LeBron who CAN do it all, that means he only gets to be his full self when he DOES do it all, which means other players being relegated to satellites. This means he's particularly prone to diminishing returns with other talent.

I don't really see how this is debatable. The only question is how severe the issue is. I think it's fine to say you don't think it's that big of a deal, but obviously I'm not willing to dismiss it so easily.

Hence my use of "just taken as an ideal". Clearly you're saying we have to do more than that, and clearly based on the amount of my writing, I am doing more than that, but we're also having issues with fundamental agreement on basic semantics so I have to lay these things down from time to time.

Writing a lot doesn't mean you're actually doing sound analysis. I'll get to that later.

I repeat: We are having fundamental issues in our discussion here. I'm having to define vocabulary terms and even when I do you respond with arguments that make clear you make all sorts of assumptions of the implication of me offering a benign clarification.

You want to dive in details but if we can't get straight exactly what we're debating we won't get anywhere.

And to be clear: The fact that that is the case is only possible because of how much knowledge we both have. It's a bit of me giving you a compliment, even as I get frustrated at how resistant you are to actually pinning down your knowledge with a precise schema.

Would the team be better if Wade was a more capable playmaker? Surely.

Do you think Westbrook is an elite level playmaker?

I mean it's working well with the two of them, but I can't really fathom the idea that if you've got a world-class off-ball player you wouldn't love to have the best playmaker you can next to him.

Let me take this same post and change some words here:

Would the team be better if Wade was a more capable shooter? Surely.

Do you think Wade is an elite level shooter?

I mean it's working well with the two of them, but I can't really fathom the idea that if you've got a world-class playmaker you wouldn't love to have the best shooter you can next to him.

Now, if you still claim that you cannot imagine how this can help LeBron and the Heat, you're not being honest.

So you're saying it's not fair to judge LeBron here until he has a teammate with all of Dwyane Wade's other abilities as well as being an elite-level shooting ability?

Or are you saying that the equivalent need for LeBron is someone who is just a world-class shooter and that makes him a superstar on that alone, in other words, we can't say LeBron has as good of a supporting cast as Durant until LeBron has someone like Durant as a sidekick?

Or are you saying LeBron doesn't need another superstar at all, just some really great shooters who aren't good enough in other areas to be called superstar sidekicks and thus don't find themselves underemployed by his presence, in other words saying that LeBron just can't make proper use of an offensive superstar next to him?

Let's consider Battier here.

This is a guy long considered probably the single greatest role player in the entire league. Does all the little things. Phenomenal defender, and certainly capable of hitting the open trey. I understand he's getting older, and maybe that's your point, but isn't he in theory exactly the type of guy you'd try to slot in next to the Big 3?

Chalmers? Hey, if you're going to have your 2 & 3 be the dictators of the offense, what are you expecting to get at the 1?

Haslem? The Big 3 took paycuts specifically so that he would stay because he seemed the precise type of player they needed.

I feel like you're looking at the things going wrong and saying "If those weren't going wrong, it would all be great", and I'm looking at them and saying "I don't think it's as easy as you think to reliably build better".

Haslem, who is usually a solid midrange shooter, is shooting -8% (compared to '12) on jumpers this season. Chalmers has never been a traditional PG even before LeBron joined the Heat, he would run the occasional play but his main talent comes from playing off-ball and shooting open shots. He hasn't shot the ball as well either compared to '12 and has even taken a step back in his free throw shooting (-10% from the line despite having the same free throw rate as '12). Are you blaming LeBron for Chalmers's regression from the line, too? Battier has shot the ball better, but he's been more of a black hole on offense and has not rebounded the ball as well. You could put some of that on Spo's insistence on small ball, but that means that the coaching staff is not utilizing Battier properly; only a fool would blame LeBron for this. LeBron loves to find the open man, and Battier doesn't have to shoot over 80% of his shots from 3 when he's not a Ray Allen from that range and he's a good mid-range shooter. He showed he could hit the mid-range shot playing with the Big 3 in '12!

Excuse you, but while you've been going on about "what-ifs" with Durant here and "easily" imaging him playing with a player with better skills and playing in a better offense; you somehow fail to see how a better Chalmers, Haslem, and Battier wouldn't help the Heat's offense. Those players have never been asked to do alot on offense and their roles in Miami are clearly defined. The world's best players are taking away attention away from them and they're not capitalizing on their shot opportunities, and in Battier's case he's not being used properly by Spoelstra who is responsible for being his team's talents together. And this is what I'm talking about when I say that you're not doing as much sound analysis as you think. You look at numbers and draw conclusions, I look at numbers AND watch the on-court actions on the floor AND the coaching personnel decisions that produce those numbers you look at and draw conclusions from. And these are good examples of how the Heat's ortg above league average is being affected by things that LeBron literally has no control over. These role players help the Heat with the lineups they can put around the Big 3; when 5-man lineups are weak in these areas it limits the effectiveness of the team, even while playing with stars.

Talking about changes to aspects in the Thunder offense and personnel as it pertains to Durant while ignoring the Heat offense with LeBron (which is what you've been doing, sorry DocMJ) is biased thinking.

I'll leave it there. Too busy to type up more lengthy, 30-page replies; and perhaps too lazy to type up everything otherwise

See, tons of knowledge in what you're saying. Totally glad I can read your thoughts.

In the end though I still think that leaves us with a kind of agree to disagree scenario.

I understand why someone says "I see the problems happening, it can be better...a LOT better". Aside from my general cautious streak that I've described above, clearly my view on role players in team building is a bit different from yours.

When Miami got their Big 3 together, there was a serious question how they could sign anyone else worth a damn. The cap made things so tight that it really handicapped them, and that's even the case despite the fact they've been able to get guys for under-market salary. On the whole, when they've made moves, people've said "Smart move", but time and again there are times when these players look terrible in the situation they are in.

Part of that has to do with fitting in with the existing talent, part of that has to do with players who maybe aren't what they used to be, and part of that has to do with the fact that beggars can't be choosers.

So your damn right I'm imagining a better version of Westbrook. Teams in the NBA can acquire a second star, and they can get a guy who fits well with the top star. It's not trivial to do, but there's nothing utterly unrealistic about it.

Talking about the 4th, 5th, and 6th guys on a roster though when you have huge amounts of money tied up with the first 3 just seems like such a pipe dream. Whoever you get is going to be flawed, and you're going to end up noticing those flaws.

[Doctor MJ]

Just replied to a couple of your posts, DocMJ. PM me if you wish, however; currently busy and can't respond to every post in detail here.

No worries mate.

Lets make this crystal-clear before I proceed: throughout his exchange (or elsewhere, for that matter), I have not considered you to be an idiot.

Noted, and right back at you. I consider you to be a very solid poster.

Now, with regards to the priors: you're right, the introduction of the priors muddies the validity of the metric. If I JUST want to see what the player in quesiton has done, then yes; non-prior data is more useful and I'll concede your point. But in actuality, stats that are great at telling you what happened don't exactly have any significant, practical purpose. (Per Alex's from Sports Skeptic great rundown on the various metrics available for public consumption, RAPM isn't even the most useful metric here). It's far better for a model to be able to predict things instead. We're halfway through the season, and Durant's "clear improvement" in 35 games or so is dwarfed by his track record from '10-'12. I cannot simply assume his much smaller sample size from this season is the new "norm". I want to know what he'll do going forward, and therefore I want those priors.

Note that priors are literally just more of "what happened". They improve prediction because they in general improve reliability, and as you conceded, there are scenarios where there are important validity issues with them. Where validity can be pinpointed to as clearly flawed you can very quickly reach the point where any reliability argument makes no sense.

If we're seeing something fundamentally new from Durant this year then there's no reason to look at previous seasons as something we expect him to regress to. One can use those seasons as a baseline for certainty of what his minimal levels habe to be, but not to use them against him.

Additionally, Skeptics' analysis is based on the idea of seeing which stat is closest to being The Lord of the Stats to rule them all. If I want one number to predict X, which will I use?

That's cool and all, but no analysis worth a salt should limit himself to one number, so it doesn't really damn RAPM to "lose" this one. I like +/- stats because of how well they work with other factors not because I can ignore other factors because of them.

Now as I say this, you're probably thinking, "Find Doc, but you talk about +/- an awful lot." Point taken, but note that I'm very careful with how I apply it. I use it prominently when I see examples that show trends to significant to ignore, but I certainly don't jump into every player comparison and side with a guy simply by throwing down a +/- stat.

We're specifically talking about a player who we both agree has changed significantly from the previous year.

Can't agree with that. His production is certainly superior (so far) but in essence his role hasn't changed significantly; he's simply refined the facets of his game at the SF position.

Oh okay. Well that would be a reason for disagreement then. Not that I really care to debate what constitutes a "role change", but when we see dramatic changes to areas of both a guy's box score and +/- stats, and this is happening for the good after his team lost a major talent from the previous year, I can't really fathom saying there hasn't been some significant change relating to that player.

I understand that sample size can be an issue here, but the season really isn't that young any more. Not that young for +/- considering how much of an outlier the numbers are, and really not that young for box score stats which tend to require far less sample size in order to achieve solid reliability.

RE: box score priors. These aren't a big issue with regards to Durant, but you brought them up relating to Nash who is tremendously underrated by box score metrics. That's the problem. If box score metrics were better then they could be used as an excellent prior...of course if box score metrics were better, why would we even bother with +/- data in the first place?

Since the goal is to make better predictions using this single stat (xRAPM), it's better to use the box score and +/- together. People do it all the time anyway, since the box score and +/- metrics present their own unique limitations in analysis. But this is a segue into your next post...

Letting you segue...

Well, by mashing everything into a black box, what he's done is create a flawed stat which can't really be used in an informed manner in conjunction with anything else with a great deal of confidence. Whereas, if you keep the the box score and the +/- separate, you've got two orthogonal indicators than an experienced analysis can factor based on their knowledge of how the two tend to work along with a variety of other factors.

Fair point. On that note, it would be interesting to see if the xRAPM blend would lead to better predictions in blends with other metrics than pure +/- and other metrics, and whether it's increased predictive power over RAPM makes up for the possible overfitting.

Then again, I use everything in analysis anyway (and this is not restricted to the metrics), so it's a bit of a "meh" issue from my point of view. And I still like xRAPM more than RAPM as a predictor despite your reservations.

I suppose what I think the key thing here is that talk of predictive power pretty much always focuses on what does better on average as if we're going to pick only one of these stats to be THE official stat to use all of the time. The reality for me is that I know where the issues are likely to be and tailor my analysis of a given player accordingly. For some comparisons, to the extent I was using a RAPM-based stat at all, the box score component wouldn't bother me in the slightest and I'd welcome the increased reliability. For others it makes the stat utterly worthless.

The fact that I can intelligently recognize when these issues come to the forefront and pick the best-of-breed at any given time makes it basically a given that I'm going to do better using my multi-tooled approach than what the Lord of the Stats winner-take-all approach could give.

On that note, you've almost certainly sensed my antipathy toward xRAPM. The issue isn't that I'm against xRAPM as a new tool added to the mix, but that it came on the seen as a replacement for RAPM, and the old data was literally deleted from the internet for no reason other than the statistician in question quite literally didn't understand why anyone would want the old data now...which said quite a lot about how different his perspective was from an experienced analyst.

If Player A has a weaker supporting cast than Player B, but Player A's team does better than Player B, then Player A has "lifted" his team more than Player B.

This statement doesn't consider the interactive effects and the phenomenon of diminishing returns in basketball. It's easier to "lift" a weaker supporting cast; harder to lift a stronger one. (It's also harder to lift a supporting cast that doesn't "fit".) One must account for these things before making any player comparison.

Where did I say I make player comparisons only on lift? In fact you've already seen me talk about lift being distinct from impact which was distinct from value which was distinct from goodness. This specialized vocabulary might seem absurd, but people use these terms to mean similar but slightly different things. What I'm doing is nailing them down not holding any of them up as the whole picture (not even "goodness" since the term implies there's a ideal numerical value that represents a player which is obviously oversimplistic).

I feel like the crux of the point is this:

You say we don't really know anything about LeBron's limitations because the fit in Miami is so terrible.

If you're talking about "fit" with Wade and Bosh, Bosh is playing his best offensive season since '10 in Toronto; and since Barkley infamously called out Wade in the Knicks blowout win vs. the Heat, Wade's skill curve has been near his performance during the '10 season. This is despite Wade being limited by with knee injuries since LeBron joined the Heat, LeBron and Wade sharing similar skill sets, and all three players changing their roles to some degree in a new system under Spoelstra. For the entire '13 season, the Big 3 together are close to what they did individually on offense by some metrics. It isn't all about RAPM (which you poster earlier), since that stat IS LIMITED. When you consider all the metrics, gametape, and diminishing returns in basketball, they're playing quite well.

No, DocMJ; what I'm referring to is your criticism of the Heat's ortg since LeBron joined the team. The Big 3 have played well, but you grossly overrate the rest of the supporting cast (players who have significant input into any team's ortg), which I will show later on.

Okay.

I say that I still expected better, and that I'm cautious about making assumptions about what ultra-outlier offensive performances LeBron could have given we haven't seen it, and that the issue of talent redundancy that we see in Miami would seem to be something that exists with any team capable of reaching those unheard of levels.

You then say "But you don't really know for sure", and I say "You're right".

I don't think you're just merely being skeptical. It's one thing to wonder how a player would hypothetically perform with a good offensive team with the ideal ingredients in place, but from what I've been reading in your posts you seem to go as far as to assume that that player isn't capable of this (LeBron) compared to another player who is playing with these ideal ingredients (Durant). And no, don't keep insisting that Durant/Westbrook is a bad "fit". Even with Westbrook's flaws, those players have more distinct, complementary games than James/Wade. This IS NOT a fair comparison, and you know it.

Put Durant with any of the other 3 players mentioned and you have a better fit than the other two guys left, and that's not a coincidence. What I'm alleging is that the advantage that you see as a coincidental unfair advantage in the comparison is pretty much what we'd expect in general.

When you have a player like LeBron who CAN do it all, that means he only gets to be his full self when he DOES do it all, which means other players being relegated to satellites. This means he's particularly prone to diminishing returns with other talent.

I don't really see how this is debatable. The only question is how severe the issue is. I think it's fine to say you don't think it's that big of a deal, but obviously I'm not willing to dismiss it so easily.

Hence my use of "just taken as an ideal". Clearly you're saying we have to do more than that, and clearly based on the amount of my writing, I am doing more than that, but we're also having issues with fundamental agreement on basic semantics so I have to lay these things down from time to time.

Writing a lot doesn't mean you're actually doing sound analysis. I'll get to that later.

I repeat: We are having fundamental issues in our discussion here. I'm having to define vocabulary terms and even when I do you respond with arguments that make clear you make all sorts of assumptions of the implication of me offering a benign clarification.

You want to dive in details but if we can't get straight exactly what we're debating we won't get anywhere.

And to be clear: The fact that that is the case is only possible because of how much knowledge we both have. It's a bit of me giving you a compliment, even as I get frustrated at how resistant you are to actually pinning down your knowledge with a precise schema.

Would the team be better if Wade was a more capable playmaker? Surely.

Do you think Westbrook is an elite level playmaker?

I mean it's working well with the two of them, but I can't really fathom the idea that if you've got a world-class off-ball player you wouldn't love to have the best playmaker you can next to him.

Let me take this same post and change some words here:

Would the team be better if Wade was a more capable shooter? Surely.

Do you think Wade is an elite level shooter?

I mean it's working well with the two of them, but I can't really fathom the idea that if you've got a world-class playmaker you wouldn't love to have the best shooter you can next to him.

Now, if you still claim that you cannot imagine how this can help LeBron and the Heat, you're not being honest.

So you're saying it's not fair to judge LeBron here until he has a teammate with all of Dwyane Wade's other abilities as well as being an elite-level shooting ability?

Or are you saying that the equivalent need for LeBron is someone who is just a world-class shooter and that makes him a superstar on that alone, in other words, we can't say LeBron has as good of a supporting cast as Durant until LeBron has someone like Durant as a sidekick?

Or are you saying LeBron doesn't need another superstar at all, just some really great shooters who aren't good enough in other areas to be called superstar sidekicks and thus don't find themselves underemployed by his presence, in other words saying that LeBron just can't make proper use of an offensive superstar next to him?

Let's consider Battier here.

This is a guy long considered probably the single greatest role player in the entire league. Does all the little things. Phenomenal defender, and certainly capable of hitting the open trey. I understand he's getting older, and maybe that's your point, but isn't he in theory exactly the type of guy you'd try to slot in next to the Big 3?

Chalmers? Hey, if you're going to have your 2 & 3 be the dictators of the offense, what are you expecting to get at the 1?

Haslem? The Big 3 took paycuts specifically so that he would stay because he seemed the precise type of player they needed.

I feel like you're looking at the things going wrong and saying "If those weren't going wrong, it would all be great", and I'm looking at them and saying "I don't think it's as easy as you think to reliably build better".

Haslem, who is usually a solid midrange shooter, is shooting -8% (compared to '12) on jumpers this season. Chalmers has never been a traditional PG even before LeBron joined the Heat, he would run the occasional play but his main talent comes from playing off-ball and shooting open shots. He hasn't shot the ball as well either compared to '12 and has even taken a step back in his free throw shooting (-10% from the line despite having the same free throw rate as '12). Are you blaming LeBron for Chalmers's regression from the line, too? Battier has shot the ball better, but he's been more of a black hole on offense and has not rebounded the ball as well. You could put some of that on Spo's insistence on small ball, but that means that the coaching staff is not utilizing Battier properly; only a fool would blame LeBron for this. LeBron loves to find the open man, and Battier doesn't have to shoot over 80% of his shots from 3 when he's not a Ray Allen from that range and he's a good mid-range shooter. He showed he could hit the mid-range shot playing with the Big 3 in '12!

Excuse you, but while you've been going on about "what-ifs" with Durant here and "easily" imaging him playing with a player with better skills and playing in a better offense; you somehow fail to see how a better Chalmers, Haslem, and Battier wouldn't help the Heat's offense. Those players have never been asked to do alot on offense and their roles in Miami are clearly defined. The world's best players are taking away attention away from them and they're not capitalizing on their shot opportunities, and in Battier's case he's not being used properly by Spoelstra who is responsible for being his team's talents together. And this is what I'm talking about when I say that you're not doing as much sound analysis as you think. You look at numbers and draw conclusions, I look at numbers AND watch the on-court actions on the floor AND the coaching personnel decisions that produce those numbers you look at and draw conclusions from. And these are good examples of how the Heat's ortg above league average is being affected by things that LeBron literally has no control over. These role players help the Heat with the lineups they can put around the Big 3; when 5-man lineups are weak in these areas it limits the effectiveness of the team, even while playing with stars.

Talking about changes to aspects in the Thunder offense and personnel as it pertains to Durant while ignoring the Heat offense with LeBron (which is what you've been doing, sorry DocMJ) is biased thinking.

I'll leave it there. Too busy to type up more lengthy, 30-page replies; and perhaps too lazy to type up everything otherwise

See, tons of knowledge in what you're saying. Totally glad I can read your thoughts.

In the end though I still think that leaves us with a kind of agree to disagree scenario.

I understand why someone says "I see the problems happening, it can be better...a LOT better". Aside from my general cautious streak that I've described above, clearly my view on role players in team building is a bit different from yours.

When Miami got their Big 3 together, there was a serious question how they could sign anyone else worth a damn. The cap made things so tight that it really handicapped them, and that's even the case despite the fact they've been able to get guys for under-market salary. On the whole, when they've made moves, people've said "Smart move", but time and again there are times when these players look terrible in the situation they are in.

Part of that has to do with fitting in with the existing talent, part of that has to do with players who maybe aren't what they used to be, and part of that has to do with the fact that beggars can't be choosers.

So your damn right I'm imagining a better version of Westbrook. Teams in the NBA can acquire a second star, and they can get a guy who fits well with the top star. It's not trivial to do, but there's nothing utterly unrealistic about it.

Talking about the 4th, 5th, and 6th guys on a roster though when you have huge amounts of money tied up with the first 3 just seems like such a pipe dream. Whoever you get is going to be flawed, and you're going to end up noticing those flaws.

[Doctor MJ]

Just replied to a couple of your posts, DocMJ. PM me if you wish, however; currently busy and can't respond to every post in detail here.

No worries mate.

Lets make this crystal-clear before I proceed: throughout his exchange (or elsewhere, for that matter), I have not considered you to be an idiot.

Noted, and right back at you. I consider you to be a very solid poster.

Now, with regards to the priors: you're right, the introduction of the priors muddies the validity of the metric. If I JUST want to see what the player in quesiton has done, then yes; non-prior data is more useful and I'll concede your point. But in actuality, stats that are great at telling you what happened don't exactly have any significant, practical purpose. (Per Alex's from Sports Skeptic great rundown on the various metrics available for public consumption, RAPM isn't even the most useful metric here). It's far better for a model to be able to predict things instead. We're halfway through the season, and Durant's "clear improvement" in 35 games or so is dwarfed by his track record from '10-'12. I cannot simply assume his much smaller sample size from this season is the new "norm". I want to know what he'll do going forward, and therefore I want those priors.

Note that priors are literally just more of "what happened". They improve prediction because they in general improve reliability, and as you conceded, there are scenarios where there are important validity issues with them. Where validity can be pinpointed to as clearly flawed you can very quickly reach the point where any reliability argument makes no sense.

If we're seeing something fundamentally new from Durant this year then there's no reason to look at previous seasons as something we expect him to regress to. One can use those seasons as a baseline for certainty of what his minimal levels habe to be, but not to use them against him.

Additionally, Skeptics' analysis is based on the idea of seeing which stat is closest to being The Lord of the Stats to rule them all. If I want one number to predict X, which will I use?

That's cool and all, but no analysis worth a salt should limit himself to one number, so it doesn't really damn RAPM to "lose" this one. I like +/- stats because of how well they work with other factors not because I can ignore other factors because of them.

Now as I say this, you're probably thinking, "Find Doc, but you talk about +/- an awful lot." Point taken, but note that I'm very careful with how I apply it. I use it prominently when I see examples that show trends to significant to ignore, but I certainly don't jump into every player comparison and side with a guy simply by throwing down a +/- stat.

We're specifically talking about a player who we both agree has changed significantly from the previous year.

Can't agree with that. His production is certainly superior (so far) but in essence his role hasn't changed significantly; he's simply refined the facets of his game at the SF position.

Oh okay. Well that would be a reason for disagreement then. Not that I really care to debate what constitutes a "role change", but when we see dramatic changes to areas of both a guy's box score and +/- stats, and this is happening for the good after his team lost a major talent from the previous year, I can't really fathom saying there hasn't been some significant change relating to that player.

I understand that sample size can be an issue here, but the season really isn't that young any more. Not that young for +/- considering how much of an outlier the numbers are, and really not that young for box score stats which tend to require far less sample size in order to achieve solid reliability.

RE: box score priors. These aren't a big issue with regards to Durant, but you brought them up relating to Nash who is tremendously underrated by box score metrics. That's the problem. If box score metrics were better then they could be used as an excellent prior...of course if box score metrics were better, why would we even bother with +/- data in the first place?

Since the goal is to make better predictions using this single stat (xRAPM), it's better to use the box score and +/- together. People do it all the time anyway, since the box score and +/- metrics present their own unique limitations in analysis. But this is a segue into your next post...

Letting you segue...

Well, by mashing everything into a black box, what he's done is create a flawed stat which can't really be used in an informed manner in conjunction with anything else with a great deal of confidence. Whereas, if you keep the the box score and the +/- separate, you've got two orthogonal indicators than an experienced analysis can factor based on their knowledge of how the two tend to work along with a variety of other factors.

Fair point. On that note, it would be interesting to see if the xRAPM blend would lead to better predictions in blends with other metrics than pure +/- and other metrics, and whether it's increased predictive power over RAPM makes up for the possible overfitting.

Then again, I use everything in analysis anyway (and this is not restricted to the metrics), so it's a bit of a "meh" issue from my point of view. And I still like xRAPM more than RAPM as a predictor despite your reservations.

I suppose what I think the key thing here is that talk of predictive power pretty much always focuses on what does better on average as if we're going to pick only one of these stats to be THE official stat to use all of the time. The reality for me is that I know where the issues are likely to be and tailor my analysis of a given player accordingly. For some comparisons, to the extent I was using a RAPM-based stat at all, the box score component wouldn't bother me in the slightest and I'd welcome the increased reliability. For others it makes the stat utterly worthless.

The fact that I can intelligently recognize when these issues come to the forefront and pick the best-of-breed at any given time makes it basically a given that I'm going to do better using my multi-tooled approach than what the Lord of the Stats winner-take-all approach could give.

On that note, you've almost certainly sensed my antipathy toward xRAPM. The issue isn't that I'm against xRAPM as a new tool added to the mix, but that it came on the seen as a replacement for RAPM, and the old data was literally deleted from the internet for no reason other than the statistician in question quite literally didn't understand why anyone would want the old data now...which said quite a lot about how different his perspective was from an experienced analyst.

If Player A has a weaker supporting cast than Player B, but Player A's team does better than Player B, then Player A has "lifted" his team more than Player B.

This statement doesn't consider the interactive effects and the phenomenon of diminishing returns in basketball. It's easier to "lift" a weaker supporting cast; harder to lift a stronger one. (It's also harder to lift a supporting cast that doesn't "fit".) One must account for these things before making any player comparison.

Where did I say I make player comparisons only on lift? In fact you've already seen me talk about lift being distinct from impact which was distinct from value which was distinct from goodness. This specialized vocabulary might seem absurd, but people use these terms to mean similar but slightly different things. What I'm doing is nailing them down not holding any of them up as the whole picture (not even "goodness" since the term implies there's a ideal numerical value that represents a player which is obviously oversimplistic).

I feel like the crux of the point is this:

You say we don't really know anything about LeBron's limitations because the fit in Miami is so terrible.

If you're talking about "fit" with Wade and Bosh, Bosh is playing his best offensive season since '10 in Toronto; and since Barkley infamously called out Wade in the Knicks blowout win vs. the Heat, Wade's skill curve has been near his performance during the '10 season. This is despite Wade being limited by with knee injuries since LeBron joined the Heat, LeBron and Wade sharing similar skill sets, and all three players changing their roles to some degree in a new system under Spoelstra. For the entire '13 season, the Big 3 together are close to what they did individually on offense by some metrics. It isn't all about RAPM (which you poster earlier), since that stat IS LIMITED. When you consider all the metrics, gametape, and diminishing returns in basketball, they're playing quite well.

No, DocMJ; what I'm referring to is your criticism of the Heat's ortg since LeBron joined the team. The Big 3 have played well, but you grossly overrate the rest of the supporting cast (players who have significant input into any team's ortg), which I will show later on.

Okay.

I say that I still expected better, and that I'm cautious about making assumptions about what ultra-outlier offensive performances LeBron could have given we haven't seen it, and that the issue of talent redundancy that we see in Miami would seem to be something that exists with any team capable of reaching those unheard of levels.

You then say "But you don't really know for sure", and I say "You're right".

I don't think you're just merely being skeptical. It's one thing to wonder how a player would hypothetically perform with a good offensive team with the ideal ingredients in place, but from what I've been reading in your posts you seem to go as far as to assume that that player isn't capable of this (LeBron) compared to another player who is playing with these ideal ingredients (Durant). And no, don't keep insisting that Durant/Westbrook is a bad "fit". Even with Westbrook's flaws, those players have more distinct, complementary games than James/Wade. This IS NOT a fair comparison, and you know it.

Put Durant with any of the other 3 players mentioned and you have a better fit than the other two guys left, and that's not a coincidence. What I'm alleging is that the advantage that you see as a coincidental unfair advantage in the comparison is pretty much what we'd expect in general.

When you have a player like LeBron who CAN do it all, that means he only gets to be his full self when he DOES do it all, which means other players being relegated to satellites. This means he's particularly prone to diminishing returns with other talent.

I don't really see how this is debatable. The only question is how severe the issue is. I think it's fine to say you don't think it's that big of a deal, but obviously I'm not willing to dismiss it so easily.

Hence my use of "just taken as an ideal". Clearly you're saying we have to do more than that, and clearly based on the amount of my writing, I am doing more than that, but we're also having issues with fundamental agreement on basic semantics so I have to lay these things down from time to time.

Writing a lot doesn't mean you're actually doing sound analysis. I'll get to that later.

I repeat: We are having fundamental issues in our discussion here. I'm having to define vocabulary terms and even when I do you respond with arguments that make clear you make all sorts of assumptions of the implication of me offering a benign clarification.

You want to dive in details but if we can't get straight exactly what we're debating we won't get anywhere.

And to be clear: The fact that that is the case is only possible because of how much knowledge we both have. It's a bit of me giving you a compliment, even as I get frustrated at how resistant you are to actually pinning down your knowledge with a precise schema.

Would the team be better if Wade was a more capable playmaker? Surely.

Do you think Westbrook is an elite level playmaker?

I mean it's working well with the two of them, but I can't really fathom the idea that if you've got a world-class off-ball player you wouldn't love to have the best playmaker you can next to him.

Let me take this same post and change some words here:

Would the team be better if Wade was a more capable shooter? Surely.

Do you think Wade is an elite level shooter?

I mean it's working well with the two of them, but I can't really fathom the idea that if you've got a world-class playmaker you wouldn't love to have the best shooter you can next to him.

Now, if you still claim that you cannot imagine how this can help LeBron and the Heat, you're not being honest.

So you're saying it's not fair to judge LeBron here until he has a teammate with all of Dwyane Wade's other abilities as well as being an elite-level shooting ability?

Or are you saying that the equivalent need for LeBron is someone who is just a world-class shooter and that makes him a superstar on that alone, in other words, we can't say LeBron has as good of a supporting cast as Durant until LeBron has someone like Durant as a sidekick?

Or are you saying LeBron doesn't need another superstar at all, just some really great shooters who aren't good enough in other areas to be called superstar sidekicks and thus don't find themselves underemployed by his presence, in other words saying that LeBron just can't make proper use of an offensive superstar next to him?

Let's consider Battier here.

This is a guy long considered probably the single greatest role player in the entire league. Does all the little things. Phenomenal defender, and certainly capable of hitting the open trey. I understand he's getting older, and maybe that's your point, but isn't he in theory exactly the type of guy you'd try to slot in next to the Big 3?

Chalmers? Hey, if you're going to have your 2 & 3 be the dictators of the offense, what are you expecting to get at the 1?

Haslem? The Big 3 took paycuts specifically so that he would stay because he seemed the precise type of player they needed.

I feel like you're looking at the things going wrong and saying "If those weren't going wrong, it would all be great", and I'm looking at them and saying "I don't think it's as easy as you think to reliably build better".

Haslem, who is usually a solid midrange shooter, is shooting -8% (compared to '12) on jumpers this season. Chalmers has never been a traditional PG even before LeBron joined the Heat, he would run the occasional play but his main talent comes from playing off-ball and shooting open shots. He hasn't shot the ball as well either compared to '12 and has even taken a step back in his free throw shooting (-10% from the line despite having the same free throw rate as '12). Are you blaming LeBron for Chalmers's regression from the line, too? Battier has shot the ball better, but he's been more of a black hole on offense and has not rebounded the ball as well. You could put some of that on Spo's insistence on small ball, but that means that the coaching staff is not utilizing Battier properly; only a fool would blame LeBron for this. LeBron loves to find the open man, and Battier doesn't have to shoot over 80% of his shots from 3 when he's not a Ray Allen from that range and he's a good mid-range shooter. He showed he could hit the mid-range shot playing with the Big 3 in '12!

Excuse you, but while you've been going on about "what-ifs" with Durant here and "easily" imaging him playing with a player with better skills and playing in a better offense; you somehow fail to see how a better Chalmers, Haslem, and Battier wouldn't help the Heat's offense. Those players have never been asked to do alot on offense and their roles in Miami are clearly defined. The world's best players are taking away attention away from them and they're not capitalizing on their shot opportunities, and in Battier's case he's not being used properly by Spoelstra who is responsible for being his team's talents together. And this is what I'm talking about when I say that you're not doing as much sound analysis as you think. You look at numbers and draw conclusions, I look at numbers AND watch the on-court actions on the floor AND the coaching personnel decisions that produce those numbers you look at and draw conclusions from. And these are good examples of how the Heat's ortg above league average is being affected by things that LeBron literally has no control over. These role players help the Heat with the lineups they can put around the Big 3; when 5-man lineups are weak in these areas it limits the effectiveness of the team, even while playing with stars.

Talking about changes to aspects in the Thunder offense and personnel as it pertains to Durant while ignoring the Heat offense with LeBron (which is what you've been doing, sorry DocMJ) is biased thinking.

I'll leave it there. Too busy to type up more lengthy, 30-page replies; and perhaps too lazy to type up everything otherwise

See, tons of knowledge in what you're saying. Totally glad I can read your thoughts.

In the end though I still think that leaves us with a kind of agree to disagree scenario.

I understand why someone says "I see the problems happening, it can be better...a LOT better". Aside from my general cautious streak that I've described above, clearly my view on role players in team building is a bit different from yours.

When Miami got their Big 3 together, there was a serious question how they could sign anyone else worth a damn. The cap made things so tight that it really handicapped them, and that's even the case despite the fact they've been able to get guys for under-market salary. On the whole, when they've made moves, people've said "Smart move", but time and again there are times when these players look terrible in the situation they are in.

Part of that has to do with fitting in with the existing talent, part of that has to do with players who maybe aren't what they used to be, and part of that has to do with the fact that beggars can't be choosers.

So your damn right I'm imagining a better version of Westbrook. Teams in the NBA can acquire a second star, and they can get a guy who fits well with the top star. It's not trivial to do, but there's nothing utterly unrealistic about it.

Talking about the 4th, 5th, and 6th guys on a roster though when you have huge amounts of money tied up with the first 3 just seems like such a pipe dream. Whoever you get is going to be flawed, and you're going to end up noticing those flaws.

[Doctor MJ]

Just replied to a couple of your posts, DocMJ. PM me if you wish, however; currently busy and can't respond to every post in detail here.

No worries mate.

Lets make this crystal-clear before I proceed: throughout his exchange (or elsewhere, for that matter), I have not considered you to be an idiot.

Noted, and right back at you. I consider you to be a very solid poster.

Now, with regards to the priors: you're right, the introduction of the priors muddies the validity of the metric. If I JUST want to see what the player in quesiton has done, then yes; non-prior data is more useful and I'll concede your point. But in actuality, stats that are great at telling you what happened don't exactly have any significant, practical purpose. (Per Alex's from Sports Skeptic great rundown on the various metrics available for public consumption, RAPM isn't even the most useful metric here). It's far better for a model to be able to predict things instead. We're halfway through the season, and Durant's "clear improvement" in 35 games or so is dwarfed by his track record from '10-'12. I cannot simply assume his much smaller sample size from this season is the new "norm". I want to know what he'll do going forward, and therefore I want those priors.

Note that priors are literally just more of "what happened". They improve prediction because they in general improve reliability, and as you conceded, there are scenarios where there are important validity issues with them. Where validity can be pinpointed to as clearly flawed you can very quickly reach the point where any reliability argument makes no sense.

If we're seeing something fundamentally new from Durant this year then there's no reason to look at previous seasons as something we expect him to regress to. One can use those seasons as a baseline for certainty of what his minimal levels habe to be, but not to use them against him.

Additionally, Skeptics' analysis is based on the idea of seeing which stat is closest to being The Lord of the Stats to rule them all. If I want one number to predict X, which will I use?

That's cool and all, but no analysis worth a salt should limit himself to one number, so it doesn't really damn RAPM to "lose" this one. I like +/- stats because of how well they work with other factors not because I can ignore other factors because of them.

Now as I say this, you're probably thinking, "Find Doc, but you talk about +/- an awful lot." Point taken, but note that I'm very careful with how I apply it. I use it prominently when I see examples that show trends to significant to ignore, but I certainly don't jump into every player comparison and side with a guy simply by throwing down a +/- stat.

We're specifically talking about a player who we both agree has changed significantly from the previous year.

Can't agree with that. His production is certainly superior (so far) but in essence his role hasn't changed significantly; he's simply refined the facets of his game at the SF position.

Oh okay. Well that would be a reason for disagreement then. Not that I really care to debate what constitutes a "role change", but when we see dramatic changes to areas of both a guy's box score and +/- stats, and this is happening for the good after his team lost a major talent from the previous year, I can't really fathom saying there hasn't been some significant change relating to that player.

I understand that sample size can be an issue here, but the season really isn't that young any more. Not that young for +/- considering how much of an outlier the numbers are, and really not that young for box score stats which tend to require far less sample size in order to achieve solid reliability.

RE: box score priors. These aren't a big issue with regards to Durant, but you brought them up relating to Nash who is tremendously underrated by box score metrics. That's the problem. If box score metrics were better then they could be used as an excellent prior...of course if box score metrics were better, why would we even bother with +/- data in the first place?

Since the goal is to make better predictions using this single stat (xRAPM), it's better to use the box score and +/- together. People do it all the time anyway, since the box score and +/- metrics present their own unique limitations in analysis. But this is a segue into your next post...

Letting you segue...

Well, by mashing everything into a black box, what he's done is create a flawed stat which can't really be used in an informed manner in conjunction with anything else with a great deal of confidence. Whereas, if you keep the the box score and the +/- separate, you've got two orthogonal indicators than an experienced analysis can factor based on their knowledge of how the two tend to work along with a variety of other factors.

Fair point. On that note, it would be interesting to see if the xRAPM blend would lead to better predictions in blends with other metrics than pure +/- and other metrics, and whether it's increased predictive power over RAPM makes up for the possible overfitting.

Then again, I use everything in analysis anyway (and this is not restricted to the metrics), so it's a bit of a "meh" issue from my point of view. And I still like xRAPM more than RAPM as a predictor despite your reservations.

I suppose what I think the key thing here is that talk of predictive power pretty much always focuses on what does better on average as if we're going to pick only one of these stats to be THE official stat to use all of the time. The reality for me is that I know where the issues are likely to be and tailor my analysis of a given player accordingly. For some comparisons, to the extent I was using a RAPM-based stat at all, the box score component wouldn't bother me in the slightest and I'd welcome the increased reliability. For others it makes the stat utterly worthless.

The fact that I can intelligently recognize when these issues come to the forefront and pick the best-of-breed at any given time makes it basically a given that I'm going to do better using my multi-tooled approach than what the Lord of the Stats winner-take-all approach could give.

On that note, you've almost certainly sensed my antipathy toward xRAPM. The issue isn't that I'm against xRAPM as a new tool added to the mix, but that it came on the seen as a replacement for RAPM, and the old data was literally deleted from the internet for no reason other than the statistician in question quite literally didn't understand why anyone would want the old data now...which said quite a lot about how different his perspective was from an experienced analyst.

If Player A has a weaker supporting cast than Player B, but Player A's team does better than Player B, then Player A has "lifted" his team more than Player B.

This statement doesn't consider the interactive effects and the phenomenon of diminishing returns in basketball. It's easier to "lift" a weaker supporting cast; harder to lift a stronger one. (It's also harder to lift a supporting cast that doesn't "fit".) One must account for these things before making any player comparison.

Where did I say I make player comparisons only on lift? In fact you've already seen me talk about lift being distinct from impact which was distinct from value which was distinct from goodness. This specialized vocabulary might seem absurd, but people use these terms to mean similar but slightly different things. What I'm doing is nailing them down not holding any of them up as the whole picture (not even "goodness" since the term implies there's a ideal numerical value that represents a player which is obviously oversimplistic).

I feel like the crux of the point is this:

You say we don't really know anything about LeBron's limitations because the fit in Miami is so terrible.

If you're talking about "fit" with Wade and Bosh, Bosh is playing his best offensive season since '10 in Toronto; and since Barkley infamously called out Wade in the Knicks blowout win vs. the Heat, Wade's skill curve has been near his performance during the '10 season. This is despite Wade being limited by with knee injuries since LeBron joined the Heat, LeBron and Wade sharing similar skill sets, and all three players changing their roles to some degree in a new system under Spoelstra. For the entire '13 season, the Big 3 together are close to what they did individually on offense by some metrics. It isn't all about RAPM (which you poster earlier), since that stat IS LIMITED. When you consider all the metrics, gametape, and diminishing returns in basketball, they're playing quite well.

No, DocMJ; what I'm referring to is your criticism of the Heat's ortg since LeBron joined the team. The Big 3 have played well, but you grossly overrate the rest of the supporting cast (players who have significant input into any team's ortg), which I will show later on.

Okay.

I say that I still expected better, and that I'm cautious about making assumptions about what ultra-outlier offensive performances LeBron could have given we haven't seen it, and that the issue of talent redundancy that we see in Miami would seem to be something that exists with any team capable of reaching those unheard of levels.

You then say "But you don't really know for sure", and I say "You're right".

I don't think you're just merely being skeptical. It's one thing to wonder how a player would hypothetically perform with a good offensive team with the ideal ingredients in place, but from what I've been reading in your posts you seem to go as far as to assume that that player isn't capable of this (LeBron) compared to another player who is playing with these ideal ingredients (Durant). And no, don't keep insisting that Durant/Westbrook is a bad "fit". Even with Westbrook's flaws, those players have more distinct, complementary games than James/Wade. This IS NOT a fair comparison, and you know it.

Put Durant with any of the other 3 players mentioned and you have a better fit than the other two guys left, and that's not a coincidence. What I'm alleging is that the advantage that you see as a coincidental unfair advantage in the comparison is pretty much what we'd expect in general.

When you have a player like LeBron who CAN do it all, that means he only gets to be his full self when he DOES do it all, which means other players being relegated to satellites. This means he's particularly prone to diminishing returns with other talent.

I don't really see how this is debatable. The only question is how severe the issue is. I think it's fine to say you don't think it's that big of a deal, but obviously I'm not willing to dismiss it so easily.

Hence my use of "just taken as an ideal". Clearly you're saying we have to do more than that, and clearly based on the amount of my writing, I am doing more than that, but we're also having issues with fundamental agreement on basic semantics so I have to lay these things down from time to time.

Writing a lot doesn't mean you're actually doing sound analysis. I'll get to that later.

I repeat: We are having fundamental issues in our discussion here. I'm having to define vocabulary terms and even when I do you respond with arguments that make clear you make all sorts of assumptions of the implication of me offering a benign clarification.

You want to dive in details but if we can't get straight exactly what we're debating we won't get anywhere.

And to be clear: The fact that that is the case is only possible because of how much knowledge we both have. It's a bit of me giving you a compliment, even as I get frustrated at how resistant you are to actually pinning down your knowledge with a precise schema.

Would the team be better if Wade was a more capable playmaker? Surely.

Do you think Westbrook is an elite level playmaker?

I mean it's working well with the two of them, but I can't really fathom the idea that if you've got a world-class off-ball player you wouldn't love to have the best playmaker you can next to him.

Let me take this same post and change some words here:

Would the team be better if Wade was a more capable shooter? Surely.

Do you think Wade is an elite level shooter?

I mean it's working well with the two of them, but I can't really fathom the idea that if you've got a world-class playmaker you wouldn't love to have the best shooter you can next to him.

Now, if you still claim that you cannot imagine how this can help LeBron and the Heat, you're not being honest.

So you're saying it's not fair to judge LeBron here until he has a teammate with all of Dwyane Wade's other abilities as well as being an elite-level shooting ability?

Or are you saying that the equivalent need for LeBron is someone who is just a world-class shooter and that makes him a superstar on that alone, in other words, we can't say LeBron has as good of a supporting cast as Durant until LeBron has someone like Durant as a sidekick?

Or are you saying LeBron doesn't need another superstar at all, just some really great shooters who aren't good enough in other areas to be called superstar sidekicks and thus don't find themselves underemployed by his presence, in other words saying that LeBron just can't make proper use of an offensive superstar next to him?

Let's consider Battier here.

This is a guy long considered probably the single greatest role player in the entire league. Does all the little things. Phenomenal defender, and certainly capable of hitting the open trey. I understand he's getting older, and maybe that's your point, but isn't he in theory exactly the type of guy you'd try to slot in next to the Big 3?

Chalmers? Hey, if you're going to have your 2 & 3 be the dictators of the offense, what are you expecting to get at the 1?

Haslem? The Big 3 took paycuts specifically so that he would stay because he seemed the precise type of player they needed.

I feel like you're looking at the things going wrong and saying "If those weren't going wrong, it would all be great", and I'm looking at them and saying "I don't think it's as easy as you think to reliably build better".

Haslem, who is usually a solid midrange shooter, is shooting -8% (compared to '12) on jumpers this season. Chalmers has never been a traditional PG even before LeBron joined the Heat, he would run the occasional play but his main talent comes from playing off-ball and shooting open shots. He hasn't shot the ball as well either compared to '12 and has even taken a step back in his free throw shooting (-10% from the line despite having the same free throw rate as '12). Are you blaming LeBron for Chalmers's regression from the line, too? Battier has shot the ball better, but he's been more of a black hole on offense and has not rebounded the ball as well. You could put some of that on Spo's insistence on small ball, but that means that the coaching staff is not utilizing Battier properly; only a fool would blame LeBron for this. LeBron loves to find the open man, and Battier doesn't have to shoot over 80% of his shots from 3 when he's not a Ray Allen from that range and he's a good mid-range shooter. He showed he could hit the mid-range shot playing with the Big 3 in '12!

Excuse you, but while you've been going on about "what-ifs" with Durant here and "easily" imaging him playing with a player with better skills and playing in a better offense; you somehow fail to see how a better Chalmers, Haslem, and Battier wouldn't help the Heat's offense. Those players have never been asked to do alot on offense and their roles in Miami are clearly defined. The world's best players are taking away attention away from them and they're not capitalizing on their shot opportunities, and in Battier's case he's not being used properly by Spoelstra who is responsible for being his team's talents together. And this is what I'm talking about when I say that you're not doing as much sound analysis as you think. You look at numbers and draw conclusions, I look at numbers AND watch the on-court actions on the floor AND the coaching personnel decisions that produce those numbers you look at and draw conclusions from. And these are good examples of how the Heat's ortg above league average is being affected by things that LeBron literally has no control over. These role players help the Heat with the lineups they can put around the Big 3; when 5-man lineups are weak in these areas it limits the effectiveness of the team, even while playing with stars.

Talking about changes to aspects in the Thunder offense and personnel as it pertains to Durant while ignoring the Heat offense with LeBron (which is what you've been doing, sorry DocMJ) is biased thinking.

I'll leave it there. Too busy to type up more lengthy, 30-page replies; and perhaps too lazy to type up everything otherwise

See, tons of knowledge in what you're saying. Totally glad I can read your thoughts.

In the end though I still think that leaves us with a kind of agree to disagree scenario.

I understand why someone says "I see the problems happening, it can be better...a LOT better". Aside from my general cautious streak that I've described above, clearly my view on role players in team building is a bit different from yours.

When Miami got their Big 3 together, there was a serious question how they could sign anyone else worth a damn. The cap made things so tight that it really handicapped them, and that's even the case despite the fact they've been able to get guys for under-market salary. On the whole, when they've made moves, people've said "Smart move", but time and again there are times when these players look terrible in the situation they are in.

Part of that has to do with fitting in with the existing talent, part of that has to do with players who maybe aren't what they used to be, and part of that has to do with the fact that beggars can't be choosers.

So your damn right I'm imagining a better version of Westbrook. Teams in the NBA can acquire a second star, and they can get a guy who fits well with the top star. It's not trivial to do, but there's nothing utterly unrealistic about it.

Talking about the 4th, 5th, and 6th guys on a roster though when you have huge amounts of money tied up with the first 3 just seems like such a pipe dream. Whoever you get is going to be flawed, and you're going to end up noticing those flaws.

[Doctor MJ]

Just replied to a couple of your posts, DocMJ. PM me if you wish, however; currently busy and can't respond to every post in detail here.

No worries mate.

Lets make this crystal-clear before I proceed: throughout his exchange (or elsewhere, for that matter), I have not considered you to be an idiot.

Noted, and right back at you. I consider you to be a very solid poster.

Now, with regards to the priors: you're right, the introduction of the priors muddies the validity of the metric. If I JUST want to see what the player in quesiton has done, then yes; non-prior data is more useful and I'll concede your point. But in actuality, stats that are great at telling you what happened don't exactly have any significant, practical purpose. (Per Alex's from Sports Skeptic great rundown on the various metrics available for public consumption, RAPM isn't even the most useful metric here). It's far better for a model to be able to predict things instead. We're halfway through the season, and Durant's "clear improvement" in 35 games or so is dwarfed by his track record from '10-'12. I cannot simply assume his much smaller sample size from this season is the new "norm". I want to know what he'll do going forward, and therefore I want those priors.

Note that priors are literally just more of "what happened". They improve prediction because they in general improve reliability, and as you conceded, there are scenarios where there are important validity issues with them. Where validity can be pinpointed to as clearly flawed you can very quickly reach the point where any reliability argument makes no sense.

If we're seeing something fundamentally new from Durant this year then there's no reason to look at previous seasons as something we expect him to regress to. One can use those seasons as a baseline for certainty of what his minimal levels habe to be, but not to use them against him.

Additionally, Skeptics' analysis is based on the idea of seeing which stat is closest to being The Lord of the Stats to rule them all. If I want one number to predict X, which will I use?

That's cool and all, but no analysis worth a salt should limit himself to one number, so it doesn't really damn RAPM to "lose" this one. I like +/- stats because of how well they work with other factors not because I can ignore other factors because of them.

Now as I say this, you're probably thinking, "Find Doc, but you talk about +/- an awful lot." Point taken, but note that I'm very careful with how I apply it. I use it prominently when I see examples that show trends to significant to ignore, but I certainly don't jump into every player comparison and side with a guy simply by throwing down a +/- stat.

We're specifically talking about a player who we both agree has changed significantly from the previous year.

Can't agree with that. His production is certainly superior (so far) but in essence his role hasn't changed significantly; he's simply refined the facets of his game at the SF position.

Oh okay. Well that would be a reason for disagreement then. Not that I really care to debate what constitutes a "role change", but when we see dramatic changes to areas of both a guy's box score and +/- stats, and this is happening for the good after his team lost a major talent from the previous year, I can't really fathom saying there hasn't been some significant change relating to that player.

I understand that sample size can be an issue here, but the season really isn't that young any more. Not that young for +/- considering how much of an outlier the numbers are, and really not that young for box score stats which tend to require far less sample size in order to achieve solid reliability.

RE: box score priors. These aren't a big issue with regards to Durant, but you brought them up relating to Nash who is tremendously underrated by box score metrics. That's the problem. If box score metrics were better then they could be used as an excellent prior...of course if box score metrics were better, why would we even bother with +/- data in the first place?

Since the goal is to make better predictions using this single stat (xRAPM), it's better to use the box score and +/- together. People do it all the time anyway, since the box score and +/- metrics present their own unique limitations in analysis. But this is a segue into your next post...

Letting you segue...

Well, by mashing everything into a black box, what he's done is create a flawed stat which can't really be used in an informed manner in conjunction with anything else with a great deal of confidence. Whereas, if you keep the the box score and the +/- separate, you've got two orthogonal indicators than an experienced analysis can factor based on their knowledge of how the two tend to work along with a variety of other factors.

Fair point. On that note, it would be interesting to see if the xRAPM blend would lead to better predictions in blends with other metrics than pure +/- and other metrics, and whether it's increased predictive power over RAPM makes up for the possible overfitting.

Then again, I use everything in analysis anyway (and this is not restricted to the metrics), so it's a bit of a "meh" issue from my point of view. And I still like xRAPM more than RAPM as a predictor despite your reservations.

I suppose what I think the key thing here is that talk of predictive power pretty much always focuses on what does better on average as if we're going to pick only one of these stats to be THE official stat to use all of the time. The reality for me is that I know where the issues are likely to be and tailor my analysis of a given player accordingly. For some comparisons, to the extent I was using a RAPM-based stat at all, the box score component wouldn't bother me in the slightest and I'd welcome the increased reliability. For others it makes the stat utterly worthless.

The fact that I can intelligently recognize when these issues come to the forefront and pick the best-of-breed at any given time makes it basically a given that I'm going to do better using my multi-tooled approach than what the Lord of the Stats winner-take-all approach could give.

On that note, you've almost certainly sensed my antipathy toward xRAPM. The issue isn't that I'm against xRAPM as a new tool added to the mix, but that it came on the seen as a replacement for RAPM, and the old data was literally deleted from the internet for no reason other than the statistician in question quite literally didn't understand why anyone would want the old data now...which said quite a lot about how different his perspective was from an experienced analyst.

If Player A has a weaker supporting cast than Player B, but Player A's team does better than Player B, then Player A has "lifted" his team more than Player B.

This statement doesn't consider the interactive effects and the phenomenon of diminishing returns in basketball. It's easier to "lift" a weaker supporting cast; harder to lift a stronger one. (It's also harder to lift a supporting cast that doesn't "fit".) One must account for these things before making any player comparison.

Where did I say I make player comparisons only on lift? In fact you've already seen me talk about lift being distinct from impact which was distinct from value which was distinct from goodness. This specialized vocabulary might seem absurd, but people use these terms to mean similar but slightly different things. What I'm doing is nailing them down not holding any of them up as the whole picture (not even "goodness" since the term implies there's a ideal numerical value that represents a player which is obviously oversimplistic).

I feel like the crux of the point is this:

You say we don't really know anything about LeBron's limitations because the fit in Miami is so terrible.

If you're talking about "fit" with Wade and Bosh, Bosh is playing his best offensive season since '10 in Toronto; and since Barkley infamously called out Wade in the Knicks blowout win vs. the Heat, Wade's skill curve has been near his performance during the '10 season. This is despite Wade being limited by with knee injuries since LeBron joined the Heat, LeBron and Wade sharing similar skill sets, and all three players changing their roles to some degree in a new system under Spoelstra. For the entire '13 season, the Big 3 together are close to what they did individually on offense by some metrics. It isn't all about RAPM (which you poster earlier), since that stat IS LIMITED. When you consider all the metrics, gametape, and diminishing returns in basketball, they're playing quite well.

No, DocMJ; what I'm referring to is your criticism of the Heat's ortg since LeBron joined the team. The Big 3 have played well, but you grossly overrate the rest of the supporting cast (players who have significant input into any team's ortg), which I will show later on.

Okay.

I say that I still expected better, and that I'm cautious about making assumptions about what ultra-outlier offensive performances LeBron could have given we haven't seen it, and that the issue of talent redundancy that we see in Miami would seem to be something that exists with any team capable of reaching those unheard of levels.

You then say "But you don't really know for sure", and I say "You're right".

I don't think you're just merely being skeptical. It's one thing to wonder how a player would hypothetically perform with a good offensive team with the ideal ingredients in place, but from what I've been reading in your posts you seem to go as far as to assume that that player isn't capable of this (LeBron) compared to another player who is playing with these ideal ingredients (Durant). And no, don't keep insisting that Durant/Westbrook is a bad "fit". Even with Westbrook's flaws, those players have more distinct, complementary games than James/Wade. This IS NOT a fair comparison, and you know it.

Put Durant with any of the other 3 players mentioned and you have a better fit than the other two guys left, and that's not a coincidence. What I'm alleging is that the advantage that you see as a coincidental unfair advantage in the comparison is pretty much what we'd expect in general.

When you have a player like LeBron who CAN do it all, that means he only gets to be his full self when he DOES do it all, which means other players being relegated to satellites. This means he's particularly prone to diminishing returns with other talent.

I don't really see how this is debatable. The only question is how severe the issue is. I think it's fine to say you don't think it's that big of a deal, but obviously I'm not willing to dismiss it so easily.

Hence my use of "just taken as an ideal". Clearly you're saying we have to do more than that, and clearly based on the amount of my writing, I am doing more than that, but we're also having issues with fundamental agreement on basic semantics so I have to lay these things down from time to time.

Writing a lot doesn't mean you're actually doing sound analysis. I'll get to that later.

I repeat: We are having fundamental issues in our discussion here. I'm having to define vocabulary terms and even when I do you respond with arguments that make clear you make all sorts of assumptions of the implication of me offering a benign clarification.

You want to dive in details but if we can't get straight exactly what we're debating we won't get anywhere.

And to be clear: The fact that that is the case is only possible because of how much knowledge we both have. It's a bit of me giving you a compliment, even as I get frustrated at how resistant you are to actually pinning down your knowledge with a precise schema.

Would the team be better if Wade was a more capable playmaker? Surely.

Do you think Westbrook is an elite level playmaker?

I mean it's working well with the two of them, but I can't really fathom the idea that if you've got a world-class off-ball player you wouldn't love to have the best playmaker you can next to him.

Let me take this same post and change some words here:

Would the team be better if Wade was a more capable shooter? Surely.

Do you think Wade is an elite level shooter?

I mean it's working well with the two of them, but I can't really fathom the idea that if you've got a world-class playmaker you wouldn't love to have the best shooter you can next to him.

Now, if you still claim that you cannot imagine how this can help LeBron and the Heat, you're not being honest.

So you're saying it's not fair to judge LeBron here until he has a teammate with all of Dwyane Wade's other abilities as well as being an elite-level shooting ability?

Or are you saying that the equivalent need for LeBron is someone who is just a world-class shooter and that makes him a superstar on that alone, in other words, we can't say LeBron has as good of a supporting cast as Durant until LeBron has someone like Durant as a sidekick?

Or are you saying LeBron doesn't need another superstar at all, just some really great shooters who aren't good enough in other areas to be called superstar sidekicks and thus don't find themselves underemployed by his presence, in other words saying that LeBron just can't make proper use of an offensive superstar next to him?

Let's consider Battier here.

This is a guy long considered probably the single greatest role player in the entire league. Does all the little things. Phenomenal defender, and certainly capable of hitting the open trey. I understand he's getting older, and maybe that's your point, but isn't he in theory exactly the type of guy you'd try to slot in next to the Big 3?

Chalmers? Hey, if you're going to have your 2 & 3 be the dictators of the offense, what are you expecting to get at the 1?

Haslem? The Big 3 took paycuts specifically so that he would stay because he seemed the precise type of player they needed.

I feel like you're looking at the things going wrong and saying "If those weren't going wrong, it would all be great", and I'm looking at them and saying "I don't think it's as easy as you think to reliably build better".

Haslem, who is usually a solid midrange shooter, is shooting -8% (compared to '12) on jumpers this season. Chalmers has never been a traditional PG even before LeBron joined the Heat, he would run the occasional play but his main talent comes from playing off-ball and shooting open shots. He hasn't shot the ball as well either compared to '12 and has even taken a step back in his free throw shooting (-10% from the line despite having the same free throw rate as '12). Are you blaming LeBron for Chalmers's regression from the line, too? Battier has shot the ball better, but he's been more of a black hole on offense and has not rebounded the ball as well. You could put some of that on Spo's insistence on small ball, but that means that the coaching staff is not utilizing Battier properly; only a fool would blame LeBron for this. LeBron loves to find the open man, and Battier doesn't have to shoot over 80% of his shots from 3 when he's not a Ray Allen from that range and he's a good mid-range shooter. He showed he could hit the mid-range shot playing with the Big 3 in '12!

Excuse you, but while you've been going on about "what-ifs" with Durant here and "easily" imaging him playing with a player with better skills and playing in a better offense; you somehow fail to see how a better Chalmers, Haslem, and Battier wouldn't help the Heat's offense. Those players have never been asked to do alot on offense and their roles in Miami are clearly defined. The world's best players are taking away attention away from them and they're not capitalizing on their shot opportunities, and in Battier's case he's not being used properly by Spoelstra who is responsible for being his team's talents together. And this is what I'm talking about when I say that you're not doing as much sound analysis as you think. You look at numbers and draw conclusions, I look at numbers AND watch the on-court actions on the floor AND the coaching personnel decisions that produce those numbers you look at and draw conclusions from. And these are good examples of how the Heat's ortg above league average is being affected by things that LeBron literally has no control over. These role players help the Heat with the lineups they can put around the Big 3; when 5-man lineups are weak in these areas it limits the effectiveness of the team, even while playing with stars.

Talking about changes to aspects in the Thunder offense and personnel as it pertains to Durant while ignoring the Heat offense with LeBron (which is what you've been doing, sorry DocMJ) is biased thinking.

I'll leave it there. Too busy to type up more lengthy, 30-page replies; and perhaps too lazy to type up everything otherwise

See, tons of knowledge in what you're saying. Totally glad I can read your thoughts.

In the end though I still think that leaves us with a kind of agree to disagree scenario.

I understand why someone says "I see the problems happening, it can be better...a LOT better". Aside from my general cautious streak that I've described above, clearly my view on role players in team building is a bit different from yours.

When Miami got their Big 3 together, there was a serious question how they could sign anyone else worth a damn. The cap made things so tight that it really handicapped them, and that's even the case despite the fact they've been able to get guys for under-market salary. On the whole, when they've made moves, people've said "Smart move", but time and again there are times when these players look terrible in the situation they are in.

Part of that has to do with fitting in with the existing talent, part of that has to do with players who maybe aren't what they used to be, and part of that has to do with the fact that beggars can't be choosers.

So your damn right I'm imagining a better version of Westbrook. Teams in the NBA can acquire a second star, and they can get a guy who fits well with the top star. It's not trivial to do, but there's nothing utterly unrealistic about it.

Talking about the 4th, 5th, and 6th guys on a roster though when you have huge amounts of money tied up with the first 3 just seems like such a pipe dream. Whoever you get is going to be flawed, and you're going to end up noticing those flaws.

[SideshowBob]

Gosh Doc, quintuple post much

[Doctor MJ]

Gosh Doc, quintuple post much

lol, jeez and of all the posts to stutter.

[Doctor MJ]

Like I said before DocMJ, if we define MVP that way then we should just give it to the best player on the weakest overall team or "best fit" roster. We don't even need to use stats for that, either.

That's all dandy if people decide to define value that way, but I don't buy it.

This is a pretty weird statement to say given that Durant's team, and their performance with him on the court, is better than LeBron's right now and that's been a part of pretty much everything I've been talking about.

[MisterWestside]

Like I said before DocMJ, if we define MVP that way then we should just give it to the best player on the weakest overall team or "best fit" roster. We don't even need to use stats for that, either.

That's all dandy if people decide to define value that way, but I don't buy it.

This is a pretty weird statement to say given that Durant's team, and their performance with him on the court, is better than LeBron's right now and that's been a part of pretty much everything I've been talking about.

Not so. I think Derrick Rose in '11 (compared to LeBron) is the best example for this. Rose was more valuable to his team largely because of the better roster "fit" in Chicago; they won more, they built around him and they also faltered more without him (compared to LeBron's Heat) on the floor. I also don't think that, despite these things, anyone would doubt that LeBron was still the better player, which is what I care about. Value can be tied into several factors that don't always have anything to do with the individual player, and I seek to filter out those factors.

Thanks for your reply and the spirited yet friendly dialogue. In part it shows that LeBron and Durant are playing at virtually even levels this season and it's hard to separate them. Unlike previous seasons LeBron has super stiff competition for best player, a statement I think we can both agree with.

[fallacy]

Like I said before DocMJ, if we define MVP that way then we should just give it to the best player on the weakest overall team or "best fit" roster. We don't even need to use stats for that, either.

That's all dandy if people decide to define value that way, but I don't buy it.

This is a pretty weird statement to say given that Durant's team, and their performance with him on the court, is better than LeBron's right now and that's been a part of pretty much everything I've been talking about.

Not so. I think Derrick Rose in '11 (compared to LeBron) is the best example for this. Rose was more valuable to his team largely because of the better roster "fit" in Chicago; they built around him and also faltered more without him (compared to LeBron's Heat) on the floor. I also don't think anyone would doubt that LeBron was still the better player, which is what I care about. Value can be tied into several factors that don't always have anything to do with the individual player, and I seek to filter out those factors.

Thanks for your reply and the spirited yet friendly dialogue. In part it shows that LeBron and Durant are playing at virtually even levels this season and it's hard to separate them. Unlike previous seasons LeBron has super stiff competition for best player, a statement I think we can both agree with.

Wait, are you arguing that the MVP should just be the "best player in the NBA" award and you should just disregard factors like record, conference, team value, etc?

[inquisitive]

at this point...i feel it is a toss up between lbj and kd...

[MisterWestside]

Wait, are you arguing that the MVP should just be the "best player in the NBA" award and you should just disregard factors like record, conference, team value, etc?

Yes. Sue me

Even if you disagree, I still have Durant right there with LeBron (or even better) for the season.

[Doctor MJ]

Not so. I think Derrick Rose in '11 (compared to LeBron) is the best example for this. Rose was more valuable to his team largely because of the better roster "fit" in Chicago; they won more, they built around him and they also faltered more without him (compared to LeBron's Heat) on the floor. I also don't think that, despite these things, anyone would doubt that LeBron was still the better player, which is what I care about. Value can be tied into several factors that don't always have anything to do with the individual player, and I seek to filter out those factors.

Thanks for your reply and the spirited yet friendly dialogue. In part it shows that LeBron and Durant are playing at virtually even levels this season and it's hard to separate them. Unlike previous seasons LeBron has super stiff competition for best player, a statement I think we can both agree with.

Appreciate your tone. I have enjoyed the exchange.

Responding to this and to your reply to fallacy, you want the "Most Valuable Player" award to have nothing to do with value. Here's where I really see that as an issue (aside from the fact that you could have said that up front):

There is a serious danger when you do this for making up your mind about someone and then just making excuses for everything that happens that goes against this.

When we did the Retro POY project on the PC board I defined the terms loosely but said that it had to be based on the good you were doing your team in that year. It worked fine until we got to Wilt Chamberlain where I had a guy who basically insisted on voting for Wilt #1 no matter what.

His reasoning was easy to understand: "I think Wilt was the best player period. The fact that stuff went wrong in a particular year is just stuff that's bound to happen in a team game where things are out of your control."

So, every year to this guy, Wilt beats Russell...despite the fact that quite literally Russell kills Wilt in most years in terms of what he's doing for his team and the reasons for this are quite clear and precise and have everything to do with weaknesses in Wilt.

LeBron's issues are not anywhere near so severe, but the fundamental principle remains.

[fallacy]

Wait, are you arguing that the MVP should just be the "best player in the NBA" award and you should just disregard factors like record, conference, team value, etc?

Yes. Sue me

Even if you disagree, I still have Durant right there with LeBron (or even better) for the season.

Nah, I just don't get why you're arguing about an award while changing the entire definition of the award. I don't understand the endgame.

[JordansBulls]

Wait, are you arguing that the MVP should just be the "best player in the NBA" award and you should just disregard factors like record, conference, team value, etc?

Yes. Sue me

Even if you disagree, I still have Durant right there with LeBron (or even better) for the season.

It would be wrong to change the award now to the best player in the league when they didn't do it for others in the past.

[fallacy]

Wait, are you arguing that the MVP should just be the "best player in the NBA" award and you should just disregard factors like record, conference, team value, etc?

Yes. Sue me

Even if you disagree, I still have Durant right there with LeBron (or even better) for the season.

It would be wrong to change the award now to the best player in the league when they didn't do it for others in the past.

They could just change the name to "The Jordan Award" and go back and give Jordan six more