Jokic GOAT Peak?

[MacGill]

I know that in the modern game of all advanced stats now most cannot abort programming if their eye results and numbers produced don't agree, so the numbers won't lie , but I will say, probably the greatest peak ever with the absolute least amount of all-star, superstar support. Clearly the rules of the league, style & tempo have shifted back into the 60's style pace but modern formula's make many conclude judgements that others just cannot see and Jokic looks like he took a 'hottub' time machine and went into the future. Let's see if he ends up playing with a few superstars before his time is up and how the usual suspects will bend the formula's against him. I dislike this era, and the celebration of underachieving designed to look like overachieving but Jokic's game and peak is that good. No teammate BS, Coach BS, right system BS, the dude is what everyone wanting to play the game should try to resemble. Of course, he gets minimal luv by the mainstream, but he absolutely deserves to be at or near the top and will probably go down as the best of the 20's when it is all said and done.

[eminence]

I'm sad that JE (and many others) have continued to go further down the rabbit hole of further tweaking the tried-and-true APMs in their quest for improvements.

There's more to life than regression, live a little.

[PooledSilver]

I'm sad that JE (and many others) have continued to go further down the rabbit hole of further tweaking the tried-and-true APMs in their quest for improvements.

There's more to life than regression, live a little.

At the end of the day people just want it to be more predictive than descriptive since one’s much more practical, as far as I know LEBRONs the current king when it comes to win projections (saw ur name in that hat there too lmfao) but in any case I get people wanting to keep things pure and somewhat tainting impact data by trying to make it an all in one thing rather than a specific tool in the toolbox, but the ability for people to practically use that is a pain and the noise makes it hard for people to buy in

[OhayoKD]

I'm sad that JE (and many others) have continued to go further down the rabbit hole of further tweaking the tried-and-true APMs in their quest for improvements.

There's more to life than regression, live a little.

At the end of the day people just want it to be more predictive than descriptive since one’s much more practical, as far as I know

Not going to pretend I have a great understanding of all the math that goes into it, but Cheema remains my favorite set if for no other reason than the methodology and math being completely transparent.

It's also convenient to have the possession-totals visualized in a graph

LEBRONs the current king when it comes to win projections (saw ur name in that hat there too lmfao) but in any case I get people wanting to keep things pure and somewhat tainting impact data by trying to make it an all in one thing rather than a specific tool in the toolbox, but the ability for people to practically use that is a pain and the noise makes it hard for people to buy in

I thought the king was DARKO?

[PooledSilver]

I'm sad that JE (and many others) have continued to go further down the rabbit hole of further tweaking the tried-and-true APMs in their quest for improvements.

There's more to life than regression, live a little.

At the end of the day people just want it to be more predictive than descriptive since one’s much more practical, as far as I know

Not going to pretend I have a great understanding of all the math that goes into it, but Cheema remains my favorite set if for no other reason than the methodology and math being completely transparent.

It's also convenient to have the possession-totals visualized in a graph

LEBRONs the current king when it comes to win projections (saw ur name in that hat there too lmfao) but in any case I get people wanting to keep things pure and somewhat tainting impact data by trying to make it an all in one thing rather than a specific tool in the toolbox, but the ability for people to practically use that is a pain and the noise makes it hard for people to buy in

I thought the king was DARKO?

The public test DARKO was up too since LEBRON changed it completely destroyed everything else once I checked

I have some issues with it on a case by case basis, but overall I like it more than most other ones. At least their methodology is all up there

[eminence]

I'm sad that JE (and many others) have continued to go further down the rabbit hole of further tweaking the tried-and-true APMs in their quest for improvements.

There's more to life than regression, live a little.

At the end of the day people just want it to be more predictive than descriptive since one’s much more practical, as far as I know LEBRONs the current king when it comes to win projections (saw ur name in that hat there too lmfao) but in any case I get people wanting to keep things pure and somewhat tainting impact data by trying to make it an all in one thing rather than a specific tool in the toolbox, but the ability for people to practically use that is a pain and the noise makes it hard for people to buy in

It's a fun little contest

As of last check I was stinking it up this season, been trying to apply my area to the predictions the last couple of years and haven't quite made it work. This season I went with a plus/minus 'all-in-one' built from hypergraph centrality as my entry, then applied much like one would an xRAPM measure.

Had my most success in the contest when I completely ignored individual players and went with team based vibes. Read - qualitative descriptors/variables of teams sourced from various sites around the web like RealGM.

[PooledSilver]

I'm sad that JE (and many others) have continued to go further down the rabbit hole of further tweaking the tried-and-true APMs in their quest for improvements.

There's more to life than regression, live a little.

At the end of the day people just want it to be more predictive than descriptive since one’s much more practical, as far as I know LEBRONs the current king when it comes to win projections (saw ur name in that hat there too lmfao) but in any case I get people wanting to keep things pure and somewhat tainting impact data by trying to make it an all in one thing rather than a specific tool in the toolbox, but the ability for people to practically use that is a pain and the noise makes it hard for people to buy in

It's a fun little contest

As of last check I was stinking it up this season, been trying to apply my area to the predictions the last couple of years and haven't quite made it work. This season I went with a plus/minus 'all-in-one' built from hypergraph centrality as my entry, then applied much like one would an xRAPM measure.

Had my most success in the contest when I completely ignored individual players and went with team based vibes. Read - qualitative descriptors/variables of teams sourced from various sites around the web like RealGM.

Curious if you beat Vegas haha

I think the all in one idea to predict team success and combining that with some outside knowledge of fit and role built in and weighting the two might be a move

As far as I know LEBRON seems pretty simple for doing as well as it has thus far, although I’ve heard they changed up the luck adjustments

Imagine the money you could make if you had these models before Vegas knew what was up

[eminence]

At the end of the day people just want it to be more predictive than descriptive since one’s much more practical, as far as I know LEBRONs the current king when it comes to win projections (saw ur name in that hat there too lmfao) but in any case I get people wanting to keep things pure and somewhat tainting impact data by trying to make it an all in one thing rather than a specific tool in the toolbox, but the ability for people to practically use that is a pain and the noise makes it hard for people to buy in

It's a fun little contest

As of last check I was stinking it up this season, been trying to apply my area to the predictions the last couple of years and haven't quite made it work. This season I went with a plus/minus 'all-in-one' built from hypergraph centrality as my entry, then applied much like one would an xRAPM measure.

Had my most success in the contest when I completely ignored individual players and went with team based vibes. Read - qualitative descriptors/variables of teams sourced from various sites around the web like RealGM.

Curious if you beat Vegas haha

I think the all in one idea to predict team success and combining that with some outside knowledge of fit and role built in and weighting the two might be a move

As far as I know LEBRON seems pretty simple for doing as well as it has thus far, although I’ve heard they changed up the luck adjustments

Imagine the money you could make if you had these models before Vegas knew what was up

Just gotta make the models yourself

I'm behind Vegas in the contest so far this season (4-2 vs Vegas in the 6 prior seasons), but I didn't put my money where my mouth was, so no skin off my teeth. I prefer retirement

Generally I found it notably easier to find a game or two a night to beat Vegas on than to beat them at the season long stuff.

Went back, and yep, the pure team vibes entry was my strongest finish - my first contest in '17-'18.

The only two model inputs were:
Team Win% the prior season
One of 5 categorical variables based off fan surveying

The categories:
Very Negative
Negative
Neutral
Positive
Very Positive

[lessthanjake]

You completely ignored what I said, but paragraph by paragraph

1. Applying an aging curve helping Lebron is valid, but Jokic falls from 1st to 7th, and we have RAPM throughout his prime too.

More than this though, you ignore the points that Jokic’s pre prime RAPM is WAY better than most people’s, it’s both a testament to his early year success being pretty high but also inflates the overall career data where almost everyone else has pre or post prime years bringing them down

2. You completely ignored what I said, your breakdown would be sensible if you looked at the actual data but it simply isn’t whatsoever. In the context of RAPM Jokic’s “pre-prime” years inflate his data relative to everyone else more than hurting him because his pre-prime years were very solid impact wise ESPECIALLY for being his “bad” years. You can’t stand by what you said because there’s nothing to stand on its blatantly being misleading and ignoring the reality based on ur preconceived notion, and that’s fine untill someone brings it up, by blatantly ignoring this point now you are wasting your own time and mine.

Is that really true though? If we look at on-off (which obviously pretty closely resembles RAPM and is something I have easily at my fingertips as I’m writing this), Jokic in his true prime is *way* higher than Jokic before that. What is true is that the years I labeled not-at-all prime aren’t worse (and are actually better) than the years I labeled weak prime. But his true prime years are, on average, a lot higher, and his true prime years are a pretty abnormally low percent of his minutes. Not sure off the top of my head whether this would translate exactly from on-off to RAPM, but I suspect it would generally be the case.

But yes, you are right that just looking at percent of one’s career that was true prime vs. weak prime vs. not-at-all prime doesn’t really fully get at exactly what the difference would be between prime/peak RAPM and career RAPM. We’d generally *expect* that someone who has a lower percent of his career being prime years in the data set and ends up higher overall in career RAPM probably had a higher RAPM in the peak or prime years. But of course, along the lines you discuss, that’s not *necessarily* the case, because maybe that player’s numbers in those non-prime years are way better than others’ non-prime years and that gets them ahead overall despite the above. However, I wasn’t suggesting that being #1 in JE’s 1997-2024 RAPM *must* mean Jokic has the best peak RAPM (though I grant that my wording might’ve been confusing on this front). I would say it’s hard to imagine someone who is #1 overall but yet didn’t have an incredibly high peak RAPM (and we also know Jokic has graded out extremely well in other impact metrics in recent years anyways, so we independently know his peak impact numbers are incredibly good). Meanwhile, ultimately we can’t really directly compare specific RAPM values from one year and another, since the scaling and whatnot isn’t the same year to year. So there’s no perfect method to directly compare RAPM values that different players have had in different seasons. What we know is that Jokic has looked incredibly good in season-by-season impact metrics in recent years, and he has looked incredibly good in a career-long RAPM measure. I think that’s consistent with Jokic having the GOAT peak, though I also think the same could be said about the “impact case” of several other players—there’s enough uncertainty and flaws with impact measures that I don’t think we could ever say there’s only one player with an “impact case” for GOAT peak.

3. this is ridiculous we literally know Lebrons RAPM data always gets a huge boost from him turning it on in the playoffs. I’m pretty sure JE showed data that Lebrons RAPM data went up by 33% in the playoffs from 2014-2017. A fair sample to really capture his peak would weight playoffs more considering he blatantly takes his foot off the pedal and his numbers all universally sky rocket in the playoffs including impact, which is absolutely not a trend we see with Jokic which is the point here

Jokic does not go down in the playoffs in box stats, but definitely has in impact stats. Does the fact that his impact numbers have actually gone down in the playoffs—where the “off” samples are absolutely tiny—mean Jokic isn’t often coasting in the regular season, even though we know his box numbers stay similarly incredible in the playoffs (when most stars have their box numbers get worse because playoff basketball is harder) and he has a reputation for mailing it in in the regular season? You’re free to draw the less favorable conclusion about Jokic based primarily on very low-sample-size data, but I don’t think it’s at all “ridiculous” to come to a different conclusion. And you’re also free to put greater weight on playoff impact, but again it’s not “ridiculous” to not layer on that sort of adjustment.

I don’t understand why you are being purposefully disingenuous here but I’m not trying to have a conversation with someone who willfully ignores and selectively reads what he wants to read

…

My interest in this disingenuous conversation isn’t particularly high, so I don’t know if I will respond anymore since I clearly can tell what kind of conversation this is going to be. Ignoring evidence from a place of genuine ignorance is fine, ignoring evidence in the face of bias and close mindedness is a pain because it’s just a waste of all of our time. Do better. I see through this type of approach.

I’m not sure what I said to you to make you respond this rudely. I thought I was perfectly polite to you (and I believe I’ve been polite to you in this response too), and I’d ask that you do the same for me (and probably won’t respond further if you don’t). We’re just discussing basketball, and I have no desire to have an acrimonious discussion with someone I’ve never previously interacted with, based on a topic that doesn’t really matter. And, to be clear, definitely no hard feelings towards you at all if this is just a one-time thing! You seem to have a lot of interesting insights, and so I’d hope to be able to keep discussing things with you, without this sort of thing.

I’m not someone that blindly follows impact data, I run it myself and see issues with it (none that would help your case here mind you, so don’t try twisting my words here…), but if you are arguing he has goat tier raw RAPM data he absolute does not unless you are purposefully not looking at it within context, which you must be doing considering it has been brought up by multiple people. Lebron and KG both are clearly better looking at it within context

I think “within context” is doing a lot of work there. Impact measures aren’t an exact science, and one significant reason for that is because there will always be significant disagreement on what factors and context to adjust for (and how to adjust for it). If you feel very strongly that we should layer on things into the analysis that end up leaving Jokic looking less good relative to others, then you may well be able to come to a personal conclusion that Jokic’s impact data isn’t GOAT-tier. To me, though, I think we’re talking about a very inexact science where data can be massaged a million different ways, and if Jokic is sitting at #1 in JE’s initial raw RAPM calculation and also has looked incredible overall in various season-by-season impact metrics, then I’m comfortable in concluding that that’s indicative of GOAT-tier impact, even if there’s other players that also have data indicative of GOAT-tier impact and there’s other ways to look at the data that doesn’t have Jokic at the top. Basically, I think we should think of these things like they have a confidence interval, and I think what I see in Jokic’s data overall suggests that having GOAT-tier impact is definitely within the confidence interval, as it is for several other players as well. And, crucially, I’m not really making an affirmative argument based on impact data. Rather, what I was doing was disputing the idea that Jokic gets “clobbered” and has “no impact case.” I think that’s wrong. But I also don’t think that impact numbers necessarily suggest on their own that Jokic is at the very top. As I’ve repeatedly said in this thread, I think Jokic has the GOAT peak primarily based on my eye test—and I think his impact numbers and box numbers are consistent with that conclusion because I think he’s within the confidence interval on those.

[PooledSilver]

You completely ignored what I said, but paragraph by paragraph

1. Applying an aging curve helping Lebron is valid, but Jokic falls from 1st to 7th, and we have RAPM throughout his prime too.

More than this though, you ignore the points that Jokic’s pre prime RAPM is WAY better than most people’s, it’s both a testament to his early year success being pretty high but also inflates the overall career data where almost everyone else has pre or post prime years bringing them down

2. You completely ignored what I said, your breakdown would be sensible if you looked at the actual data but it simply isn’t whatsoever. In the context of RAPM Jokic’s “pre-prime” years inflate his data relative to everyone else more than hurting him because his pre-prime years were very solid impact wise ESPECIALLY for being his “bad” years. You can’t stand by what you said because there’s nothing to stand on its blatantly being misleading and ignoring the reality based on ur preconceived notion, and that’s fine untill someone brings it up, by blatantly ignoring this point now you are wasting your own time and mine.

Is that really true though? If we look at on-off (which obviously pretty closely resembles RAPM and is something I have easily at my fingertips as I’m writing this), Jokic in his true prime is *way* higher than Jokic before that. What is true is that the years I labeled not-at-all prime aren’t worse (and are actually better) than the years I labeled weak prime. But his true prime years are, on average, a lot higher, and his true prime years are a pretty abnormally low percent of his minutes. Not sure off the top of my head whether this would translate exactly from on-off to RAPM, but I suspect it would generally be the case.

But yes, you are right that just looking at percent of one’s career that was true prime vs. weak prime vs. not-at-all prime doesn’t really fully get at exactly what the difference would be between prime/peak RAPM and career RAPM. We’d generally *expect* that someone who has a lower percent of his career being prime years in the data set and ends up higher overall probably had a higher RAPM in the peak or prime years. But of course, along the lines you discuss, that’s not *necessarily* the case, because maybe that player’s numbers in those non-prime years are way better than others’ non-prime years and that gets them ahead overall despite the above. However, I wasn’t suggesting that being #1 in JE’s 1997-2024 RAPM *must* mean Jokic has the best peak RAPM (though I grant that my wording might’ve been confusing on this front). I would say it’s hard to imagine someone who is #1 overall but yet didn’t have an incredibly high peak RAPM (and we also know Jokic has graded out extremely well in other impact metrics in recent years anyways, so we independently know his peak impact numbers are incredibly good). Meanwhile, ultimately we can’t really directly compare specific RAPM values from one year and another, since the scaling and whatnot isn’t the same year to year. So there’s no perfect method to directly compare RAPM values that different players have had in different seasons. What we know is that Jokic has looked incredibly good in season-by-season impact metrics in recent years, and he has looked incredibly good in a career-long RAPM measure. I think that’s consistent with Jokic having the GOAT peak, though I also think the same could be said about the “impact case” of several other players—there’s enough uncertainty and flaws with impact measures that I don’t think we could ever say there’s only one player with an “impact case” for GOAT peak.

3. this is ridiculous we literally know Lebrons RAPM data always gets a huge boost from him turning it on in the playoffs. I’m pretty sure JE showed data that Lebrons RAPM data went up by 33% in the playoffs from 2014-2017. A fair sample to really capture his peak would weight playoffs more considering he blatantly takes his foot off the pedal and his numbers all universally sky rocket in the playoffs including impact, which is absolutely not a trend we see with Jokic which is the point here

Jokic does not go down in the playoffs in box stats, but definitely has in impact stats. Does the fact that his impact numbers have actually gone down in the playoffs—where the “off” samples are absolutely tiny—mean Jokic isn’t often coasting in the regular season, even though we know his box numbers stay similarly incredible in the playoffs (when most stars have their box numbers get worse because playoff basketball is harder) and he has a reputation for mailing it in in the regular season? You’re free to draw the less favorable conclusion about Jokic based primarily on very low-sample-size data, but I don’t think it’s at all “ridiculous” to come to a different conclusion. And you’re also free to put greater weight on playoff impact, but again it’s not “ridiculous” to not layer on that sort of adjustment.

I don’t understand why you are being purposefully disingenuous here but I’m not trying to have a conversation with someone who willfully ignores and selectively reads what he wants to read

…

My interest in this disingenuous conversation isn’t particularly high, so I don’t know if I will respond anymore since I clearly can tell what kind of conversation this is going to be. Ignoring evidence from a place of genuine ignorance is fine, ignoring evidence in the face of bias and close mindedness is a pain because it’s just a waste of all of our time. Do better. I see through this type of approach.

I’m not sure what I said to you to make you respond this rudely. I thought I was perfectly polite to you (and I believe I’ve been polite to you in this response too), and I’d ask that you do the same for me (and probably won’t respond further if you don’t). We’re just discussing basketball, and I have no desire to have an acrimonious discussion with someone I’ve never previously interacted with, based on a topic that doesn’t really matter. No hard feelings towards you at all if this is just a one-time thing!

I’m not someone that blindly follows impact data, I run it myself and see issues with it (none that would help your case here mind you, so don’t try twisting my words here…), but if you are arguing he has goat tier raw RAPM data he absolute does not unless you are purposefully not looking at it within context, which you must be doing considering it has been brought up by multiple people. Lebron and KG both are clearly better looking at it within context

I think “within context” is doing a lot of work there. Impact measures aren’t an exact science, and one significant reason for that is because there will always be significant disagreement on what factors and context to adjust for (and how to adjust for it). If you feel very strongly that we should layer on things into the analysis that end up leaving Jokic looking less good relative to others, then you may well be able to come to a conclusion that Jokic’s impact data isn’t GOAT-tier. To me, though, I think we’re talking about a very inexact science where data can be massaged a million different ways, and if Jokic is sitting at #1 in JE’s initial raw RAPM calculation and also has looked incredible overall in various season-by-season impact metrics, then I’m comfortable in concluding that that’s indicative of GOAT-tier impact, even if there’s other players that also have data indicative of GOAT-tier impact and there’s other ways to look at the data that doesn’t have Jokic at the top. Basically, I think we should think of these things like they have a confidence interval, and I think what I see in Jokic’s data suggests that having GOAT-tier impact is definitely within the confidence interval, as it is for several other players as well. And, crucially, I’m not realy making an affirmative argument based on impact data. Rather, what I was doing was disputing the idea that Jokic gets “clobbered” and has “no impact case.” I think that’s wrong. But I also don’t think that impact numbers necessarily suggest Jokic is at the very top. As I’ve repeatedly said in this thread, I think Jokic has the GOAT peak primarily based on my eye test—and I think his impact numbers are box numbers are consistent with that conclusion because I think he’s within the confidence interval on those.

Yeah just reading the first bit alone I refuse to waste my time with disingenuous discussion disguised by fake politenes. “Is this really true though” when it’s been referenced by multiple people simply shows me the type of conversation this is going to be. If you are going to argue about the data you should understand it first. I wasted my time getting the dataset for nothing and it’s very irritating that it’s been clarified multiple times, but I’ll just know this if I interact with you further

I refuse to waste my time when the other person simply ignores the same statements and clarifications over and over again, not worth the effort, OhayoKD can continue the conversation if he wishes i might help out if it gets too blatant

[PooledSilver]

It's a fun little contest

As of last check I was stinking it up this season, been trying to apply my area to the predictions the last couple of years and haven't quite made it work. This season I went with a plus/minus 'all-in-one' built from hypergraph centrality as my entry, then applied much like one would an xRAPM measure.

Had my most success in the contest when I completely ignored individual players and went with team based vibes. Read - qualitative descriptors/variables of teams sourced from various sites around the web like RealGM.

Curious if you beat Vegas haha

I think the all in one idea to predict team success and combining that with some outside knowledge of fit and role built in and weighting the two might be a move

As far as I know LEBRON seems pretty simple for doing as well as it has thus far, although I’ve heard they changed up the luck adjustments

Imagine the money you could make if you had these models before Vegas knew what was up

Just gotta make the models yourself

I'm behind Vegas in the contest so far this season (4-2 vs Vegas in the 6 prior seasons), but I didn't put my money where my mouth was, so no skin off my teeth. I prefer retirement

Generally I found it notably easier to find a game or two a night to beat Vegas on than to beat them at the season long stuff.

Went back, and yep, the pure team vibes entry was my strongest finish - my first contest in '17-'18.

The only two model inputs were:
Team Win% the prior season
One of 5 categorical variables based off fan surveying

The categories:
Very Negative
Negative
Neutral
Positive
Very Positive

Haha that’s pretty cool, I’ve wanted to try my hand at making a model beating the spread on a night to night basis when let’s say there’s a huge discrepancy

The vibes one winning out honestly doesn’t suprise me that much, I’m curious at how the specific errors were if there are patterns between teams that overpreform and underpreform vs other more model based ones, probably a solid mix one way or another.

[Peregrine01]

You completely ignored what I said, but paragraph by paragraph

1. Applying an aging curve helping Lebron is valid, but Jokic falls from 1st to 7th, and we have RAPM throughout his prime too.

More than this though, you ignore the points that Jokic’s pre prime RAPM is WAY better than most people’s, it’s both a testament to his early year success being pretty high but also inflates the overall career data where almost everyone else has pre or post prime years bringing them down

2. You completely ignored what I said, your breakdown would be sensible if you looked at the actual data but it simply isn’t whatsoever. In the context of RAPM Jokic’s “pre-prime” years inflate his data relative to everyone else more than hurting him because his pre-prime years were very solid impact wise ESPECIALLY for being his “bad” years. You can’t stand by what you said because there’s nothing to stand on its blatantly being misleading and ignoring the reality based on ur preconceived notion, and that’s fine untill someone brings it up, by blatantly ignoring this point now you are wasting your own time and mine.

Is that really true though? If we look at on-off (which obviously pretty closely resembles RAPM and is something I have easily at my fingertips as I’m writing this), Jokic in his true prime is *way* higher than Jokic before that. What is true is that the years I labeled not-at-all prime aren’t worse (and are actually better) than the years I labeled weak prime. But his true prime years are, on average, a lot higher, and his true prime years are a pretty abnormally low percent of his minutes. Not sure off the top of my head whether this would translate exactly from on-off to RAPM, but I suspect it would generally be the case.

But yes, you are right that just looking at percent of one’s career that was true prime vs. weak prime vs. not-at-all prime doesn’t really fully get at exactly what the difference would be between prime/peak RAPM and career RAPM. We’d generally *expect* that someone who has a lower percent of his career being prime years in the data set and ends up higher overall probably had a higher RAPM in the peak or prime years. But of course, along the lines you discuss, that’s not *necessarily* the case, because maybe that player’s numbers in those non-prime years are way better than others’ non-prime years and that gets them ahead overall despite the above. However, I wasn’t suggesting that being #1 in JE’s 1997-2024 RAPM *must* mean Jokic has the best peak RAPM (though I grant that my wording might’ve been confusing on this front). I would say it’s hard to imagine someone who is #1 overall but yet didn’t have an incredibly high peak RAPM (and we also know Jokic has graded out extremely well in other impact metrics in recent years anyways, so we independently know his peak impact numbers are incredibly good). Meanwhile, ultimately we can’t really directly compare specific RAPM values from one year and another, since the scaling and whatnot isn’t the same year to year. So there’s no perfect method to directly compare RAPM values that different players have had in different seasons. What we know is that Jokic has looked incredibly good in season-by-season impact metrics in recent years, and he has looked incredibly good in a career-long RAPM measure. I think that’s consistent with Jokic having the GOAT peak, though I also think the same could be said about the “impact case” of several other players—there’s enough uncertainty and flaws with impact measures that I don’t think we could ever say there’s only one player with an “impact case” for GOAT peak.

3. this is ridiculous we literally know Lebrons RAPM data always gets a huge boost from him turning it on in the playoffs. I’m pretty sure JE showed data that Lebrons RAPM data went up by 33% in the playoffs from 2014-2017. A fair sample to really capture his peak would weight playoffs more considering he blatantly takes his foot off the pedal and his numbers all universally sky rocket in the playoffs including impact, which is absolutely not a trend we see with Jokic which is the point here

Jokic does not go down in the playoffs in box stats, but definitely has in impact stats. Does the fact that his impact numbers have actually gone down in the playoffs—where the “off” samples are absolutely tiny—mean Jokic isn’t often coasting in the regular season, even though we know his box numbers stay similarly incredible in the playoffs (when most stars have their box numbers get worse because playoff basketball is harder) and he has a reputation for mailing it in in the regular season? You’re free to draw the less favorable conclusion about Jokic based primarily on very low-sample-size data, but I don’t think it’s at all “ridiculous” to come to a different conclusion. And you’re also free to put greater weight on playoff impact, but again it’s not “ridiculous” to not layer on that sort of adjustment.

I don’t understand why you are being purposefully disingenuous here but I’m not trying to have a conversation with someone who willfully ignores and selectively reads what he wants to read

…

My interest in this disingenuous conversation isn’t particularly high, so I don’t know if I will respond anymore since I clearly can tell what kind of conversation this is going to be. Ignoring evidence from a place of genuine ignorance is fine, ignoring evidence in the face of bias and close mindedness is a pain because it’s just a waste of all of our time. Do better. I see through this type of approach.

I’m not sure what I said to you to make you respond this rudely. I thought I was perfectly polite to you (and I believe I’ve been polite to you in this response too), and I’d ask that you do the same for me (and probably won’t respond further if you don’t). We’re just discussing basketball, and I have no desire to have an acrimonious discussion with someone I’ve never previously interacted with, based on a topic that doesn’t really matter. No hard feelings towards you at all if this is just a one-time thing!

I’m not someone that blindly follows impact data, I run it myself and see issues with it (none that would help your case here mind you, so don’t try twisting my words here…), but if you are arguing he has goat tier raw RAPM data he absolute does not unless you are purposefully not looking at it within context, which you must be doing considering it has been brought up by multiple people. Lebron and KG both are clearly better looking at it within context

I think “within context” is doing a lot of work there. Impact measures aren’t an exact science, and one significant reason for that is because there will always be significant disagreement on what factors and context to adjust for (and how to adjust for it). If you feel very strongly that we should layer on things into the analysis that end up leaving Jokic looking less good relative to others, then you may well be able to come to a conclusion that Jokic’s impact data isn’t GOAT-tier. To me, though, I think we’re talking about a very inexact science where data can be massaged a million different ways, and if Jokic is sitting at #1 in JE’s initial raw RAPM calculation and also has looked incredible overall in various season-by-season impact metrics, then I’m comfortable in concluding that that’s indicative of GOAT-tier impact, even if there’s other players that also have data indicative of GOAT-tier impact and there’s other ways to look at the data that doesn’t have Jokic at the top. Basically, I think we should think of these things like they have a confidence interval, and I think what I see in Jokic’s data suggests that having GOAT-tier impact is definitely within the confidence interval, as it is for several other players as well. And, crucially, I’m not realy making an affirmative argument based on impact data. Rather, what I was doing was disputing the idea that Jokic gets “clobbered” and has “no impact case.” I think that’s wrong. But I also don’t think that impact numbers necessarily suggest Jokic is at the very top. As I’ve repeatedly said in this thread, I think Jokic has the GOAT peak primarily based on my eye test—and I think his impact numbers are box numbers are consistent with that conclusion because I think he’s within the confidence interval on those.

Yeah just reading the first bit alone I refuse to waste my time with disingenuous discussion disguised by fake politenes. “Is this really true though” when it’s been referenced by multiple people simply shows me the type of conversation this is going to be. If you are going to argue about the data you should understand it first. I wasted my time getting the dataset for nothing and it’s very irritating that it’s been clarified multiple times, but I’ll just know this if I interact with you further

I refuse to waste my time when the other person simply ignores the same statements and clarifications over and over again, not worth the effort, OhayoKD can continue the conversation if he wishes i might help out if it gets too blatant

OhayoKD, is that you?

[PooledSilver]

Is that really true though? If we look at on-off (which obviously pretty closely resembles RAPM and is something I have easily at my fingertips as I’m writing this), Jokic in his true prime is *way* higher than Jokic before that. What is true is that the years I labeled not-at-all prime aren’t worse (and are actually better) than the years I labeled weak prime. But his true prime years are, on average, a lot higher, and his true prime years are a pretty abnormally low percent of his minutes. Not sure off the top of my head whether this would translate exactly from on-off to RAPM, but I suspect it would generally be the case.

But yes, you are right that just looking at percent of one’s career that was true prime vs. weak prime vs. not-at-all prime doesn’t really fully get at exactly what the difference would be between prime/peak RAPM and career RAPM. We’d generally *expect* that someone who has a lower percent of his career being prime years in the data set and ends up higher overall probably had a higher RAPM in the peak or prime years. But of course, along the lines you discuss, that’s not *necessarily* the case, because maybe that player’s numbers in those non-prime years are way better than others’ non-prime years and that gets them ahead overall despite the above. However, I wasn’t suggesting that being #1 in JE’s 1997-2024 RAPM *must* mean Jokic has the best peak RAPM (though I grant that my wording might’ve been confusing on this front). I would say it’s hard to imagine someone who is #1 overall but yet didn’t have an incredibly high peak RAPM (and we also know Jokic has graded out extremely well in other impact metrics in recent years anyways, so we independently know his peak impact numbers are incredibly good). Meanwhile, ultimately we can’t really directly compare specific RAPM values from one year and another, since the scaling and whatnot isn’t the same year to year. So there’s no perfect method to directly compare RAPM values that different players have had in different seasons. What we know is that Jokic has looked incredibly good in season-by-season impact metrics in recent years, and he has looked incredibly good in a career-long RAPM measure. I think that’s consistent with Jokic having the GOAT peak, though I also think the same could be said about the “impact case” of several other players—there’s enough uncertainty and flaws with impact measures that I don’t think we could ever say there’s only one player with an “impact case” for GOAT peak.



Jokic does not go down in the playoffs in box stats, but definitely has in impact stats. Does the fact that his impact numbers have actually gone down in the playoffs—where the “off” samples are absolutely tiny—mean Jokic isn’t often coasting in the regular season, even though we know his box numbers stay similarly incredible in the playoffs (when most stars have their box numbers get worse because playoff basketball is harder) and he has a reputation for mailing it in in the regular season? You’re free to draw the less favorable conclusion about Jokic based primarily on very low-sample-size data, but I don’t think it’s at all “ridiculous” to come to a different conclusion. And you’re also free to put greater weight on playoff impact, but again it’s not “ridiculous” to not layer on that sort of adjustment.



I’m not sure what I said to you to make you respond this rudely. I thought I was perfectly polite to you (and I believe I’ve been polite to you in this response too), and I’d ask that you do the same for me (and probably won’t respond further if you don’t). We’re just discussing basketball, and I have no desire to have an acrimonious discussion with someone I’ve never previously interacted with, based on a topic that doesn’t really matter. No hard feelings towards you at all if this is just a one-time thing!



I think “within context” is doing a lot of work there. Impact measures aren’t an exact science, and one significant reason for that is because there will always be significant disagreement on what factors and context to adjust for (and how to adjust for it). If you feel very strongly that we should layer on things into the analysis that end up leaving Jokic looking less good relative to others, then you may well be able to come to a conclusion that Jokic’s impact data isn’t GOAT-tier. To me, though, I think we’re talking about a very inexact science where data can be massaged a million different ways, and if Jokic is sitting at #1 in JE’s initial raw RAPM calculation and also has looked incredible overall in various season-by-season impact metrics, then I’m comfortable in concluding that that’s indicative of GOAT-tier impact, even if there’s other players that also have data indicative of GOAT-tier impact and there’s other ways to look at the data that doesn’t have Jokic at the top. Basically, I think we should think of these things like they have a confidence interval, and I think what I see in Jokic’s data suggests that having GOAT-tier impact is definitely within the confidence interval, as it is for several other players as well. And, crucially, I’m not realy making an affirmative argument based on impact data. Rather, what I was doing was disputing the idea that Jokic gets “clobbered” and has “no impact case.” I think that’s wrong. But I also don’t think that impact numbers necessarily suggest Jokic is at the very top. As I’ve repeatedly said in this thread, I think Jokic has the GOAT peak primarily based on my eye test—and I think his impact numbers are box numbers are consistent with that conclusion because I think he’s within the confidence interval on those.

Yeah just reading the first bit alone I refuse to waste my time with disingenuous discussion disguised by fake politenes. “Is this really true though” when it’s been referenced by multiple people simply shows me the type of conversation this is going to be. If you are going to argue about the data you should understand it first. I wasted my time getting the dataset for nothing and it’s very irritating that it’s been clarified multiple times, but I’ll just know this if I interact with you further

I refuse to waste my time when the other person simply ignores the same statements and clarifications over and over again, not worth the effort, OhayoKD can continue the conversation if he wishes i might help out if it gets too blatant

OhayoKD, is that you?

I think WOWY is very stupid, but that person seems much more comitted to this post anyways

On the other hand I’ve been on enough forums to know to leave a conversation when it’s clear the other side isn’t genuine

I don’t have as much of an issue with the side the person is on as much as him (or her!) making this entire conversation a ridiculous waste of time when it was clear he wasn’t genuinely reading through responses by reiterating the same talking points. If someone is being disingenuous under the guise of fake politeness it’s easy to see through

[lessthanjake]

Yeah just reading the first bit alone I refuse to waste my time with disingenuous discussion disguised by fake politenes. “Is this really true though” when it’s been referenced by multiple people simply shows me the type of conversation this is going to be. If you are going to argue about the data you should understand it first. I wasted my time getting the dataset for nothing and it’s very irritating that it’s been clarified multiple times, but I’ll just know this if I interact with you further

I refuse to waste my time when the other person simply ignores the same statements and clarifications over and over again, not worth the effort, OhayoKD can continue the conversation if he wishes i might help out if it gets too blatant

I genuinely don’t know what you’re referring to in the bolded part. As far as I can tell, the only data you posted was JE’s RAPM numbers that included additional adjustments, and I addressed that at length in a response to you. Not coming to the same bottom-line conclusion as you based on information you post doesn’t mean someone is being “disingenuous.” Am I disingenuous unless I preference all the same adjustments as you do, to the exclusion of other data that does not use those adjustments? I really don’t think so. Meanwhile, other than that, I don’t think you’ve posted any data—you mentioned Jokic’s RAPM being abnormally good in his early years, and because you didn’t actually post any specific data on that and I didn’t have any at my fingertips, I responded with a nuanced point using related data I did readily have (i.e. on-off), while explicitly acknowledging that it’s possible what I was saying doesn’t entirely translate from on-off to RAPM. I don’t see how that could fairly be described as disingenuous. Finally, I’m not sure what you mean by “it’s been referenced by multiple people” and that I have “ignore[d] the same statements and clarifications over and over again.” As far as I’m aware, I’ve pretty thoroughly responded to every post directed at me in this thread. I’ve not read every other post in this thread, so it’s possible there’s something relevant to our discussion that was not directed at me and that I’ve not read (and if there is, feel free to actually precisely point it out, rather than vaguely referring to its existence), but I hardly think I’m obliged to read and account for everything people have said in a thread before posting in order to not receive this sort of response. In any event, maybe there’s something I’m missing here that the above doesn’t cover—it’s genuinely hard for me to tell what you’re referring to, and I’m a bit taken aback and confused at your seemingly angry reaction to my posts. If there’s something important I’ve missed and/or not responded to (or perhaps just an important point that you think I’ve not addressed thoroughly enough), please let me know specifically what it is.

More generally, please do try to be polite. In my last post, I said I probably wouldn’t respond if you made an acrimonious post again, but I decided to respond again here because you seem like you are a sophisticated poster who has interesting insights and access to data, and so I think it’d be a shame to entirely disengage with you based upon our first interaction. There’s no need to have this kind of discussion. And while we may disagree on this particular thing, I imagine there’s other things we’d agree on!

[PooledSilver]

Shot charts. Last time.

I would probably assume 21-24 Jokic has similar level of RAPM as 2015-2018 curry and below peak KG and the peak Lebron years although his are scattered, but I wouldn’t suprised if only 2022 and 2023 were truly all time level specifically

Does Jokic have incredibly strong impact data? Yes

Is it goat tier? No, in the context that he’s a tier below the Bron and KG tier. (And really that tier has a bit of a gap between bron and KG when you consider the playoff RAPM that got shown)

I mean I have KG not sniffing my personal top ten

That’s not that but of a deal in general but my issue is implying that this specific aspect is and explanations which ignore the things referenced like the shot charts single year data and multiyear data

[OhayoKD]

Is that really true though? If we look at on-off (which obviously pretty closely resembles RAPM and is something I have easily at my fingertips as I’m writing this), Jokic in his true prime is *way* higher than Jokic before that. What is true is that the years I labeled not-at-all prime aren’t worse (and are actually better) than the years I labeled weak prime. But his true prime years are, on average, a lot higher, and his true prime years are a pretty abnormally low percent of his minutes. Not sure off the top of my head whether this would translate exactly from on-off to RAPM, but I suspect it would generally be the case.

But yes, you are right that just looking at percent of one’s career that was true prime vs. weak prime vs. not-at-all prime doesn’t really fully get at exactly what the difference would be between prime/peak RAPM and career RAPM. We’d generally *expect* that someone who has a lower percent of his career being prime years in the data set and ends up higher overall probably had a higher RAPM in the peak or prime years. But of course, along the lines you discuss, that’s not *necessarily* the case, because maybe that player’s numbers in those non-prime years are way better than others’ non-prime years and that gets them ahead overall despite the above. However, I wasn’t suggesting that being #1 in JE’s 1997-2024 RAPM *must* mean Jokic has the best peak RAPM (though I grant that my wording might’ve been confusing on this front). I would say it’s hard to imagine someone who is #1 overall but yet didn’t have an incredibly high peak RAPM (and we also know Jokic has graded out extremely well in other impact metrics in recent years anyways, so we independently know his peak impact numbers are incredibly good). Meanwhile, ultimately we can’t really directly compare specific RAPM values from one year and another, since the scaling and whatnot isn’t the same year to year. So there’s no perfect method to directly compare RAPM values that different players have had in different seasons. What we know is that Jokic has looked incredibly good in season-by-season impact metrics in recent years, and he has looked incredibly good in a career-long RAPM measure. I think that’s consistent with Jokic having the GOAT peak, though I also think the same could be said about the “impact case” of several other players—there’s enough uncertainty and flaws with impact measures that I don’t think we could ever say there’s only one player with an “impact case” for GOAT peak.



Jokic does not go down in the playoffs in box stats, but definitely has in impact stats. Does the fact that his impact numbers have actually gone down in the playoffs—where the “off” samples are absolutely tiny—mean Jokic isn’t often coasting in the regular season, even though we know his box numbers stay similarly incredible in the playoffs (when most stars have their box numbers get worse because playoff basketball is harder) and he has a reputation for mailing it in in the regular season? You’re free to draw the less favorable conclusion about Jokic based primarily on very low-sample-size data, but I don’t think it’s at all “ridiculous” to come to a different conclusion. And you’re also free to put greater weight on playoff impact, but again it’s not “ridiculous” to not layer on that sort of adjustment.



I’m not sure what I said to you to make you respond this rudely. I thought I was perfectly polite to you (and I believe I’ve been polite to you in this response too), and I’d ask that you do the same for me (and probably won’t respond further if you don’t). We’re just discussing basketball, and I have no desire to have an acrimonious discussion with someone I’ve never previously interacted with, based on a topic that doesn’t really matter. No hard feelings towards you at all if this is just a one-time thing!



I think “within context” is doing a lot of work there. Impact measures aren’t an exact science, and one significant reason for that is because there will always be significant disagreement on what factors and context to adjust for (and how to adjust for it). If you feel very strongly that we should layer on things into the analysis that end up leaving Jokic looking less good relative to others, then you may well be able to come to a conclusion that Jokic’s impact data isn’t GOAT-tier. To me, though, I think we’re talking about a very inexact science where data can be massaged a million different ways, and if Jokic is sitting at #1 in JE’s initial raw RAPM calculation and also has looked incredible overall in various season-by-season impact metrics, then I’m comfortable in concluding that that’s indicative of GOAT-tier impact, even if there’s other players that also have data indicative of GOAT-tier impact and there’s other ways to look at the data that doesn’t have Jokic at the top. Basically, I think we should think of these things like they have a confidence interval, and I think what I see in Jokic’s data suggests that having GOAT-tier impact is definitely within the confidence interval, as it is for several other players as well. And, crucially, I’m not realy making an affirmative argument based on impact data. Rather, what I was doing was disputing the idea that Jokic gets “clobbered” and has “no impact case.” I think that’s wrong. But I also don’t think that impact numbers necessarily suggest Jokic is at the very top. As I’ve repeatedly said in this thread, I think Jokic has the GOAT peak primarily based on my eye test—and I think his impact numbers are box numbers are consistent with that conclusion because I think he’s within the confidence interval on those.

Yeah just reading the first bit alone I refuse to waste my time with disingenuous discussion disguised by fake politenes. “Is this really true though” when it’s been referenced by multiple people simply shows me the type of conversation this is going to be. If you are going to argue about the data you should understand it first. I wasted my time getting the dataset for nothing and it’s very irritating that it’s been clarified multiple times, but I’ll just know this if I interact with you further

I refuse to waste my time when the other person simply ignores the same statements and clarifications over and over again, not worth the effort, OhayoKD can continue the conversation if he wishes i might help out if it gets too blatant

OhayoKD, is that you?

Yes

[PooledSilver]

Yeah just reading the first bit alone I refuse to waste my time with disingenuous discussion disguised by fake politenes. “Is this really true though” when it’s been referenced by multiple people simply shows me the type of conversation this is going to be. If you are going to argue about the data you should understand it first. I wasted my time getting the dataset for nothing and it’s very irritating that it’s been clarified multiple times, but I’ll just know this if I interact with you further

I refuse to waste my time when the other person simply ignores the same statements and clarifications over and over again, not worth the effort, OhayoKD can continue the conversation if he wishes i might help out if it gets too blatant

OhayoKD, is that you?

Yes

Kys

[lessthanjake]

Shot charts. Last time.

I would probably assume 21-24 Jokic has similar level of RAPM as 2015-2018 curry and below peak KG and the peak Lebron years although his are scattered, but I wouldn’t suprised if only 2022 and 2023 were truly all time level specifically

Does Jokic have incredibly strong impact data? Yes

Is it goat tier? No, in the context that he’s a tier below the Bron and KG tier. (And really that tier has a bit of a gap between bron and KG when you consider the playoff RAPM that got shown)

I mean I have KG not sniffing my personal top ten

That’s not that but of a deal in general but my issue is implying that this specific aspect is and explanations which ignore the things referenced like the shot charts single year data and multiyear data

Okay, I didn’t realize you were referring to NBAshotcharts data. You’d only briefly mentioned that in posts directed at me, and as far as I can tell no other poster mentioned it in a post directed to me. My apologies for the confusion, but I do think you were a bit overaggressive to respond in the way you did just because I didn’t specifically address something you’d only briefly mentioned to me in earlier posts. I’ve responded to virtually everything you and others said to me and one thing slipped through the cracks, from a post you made that was almost entirely focused instead on JE’s additional adjustments (which I did respond to). Immediately jumping to the conclusion that I’m “disingenuous” based on that is pretty uncalled for IMO.

Anyways, don’t you think the NBAshotcharts data is consistent with the response I already gave you about Jokic’s impact numbers over the course of his career? In my post that responded to you on that issue (https://forums.realgm.com/boards/viewtopic.php?p=111820519#p111820519), I pointed out that while Jokic’s not-at-all prime years had better on-off than his weak-prime years, his true-prime years have also been way better on average than the prior years, and that his true prime years are also a pretty abnormally low percent of his minutes. The NBAshotcharts data seems pretty supportive of that (unsurprising, since RAPM and on-off are very closely linked). Jokic’s league rank in NBAshotcharts RAPM in chronological order, ending before the current season is: 32nd, 24th, 14th, 44th, 70th, 39th, 3rd, and 2nd (note for others: NBAshotcharts RAPM is very noisy—presumably because it’s a raw measure—so this measure isn’t something anyone is farming 1st places in). Those last three numbers (39th, 3rd, and 2nd) are the true-prime years and the two years before that are the weak prime. This is all consistent with what I said in my prior post. The not-at-all prime years actually do look better in terms of impact than the weak-prime years (not surprising, since that’s true for on-off as well, as I pointed out in my prior post), but the true-prime years are on average a large step up overall (even without this season, which would very likely be another high rank, given Jokic’s +19.6 on-off). Given this, I guess I’m having trouble seeing why you responded to my post as you did (calling my response disingenuous, etc.). I realize my response didn’t utilize the RAPM source you’d mentioned earlier in the thread, but if you read and digested the point I made, I think you’d have realized that the point I made applied essentially exactly the same even if I’d used that RAPM source anyways. And, in any event, if you looked at the second paragraph in that post, you’d see I addressed the discussion more generally, in a way that I think was sensible and acknowledged the general point you’d made while further clarifying what I was arguing. I really don’t think my post deserved the type of response you gave it.

[PooledSilver]

In the context of his number not inflated because his pre prime years not being good, absolutely not.

Jokic lacks the bad first year that almost everyone has since they’re high picks not ready for leading roles, which is why age adjustments change his place in that dataset so much

Most early and end of prime years are big negatives. Jokic’s rankings being as high as they are early on is why his career marks are great, he doesn’t have the really bad early or late years bringing it down

Lebrons first year was 224th for example

Jokic’s RAPM this year is 11th iirc.

To say Jokic is adversely effected because he is in the middle of his prime isn’t true though, because of the nature of his pre prime/peak years being insanely good in the context of impact data because he didn’t have the 1 or 2 horrible years almost everyone has

You can actually get just his prime years on some rshiny website including only 2022, 2023, and some of 2024 (so it’ll have more of the 2022 and 2023 sample weighted towards it) where he is a clear first in that three year stretch (only one he is first in), but also not a standout within 3 year stretches on the site , definately less impressive than currys 3 year stretches from 2015-2018 and 2016-2019 and kawhis 2020-2022 if you are talking about the raw data and relation to others

In the context of raw RAPM Jokic definately doesn’t hVe a case over lebron or KG though, the idea that career data hurts him because of where he is in his career simply doesn’t hold true because we can take a sample of just his peak years that’s biased towards the better years and he still falls short of curry’s likely second best 3 year stretch who falls short of the other two.

Like it makes no sense to say that Jokic’s data is deflated by his place in his career because we can get the 3 year peaks of different players and Jokic goes from GOAT tier to more top 5-7 tier, which is perfectly fine, but adding the rest of the years brings everyone’s data down but more so the others than Jokic’s because of years like lebrons rookie year (224th, nearly the same rank sum of Jokic’s entire career)

None of that means all that much to me personally, but in the context of how is his RAPM data is it GOAT tier the answer is it’s a tier or two below that depending on if you view bron as a seperate tier from KG with the 14-17 run which while lowering his overall RAPM almost certainly also for sure has the highest playoff peak in that timespan given his RS to playoff RAPM jumps from (according to J.E. Numbers so different scale so just focus on he % increase) 9 to 12.2

[Jaqua92]

We have a lot of scientists and researchers in here, apparently.

Wait, nvm. Real scientists actually acknowledge that the data they ACCURATELY assess with the scientific method is often inconclusive and "requires further analysis"when contemplating conclusion.

They should consult with RealGM's stat people, they're obviously smarter.

The dialogue these last few pages is embarrassing