A player whose impact doesn't match his reputation (yet)

[Bob8]

You can believe that all you want, but I don't post agenda-based threads (with the possible exception of my Embiid threads which are at least grounded in facts). I don't hate Luka, don't hate the Spurs, always rooted for the Nets to succeed, yada yada yada. It doesn't stop me from seeking objective truth.

I believe you’re wrong, nothing else.

I won't fault you for that. Even in my infinite wisdom I've been known to be wrong from time to time

I will tell you in advance what your model will discover. Mavs team is constructed poorly around Luka. Luka is poor defender and great passer, so he should be surrounded with 3&D players and rim protector. There isn’t a single 3&D player in Mavs roster and KP could be rim protector but is not ready yet. So what happens with his unit, that’s what you see through your data, bad D. means a lot points received, bad shooting too little points from players not named Luka. If they want be serious playoffs team, they need to add 2 3&D starters and wait till KP is ready.

[Wooderson]

Those numbers are derived from individual box metrics. You can find comprehensive on/off data here: https://www.basketball-reference.com/players/d/doncilu01/on-off/2020

I'd recommend using nba.com on/off stats instead of BBRef.

nba.com uses true play by play data for possessions while BBRef is still only estimates I believe.

I believe the estimates you are referring to are ORtg and DRtg? What I used were +/- and on/off (per 100 possessions) derived from play by play data.

I'm speaking to +/- and on/off numbers not the individual ORtg/DRtg. The sites track possessions differently, which is why you see a difference across the board for all players with nba.com vs basketball reference. nba.com changed their on/off calculations last year using their play by play data to give "true possession data" and better reflect what's happening on the court. Here's an article on it. Pretty sure basketball reference still estimates the number of possessions with the box score which nba.com used to do.

[GeorgeMarcus]

I'd recommend using nba.com on/off stats instead of BBRef.

nba.com uses true play by play data for possessions while BBRef is still only estimates I believe.

I believe the estimates you are referring to are ORtg and DRtg? What I used were +/- and on/off (per 100 possessions) derived from play by play data.

I'm speaking to +/- and on/off numbers not the individual ORtg/DRtg. The sites track possessions differently, which is why you see a difference across the board for all players with nba.com vs basketball reference. nba.com changed their on/off calculations last year using their play by play data to give "true possession data" and better reflect what's happening on the court. Here's an article on it. Pretty sure basketball reference still estimates the number of possessions with the box score which nba.com used to do.

Appreciate the find- I’ll read the article when I get a chance. I recently had a conversation with my buddy on who they attribute +/- to when players are subbed between FT’s. Obviously in that scenario they had no bearing on that player getting to the line, and shouldn’t be punished for it.

[KqWIN]

Reading through this thread...I think somethings being missed here. The first thing that comes to mind when this stats/reputation doesn’t match impact is defense. Defense isn’t tracked as well by numbers and people don’t care enough about it to begin with.

PS: pbpstats.com is the best site out there. Give that man some support!

[Pumpkin17]

First of all, you cannot say you are proving anything through numbers if applying your theory you end up with absurd results.
THJ would be Mavs best player according to your "model" and we all know that's not the case. So let's forget about the presumption of some kind of model-given proof.

Now, numbers are interesting and we can investigate these. We can think about rotations, lineups and so on and find some reasonable explainations about the phenomenon. We can even conclude Doncic needs to change some habits and improve, which he certainly needs to do. But this is more an optimization/team building kind of reasoning.

But, again, you are dealing with flawed statistics and you are using these according to your interpretation, cause there is no general consensus about these statistics being an appropriate measure of impact.

I suggest you take a more reasonable approach in order to develop a proper model capable of measuring impact: start from the evidence.
The Earth is a globe: you can find a lot of evidence supporting this and you can also imagin ubercomplex models allowing the assumption that this is the opposite according to some flawed line of reasoning. But this does not change the fact that the Earth is a globe, so if your reasoning gives you a flat earth, you have a fallacy somewhere. Change the reasoning, change the model.
Doncic is a great player which is the backbone of his team: you can also find a lot of evidence supporting this and Carlisle would play him even if some measure is not that good. If your "model" gives you an indication of the opposite, again, change your reasoning, change the model.

Your numbers are genuinely interesting and worth investigating, but your model gives you an absurd result. I think it is more reasonable to think that your reasoning is flawed than that Doncic is not an impact player.

If we look into THJ's history, he's been more of a negative than a positive using the same "model" (though I'm not sure I'd call it a model). I think you and others are misunderstanding how I'm using those numbers. I'm not saying "rank the players based on these 2 values" but I addressed the direct implications of the values, and how we can interpret them. For example, if a player is regarded as clearly the most impactful on his team, it doesn't really make sense for him to have a negative on/off. I stand by that opinion but, as I said in other posts, I'm willing to give Luka the benefit of the doubt in terms of the sample size so far this season. I wouldn't even have addressed a 10 game sample size if it wasn't backed up by the rookie data, which again is understandable for a rookie.

It is not that it does not make sense, is just need to be interpreted, and different posters gave you a possible explaination, or it is a sign of the fallacy of the model.

With all the due respect, the fact that you are willing to give him the benefit of the doubt based on this statistics sounds pretty arrogant.
The whole NBA would kill to have Doncic and also numerous stats support this. If you want to be a doubter you can criticize lot of things about his game. But if you pretend to have found two numbers telling you a hidden truth while the other stats and evaluations, not to mention the whole NBA world made of professionals, tell you the opposite, it is just you putting to much weight on it.

Statistics describe the world and we need to find models fitting reality. Statistic do NOT CAUSE the world to transform according to our twists.

Doncic not going to regress to a bench level player just cause some stats tell you so. Also, you are not giving him the benefit of the doubt. We are giving you the benefit of the doubt and maybe you will change your opinion.

No one in the world cares if you (or me obv.) don't think he is an impact player based on these stats, he is according to basically all the other stats, other coaches, other gms and other players. It is just your very personal tweak of reality. To each his own i guess

[GeorgeMarcus]

First of all, you cannot say you are proving anything through numbers if applying your theory you end up with absurd results.
THJ would be Mavs best player according to your "model" and we all know that's not the case. So let's forget about the presumption of some kind of model-given proof.

Now, numbers are interesting and we can investigate these. We can think about rotations, lineups and so on and find some reasonable explainations about the phenomenon. We can even conclude Doncic needs to change some habits and improve, which he certainly needs to do. But this is more an optimization/team building kind of reasoning.

But, again, you are dealing with flawed statistics and you are using these according to your interpretation, cause there is no general consensus about these statistics being an appropriate measure of impact.

I suggest you take a more reasonable approach in order to develop a proper model capable of measuring impact: start from the evidence.
The Earth is a globe: you can find a lot of evidence supporting this and you can also imagin ubercomplex models allowing the assumption that this is the opposite according to some flawed line of reasoning. But this does not change the fact that the Earth is a globe, so if your reasoning gives you a flat earth, you have a fallacy somewhere. Change the reasoning, change the model.
Doncic is a great player which is the backbone of his team: you can also find a lot of evidence supporting this and Carlisle would play him even if some measure is not that good. If your "model" gives you an indication of the opposite, again, change your reasoning, change the model.

Your numbers are genuinely interesting and worth investigating, but your model gives you an absurd result. I think it is more reasonable to think that your reasoning is flawed than that Doncic is not an impact player.

If we look into THJ's history, he's been more of a negative than a positive using the same "model" (though I'm not sure I'd call it a model). I think you and others are misunderstanding how I'm using those numbers. I'm not saying "rank the players based on these 2 values" but I addressed the direct implications of the values, and how we can interpret them. For example, if a player is regarded as clearly the most impactful on his team, it doesn't really make sense for him to have a negative on/off. I stand by that opinion but, as I said in other posts, I'm willing to give Luka the benefit of the doubt in terms of the sample size so far this season. I wouldn't even have addressed a 10 game sample size if it wasn't backed up by the rookie data, which again is understandable for a rookie.

It is not that it does not make sense, is just need to be interpreted, and different posters gave you a possible explaination, or it is a sign of the fallacy of the model.

With all the due respect, the fact that you are willing to give him the benefit of the doubt based on this statistics sounds pretty arrogant.
The whole NBA would kill to have Doncic and also numerous stats support this. If you want to be a doubter you can criticize lot of things about his game. But if you pretend to have found two numbers telling you a hidden truth while the other stats and evaluations, not to mention the whole NBA world made of professionals, tell you the opposite, it is just you putting to much weight on it.

Statistics describe the world and we need to find models fitting reality. Statistic do NOT CAUSE the world to transform according to our twists.

Doncic not going to regress to a bench level player just cause some stats tell you so. Also, you are not giving him the benefit of the doubt. We are giving you the benefit of the doubt and maybe you will change your opinion.

No one in the world cares if you (or me obv.) don't think he is an impact player based on these stats, he is according to basically all the other stats, other coaches, other gms and other players. It is just your very personal tweak of reality. To each his own i guess

The point in question is whether or not Doncic provides superstar level impact right now. And you think it's arrogant to believe that maybe he doesn't?

I know what statistics do- the problem is that most people who deny impact metrics tend to rely on box statistics which are far, far less reliable when it comes to contributing to team success.

[GeorgeMarcus]

Reading through this thread...I think somethings being missed here. The first thing that comes to mind when this stats/reputation doesn’t match impact is defense. Defense isn’t tracked as well by numbers and people don’t care enough about it to begin with.

PS: pbpstats.com is the best site out there. Give that man some support!

Wish I had known about this site sooner

[Pumpkin17]

If we look into THJ's history, he's been more of a negative than a positive using the same "model" (though I'm not sure I'd call it a model). I think you and others are misunderstanding how I'm using those numbers. I'm not saying "rank the players based on these 2 values" but I addressed the direct implications of the values, and how we can interpret them. For example, if a player is regarded as clearly the most impactful on his team, it doesn't really make sense for him to have a negative on/off. I stand by that opinion but, as I said in other posts, I'm willing to give Luka the benefit of the doubt in terms of the sample size so far this season. I wouldn't even have addressed a 10 game sample size if it wasn't backed up by the rookie data, which again is understandable for a rookie.

It is not that it does not make sense, is just need to be interpreted, and different posters gave you a possible explaination, or it is a sign of the fallacy of the model.

With all the due respect, the fact that you are willing to give him the benefit of the doubt based on this statistics sounds pretty arrogant.
The whole NBA would kill to have Doncic and also numerous stats support this. If you want to be a doubter you can criticize lot of things about his game. But if you pretend to have found two numbers telling you a hidden truth while the other stats and evaluations, not to mention the whole NBA world made of professionals, tell you the opposite, it is just you putting to much weight on it.

Statistics describe the world and we need to find models fitting reality. Statistic do NOT CAUSE the world to transform according to our twists.

Doncic not going to regress to a bench level player just cause some stats tell you so. Also, you are not giving him the benefit of the doubt. We are giving you the benefit of the doubt and maybe you will change your opinion.

No one in the world cares if you (or me obv.) don't think he is an impact player based on these stats, he is according to basically all the other stats, other coaches, other gms and other players. It is just your very personal tweak of reality. To each his own i guess

The point in question is whether or not Doncic provides superstar level impact right now. And you think it's arrogant to believe that maybe he doesn't?

I know what statistics do- the problem is that most people who deny impact metrics tend to rely on box statistics which are far, far less reliable when it comes to contributing to team success.

No, i believe it is arrogant to pretend that "you are giving him the benefit of the doubt" just because you decided to consider flawed statistics,just your choice, as the only basis to assess if he is or he is not an impact player.

The rest of the world, NBA professionals and statistics up to uneducated fans, already know he does have a superstar impact on the basis of a far broader range of stats and qualitative evaluations.

You should have just said Imho that you've found these interesting numbers and ask for possibke explainations. But none of these should lead to an absurd conclusion. Again, you are not convincing Carlisle to bench him, nor Dallas quants, nor LeBron to not consider him a great player.

You can go on finding a model telling us evolution theory is bull or people can fly, not my problem how you spend your time, in the meantime real world happens though.

[Sactowndog]

According to BasketballReference, Doncic ORTG is 119 and his DRTG is 109.

Só, how is he a negative-negative?

Offensive and Defensive rating emphasize counting stats. On Defense, Doncic does nothing but hunt steals and rebounds. A perfect example was the Knicks game tonight where he gave up a couple drives. When he wasn’t doing that, the Knicks put him in a pick and roll where Doncic made no effort to fight over the pick or recover. Over that time the Knicks significantly increased their lead.

[GeorgeMarcus]

It is not that it does not make sense, is just need to be interpreted, and different posters gave you a possible explaination, or it is a sign of the fallacy of the model.

With all the due respect, the fact that you are willing to give him the benefit of the doubt based on this statistics sounds pretty arrogant.
The whole NBA would kill to have Doncic and also numerous stats support this. If you want to be a doubter you can criticize lot of things about his game. But if you pretend to have found two numbers telling you a hidden truth while the other stats and evaluations, not to mention the whole NBA world made of professionals, tell you the opposite, it is just you putting to much weight on it.

Statistics describe the world and we need to find models fitting reality. Statistic do NOT CAUSE the world to transform according to our twists.

Doncic not going to regress to a bench level player just cause some stats tell you so. Also, you are not giving him the benefit of the doubt. We are giving you the benefit of the doubt and maybe you will change your opinion.

No one in the world cares if you (or me obv.) don't think he is an impact player based on these stats, he is according to basically all the other stats, other coaches, other gms and other players. It is just your very personal tweak of reality. To each his own i guess

The point in question is whether or not Doncic provides superstar level impact right now. And you think it's arrogant to believe that maybe he doesn't?

I know what statistics do- the problem is that most people who deny impact metrics tend to rely on box statistics which are far, far less reliable when it comes to contributing to team success.

No, i believe it is arrogant to pretend that "you are giving him the benefit of the doubt" just because you decided to consider flawed statistics,just your choice, as the only basis to assess if he is or he is not an impact player.

The rest of the world, NBA professionals and statistics up to uneducated fans, already know he does have a superstar impact on the basis of a far broader range of stats and qualitative evaluations.

You should have just said Imho that you've found these interesting numbers and ask for possibke explainations. But none of these should lead to an absurd conclusion. Again, you are not convincing Carlisle to bench him, nor Dallas quants, nor LeBron to not consider him a great player.

You can go on finding a model telling us evolution theory is bull or people can fly, not my problem how you spend your time, in the meantime real world happens though.

The statistics I used aren't "flawed". If you think it's a matter of sample size, then you would still acknowledge what it meant for his rookie season (which was the bulk of my argument if you read it).

Brandon Ingram is averaging 26ppg (more efficiently than Doncic), 7 boards and 4 assists. And guess what? He's still not close to a star-level player, and arguably not even a net positive. People who deny the kind of data I used are generally the same that wonder why players like Ingram always play on bad teams.... And the same that thought Lowry was some kind of playoff choke when he was always the reason the Raps made it as far as they did.

[Pumpkin17]

The point in question is whether or not Doncic provides superstar level impact right now. And you think it's arrogant to believe that maybe he doesn't?

I know what statistics do- the problem is that most people who deny impact metrics tend to rely on box statistics which are far, far less reliable when it comes to contributing to team success.

No, i believe it is arrogant to pretend that "you are giving him the benefit of the doubt" just because you decided to consider flawed statistics,just your choice, as the only basis to assess if he is or he is not an impact player.

The rest of the world, NBA professionals and statistics up to uneducated fans, already know he does have a superstar impact on the basis of a far broader range of stats and qualitative evaluations.

You should have just said Imho that you've found these interesting numbers and ask for possibke explainations. But none of these should lead to an absurd conclusion. Again, you are not convincing Carlisle to bench him, nor Dallas quants, nor LeBron to not consider him a great player.

You can go on finding a model telling us evolution theory is bull or people can fly, not my problem how you spend your time, in the meantime real world happens though.

The statistics I used aren't "flawed". If you think it's a matter of sample size, then you would still acknowledge what it meant for his rookie season (which was the bulk of my argument if you read it).

Brandon Ingram is averaging 26ppg (more efficiently than Doncic), 7 boards and 4 assists. And guess what? He's still not close to a star-level player, and arguably not even a net positive. People who deny the kind of data I used are generally the same that wonder why players like Ingram always play on bad teams.... And the same that thought Lowry was some kind of playoff choke when he was always the reason the Raps made it as far as they did.

They are very flawed, cause they are telling you, as an example, that Tim Hardaway Junior is a great player while Doncic is not. They are just statistics, but if you want to use these as a basis of a model and the model give you absurd results, again, your reasoning is flawed and your model as well.

Models need to fit reality. If your model doesn't, first consideration should be your model is not that good.

In this particular case, also, given the strnght of your assumption, i.e. Doncic is not a superstar level inpact player, a more proper approach would be collect A LOT of data and A LOT of statistics and then make your conclusions.

You cherry picked two interssting stats and are ignoring basically all the others which describe Doncic as a superstar. Also, you are completely ignoring qualitative judments made by coaches, professionals,players. You decide to oversimplify a question and try to find a solution using a model you created which also give you absurd results. It is like assuming water is always in a solid state cause you once had ice in your drink.

Want to use statistics? Consider a wide range of stats and see which result you end up with.

[GeorgeMarcus]

No, i believe it is arrogant to pretend that "you are giving him the benefit of the doubt" just because you decided to consider flawed statistics,just your choice, as the only basis to assess if he is or he is not an impact player.

The rest of the world, NBA professionals and statistics up to uneducated fans, already know he does have a superstar impact on the basis of a far broader range of stats and qualitative evaluations.

You should have just said Imho that you've found these interesting numbers and ask for possibke explainations. But none of these should lead to an absurd conclusion. Again, you are not convincing Carlisle to bench him, nor Dallas quants, nor LeBron to not consider him a great player.

You can go on finding a model telling us evolution theory is bull or people can fly, not my problem how you spend your time, in the meantime real world happens though.

The statistics I used aren't "flawed". If you think it's a matter of sample size, then you would still acknowledge what it meant for his rookie season (which was the bulk of my argument if you read it).

Brandon Ingram is averaging 26ppg (more efficiently than Doncic), 7 boards and 4 assists. And guess what? He's still not close to a star-level player, and arguably not even a net positive. People who deny the kind of data I used are generally the same that wonder why players like Ingram always play on bad teams.... And the same that thought Lowry was some kind of playoff choke when he was always the reason the Raps made it as far as they did.

They are very flawed, cause they are telling you, as an example, that Tim Hardaway Junior is a great player while Doncic is not. They are just statistics, but if you want to use these as a basis of a model and the model give you absurd results, again, your reasoning is flawed and your model as well.

Models need to fit reality. If your model doesn't, first consideration should be your model is not that good.

In this particular case, also, given the strnght of your assumption, i.e. Doncic is not a superstar level inpact player, a more proper approach would be collect A LOT of data and A LOT of statistics and then make your conclusions.

You cherry picked two interssting stats and are ignoring basically all the others which describe Doncic as a superstar. Also, you are completely ignoring qualitative judments made by coaches, professionals,players. You decide to oversimplify a question and try to find a solution using a model you created which also give you absurd results. It is like assuming water is always in a solid state cause you once had ice in your drink.

Want to use statistics? Consider a wide range of stats and see which result you end up with.

A LOT of data went into the numbers I used. You're fundamentally misunderstanding my approach on multiple levels. If you understood what I did and the significance of those values, your defense would be to present confounding variables that influence those results. Not deny them altogether. There's a reason why he was the only player out of the top 61 to meet the criteria.

[AIfan3]

Porzingis. And I honestly thought he'd be the more impact of the two. Boy was I wrong.. Guy is playing like a complete BUM.

[Pumpkin17]

The statistics I used aren't "flawed". If you think it's a matter of sample size, then you would still acknowledge what it meant for his rookie season (which was the bulk of my argument if you read it).

Brandon Ingram is averaging 26ppg (more efficiently than Doncic), 7 boards and 4 assists. And guess what? He's still not close to a star-level player, and arguably not even a net positive. People who deny the kind of data I used are generally the same that wonder why players like Ingram always play on bad teams.... And the same that thought Lowry was some kind of playoff choke when he was always the reason the Raps made it as far as they did.

They are very flawed, cause they are telling you, as an example, that Tim Hardaway Junior is a great player while Doncic is not. They are just statistics, but if you want to use these as a basis of a model and the model give you absurd results, again, your reasoning is flawed and your model as well.

Models need to fit reality. If your model doesn't, first consideration should be your model is not that good.

In this particular case, also, given the strnght of your assumption, i.e. Doncic is not a superstar level inpact player, a more proper approach would be collect A LOT of data and A LOT of statistics and then make your conclusions.

You cherry picked two interssting stats and are ignoring basically all the others which describe Doncic as a superstar. Also, you are completely ignoring qualitative judments made by coaches, professionals,players. You decide to oversimplify a question and try to find a solution using a model you created which also give you absurd results. It is like assuming water is always in a solid state cause you once had ice in your drink.

Want to use statistics? Consider a wide range of stats and see which result you end up with.

A LOT of data went into the numbers I used. You're fundamentally misunderstanding my approach on multiple levels. If you understood what I did and the significance of those values, your defense would be to present confounding variables that influence those results. Not deny them altogether. There's a reason why he was the only player out of the top 61 to meet the criteria.

I am not misunderstanding anything. You are misunderstanding your own work and fail to uderstand the basic concepts of statistics, building a model and interpretation of results. In a scientific environment, your approach would result only in a lot of laughs and you are failing to recognise it. You would not pass a college exam with this kind of approach.

Anyway, go on and best of luck, I am done here and for sure I don't want to be the one discussing against the "prophet of the +/- impact model" which is going to relegate Doncic on the bench. I am sure your crusade will lead to Doncic being cut by Mavs by Christmas and being forgotten by general public, you just need a bigger audience.

Sure, we are fundamentally misunderstanding your great model, it's not that you are looking at a single stats without context. Please make some Youtube videos so that you can spread your theory to the world. We are so much in need to be educated and we all wish so much to see THJ winning the MVP.

[Baski]

It's amazing what can be read at this forum lately.

- Luka is actually bad because of his +/-, the coaches, opposing players and fans are just running a propaganda machine
- Steph was always overrated, and he is now delaying his return because obviously a broken hand is not a big deal since Hayward can come back sooner
- Draymond is the most overrated guy in history, every coach and media member that kept voting him all-nba and all-star was clearly part of a conspiracy
- Kawhii sucks because he is load managing himself, he was never that good in the first place, again propaganda.

On the other hand we have
- Pacers are title contenders
- The Bucks didn't beat anyone last year in the playoffs since Pistons/Celtics don't count
- LeBron sucks cause he is getting blocked more.

It seems the only important thing is to have a hot take that riles people up for debate. I'm actually starting to doubt people watch more than highlights and go further than ESPN titles to have what is supposed to be an informed opinion about basketball.

One of those is a misconception

[Baski]

I partly agree with you. The problem is that this thread was made yesterday, using 10 games numbers, with only 1 clear objective, show how bad Luka is. OP was pretty sure what this numbers mean. Luka is massively overrated player.
+/- is and will always be wrong stat to measure player impact. You might use it just as complementary data, but that’s more or less it. What I can see from this lineups comparison stats is, that Mavs are far from playoffs team. They need another 2 solid starter that will enable starting unit to be competitive against the best teams. And there is another problem for the Mavs, it’s highly unlikely that Luka can play like that the whole season. I expect a drop in his performance.

It worries me whenever I see "this stat is not right" because of one player. With all due respect, surely you have to see the flaw in dismissing a stat that correlates positively with good players except for one guy, because of that one guy. Why not discuss the reasons for Luka's poor stats (which I see you've done) and apply context instead of dismissing the stat as a whole?

It's even more worrying because we all know that in a few year's time, if Luka becomes as good as we all think, this same stat will paint him in a positive light right alongside the other stars, and there won't be anyone questioning it as you're doing now.

[Alyosha12]

Watch the game @Knicks and you will have all the answers. Luka is struggling from the field but without him the Mavs would be down double digits. Nobody can hit a shot. They miss open layups, uncontested putbacks and wide open 3s.
Luka scored 13 of 23 points and had 2 assists. Responsible for 80% of the Mavs offense. That´s not how it should be but there is no other option right now.

The bolded should reveal itself in the on/off data. Perhaps it will overtime, but it hasn't yet.

But you see, in this logic lays your fault. It is not true that the bolded would reveal itself in on/off. Not necessarily at least.

Firstly you are using average stas, not median stats. One game could throw your whole metric off balance, which it did (the Denver game). Luka's on/off for the first 10 games has been positive 6 times and negative 4 times. So he is a net positive player on average.

Even with that, both on/off and+/- have the same flaw, which is that the premise for them both is, all things being equal if a player is on court the team scores this much if he is not, they score or are scored on this much. The fundamental flaw with this is, that all things are never equal. The same situation in game never happens. Both +/- and on/off don't take into account game flow and momentum.

Who is to say that in the first 5 minutes of the game and Luka plays and Dallas is losing by 10 (his +/- then being -10) Dallas wouldn't be losing by 20 if THJ was playing instead of him? You would say, well THJ's +/- numbers say that ofcourse, which is incorrect. THJ's +/- and on/off only say that the next 5 minutes he plays the team does better. But those 5 minutes of the game are not the same as the previous ones. May be the opposite team exhausted themselves trying desperately to take that 20 pt run, or may be their run was broken by a timely pass, a timely 3pt etc. The point is, you can't know, how THJ would have done in the first 5 minutes instead of Luka. YOu can only know, how he does after Luka based on on/off and +/-.

The second fault, from an impact perspective is that +/- tracks score shifts, but doesn't track runs. It doesn't track how a player can take over a game. For example, lets say Luka plays the first 10 minutes, Dallas is loosing by 10 for those 10 minutes and his +/- being -10. He goes and sits, THJ comes in, the score shifts by 2, THJ is at +2. Luka comes back in, the teams trade baskets but the score doesn't shift, his +/1 stays -10. THJ comes in again, Kleber makes a 3 to end the second q, THJ's +/- is now +5. Dallas is still down 5. THJ starts the third, Dallas starts strong shifting the score by 3, THJ'S +/- is now +8, he fouls the opposing player, who is rewarded 2 FTS. Before he takes them, THJ is subbed out and Luka is Subbed in, the opposing player makes both FTs, Luka's +/- is now -12. Dallas is down by 4 now.

Luka takes over the game and is crucial in a 10-0 run. Dallas takes the lead by 6, and is able to hold the lead till the end of the game. Luka has a +/- of -2, THJ has a +/- of +8. The impact Luka had on the game is huge, he was responsible for the lead change and Dallas taking over the game, however +/- doesn't show that. It doesn't show the impact, it doesn't show the influence Luka had on the game momentum. That is impact, that is superstar power. Taking over a game. Imposing your will on the other team, your rhythm. IMO that is why your model is flawed.

Though TBH, i do think you are on to something, and that is, that Dallas is not using Luka correctly, and i still stand by my statement, that having him at PG is not the best thing for any team. He is not a true PG and game flow does seem to suffer because of it. Thats what I think at least. Him at SG and a an aggressive PG like Rose, Dragic, CP3 or Fox, would be amazing.

Though it could be like someone said in another thread that Carlisle is grooming him into a superstar, that is why he is giving him the usage he has atm. Which also might be true.

[Bob8]

I partly agree with you. The problem is that this thread was made yesterday, using 10 games numbers, with only 1 clear objective, show how bad Luka is. OP was pretty sure what this numbers mean. Luka is massively overrated player.
+/- is and will always be wrong stat to measure player impact. You might use it just as complementary data, but that’s more or less it. What I can see from this lineups comparison stats is, that Mavs are far from playoffs team. They need another 2 solid starter that will enable starting unit to be competitive against the best teams. And there is another problem for the Mavs, it’s highly unlikely that Luka can play like that the whole season. I expect a drop in his performance.

It worries me whenever I see "this stat is not right" because of one player. With all due respect, surely you have to see the flaw in dismissing a stat that correlates positively with good players except for one guy, because of that one guy. Why not discuss the reasons for Luka's poor stats (which I see you've done) and apply context instead of dismissing the stat as a whole?

It's even more worrying because we all know that in a few year's time, if Luka becomes as good as we all think, this same stat will paint him in a positive light right alongside the other stars, and there won't be anyone questioning it as you're doing now.

This my last comment. +/- is 5 man lineup stat, not individual stat. I didn’t comment Luka’s poor stat, but problems in Mavs starting lineup. There’s just to many variables which influence single player numbers. Let’s look at Boston game, Tatum 1/18 shooting, but he had +16, 3rd best +/- in his team. According +/- he had fantastic game, on pair with Kemba. Btw. far the best was Theis, playing only 21 minutes. Luka on the other hand had 34/7/9 on 67 TS% only 1 TO and - 17 +/-. Was he really that disastrous in that game? You will say, everything is normall, they have lost. Ok lets look at Hardaway, 9/4/1, 3/11 shooting, 36% TS and +8. Had Hardaway a good game? Of course not. Do you know, what had the biggest impact on this 3 players’s +/-? Kemba going supernova. 10 points in 90 seconds. Luka and Tatum being on court and Hardaway not. What has fantastic 90 seconds of Kemba with individual performance of Luka, Tatum and Hardaway? Not much. But +/-, on/off says everything. Even for Hardaway, who wasn’t even in the court. Yes, it’s only 1 game, but this data is total garbage. I know, only 1 game, on long run all things even up. Do they? With more data we probably won’t see extremes like this, but more garbage data is still garbage data, no matter how you manipulate it or compare with other data. +/- is useable to compare different lineups, even how single players function in different lineups, but not what is their individual impact on the team.

We will very soon have out RPM and similar advanced data, which are made to show players individual performance. I can bet whatever you want, that Luka will be light years better than THJ, who has double positive and elite +/-, on/off. I’m sure that Luka will have great offensive advanced stats, among the best in his position. But it’s also true that his defensive stats will imho be very bad. What’s even a point to lose time with so flawed stats, for measuring individual performance, like +/- is, if we have a tonne of other advanced stats build for measuring individual impact?

[Archx]

According to BasketballReference, Doncic ORTG is 119 and his DRTG is 109.

Só, how is he a negative-negative?

Offensive and Defensive rating emphasize counting stats. On Defense, Doncic does nothing but hunt steals and rebounds. A perfect example was the Knicks game tonight where he gave up a couple drives. When he wasn’t doing that, the Knicks put him in a pick and roll where Doncic made no effort to fight over the pick or recover. Over that time the Knicks significantly increased their lead.

You know Luka is a superstar when people start cherry picking dumb things about him and even start making stuff up.

Looking at the NYK game, what did you want him to do? Because it felt like he was literally doing everything this game. Yes, agreed he could have shot better BUT he provided 80% of the Mavs offense in the 1st Q, he helped them with assists in the 2nd Q, Mavs at the end never trailed by a large margin. He also kept them in the game in the 3rd with 14 straight points if i am not mistaken. He got so many Knicks players in foul trouble that one of them fouled out at the end. Also had 3 steals and a block and only 3 TO's. Mavs had 21 assists, 11 of those came from Luka. Bench couldn't provide any reasonable playmaking, at some point in the 4th Q, Wright simply stopped dribbling, and because he couldn't get the shot off, he simply waited till the shot clock runs out, they were that bad.

Trying to pin point this loss on Luka because he missed a couple of defensive assignments is mind boggling. NYK shot 43% from 3, that's also Doncic's fault, right?

[marco102]

According to BasketballReference, Doncic ORTG is 119 and his DRTG is 109.

Só, how is he a negative-negative?

Offensive and Defensive rating emphasize counting stats. On Defense, Doncic does nothing but hunt steals and rebounds. A perfect example was the Knicks game tonight where he gave up a couple drives. When he wasn’t doing that, the Knicks put him in a pick and roll where Doncic made no effort to fight over the pick or recover. Over that time the Knicks significantly increased their lead.

You know Luka is a superstar when people start cherry picking dumb things about him and even start making stuff up.

Looking at the NYK game, what did you want him to do? Because it felt like he was literally doing everything this game. Yes, agreed he could have shot better BUT he provided 80% of the Mavs offense in the 1st Q, he helped them with assists in the 2nd Q, Mavs at the end never trailed by a large margin. He also kept them in the game in the 3rd with 14 straight points if i am not mistaken. He got so many Knicks players in foul trouble that one of them fouled out at the end. Also had 3 steals and a block and only 3 TO's. Mavs had 21 assists, 11 of those came from Luka. Bench couldn't provide any reasonable playmaking, at some point in the 4th Q, Wright simply stopped dribbling, and because he couldn't get the shot off, he simply waited till the shot clock runs out, they were that bad.

Trying to pin point this loss on Luka because he missed a couple of defensive assignments is mind boggling. NYK shot 43% from 3, that's also Doncic's fault, right?

Nah, what the original poster is saying is stop calling Luka a superstar when he obviously hasn't earned it.

Devin Booker is putting up a much more efficient season and it's having an impact on winning, yet you guys will NEVER say Devin is a superstar.

Superstars are made in the playoffs. Once Luka drags the Mavs there and puts up numbers to help then succeed, we can then start having the superstar conversation.

It's just annoying you guys are throwing Superstar around and being so damn hyperbolic. Luka's great by the way, but his stans are worst thing to happen to the nba in a while.