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

[Richard Miller]

Trying not to be triggered for using who's instead of whose

He used that by purpose, because he knows that many of his supporters are from Slovenia, English being 3rd or 4th language.

Deliberately teaching bad English to them? Oh, I see it's already corrected, thanks mods.

[Bob8]

Trying not to be triggered for using who's instead of whose

He used that by purpose, because he knows that many of his supporters are from Slovenia, English being 3rd or 4th language.

Deliberately teaching bad English to them? Oh, I see it's already corrected, thanks mods.

He knows that everyone knows who’s, but whose is a different kind of animal, similar like using +/- to measure individual player’s impact, not many know that too.

[GeorgeMarcus]

Regarding RPM: you know what that stat is designed to do? It's designed to approximate RAPM, so it can be used in smaller sample sizes. If we direct our attention to the number it's trying to emulate, Luka's RAPM last season was -0.3068. Even if we took RPM as scripture, Luka ranked 86th last season at +1.29. Significantly worse than any other player regarded as a star. I repeat: I'm not saying Luka is a negative impact player. I'm saying he doesn't provide star impact (yet).

Hey GeorgeMarcus, great OP and great thread.

I agree with you 100% about last season. As I stated in the first week of this season, I didn't/don't believe Doncic was a top 30 player last season.

The fatal flaw in your idea, in my opinion, is that you're putting last season's sample size of 2,318 minutes and this season's sample size of 348 minutes on apparently equal footing. Certainly, this season hasn't contradicted the idea that he's not yet a superstar, but I think you're jumping the gun.

This season, I think that over time you'll find he's made a huge leap. (I'm assuming you haven't watched more than one or two of his games this season?) The tweet posted about Porzingis' minutes and +/- is relevant, no? Assuming that KP and Doncic have been sharing the floor a lot? And given that this time with Porzingis is already weighing down this season, it may take a little longer for Doncic to dig out of this +/- hole.

For sure, many are overrating him this season as they did last season (Bill Simmons having him as a top 2 MVP candidate is laughable), but I think he's much closer to superstar impact than this single, small-sample-sized metric implies, and I think that by the All-Star break, even this metric will make that clear.

EDIT: I'm reminded of Westbrook being -46 against Miami and +40 against Golden State this season. Imagine if he'd rested one of those games. In short, it's just too early to use +/- for this season.

Thanks man, I think you addressed the 2 most acceptable explanations: KP and sample size from this year’s dataset. KP has played 64.4% of Luka’s on minutes and 43.1% of Luka’s off minutes, which is enough of a correlation to affect Luka’s on/off. For Luka to be double digit negative is still note worthy, but palatable this early in the season.

I do think the data last year shows that he wasn’t the star some touted him to be as a rookie, but he’s certainly on the trajectory to become one.

[GeorgeMarcus]

GeorgeMarcus, just to clear up any confusion again, is this thread about players who's impact doesn't match their reputation in general...or specificly the player (luka doncic) who's impact is so far away from his reputation it could be considered an outlier? Or is this thread about both of these things?

Lol this time it’s specifically about Luka

[GeorgeMarcus]

Trying not to be triggered for using who's instead of whose

I wish you had posted that earlier. It’s bothering me even after the edit

[scrabbarista]

Try this formula in every game instead of on off.

OffRtg - Drtg = x (not net rating)

This formula penalizes you for turnovers. Learned this from AirP. Been mindblowingly awesome to watch games and then the next day on bbref, watch that calculation agree with my eye test.

Yeah, at one point I was doing a single-season peak formula and went through tons of examples of this little equation. Since then it's one I often glance at on bbref. The GOATs are like +20-25, and the serious contributors are usually around +10-12, if I'm not mistaken.

Here are four random selections:

2016 Curry = +22
2020 Gordon Hayward = +16
2020 Doncic = +11
2019 Doncic = -2

Since this thread is about Luka and I find it really strange that he could be a minus anything on the Mavs, I checked the game logs for the simple formula Ortg - Drtg. Mind blowing as KingDavid states.

Luka had a minus in two wins using that simple formula.
Dallas at Denver 10-29, -33 by that formula.
Orlando vs Dallas 11-06, -5 by that formula.
The only other minus is a game in which the Mavs lost and he had a -1 on 10-27 against Portland.

I am guessing that -33 is a huge outlier just by me never being able to see Luka that bad a negative; and maybe could explain a lot about the stats of on/off having him in a negative for 2019. I do not go through stats, so its just a guess on my part; but -33 seems huge.

No, you're confused. The formula KingDavid is talking about and the one the OP is talking about are two different stats. In the one KingDavid is talking about (as stated in my post) Doncic is +11, not negative. That's the one you looked at.

The one the OP is talking about is a different stat. Or, actually, two different stats. In both of those stats, Doncic is currently negative. If you go to basketball-reference.com and go down to "Play-by-Play" and look in the middle, you'll see something marked "+/- Per 100 Possessions." The two stats the OP is referring to are the "On Court" stat and the "On/Off" stat.

But you are getting at a simple criticism that still applies to the stats the OP is referring to, which is that the sample size is very small. For example, last season Doncic's On/Off was -3.9, and this season it's -14.5. Even at first glance, for anyone who's watched Doncic play this season, this is obviously an extreme result that is probably tied to small sample size.

[Bob8]

Yeah, at one point I was doing a single-season peak formula and went through tons of examples of this little equation. Since then it's one I often glance at on bbref. The GOATs are like +20-25, and the serious contributors are usually around +10-12, if I'm not mistaken.

Here are four random selections:

2016 Curry = +22
2020 Gordon Hayward = +16
2020 Doncic = +11
2019 Doncic = -2

Since this thread is about Luka and I find it really strange that he could be a minus anything on the Mavs, I checked the game logs for the simple formula Ortg - Drtg. Mind blowing as KingDavid states.

Luka had a minus in two wins using that simple formula.
Dallas at Denver 10-29, -33 by that formula.
Orlando vs Dallas 11-06, -5 by that formula.
The only other minus is a game in which the Mavs lost and he had a -1 on 10-27 against Portland.

I am guessing that -33 is a huge outlier just by me never being able to see Luka that bad a negative; and maybe could explain a lot about the stats of on/off having him in a negative for 2019. I do not go through stats, so its just a guess on my part; but -33 seems huge.

No, you're confused. The formula KingDavid is talking about and the one the OP is talking about are two different stats. In the one KingDavid is talking about (as stated in my post) Doncic is +11, not negative. That's the one you looked at.

The one the OP is talking about is a different stat. Or, actually, two different stats. In both of those stats, Doncic is currently negative. If you go to basketball-reference.com and go down to "Play-by-Play" and look in the middle, you'll see something marked "+/- Per 100 Possessions." The two stats the OP is referring to are the "On Court" stat and the "On/Off" stat.

But you are getting at a simple criticism that still applies to the stats the OP is referring to, which is that the sample size is very small. For example, last season Doncic's On/Off was -3.9, and this season it's -14.5. Even at first glance, for anyone who's watched Doncic play this season, this is obviously an extreme result that is probably tied to small sample size.

Stat that Tim Hardaway has + 17.6, which should explain everything about representative value of this stat at this moment.

[scrabbarista]

Since this thread is about Luka and I find it really strange that he could be a minus anything on the Mavs, I checked the game logs for the simple formula Ortg - Drtg. Mind blowing as KingDavid states.

Luka had a minus in two wins using that simple formula.
Dallas at Denver 10-29, -33 by that formula.
Orlando vs Dallas 11-06, -5 by that formula.
The only other minus is a game in which the Mavs lost and he had a -1 on 10-27 against Portland.

I am guessing that -33 is a huge outlier just by me never being able to see Luka that bad a negative; and maybe could explain a lot about the stats of on/off having him in a negative for 2019. I do not go through stats, so its just a guess on my part; but -33 seems huge.

No, you're confused. The formula KingDavid is talking about and the one the OP is talking about are two different stats. In the one KingDavid is talking about (as stated in my post) Doncic is +11, not negative. That's the one you looked at.

The one the OP is talking about is a different stat. Or, actually, two different stats. In both of those stats, Doncic is currently negative. If you go to basketball-reference.com and go down to "Play-by-Play" and look in the middle, you'll see something marked "+/- Per 100 Possessions." The two stats the OP is referring to are the "On Court" stat and the "On/Off" stat.

But you are getting at a simple criticism that still applies to the stats the OP is referring to, which is that the sample size is very small. For example, last season Doncic's On/Off was -3.9, and this season it's -14.5. Even at first glance, for anyone who's watched Doncic play this season, this is obviously an extreme result that is probably tied to small sample size.

Stat that Tim Hardaway has + 17.6, which should explain everything about representative value of this stat at this moment.

Over his career, THJ is a -3.7 on OnCourt and a 0.0 on On-Off. The stat is useful and meaningful, but it's not a single number to tell you "how good" a player is, and it definitely requires more than 10 games to become a "real" representation of a player's performance. This is just a small sample. I say we revisit this thread around the All-Star Break and again at the end of the regular season. We should also keep an eye on Dallas' team record and whether they make the playoffs. All of this, combined with the eye-test and Luka's other traditional and advanced numbers, will tell us the story of how good he is. For now, let's just keep an eye on this stat and enjoy the ride that Luka and the Mavs are on.

[GeorgeMarcus]

I don't think +/- as a stat should be used t evaluate individual players/player comparison. It should be used by teams to evaluate and analyse which lineups work and which don't. It shows the cohesiveness of units and what they specialize in. Of-course, you can use it in conjunction with other stats and perform a detailed analysis to gauge the player better, but it's a poor stat to adjudge an individual.

We agree and disagree to an extent. Pure impact metrics (iterations of +/- like the ones I used, RAPM, etc) are extremely predictive and transferable across different settings. However, that stops being the case if a player is being used in a fundamentally different role which you alluded to. If the bulk of a player’s impact comes from his scoring ability and his scoring volume is cut in half, then he won’t be nearly as impactful as he could be. If a player provides the most impact as a rim protector and he’s forced to defend wings, then he won’t be nearly as impactful as he could be.

Navigating “impact” still makes more sense then raw production though. Those numbers don't depend on a player’s role/fit any less. There are so many variables that get wrongly dismissed when it comes to counting stats, but nothing gets dismissed under the impact umbrella. Influenced by external factors? Sure, but not dismissed. And again- raw production is influenced by external factors too. We just have to acknowledge and investigate those factors to contextualize the data. At the end of the day, impact on winning is the only thing that matters.

[Bob8]

No, you're confused. The formula KingDavid is talking about and the one the OP is talking about are two different stats. In the one KingDavid is talking about (as stated in my post) Doncic is +11, not negative. That's the one you looked at.

The one the OP is talking about is a different stat. Or, actually, two different stats. In both of those stats, Doncic is currently negative. If you go to basketball-reference.com and go down to "Play-by-Play" and look in the middle, you'll see something marked "+/- Per 100 Possessions." The two stats the OP is referring to are the "On Court" stat and the "On/Off" stat.

But you are getting at a simple criticism that still applies to the stats the OP is referring to, which is that the sample size is very small. For example, last season Doncic's On/Off was -3.9, and this season it's -14.5. Even at first glance, for anyone who's watched Doncic play this season, this is obviously an extreme result that is probably tied to small sample size.

Stat that Tim Hardaway has + 17.6, which should explain everything about representative value of this stat at this moment.

Over his career, THJ is a -3.7 on OnCourt and a 0.0 on On-Off. The stat is useful and meaningful, but it's not a single number to tell you "how good" a player is, and it definitely requires more than 10 games to become a "real" representation of a player's performance. This is just a small sample. I say we revisit this thread around the All-Star Break and again at the end of the regular season. We should also keep an eye on Dallas' team record and whether they make the playoffs. All of this, combined with the eye-test and Luka's other traditional and advanced numbers, will tell us the story of how good he is. For now, let's just keep an eye on this stat and enjoy the ride that Luka and the Mavs are on.

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.

[GeorgeMarcus]

He used that by purpose, because he knows that many of his supporters are from Slovenia, English being 3rd or 4th language.

Deliberately teaching bad English to them? Oh, I see it's already corrected, thanks mods.

He knows that everyone knows who’s, but whose is a different kind of animal, similar like using +/- to measure individual player’s impact, not many know that too.

I’ve been doing this a long time guy, and mostly to the tune of disgruntled fans like yourself who don’t like my findings. They try to undermine my methods based on misguided interpretations of the numbers and their implications. I heard it from Celtics fans last year when they got Kyrie, and Nets fans again this year. I heard it from Spurs fans when I emphasized the importance of Davis Bertans, and I heard it from Raptors haters when I praised the impact of Kyle Lowry. The list goes on but rest assured, I’m not seeking anyone’s approval. If I was I wouldn’t post a Doncic criticism on RealGM. Ultimately I know the legitimacy of my research, and I posted my findings in order to delve deeper into the root causes.

[VanWest82]

Jayson Tatum.

You mean the guy with the hightest +/- of anybody in the league (by far) Jayson Tatum? Or is there another Jayson Tatum that you are thinking of?

Yup, and this is partly what I’m talking about. The whole Celtics starting unit has gaudy +/- #s. If you’re referring to his OFF numbers which make it look like the Celtics offense just can’t possibly function without his 39% FG% and middling to below average play making then yeah I call BS.

He’s been Celtics 4th best player so far this year. To hear Celtics fans talk about it he’s this untradable all star, future MVP candidate and I just don’t see it. I don’t see that type of impact when I watch him play. He’s a good three point shooter and system defender. He doesn’t get to the rim or play make for others. He’s a one to two dribble player with a tendency to float in the offense. Obviously he has talent and I’m not saying he can’t one day be a great player with a lot of improvement but as the thread title says right now he’s someone who’s on court impact doesn’t match the reputation he has among NBA fans as this sure fire potential top 10 guy.

[Bob8]

Deliberately teaching bad English to them? Oh, I see it's already corrected, thanks mods.

He knows that everyone knows who’s, but whose is a different kind of animal, similar like using +/- to measure individual player’s impact, not many know that too.

I’ve been doing this a long time guy, and mostly to the tune of disgruntled fans like yourself who don’t like my findings. They try to undermine my methods based on misguided interpretations of the numbers and their implications. I heard it from Celtics fans last year when they got Kyrie, and Nets fans again this year. I heard it from Spurs fans when I emphasized the importance of Davis Bertans, and I heard it from Raptors haters when I praised the impact of Kyle Lowry. The list goes on but rest assured, I’m not seeking anyone’s approval. If I was I wouldn’t post a Doncic criticism on RealGM. Ultimately I know the legitimacy of my research, and I posted my findings in order to delve deeper into the root causes.

Interesting. It’s kinda strange though that you have admitted few minutes back, how KP and not enough data crucially influenced your findings about Luka. It seems to me that you posted your findings at least a bit premature. And it showed that your methodology has a big problems, if a teammate of a player is hugely negative as Kp is.
I’m disgruntled fan, because I don’t agree with yours flawed methodology, which you even admitted yourself? That’s something new. Nobody force you to post premature data and make premature conclusions. Now it is what it is. I’m 100% sure, that you wouldn’t made that conclusion, if you had watched all Mavs games, instead of only looking at numbers. And would rather find the problem in calculations.

[Hesh]

I'm just waiting to see the Booker discussion here, it should be coming.

[Bob8]

I'm just waiting to see the Booker discussion here, it should be coming.

no, Booker is double positive in a good way of course.

[Hesh]

I'm just waiting to see the Booker discussion here, it should be coming.

no, Booker is double positive in a good way of course.

[Alyosha12]

I don't think +/- as a stat should be used t evaluate individual players/player comparison. It should be used by teams to evaluate and analyse which lineups work and which don't. It shows the cohesiveness of units and what they specialize in. Of-course, you can use it in conjunction with other stats and perform a detailed analysis to gauge the player better, but it's a poor stat to adjudge an individual.

We agree and disagree to an extent. Pure impact metrics (iterations of +/- like the ones I used, RAPM, etc) are extremely predictive and transferable across different settings. However, that stops being the case if a player is being used in a fundamentally different role which you alluded to. If the bulk of a player’s impact comes from his scoring ability and his scoring volume is cut in half, then he won’t be nearly as impactful as he could be. If a player provides the most impact as a rim protector and he’s forced to defend wings, then he won’t be nearly as impactful as he could be.

Navigating “impact” still makes more sense then raw production though. Those numbers don't depend on a player’s role/fit any less. There are so many variables that get wrongly dismissed when it comes to counting stats, but nothing gets dismissed under the impact umbrella. Influenced by external factors? Sure, but not dismissed. And again- raw production is influenced by external factors too. We just have to acknowledge and investigate those factors to contextualize the data. At the end of the day, impact on winning is the only thing that matters.

Based on your explanation, you came to the wrong conclusion. If what you say is what you believe (even if i do not share your opinion), then the stats you used, don't show the players impact, but rather if a player is being utilized correctly. If a player can have a positive +/- or off-on in one role, and a negative in another, and still be the same player, then he is not being utilized to his full potential or is being utilized incorrectly.

I personally do not put any stock into +/- as a stat to judge a players impact simply because there are too many variables in a game that swing a result. Like Kobe said, basketball is a game of momentum, and it is a lottery when a team can go on a run.

However i do believe even with Luka's amazing numbers, he is not being utilized correctly. I do not think he should be the primary ballhandler and PG on any team, but rather a SG which is his natural position. Would he get less assists? Sure, however his gravity would open the flow of the offense more, he would get a few more open looks, and he wouldn't destroy game flow, with desperation step back threes as much. The team would thrive with a PG, who can penetrate a lot and plays great defense and Lua at the SG.

Every coach Lebron had up until now and Lebron himself understood that as well, which is why he was always pared with such a PG.

Can someone please check what the team +/- is when Luka and Brunson are on the court together, or Barea or Wright?

[Bob8]

I'm just waiting to see the Booker discussion here, it should be coming.

no, Booker is double positive in a good way of course.

Seriously. Don't worry, this thread was fabricated just for Luka.

[Bob8]

I don't think +/- as a stat should be used t evaluate individual players/player comparison. It should be used by teams to evaluate and analyse which lineups work and which don't. It shows the cohesiveness of units and what they specialize in. Of-course, you can use it in conjunction with other stats and perform a detailed analysis to gauge the player better, but it's a poor stat to adjudge an individual.

We agree and disagree to an extent. Pure impact metrics (iterations of +/- like the ones I used, RAPM, etc) are extremely predictive and transferable across different settings. However, that stops being the case if a player is being used in a fundamentally different role which you alluded to. If the bulk of a player’s impact comes from his scoring ability and his scoring volume is cut in half, then he won’t be nearly as impactful as he could be. If a player provides the most impact as a rim protector and he’s forced to defend wings, then he won’t be nearly as impactful as he could be.

Navigating “impact” still makes more sense then raw production though. Those numbers don't depend on a player’s role/fit any less. There are so many variables that get wrongly dismissed when it comes to counting stats, but nothing gets dismissed under the impact umbrella. Influenced by external factors? Sure, but not dismissed. And again- raw production is influenced by external factors too. We just have to acknowledge and investigate those factors to contextualize the data. At the end of the day, impact on winning is the only thing that matters.

Based on your explanation, you came to the wrong conclusion. If what you say is what you believe (even if i do not share your opinion), then the stats you used, don't show the players impact, but rather if a player is being utilized correctly. If a player can have a positive +/- or off-on in one role, and a negative in another, and still be the same player, then he is not being utilized to his full potential or is being utilized incorrectly.

I personally do not put any stock into +/- as a stat to judge a players impact simply because there are too many variables in a game that swing a result. Like Kobe said, basketball is a game of momentum, and it is a lottery when a team can go on a run.

However i do believe even with Luka's amazing numbers, he is not being utilized correctly. I do not think he should be the primary ballhandler and PG on any team, but rather a SG which is his natural position. Would he get less assists? Sure, however his gravity would open the flow of the offense more, he would get a few more open looks, and he wouldn't destroy game flow, with desperation step back threes as much. The team would thrive with a PG, who can penetrate a lot and plays great defense and Lua at the SG.

Every coach Lebron had up until now and Lebron himself understood that as well, which is why he was always pared with such a PG.

Can someone please check what the team +/- is when Luka and Brunson are on the court together, or Barea or Wright?

You're describing Slovenian golden team, with Dragic as Pg and Luka as Sg. I have to agree there, the problem is, that Mavs don't have player like that.

[Alyosha12]

We agree and disagree to an extent. Pure impact metrics (iterations of +/- like the ones I used, RAPM, etc) are extremely predictive and transferable across different settings. However, that stops being the case if a player is being used in a fundamentally different role which you alluded to. If the bulk of a player’s impact comes from his scoring ability and his scoring volume is cut in half, then he won’t be nearly as impactful as he could be. If a player provides the most impact as a rim protector and he’s forced to defend wings, then he won’t be nearly as impactful as he could be.

Navigating “impact” still makes more sense then raw production though. Those numbers don't depend on a player’s role/fit any less. There are so many variables that get wrongly dismissed when it comes to counting stats, but nothing gets dismissed under the impact umbrella. Influenced by external factors? Sure, but not dismissed. And again- raw production is influenced by external factors too. We just have to acknowledge and investigate those factors to contextualize the data. At the end of the day, impact on winning is the only thing that matters.

Based on your explanation, you came to the wrong conclusion. If what you say is what you believe (even if i do not share your opinion), then the stats you used, don't show the players impact, but rather if a player is being utilized correctly. If a player can have a positive +/- or off-on in one role, and a negative in another, and still be the same player, then he is not being utilized to his full potential or is being utilized incorrectly.

I personally do not put any stock into +/- as a stat to judge a players impact simply because there are too many variables in a game that swing a result. Like Kobe said, basketball is a game of momentum, and it is a lottery when a team can go on a run.

However i do believe even with Luka's amazing numbers, he is not being utilized correctly. I do not think he should be the primary ballhandler and PG on any team, but rather a SG which is his natural position. Would he get less assists? Sure, however his gravity would open the flow of the offense more, he would get a few more open looks, and he wouldn't destroy game flow, with desperation step back threes as much. The team would thrive with a PG, who can penetrate a lot and plays great defense and Lua at the SG.

Every coach Lebron had up until now and Lebron himself understood that as well, which is why he was always pared with such a PG.

Can someone please check what the team +/- is when Luka and Brunson are on the court together, or Barea or Wright?

You're describing Slovenian golden team, with Dragic as Pg and Luka as Sg. I have to agree there, the problem is, that Mavs don't have player like that.

Also Real's team, with Capazzo and Doncic.