RPM released

[dirkfansince1998]

http://www.espn.com/nba/statistics/rpm

Noteable numbers.

Luka ranks 15th in the league. 5th among SFs. He has the worst DRPM among regular rotation players at -0.9. On offense he ranks 7th in the entire league.

Brunson really suprises with the 17th best overall RPM. 3rd among PGs. Nearly a +2 on defense.
Delon ranks 15th among PGs. JJB manages to sneak in right in front of him. Ranking 13th among PGs in a really small sample size.

Curry ranks 23th among SG. Only + 0.13 ORPM but +1.26 DRPM.
THJ is right behind him at 26. His numbers are the direct opposite of Currys. +1.26 ORPM, -0.05 DRPM.

DFS is the only regular Mavs rotation with a negative RPM (-0.23). Him being the 2nd worst defender (-0.85 DRPM) among them is a real suprise. Right now he is 38th among SFs.
Jackson is clearly in front of him. Negative on defense as well (-0.59 DRPM) but a big plus on offense (+1.28 ORPM). Ranking 24th among SFs.

Kleber and Powell are clearly in front of KP. Powell has balanced numbers (+0.6 ORPM on offense, +0.7 DRPM on defense). Ranking 13th in the league.
Kleber shows himself as a defensive specialist. He is by far the worst offensive player in the rotation (-1.42 ORPM) but also by far the best defender (+2.11 DRPM). Overall that leaves him with rank 22 among PFs. 9th on defense.
KP ranks 36th among PFs. -0.76 ORPM / 0.94 DRPM. Not a big suprise to see him that low but I expected him to be closer to Maxi on defense and even worse on offense.

Boban ranks 18th among centers. The numbers agree with the eye test. He ranks 6th on offense (+1.08 ORPM) but is a slight negative on defense (-0.03).



Some non Mavs related numbers. AD only ranks 61th in the league and 143rd on defense. Trae Young is the 4th best in ORPM and the 3rd last in DRPM. Will Barton and Jayson Tatum lead the league in DRPM. James Harden is a +2 on defense.


Interpretation: Always take RPM with a grain of salt. Despite Engelmanns attempts to sell it it is not the holy grail of stats. As long as they don´t release the formula we simply don´t know how important prior data and boxscore numbers are.
Especially early in the year there is a lot of noise. Based on the rankings of previous years boxscore numbers and prior data have a bigger impact early in the season. That leads to strange results. Especially on defense. Good example would be Hardens DRPM. He always starts out as positive but the numbers regress as the season progresses.
Next important aspect is the best way to use those numbers. Does it make sense to compare a high usage superstar to a role player? No. Does it make sense to compare players with similar roles? Yes.
For example. Is DFS the worst defender on the roster? No. Is he worse than Curry? No.
DFS defends the best opposing scorer every night. That obviously influences his numbers. Switching their roles wouldn´t help the Mavs. Curry would have no chance to stop LeBron, Kawhi or Harden.
DRPM tells us that Curry is achieving better defensive results in his given role but that doesn´t make him the better defender.

[Thukydides]

Never been a fan of RPM, but this year it is nothing but a joke.

[DanSchwartzgan]

 

Noteable numbers.

Luka ranks 15th in the league. 5th among SFs. He has the worst DRPM among regular rotation players at -0.9. On offense he ranks 7th in the entire league.

Brunson really suprises with the 17th best overall RPM. 3rd among PGs. Nearly a +2 on defense.
Delon ranks 15th among PGs. JJB manages to sneak in right in front of him. Ranking 13th among PGs in a really small sample size.

Curry ranks 23th among SG. Only + 0.13 ORPM but +1.26 DRPM.
THJ is right behind him at 26. His numbers are the direct opposite of Currys. +1.26 ORPM, -0.05 DRPM.

DFS is the only regular Mavs rotation with a negative RPM (-0.23). Him being the 2nd worst defender (-0.85 DRPM) among them is a real suprise. Right now he is 38th among SFs.
Jackson is clearly in front of him. Negative on defense as well (-0.59 DRPM) but a big plus on offense (+1.28 ORPM). Ranking 24th among SFs.

Kleber and Powell are clearly in front of KP. Powell has balanced numbers (+0.6 ORPM on offense, +0.7 DRPM on defense). Ranking 13th in the league.
Kleber shows himself as a defensive specialist. He is by far the worst offensive player in the rotation (-1.42 ORPM) but also by far the best defender (+2.11 DRPM). Overall that leaves him with rank 22 among PFs. 9th on defense.
KP ranks 36th among PFs. -0.76 ORPM / 0.94 DRPM. Not a big suprise to see him that low but I expected him to be closer to Maxi on defense and even worse on offense.

Boban ranks 18th among centers. The numbers agree with the eye test. He ranks 6th on offense (+1.08 ORPM) but is a slight negative on defense (-0.03).


Interpretation: Always take RPM with a grain of salt.  

Brunson is an advanced stats darling this year.  The metrics at 538.com make him look good.  His PER is over 17 this year and his WS/48 is over .150.  The most interesting thing is he gets most of his minutes opposite Luka.  In fact, the only real knocks you can make against his numbers are that the sample size is small and he achieves those results mainly against bench players.

There is a narrative around here that this is a team of "6th/7th men" type players.  According to this, we may not have a "3rd man" type, but we have a team of 4th/5th man types, not 6th/7th.  One has to wonder what some of those offensive ratings would look like if Luka were replaced by an average PG.

Actually, we don't currently have a #2.  KP has never been good at O-RPM.  But, his D has typically been much better by this metric.  His prior two seasons had him ranked 9th and 14th best at PF, mostly on the strength of his defensive showing.  If the expectation is KP will soon return to being a 22 ppg all-star, take a look at his FG% and the number of shot attempts it took to get him to 22 points.  That has never been who he is.  In fact, his .500 TS% is by far the worst on the team (.067% worse than JJB, the next closest at .567%).  In fact, his career best TS% would be the worst on this team by .021%.  I'm not saying there isn't room for improvement with greater comfort and role definition.  But KP never was AD and probably never will be.

Whether we like it or not, the Mav's own analytics seem to have an element of this in it.  Matthews and Parsons were both RPM studs before arriving here.  However, TJ Warren has ranked 29th and 47th the last two years among SF's and is currently ranked 10th.  Sometimes just getting into the right situation helps.  He's having by far his best +/- year.  I wonder is something in the data showed that he could be a good player in the right situation.  His defensive numbers are good (I've never thought of him as a defender).  And he hits 40% of his 3's.  Imagine what this offense looks like with him in place of DFS.

Note Christian Wood's ranking as the 5th best PF.  Good for him.

[Kammrath]

DFS is the only regular Mavs rotation with a negative RPM (-0.23). Him being the 2nd worst defender (-0.85 DRPM) among them is a real suprise.

You know this honestly doesn't surprise me. I love DFS's heart, energy, and attitude and am very grateful he is on the team....BUT I do think he is a vastly overrated defender. Similar to Marion for a number of years.

[dirkfansince1998]

DFS is the only regular Mavs rotation with a negative RPM (-0.23). Him being the 2nd worst defender (-0.85 DRPM) among them is a real suprise.

You know this honestly doesn't surprise me. I love DFS's heart, energy, and attitude and am very grateful he is on the team....BUT I do think he is a vastly overrated defender. Similar to Marion for a number of years.

Agree. DFS outstanding ability is his offensive rebounding. Nearly 2 OREBs per game for a 6`7 wing is impressive.
He is the only athletic wing defender on the roster. Making him the only option against players like LeBron or Kawhi by default. A wing sized 3&D player remains among the biggest needs.

[ItsGoTime]

BUT I do think he is a vastly overrated defender.

Has he shown better stat-wise in seasons past? Does that have anything to do with going from mostly a bench player to being a starter? Defending the bench vs starters could well be a big part of the reason, unless he didn't show up well in seasons past.

[fifteenth]

DFS guarding LeBron vs Seth guarding Danny Green isn't an apples to apples comparison. If that's the data that's inside of the metric then you can't use that metric to compare Mavs defenders to each other. Really, you'd have to do a decent amount of work to figure out who you could compare any one player to.

This is always my comment about many of these formulas. Context is required to make sense of this stuff, and sometimes we mention that, but when all is said and done we use each metric as a linear player ranking system.

"yeah, good point, context, right...now back my player ranking point...'

[ItsGoTime]

Context is required to make sense of this stuff, and sometimes we mention that, but when all is said and done we use each metric as a linear player ranking system.

I agree. I think the main reason the statement is made of "small sample size" is because of this as well. At the end of the season, when every team has played every team roughly the same amount, or at least as much as they can, you are then able to say "this guy defended the best player 2-4 in every matchup" and then find those guys like that around the league that did the same and you get a bit of the truth to it all. I say a bit, because injuries and playing against LeBron/Kawhi/George happens 2, 3 and 4 times for some teams. All of those things skew the numbers.

[mavsluvr]

Always take RPM with a grain of salt.

This. These ESPN secret stats are strictly for entertainment value.

[dirkfansince1998]

Always take RPM with a grain of salt.

This. These ESPN secret stats are strictly for entertainment value.

They need to release the formula. We know that xRAPM uses prior informed data. We know that RPM adds boxscore data. But they refuse to publish the formula.
It seems like they changed something this year and it is either the boxscore influence or something about the prior information.
In recent years bigman had a big advantage when it comes to DRPM. It seems like that is gone this year. I would love to hear an explanation for the change.

In the end RAPM is a more reliable number because we know how it is created. Regression of lineups and on/off that can happen with or without prior information.

[mavsluvr]

Always take RPM with a grain of salt.

This. These ESPN secret stats are strictly for entertainment value.

They need to release the formula. We know that xRAPM uses prior informed data. We know that RPM adds boxscore data. But they refuse to publish the formula.
It seems like they changed something this year and it is either the boxscore influence or something about the prior information.
In recent years bigman had a big advantage when it comes to DRPM. It seems like that is gone this year. I would love to hear an explanation for the change.

In the end RAPM is a more reliable number because we know how it is created. Regression of lineups and on/off that can happen with or without prior information.

What are people relying on them for? 

I would be very surprised if any front office/coaching staff in the NBA pays any attention at all to any of these stats. I don't know that for sure, of course. 

I would say that the stats are pretty reliable entertainers however, which is why ESPN keeps putting them out there. A number of fans seem to enjoy parsing them, and a few may even take them seriously. But I don't know if the stats are being relied on for anything of consequence. Fantasy basketball, maybe?

I have no objection to people discussing them or anything, if they find it amusing. Just think that as far as using them to reach some meaningful analytical result, it's pretty much a case of GIGO.

[fifteenth]

What are people relying on them for? 

I would be very surprised if any front office/coaching staff in the NBA pays any attention at all to any of these stats. I don't know that for sure, of course. 

I would say that the stats are pretty reliable entertainers however, which is why ESPN keeps putting them out there. A number of fans seem to enjoy parsing them, and a few may even take them seriously. But I don't know if the stats are being relied on for anything of consequence. Fantasy basketball, maybe?

I have no objection to people discussing them or anything, if they find it amusing. Just think that as far as using them to reach some meaningful analytical result, it's pretty much a case of GIGO.

Are you talking about just what ESPN puts out? Are you talking about anything that makes it into the public that's not directly from a team's analytics department?

[ItsGoTime]

What are people relying on them for? 

I would be very surprised if any front office/coaching staff in the NBA pays any attention at all to any of these stats. I don't know that for sure, of course. 

To be honest, why are so many teams relying so heavily on them? Since stats have become so prevalent, has that changed the amount of bad contracts that go out? The amount of bad deals that get done? Has it improved how many good teams there are in the league (I get that there has to be winners and losers in a game, but if every game was close, that would say something to me)? I'm sure it helps some, I do wonder how much a guy like Jerry West relies on them though?

[Fuerza1]

Not surprised some are downplaying this highly advanced, analytic statistic since it only places Luka as a top 15 player rather than top 5. I'm guessing you won't see this stat in the HALLELUKA thread.

I also see " Mav fans" are also awfully quiet about Brunson's ELITE play based on this advanced statistic. Some used this very same stat to trash Brunson this past summer. Where you at?

[mavsluvr]

What are people relying on them for? 

I would be very surprised if any front office/coaching staff in the NBA pays any attention at all to any of these stats. I don't know that for sure, of course. 

I would say that the stats are pretty reliable entertainers however, which is why ESPN keeps putting them out there. A number of fans seem to enjoy parsing them, and a few may even take them seriously. But I don't know if the stats are being relied on for anything of consequence. Fantasy basketball, maybe?

I have no objection to people discussing them or anything, if they find it amusing. Just think that as far as using them to reach some meaningful analytical result, it's pretty much a case of GIGO.

Are you talking about just what ESPN puts out? Are you talking about anything that makes it into the public that's not directly from a team's analytics department?

I was referring to ESPN's collection of "proprietary statistics."

---

What are people relying on them for? 

I would be very surprised if any front office/coaching staff in the NBA pays any attention at all to any of these stats. I don't know that for sure, of course. 

To be honest, why are so many teams relying so heavily on them? Since stats have become so prevalent, has that changed the amount of bad contracts that go out? The amount of bad deals that get done? Has it improved how many good teams there are in the league (I get that there has to be winners and losers in a game, but if every game was close, that would say something to me)? I'm sure it helps some, I do wonder how much a guy like Jerry West relies on them though?

I am not saying that NBA teams don't rely on statistics. Of course, they do. I'm saying that I don't think any team rushes to check the ESPN proprietary stats every time they come out and adjust their roster construction and coaching decisions on them. In fact, I am saying I doubt that the analytics departments and coaching staffs even follow ESPN ratings at all. 

Now, as far as whether the various teams are using the right stats in the right way is a much more complex question, and is an area of continual study and adjustment.

[dirkfansince1998]

What are people relying on them for? 

I would be very surprised if any front office/coaching staff in the NBA pays any attention at all to any of these stats. I don't know that for sure, of course. 

I would say that the stats are pretty reliable entertainers however, which is why ESPN keeps putting them out there. A number of fans seem to enjoy parsing them, and a few may even take them seriously. But I don't know if the stats are being relied on for anything of consequence. Fantasy basketball, maybe?

I have no objection to people discussing them or anything, if they find it amusing. Just think that as far as using them to reach some meaningful analytical result, it's pretty much a case of GIGO.

Are you talking about just what ESPN puts out? Are you talking about anything that makes it into the public that's not directly from a team's analytics department?

I was referring to ESPN's collection of "proprietary statistics."

---

What are people relying on them for? 

I would be very surprised if any front office/coaching staff in the NBA pays any attention at all to any of these stats. I don't know that for sure, of course. 

To be honest, why are so many teams relying so heavily on them? Since stats have become so prevalent, has that changed the amount of bad contracts that go out? The amount of bad deals that get done? Has it improved how many good teams there are in the league (I get that there has to be winners and losers in a game, but if every game was close, that would say something to me)? I'm sure it helps some, I do wonder how much a guy like Jerry West relies on them though?

I am not saying that NBA teams don't rely on statistics. Of course, they do. I'm saying that I don't think any team rushes to check the ESPN proprietary stats every time they come out and adjust their roster construction and coaching decisions on them. In fact, I am saying I doubt that the analytics departments and coaching staffs even follow ESPN ratings at all. 

Now, as far as whether the various teams are using the right stats in the right way is a much more complex question, and is an area of continual study and adjustment.

There certainly has to be some correlation. The creator of RPM worked for NBA teams before he joined ESPN. Same for some of the guys that are involved in the RAPTOR or DRAYMOND rating. Engelmann is actually working for the Mavs right now.

[fifteenth]

Not surprised some are downplaying this highly advanced, analytic statistic since it only places Luka as a top 15 player rather than top 5. I'm guessing you won't see this stat in the HALLELUKA thread.

I also see " Mav fans" are also awfully quiet about Brunson's ELITE play based on this advanced statistic. Some used this very same stat to trash Brunson this past summer. Where you at?

There are lots of useful numerical ways to parse basketball exist. It'd be fun to know exactly what teams are using. 

What makes ESPN's offering "highly" advanced? I had questions about their versions of the attempts to apply lineup data to individuals before Luka was in the league. 

You're not the only Brunson fan around here. Most seem to like him quite a bit.

[mavsluvr]

Not surprised some are downplaying this highly advanced, analytic statistic since it only places Luka as a top 15 player rather than top 5. I'm guessing you won't see this stat in the HALLELUKA thread.

I also see " Mav fans" are also awfully quiet about Brunson's ELITE play based on this advanced statistic. Some used this very same stat to trash Brunson this past summer. Where you at?

It is not possible to evaluate this stat, since we don't know how it is put together. However, some of the counterintuitive results leave room for doubt. 

No person relying on statistics for serious purposes should ever evaluate a stat based on whether it supports some result he/she is emotionally attached to. The usefulness of statistics is to reach a better understanding of the truth -- to confirm/deny an experience-based analysis, or identify trends that may justify a closer look, for example. 

Of course, sports fans are free to use stats any way they want, and if they want to judge statistics based on whether they make particular players look good or bad, I guess they can have at it and enjoy themselves!

[fifteenth]

I am saying I doubt that the analytics departments and coaching staffs even follow ESPN ratings at all. 

yes, for sure...no way NBA teams use ESPN's stuff

[dirkfansince1998]

https://twitter.com/JerryEngelmann/statu...3865182208