A German magazine interview with Jeremias Engelmann, creator of RPM

[Slava]

Engelmann is kind of the anti-Barkley in the sense that he leans a lot more extremely to the analytics side. This is an interview he did recently with a German magazine and there are a few hot-takes.

Mind, most of this is from google translate.

On his initial days with the Suns front office:

SPOX : Let's go back to Phoenix. How were you there as a "double outsider" - as a European, who does not come from professional basketball, so is not a "basketball guy"?

Engelmann : (laughs) As I was recorded at the Suns explains why I quit after three months. My direct boss, who has now landed at Chelsea, was great, the daily work was really fun. But as with many teams, the Suns' Analytics department was launched by the owners, while the front office felt more threatened by them. After all, it had the sole power over the scouting and the squad.

My player rating systems were not really good. Also for trade proposals I received much negative feedback. We were simply not at all on a line. I had the feeling that some of the management had no other view of basketball than the average Joe Blow from the road: Carmelo is super, Kyrie is super, and who is Amir Johnson at all?

He was asked about different players ranks as per RPM:

Jokic:

SPOX : And Jokic? He was the number one after the best center and the sixth best player in the league.

Engelmann : At least in the top 20 I would certainly see him. With him, the nuggets were offensive at the level of the Warriors and Denver has now really not the greatest offensive players. He makes the points for his teammates very easy and is also an incredibly efficient pitcher himself.

On the defensive ... (laughs) I admit that he may be over-rated. In fact, the Nuggets have better defended when he was on the field - though still relatively terrible. He is at least a good rebounder and brings many steals. RPM also has a tendency to evaluate great players better.

Schröder:

SPOX : I have one last player on offer: Dennis Schröder. Provocatively asked: Is he one of the worst point guards in the league?

Engelmann : Unfortunately, there are very few arguments that are currently being discussed. A big problem is its high turnover rate, especially the many so-called live ball turnovers, which lead to opposing fast-attack. They pull it statistically extremely down. But otherwise the numbers do not look good.

A coaching evaluation metric and Brad Stevens, David Blatt, Byron Scott:

Maybe another hot take to the Celtics: I also have a metric that tries to evaluate coaches. After that, Brad Stevens is only a fairly average coach.

SPOX : Oh ... Can you briefly explain this metric?

Engelmann : She puts the coach almost as a sixth player on the field to calculate his influence on the offensive and defensive. The results are broadly in line with what one would expect. The Van Gundys are on top, as does Quin Snyder from Utah Jazz, Phil Jackson or Gregg Popovich.

The former Lakers coach Byron Scott, on the other hand, is terrible. A total of about 140 coaches are involved, and the worst 50 currently has no job at all. David Blatt is also very far down in the ranking.

some Jokic vs Embiid and Joel Embiid in general:

SPOX : Is Nikola Jokic a better defender than Joel Embiid? His defensive Real Plus minus was slightly higher in the last season.

Engelmann : That's because Embiid played only a few minutes. As a result, the number is much less precise than in the case of players who have played a lot. A "regression to the middle" is made for players with a small number of minnows, as they are expected not to "hold" extreme stats. Had Embiid brought the same achievements and stats over 60 or 70 games, RPM would see him as one of the best defenders of the league.

SPOX : The Celtics could be disappointed. Are there any other candidates?

Engelmann : The health of Joel Embiid naturally worries me at the Philadelphia 76ers. If someone kinks and tapes torn, I do not worry in the long term. Even with a cruciate ligament crack, I do not have much more concern today that people are coming back.

But stress factors like Embiid in the foot and also in the back can be a sign that the body does not cope with the sport itself and the stress. Especially for extremely big players. I would naturally wish him to be completely fit, but I am not very confident.

Link to full interview

[jimmy_smith]

Interesting about coaching evaluation metric.
Van Gundys, Quin Snyder, Phil Jackson and Gregg Popovich are at the top.
Brad Stevens is average is a surprise for me but no surprise for Byron Scott at the bottom.

[the sea duck]

What kind of track record does Engelmann have with these ratings?

[jinxed]

I'd like to see the coaches ranking become public again. He used to have it on his old site. Thibs was rated very highly as well at the time. It seemed like defense is where coaches brought the most juice.

[jinxed]

What kind of track record does Engelmann have with these ratings?

As he states in the interview, it used to be that you could use RPM numbers to bet against Vegas and make a pretty penny. Vegas has gotten a lot smarter now. Still, RPM is probably the best available stat in terms of predictions.

[caliban]

What kind of track record does Engelmann have with these ratings?

As he states in the interview, it used to be that you could use RPM numbers to bet against Vegas and make a pretty penny. Vegas has gotten a lot smarter now. Still, RPM is probably the best available stat in terms of predictions.

If one knows where to look it hasn't been the best at predictions. See the apbr team prediction contest the last years for evidence. Have you tested it yourself? It has a very good track record, not the best which is a bit strange considering the espn time and money backing the projection.

[jazzfan1971]

Interesting about coaching evaluation metric.
Van Gundys, Quin Snyder, Phil Jackson and Gregg Popovich are at the top.
Brad Stevens is average is a surprise for me but no surprise for Byron Scott at the bottom.

This is really interesting. I have to question any list where Popovich isn't #1 though. Ranking coaches is sooooo difficult IMHO. It's nice that there is an objective standard out there, even if it's not close to perfect, it's probably miles better than just going with the groupthink of which coach is good.

Would love to hear more about this. Would have liked to read the whole article by my German is Nicht Gut.

[12footrim]

Interesting about coaching evaluation metric.

Brad Stevens is average is a surprise for me

I'm not. This is the same guy that had IT coming off the bench until Smart was injured. It was pretty obvious what he could be if he was given the opportunity. I went 10 rounds with Boston fans on their boards before who had drank koolaid and defended that crap that he couldn't be a full time player etc etc and last year he was top 5 in the MVP voting.

[76ciology]

RPM and Jokic mystery solved

[clyde21]

I'd like to real more about that coaching metric. I like Engelmann's work.

[OkcSinceSGA]

Interesting thread, thanks. I'm wondering how Doc fairs on the coaching rank lol. I'd imagine 15-20 despite top 5-10 hype at times.

[JonFromVA]

We were simply not at all on a line. I had the feeling that some of the management had no other view of basketball than the average Joe Blow from the road: Carmelo is super, Kyrie is super, and who is Amir Johnson at all?

Yeah, there seems to be a presumption that if you're around the game that you should be capable of making sound decisions. This tends to re-enforce group-think. Funny thing is decision making can be evaluated well in advance of application, and that's probably the best part of the analytics movement - in that it can be used to make predictions and judged.

An analyst should be able to out-perform a model, but it seems that's rarely tested (except when they're given a job).

It would be interesting after a few years of ESPN panel predictions to go back and see if any members of the panels actually outperformed the rest with any consistency.

[jinxed]

What kind of track record does Engelmann have with these ratings?

As he states in the interview, it used to be that you could use RPM numbers to bet against Vegas and make a pretty penny. Vegas has gotten a lot smarter now. Still, RPM is probably the best available stat in terms of predictions.

If one knows where to look it hasn't been the best at predictions. See the apbr team prediction contest the last years for evidence. Have you tested it yourself? It has a very good track record, not the best which is a bit strange considering the espn time and money backing the projection.

Most entrants in the contest did some kind of RPM/BPM blend, similar to RPM or just used RPM. Most were close enough in the final total that minutes distribution/chance could have decided the winner. I know PT-PM RAPM blend has won in the past.

If you know of a single public metric that is better than RPM let me know!

[dYYYh.k²]

surprisingly, they didnt ask him any questions regarding nowitzki.!

[LakersLegacy]

surprisingly, they didnt ask him any questions regarding nowitzki.!

Or David Hasslehof.

I'd like to see his complete coaching rankings.

[Koreandoo]

Here some other answers (with translator)

SPOX: How did the idea arise to your own metric, the real plus minus, then?
Engelmann: Already before the Suns have asked me, I had tried to offer German basketball and soccer teams the RAPM.I was then invited to an interview by three national league teams, among other things Borussia Mönchengladbach.
One had a rather cool internal assessment {weighting} system there: Scouts carried subjective "Stats" for all possible youth players for dribbling ability, points from 1 to 10 allocated one.
The thought that it would make sense not to leave such information completely unused as the case is at the RAPM, and to feed her in the algorithm instead anyway befitted me there.

SPOX: Which fundamental advantage do statistics have opposite the Eye test, then?
Can't I recognize more yet if I simply look at the games?
Engelmann: On the one hand, it is primarily hardly possible in the NBA of course to see every game actually.
There are 1230 games per season.
Even for hardcore fans it is probably hard to look at at all the 82 games of the team of one's own.
If somebody, however, would sit down every day, looking at games for eight hours and building his own, subjective assessment system ... I could imagine, that this would be perhaps better I have the feeling as RPM. Others that the human perception is often distorted.
People for example of many and a player comes off particularly there, they are aware perhaps also in the reality as better of him when he is actually if NBA2K gamble.
It is spectacularly similar dunken which one sees through this once in a while, too this one at players which often happen in highlights.
Stats often provide a more objective and also more exact assessment there.
In addition, one is not aware of also many things so exactly at all in the defensive.
One probably should look with half speed at games there to see at every player at every time what he does correctly or wrongly.

SPOX: The perception parts at which players particularly far, then, who gets special in public {attention} about or undervalued?

Engelmann: Kobe Bryant was teams in one of the all defensive twelve times.
He has often only then put much energy into the Defense if the Lakers have played on television or he had a particularly known opponent.
Altogether, have defended the Lakers without him in the field considerably better within most years, however.
This cannot be somehow. There is a great discrepancy there because you often simply look only on the ball.
Things are pursue after a ball loss then neglected like Help Defense or this.
A current example would be Avery Bradley. Also he plays a good defense at the ball - this looks also rather stressful for the ball leading player because Bradley "sticks" to his opponent so.But he just isn't the biggest.
So opponents can often simply throw about him and of course this is a problem also for his Help Defense.
But who pays attention already to the Help Defense of a Shooting Guards if he looks at a game with friends?

SPOX: In the offensive Eye test and Stats usually are apart less far. It is easy to observe more easily who makes a large contribution there and statistics can measure this considerably more exactly also.
Nevertheless are there examples of a similar discrepancy also in the Offense between Stats and the public perception?
Engelmann: I am not as safe as in the defense here.
But I think that volume Scorer which make simply many points per game are overvalued by many fans.
Playing players at all events opposite just less which although establish good statistics per minute, however.
For me Amir Johnson who always had very good Stats per minute was always the prime example there.
But the numbers per game still seem for many to be decisive although this doesn't play a large role in my world. Also Cody Zeller or Otto Porter is players which are very efficient but undervalued still.
Porter was last season e.g. of one of the most efficient throwers of the league.
Such players are the simply much perhaps nevertheless underestimates points out good hidings in comparison with players like Carmelo Anthony, Kyrie Irving or DeMar DeRozan. Kobe Bryant is an example again also there.
His points still felt often complicated to this, because he has made the life difficult for himself himself.
Much isolation {confinement}, credit fakes, throws against three defenders ... This looks cool if it goes in.
But it is the question of course whether this was always necessary.

[TankCommander17]

Here some other answers (with translator)

SPOX: How did the idea arise to your own metric, the real plus minus, then?
Engelmann: Already before the Suns have asked me, I had tried to offer German basketball and soccer teams the RAPM.I was then invited to an interview by three national league teams, among other things Borussia Mönchengladbach.
One had a rather cool internal assessment {weighting} system there: Scouts carried subjective "Stats" for all possible youth players for dribbling ability, points from 1 to 10 allocated one.
The thought that it would make sense not to leave such information completely unused as the case is at the RAPM, and to feed her in the algorithm instead anyway befitted me there.

SPOX: Which fundamental advantage do statistics have opposite the Eye test, then?
Can't I recognize more yet if I simply look at the games?
Engelmann: On the one hand, it is primarily hardly possible in the NBA of course to see every game actually.
There are 1230 games per season.
Even for hardcore fans it is probably hard to look at at all the 82 games of the team of one's own.
If somebody, however, would sit down every day, looking at games for eight hours and building his own, subjective assessment system ... I could imagine, that this would be perhaps better I have the feeling as RPM. Others that the human perception is often distorted.
People for example of many and a player comes off particularly there, they are aware perhaps also in the reality as better of him when he is actually if NBA2K gamble.
It is spectacularly similar dunken which one sees through this once in a while, too this one at players which often happen in highlights.
Stats often provide a more objective and also more exact assessment there.
In addition, one is not aware of also many things so exactly at all in the defensive.
One probably should look with half speed at games there to see at every player at every time what he does correctly or wrongly.

SPOX: The perception parts at which players particularly far, then, who gets special in public {attention} about or undervalued?

Engelmann: Kobe Bryant was teams in one of the all defensive twelve times.
He has often only then put much energy into the Defense if the Lakers have played on television or he had a particularly known opponent.
Altogether, have defended the Lakers without him in the field considerably better within most years, however.
This cannot be somehow. There is a great discrepancy there because you often simply look only on the ball.
Things are pursue after a ball loss then neglected like Help Defense or this.
A current example would be Avery Bradley. Also he plays a good defense at the ball - this looks also rather stressful for the ball leading player because Bradley "sticks" to his opponent so.But he just isn't the biggest.
So opponents can often simply throw about him and of course this is a problem also for his Help Defense.
But who pays attention already to the Help Defense of a Shooting Guards if he looks at a game with friends?

SPOX: In the offensive Eye test and Stats usually are apart less far. It is easy to observe more easily who makes a large contribution there and statistics can measure this considerably more exactly also.
Nevertheless are there examples of a similar discrepancy also in the Offense between Stats and the public perception?
Engelmann: I am not as safe as in the defense here.
But I think that volume Scorer which make simply many points per game are overvalued by many fans.
Playing players at all events opposite just less which although establish good statistics per minute, however.
For me Amir Johnson who always had very good Stats per minute was always the prime example there.
But the numbers per game still seem for many to be decisive although this doesn't play a large role in my world. Also Cody Zeller or Otto Porter is players which are very efficient but undervalued still.
Porter was last season e.g. of one of the most efficient throwers of the league.
Such players are the simply much perhaps nevertheless underestimates points out good hidings in comparison with players like Carmelo Anthony, Kyrie Irving or DeMar DeRozan. Kobe Bryant is an example again also there.
His points still felt often complicated to this, because he has made the life difficult for himself himself.
Much isolation {confinement}, credit fakes, throws against three defenders ... This looks cool if it goes in.
But it is the question of course whether this was always necessary.

In that last quote did he try to say that Kobe was a prime example of an inefficient player and was overvalued?

[dhsilv2]

Here some other answers (with translator)

SPOX: How did the idea arise to your own metric, the real plus minus, then?
Engelmann: Already before the Suns have asked me, I had tried to offer German basketball and soccer teams the RAPM.I was then invited to an interview by three national league teams, among other things Borussia Mönchengladbach.
One had a rather cool internal assessment {weighting} system there: Scouts carried subjective "Stats" for all possible youth players for dribbling ability, points from 1 to 10 allocated one.
The thought that it would make sense not to leave such information completely unused as the case is at the RAPM, and to feed her in the algorithm instead anyway befitted me there.

SPOX: Which fundamental advantage do statistics have opposite the Eye test, then?
Can't I recognize more yet if I simply look at the games?
Engelmann: On the one hand, it is primarily hardly possible in the NBA of course to see every game actually.
There are 1230 games per season.
Even for hardcore fans it is probably hard to look at at all the 82 games of the team of one's own.
If somebody, however, would sit down every day, looking at games for eight hours and building his own, subjective assessment system ... I could imagine, that this would be perhaps better I have the feeling as RPM. Others that the human perception is often distorted.
People for example of many and a player comes off particularly there, they are aware perhaps also in the reality as better of him when he is actually if NBA2K gamble.
It is spectacularly similar dunken which one sees through this once in a while, too this one at players which often happen in highlights.
Stats often provide a more objective and also more exact assessment there.
In addition, one is not aware of also many things so exactly at all in the defensive.
One probably should look with half speed at games there to see at every player at every time what he does correctly or wrongly.

SPOX: The perception parts at which players particularly far, then, who gets special in public {attention} about or undervalued?

Engelmann: Kobe Bryant was teams in one of the all defensive twelve times.
He has often only then put much energy into the Defense if the Lakers have played on television or he had a particularly known opponent.
Altogether, have defended the Lakers without him in the field considerably better within most years, however.
This cannot be somehow. There is a great discrepancy there because you often simply look only on the ball.
Things are pursue after a ball loss then neglected like Help Defense or this.
A current example would be Avery Bradley. Also he plays a good defense at the ball - this looks also rather stressful for the ball leading player because Bradley "sticks" to his opponent so.But he just isn't the biggest.
So opponents can often simply throw about him and of course this is a problem also for his Help Defense.
But who pays attention already to the Help Defense of a Shooting Guards if he looks at a game with friends?

SPOX: In the offensive Eye test and Stats usually are apart less far. It is easy to observe more easily who makes a large contribution there and statistics can measure this considerably more exactly also.
Nevertheless are there examples of a similar discrepancy also in the Offense between Stats and the public perception?
Engelmann: I am not as safe as in the defense here.
But I think that volume Scorer which make simply many points per game are overvalued by many fans.
Playing players at all events opposite just less which although establish good statistics per minute, however.
For me Amir Johnson who always had very good Stats per minute was always the prime example there.
But the numbers per game still seem for many to be decisive although this doesn't play a large role in my world. Also Cody Zeller or Otto Porter is players which are very efficient but undervalued still.
Porter was last season e.g. of one of the most efficient throwers of the league.
Such players are the simply much perhaps nevertheless underestimates points out good hidings in comparison with players like Carmelo Anthony, Kyrie Irving or DeMar DeRozan. Kobe Bryant is an example again also there.
His points still felt often complicated to this, because he has made the life difficult for himself himself.
Much isolation {confinement}, credit fakes, throws against three defenders ... This looks cool if it goes in.
But it is the question of course whether this was always necessary.

In that last quote did he try to say that Kobe was a prime example of an inefficient player and was overvalued?

That's how I read it. Basically saying Kobe made the game harder for himself than it should have been by playing so much iso/post play. I think at least his last few years in the league, he really did do way too much iso and it was very much a problem.

[Manute Lol]

Oh man...this guy really believes his own hype (I can read German). Among his other hot takes, he also said that DeRozan was a league average player last year and no better than Jason Terry, and that he doesn't think the Wolves or Sixers will crack 55 wins in the next four years.

He also has a big sad face for Boston because they lost their three best players by RPM: Crowder, IT and Amir Johnson, which he doesn't seem to realize is more of an embarrassment to RPM than to the Celtics, but whatever.

[the sea duck]

Oh man...this guy really believes his own hype (I can read German). Among his other hot takes, he also said that DeRozan was a league average player last year and no better than Jason Terry, and that he doesn't think the Wolves or Sixers will crack 55 wins in the next four years.

He also has a big sad face for Boston because they lost their three best players by RPM: Crowder, IT and Amir Johnson, which he doesn't seem to realize is more of an embarrassment to RPM than to the Celtics, but whatever.

It does seem that he kind of makes the assumption that his rating is accurate and therefore his less popular opinions should be celebrated. When in reality, I'm left questioning whether the ratings are really all that good because of what they say. That's why I asked about his track record.

I came up with a pretty BS rating system myself a few years ago and tracked players for partial seasons. They were mostly spot on, but sometimes a weird player here or there would be seemingly out of place. I assumed my ratings were garbage (because I knew they were), but I guess I could have just acted like I knew something no one else did.