ESPN RAPM finally posted

[juju151111]

The best stat to use along with eye test http://espn.go.com/nba/statistics/rpm/_/sort/RPM


Interesting notes. Curry is a absurd +9 offensively, but is playing less defense this year . He has been getting lit up by PG this year.

Duncan still one of the best defenders.

[fpliii]

That's RPM (or xRAPM), which means it has a box score prior (which also takes into account other variables like height and age).

J.E. hasn't released pure RAPM yet but on his twitter (https://twitter.com/JerryEngelmann) gave out:


Jeremias Engelmann ‏@JerryEngelmann Nov 23
Single year RAPM rankings (in order):
Draymond Green, Iguodala, Curry, Ginobili, LeBron James, Marcus Morris, Dellavedova, Lavoy Allen

You can probably expect the first full pure RAPM rankings around all-star break.

[Terahite]

The best stat to use along with eye test http://espn.go.com/nba/statistics/rpm/_/sort/RPM


Interesting notes. Curry is a absurd +9 offensively, but is playing less defense this year . He has been getting lit up by PG this year.

How could Curry possibly be playing less defense than last year? Might have to click this link...

[TaLvsCuaL]

Jae Crowder, Pachulia and Marvin Williams :wtf:

[warriorfan]

King Kurry better defender than WestShit

Deal with it

[juju151111]

How could he be possibly be playing less defense than last year? Might have to click this link...
Les defense then last year? CURRY was playing great defense last year and was pointed out on many articles. This year everyone is having a great game against him and he missong some switches. He seemsto concentrate on offense this season.

[kuniva_dAMiGhTy]

Hey fpliii,

Any idea why JE doesn't just use RAPM for ESPN (supposedly less noise and more accurate) and trash RPM?

The results for RPM and RAPM are somewhat inconsistent, so its kind of confusing as to who legitimately has greater impact.

[dhsilv]

That's RPM (or xRAPM), which means it has a box score prior (which also takes into account other variables like height and age).

J.E. hasn't released pure RAPM yet but on his twitter (https://twitter.com/JerryEngelmann) gave out:



You can probably expect the first full pure RAPM rankings around all-star break.

I'm struggling to buy Manu being that high. It looks as though xRAMP is doing what it's supposed here. I would like to see a prior influenced xRAMP at this stage of the season though.

[no pun intended]

Mudiay and Okafor are second and seventh to last, respectively lmao

[FKAri]

Les defense then last year? CURRY was playing great defense last year and was pointed out on many articles. This year everyone is having a great game against him and he missong some switches. He seemsto concentrate on offense this season.

It might be sample size or it might honestly just be not trying as much. Are you going to go hard on D when you're up 20 or when it's a tie game?

[Achilleas]

is the right;
monta 3 best pg on def
and calderon 12 on def
wall 25
paul 40

[juju151111]

That's RPM (or xRAPM), which means it has a box score prior (which also takes into account other variables like height and age).

J.E. hasn't released pure RAPM yet but on his twitter (https://twitter.com/JerryEngelmann) gave out:



You can probably expect the first full pure RAPM rankings around all-star break.
Pure Rapm is better for single season right. And where does he post the pure rapm

[fpliii]

Hey fpliii,

Any idea why JE doesn't just use RAPM (supposedly less noise and more accurate) and trash RPM?

The results for RPM and RAPM look a bit different, so its kind of confusing as to who legitimately has greater impact.
xRAPM in a previous year is superior at predicting xRAPM the next year than RAPM in a previous year is at predicting RAPM in the next year (since the blend includes elements of the box score that are consistent). So in terms of statistical 'reliability' (which is the consistency of a measure), xRAPM would be superior.

In terms of statistical validity (degree of correctness), RAPM is superior. It has superior explanatory value, since it is a mathematical isolation of an individual's impact on scoring margin.

http://www.documentingexcellence.com/stat_tool/images/targets4.jpg

J.E. isn't interested in answering which player is better between X and Y most accurately, his goal is to put together the best predictive model.

[Giaodollo]

Kobe bryant 381th best player in the league.

[HenryGarfunkle]

Kobe bryant 381th best player in the league.
too high

these advanced stats are bullshit

[fpliii]

I'm struggling to buy Manu being that high. It looks as though xRAMP is doing what it's supposed here. I would like to see a prior influenced xRAMP at this stage of the season though.
Well he uses a prior in xRAPM, which is comprised of a SPM (statistical plus minus...something like his own personal BPM as listed on basketball-reference). Though the numbers on ESPN only use on year of lineup data.

Last year I believe he posted multi-year RPM/xRAPM on Insider.

Pure Rapm is better for single season right. And where does he post the pure rapm
Check apbr.org/metrics/ around the all-star break, or his twitter (linked above).

[ISHGoat]

It might be sample size or it might honestly just be not trying as much. Are you going to go hard on D when you're up 20 or when it's a tie game?

But are his opponents likely to go hard and play smart offense when they are in these blow out conditions? It evens out IMO.

[JtotheIzzo]

So if you replace Andrew Wiggins with Austin Rivers the Wolves are a better team?

Advanced stats are so retarded, no wonder Houston is such shit.

[Blue&Orange]

plus minus is the most inaccurate and irrelevant stat imaginable, any "advanced" form of the former will always be well, inaccurate and irrelevant.

seriously can't think of a more worthless stat.

[fpliii]

plus minus is the most inaccurate and irrelevant stat imaginable, any "advanced" form of the former will always be well, inaccurate and irrelevant.

seriously can't think of a more worthless stat.
Don't compare players playing different roles to one another.

Ask yourself...why does almost every front office have analytics teams that heavily utilize it in personnel or lineup decisions?

Or, why do plus-minus based models lap the field when predicting wins?

[Blue&Orange]

Ask yourself...why does almost every front office have analytics teams that heavily utilize it in personnel or lineup decisions?

Or, why do plus-minus based models lap the field when predicting wins?
Those front offices must be really successful. :lol

i bet the sixers are one.

[fpliii]

Those front offices must be really successful. :lol

i bet the sixers are one.
Actually, it's every front office (http://espn.go.com/espn/feature/story/_/id/12331388/the-great-analytics-rankings#nba-tor), except for the Lakers, Knicks, and Nets.


All IN
Dallas Mavericks
Houston Rockets
Philadelphia 76ers
San Antonio Spurs
BELIEVERS
Atlanta Hawks
Boston Celtics
Cleveland Cavaliers
Detroit Pistons
Golden State Warriors
Memphis Grizzlies
Oklahoma City Thunder
Portland Trail Blazers
ONE FOOT IN
Charlotte Hornets
Indiana Pacers
Miami Heat
Milwaukee Bucks
Orlando Magic
Phoenix Suns
Sacramento Kings
Toronto Raptors
Utah Jazz
SKEPTICS
Chicago Bulls
Denver Nuggets
Los Angeles Clippers
Minnesota Timberwolves
New Orleans Pelicans
Washington Wizards
NONBELIEVERS
Brooklyn Nets
Los Angeles Lakers
New York Knicks

[Blue&Orange]

Actually, it's every front office (http://espn.go.com/espn/feature/story/_/id/12331388/the-great-analytics-rankings#nba-tor), except for the Lakers, Knicks, and Nets.
i clicked the link choose NBA searched for RAPM was nowhere to be found, bu thanks for the link.

[fpliii]

i clicked the link choose NBA searched for RAPM was nowhere to be found, bu thanks for the link.
RAPM is regularized adjusted plus minus. The link just talks about analytics integration in the NBA. If you're looking for more information check out here:

apbr.org/metrics/

In particular, one poster is looking to do put together a history of RAPM (http://www.apbr.org/metrics/viewtopic.php?f=2&t=8995).

Anyhow though, just for starters, the Mavs have used APM (predecessor of RAPM using slightly different techniques) for well over a decade. Jeff Sagarin and Wayne Winston were the developers.

[juju151111]

RAPM is regularized adjusted plus minus. The link just talks about analytics integration in the NBA. If you're looking for more information check out here:

apbr.org/metrics/

In particular, one poster is looking to do put together a history of RAPM (http://www.apbr.org/metrics/viewtopic.php?f=2&t=8995).

Anyhow though, just for starters, the Mavs have used APM (predecessor of RAPM using slightly different techniques) for well over a decade. Jeff Sagarin and Wayne Winston were the developers.
So xRapm isn't the best for individual impact. Say you want to compare players with similar roles like Durant and LJ. We would have to wait till the pure Rapm comes out

[fpliii]

So xRapm isn't the best for individual impact. Say you want to compare players with similar roles like Durant and LJ. We would have to wait till the pure Rapm comes out
I guess, but I still wouldn't use it to compare them directly. It's just one extra tool to have (albeit a very good one). They have similar roles (so it has a good deal of use), but they're not identical.

xRAPM/RPM isn't bad though, I don't have any problem with it. It's an incredible predictive metric. If someone wants to use it for comparisons it's not the worst idea. But I just think it's important for someone to understand what exactly it is, and what it isn't.

[juju151111]

I guess, but I still wouldn't use it to compare them directly. It's just one extra tool to have (albeit a very good one). They have similar roles (so it has a good deal of use), but they're not identical.

xRAPM/RPM isn't bad though, I don't have any problem with it. It's an incredible predictive metric. If someone wants to use it for comparisons it's not the worst idea. But I just think it's important for someone to understand what exactly it is, and what it isn't.
Of course you have to use the eye test too and context.

[fpliii]

Of course you have to use the eye test too and context.
Agree 100%. Watching games, and looking at both data and scouting reports is the best way to evaluate players and teams.

[Shih508]

Kobe is ranked at 381 out of 409 players. He overrated himself thinking he's still top 200th player in the league.