2017 RAPM/RPM/etc. Thread 2017 S-Y NPI RAPM Added (6/26)

[Arman_tanzarian]

Does JE have a link to RAPM on a year-by-year basis? I've seen a bunch online but can't tell which one are his.

[dontcalltimeout]

Did we ever get a multiyear RAPM for last season?

[SideshowBob]

Did we ever get a multiyear RAPM for last season?

Nope. Followed up with JE several times, to no avail.

[dontcalltimeout]

Did we ever get a multiyear RAPM for last season?

Nope. Followed up with JE several times, to no avail.

[mysticbbb]

If anyone is interested here is an excel file with RAPM since 1997 with added box score stats, positions played, teams and age (source too, but mostly it is what Engelamann published), so it is quite useful for any kind of research:

https://docs.google.com/spreadsheets/d/1NxcRTaNfTDuD288wI16myUwHUM0dDjkb5UebrI412pY/edit?usp=sharing

[SideshowBob]

If anyone is interested here is an excel file with RAPM since 1997 with added box score stats, positions played, teams and age (source too, but mostly it is what Engelamann published), so it is quite useful for any kind of research:

https://docs.google.com/spreadsheets/d/1NxcRTaNfTDuD288wI16myUwHUM0dDjkb5UebrI412pY/edit?usp=sharing

Have you been running your own numbers the last few seasons?

[mysticbbb]

If anyone is interested here is an excel file with RAPM since 1997 with added box score stats, positions played, teams and age (source too, but mostly it is what Engelamann published), so it is quite useful for any kind of research:

https://docs.google.com/spreadsheets/d/1NxcRTaNfTDuD288wI16myUwHUM0dDjkb5UebrI412pY/edit?usp=sharing

Have you been running your own numbers the last few seasons?

I'm not, and I'm not the original mysticbb (2 bs, not 3 ). My nickname was created to honor him.

[SideshowBob]

If anyone is interested here is an excel file with RAPM since 1997 with added box score stats, positions played, teams and age (source too, but mostly it is what Engelamann published), so it is quite useful for any kind of research:

https://docs.google.com/spreadsheets/d/1NxcRTaNfTDuD288wI16myUwHUM0dDjkb5UebrI412pY/edit?usp=sharing

Have you been running your own numbers the last few seasons?

I'm not, and I'm not the original mysticbb (2 bs, not 3 ). My nickname was created to honor him.

My bad

[Ryoga Hibiki]

I have three comments:
- defensive contribution is enormously dependent to context, more so than offence. The right system, and the ability to implement it, can make a huge difference for guys who are not Gobert like rim protectors
- I see that many of the 3&D guys are actually surprisingly positive contributors on offence, often more than on defense. Are we still underestimating the impact that a good 3 point threat who doesn't do stupid things can have?
- to all people rejecting a stat because of the outliers...
1) you must be sure you understand what the stat is measuring, what's the scope of it. If you're using it for things it has not been designed for it's your problem
2) don't focus that much on the number itself, embracing or rejecting it, but try to understand the context, why you get such results! it can be statistical noise, but you might find something surprising digging further. Especially when you don't agree initially with the results!
3) all stats have outliers. the top10 in PPG are not the top10 best scorers. nor the top10 apg the best passers

[Pillendreher]

This is new, right?

https://94feetreport.com/an-introduction-to-four-factors-rating-3e9ee475ceed

[Dr Spaceman]

This is new, right?

https://94feetreport.com/an-introduction-to-four-factors-rating-3e9ee475ceed

It used to exist in the form of “FFAPM” on GotBuckets but it was to rate the effect of individual players. As far as I’m aware, yes this is new, at least as far as applying to whole teams.

[Dr Spaceman]

I have three comments:
- defensive contribution is enormously dependent to context, more so than offence. The right system, and the ability to implement it, can make a huge difference for guys who are not Gobert like rim protectors
- I see that many of the 3&D guys are actually surprisingly positive contributors on offence, often more than on defense. Are we still underestimating the impact that a good 3 point threat who doesn't do stupid things can have?
- to all people rejecting a stat because of the outliers...
1) you must be sure you understand what the stat is measuring, what's the scope of it. If you're using it for things it has not designed for it's your problem
2) don't focus that much on the number itself, embracing or rejecting it, but try to understand the context, [b]why you get such results! can be statistical noise, but you might find something surprising digging further. Especially when you don't agree initially with the results!
3) all stats have outliers. the top10 in PPG are not the top10 best scorers. nor the top10 apg the best passers[/b]

Im late to respond to this but I agree with everything above and would add even one more point:

To me, the outliers are a benefit. The “advanced” stats, that we think of, were designed with specific results in mind. EG. PER was designed to have Michael Jordan be the best. RAPM is notable in that it doesn’t do that. What you’re getting is not necessarily what you’d expect, and to me that gives validity to the data. It doesn’t mean you take the results as gospel, but in large part I think what people don’t keep in mind is that the statistical world contains the biases of the creators, but RAPM much less so. RAPM May be more likely to get it wrong in small samples, but as a tool it is unique in the sense of finding truths that other tools couldn’t.

More succinctly: the reason Nick Collison isn’t a top player in other metrics is because nobody would think of Nick Collison as a top player. There’s power in the fact it does that, even if we shouldn’t take it 100% on faith.

[dontcalltimeout]

Read on Twitter

[oldschooled]

You can really make a case for Curry being the best player in the league for the past 3 years.

[SpreeS]

Curry is GOD in these kind of stats. Nothing new here.

[trex_8063]

There doesn't appear to be any particular explanation of what the number actually means or represents.

Curry's effect on the win probability is 9.3 per 100 possessions; meaning??......

[QRich3]

ESPN just posted the first batch of RPM for the season:

https://www.espn.com/nba/statistics/rpm/_/sort/RPM

Marc Gasol first, and then some of the regulars. Nothing too surprising really, good to see role guys playing well like Siakam or Bjelica do well here, and the emergence of young guys like Richardson, Sabonis or Harrell.

[Clyde Frazier]

I just bumped the below thread thinking it was on the PC board, but figured i'd post it here, too if anyone's interested:

Raw plus/minus for 93-94, 94-95, 95-96 seasons

That data was then used to create some net rating and on/off #s

viewtopic.php?f=344&t=1343246