Playoff RAPM (Regular Season Informed)

Home for all your discussion of basketball statistical analysis.
Post Reply
sbs
Posts: 17
Joined: Fri Oct 19, 2012 7:25 am

Playoff RAPM (Regular Season Informed)

Post by sbs » Fri Jul 19, 2019 12:21 am

1998-2018 Playoff RAPM that uses Regular Season RAPM as a prior. It's only single component so there's no O/D split-out.

Here's the top/bottom 30 changes from regular season to playoffs with 8000 playoff possessions:

Top
Image

Bottom
Image

https://docs.google.com/spreadsheets/d/ ... sp=sharing

Crow
Posts: 6248
Joined: Thu Apr 14, 2011 11:10 pm

Re: Playoff RAPM (Regular Season Informed)

Post by Crow » Fri Jul 19, 2019 5:59 am

Pretty small differences for tails.

Might be a bit more digestible if it were only to 1-2 decimals instead of 6.

Wonder what you'd find is you summed by coach.

Mike G
Posts: 4429
Joined: Fri Apr 15, 2011 12:02 am
Location: Asheville, NC

Re: Playoff RAPM (Regular Season Informed)

Post by Mike G » Fri Jul 19, 2019 11:35 am

What is the number after the name?
The link only leads to the same 60 players. Is there more somewhere?

It's surprising to see Tyson Chandler among the playoff improvers. Always his numbers drop off in postseason, often by a lot.
Ray Allen also declined more than the norm, in boxscores. And Marion, and Zach.

The 30th most-improved is only +0.42, and the #30 least is just -0.16
Does this disparity carry throughout the list?

sbs
Posts: 17
Joined: Fri Oct 19, 2012 7:25 am

Re: Playoff RAPM (Regular Season Informed)

Post by sbs » Fri Jul 19, 2019 4:19 pm

The number in the player name is their ID from stats.nba.com.

There's going be some confounding influences in here for sure, some of which could probably be handled.

Taking a guess at Chandler, his playoff possessions come from his seasons where he is close to his prime (Age 22-32) and the bulk come from his 2011 run with Dallas (~28%, Age 28). His regular season RAPM includes two stretches, Age 19-21 and Age 33-35, that certainly bring down his prior.

The disparity does carry throughout the list which I suppose can be interpreted that most players perform very similarly in the playoffs as they do in the regular season.

shadow
Posts: 220
Joined: Fri May 29, 2015 3:38 am

Re: Playoff RAPM (Regular Season Informed)

Post by shadow » Tue Jul 23, 2019 2:56 pm

Would you be able to share any insight as to how you used Regular Season RAPM as a prior?

Crow
Posts: 6248
Joined: Thu Apr 14, 2011 11:10 pm

Re: Playoff RAPM (Regular Season Informed)

Post by Crow » Tue Jul 23, 2019 3:14 pm

Good concept. We want the best playoff RAPM possible. Regular season priors should help.

sbs
Posts: 17
Joined: Fri Oct 19, 2012 7:25 am

Re: Playoff RAPM (Regular Season Informed)

Post by sbs » Wed Jul 24, 2019 8:31 pm

shadow wrote:
Tue Jul 23, 2019 2:56 pm
Would you be able to share any insight as to how you used Regular Season RAPM as a prior?
Basically the same process as a SPM prior. Adjust the possession expectation based on the player's priors, run the regression, and then add the priors back in at the end.

Post Reply