2013 RAPM up to the All Star break

[v-zero]

Rookies get a fixed (negative) prior.

[colts18]

You said this built from 2001-onwards. Does that mean you have RAPM data for 2001-2012?

[v-zero]

I have stint data (matchup data, whatever you want to call it) and an extended box score from play by play over that period.

[EvanZ]

There are some seriously puzzling names at the top there. Salmons? Ellington?

[v-zero]

Like I said, single season NPI RAPM is crap.

[permaximum]

Like I said, single season NPI RAPM is crap.

Then how are we going to judge player performances season by season?

On the contrary, single season non prior informed RAPM is the best for my needs. I want to analyse players' seasonal performances.

[DSMok1]

Like I said, single season NPI RAPM is crap.

Then how are we going to judge player performances season by season?

On the contrary, single season non prior informed RAPM is the best for my needs. I want to analyse players' seasonal performances.

You can't analyze players' seasonal performances with single-season NPI RAPM. The statistical noise exceeds the signal, period. It is impossible to disentangle with that sample size and the collinearity present in the sample.

[v-zero]

Then how are we going to judge player performances season by season?

On the contrary, single season non prior informed RAPM is the best for my needs. I want to analyse players' seasonal performances.

It isn't best for your needs, unless what you need is an extremely poor guesstimate of player ability.

It's not that it's a little noisy, or suffers a bit from problems, it's that the answers it gives are wrong, often very.

[Crow]

(Slightly revised to correct one point)


Using the data from this source (prior informed I believe) from a few weeks ago I looked at the top 100 players. I sorted by big man, wing and pg using a bit of judgment who played like a big man or a wing for a few tweener forwards. The group split right at 40 bigs, 40 wings and 20 pgs. The bigs averaged about +3 RAPM and that impact came almost entirely from defense. The pgs averaged about +3 RAPM and almost entirely from offense. The wings averaged about +3 RAPM with +2 coming from offense and +1 from defense.


If / when I have time I might sort these 3 groups (from the top 100, 150 or all players) into 3 tiers each and average their boxscore stats and SPM inferred direct contributions and see how things shake out by player type and see how much of the RAPM estimated value is coming from beyond the boxscore. One could then compare individual players to the averages of the 3 tiers for their peers and see how much they vary in the box score stats and outside the boxscore impact from average or elite. That might help in considering what they need to "work on". One could also do "side by side" comparison of their videotape and synergy and sportsvu data with that of the elite, especially those with the most similar physique / role / age, etc. Even if one didn't put that much stock in the RAPM data itself, it could still be useful for sorting and analyzing and getting down to real insight or advice.

[permaximum]

@DSMok1 and v-zero

I can't agree. I calculated yearly RAPM for past seasons a few months ago and subjectively I was satisfied with the results except the lockout season's were a little bit out of place because of less data.

@Crow

I scrapped my work so you can only take my word on this one but I believe, box score (not advanced) explains 30-35% of player performances (RAPM). However I would like to see the difference between bigs, wings and pgs.

[Crow]

Well, thanks for sharing the initial data set.

[Crow]

So I'm doing 1-yr RAPM with no prior and finding something very, very strange (or maybe not considering the post above):

https://docs.google.com/spreadsheet/pub ... utput=html

No, not that Amir Johnson is near the top. Or Chris Paul or Tim Duncan or Kevin Durant.

What is strange is that LeBron is at #363 with a rating of -1.70.

I'm calculating RAPM exactly the way I've done it in the past (using cv.glmnet to find lambda that gives minimum cv-error). The lambda I'm getting is ~100. This is puzzling, to say the least.

v-zero, have you found something similar when you do 1-yr non-prior-informed ridge regression?

I was moderately troubled by these results (because they vary a fair amount from other RAPM and my subjective expectations) but didn't talk myself into saying it til now. Did you handle low minute players similar to JE and others? Any further analysis of the results or changes to the method?

[permaximum]

Does anyone have matchup data of the complete 2012-13 season?

[colts18]

This is basically a data dump as a result of incomplete work I'm doing - I know people are missing xRAPM and IPV (and most largely the lack of a good defensive metric) so hopefully this will help. It's built from 2000-01 onwards, with the playoffs, and with each year's ratings as the prior for the next. There is no box-score information involved, and new players to the league receive a prior according to their percentage of possible possessions played - so in 2000-01 everybody got a prior based on their percentage playing time, basically. Lambdas were found via cross-validation until it was clear that lambda settled (i.e. that the ratio of useful information contained in the prior to useful information contained in the in-season measurements became constant, which more or less happened in 2003-04). The ratings are such that possession weighted offensive ratings sum to zero, and possession weighted defensive ratings sum to zero, so that these can be interpreted with zero as the origin for for defence and offence.

Single Year Prior Informed RAPM:
https://docs.google.com/spreadsheet/ccc ... sp=sharing

Single Year Non Prior Informed RAPM:
https://docs.google.com/spreadsheet/ccc ... sp=sharing

N.B. I'm not suggesting anybody have too much faith in these, given that they aren't my end goal, but they should be as good as the estimates JE made with RiRAPM.


EDIT:

These have been discussed somewhat at gamefaqs and possibly elsewhere, I don't post there so I'll post here: Don't trust these ratings as prescriptive. They're a work in progress. Specifically intelligently decaying 'old' information (data from prior years) and weighting priors by sample size (in a basic sense) is yet to be properly done, so when you see KG and Amir ridiculously high on D you should assume the number is too high, but that they're pretty excellent. As for guards being underrated on D: size matters, guards are nothing like as useful as bigs on the interior, and since shooting% is roughly inversely proportional to distance it's fair to assume that great interior D beats great perimeter D. However, guys like Battier whose awesome perimeter D doesn't show up in the box score will show up here.

EDIT 30/03/2013:

Updated with most up to date pbp data and a vast improvement on the priors.

Do you have the 2013 season updated?

[v-zero]

I'll update them as soon as I can, should be this week.