Over the years I have suggested many, many dozens of specialized RAPM splits. Search RAPM splits and you can see them. I could probably come up with more, including those linked to player tracking data (simple to very detailed). Basically no one ever showed public interest in them. Sample size can be a concern but multi-year would lessen the concern. The choice has always been look / consider or not and the former always seemed the better answer to me. Time is a constraint for outside analyst but teams can make as much time as they are willing to, willing to spend. (9 figure revenue enterprises pursuing 8 and 9 figure outcome prizes can afford a lot.)
I am not going to go long on the overall subject now, but if anyone ever were interested in RAPM splits or did them or might explore doing some, this is a chance to mention that or say more, if interested.
The main thing I am going to ask right now is whether anyone has or would be willing to produce and then share pair rapm for last season (and / or next), hopefully for offensive / defensive splits or to factor level. If so, I would make use of it and be appreciative. Pair RAPM has occasionally been available in public, but not recently. For any set up and capable of doing it carefully, it would be something I'd think others would notice.
On specialized RAPM products
Re: On specialized RAPM products
Would be very interested in single-season regular season only vanilla RAPM with off/def splits from 96-97 to 22-23. Not sure if this is specialized, but there isn't a single great source.
Re: On specialized RAPM products
RAPM and RAPM factors gone from Ryan Davis' site (at least for now).
Hope for return of RAPM factors, from somebody. Rough guesses might be inferred from raw data but this is a big step back.
I never hear anyone else mention RAPM factors. Do any insiders "use" them? I dunno I'd hope so but it may be rare or non-existent.
RPM available, fwiw.
2022-23 RAPM offensive and defensive rollups might still be available somewhere... but I didn't find at the previous secondary source, craftednba.
Hope for return of RAPM factors, from somebody. Rough guesses might be inferred from raw data but this is a big step back.
I never hear anyone else mention RAPM factors. Do any insiders "use" them? I dunno I'd hope so but it may be rare or non-existent.
RPM available, fwiw.
2022-23 RAPM offensive and defensive rollups might still be available somewhere... but I didn't find at the previous secondary source, craftednba.
Re: On specialized RAPM products
SPM and RAPM are fairly well correlated.
SPM broken into the 4 factors vs. RAPM factors? I don't immediately recall seeing this in academic literature. Remind if it is.
How good the correlation on each at league level? Team level, player level? Which factors have highest and lowest correlations at each level?
I dunno. I probably did a few individual comparisons in past. The variance between the two could be real uncounted by boxscore impact or error by one or both or almost certainly some of each, so it is hard to say what it means, but it is useful for raising questions and prompting more research. (Said to be a main purpose of analysis).
But incentive to do such independent public research may decline or to even talk more much about it when no one else talks about such a task and teams don't appear to be looking to add such analysis.
The strengthening of my interest in this work unfortunately also coincides with the temporary or permanent departure of Ryan Davis' data from public view. I have downloaded some data sources in past out of fear of losing access but I didn't get such a fear in this immediate case.
Would it be available by request? This is the only place I currently knows to ask the questíon. Did anyone else store a comprehensive copy of the database?
If such comparison / study is happening within team offices, somebody that knows could say so. Maybe it is, but it hasn't been said exactly to my knowledge (including in a recent book) and probably won't unless it is so common that it isn't really news, isn't revealing a secret / work edge (an attitude that hasn't developed to date).
Oh well. For now.
SPM broken into the 4 factors vs. RAPM factors? I don't immediately recall seeing this in academic literature. Remind if it is.
How good the correlation on each at league level? Team level, player level? Which factors have highest and lowest correlations at each level?
I dunno. I probably did a few individual comparisons in past. The variance between the two could be real uncounted by boxscore impact or error by one or both or almost certainly some of each, so it is hard to say what it means, but it is useful for raising questions and prompting more research. (Said to be a main purpose of analysis).
But incentive to do such independent public research may decline or to even talk more much about it when no one else talks about such a task and teams don't appear to be looking to add such analysis.
The strengthening of my interest in this work unfortunately also coincides with the temporary or permanent departure of Ryan Davis' data from public view. I have downloaded some data sources in past out of fear of losing access but I didn't get such a fear in this immediate case.
Would it be available by request? This is the only place I currently knows to ask the questíon. Did anyone else store a comprehensive copy of the database?
If such comparison / study is happening within team offices, somebody that knows could say so. Maybe it is, but it hasn't been said exactly to my knowledge (including in a recent book) and probably won't unless it is so common that it isn't really news, isn't revealing a secret / work edge (an attitude that hasn't developed to date).
Oh well. For now.
Re: On specialized RAPM products
An example of actually checking RAPM and RAPM factors:
https://x.com/bballstrategy/status/1701 ... 89569?s=20
Also checked Dort's estimated rebounding impact recently.
https://x.com/bballstrategy/status/1701 ... 89569?s=20
Also checked Dort's estimated rebounding impact recently.
Re: On specialized RAPM products
Stay tuned, I think.