• Moderators, please send me a PM if you are unable to access mod permissions. Thanks, Habsy.

OT: The Official Toronto Raptors Thread

Poor Swaggy Z

iu


Mate....

RAPM is present in all of the metrics that exist. It's the grandfather of NBA advanced metrics, the thing you copied shows that here:

1776383618154.png

Like, over a decade ago I used to post Hollinger's ESPN version of RAPM on here regularly.

Where it (and metrics like LEBRON) are now considered weak is that they don't utilized modern tracking data, as is mentioned in the EPM bit you quoted right here:

1776383697134.png

Which is why I made the corsi joke. Corsi is outdated, but the logic of corsi is still present in more modern xG metrics.

LEBRON is a refreshed RAPM, but still with it's core weaknesses that BPM, EPM, Crafted (which is really just a weighted value of the other plus minus stats) all address. You don't see me use those and haven't for a bunch of years now because the newer metrics (epm, bpm 2.0, etc) are just better.

Fwiw, the LEBRON developers themselves have said that they plan to add tracking data into an updated version.
 
You haven't seen me use a RAPM based metric on here for years

I misspoke.

When I think of epm and bpm, I consider them to be modern updated metrics that rely heavily on tracking data. Yeah, you've got me that they're technically derived from a nearly 20 yr old metric but you're (intentionally) missing the point I'm making. Metrics like LEBRON are considered flawed compared to newer metrics for reasons already stated.

It would be like you and I arguing NHL analytics that include shot distance/location data and older metrics that don't, and then my misspeaking and saying that I don't use metrics that rely on corsi. Of course the newer metrics include shot differential, but they've added newer features to provide significantly more context and cover for previous weaknesses in the metrics.

as for your peeling back the onion on crafted, I've never looked at their weightings before and just understood it to be a weighted plus minus metric based on others. I'll take that into account going forward.

All of this nonsense is you running cover fire for an older metric with a known weakness so that you can include it in a non weighted average.

In the bit on bball index you grabbed screenshots from, you left one out:

1776385658986.png

Spoiler: They've never added that, and the statement is from 2022.
 
It's the difference between an 81st percentile player and 84th percentile in this specific case.

That's not what a rounding error is.
 
Back
Top