Crow wrote:I am more comfortable with top talent being scarce than I am with the bottom being so heavy but you are right to recommend comparing the distributions on other metrics. I've known that median performance is below mean but seeing 1000 of 1280 below neutral was more than I expected. Not a bell curve.
Nope, not a bell curve. Highly skewed.
See Tom Tango's seminal piece on talent distribution, written over 10 years ago. Amazing how many sports statisticians don't realize what is going on. http://www.tangotiger.net/talent.html
Developer of Box Plus/Minus
APBRmetrics Forum Administrator Twitter.com/DSMok1
I may have seen this or something similar long ago, but... I forget things.
It is a "seminal" piece but the author says it is 10% done and partially based on conjecture ( "estimates, some data and intuition") Is there follow-up also worth reading, more complete and fully fact based?
Chart 6 is not that different a shape that Chart 1. Or Mike's minutes table.
This is a bigger question, but for folks who had or moved to zero interest in baseball and basically never or hardly ever read anything in the field of sabermetrics, what are the top 5-10 things written there with major applicability to study and understanding of the NBA? Again for someone with zero interest in baseball per se.
I guess about 70% were below neutral in 2013-4 Rapm, similar to the distribution in the 14 yr set. So many measures to make, compare, get used to.
Neutral on 1 or 14 yr rapm? On this scale you are a high performer. Easy to make a big misinterpretation, you know with the notion than a zero might be related to a median player. Nope, mean performance. You can know that and occasionally slip.
Some concerns expressed at recycling from 6-7 months ago, so maybe I should lay them down.
But... in further commentary I see medial performance on 14 yr is about -1.8. In one yr RAPM is was -0.9. In some ways it would be easier if median was zero, but it is not and apparently won't be, so remember, remember, calibrate.
RAPM total should be 0 for (RAPM_Player1/200*Poss_Player1)+.....
In this regard, the total of 14-year RAPM value for 1383 players is 68. And that's because there's only one digit after decimal. So, it's 0 in fact. So there aren't any priors (age adjusted and then de-adjusted)
For 2013-14 xRAPM the total is 4552 for 474 players. That's the prior, adjustment, weighting effect. So I don't think making a comparison between pure RAPM and xRAPM values is really meaningful.