Also Beverley's ranking is based on a very small career sample size. He only played 713 minutes in 41 games last year(Lin played 2640) and Beverley has missed a lot of games this year against teams in the playoff race(DAL,LAC,DAL,SAS,MEM,OKC,ATL,WAS,OKC,WAS,LAC,BRK,TOR,OKC) and he might miss another SAS game coming up. This stat and xRAPM start with on/off stats if you play only 713 minutes the previous season and miss a dozen or so games against playoff bound teams your on/off stats are better than they would be if you played more minutes last season and more minutes against tough teams this year. Playoff teams usually have better defenses and offenses than non playoff teams and just playing games against them usually negatively impacts most players on/off numbers. These stats are supposed to be adjusted for toughness of opponents but I'm guessing they probably don't go far enough to weed out inflated rankings.
Omer's and Howard's DRPM are virtually the same. From the eye test, i think Omer is a better one on one defender but Howard has ability to block shots and just his presence alone has a profound impact. But i like Omer better as a defender. Just my two cents
Look at where Howard, Harden and Asik were last year http://stats-for-the-nba.appspot.com/ratings/2013.html
Holy crap Dwight Howard's Pure Ramp 2008 4th @ 7.4 2009 4th @ 8.1 2010 2nd @ 9.5 2011 1st @ 9.7 2012 2nd @ 8.8 2013 6th @ 5.5 (W/Lakers) 2014 31st @ 2.7 What's wrong with Dwight and how is he worse than last year?
This is cool. I would say its the best single number player rating out there which I am aware. Its comprehensive, and the results generally look right. That said, trying to sum up how good a player is in a single number will inevitably be flawed. Adding standard errors, so people could (more appropriately) understand these ratings as a probabilistic range, might have been better.
You are comparing pure RAPM 2014 to xRAPM 2008-2013. His xRAPM for 2014 is 4.0. xRAPM takes multiple years into account to smooth out the data as a year is still a small sample size to gauge a player's value. While he may be better in 2014, his 2012-2014 performance is probably worse than his 2011-2013 performance.
Ya think?? Cuz we all know Bev is the 3rd best PG in the league. Like Bev a lot, but that's redonkulous.
I think the disconnect people have with stats like this really comes down to the language we use when it comes to player performance evaluation. The wrong way to interpret these stats is: "Player A has a higher RPM than Player B, therefore Player A is better than Player B." I'm pretty certain even the authors of RPM/xRAPM do not believe such a thing. No one should mistake it for a holy grail. Its a model which provides one estimation of a player's value to his team. The goal is for it to closely capture player value as often as possible, but the hit rate will never be 100% because much of the game is inherently random. That's why cherry-picking certain ratings which don't seem right is not a particularly compelling critique of the methodology, in my view.
