First Step wrote:Any statistic that has LeBron James ranked 81st is highly suspect. I don't care what context it is.
It's an amazing stat. But it's not something that can be taken at face value. Role has to be accounted for and it's very prone to errors in the case of smaller samples, which is why single-year RAPM isn't very accurate, and prior-informed is much more valuable. In the case of LeBron, he's played a league-leading 2709 minutes this year and he's been a part of some horribly performing lineups involving Thomas/Crowder/Tristan/JR/Rose/etc that haven't seen enough minutes
without LeBron due to the fact that he hasn't missed any time and Lue has very odd rotation patterns that have involved almost the entire bench coming in for the starters.
Since the bench has played so well all year while the starters other than LeBron and Love have been awful, RAPM needs to assign some negative values to the starters, but unfortunately there isn't a good enough sample size of "off the court" data for LeBron for it to accurately isolate his impact. In the 2014-15 he missed a good number of games and the team completely fell apart without him, which explains why he posted an extremely high +8.77, far higher than anyone else in the league that year. Though you could definitely argue that he's been a more impactful player this year than that season (worse on defense, but also clearly better on offense).
With guys like Curry and Paul and Embiid, they've all missed significant time, so we've gotten the chance to see just how valuable they are to their teams. Furthermore, they all lead amazing starting units while LeBron has played with some pretty extreme liabilities this season in the starting lineup, which is difficult for RAPM to truly recognize without a greater sample.