... he was dead last on his team in dRAPM
I don't have any faith in this stat. It puts Lebron, for instance, as the 162nd best defender in the NBA, wedged between Hedo Turkuglu and Jeremy Lin. Klay Thompson is 215th, very close to Steph Curry (231st).
There are dozens and dozens of similar examples.
Ah, I see. You misunderstood me. I wasn't referencing RPM, which I agree is fishy if only because ESPN have been hesitant to explicitly describe the calculation process. But RAPM. The two are similar in their attempt to reduce the noise of +/- and APM, but RPM makes use of box scores to try, similar to xRAPM but with some other additions, to improve the prediction quality.
Still, your argument still stands - if inadvertently - because there is a high correlation between RPM and RAPM. RAPM has issues, but it's defensive split is one of the better measures of defensive performance. James' dRAPM for this season is still a negative, just like his defensive RPM. The best explanation I can offer is that he was coasting throughout the majority of the season, something most can attest to. During the Playoffs, his dRAPM was much higher.
I hope all this doesn't deter us from my original point, which was, Irving was very bad last season.