Interesting stuff, but the muddling factor here, which 538 mentioned but didn't seem to fully account for, is that the game had to be both late and close. So having more calls go against us could be partly or entirely a function of being in those clutch situations more often than most teams.
Might be better to measure as number of blown calls per minute in those situations. That's still influenced by pace, intentional fouling situations, etc, but seems like a cleaner measure. It'd also be nice to see our bad calls compared to how many bad calls we benefited from, in case our close games just had refs prone to swallow the whistle more.