Plus-Minus: A misleading statistic?
Let me suggest a plausible reason that may account for Monta Ellis's very poor adjusted plus-minus while the Warriors looked very good with him on the bench: the effects of garbage time.
It works like this: the Warriors are/ were a terrible team. They often get beaten badly. During the part of the game that their opponents are accumulating their very big lead, Ellis was in the game and so were the opponents' best players. Ellis's number therefore looked very bad.
Once the game was so far out of hand that both teams emptied their benches, the Warriors bench proved relatively efficient at outscoring their opponents with meaningless points.
Unless the system tracks the statistics deeply enough to prove that the above is incorrect, then the statistic is misleading, and minimally helpful.
After all, if the statistic is actually meaningful as is, then that would suggest that loading up Ellis with a bunch of DNP-Coaches Decisions would have made the Warriors about the number 4 or number 5 seed in the West-- that's what having a team +4 differential will do for you. And clearly that's implausible.
Also, I'm trying to figure out why you're trying to "introduce" a statistic that you refer to as the "most understood". If you think this is the "most understood-statistic" then what's the point of writing about it?
Poor word choice?