There will be anomolies with plus/minus. Its an inherently noisy stat. You can still tease out potentially useful information from it, but you shouldn't rely exclusively on it. Its a stat to complement or reinforce what you observe about the player. And sometimes, what it says just won't make much sense. At least on the surface, without digging deeper.
Yeah its not completely useless. Its a fun stat to look at. I think it certainly plays a role to determine whether a guy can push your team into the negative realm while he is playing. For good players on losing teams I think its even harder of a stat to prove if the good player has bad player and coaches around him already. So its good to reassure who the good players are but harder to determine from the bad teams who the good players are and aren't.
OK. You're not understanding this. I'm not saying you only apply the +/- stat to evaluate a player or determine their PT. I was saying you were doing that and assuming that's what it for. There may be a random poster here and there who implies that because they're not familiar with how it works, but i sure have never used it that way and posters who use it frequenlty like durvasa have always agreed it's not to be used alone. You're going on this crusade to prove that advanced stats alone don't tell the whole story, but you're so caught up in it you failed to realize the majority, if not all, of these "stat geeks" aren't even claiming that. They're using the stat, along with their observations and other evidence, to prove their points. However, a poster may ask or say something that can be countered with just a plus/minus, and in those cases it it does make sense to reply with just that stat. Maybe that's where you're getting confused, but it doesnt mean they're saying it's the only tool to you need in every situation.
People act like Morey runs around spewing numbers and making that the only say-so in plalyer evaluation. He's said many times they evaluate based upon a combination of factors which involve stats and on-court/basketball player evaluation. Many articles on Morey have stated that he understands the limits of modeling players purely statistically. Instead, he apparently lets basketball action try to validate what stats show and vice versa - is what the numbers show really what's happening and is what's happening really reflected in the numbers? As for the +/- stat, Morey uses a version of adjusted +/- as part of his evaluation. The basic +/- can be a goofy stat and I doubt anybody uses it as a sole basis for player evaluation. The "noise" in the pure +/- stat everyone refers to in this thread comes from the fact that it doesn't take into account the quality of the players the player in question played against or the ones he played with while on the court. Adjusted +/- tries to correct this using regression analysis. Is it perfect? Of course not, but it's a much better stat than pure +/-.
