Better than the literal worst among his peers seems like a good baseline.
You keep bringing up Hill, but the funny thing about Hill is that his stats didn't really fluctuate much, while he was in Sacramento. Even his counting stats, per/36 and per/100p, were surprisingly in-line with his career numbers. If anything, Hill's tenure in Sacramento demonstrates that he's a guy who can make a good team better, but can't help a bad team become good. As to why bring up advance stats, because they indicate a disturbing trend, that's why; they indicate that he's a net negative on a bad team and, at best, a non-factor on a good team.