Is there anything wrong with the stats or methodology? People choose to focus in on a very small set of games, instead of looking at the whole picture. The data is there.
Last season, Millsap's game against Miami (along with a couple more big games early) were hard to forget about. When he couldn't replicate his Miami performance, people assumed that that somehow represented a season-long trend. Obviously, judging by the monthly splits anyway, this was not the case.
The same thing has happened this season. After a middling first 8 games, Millsap had 9 games where he put up big numbers. Over the last 8, his shot has crashed down to earth (I would expect this to correct upward), and his numbers look a lot like his first 8 games. Unfortunately, again, Jazz fans have completely forgotten about the first 8 games.
It's hard to spot an obvious trend. What I feel comfortable saying is that there's a high variance in Millsap's offensive production. He'll have stretches of great games and stretches of mediocre ones. He's not a 1st option, obviously, and probably not a very dependable 2nd option either. He is, however, and extremely useful player who does loads to help his teams win. It's hard to argue with his production v. his opposite number and his adjusted (for teammates and opponents) +/- over the last 1 and 2 years (top 5 in the NBA).