Absolutely I was talking about career stats. Statistics says there will be people who under/over perform consistently. That's why tails will come up 9 times in a row when you flip a coin 100 times. That is the definition of a bell curve. If there weren't outliers then I would be worried.
Is FIP perfect? No. Does a player's ERA outperforming their FIP for their career mean FIP is broken? No. Could there be something FIP isn't catching? Sure, but there better be more proof for it than "Random player X is out performing his FIP, it must be broken".
You're confusing your analogies. The coin coming up heads nine times in a row is the equivalent of a single season. It's a statistical aberration but not unreasonably so (1 in 512 chance).
The coin coming up heads 70 times out of 100 is the equivalent of a career (1 in 43,000-something chance). There's a huge statistical difference there. The more data points, the more the odds of unusual performance plummet.
When you apply that to a real world person who plays for multiple teams, in multiple leagues, in multiple divisions, over the course of a 12 year career, and it happens almost every year, the odds are so low they're not worth mentioning.