It's getting a little noisy in the Oswalt thread. Might as well discuss it here.
To me, defensive independent pitching stats that show what the pitcher did in isolation from his defense, manager, the official scorer and luck are much more indicative of what a pitcher's true performance was and what it should likely be going forward.
Arroyo's ERA over the past 3 years has been about 4% worse than the rest of the league. His FIP* has been about 24% worse than the rest of the league. I think it's likely he pitches better next year but more to the tune of a 5-ish FIP rather than a 5.71 FIP. Improved but still poor.
Sidebar: Bronson Arroyo is looking an awful lot like Eric Milton in more ways than one.
Amongst starting pitchers over the past three seasons, no peripheral, hit rate or "stat" correlates higher (0.738) to BABIP than the difference between ERA and FIP.
While BABIP may not explain the difference 100%, it explains it around 70-80% (I think this is highly generous but it's almost Valentine's Day so, what the hell.) So, great: 25% of the difference was keeping hitters off-balance or whatever it is one might think makes him "good."
I think many of these macro stats are show in his pitch selection. Check out his PitchFx stats. Fastball velocity is down which makes the offspeed pitches less effective. He threw the cut in his "good" years and now he's throwing it almost a third as much as he used to. In fact, the 25% of "good" from 2010 might be his changeup which he threw 22.5% vs. 16.4% of the time since 2007; it had a lot of horizontal movement.
But 75% of it was simply the ball finding a glove at a higher rate than usual. I don't call this luck, I call it nature.
Problem solved, high five!
*A note about FIP. When I look at FIP, it's shorthand for a great number of other things to check out: component peripherals, other defense independent metrics, BABIP, hit types, etc. Even though I post "FIP indicates," it doesn't mean it's the _only_ thing looked at. It means that it's shorthand and a time saver from having to type basically the same thing over and over again.