Matt Cain is a commonly used example of why the DIPS theory is invalid. Of course, one counterexample shouldn't be enough to invalidate a whole entire theory, but even if it were, hopefully people would at least pick a valid counterexample. Matt Cain is not one.
People are quick to point out that Cain's career BABIP is .275 which is statistically significant because of his career 1028 IP, and so his career ERA of 3.46 is a full .41 lower than his FIP of 3.87, which is also meaningful. This discrepancy isn't going to vanish anytime soon either, but to claim that it exists because the DIPS theory is wrong and that Cain has significant control over balls after they've gone into play because of skill is baseless when one considers his home/road splits:
Cain has well over 450 innings pitched on both home and road games so these numbers are probably stable at this point. While it's repeatable, this extraordinarily low BABIP doesn't seem to be a skill. The claim that Cain has long-term control over his BABIP to a significant degree appears largely to be merely an abnormality of his pitcher-friendly home park.