Your statement was that "starters as a whole improve drastically when they leave the Twins." Crain and Guerrier are relievers. Dickey started exactly one game for the Twins. Why is K/9 your only measure of whether or not a pitcher has improved? Two of the three new guys are coming off surgery, and all three are coming from the National League. Facing pitchers helps the K rate compared to facing DHs, though I don't know the expected amount. Some adjustment is called for when pitchers change leagues. According to Baseball Reference, Lohse pitched 908.1 innings for the Twins, with 561 K's. That works out to 5.56 K/9, virtually the same as his post-Twins career (all of which was in the NL). Garza was a kid who was likely to improve- he had pitched 133 total innings for the Twins. What about Slowey since he left the Twins, not just the last few months? Is a change of 0.5 even statistically significant without factoring in age, league, etc? Jason Marquis, really? You're arguing that the Twins' alleged philosophy had a huge impact for 7 starts on a guy who had made over 250 starts in his career? A guy who wasn't even in spring training? He was barely on the club long enough to learn the name of the clubhouse attendants. You forgot Eric Milton, but he went the opposite direction. We'd all like to forget Carlos Silva, who never struck anyone out anyway. I could argue the Twins gave up too early on Jim Kaat and Jerry Koosman and I'd be right, but it wouldn't be relevant. Pavano improved by 1.1 K/9 his first year with the Twins, then fell apart. Andersen's fault? I doubt it. I don't think the claim that starting pitchers improve when they leave the Twins has very much support, personally. I think the claim that the Twins have had a lot of crappy pitchers is much more plausible. You could also argue that they have given up on a few guys they should have kept, which is certainly true (and true of an awful lot of teams, but no less true here). But I sure don't see a long track record of starters going off to be wildly successful elsewhere.