D: Because 2-3 months of good results outweigh 3 appearances. The "good" results were somewhat overrated, and no, they don't necessarily do so in this case. I know people love to spam "sample size," but there's also the matter of extremity. Extremely bad results are less likely to be sample-size flukes than ordinarily bad ones (and the same goes for good). Same argument as we had last year when Brett Jackson had a million Ks in 50 PAs or whatever. 13 batters faced: 2 BB 2 HBP 2 HR 2 K 5 BIP That's bad enough to be legitimately concerned over, sample size or no. Legitimately concerned, sure. Cut and run at first sign of difficulty, no. Velocity seems normal, location seems off, front office likely thinks he should get time to try to get control straightened out. First sign? You'd have a point if that wasn't an absurdly wrong notion.