R.A. Dickey, May-July: 7-4, 2.32 ERA, 14 GS, 93 IP, 62 K, 24 BB, 5 HR
R.A. Dickey, Aug.-Sept.: 4-3, 3.74 ERA, 10 GS, 67 1/3 IP, 33 K, 15 BB, 7 HR
OK, lots of things at play here. First of all, pretty arbitrary endpoints, and I don’t think anyone reasonably expected Dickey to be as good going forward as he was in the first two and a half months of his Mets career.
Plus since we’re dealing in 14- and 10-start samples, all the trappings of small sample size are in play. It’s really hard to draw any firm conclusions from any of the information above.
But it does look as though Dickey is regressing a bit with exposure, which probably could be expected. Again, it’s not a perfect comparison because they’re hardly identical knuckleballers, but Dickey’s initial run of enormous success bears some resemblance to Tim Wakefield’s in his first time through the National League in 1992 and then the American League in 1995.
Obviously Wakefield settled into a nice career as a solid Major League innings-eater, which Dickey certainly seems fit to become as long as he can control his knuckleball and yield a ton of groundballs.
Plus he has the funny pitching face and the love of literature and all that, which is cool.
Tedquarters 
I’m going with small sample set. 12 of the 28 Aug-Sep runs were on 4 3-run homers, 3 of them with 2 outs. His 2 best starts were Aug-Sep, including a complete game 1-hitter the 3rd time he saw the Phils. The worst start (by ERs) was the first time he saw the Cubs. His ERA was much better in July than in May or June. He’s not as good as he looked at the beginning or as bad as he looks now. Not enough evidence of regressing with exposure, IMO.
A series of good points.