I was at The Book Blog the other day, and MGL made a comment that I found very interesting:
And there is NO evidence that “trends” mean anything for pitchers or batters, other than how they affect a Marcel-type weighted average. If I am wrong about that, I invite Steve or anyone else to enlighten me.
IOW, a pitcher with a K per 9 (for convenience sake, as opposed to K per PA), over the last 4 years, of:
should have roughly the same K/9 projection as someone with:
I may be reading this wrong, but it appears that he is saying that trends do not predict future performance. I know that hot streaks do not predict performance (thanks to Dave Cameron), but I always thought it was different in a larger sample size of a full season.
I would be intersted if someone here could point me in the direction of an article or something about this.