donchoi said:Savin Hillbilly, that analogy makes some sense, but I think it's different when you apply it to a group of players rather than individuals (as shown above), since you expect statistical regression to the mean to cancel out a bit, countering the impact of "outlier" seasons like Verlander's 2014. I think that the numbers do show that this year's front five are not as good as last year's (which we all agree on). But I think that the effect of increased groundballs in front of our solid infield defense will be cancelled out, and probably overwhelmed by fewer strikeouts and more walks. Those have been shown to have a much stronger influence on FIP and ERA than increasing GB% from say 45% to 50%.
I don't think you understood my analogy, though. It has nothing to do with whether Verlander's season was an outlier. (I should probably have left the word "surprisingly" out, since it's not really relevant.) The thing I was trying to get at was your suggestion that the performance of the 2014 starters relative to their initial projection should affect our thinking about the likelihood of various possible outcomes for the 2015 starters. Here's what you said:
As I mentioned above, you're right that those numbers aren't really a fair representation of what that Opening Day rotation could do. But in fact, it is what happened with that starting five. Now we are looking at a top five which are worse than those guys. Is it reasonable to think that they will outperform last year? Based on what? Blind optimism? Homerism?
No, based on the fact that a projection defines the midpoint of a range of possible outcomes. The 2014 starters were projected for an ERA of 4.10 and wound up with a 4.36; assuming for a moment that 4.36 was the worst reasonably likely outcome (however you want to define the limits of "reasonably likely"), then presumably the best reasonably likely outcome would have been something like 3.85. This year's starters are projected for 4.25; if we assume the same range of reasonably likely divergences from that projection as last year, the range would be between 4.00 and 4.50. Therefore about 70% (35/50) of reasonably likely outcomes for this year's rotation are better than the outcome for last year's rotation.
Can you point me to the "blind optimism or homerism" in the above?