We like stories and second-half splits tell us one: pitchers are good early in the year but tire late. They fade.
Give me last year's trailers in ERA and WHIP during the second half and I will certainly find examples that seem to support this theory. If you are so inclined, you will bet against them. Except that when you look at three-year splits, you find what you'd expect to find if second-half (or post-All Star) splits meant nothing – the bad second-half pitchers are mostly the bad overall pitchers and vice versa.
Some people even bet that pitchers who improved last year in the second half will improve this year, too. I'm not sure what the story even is there. Okay, they don't tire. But how can they get better as the season wears on?
Let's illustrate with leaders and trailers, post-All-Star break from 2009-2011 (minimum 200 second-half innings pitched). We'll start with leaders first, even though, again, what's the theory here even? I'm not sure. If you are a believer, please share in the comments or tweet me @MichaelSalfino.
Just a bunch of good pitchers. If second-half stats were not arbitrary endpoints, wouldn't you expect to see some pitchers who just got better in the second half? Maybe you say, "Well, no, because it doesn't make sense to get better as the season gets longer; only worse." More on that later. But for now, let's just dismiss the notion of better second-half splits for, say, last season having any predictive value.
And if you're still holding on to the notion that some pitchers on this list have pitched better in the second half than in the first, drop it. Look, if we were to stipulate that second-half performance is totally arbitrary and we merely have to look at the quality of the pitcher over the biggest relevant sample size to project him, we would still expect some variance. We certainly would not expect identical performance for first half during the period and second half, right?
Now on to the second-half trailers from 2009-11 (three years). Are they mostly the bad pitchers, period?
Mostly, yes. Remember, Roberto Hernandez is Fausto Carmona. Also remember that Hammel was bad, period, until this year. Okay, Cahill is up about a run higher but he's a young pitcher and you can understand why guys prior to their mid-20s may not have fully developed man strength/stamina. For those who say, "Young pitchers present risk that they may fade as the season wears on," I can't argue. Gallardo also was 25 or under for the period. The Mets Jonathan Niese doesn't make the list, but same story with him (though this is his age 25 season, were we should stop worrying, generally). Shields and Beckett could definitely fit the story that they just get tired in the second half. But then why are their strikeout rates steady? Shields is off more than a run though. But last year, his second-half ERA was 3.35, lower than this first-half ERA for the period. Either way, I'm not going to hang an entire theory of projecting pitching performance on the thin reed of James Shields.
I provide this information because people want it and am doing so in a manner that I hope dissuades you from making transactions based on it. In fact, I encourage you to take advantage of your leaguemates who may overreact to the second-half meme. There are profits there to be made – with the possible caveat pricing more risk into younger, generally less physically developed hurlers.