Pitchers: Whose Stats Underachieved?

Luck, as Branch Rickey famously observed, is the residue of good fortune, and it seems to us that a lot of what we Fantasists do amounts to determining who’s been lucky and who hasn’t. This is the stock-in-trade of one genre of preseason Fangraphs article that we, for two, are suckers for: Player A, the article will assert, had bad (or good) Fantasy-relevant numbers last season, but a massage of those numbers or an examination of more granular stats suggests that his performance wasn’t as bad as (or was worse than) his Fantasy outcomes.

The closer look or the more granular stats, the article will continue, reflect the guy’s true performance, whereas the Fantasy numbers are artifactual, and largely produced by the guy’s luck. Since luck evens out, the article will conclude, the guy will do better (or worse) than people who haven’t looked closely at the numbers think, and will be worth more (or less) than the market thinks he is.

One of our relatively accurate forecasts of our rather pitiful 2016 season derived from this approach. At mid-season, we opined that Danny Salazar (first half ERA: 2.75) would decline sharply thereafter, whereas Carlos Rodon (first half ERA: 4.50) would improve significantly. And so it turned out. We reached these conclusions by asking: which starting pitchers, if any, were in the highest (i.e. worst) quartile of Batting Average on Balls in Play and Home Run to Fly Ball Ratio, and in the bottom (i.e. best) quartile of Hard-Hit Ball Percentage?

And which starters, conversely, were in the lowest quartile of BABIP and HR/FB and the top quartile of HH%? Our reasoning wasn’t abstruse: if a guy’s not getting hit hard, and yet is giving up a disproportionate number of hits and home runs, maybe he’s been unlucky, and if his only problem is that he’s been unlucky, maybe his luck will change. And, on the other hand, maybe the luck of a guy who’s getting hit hard but doesn’t yet have the scars to show for it will run out.

We wondered whether a review of 2016 from this perspective would produce anything interesting, and it does. Of ERA-qualified pitchers, two appear to have been unlucky. One of them is Rodon; even his good second half couldn’t neutralize the first-half numbers. The other is Dallas Keuchel, but it’s hard to say whether that information should lead you to do something with Keuchel that other people aren’t. Anyone who’s reading an article like this one in Fangraphs knows that (1) Keuchel won the AL Cy Young Award in 2015, but (2) had a severely disappointing, though not horrific, 2016, evidently as a result of his season-long struggle with (3) shoulder inflammation. If you think that Keuchel’s healthy, you’re not just going to pay for a luck-adjusted 2016 Keuchel; you’re going to pay for a decent shot at 2015 Keuchel. And if you think he’s not healthy, and that the best you’ll do is a luck-infused version of 2016, you’re not going to want him, except at a price that your league won’t let you get away with.

OK—that’s a dead end. Were any of the ERA qualifiers lucky? Yes—Julio Teheran. He was, to judge by Fangraphs WAR, the 24th most successful starting pitcher in the majors last season, and the market, which has made him the 27th most popular starter, envisions little or no regression. We don’t want him, because with ordinary luck, he could duplicate his performance of 2016 in 2017 and still produce numbers that look more like his 2015 (4.04 ERA, 1.31 WHIP). Sure, Teheran’s a good pitcher, and he could get a lot less lucky and still be a top-40 starter, but he’s going to cost you top-30 money no matter what, so why bother?

We found some other interesting stuff when we tweaked our requirements a bit. When we expanded the population to include all pitchers with 40 or more innings as starters, we discovered that Mr. Misfortune of 2016 was Dillon Gee. Again, we’re not quite sure what to do with this information. We’re not taking Gee, who (as noted in our Salazar/Rodon article) was the object of desire in what was quite possibly the single worst transaction in all of Fantasy Baseball last year. Moreover, there’s reason—even aside from the fact that he’s never had a truly commanding season in six years of trying—to be wary. He doesn’t miss many bats, so he needs some good fielding behind him, and the move from Kansas City (even the impaired version of the team that played for most of last season) to Texas, with Rougned Odor at second base and Nomar Mazara in right field, won’t help. His upside is that he finds his way into the (quite underwhelming) back end of the Texas rotation, and provides replacement-level stats. Still, if it’s the 45th round of your slow draft and you want a veteran starting pitcher who’s not an entirely lost cause, Gee is a possibility.

More interesting still is something we found when we looked at who’d been unlucky in the second half of the season. One guy: Jimmy Nelson. He was pretty good in the first half (3.92 ERA, 1.38 WHIP, 1.01 HR/9) and pretty bad in the second half (6.10 ERA, 1.71 WHIP, 1.62 HR/9). Our Fangraphs colleague Scott Strandberg wrote two articles about Nelson last season, one in May and the other in October. Both were carefully observed, thoroughly researched, and closely reasoned, and taken together accounted for Nelson’s initial success and subsequent decline by focusing (to simplify drastically but not distort Strandberg’s analysis) on Nelson’s pitch location, which was good in the first half and bad in the second.

Fair enough, but we wonder whether fortune had something to do with it as well. Nelson wasn’t immoderately lucky in the first half, but he came close. In the second half, it just rained pigs on him. True, he walked more guys, and gave up a bunch of home runs that would have been out of any ballpark in the country, not just Miller Park. But the facts are that (1) Nelson’s Hard-Hit Percentage declined significantly in the second half, going from 35.4% to 28.8%, and yet (2) his BABIP increased significantly in the second half, going from .279 to .327, and (3) even against left-handed hitters, always his nemesis, his hard-hit percentage decreased and his BABIP increased. We posit that, whatever else the 2016 Nelson did wrong, he was more unfortunate in the second half than he was fortunate in the first, such that a return to the modest usefulness of 2015 is in the cards. A reserve-round selection (his ADP is 514), or perhaps even a $1 bid, shouldn’t go amiss.

And finally, and perhaps most interestingly, we looked at relief pitchers. Most of what we learned was useless for present purposes: who cares that Antonio Bastardo was unlucky last season? But it’s worth reporting on one guy who, it appears, got lucky. That guy is Seung-hwan Oh, and it occurs to us that, in a strange way, he fits the profile of a kind of pitcher we’re wary of: the reliever of a certain age (Oh will turn 35 this season) who has one excellent season as a closer and then, having seduced everyone into paying top dollar for him, turns into a pumpkin. Yes, we know: before Oh joined the Cardinals last season, he’d spent 11 seasons being the Trevor Hoffman of Korea. But it’s not like the Cardinals don’t have anyone to take over if Oh turns out to be unreliable. At the moment, Oh is the fifth closer off the board, and there are at least half a dozen guys who are getting taken later that we prefer.