The Case Against Marco Estrada, Reliever

Way back on November 1, the Blue Jays began the offseason by sending Adam Lind to the Brewers in exchange for Marco Estrada, a move many (including yours truly) found puzzling. After all, Estrada had bombed out of the Brewers' rotation in 2014 after giving up 27 home runs in 107 innings and posted mediocre peripherals though good results in the bullpen. The Jays were returning their entire 2014 starting rotation, so Estrada was destined for the bullpen. Worse, he was projected to make $4.7-million through arbitration in 2015 by MLBTR, a pretty big hit against the expect remaining budget.

Shortly after, Nick laid out options the Jays had: put him in the bullpen; move J.A. Happ (quite prescient indeed) and use Estrada in his stead; trade him; or non-tender him. At this point, non-tendering him is moot, and a trade is unlikely. Since it appears he will be on the 2015 roster, under contract for $3.9-million, the question is then bullpen or rotation, especially now that there's an open spot which appears to me wide open for the taking.

Of course, the even more important question is: can Estrada provide value in either capacity? Two years ago he seemed to have emerged as a decent mid-rotation starter, before going significantly backwards in both of the last two years. At the very least, it's not intuitively clear that he belongs in a major league rotation. But the root causes of that decline also didn't appear intuitively obvious to me either. He didn't suddenly lose velocity, or forget how to throw his signature change-up. Was 2012 a fluke? Or maybe 2014? Something in between? This I leave for another day, as today we'll dive into his potential as a reliever (though normally the bullpen is a fallback once you've established whether a pitcher is a viable starter).

Estrada doesn't fit the profile of a power reliever, as an undersized righty with a fastball in the low 90s. He does have a strong secondary weapon in his change-up, though even that goes against the typical power breaking ball secondary (not to suggest this is a demerit; it's not, just different).

Most of the enthusiasm for Estrada as a reliever seems premised on his performance in Milwaukee's bullpen in the second half of 2014. In 43.2 innings, he allowed just a 2.89 ERA, 2.81 FIP and .217/.261/.350 triple slash against, all excellent marks. Unfortunately, on closer scrutiny, things aren't so rosy:

In relief, Estrada did cut back on his walk rate (and more in line with his career rate), but his strikeout rate was no better. He generated fewer ground balls too, but considering the sample size nothing extraordinary. But there was a stunning difference in the rate at which he allowed fly balls on home runs. With a league average of 9.5% of fly balls leaving the park in 2014 (Miller Park is friendly, so a park adjusted average would be more like 10-11%), Estrada was hammered as a starter, but as a reliever barely touched. It is widely known that this statistic is really noisy, and so we should not expect this gap to persist. Moreover, intuitively it doesn't make sense that going to the bullpen should flip a magic switch that lowers a pitcher's home run rate five-fold.

Looking at xFIP - which normalizes this difference - there's a much smaller difference between Bullpen Estrada and Rotation Estrada, mostly driven by the difference in walk rate. In fact, Bullpen Estrada's xFIP of 4.03 would be pretty mediocre, at best that of a middle reliever. Something like the 5th guy in a team's bullpen, not really worth $3.9-million especially for a budget constrained team. But that's only 2014, and over his career Estrada has 130 relief innings (vs. 411 as a starter), enough for a meaningful comparison:

On his career, Estrada's ERA is slightly higher as a reliever than a starter, though his peripherals indicate he's better from the pen. Or more accurately, one peripheral, his home run rate. Curiously, though most pitchers have better strikeout rates in relief, Estrada gets no bump. He actually walks more in relief too, though this may be noise (I can't think of a good reason he should walk more in relief). He doesn't get an improvement in ground ball rate or BABIP either. The only improvement is allowing a lower rate of home runs, 9% as a reliever compared to 13% as a starter.

Some of that difference is undoubtedly real: league wide, relievers do allow home runs at a slightly lower rate, about 1% over the last 5 years. So unless there's some inflection point special to Estrada, then even this difference seems mostly ephemeral. Basically, Estrada is almost the same pitcher in relief that he was a starter. At any rate, the difference is much less than for the average pitcher from starting to relieving. And most importantly, as a reliever he has not been very good, certainly not a guy to be pitching important innings.

The typical candidate to convert from starter to reliever is a pitcher whose stuff can tick up in short outings, and usually a pitcher who struggles with control or only has one good secondary and therefore has platoon splits. In the pen, their strikeout rate is much higher, they can generate weaker contact, and a higher walk rate is less important. A great example is Brett Cecil:

How did he do it? His stuff is dynamite in relief, as is obvious looking at the trend in his velocity at Brooks Baseball. It's clear when he went to the pen, and what a difference. You can see the same for other pitchers who became much more effective in the pen: Wade Davis, Andrew Miller, Zach Britton as some of the best examples.

This isn't to suggest the Estrada has to be as good as those guys to provide value. But looking at his velocity trend, there's not nearly the same kind of spike - or really much of a spike at all. Combining this with his historical results, unless he suddenly does spike up or find some sort of new trick (either would be unexpected, though neither unprecedented), acquiring Estrada to be a $3.9-million reliever makes little sense. Even for a team as lacking in bullpen options from the right side, the value is not there because he's at best a middle reliever.

In the second part, we'll look at Marco Estrada as a starting option. Happily, in that respect I see far more reason for optimism.