clock menu more-arrow no yes

Filed under:

The Yankees bullpen has suffered from some bad luck

New, comments

That doesn’t excuse poor performance, but there will likely be positive regression as well.

MLB: Spring Training-Philadelphia Phillies at New York Yankees Kim Klement-USA TODAY Sports

I’m not here to excuse what has been a pretty awful period of time for the relief corps. Pitching to the tune of a 4.48 ERA, the eight-highest in baseball, and allowing the fourth-highest HR/FB% of 16% doesn’t exactly warrant sympathy, but hear me out.

One thing that has perplexed sabermetricians is the ERA-FIP differential. This was somewhat cured by the introduction of Deserved Run Average, which sorts through the “fault” of contact events fairly well, but unfortunately that data is not available until later in the year.

There’s a familiar adage based on the older metrics: If a pitcher’s context-neutral events are going well, but context-dependent ones are fairing poorly, the future performance will likely be closer to the former than the latter. If you take what you intuitively know about the bullpen, and look at said differential, things don’t add up. Consider the chart below, where the Yankees’ node is selected:

That’s 40 points! What’s even funnier is that not only is the difference so great, but it is tied for the lowest FIP- in baseball (70).

That doesn’t mean the bullpen is faultless, obviously. They have the third-highest hard hit-rate in baseball, and some have been objectively awful. Dellin Betances has looked totally lost, and Tommy Kahnle had a poor performance that was the prelude to a disabled list stint.

Yet there are still even more context-dependent factors that just cannot continue based on the group’s true talent level, which is sky-high. Here’s how they rank in some other context-dependent categories:

  • LOB%: 29th (59.3%)
  • BABIP: 2nd (.331)

There are also some other, context-neutral statistics:

  • K/9: 1st (13.43)
  • BB/9: 20th (3.50)
  • WPA/LI: 12th

The last one measures “context neutral reliever wins,” which does not include leverage. This is likely a better indicator of talent level than merely looking at who blew a certain amount of games.

There’s a bit more. Not only is the relief corps tied for first in fastball velocity at 94.6 mph, but they are also 26th (!) in fastball percentage. Not only are they throwing hard, they are also taking a deliberate strategy that separates them from every other staff in the league.

That being said, it doesn’t mean it’s set-it-and-forget-it. Bullpen performance, past tense, ultimately revolves around how well you did in high leverage situations. It’s the pyramid game of relief, and it’s why even though conceptions of leverage have changed — from closer/setup/seventh inning guy/etc. to “firemen” — the hierarchy is still the same. If you don’t produce, you find yourself on the Scranton shuttle.

Yet based on this information, it’s nearly impossible to imagine a scenario where they don’t improve demonstrably. An incredibly high strikeout rate, low walk rates, and blazing fast fastballs almost always means an effective bullpen. Injuries happen, so this is not a fixed state, and relievers’ talent levels are notoriously hard to pin down. All signs, though, point to progress as the Yankees head further into the season.