What Twins Pitchers Get the Toughest Assignments?

We're approaching the halfway point of Rocco Baldelli's first season as Twins manager. It's not controversial to say it's going well. With the best offense in the majors and a sizable lead atop the AL Central, the Twins are near-locks to make the postseason and the definitive favorite to win the division.

If there's a problem though, it's been the bullpen. The Twins have used 16 relief-only pitchers this season. If you ignore Tyler Duffey and Trevor May's combined two spot-starts last season, then 16 is the number of relief-only pitchers they used in 2019 as well. (I'm leaving out the four position players who pitched.) As a whole, the bullpen has been mediocre to bad in most measurements of overall performance.

Still, looking at individual stats shows there are some bright spots. Just taking a glance at ERA gives one the impression that Baldelli has a few at least decent options in the pen.

(All stats through June 22.)

You may notice that the bars on the above graph are not ordered by ERA or alphabetically. It's sorted by total batters faced. (Pitchers with fewer batters faced than Zack Littell's 37 have even more unreliably small samples than we often get in analysis of relievers.) Sorting the graph this way shows us who on staff might deserve more innings than he gets right now. For example, Tyler Duffey has an ERA of 1.99. His fielding-independent ERA predictors (FIP, xFIP, etc.) suggest that it's not a fluke and that maybe he needs to pitch more often than Matt Magill and Blake Parker.

But not all innings are the same. Data repeatedly tell us players aren't "clutch," but that doesn't mean that some situations aren't more important than others. Fortunately, we have a way of measuring that. FanGraphs tracks the importance of each moment in a game using the Leverage Index (LI). You can read more about it here, but here's the short of it: A Leverage Index of 1 indicates a situation of average importance. Anything more than 1 is more important, and anything less than 1 is less important.

By looking at the LIs of the situations into which each pitcher enters, we can get a general idea of how much Baldelli trusts each of his relievers. A LI above the axis means that a pitcher, on average, has entered the game with more on the line; a LI below the axis means the inverse. FanGraphs defines high-leverage situations as those with LIs of 2 or more, and low-leverage situations as those with LIs of less than 0.85. The rest is medium-leverage.

(via FanGraphs)

Most Twins relievers enter games in fairly average situations, but we can see that a few have gotten bigger assignments. Trevor Hildenberger, previously a rock, pitched a lot of important innings until poor performance led to his demotion in mid-May. Fernando Romero, conversely, has been sheltered when he's taken the mound, with an average LI of just 0.08 when he enters the game.

Among relievers on the active roster, Taylor Rogers has seen the biggest spots. His gmLI is 1.73, significantly more than any Twins reliever this year other than Hildenberger. Only a fifth of the plate appearances he's pitched in have been defined as low-leverage, the lowest portion of anyone in the bullpen.

(via FanGraphs)

Unlike Hildenberger, Rogers has proven he deserves to pitch in those spots. With a 2.12 ERA through Saturday and a difference in strikeout and walk rates of nearly 25 percent, he's been the best relief pitcher on the team and one of the best in baseball.

Duffey and Mike Morin have also posted low ERAs but not received the same level of trust. Morin is understandable: He doesn't strike out many hitters, and predictive stats dislike him for it – even those that take into account his slightly above-average ability to generate ground balls, like xwOBA. But Duffey's ERA looks less like a mirage, yet he hasn't seen important innings. Is there something we're missing?

LI considers general elements of the situation, like how many runners are on base, the inning, and how many outs there are. Advanced metrics like wOBA consider park effects – how the ballpark influences offensive output. But what most baseball statistics don't do is consider players' opponents. We know that having a runner on third with one out and a one-run lead in the ninth is much easier with Jeff Mathis at the plate than it is with Gary Sánchez at the plate. The same is true for a bases-empty, two-out situation in the fourth with a six-run lead, even if the importance is lesser.

Along the same lines as wRC+ and OPS+, Baseball Prospectus has a stat called Runs per Plate Appearance Plus (RPA+), a measurement of overall offensive performance. Like other "plus" stats, RPA+ is scaled so that 100 is average, and anything greater than 100 is above average. BP not only has RPA+ for hitters; it has the stat for pitchers' opponents. That's important for our purposes, as it can show us how good the hitters are that Twins pitchers have faced.

(via Baseball Prospectus)

Tyler Duffey has faced hitters with the worst average RPA+ of any Twins relief pitcher. (Rogers and Littell don't have bars because they each have an oppRPA+ of exactly 100.) The quality of his opponents has likely influenced his low ERA to some degree. After all, looking at these relievers, you can see a loose but positive correlation between oppRPA+ and ERA: Those who have faced tougher hitters have generally given up more runs.

It's hard to tell how much hitters' ability has influenced Duffey's performance, however. Opposing hitters have only been 13 percent worse than average; give his ERA a 13 percent boost, and it's still a solid 2.25.

Even if an opponent adjustment doesn't explain how Duffey has pitched so well, this exercise does tell us more about how Baldelli handles his bullpen. If nothing else, we can see that not only is Rogers facing some of the toughest situations; he's facing some of the toughest hitters, too.

If there is a reason that Duffey isn't trusted with the best hitters or the most crucial situations, it might be that he's allowed four home runs in 18 appearances. That may not sound like a lot, but his 1.59 HR/9 is among the worst figures in this sample of pitchers. Perhaps that has stuck in Baldelli's mind and kept Duffey off the mound. However, his other numbers suggest he's much better than Parker, Romero, and Adalberto Mejia, his homer-happy peers in the pen, so hopefully with time, Duffey takes the mound in bigger situations.

Future analysis of how difficult it is for Twins relievers might be handedness. Rogers has faced more than twice as many righties than he has lefties this season. Traditionally, that's a much harder assignment for a lefthanded pitcher. Perhaps digging further can tell us who on the roster has had the biggest platoon advantages and disadvatages. (Rogers has excelled against both righties and lefties.)

Based on leverage and quality of opponent, however, we can learn some things about how difficult each Twins reliever's assignments have been – even if the main lesson is that Taylor Rogers is as good as we think he is.