Andrew Cashner and Theoretical Home Run Shenanigans

The Boston Red Sox and the Baltimore Orioles completed a trade over the weekend, with Baltimore sending pitcher Andrew Cashner to Boston in return for center fielder Elio Prado and third baseman Noelberth Romero.

At 28-65, the Orioles appear likely to be eliminated from the playoff race sometime in August. Andrew Cashner is a free agent at the end of the season, and even if Baltimore had a less implausible shot at the playoffs, it makes a lot of sense to get something in return for the right-hander while the getting is good. In this case, the getting is two very deep dives into the Red Sox organization. Prado and Romero are both 17 year-olds out of Venezuela. Neither player is anywhere near the top of the prospect radar at this point. To grab a couple of lottery picks, the Orioles agreed to pay half of Cashner’s salary to the Red Sox, a figure just a bit under $2 million. If either prospect works out, it won’t be a new experience for Cashner, who has been swapped for Anthony Rizzo and Luis Castillo in previous trades.

Cashner has had a decent season on paper, but the Orioles’ return suggests that there is a good deal of skepticism surrounding his 3.83 ERA and 4.25 FIP. The bump in Cashner’s peripherals in 2019 is at least enough for ZiPS to think of him as a one-win player. That’s par for the course for a fifth starter, and it just so happens that’s exactly what the Red Sox were in the market for. It isn’t something that will show up well in playoff projections, but remember that teams can no longer pick up major league-caliber fourth and fifth starter types in August, which means that teams ought to take more care to prepare for emergencies now. And pitchers famously have lots of emergencies. Brian Johnson is currently out due to an intestinal issue, and given that he’s been out for weeks, it seems to be something a good bit more serious than overindulging in spicy chili. He has thrown a couple of bullpens, but his trip to the IL creates some uncertainty, which isn’t a good state of being for a contending team. And Cashner is likely a safer below-average pitcher than Hector Velazquez.

Cashner has had a bit of an odd career, generally struggling to strike out batters when he’s not throwing 98 mph. Projecting his BABIP has been a hilarious nightmare, as he’s yet to put up a full season with a BABIP between .274 and .311. His .256 BABIP in 2019 is especially befuddling given that there’s no convenient defense-based reason, with Baltimore sitting at the bottom of the league in both DRS and UZR. ZiPS is understandably suspicious and without a sudden change in the characteristics of his changeup, I’m not completely buying his improved results with that pitch.

There is one thing that Cashner has always done quite well for a below-average pitcher: keep the ball in the park. Cashner keeps the ball down well without an explicit sinkerball and so far, batters have struggled to golf his low pitches out of the park the way they’ve been able to against, for example, Ivan Nova. That Cashner is on pace to allow fewer than 20 homers in 2019 is not a fluke.

To estimate home runs tables, ZiPS calculates something I call zHR (I’ll let you figure out what the Z stands for), which estimates how many home runs a pitcher “should have” allowed based on a variety of advanced data that has a predictive relationship with home runs. Some of the most significant factors are exit velocity data, pull tendency, flyball percentage, and fastball velocity. If you use only a mix of one-year zHR rate and one-year actual HR rate to predict next year’s home run rate, for pitchers from 2002-2018, the best mix comes out at 88% zHR rate and 12% HR rate. Home runs are volatile, but not hard to predict over the long haul; they’re a bit like the weather that way. In Andrew Cashner’s case, he has 120.6 zHR in his career, compared to 122 homers allowed. This measure also uses league and park factors as an input, so it’s aware of changes in home run environment.

Cashner’s performance in these stats has improved in 2019 compared to 2018; his velocity is up, his grounders are up, and batters have been less able to pull the ball against him. ZiPS thinks that Cashner should have allowed 11.9 homers this season, instead of his 11 actual homers. When the road to the World Series will likely involve beating at least two of the Yankees, Twins, and Astros, that’s a useful skillset to have.

League-wide, zHR can be calculated going back to 2002, the year that some of the data required came into public existence. Below are the largest one-year over and underachievers in expected home runs.

Just as with BABIP, there appears to be a bit of a ceiling for home run “ability” for basically competent professional pitchers. Dan Straily allowed 9.3% of batters to hit a home run this year, but even in this environment, it strikes me as unlikely that Straily would have actually allowed 70 home runs if given 150 innings. Home run percentages over five percent seem extremely difficult, historically, to maintain.

As you can see from the above list, the vast majority of over and underachieving pitchers subsequently allowed home run totals far closer to what was expected.

Looking at 2019, one can easily see why the Orioles are headed toward an all-time record for home runs allowed, with the pitching staff teetering on the unfortunate intersection of terrible and unlucky.

The Andrew Cashner trade is hardly a blockbuster, but in the end, both teams get what they want. ZiPS does not see this trade as one that will fundamentally change the AL East race by itself; the acquisition shifts Boston’s playoff percentage from 55.8% to 56.1%. It’s a useful move, just not one that should preclude Boston from making a larger score before the trade deadline.