After an atrocious outing Wednesday night in which he allowed seven earned runs without making it out of the third inning, Joe Kelly doesn’t have too many valid arguments to make against his demotion to Triple-A. While he sports a 2-0 record, his 8.46 ERA in six starts is not acceptable for a major league pitcher.
The Sox have announced that they will call up 26-year-old Noe Ramirez to replace Kelly. Ramirez, who has pitched in nine big league games this year, will be a bullpen option for the Sox, since they have three days off over the next 12 and won’t require a fifth starter during that time.
“Obviously I didn’t see it coming,” Kelly said. “It’s a move that they decided to make and I’m just going to go down there and just continue to get better with commanding the baseball and work on all my stuff and try to get back as soon as I can.”
I did some analysis with the help of fangraphs to try to understand how Kelly got to this point, and whether he’s capable of returning to the form he showed late last season. I believe there are three main reasons Kelly has disappointed this year.
1. He’s been terrible on the road
This one really stands out; even though Kelly has been a better pitcher at home than on the road during his career (3.58 ERA at home vs. 4.52 ERA away), the disparity this year is absurd. It’s obviously a limited sample size, a warning that applies to all the following analysis, but Kelly’s ERA is 1.46 at Fenway compared to 17.10 on the road in 2016. If he eventually returns to the Sox, expect Kelly to “regress to the mean,” pitching a bit worse in Boston and a lot better away from home than he has thus far.
2. Bad control
Kelly has been getting behind in the count too many times this season, and he hasn’t exactly excelled in those situations. Overall, he’s thrown 56.1% strikes and 43.9% balls (267 strikes, 208 balls), significantly worse than his career totals of 61.6% strikes/38.4% balls. As opposed to last season, when Kelly compiled 110 strikeouts and 49 walks (with only 3.28 walks per nine innings), he’s walked 27 and struck out 19 (7.6 BB/9) this year.
According to Kelly himself, poor control is the root of his issues: “I wasn’t commanding the baseball at all,” said Kelly about his start Wednesday. “Obviously from the walks and falling behind hitters, it was just a lack of fastball command again where I wasn’t putting the ball where it needed to be. And from there, falling behind hitters and trying to get back into counts, and it just wasn’t happening for me.”
Most of the stats about Kelly’s stuff, such as the velocity of his fastball and the movement of his off-speed pitches, are similar to previous years. The only outlier is that hitters are batting .500 (17-34) against his two-seam fastball with a 89.7% contact percentage. His other pitches have been relatively effective, although he has generally relied on his fastballs, two and four-seam, which he has thrown 63.9% of the time. Though that’s typical for Kelly, the fact that he’s fallen behind in the count more often this season is a major reason why those fastball are getting pounded. In all non-hitter’s counts (0-1, 0-2, 1-1, and 1-2), Kelly has held the opposition to a .211 batting average or lower. When it’s 1-0 or 2-0, hitters average .370 and .476 respectively. Of the 107 hitters Kelly has faced that didn’t make an out or get on-base the first pitch of an at-bat, 60 have taken ball one, with just 47 first-pitch strikes. At Pawtucket, Kelly will undoubtedly need to focus on improving that ratio and taking command of at-bats before he places himself in situations advantageous to the hitter.
3. Bad luck
All that said, several stats suggest that, despite Kelly’s ample run support, his luck hasn’t been the greatest this season. 36.6% of balls in play against him have been hit hard, just a percentage point higher than last year. His breakdown of line drives, ground balls and fly balls-25.7%, 48.6%, and 25.7% respectively- is also very close to 2015 (when it was 25.1%/45.6%/29.4%).
The big difference lies in (1.) when those balls are being hit, and (2.) where they’re hit. In 2015, the opposition hit .286 with the bases empty against Kelly, compared to .255 with men on base. Hitters are still averaging .286 with the bases empty in 2016 against Kelly, but their average with men on base has skyrocketed .092 points, to .347. However, Kelly has actually conceded less hard-hit balls from the stretch than in the wind-up (28.6% vs. 44.4%).
This points to a larger trend, which is that a lot of balls not hit hard have found gaps against Kelly in 2016. In fact, hitters’ BABIP (batting average on balls in play) versus the Sox’ righty is .404, well above his career number of .300 as a starting pitcher. To illustrate his poor fortune, look at Kelly’s aforementioned stats for hard-hit balls and line-drives in 2015 and 2016. Kelly’s .319 BABIP last season is much more in line with what we should expect over the course of a full season than his current .404 line, given that he’s performing almost identically in terms of how often he allows hard contact and line drives.
John Farrell has to weigh that subpar luck along with Kelly’s lackluster control, his performance at Triple-A, and Clay Buchholz‘s showings from the bullpen over the next couple weeks as he figures out which of those two guys (if any) should re-join a somewhat shaky starting rotation.