/cdn.vox-cdn.com/uploads/chorus_image/image/54978169/usa_today_10072468.0.jpg)
Some of life’s toughest questions are ones you never thought you’d have to answer. If you had asked me a year or two years ago if Joe Kelly is good I would confidently say absolutely not. Now, I’m not so sure.
Kelly has been a fairly important part of the Red Sox’s bullpen this season. After he failed to stick as a starter in Boston, his live arm and sometimes overpowering stuff made him an easy player to slot into the pen. The returns last year when the shift was made certainly show that was the right idea. He posted a 1.02 ERA in 17 2⁄3 innings, striking out 21 batters along the way.
This year, as Craig Kimbrel strikes out everything that moves, Kelly has been the second-best reliever for the cause, by some metrics, but is he actually good? His 1.61 ERA says yes, but there are a lot of other numbers that would indicate, if not the opposite, at least that Kelly isn’t the right-handed Andrew Miller. Plus there is the always non-scientific eye test. When I watch Kimbrel, I am confident. When Kelly is out there, I’m just not. I could get there, sure, but it will take some more time. So let’s try to answer that question I posted above: is Joe Kelly a good setup man?
The arguments in Kelly’s favor are easy to find, especially to the non-discerning eye. A 1.61 ERA is excellent in a game that is all about run prevention. In addition to stopping runs from scoring, Kelly is posting favorable numbers in ground ball rate (57.8 percent) and he has yet to allow a home run in 19 2⁄3 innings.
However, some of that has been luck. His left-on-base rate (83.3 percent) is fairly high and has been helped by a lowly .234 batting average on balls in play. Its likely that number will go up at least some especially considering his career mark (.300) is well above his current one.
:no_upscale()/cdn.vox-cdn.com/uploads/chorus_asset/file/8588817/669170936.jpg)
But batters have not been flailing at Kelly’s offerings and have been patient with a hurler that can reach 100 on the gun. He is walking 3.6 batters per nine innings, which is down from his weak 2016 campaign but still higher than you’d like to see. In addition, despite having such a fiery arsenal, Kelly just isn’t striking anybody out. Last year, when he struggled to a 5.18 ERA, he was ringing up 10.8 batters per nine. That number has nearly been chopped in half, while his groundball rate has grown by roughly 13 percent, suggesting a possible change to Kelly’s approach.
Some apparent changes to that approach are the righty’s increased usage of his slider. He’s thrown his slide piece 20 percent of the time this year, up slightly from 18.1 percent a year ago. It’s his second-most utilized offering, after being third in the pecking order in 2016. That change is nothing compared to his two-seam usage. A pitch he used primarily from 2012 to 2015, Kelly threw it 161 times last season, half as often he threw a generic fastball. However, this season, he’s tossed 144 two-seam heaters, while a garden variety fastball has been fired 69 times. (#nice).
However, shifting up his pitching choices hasn’t led to a big uptick in soft contact. Only 10.9 percent of his pitches hit have been categorized as soft (per Fangraphs). Meanwhile, his medium and hard hit percentages have both risen from last season, although the change in medium contact is pretty marginal. Adding further to the confusion is the fact that Kelly has an excellent FIP (2.90) and solid results in xFIP (3.82) and DRA (3.65) outside of that realm.
So, with all the data we’ve just looked over we can say a few things. Kelly is not striking anybody out, giving up far more solid contact that weak contact, walking too many and using his slider more often. Yet somehow he is still sporting an All Star caliber ERA, has yet to give up a home run and, to steal a line from Jonah Keri, killing more worms than a year ago.
With a Red Sox track record that has not been particularly strong, the logical conclusion to make is that Kelly’s two-month surge is smoke and mirrors and as that BABIP goes up, even with an accompanying K/9 boost, Kelly will more resemble the pitcher we’ve become used to seeing. However, there’s no certainty in that, especially considering we are now looking at nearly 40 innings of impressive relief work if you count last season.
When I started writing this post I was out to find an answer to a question. Unfortunately I have not reached any conclusion other than Joe Kelly is one of the most confusing relief pitchers in the game. For now he’s also one of the better ones, at least by simple results.