Prior to the 2014 season, Tanner Roark won the battle with Ross Detwiler for the final spot in the rotation. Roark emerged as one of the better starters in the National League. Detwiler was essentially banished into mop-up duty. One year later, Roark has been relegated to the bullpen himself because of the Max Scherzer signing. He began the year filling Detwiler's former role in the bullpen, but injuries to Craig Stammen and Casey Janssen have left the Nats with a void in late inning leverage spots. Can Roark continue to serve as a part of the three-headed setup role? Given that the Nats would like to keep him stretched out in case of a long-term injury in the rotation, should he?
The next Ross Detwiler?
Let's start by talking a bit about Ross Detwiler and his performance/role last season. If you look simply at surface stats, they'll tell you that Detwiler didn't get off to that poor of a start. In 14.2 April innings, Detwiler allowed just three earned runs (1.84 ERA). Of course, that doesn't tell us the whole story. Detwiler was pretty decent in his first few outings. He didn't yield a run until his fifth appearance of the season. He would go on to allow three earned and five unearned runs in the first month of the season. This means that there were a few errors behind him, but not all unearned runs are created equally. Three of those unearned runs came on a home run in an inning where he pretty much got tagged, but there were two outs and Ian Desmond had made an error earlier in the inning, so none of the runs were earned.
Detwiler did have some problems in the bullpen from the get-go. He walked at least one hitter in each of his first five appearances, finishing with nine walks in just 14.2 April innings. He had a 1.64 WHIP, which sure didn't indicate that the 1.84 ERA was going to last for long (it didn't... 10.38 ERA in 8.2 May innings). Detwiler quickly went from being a versatile reliever that could throw multiple innings (or come in to face a tough lefty) to a guy that nobody in their right mind would want to see unless the Nats were either five runs up or five runs down.
He did end up settling in a bit the rest of the way, particularly during the middle of the season. In June and July, Detwiler had a 2.07 ERA and a 0.88 WHIP. Still, the damage was done. At this point, Detwiler wasn't seeing any leverage opportunities (per Fangraphs, 46.1 of Detwiler's 63 innings last season were in low leverage spots). It was going to take more than a month and a half of pitching well in mop-up spots to earn his manager's trust with the game on the line.
How are they treating Roark differently from Detwiler?
At first, Matt Williams wasn't really using Roark all that differently from how he used Detwiler. Roark saw three appearances in the first week and a half, spanning six innings. He then didn't see action for another eight days. In just over a month, Roark has had three instances where he's had at least four days off between appearances. He's thrown three innings or more twice and multiple innings three times. In two of those games, the Nats were already down by more than six runs when he entered the game... even if one of those games was the massive comeback victory in Atlanta. Essentially, Roark should have been considered one of the better relievers in the bullpen to begin the year... particularly because the bullpen was a bit depleted. Instead, he was used as an inning eating arm to save the rest of the bullpen when a few games had gotten out of hand.
Like Detwiler, Roark didn't exactly get off to a great start. He allowed a run in both his second and third appearances of the season, including a walkoff extra inning loss in Philadelphia. He allowed at least one run in four of his seven April appearances. He didn't have quite the walks problem that Ross Detwiler did (just three walks in 12.1 April innings), but get this.... Tanner Roark didn't strike out a single hitter this season until this past Sunday. There are some pitchers (Doug Fister comes to mind) who can get by without having a big strikeout rate, but it's difficult for any pitcher to get by if they're not fanning anyone.
Still, with the bullpen shaping up as it has, it was time for Matt Williams to try something different with Roark. Craig Stammen, who could have filled that role, is done for the year. Casey Janssen, who was signed in the offseason to be the primary right-handed bridge to Drew Storen, is about to start a rehab assignment after missing the first month of the season. Blake Treinen has been... hmmm... disappointing in the role, and that phrasing is just me trying to be nice. Aaron Barrett has been terrific, but he can't pitch every day. Matt Thornton and Matt Grace have been fine from the left side, but having another right-hander that can be trusted in leverage spots should help settle the bullpen down some.
Playing up in the bullpen
There's a theory that starters' "stuff" will often play up a bit when they move to the bullpen. They don't have to worry about saving their energy so that they can reach back for a big fastball when they need it. Instead, they just throw the big fastball the whole time since they're liable to only pitch an inning or two. We hadn't seen Roark do this early in the season. Fangraphs has his average fastball at 92.1 MPH, 1 MPH better than his average fastball in 2014 and actually a tad lower than his average fastball from 2013. However, in his three appearances in May, we're starting to see it......
Even though 5/15 is just a three game sample at this point, it's what we want to see. I don't know how to explain the fact that his velocity didn't play up at all in April. It's possible that he just took a little more time to find his rhythm than expected. It's possible that he wasn't really taking full advantage of throwing in shorter bursts and was attacking hitters/pitching with the same mindset as he did when he was starting. Given that he did throw some longer outings in April, it's possible he was trying to pace himself rather than going max effort and trying to dominate for an inning.
Either way, it's good to see that he's popping the fastball at 95 MPH in his past few outings. It's great to see that his velocity is up higher than it's ever been with the slider, too. That's been a huge pitch for him in his past few outings, and the uptick in velocity could be a factor. If he's focusing more on throwing in shorter bursts in late and higher leverage innings, this should become the norm for him for the remainder of 2015.
Pulling the trigger on more optimal usage
On Sunday, Williams decided to put Roark in a big spot. After Doug Fister allowed a one out double in the seventh inning of a 1-0 game, he called on Roark to work as a short reliever in a high leverage spot. Roark responded, suddenly having a lot more life on both his fastball and his slider. He notched his first strikeout of the season in that game, retiring the only two batters he faced.
Williams kept it going on Monday, calling upon Tanner Roark in the ninth inning of a two run game. Both Drew Storen and Aaron Barrett had pitched on consecutive days, so Roark got the call in a "save" situation. He responded again, striking out two, including Giancarlo Stanton representing the tying run to end the ballgame. That same life on his fastball and slider were there, and suddenly Tanner Roark was back!
Roark again got the call on Wednesday after Max Scherzer had allowed a three run homer to cut the Nats' lead to 7-5. He didn't look quite as overpowering as he had in his previous two appearances, but he still got the job done. Roark allowed consecutive one out singles by Justin Bour and J.T. Realmuto, but then struck out Ichiro Suzuki (tough to do!) and got a grounder to third off the bat of Adeiny Hechavarria to escape the jam.
Some Nats twitter stuff
I did see a little speculation on Nats twitter Wednesday about whether Roark should have remained in the game....
Thinking out loud and in generalizations here: was it really necessary to go to Storen? Roark threw 18 pitches. He can't go two?
— Red Porch Report (@RedPorchReport) May 6, 2015
Someday we'll live in a world where relievers are allowed to pitch more than one inning.
— Half St Heart Blog (@HalfStreetHeart) May 6, 2015
Initially I was going to call these overreactions, but if you follow the rest of the conversation, they seemed to be questioning the utility of relievers in general more than saying, "Roark should be out there instead of Storen." In truth, there was a little more to this situation than these two tweets bring up. Michael Taylor had an infield single with two outs in a one run game, bringing up Tanner Roark's spot in the order. Essentially, this leaves Matt Williams with two choices....
- He could let Roark bat, which pretty much means he's deciding that he's not really playing for another run to extend the lead beyond two runs
- He could pinch hit and go to a fresh reliever who is as good as Roark... or better, if you want to believe the small sample size we have so far this season. He'd also give himself a better shot at plating that insurance run
Yeah... That's not really a choice at all. Unless it's an extended extra inning game or the benches and/or bullpen are getting thin, you take the shot at that run and turn the game over to Storen. Having watched Williams manage games for over a year now, it's pretty safe to say that Williams would have gone to Storen even if the pitcher's spot hadn't come up in the bottom of the eighth. The fact that the pitcher's spot did come up kind of nullifies the argument, though.
Still, I'll take my crack at the argument. Roark was "meh" for about a month before his recent hot stretch out of the bullpen. I do believe it's real, but it's a small sample size. If we want to focus on small sample sizes (not saying they were), Storen pitched in three of the four games in the Mets series. Two were high leverage spots in what ended up being 1-0 saves. He didn't allow a baserunner in those three appearances... Roark allowed two in the eighth inning.
To respond to the question of relievers throwing multiple innings, those days mainly seem to be in the past. As the game has become more specialized, we see managers play matchups more and more often, which leads to shorter outings from the bullpen. This is a double-edged sword. While teams often end up going to inferior pitchers and use up their bullpen faster in individual games, it also means that most relievers have a shorter recovery time and are capable of throwing on consecutive days more regularly. Had Roark gone two innings, he would likely have been unavailable for a day or two (they're off Thursday, so it probably wouldn't have mattered that much). Since he only threw one inning, it's pretty safe to say he'll be ready if necessary for Friday's game against Atlanta. Considering that the Nats' bullpen is a little undermanned right now, I'd rather have Roark ready to go on Friday.
Back on topic with a conclusion
Close and late games have been a bit of a sore spot for the Nats this season. Williams has been emphasizing matchups a bit more over the past week or so, which should help to correct some of their issues. Having another reliever who looks comfortable pitching in these situations is important. I'm not nominating Tanner Roark for "eighth inning guy" since I've said many times this season that this is a position that simply doesn't exist. However, the more guys that Williams feels comfortable using in the eighth inning, the better. It's a lot easier to manage a bullpen based on matchups if you have a handful of pitchers that you're comfortable using. They should continue to use Roark in shorter (one inning) stints.
- His stuff will play up
- He'll stay sharper because he won't have those droughts where he goes five or six days without pitching
- The dearth of pitchers that look like they can handle leverage spots in the Nats bullpen means they need to add Roark to the mix of short relievers or continue to try other pitchers in that role who are either inferior to Roark or inexperienced at the big league level