102.2 miles per hour. That’s squarely in Aroldis Chapman territory. And as eye-popping as that is, there are a few key things you should notice. Firstly, had Aaron Judge not fouled it off, it would’ve been a ball. There’s no debating that. It wasn’t going to be in the zone for any umpire, even C.B. Bucknor’s. But before you tell me it was supposed to be there, let me bring up my second point: it’s way off target. Christian Vazquez sets up on the inside corner, and Kelly launches his firecracker up around Judge’s collarbone. Frankly, I’m surprised Vazquez had his glove up there in time, given the velocity and where he originally wanted it to go. Lastly, observe Judge actually fouling that off. If you’re throwing 102 that high and it’s still getting fouled off or hit into play, you’ve screwed up. They’re either counting on it or it’s predictable — maybe both! It’s not as bad as throwing 103 in the zone and watching your lead disappear, but still, it’s pretty bad.

That’s Joe Kelly in a nutshell. His velocity is there, the Stuff is Great, but the results you’d expect never materialized, due to either bad control or, like last year, a frustrating inability to strike guys out. Nearly a year ago to this day, Chris Teeter wrote about Kelly’s lack of strikeouts in these same webpages, and noted that Kelly was just one strike away from really becoming a reliable contributor in the bullpen. The ineffectiveness with two strikes really did show last year, as Kelly routinely could not turn those advantageous counts into outs. He was fine on 0-2, with a .171 opposing OBP on batters through that count. But anything more than that, and you’d be hard-pressed to find good numbers. Through 1-2 counts, over 30 percent of batters reached base. When it went to 2-2, that number jumped to 35. In full counts, Kelly allowed 25 of 51 total batters faced to reach base. He had 21 walks and seven strikeouts. That’s downright repulsive.

So far in 2018, Kelly’s been much better. Ever since the Opening Day implosion, Kelly’s faced 49 batters, allowed six hits with one(!!) walk, and struck out 15 of them while allowing zero runs, earned or otherwise. That’s good for a 30.6 percent strikeout rate, a two percent walk rate, and a 1.45 FIP over that span. And he’s doing all that while still throwing as hard as he ever was. It’s seemingly the same old Joe Kelly, but what’s really going on here?

Pitch selection is key, so let’s start there. Go back to that video of the 102 mph bullet Kelly threw. The ending clip of that is a good slider that Judge simply has no chance on. The slider wasn’t Kelly’s go-to secondary pitch for the vast majority of 2017 — he liked using a curveball that didn’t really do much of anything. It was fairly middling in terms of results. The slider, however, raked in the whiffs.

Other than August, his curveball was really bad at getting swinging strikes. Last June, he threw it 35 times and got zero whiffs, then last October he threw it 10 times with zero whiffs, and so far this month, he’s thrown it four times and gotten (surprise!) zero whiffs. At this point, you’re clamoring for him to use something, anything else. And so, Joe Kelly adjusted.

His slider became his best friend. Couple that with changeup that getting a swing and a miss 33 percent of the time and he’s actually got a pair of useful, potentially wipeout pitches. The changeup is an especially good development, since it seems like common sense to pair a fastball that sits 98 with a changeup — a pitch designed to look like a fastball until the last 15 feet — that averages an 11 mph difference between it and his fastball. With that kind of velocity, you could get whiffs on those two pitches alone. Add a slider in there that isn’t too shabby either, and that’s a potentially great pitch repertoire. At the very least, it’s shown to be quite effective so far in 2018, as his 11.8 percent whiff rate is the highest of his career, while his 72 percent contact rate is the lowest of his career.

It might just be a case of Kelly being more emotive, but he seems to be pitching smarter as well, since trying to beat batters on stuff alone isn’t always a good idea. Back during the Anaheim series, the Red Sox had a plan to exploit Shohei Ohtani at the plate, and Kelly followed suit, blasting a 98 mph fastball down and in for a swinging strikeout. Kelly tapped his head to acknowledge a smart plan of attack — maybe a plan by Vazquez, and then executed by Kelly. That same gesture made an appearance once more in Texas, during a nerve-wracking Joey Gallo plate appearance. He blew past Gallo with a — you guessed it! — 98 mph heater high, and did the same thing afterwards. Stuff alone won’t make you a good pitcher, but stuff in the right spot? That’ll get you places.

This isn’t going to be all optimism, obviously. Kelly has only thrown 14.1 innings so far this year, so we’re stuck dealing with a small sample size until, well, August. Then we’ve got to deal with a .212 BABIP, a 33 percent line drive rate, and a 42 percent hard-hit rate — a rise of eight percentage points from last year. The groundballs disappearing and him throwing first-pitch strikes less than 50 percent of the time are both something we need to keep eyes on. Everything here is more than a little foreboding.

But right now, he’s striking out and walking batters at career-best rates, and finally seems to be rounding into a really good and potentially reliable late-inning arm during a point in time where the Red Sox desperately need one. It’s about time Joe Kelly started, for lack of a better term, punching people out.

Header photo by Kelley L Cox — USA TODAY Sports

Getting your teeth cleaned. Scheduling an oil change. Filing taxes. Shopping for Valentine’s Day gifts. We all have things we put off until we absolutely have to do them. It’s part of what makes us human. But sooner or later, we realize that yes, our teeth really do need attention, our engine really is overheating, the IRS is real and our significant others will be mad even if they say they don’t want gifts (an especially painful lesson).

For years and years the Red Sox put off moving Joe Kelly to the bullpen. They knew they’d need to get around to it sooner or later, but by gosh work just kept popping up and they had prospects to trade and divisions to win, ya know? But 2016 was finally the season. And folks, let me tell you, when Joe Kelly is a reliever, my god does Joe Kelly Have Great Stuff™.

What Went Right in 2016

Joe Kelly in the bullpen went right. After six bad starts to begin the season, Kelly was banished to the minors in early June. He’d remain there for six weeks, emerging as a bullpen option on July 25. In his first outing as a reliever, Kelly allowed a run and coughed up two hits, but foreshadowing doesn’t always pay off. That was one of just two runs Kelly allowed in his 17.2 innings as a fireman, and John Farrell began trusting Kelly with higher-leverage situations as the season progressed.

In the end, Kelly posted a 1.02 ERA and held batters to a .203/.261/.297 line as a reliever. He struck out more than 30 percent of the batters he faced, his walk and homer rates plummeted and his fastball routinely touched the upper 90s. The stuff was great. The results were great. He looked great.

What Went Wrong in 2016

Joe Kelly in the rotation continued to go wrong. Those six starts we mentioned? Kelly allowed 21 earned runs in 22.1 innings. He walked 16 percent of batters faced. He gave up four homers. It was, to use a scouting term, woof city.

You can sort of understand why the Red Sox gave Kelly one more shot in the rotation. It’s easy to forget now, but he ended 2015 on a dominant eight-game stretch. Eduardo Rodriguez was on the DL. Clay Buchholz was Clay Buchholz. It’s not as though Boston was overflowing with talented starting options. Still, we knew this was probably coming. Scouts have thrown the reliever label on Kelly dating back to his days as a prospect. Sox fans have been screaming for this move since early 2015. It was past due. 

Outlook for 2017

The ninth inning belongs to Craig Kimbrel and the eighth to Tyler Thornburg, but the seventh inning could very well become the property of Kelly. You can’t expect him to be as dominant in the bullpen for a full season as he was for 17.2 innings last season, but it wouldn’t be shocking to see him develop into a talent worthy of serving as a primary set-up guy on a first-division roster. We’ll see if Kelly can hold up for a full season as a reliever, and we’ll see if his walk issues are truly behind him. If so, he’ll be a key piece in a suddenly deep Red Sox bullpen. If not, we’ll basically have two Matt Barneses (Barnesai?). That’s not exciting, but it could be worse.

I can’t help myself. pic.twitter.com/lRJtEpYoLx

— Ben Carsley (@BenCarsley) August 13, 2015

You know what we won’t have? We won’t have any more of Joe Kelly in the rotation. No more 24-pitch innings with three walks. No more 94 mph meatballs down the middle. No more Cy Young jokes.

Just Joe Kelly. Just Joe Kelly and his Great Stuff™.

Photo by Bob DeChiara/USA Today Sports Images

Want to guess how many innings Craig Kimbrel threw all series? 1 1/3. Yeah. So. Anyway.

When you take two occurrences: 1) Francona’s lethal (to the other team) bullpen usage and 2) Buck Showalter’s lethal (to his own team) refusal to use Zach Britton in an elimination game he eventually lost because no save situation had presented itself yet, it seems we are at one of those tipping points in baseball where before things were a certain way and now they are different. This, combined with the bullpen-based success enjoyed by the Kansas City Royals over the previous two seasons, means that we are going to see a change in the way teams value and presumably treat relief pitchers.

The point is relievers were already valued highly, but after this season, they’re going to be even more expensive to get. Kenley Jansen and Aroldis Chapman will both be free agents this offseason, but they’re going to get contracts that heretofore would have been unheard of for a reliever. The reality that signing either will likely cost the signing teams their first-round draft pick as well means these are going to be expensive additions to any ballclub.

Fortunately the Red Sox already have Craig Kimbrel. Whatever you think of the deal that brought him to Boston and the way he pitched at the very end of the regular season, Kimbrel is an elite-level closer and should be up to filling that role in 2017. That doesn’t mean the Red Sox can’t or shouldn’t sign Jansen, but it means they don’t absolutely have to. In fact, despite popular perception, the Red Sox are in a pretty good spot with their bullpen.

In fact, to prove this point, let’s look at how the Red Sox could build a good bullpen from players currently available to them. Any bullpen requires depth; the more capable guys you have, the better your team will be able to overcome the inevitable injuries that crop up during a 162-game season. That said, to avoid getting too far into the weeds, I’m going to focus on the opening-day roster. The Red Sox started last season with eight relievers on their roster, but now with the retirement of David Ortiz, it’s possible John Farrell decides to expand the bench by one for better flexibility.

Going into the 2017 season, the Red Sox are virtually guaranteed to have the following in their bullpen:

Craig Kimbrel (Closer)
Robbie Ross (LHP)
Matt Barnes (RHP)

Despite the way the postseason has played out, we shouldn’t expect anything more from Kimbrel than to pitch the ninth inning during the regular season. Ross was a bit under the radar this year but was an effective weapon against both right-handed hitters and lefties, holding the former to a .660 OPS and the latter to a .545 OPS. He’s a lefty but doesn’t need to be used only against his own kind, meaning he’s likely the seventh inning guy who can step into the eighth or sixth inning if circumstances dictate it. That opens the door to a different lefty specialist if the Red Sox want.

Then there’s Matt Barnes. He was markedly better this season than last, bringing up the strikeouts and dropping the home-run rate. There’s still a walks problem there, though, which makes him tough to trust in important situations. Yet given his youth and inexperience out of the bullpen, it’s not crazy to think there’s maybe another gear in there somewhere. Even if he only replicates 2016, Barnes is worth rostering.

That’s three. The fourth is probably Joe Kelly. Considering his success out of the pen at both the end of the season (14 innings, 20 strikeouts, 0.64 ERA) and in the playoffs (3 2/3 innings, three strikeouts, no baserunners allowed), it’s intriguing to think what he could do over a full season. Between he, Barnes and Kimbrel, the Red Sox have three guys who can hit the upper-90s with ease.

That’s a pretty solid top four, but that wasn’t in doubt. The trick is finding quality for the rest of the spots, as those guys are going to have to pitch too. Junichi Tazawa and Koji Uehara are both free agents. One or both could be brought back, but for now, let’s let them both go and see what we can do internally.

Although Robby Scott didn’t make the postseason roster, he did pitch well for Boston during the stretch run (six innings, five strikeouts, two walks, no runs). His time in Triple-A has shown he can get hitters from both sides of the plate out, but his funky side-arm delivery is probably best against mostly — if not exclusively — left handed hitters. Still, he’s a young guy who has had success and represents a different look than the high heat from most of the other bullpen members listed above.

There’s something about the way Heath Hembree throws. Whatever it is makes it near impossible for right-handed hitters to hit, but lefties all turn into Jose Altuve against him. Still, considering most batters are right-handed, and Hembree’s ability to give multiple innings in games where the result has already been decided one way or the other, he has value. If Farrell starts to shelter him against lefties a bit more than he has previously done, his results might even improve.

So that’s four top relievers and two platoon specialists. After that, the Red Sox have some choices. Fernando Abad could be tendered a contract, though if Scott is going to be on the major-league roster, the Red Sox don’t need another lefty specialist, especially one who projects to cost four-to-six times more. Roenis Elias is another option if the team wants to move him to relief full time. His potential in the pen is intriguing, but it’s difficult to count on a guy in a role he’s never played before.

And now having said that, I’m going to contradict myself. Kyle Martin isn’t on the 40-man roster, so he didn’t get a call-up in September, but the 6-foot-7 reliever was very effective out of the pen in Triple-A, striking out 78 in 66.2 innings. He has the control and command to back those strikeouts up with a reasonable walk rate, and he only surrendered five homers last season. Martin’s the perfect last reliever in that he could be good, or the Red Sox can easily send him back down without losing him.

In addition, Brandon Workman could be back and healthy at the start of the season. At some point, potentially around the first month or so of the season, the Red Sox could see the return of Carson Smith as well.

It’s undoubtedly true that the Red Sox bullpen would be improved with the addition of either Chapman or Jansen. Any bullpen would. But the cost, the loss of available funds to devote to other parts of the roster, the loss of a draft pick and the money associated with that pick, combine to make doing so significantly expensive. That doesn’t mean the Red Sox shouldn’t explore the free-agent market for relievers. They could bring back Uehara, Tazawa or even Brad Ziegler, any of which would move either Martin or Scott to Triple-A.

But the important thing is the Red Sox don’t have to do any of those things. The bones and flesh of a good bullpen are already here. With the departure of David Ortiz and the arbitration raises due to the team’s young stars, if there is money to be spent, it should probably go to a player who can contribute in the middle of the lineup with the bat, not one who will throw just 60 innings.

Photo by Bob DeChiara/USA Today Sports Images

Triple-A Pawtucket: Joe Kelly (RHP)

Hey. I’m here to talk about Joe Kelly. No I’m not thrilled about it either. He’s 28, so it’s not really a Fenway’s Future thing as much as it is a Fenway’s Please Help In The Present thing. The Red Sox bullpen struggles of late have been both completely predictable and borderline unbelievable, which would seem like a pretty hard duality to achieve.

Enter Joe Kelly’s Great Stuff. In 34 innings at Triple-A, Kelly has a 1.58 FIP and a 1.08 WHIP. He’s averaging 11.1 K/9 and 1.6 BB/9. If you want to be skeptical of how useful Joe Kelly can be for a major league team, fine – there’s more than likely a very strong argument there. But what do the Red Sox have to lose? He’s pitched well in Pawtucket and it’s not like the Sox are drowning in bullpen help. Here were are, guys: hoping Joe Kelly can save the bullpen.

Double-A Portland: Yoan Moncada (3B)

Ohhhhh it’s finally happening. I know that there are probably at least a dozen interesting prospects or storylines from Portland this season that should be talked about, but Yoan Moncada is getting called up to the majors so that’s what we’re going to talk about.

Travis Shaw is slashing .189/.268/.360 with a .629 OPS in the second half. Aaron Hill has been even worse, hitting .185/.267/.247 with a .514 OPS in the second half as well. Maybe he’s not as ready as everyone would like him to be, but he’s clearly not going to play everyday unless he is the next Andrew Benintendi (miss you, man). He’s proven himself a capable hitter in Portland, so it’s hard to see the risk in doing this. At the worst,  he’s a pinch runner, though his ankle injury may compromise his top-end speed for now. Anything more than that is gravy. And like many people have pointed out, even if everything goes wrong this September, at some point the Sox will throw out a lineup with Betts, Bogaerts, Benintendi, Bradley, and Moncada, and that’s super exciting.

Single-A: Rafael Devers (3B)

There is nothing new to add to the Devers narrative. He’s continuing to crush the ball in the second half. He’s hitting .335/.376/.567 with a .944 OPS in the second half, including hitting .474/.500/632 with a 1.132 OPS over the last week. He has seven home runs in the second half of the year, and at this point it any sort of struggles are a distant memory. He was also named a Carolina League Post-Season All Star, so that’s cool too.

Bonus Jason Groome Update!

He’s already been promoted to Lowell and that noise you hear is the sound of my expectations for him soaring to unrealistic heights. In his second pro outing he threw two innings while allowing two hits, one run and striking out five. He reported to Lowell yesterday and will make his first start for them on Friday.

Top Play (WPA)

Old friend Jose Iglesias hit a home run out of Fenway Park in the sixth inning (.285), the only real damage done to Drew Pomeranz all night. It was his fourth of the season – a season where he’s hitting .255/.208/.335 with a .240 TAv. The next three top plays all belonged to the Red Sox. You can watch it while reminiscing below:

http://m.mlb.com/video/v970368083/detbos-iglesias-drills-a-tworun-home-run-to-left/?query=iglesias

Bottom Play (WPA)

Mookie Betts, up in the 8th inning with the bases loaded and two outs, grounded into a fielder’s choice to end the inning (-.156). Not unsurprisingly, the next two worst plays were also outs in that inning, when both Bryce Brentz (-.113) and Brock Holt (-.146) struck out. Speaking in an entirely non-scientific manner, it certainly felt like the offense missed a lot of good pitches to hit last night.

Trend to Watch

The Clay Buchholz Showcase has begun. After being noticeably absent from the field for most of July, Buchholz has pitched three times in the last four games. After handling mop-up duty the last two appearances, last night he came into a two-run game. Is it a showcase to other teams or the Sox figuring out what kind of bullpen role fits him best? If Buchholz isn’t the backup plan, then who is? Henry Owens or Roenis Elias? Don’t think about it for too long. It’s all so bleak.

Joe Kelly pitched tonight.

It was predictable.

What’s Next

Steven Wright vs. Mike Pelfrey, that’s what. How the absence of Sandy Leon’s bat will affect the Sox offensive production is yet to be seen.

Photo by Bob DeChiara/USA Today Sports Images

Top Play (WPA)
This was a bizarre game, so of course the Red Sox had the most important play by WPA despite losing by four runs. In a third inning that lasted roughly an hour, Ryan Hanigan came to the plate with the bases loaded and two outs in a tie ball game. He smacked a single back up the middle to score two runs (+.183) and give the Red Sox a 7-5 lead. It’s in an entirely insignificant number of plate appearances, but Hanigan now has a wRC+ over 100 in high-leverage situations compared to a 36 overall mark on the year.

The top play for the victorious Orioles came all the way back in the first inning. Mark Trumbo came up with the bases loaded and one out with his team down 1-0 and he came through with a two-run, go-ahead single. It’s very odd for a back-and-forth game like this to have a top play come in the first inning for the winning team.

Bottom Play (WPA)

David Ortiz obviously comes through in the big situations more often than not, but it wasn’t the case in last night’s game. With the game tied at eight in the sixth inning and the Red Sox putting together what looked to be a big inning, David Ortiz came to the plate with runners on the corners and one out. One could’ve easily envisioned his second homer of the game to give the Red Sox a bit of a cushion. Instead, he hit a routine double play (-.176) to end the inning. Baseball is dumb sometimes.

Joe Kelly’s Leash Has to be Getting Smaller

Regardless of how you feel about his stuff, Kelly is proving time and time again that he’s just not a good starting pitcher. Anyone can struggle against this Orioles lineup, but that’s not an excuse for a guy who has pitched like Kelly has over the last couple of seasons. His command was all over the place, and he gave up hard contact in just about every at bat. Given how the Red Sox have treated him since acquiring him two summers ago, who knows how many more starts he’ll get. There’s not much reason to give him many more, even with the acknowledgement that the other options aren’t much better. Not long after the game, Kelly was optioned to Pawtucket. Which leads us to our next section.

Clay Buchholz Was Solid out of the Bullpen

Part of this has more to do with how low Buchholz’s bar has been set than his actual performance. He was by no means his former dominant self. With that being said, it was a solid outing that helped save the bullpen from being any more blown up than it was after Kelly’s short outing. Those four runs clearly jump out as bad, but it’s not as discouraging as it seems. He was mainly hurt by a poor sixth inning, for which he deserves plenty of blame after starting the frame off with eight straight balls. However, he was also undone by a rare error from Dustin Pedroia on what should’ve been an easy double play. He was then hurt by a small strikezone in the next inning. He’s not “fixed” by any stretch of the imagination, but I’d have more faith in him than Kelly at this point. The good news for the Red Sox is that they have two days off next week, so they have some time to figure this out.

Mookie Betts Does It Again

As bleak as things may sound from the above, the reality is that the Red Sox are still a very good baseball team. Much of that has to do with the offense, and much of that has to do with Mookie Betts right now. He hit two more homers on Wednesday night and became the first player in history to hit dingers in the first two innings of two consecutive games. Remember when he was off to a slow start? He’s now hitting .288/.335/.549 on the year.

Xander Bogaerts’ Hit Streak Continued

It’s now up to 25 games. There is no reason to read any more into the following observation: Bogaerts really cares about this hit streak. He’s visibly celebrating after he extends it every night, in a sharp contrast from Bradley’s nonchalant approach.

Coming Next

The Red Sox look to win this four-game set in Baltimore tonight with Rick Porcello taking on Ubaldo Jimenez. First pitch is at 7:05 ET.

Last Friday night, Kelly was building momentum. Through three innings in Toronto, facing the league’s most powerful offense in hitter-friendly Rogers Centre, he’d recorded four strikeouts and allowed just one run. Go back further and Kelly had six solid starts in spring training, posting a 2.63 ERA with 22 strikeouts and eight walks. Go back even further and Kelly had a fine second half in 2015, notching a 3.77 ERA and a 2.78 strikeout-to-walk ratio.

Then the fourth inning happened. Kelly faced six batters without recording an out, quickly erasing those positive feelings you were developing toward the righty. “Hey, this guy might be alright after all,” you were probably thinking, only to have your hopes dashed by a string of hard-hit singles, a hit-by-pitch, and a 387-foot Josh Donaldson exclamation point.

What went wrong with Kelly, as it often does, was his command—the stuff’s great, of course. Though the results were solid in innings one through three, much of his early success on Friday night could be classified as a mirage. Kelly threw a first pitch strike to nine of 14 batters in those innings, but during the 12 at-bats that went at least three pitches, he was ahead in the count (after pitch three) in just four of them. He was consistently falling behind against a good offense, and when that happens bad things are bound to happen. In the fourth inning, another bout of Kelly’s wildness caught up to him, as the Jays finally made him pay for missing his spots.

I examined Kelly’s outing on Friday night and classified each pitch into a bucket based on whether Kelly hit catcher Ryan Hanigan’s target or not—let’s call it command percentage.

Kelly started off with poor command and it trended downhill toward the end of his outing. In the third and fourth innings, Kelly hit his target just 22 percent of the time. In those two innings, just 15 percent of his fastballs went to Hanigan’s target. Without establishing the fastball early in the count, Kelly couldn’t stay ahead of hitters—and it’s not like his breaking pitches are of the get-me-over variety.

Whenever Hanigan called for a fastball up in zone, Kelly delivered one so far up that it served little purpose other than to put Toronto’s hitters in better counts—the up-and-in fastball that glazed off Kevin Pillar’s helmet being the most extreme example. Further, whenever Hanigan called for a slider down in the zone, Kelly threw a pitch so low that Jays hitters rarely offered at it, particularly when they were ahead in the count. In the end, Kelly’s struggle to get ahead in the count forced him more toward the heart of the plate, and that’s when Toronto did its damage.

The inevitable fourth inning hit parade provides a perfect example of Kelly’s downfall. Here’s a 2-0, 95 mph two-seam fastball to Russell Martin:

Hanigan wanted the pitch on the outside edge, but instead it leaked over the inner half allowing Martin to redirect it back at Kelley. Next up, here’s a first-pitch 96 mph two-seamer to the following batter, Ryan Goins:

Hanigan wants this fastball up, and he comes out of the crouch to exaggerate that desire. This time Kelly’s fastball ends up belt high and out over the plate, and Goins promptly rips it into center field to load the bases. After falling behind 2-0 with two sliders to the next batter, Darwin Barney, here’s what happened on pitch No. 3:

Hey, look, it’s a carbon copy of the first image. Again, Hanigan is set up on the outside edge of the plate, and again Kelly’s two-seamer ends up on the inner half, right in Barney’s wheelhouse, to the extent that Barney has a wheelhouse. We could go on, of course, and we will. Just one more.

Perhaps the most amusing (or disheartening) example of Kelly’s lack of command came in the third inning against Edwin Encarnacion, where Hanigan set up in the same spot—I mean, pretty much exactly the same spot—for six straight fastballs. Kelly came close to hitting the glove once:

The cluster of colored dots around Hanigan’s glove represent the approximate area where he set up on each pitch of the at-bat, and the dots of corresponding colors scattered all over represent the approximate area where each pitch ended up. On Kelly’s sixth and final try against Encarnacion (the white dots) he finally entered the vicinity of Hanigan’s target, but the rest of his efforts all missed by significant margins, one high and four others low and/or inside. He was probably lucky to escape with a walk.

Kelly’s maddening because he throws hard and—all joking aside—appears to have solid stuff. It’s just that most of the time he doesn’t have the ability to execute any type of consistent game plan. When the stuff is on, the command is off. When the stuff is off, forget it. There’s still hope Kelly can improve because he’s relatively young and he’s put up solid results in the past, but each passing disastrous start adds to a mounting pile of evidence that says maybe he just isn’t going to put it all together.

Photo by Bob DeChiara/USA Today Sports Images

Instead, they kept their faith that Joe Kelly is deserving of a major-league rotation spot. This will be a big season for the former Cardinal, who has the [editor’s note: great] stuff people dream on but has yet to put up the consistent production to go with it. In all likelihood, 2016 will be Kelly’s last chance at a starting gig in the majors. If he can’t prove himself as at least a consistent back-end starter, it’ll be time to see what he can do in a full-time relief role.

But let’s talk about that stuff for a minute. It’s fun (for some) to joke about how great his stuff is, but there’s no denying Kelly possesses an expansive and intriguing arsenal. Unfortunately, for all the pitches he has, the 27-year-old has turned into an incredibly predictable pitcher. Per Brooks Baseball, the righty has thrown his sinker over 56 percent of the time in his career, and he threw it 57 percent of the time in 2015.

Unfortunately, for all the pitches he has, the 27-year-old Kelly has turned into an incredibly predictable pitcher.

All that predictability leads to a version of Kelly that’s average at best, and below average much more often than that. Over the course of his career, Kelly has pitched 461.2 innings. In that time, he’s put up a 3.82 ERA with a lackluster 4.13 FIP, 4.43 DRA and 111 cFIP. He was more or less the same guy last year, with an inflated ERA and FIP but improved DRA and cFIP. Along with that, his strikeout rate rose from his career norms, as Kelly struck out 18.8 percent of the batters he faced.

Unfortunately, his swinging-strike rate didn’t rise appropriately. The 19.1 percent whiff rate he garnered per swing ranked 93rd among the 110 starters who threw at least 2000 pitches in 2015. That’s what happens when you rely so heavily on one pitch. To make matters worse, that one pitch didn’t do it’s most important job last season. Despite leaning so heavily on the sinker, Kelly’s ground-ball rate fell all the way down to 46 percent in 2015, his first season below 52 percent.

If he’s going to stick as a starter, something has to change. Everything detailed above describes a pitcher who’s desperately in need of adaptation. It’s clear to me that Kelly needs to become less predictable, which means relying more heavily on his secondary pitches. His sinker just isn’t effective any more. Major-league hitters adjust, and they whiffed on less than 10 percent of his sinkers last season. More importantly, the ground-ball rate on the offering fell all the way down to 44 percent last year. That reality right there should be the final straw for Kelly’s repertoire approach.

So, now the task is to figure out which pitch he should turn to. The simple answer is all of his offspeed offerings — a changeup, a slider and a curveball. Amazingly, he didn’t throw a single one of those more than 15 percent of the time last season, despite enjoying some good results. All three pitches induced grounders on at least half the balls that were put in play against them and garnered whiff rates of over 25 percent.

It wasn’t just last season, either. Kelly’s secondaries have continued that trend over his entire career.

It’s possible that Kelly has something special in his changeup, and he’s just completely neglecting it.

There’s one pitch in particular, though, that Kelly should be focusing on. It’s possible that he has something special in his changeup, and he’s just completely neglecting it. In 2015, he threw the offering only 11 percent of the time despite great results. To whit, exactly half of balls put in play against his changeup ended up on the ground, while batters also whiffed 31 percent of the time. Additionally, Kelly didn’t give up any home runs with the pitch.

If Kelly really wants to shake things up, he can start using it against right-handed hitters. Typically, changeups are saved to fight off a platoon advantage for the hitter, but Kelly is in need of a drastic change. As it turns out, he has acknowledged this himself in a conversation with Brian MacPherson. He mentions a new technique for the pitch, but really all he needs to do is use it at a higher rate. As MacPherson mentions in his piece, Kelly hardly ever threw the changeup against righties. In fact, he only did it 46 times last year — under four percent in total — and has only turned to the offering five percent of the time throughout his career.

When Kelly has thrown it, however, his changeup has been successful. Over his career, looking at the biggest available sample, righties have whiffed 31 percent of the time against Kelly’s changeup, while 59 percent of their batted balls have been on the ground. It is worth noting that there have been some home run problems involved, as righties also have a 22 percent HR/FB+LD ratio against the offering.

The obvious caveat is that Kelly’s changeup — and all his other secondaries, for that matter — might be successful because he throws them so sparingly. Once batters adjust to a new strategy, the results could be ugly. Nevertheless, he’s a boring and predictable pitcher right now, and the results haven’t been good enough. Batters know they can wait for their pitch. Despite hitting the zone at roughly an average rate in 2015, opponents swung at pitches from Kelly less often than all but two of the 113 qualified pitchers who threw at least 2000 pitches. On the plus side, this leads to a ton of two-strike counts. Unfortunately, largely due to his lack of secondary pitches, he had the 11th worst K/BB ratio in two-strike counts among hurlers with 200 such plate appearances.

All this brings us back to the same point I’ve been hammering home: Kelly needs to change something. The most obvious (and potentially most lucrative) chance would be a greater reliance on his secondaries. That can mean different pitches on different nights, but Kelly’s changeup should be the one he leans on the most, even against right-handed hitters.

There’s no guarantee it will work, but Kelly has little choice but to try something new. It’s time for him to make a change and swing for the fences with that impressive repertoire we’ve heard so much about.

Photo by Troy Taormina/ USA Today Sports Images

As I detailed earlier in the season, Boston’s in a rather stable position going into the 2015-2016 arbitration window. With most of the team’s most critical players either still playing out the string on league-minimum deals (Mookie! Xander! Hooray!) or in the midst of pricey open-market deals and extensions (Porcello! Pablo! Boo!), the Sox are a team without a whole lot of uncertainty going into the arb process.

Recently, Matt Swartz came out with his offseason arbitration projections over at MLB Trade Rumors. Predicting things is a terrible business, but just like the old BP Annuals, Matt’s projections are deadly accurate. Using these arbitration projections as a guide to what a player will end up with is almost always a safe bet, and a great way to manage expectations.

Today, I thought it might be fun to take a look at the team’s current arb-eligible players, and find out what tact the team may take when it comes to the offseason. Knowing what we know now — that the team will likely be looking to build a fresh bullpen and hopefully reload to leap back into contention — I think we can make a reasonable guess as to which players will return, barring a trade or three. And that’s especially true given that all seven players for which the Sox need to make an arbitration decision come from the team’s sketchy bullpen.

The Definites: Junichi Tazawa, Joe Kelly and Robbie Ross

These three relievers (well, two relievers and a should be reliever) are, perhaps, the team’s best relievers under the age of 40, and for that reason it is extraordinarily unlikely that the team would non-tender any of them. In addition, none of these players has the counting stats that become overvalued in the arbitration process: namely saves and innings pitched. Let’s try to break them down one-by-one.

Junichi Tazawa — Projected 2016 Salary: $3.3 million

Unlike most of the pitchers on this list, I think there’s a pretty solid consensus that Tazawa is both an above-average reliever and under-valued compared to the open market. Junichi is heading into his final arbitration season, and for a player with both his pedigree and time in the league, this is a great value.

More than any other pitcher in the Boston bullpen, Tazawa has a proven “true talent” ability. cFIP, which measures this, posits that Tazawa has a score of 84 for his big league career, which is solidly above-average. While his seasonal ERA and DRA were down in ‘15, he got BABIPed to death (.349) and had tough luck with his strand rate (71%). He should be a keeper, and be a fine late-relief option, if not a dominant relief ace.

Joe Kelly — Projected 2016 Salary: $3.2 million

Here’s the thing: if Kelly does somehow transform into a relief ace, then his arbitration cost is just fine. If Kelly continues to be used in the rotation, pretty much at any non-disaster level of performance, then his arbitration cost is fine. But if he’s a bullpen JAG (just another guy), then while 2016 may be an okay term, the team will almost certainly have to trade or non-tender him after next season.

As a swingman, well, you’re probably not very excited about Kelly. Neither am I. I’d love to see if his already-nice fastball picks up a few ticks in short work, and maybe he can mothball his not-so-nice curveball. We already know that Dombrowski has come out in favor of Kelly as a starter, which is fine, I suppose, but on a team that could use bullpen weapons and has good-ish starters galore, I’d like to see them consider converting him.

He’ll be fine. And either way, he’s likely worth the money.

Robbie Ross — Projected 2016 Salary: $1.1 million

Robbie Ross may have been the team’s closer and best reliever near the end of the season, but there’s no circumstance where the team wishes for him to be the ninth-inning guy in 2016. Ross is left-handed and okay, which is great, since no one else currently in the bullpen is both left-handed and okay. His overall numbers for 2015 are pretty average: a cFIP of 99 says that he’s about league-average in terms of true talent, and a DRA of 4.07 says that he about got what he deserved in terms of runs against. Of course, Ross seemed to improve in the second half of the season, and it’s possible that he’ll beat his 2015 numbers rather than fall apart.

Ross’s salary projection is $1.1 million, which is chump change for a slightly above-average reliever. Boston should end up paying this in a heartbeat, if they don’t figure out some sort of short-term extension instead. Going year-to-year on Ross is fine, but extending him on a value contract is great too, as the stability of having a solid ‘pen lefty is a nice thing to have. He’ll be back.

The Maybes: Anthony Varvaro, Jean Machi, and Ryan Cook

Anthony Varvaro — Projected 2016 Salary: $700k

I’m inclined to dump off Varvaro, but mainly because of his injury issue. To put a fine point on it, Varvaro’s flexor tendon tear, which took him out of the bullpen in May after just 11 innings, caused Varvaro to be waived, claimed by the Cubs, and then returned to the Sox after his injury was found to be more serious than anticipated.

Varvaro had been pretty good in the two seasons before coming to Boston, where he saw his walk rate and BABIP get much, much worse. But given his injury issues, I’m not certain that the Red Sox would need to offer him anything above the league minimum to retain him. Varvaro’s reverse platoon split is nice to have when your bullpen isn’t exactly stacked with lefty-killers, but there’s already been a lot of performance variance, and the injury issues are a red flag. I’d think the team could let him fly, and perhaps bring him back on a minor-league show-me deal instead.

Jean Machi — Projected 2016 Salary: $900k

Machi’s ERA in his two seasons prior to this most recent one is immaculate, but his 2015 performance was … not. It’s a long walk from his ERA from being in the mid-twos to five-plus, but Machi’s underlying peripherals didn’t change all that much. You saw how his Boston run was damaged by giving up lots of hits and homers, something that perhaps an improved Boston defense could help mitigate in 2016. At the same time, Machi has never had dominant stuff, and his career cFIP of 100 basically screams league-average.

Though Machi hasn’t looked sharp in his short time with Boston, his 1.3 WARP in each of the two previous seasons shows he’s got the potential to be better than just an okay bullpen piece — those are really good numbers. I’d expect somewhere between half a win and three quarters next season and, on a cost of less than a million, I’d keep that.

Ryan Cook — Projected 2016 Salary: $1.4 million

This is the toughest decision, for me at least. Even though Cook looked pretty good as recently as 2014, his 2015 was a hot mess. You can forgive his terrible eight-and-two-thirds innings at the big league level as a small sample size, but his run at Nashville before coming over to the Sox was pretty shoddy as well. However, his time in Pawtucket was very, very strong in a limited sample.

Cook has potential, sure, but I have to think a non-tender is the right move here. Yeah, a million and a half is chump change for this team, but it’s also an unnecessary risk for a squad that needs a complete overhaul. I wouldn’t be surprised at all if a non-tendered Cook finds his way onto a roster only as a minor-league contract guy, so the team should save the money and perhaps put it toward a less risky bullpen option. But I’m not a scout, I’m looking at the stats. The real questions are: (1) are there better options out on the market somewhere and (2) is Cook going to stay healthy AND effective?

My best estimate is that the team rolls the dice on the dicey Machi, lets Varvaro fly, and Cook is a bit of a wild card. The team’s internal scouts and talent evaluators are going to make judgement calls here, and either way they probably can’t go wrong on Machi and Varvaro. When you’re talking about $700k-$900k, arbitration doesn’t drag a player too much up from the league minimum, so risk is low.

Cook’s a different issue. The team acquired him from Oakland despite an awful run there, so they must see something in his work that makes them want to take the risk. Both his bad run in Boston and his good run in Pawtucket are small samples, so I have to imagine there’s something the team likes here, and they keep him, even though I’d personally be more skeptical. Get a guy like this on a minor-league deal, and save a million.

The No: Alexi Ogando

Alexi Ogando — Projected 2016 Salary: $2.4 million

Ogando’s 3.99 ERA belies just how bad his peripherals were last year. Alexi gave up 12 homers in about 65 innings (not good) and walked too many hitters (3.9 BB/9). In fact, FanGraphs’ FIP-based wins above replacement pegged him at -0.9 fWAR, which is only the second-worst career mark of any Red Sox pitcher … in the team’s history. DRA is slightly kinder, as Ogando’s DRA-based WARP put him at 0.1, or right around replacement level.

But whether’s he’s the FIP-centered disaster or the DRA-based replacement-level pitcher, it seems as if investing more money in Ogando isn’t the best option. It’s not like this was an exceptionally down season; his cFIP in each of the last three seasons has been over 100, meaning his true talent level is less than league-average by this metric. There’s little reason to invest money in a low-upside option in the bullpen when low-upside options are in no short supply. The difference between a Jonathan Aro and Ogando is probably only $2 million. He should be non-tendered as well.

Hrm.

So with my guesses on who gets non-tendered, paired with Matt Swartz’s projections, it looks like the Sox will have to drop about $8.5 million in arbitration salaries on four bullpen pieces … and Cook is a $1.5 million wild card, barring trades. Investing in this team’s bullpen is a necessary evil, and all three of Tazawa, Kelly, and Ross could very well be above-replacement options. They’re locks. From a context-free perspective, spending $10 million of five bullpen pieces is great! Look at all the money they’re saving!

However, with all these pieces ranging from uninspiring to shruggie-guy-emoticon, saving a couple million here and leaving Cook / Varvaro / Machi off the roster is less about saving the money, and more about freeing up roster spots that could go to higher-upside or lower-risk options from trade or free agency.

In the end, that’s what Boston’s bullpen arbitration decisions are all about. There are some no-brainers here, but Boston’s choices will be keeping around a good chunk of the existing bullpen at a discount price, or opening up cash and spots for some new faces.

Photo by Bob DeChiara/USA Today Sports Images

Going Deep

The Red Sox’s 2015 pitching staff has been a disaster. They rank 27^th, 23^rd, and 14^th by ERA, FIP, and DRA, respectively. It is clear that Dave Dombrowski has work to do this offseason to fix this significant flaw. While an established way to build a bullpen remains elusive, having a strong starting rotation can have positive effects on the group that relieves them. Last Friday, Matthew Kory examined options for improving the rotation that could come via trade, while John Tomase of WEEI.com assessed players that will be available to the Sox through free agency. The big names that will be available are Johnny Cueto, Zack Greinke, David Price and Jordan Zimmerman. To me, Greinke and Price are the most attractive of this group, but will likely come with the highest price tag, making them that much more difficult to lock down. If one of those marquee players is not signed, Dombrowski could aim to get someone (or a couple of someones) from the second tier, which includes Scott Kazmir, Jeff Samardzija, Doug Fister and Matt Latos. Ben Cherington not signing a true ace for the 2015 team was a focus for much of this season, and certainly in some circles is considered a primary reason for why the team failed this year, but that is not an accurate assessment and as such is not really a valid reason to avoid targeting second- or third-tier guys again. Adding an arm or two to the starting rotation would allow some players who are being considered for a rotation spot to be moved to the bullpen (e.g., Joe Kelly, Matt Barnes, Brian Johnson, Steven Wright), which should improve that aspect of the team without additional financial commitment. It is evident that there are many, many interesting decisions to be made by the Red Sox’s front office – whoever that may include – this offseason to piece together the 2016 roster.

***

One decision regarding the 2016 roster that appears to already be progressing is moving Hanley Ramirez out of left field. As you may have heard, Hanley has been an unmitigated disaster in left field this season. Moving him off the position allows the team to play Mookie Betts, Jackie Bradley Jr. and Rusney Castillo in the outfield full time, providing a huge upgrade defensively. Peter Abraham of The Boston Globe documented how Brian Butterfield and David Ortiz have been working with Hanley at first base. Thus, it appears clear that Hanley will continue his descent down the defensive spectrum for at least another season, before likely taking the designated hitter spot when Ortiz is done. There is considerable risk that comes with playing Hanley at first base, as he now has a defensive impact on more parts of the game than he did while out in left field. I recognize the risk, but I think it is the right decision. I do not mean to suggest that it will be an easy transition, but it should be easier. Having a healthy Hanley – he has reportedly been playing through injury for much of the season – in the lineup everyday is important for the Sox’s success, so hopefully this new move works better than the last one. An interesting part of the decision to move Hanley to first is that it leaves Travis Shaw in limbo for (at least) another season. Since being called up to the big leagues Shaw has done nothing but perform well: .301 TAv in 112 plate appearances (PA), and 1.3 fielding runs above average (FRAA) at first base. Hanley moving to first keeps Shaw without a clear everyday role on the 2016 major league team, but Jason Mastrodonato of BostonHerald.com suggests that Shaw is also an option at third base. However, having Shaw at third requires trading one of Hanley or Pablo Sandoval, which is another can of worms. For next year, Shaw will end up as a bench option on the major league club or spend the season at Pawtucket, and his role will be re-evaluated as part of the plan for the 2017 Red Sox.

Quick Hits

Transitioning to a new team, city, fanbase, and manager can take different periods of time for different players. Nick Cafardo of The Boston Globe writes that Rick Porcello is a player for whom the acclimation process took longer than expected, and that may have contributed to his early season performance woes.

Through Monday’s game Xander Bogaerts has a .405 slugging and a .090 isolated power, which are both considerably lower than was expected of him, but Red Sox hitting coach Chili Davis does not care. Tim Britton of the Providence Journal describes how the team is happy with Xander’s approach at the plate, in which he consistently makes contact and shows a willingness to drive the ball the other way.

Joe Kelly had a really nice month of August, earning a win in each of his six starts, which has Sean McAdam of CSNNE.com comparing Kelly to Pedro Martinez. While Kelly has been better of late, we should probably pump the brakes on him now being an elite starter. In his August starts, Kelly struck out fewer batters than is typical, walked more, induced fewer ground balls, and stranded a really high percentage of baserunners.

As mentioned, it has been a tumultuous month for the Red Sox organization, placing many people in difficult and likely temporary positions. Acting manager Torey Lovullo is one such person. But as Peter Abraham of The Boston Globe notes in his excellent feature of Lovullo, he has handled the change admirably and should be considered this offseason for any available managerial positions.

Three Good Game Stories

The Sox earned a victory over the Mets and Matt Harvey on Friday night, a win that Rob Bradford of WEEI.com notes provided one of the few positive memorable moments of the season.

The Red Sox let a sweep of the Mets slip through their fingers on Sunday. Sean Mcadam suggests that poor execution in the series finale brought the Red Sox’s spoiler ways to a temporary halt.

Eduardo Rodriguez and the Red Sox squeaked out a win over the Yankees on Monday night. Rodriguez, given extra rest coming into the start, was not at his best, but Michael Silverman of the BostonHerald.com reminds us that Rodriguez has pitched well against the difficult Yankees’ lineup three times in his young career.

Photo by Jason Getz/USA Today Sports Images