So What Do We Think About Bullpenning Now?

For most of this postseason, the primary story tying all these disparate series together has been the significant change in the way pitchers are deployed. After Andrew Miller’s dominance last October, aggressive bullpen usage has become the norm. The tone was set in the very first Wild Card game, when Luis Severino got one out but the Yankees advanced anyway, thanks to 8.2 dominant innings from their relief corps.

But now, here we are at the end of the month, and the two bullpens left standing combined to give up 15 runs last night. To be fair, the two starters combined to give up 10 runs, so they didn’t exactly impress either, but neither bullpen had any ability to hold any kind of lead last night. And the players entrusted with those opportunities just looked exhausted.

Coming into the series, the Dodgers were supposed to have a big bullpen advantage. Their relievers had been unhittable in the first two rounds of the postseason. The right-handed trio of Kenta Maeda, Brandon Morrow, and Kenley Jansen were just slamming doors on Arizona and Chicago, and with easy series wins, no one was pushed all that hard on the way to the World Series.

In this matchup against Houston, though, Dave Roberts has leaned almost exclusively on those three pitchers. Last night, the usage seemed to catch up to them.

With Clayton Kershaw struggling and the right-handed top of the Astros order due up, Maeda was the first reliever on whom Roberts called in Game 5. He’d been nails all month, his fastball sitting 94 and regularly touching 96 after having averaged something more like 92 mph in a starting role. He’d worked his slider and cutter off that pitch, giving hitters three tough and distinct looks. Here’s how Maeda attacked Jose Altuve upon entering.

Maeda vs Altuve
Pitch Velocity Result
Slider 85 Whiff
Slider 85 Ball
Cutter 87 Ball
Cutter 87 Foul
Cutter 86 Ball
Cutter 87 Foul
Fastball 94 Home Run

Maeda came in and tried to get Altuve out with his soft stuff, probably knowing he wouldn’t have his usual oomph after throwing 42 pitches on Friday night. After Altuve torched that last cutter — MLB’s pitch-classification algorithm calls it a two-seamer, but at that velocity, it was probably just a terrible cutter — he finally decided to try and throw one past Altuve. Instead, he only got it up to 94. Altuve, meanwhile, hit it 105 off the bat for a three-run homer.

That 94 mph fastball would be the hardest one thrown by Maeda all night. He threw a pair at 92 mph to Carlos Correa and Yuliesky Gurriel before going back to the slider to end the fifth inning, then threw 92 and a 93 to Josh Reddick in the sixth before attacking Evan Gattis exclusively with offspeed stuff between 80 and 87. Maeda threw 25 pitches last night; only five of them cracked 90 mph.

After Tony Watson got out of Maeda’s jam and the Dodgers retook the lead in the top of the seventh, Roberts summoned Brandon Morrow — whom he had declared unavailable before the game started, due to the fact that Morrow had pitched in every game in this series — to go after the top of the Astros lineup. Like Maeda, Morrow has been nails for LA this entire postseason, pounding high-90s fastballs and high-80s breaking balls past nearly every hitter he’s faced.

Here’s a complete log of the pitches Morrow threw last night.

Brandon Morrow’s Outing
Pitch Velocity Result
Fastball 95 Home Run
Slider 87 Single
Slider 87 Called Strike
Fastball 97 Double
Slider 90 Ball
Fastball 97 Home Run

A guy who has averaged 98 with his fastball this month topped out at 97 and got torched for four runs on six pitches. Credit to the Astros hitters, of course, but Morrow wasn’t throwing the same stuff that has allowed him to overpower hitters this postseason.

Finally, after the Dodgers insanely tied the game in the top of the ninth, Roberts once again summoned Kenley Jansen. Jansen, who threw one inning last night but had two days off before that, was pretty much his normal self for the first 25 or so pitches of last night’s outing. But his 27th pitch hit Brian McCann in the wrist, then he issued a five-pitch walk to George Springer, and then his 33rd pitch of the night was lined for a game-winning single by Alex Bregman.

Jansen walked seven guys during the regular season; in the postseason, his strikeout-to-walk ratio stood at 15:1 entering Sunday’s game. But last night, approaching a pitch count of 50 over the previous 24 hours, he nailed McCann and walked Springer consecutively, then threw Bregman a 92 mph cutter that ended the game. His stuff wasn’t as obviously impacted by recent workload as Morrow and Maeda, but we’re used to Jansen forcing hitters to get themselves on base. Last night, though, the Astros’ winning run reached, and then was put in scoring position, by hitters who didn’t have to swing the bat.

For the second time in this series, the Dodgers lost with Jansen on the mound. Houston is hitting .292/.346/.635 against the Dodgers bullpen in the World Series. And they probably still have more of their manager’s trust than Houston’s relief corps.

Fill-in closer Chris Devenski was terrible again while replacing Ken Giles, who has been terrible all postseason. A.J. Hinch tried to avoid his normal relievers so much that he asked Brad Peacock — who also wasn’t supposed to pitch on Sunday — to throw 39 pitches.

So, this all brings up the obvious question: if this is what you get for riding your bullpen in October, then are teams being penny wise and pound foolish by aggressively pulling their starters and treating every round like the World Series? Given what these relievers have done so far this series, is the hot new trend next year going to be just giving your starters the ball and trusting them to get six innings regardless of pitch count or times through the order? Has John Smoltz and his band of grumpy men been proven correct. Will bullpenning be remembered as a failed experiment pushed by nerds who don’t understand baseball?

I think the answer is still no. There’s simply too much data to illustrate the diminishing returns provided by a tiring starter against good hitters a third time through the order. Lest we forget, Kershaw walked both of the hitters he faced the third time through last night, Alex Wood gave up a home run to the one guy he faced a third time the day before, and Yu Darvish got pulled in the second inning on Friday because he couldn’t get anybody out. In Houston, these weren’t controversial early pulls like we saw with Rich Hill in Game 2.

The story here isn’t just that relievers aren’t getting the job done; it’s that most pitchers aren’t really getting outs this series. The Astros stuck with Justin Verlander in Game 2, and he gave up a couple of home runs that put his team in a hole. Kershaw gave up a four-run lead on Sunday, while Dallas Keuchel threw 86 pitches and got just 11 outs. Charlie Morton’s brilliant outing on Saturday aside, this hasn’t exactly been a series in which starting pitchers have made a compelling case to be pushed deep into the game.

This series hasn’t been solely a story of bullpen failures; it’s been a series of across-the-board pitching failures, due in large part to the fact that these are the two best offenses in baseball. Between the heat in LA and the dimensions in Houston, the first five games have been played in hitter-friendly environments.

But I do think that 2017 probably represents something like Bullpenning 1.0, and there will need to be adjustments made in future years by teams hoping to get to the end of the World Series without exhausted relievers blowing leads. And, perhaps to Smoltz’s chagrin, I think the most necessary change is to use more, not fewer, relievers.

In general, I think quite highly of Dave Roberts, and believe he’s an upper-tier manager in MLB, if not the very best guy out there. But I think it’s fair to question how he ended up with an exhausted, overworked bullpen by Game 5 of the World Series after the team enjoyed easy DS and LCS victories that gave them plenty of rest between each series. And the adjustment that probably needs to be made is the acceptance that, even in the World Series, there are still innings that should belong to a club’s depth arms.

The Dodgers are carrying eight relievers on their roster this series. Unless you watched the end of Game 2, though, you wouldn’t know it. In the team’s first five games, Josh Fields and Brandon McCarthy have combined to throw 27 pitches to eight batters. Ross Stripling has also faced eight batters and thrown 27 pitches. The lefty Tonys — Watson and Cingrani — have been used a little more regularly, but in short stints; they’ve been called in to pitch six times, but have only combined to face a total of 20 batters.

Those five pitchers have faced 36 batters in this series, despite one starter having failed to make it out of the second inning and two other games going into extra innings. Meanwhile, Jansen, Morrow, and Maeda have combined to face 65 batters and throw 239 pitches. This isn’t so much an eight-man bullpen as it is a three-man bullpen with some other guys who might get asked to get an out or two here or there.

On this point, I would agree with Smoltz; that’s not a sustainable model for keeping a pitching staff going. Unless your starters are just absolutely nails every time out, it’s very difficult to win the World Series with a seven-man pitching staff, which is effectively what Roberts has used. At some point, it’s necessary to trust the other guys, too.

The primary source of the Dodgers’ current bullpen issues stem from how Roberts utilized his relievers after Darvish got torched early. Clearly, there’s some logic to not just punting a World Series game down by a few runs early, but pushing Maeda to throw 42 pitches, and then using Morrow to pitch in the sixth and seventh innings when trailing by a pair of runs, seems questionable.

No postseason manager has been willing to his use B-team relievers to protect any size lead, so nearly every win involves the same few pithcers on the mound for the last few innings. In the 10 postseason games Jansen had pitched prior to last night, he’d entered with a five-run lead three times and a 10-run lead once. He also was asked to get multi-inning saves with a three-run lead on two other occasions. This isn’t just a Dave Roberts thing either; pretty much every manager in this postseason has used his best relievers to protect high-win-probability leads late in games rather than use those opportunities to get some work for the guys lower on the depth chart.

The result of using the best arms to protect any size lead, while also being unwilling to ride a struggling starter to a big early deficit, means that there just aren’t innings left to be categorized as low-leverage affairs. Even when Darvish got yanked in the second inning and the team trailed the whole way, the high-leverage setup guys threw 55 pitches while the supposed long guys — Stripling and McCarthy — threw 15.

The data supporting earlier hooks for starters isn’t supposed to encourage managers to pull them for just the same two or three guys regardless of score or situation. The theory of heavy reliever usage in October is based on data showing that tiring starters aren’t any more effective than even most team’s sixth- or seventh-inning guy, and that it’s unnecessary to ride a starter the third time through the order when a multitude of fresh arms could perform just as well. The problem Roberts is running into is that he’s not using his multitude of arms.

Now, obviously, no Dodgers fan really wants to see Josh Fields pitch in the World Series again after he threw BP in the 11th inning of Game 2, but even if he’s persona non grata after that outing, there’s been more room for work from Stripling, McCarthy, Watson, and Cingrani. And if the Dodgers had shifted some of the Game 3 pitches from Maeda and Morrow to that combination on Friday night, it’s reasonable to think that Game 5 might have gone differently.

Quick hooks for starting pitchers in the postseason probably aren’t going away, despite how much protestation comes from one-half of the Fox announcer team. But the only way quicker hooks for starting pitchers can work is if a manager deploys five, six, or seven arms regularly, and doesn’t just lean on a chosen few to handle all the innings previously thrown by those starters.

When a club is trailing early or has a big lead late in the game, a manager should use his depth arms. Save the best relievers for situations where they matter most. That’s how bullpenning is supposed to work: trusting that even the fifth or sixth guys on the depth chart can be as effective as a starter turning the lineup over the third time. This only-my-best-guys-at-all-times approach isn’t viable, nor does the data suggest it’s necessary.

Dave is the Managing Editor of FanGraphs.

Newest Most Voted
Inline Feedbacks
View all comments
5 years ago

While Roberts should probably use Watson and Cingrani more often, I think this is just a combination of tough luck and the Dodgers’ one roster flaw. Not only is Houston one of the best offenses in recent memory, but their best players are almost exclusively right-handed, and the Dodgers’ only reliable RHP before Jansen is probably Morrow. It’s tough to justify leaving decent LHP in against the murderers row at the top of Houston’s order, but obviously you don’t want Fields or Stripling out there either. The Friedman FO has been fantastic overall, but I do think they deserve a bit of criticism for not pursuing another RH reliever when it became clear that Pedro Baez was just running peripheral overperformance and couldn’t be trusted in the postseason

5 years ago
Reply to  aff10

That makes sense, but I think one of Dave’s points is that you can justify using a LHP against the murderer’s row sometimes, because using them means fresher RHP in other more high leverage situations.

5 years ago
Reply to  lesjcarter

You don’t even need to use the LHP against the murderer’s row. Neither Cingrani nor Watson are pure loogy’s like an Avilan would be, so you can very easily keep them in to face even the bottom 5 guys in the Astros lineup. Gurriel hasn’t murdered lefties in his MLB career (actually showing reverse splits, but those can’t be trusted), so you could easily plan a system where you use a righty for the top 4-5 guys, then go to Watson for the bottom 4-5.

ie. SP faces the first 18 batters of the game. Then Maeda comes in for 1-4 or 1-5. Then Watson on for 5-9 or 6-9. Then Morrow on for 1-4 or 1-5 again. Then Cingrani on for 5-9 or 6-9 again, leading way to an inning for Jansen whenever you get to that spot in the lineup.

Dave T
5 years ago
Reply to  Matt

What you are saying sort of looked like the plan for Game 2. Maeda came in with the top of the order due up and then faced 5 RH hitters, getting 4 of them out. Then it was Watson for McCann, but the pitcher’s spot was due up 2nd after Watson got McCann out. Where it seemed to get odd was in the 7th, where Stripling was brought in but given only 1 hitter once he walked that hitter (Gonzalez) and then Morrow was brought in the for the next LH hitter (Reddick) rather than just letting Cingrani start the inning

In the NLCS, Roberts would use Watson and Cingrani for 1-2 outs each, with a lead, both to keep Maeda away from LH hitters and to reduce the number of outs that he was expecting from Morrow and Jansen. Games 1, 2, and 3 of the NLCS all showed some of that usage pattern. Of course, that usage pattern also looks great after the fact when the pitchers who are put in good match-ups get basically everybody out.

5 years ago
Reply to  lesjcarter

but the thing is, there really haven’t been many low leverage situations. The closest thing was game 3 for the Dodgers. But you understand that they were so close and thought keep the pressure on them and eventually we’ll get this game. It turned into one of those fools gold games though.

5 years ago
Reply to  stever20

I have to say that in this series nothing feels like a low leverage situation no matter what the general stats are. The Astros’ 3-run lead at the beginning of the 9th felt very high leverage even before anyone got on base.

5 years ago
Reply to  stever20

There have been low leverage situations in the DS and CS rounds. Especially for LAD. There has definitely been some gratuitous overuse of Morrow and Jansen.

5 years ago
Reply to  aff10

Maeda became that other reliable righthanded reliever for the postseason. The problem is that Roberts ran him into the ground keeping Game 3 relatively close. Oops.

Nelson S.
5 years ago
Reply to  aff10

I dont really get the aversion to using Fields? He had 1 bad appearance in Game 2, ok, but generally hes been a pretty good reliever the past few years, including this year

5 years ago
Reply to  Nelson S.

Josh Fields is basically a TTO pitcher, and against the Astros at MMP I would rather have McCarthy pitching, at least he consistently induces soft contact.

Fields is horrifying in any kind of small or hitters ballpark because all of his batted balls allowed are fairly well struck fly balls and popouts, so he has a homer problem.

5 years ago
Reply to  aff10

I think the goal was for Walker Buehler to fill this role, but he was definitely exhausted by the end of the year and his control was never there at the MLB level, and then Maeda emerged into the role.

Really the issue is that there are no innings for the LHP against the Astros lineup.

Dave T
5 years ago
Reply to  mikejunt

As Matt says above, I think that the Dodgers LH relievers have spots against the bottom of the Astros order. The Astros’ lineups have had two LH hitters, Reddick and McCann, separated either by 1 batter in between them in the LA games (Gonzalez) or 2 batters (Gonzalez and Gattis) in the AL park games. That’s the part of the order where it looks realistic to try to use Watson or Cingrani to get some outs.

5 years ago
Reply to  Dave T

They have used them like that, but what has frequently happened is that they have gotten 2 out of 3 of those guys, let one on base, and turned over the lineup and forced Roberts to go to someone good for George Springer.

There’s no planet where Tony Cingrani gets to pitch to George Springer in the World Series with runners on base.

Even last night when Cingrani looked like he was turning in the best pitching performance of the night for the Dodgers, he then gave up a rando homer to McCann left on left which ended up being decisive in sending the game to extras.

They have have had those opportunities and they haven’t got full IP out of them most of the time.

Its true that they should be able to carve through Gonzales/Reddick/McCann at 7-8-9, but they just haven’t ever done that.

Dave T
5 years ago
Reply to  mikejunt

If we’re going by results, Watson has appeared in 4 World Series games and faced 11 hitters to get 10 outs, including a GIDP, with 2 singles allowed (1 of them an infield single) and 1 unearned run allowed. Cingrani has faced 12 hitters across 3 games to get 9 outs, including a GIDP, while allowing 2 walks (1 intentional), 1 run (on a HR), and an infield single.

The better comparison should obviously be wOBA allowed and FIP over much larger samples, and I agree that Watson and Cingrani are not as good as Jansen, Morrow, or Maeda. I also agree with keeping both Watson and Cingrani away from the dangerous RH hitters at the top of the Astros order in any close game.

Dave’s point still stands, though, that those two, plus Stripling and McCarthy, should be perfectly reasonable options to pitch in some combination when down in a game needing a lot of innings after the starter is pulled (Game 3) or when needing only 3 outs with a 5 run lead (Game 4). The alternative has proven to be pitching Morrow in every game this series, pitching Jansen in 4 of the 5 games with 2 multi-inning appearances, and bringing Maeda back with 1 off-day after throwing 42 pitches, usage patterns that have their own not surprising downside.