The Players Teach Us How to Start a Reliever

The first of two Wild Card games is scheduled for tonight. In addition to must-win baseball, this time of year is also typically marked by the appearance of a Dave Cameron piece on the merits of “bullpen-ing” a game — that is, the practice of using nothing but relievers in a single contest, of attempting to exploit matchups in order to maximize the chances of winning.

While the logic of “bullpen-ing” is sound in theory, it also fails to account for the comfort of pitchers who’ve potentially become attached to their roles. To get a better idea of how they might adapt to such an approach and how it might be handled in practice, I asked some actual players about it. Turns out, there’s a particular type of reliever who’s best suited to take the ball in the first few innings of a win-or-go-home game. And a particular type of pitcher who should follow him.

The first thing revealed by my inquiries is that relievers love the idea. “I’m down for whatever,” said Giants reliever Hunter Strickland with a smile. Nationals closer Sean Doolittle just laughed for a while. “Would I get paid like a starting pitcher?” he finally asked after the laughter had subsided.

Relievers would be fine with it because they’re accustomed to answering the call whenever. “We’re used to throwing in whatever inning, [if] not usually the first,” said Strickland. Added Miami’s Brad Ziegler: “I don’t think it would be very different for me, as much as it would be for the starter coming into the game [in the later innings]. His whole routine would have to change.”

And a starter probably would have to throw a couple innings in such a game — in order to reach a full complement of nine and still leave some arms for extras, that is. So the question is probably which kind of starter would adapt effectively to an otherwise unusual arrangement.

The answer? Probably a young one. Older starters are more married to their routines. “It’s very hard for me personally,” said Brandon McCarthy regarding the idea of starting a game in any other inning but the first. “My routine as a starter is fixed to the minute and a lot of guys are like that. It’s certainly not something impossible to deal with but could make a team nervous.”

Younger starters are those most likely to remember “starting” a game in the middle innings as a prospect. “I piggyback-started in the minors, and we saw that with rehabbing starters, too,” said Florida’s Justin Nicolino.

Added San Francisco starter Ty Blach: “When I was in High-A, Santiago Casilla was rehabbing and he threw an inning, and then I came in and basically went through my whole routine as a starter but came in in the second inning.”

If you talk to guys who’ve done both, though, they’re very concrete about how it would change their approach. “I’d probably go through a regular warmup as if I was starting the game, and then start getting loose probably after once the first out was recorded maybe, and then treat it like a regular start,” said Oakland’s Andrew Triggs, who’s done both. “Instead of being ready at 6:40 for a 7:10 I’d probably start getting ready at 6:50.”

Even after a few responses, a plan for the ideal bullpen game begins to emerge: start with whichever reliever is willing and then turn to a swing man or young starter ready to go next.

That’s only the beginning, though. It’s important to consider handedness, obviously. And there are ways to leverage handedness in a manner that could cause trouble for an opposing club. Like beginning with a lefty reliever and then turning to a righty starter after that.

“You put the other team in a tough spot,” said Doolittle of such a plan and what it would do to their lineup decisions. “Are you willing to wave the white flag for two innings?”

Teams that platoon regularly often put out completely different lineups depending on the handedness of the opposing starter. Look at the list of this year’s platoon-advantage plate appearances by club, for example, and you’ll see the playoff-bound Indians, Twins, Cubs, Dodgers, and Nationals near the top.

By starting a guy like Doolittle to begin a game against the Dodgers, the Nationals could potentially change the complexion of the opposing lineup drastically. If Los Angeles were to respond with a lineup full of right-handed batters, the Nationals could retaliate with Tanner Roark, potentially creating roster-tightening matchup decisions earlier in the game.

In order to take advantage of the shifting face of the lineups, it would probably be ideal to have a couple pitchers with big platoon splits, too. That would cause an opponent to manipulate their lineup to avoid as many near-definite outs as possible. “You’re assuming that the two guys you have out there have dramatic splits,” thought Triggs.

So, starting with a left-hander and turning to right-hander early in the game might have its advantages. On the other hand, it might elicit less change from an opponent than expected. “If I know that’s what you’re doing, then I just set my lineup for the right-hander. You’re only going to get one inning,” said the Marlins’ Dan Straily. “It’s no secret that if I’m starting tonight I’m not going six innings,” said Ziegler.

That’s good point. If a club saw that an opponent intended to throw an important reliever early in the game, they’d probably accept a low-leverage set of zeroes in the opening innings if it meant not having to face that same reliever later with the game on the line. So they’d let their righties face Ziegler and lefties face Doolittle, and look to the third inning and beyond.

If your intention, as a manager, is to simultaneously (a) play the most tricks on an opposing lineup and also (b) leave some good relievers in the pen for the later, high-leverage innings, it looks like the best option might be to start a game with a couple of pitchers who can go multiple innings but who throw with different arms.

Recent history actually present a situation in which such an arrangement might have been effective. In last year’s playoffs, the Nationals were headed to a Game Four, up two games to one, in the National League Division Series. Their nominal starter, Joe Ross, was young, right-handed, and boasted pretty much only a slider and a fastball, which should theoretically create a big split. (It’s certainly created a big split in reality, with lefties recording a .364 wOBA against him, righties a .268.) But the team also had a young lefty who’d started recently in Sammy Solis, who has held lefties to a .274 wOBA in his career.

That’s your ideal situation. Solis could have gone multiple innings and may have had an effect on the lineup early in the game. Ross could go multiple innings and blow it by any of those right-handed batters who started against Solis. You might get four or five clean innings if you put these guys in the right order.

Instead, Ross started the game, gave up four runs to get eight outs, and Solis got the final Dodger out in a loss. Adrian Gonzalez, a lefty, hit a home run off Ross in the first, who also hit two lefties and walked two more.

Who knows what could have happened. In any case, that game provides some idea as to the appropriate situation for such a thing. Which leads to a relevant question: do we have one this year?

Perhaps, but it may not come in the Wild Card game, because most of the teams in both leagues have an ace worth starting. Jon Gray, Zack Greinke, and Luis Severino are all among the league’s top-30 starters, at least.

What about the Twins, though? They don’t have a similarly effective option in the rotation. Could the Twins try something different by starting a lefty before Ervin Santana? Santana typically throws a slider and a four-seam fastball — two pitches that lefties like to see from righties — and has had very large platoon splits in the past. Perhaps they could force the Yankees into a tough situation, lineup wise, if they started Taylor Rogers or Adalberto Mejia. Rogers, for example, had held lefties to a .243 wOBA.

But it’s not a great fit. For one, it’s been a long time since Santana has started anything but the first inning. And for two, if you start Rogers, you’re starting a guy who’s gotten more than four outs just once this season and hasn’t started a game since two occasions in Triple-A at the beginning of 2016. And you leave Mejia as the only lefty in the pen. If you start Mejia, you’re starting a pitcher who cedes a great deal of talent to Santana.

If the Rockies win their Wild Card game, however, we’ll have the team best situated to manipulate the handedness of their first two starters, use young guys who have some length to them, and take advantage of the bullpen game in a way that fits the actual players on their roster.

Lefty Kyle Freeland could start the game. He has struck out 22% of the lefties he’s seen this year, against a 5% walk rate. Those numbers drop to 12% and 10% against righties, respectively, so he’s not ideal for a full game. Their lineup features short-stint lefties Mike Dunn and Jake McGee, so they can afford to offer up Chris Rusin for two well-timed lefty innings in the middle. And that way, German Marquez can take the second two innings and go to battle against mostly righties, whom he’s struck out 23% of the time against 6% walks.

It’s probably worth noting, in this particular case, that Marquez has shown reverse splits for his career. While we probably shouldn’t project a reverse platoon split for Marquez, it’s also possible that teams would set their lineups according to the observed split data. A pitcher to whom I spoke for this piece, Dan Straily, has recently shown reverse splits, as well. His comments shed some light on the matter. In Straily’s case, he attributes it to throwing “throwing both pitches to both sides of the plate,” adding that he’s actually seen reverse platoon lineups thrown at him. “I had seven righties against the Astros, and seven lefties against the Dodgers,” he said this year. The opposing team, in other words, may just put a bunch of righties in and keep them in for Marquez.

In any case, we’ll probably see fewer innings from our starters this October, as teams want to avoid their pitchers seeing batters for a third time.

It wasn't until 2016 that we really saw Oct pitcher usage change. In 2017, expect to see a ton of #bullpenning

MORE https://t.co/VBH3uceTUO pic.twitter.com/8aIBihpErt

— Mike Petriello (@mike_petriello) September 26, 2017

But the shape of the game may not differ wildly from past years this postseason — other than a greater number of short hooks. There are so many conditions that need to be met in order for a reliever to start a postseason game. As we’ve observed here, it’d probably require starting the game with a reliever capable of throwing multiple innings, replacing him with a presumably young and certainly flexible starter-turned-reliever, and then saving a lefty or two for the late-inning endgame.

“It’s a fun idea, that’s for sure,” smiled Doolittle. “Might work for one day,” thought Straily. “I’m down to try it, I don’t think it would happen,” said Strickland. “It would be bizarre, for sure. You’d have to hope it works out,” added Ziegler. If it didn’t, the blame would clearly land on the manager’s shoulders.

Maybe that’s why we haven’t seen more of this yet.