The AL Wild Card Game Should Be a Bullpen Affair

If you’ve been reading FanGraphs for a few years, you probably know what’s coming. Since MLB instituted the one-game Wild Card play-in contests, I’ve been a loud advocate for the idea of abandoning the traditional starter/reliever pitching format for these winner-take-all games. After all, relievers are more effective on a per-batter basis than starting pitchers, and most teams don’t have the kinds of starting pitchers who can justify being left in the game a third time through the batting order, especially when their team’s season hangs in the balance.

The National League this year is an exception, though. The Cubs and Pirates both feature legitimate aces, and if you have Jake Arrieta or Gerrit Cole, it is perfectly reasonable to go with a more traditional approach and let them attempt to dominate their opponents. After all, hitters posted a .540 OPS against Jake Arrieta the third time through the order this year; Cole was at .630. Those are the kinds of pitchers that you can comfortably ask to go six or seven innings even in an elimination game without putting your season at risk while a clearly better alternative sits in the bullpen.

But in the AL Wild Card game on Tuesday, the conditions are perfect for both teams to manage their pitching staff in unconventional manners.

The Astros certainly have an ace of their own in Dallas Keuchel. Like Arrieta and Cole, Keuchel was excellent against hitters as the game wore on this year; his .553 OPS allowed against hitters the third time through the order was fifth best in baseball, in fact. But because the Astros were playing for their postseason lives over the weekend, Keuchel last pitched on Friday night, throwing 99 pitches in the team’s 21-5 win over the Diamondbacks. That means he’ll be taking the mound tomorrow on three days rest.

Dallas Keuchel has never pitched on three days rest before, and the history of pitchers going on three days rest in the postseason isn’t very good. Quoting from that piece, written in 2013.

Really at issue is Kershaw on short rest. We can’t compare him to himself, because he’s never done this before. We do have a history of other starters going on short rest in the playoffs, so that’s going to have to do. Below, you’ll see statistics for two groups, in postseason play between 1995-2013. The first group is starters who went on three days’ rest. The second group is starters who didn’t. Understand that the first group should be selective for better arms, because you don’t put mediocre starters on short rest in October, or ever. Typically, it’s aces who go on three days’ rest, and the numbers are telling:

Short-rest group: 4.66 ERA, 5.13 RA
Other group: 3.99 ERA, 4.30 RA

Last year, a few more starting pitchers have pitched in the postseason on three days rest. Kershaw did it again against St. Louis in the division series, in fact, giving up three runs in six innings as the Cardinals beat him for a second time. But then there was Madison Bumgarner, finishing game seven of the World Series on two days rest and throwing five shutout innings, allowing just two hits in the process. It isn’t so cut-and-dried that starting a guy on three days rest is definitely going to be a disaster, and the guys who tend to get selected for short-rest work are often high-quality hurlers.

But even with that group containing only elite pitchers, the data shows that short-rest starters have generally performed pretty poorly in the postseason. At the very least, the Astros shouldn’t count on Dallas Keuchel pitching like regular season Dallas Keuchel tomorrow, and he should be on a significantly shorter leash than he would be if he was working on regular rest. And that’s without even considering the variables that should incentivize both the Astros and Yankees to turn tomorrow’s contest into a battle of the bullpens.

Because MLB treats the Wild Card game as it’s own separate playoff “round”, teams are allowed to set a roster for just that one game, but teams are still given the full complement of their normal 25 man roster. Because you only need to carry two starting pitchers — the guy who is starting that day, and a backup in case he gets injured or the game goes into extra innings — each team essentially gets three extra roster spots to play with. Most of the time, teams have chosen to carry a specialty pinch-runner (think Terrance Gore) and sometimes a third catcher if they have a weak-hitting backstop and want to be able to pinch-hit, but that also allows teams to carry eight or nine relief pitchers while still maintaining a full complement of position players.

And because there are off days both before and after the Wild Card game, teams can aggressively use their relievers in the Wild Card game without causing significant harm to their chances in the division series. Luke Gregerson and Will Harris haven’t pitched since Saturday; Tony Sipp, Chad Qualls, and Pat Neshek were used very lightly on Sunday, and should be ready to pitch multiple innings if need be. Toss in the group of Joe Thatcher, Josh Fields, Oliver Perez, and Vincent Velasquez, along with Mike Fiers or Scott Kazmir as the starter who’d be hanging around in case of emergency, and you have up to 10 arms available after Keuchel to throw at the Yankees. More likely, the Astros will only carry nine of those 10 pitchers in order to get a 15th position player on the team, but either way, that’s a deep reservoir of arms to have sitting around watching a guy pitch on three days rest for the first time in his career.

In an ideal world, where routines and egos didn’t come into play, the team would hold Keuchel in reserve rather than starting him at the beginning of the game. If the Astros led off with a series of relievers, the Yankees wouldn’t be able to optimize their line-up to get the platoon advantage against the left-handed starter, and the team could keep Keuchel off the mound entirely — saving him for game one of the division series on full rest — if the team took a big early lead and the string of relievers proved effective. The ideal scenario would be winning without using Keuchel, which could only happen if he didn’t pitch the first inning.

But because pitchers are creatures of habit, and it may very well effect Keuchel to not know ahead of time that he’s going to take the mound, I understand that it’s a big risk to start with the relievers and finish with a starter if necessary. It’s out there, even for a progressive team like the Astros. But given the depth of arms they’re going to have, they should not let Keuchel dig much of a hole before turning to their relievers. After all, many of their relievers were just as effective this season as Keuchel.

Astros OPS Allowed
Player OPS
Will Harris 0.525
Luke Gregerson 0.573
Dallas Keuchel 0.575
Josh Fields 0.602
Tony Sipp 0.606
Chad Qualls 0.681
Pat Neshek 0.709
Vincent Velasquez 0.720
Joe Thatcher 0.726
Oliver Perez 0.798

Add in some performance degradation for Keuchel pitching on three days rest, and he doesn’t have a clear advantage over most of the Astros bullpen tomorrow, especially because relievers can be swapped in and out depending on the handedness of the opposing hitter. The Yankees are likely to start Chris Young in the outfield with a southpaw starting, and perhaps may start Rob Refsnyder at second base as well, given how they’ve been running their second base platoon of late. Swapping in a right-handed reliever early would force Joe Girardi to either deplete his bench in the early part of the game, or allow hitters to face an RHP who were only in the line-up to get the advantage against an LHP.

At most, I’d think the Astros should be looking at 3-4 innings from Keuchel, then begin deploying the army of relievers to keep the Yankees from getting repetitive looks in a game where there is no such thing as a low leverage situation. When every at-bat could have this large an impact on the team’s chances of advancing in the postseason, the Astros shouldn’t allow the traditional starter/reliever dynamic to put them in an early hole.

Pretty much all of this is true for the Yankees as well, of course. While Masahiro Tanaka is not pitching on short rest, he only threw five innings in his start against the Red Sox last week after having 12 days off between starts due to a hamstring issue. The Yankees can’t be sure that Tanaka is going to be at 100% effectiveness for the Wild Card start, and given that they’ve got Dellin Betances, Andrew Miller, and Justin Wilson down in the bullpen — plus Adam Warren, most likely, who has been excellent in relief as well — the Yankees should also be looking to lean heavily on their relievers to advance to the division series. Since they don’t quite have Houston’s stockpile of bullpen options, perhaps they’ll try to get five from Tanaka, but there’s no reason to push him beyond that.

The idea of a match-up of aces is quite appealing from a narrative standpoint, but the Astros and Yankees shouldn’t see this as a duel between Keuchel and Tanaka. Baseball might not be ready to move away from having a starting pitcher in a one-game playoff, but both teams should plan on going to their bullpens early and often tomorrow. The stakes are too high, and the rules too clearly slanted in this direction, to not take advantage of the opportunity to use a never-ending supply of relievers to try and advance to the ALDS.





Dave is the Managing Editor of FanGraphs.

newest oldest most voted
Marco
Guest
Marco

This makes good logical sense. Moral hazard is at play here though.

If Keuchel never sees the mound and they lose, the manager could easily lose his job. With millions of dollars at stake, it’s hard for me to fault the guy if he decides to go the “safe” route.

Bryz
Guest

If the situation is crucial enough, Keuchel should be entering the game, regardless of score.

jpg
Guest
jpg

Yeah but it’s not always that easy. Let’s say the Astros are up 1-0 in the bottom of the fourth. A fresh reliever, say Josh Fields, is brought in and he quickly gets the first two outs. Then say, there’s a throwing error which allows a guy to reach. No biggie. Fields then walks the next guy on a 3-2 pitch. Now Hinch makes the call to get Keuchel warming and sends Strom to the mound to kill time. And then Fields’ next pitch plunks the next batter. So suddenly in a matter of seven or eight pitches, the bases are now loaded. And Keuchel who probably has a much different routine when he’s preparing to start might not be able to get loose as quickly as a full time relief pitcher. So in the game’s biggest spot may come and go without enough time to have Keuchel ready. That’s why I don’t think using Keuchel as a fire man is a good idea.

If anything, he’d probably be best deployed like Bumgarner was in Game 7 last year. Let him start the 5th and let him bring it home.

Yinka Double Dare
Guest
Yinka Double Dare

I think the only way a manager does this is if the idea comes from the front office and the manager is on board with it, or they all sit down in a room and decide to do it. I can’t imagine a manager doing this on his own unless he’s a long-timer that knows he isn’t going to be fired and has the cachet to pull such a move.

Surrealistic Pilliow
Guest
Surrealistic Pilliow

This is not moral hazard.

Brent Henry
Member
Brent Henry

*googles moral hazard*

this guy’s right.

joser
Guest
joser

Indeed. I’d probably term it a “perverse incentive”

matt w
Guest
matt w

Yeah, or maybe a principal-agent problem–the manager is making decisions for the team, but his incentive isn’t necessarily to do what has the best expected outcome for the team.

a eskpert
Guest
a eskpert

Moral Hazard is a type of principal-agent problem, but it specifically refers to situations in which the agent has an incentive to take excessive risks.

JayT
Guest
JayT

I think he’s saying that the manager is taking a risk (not doing a bullpen game) because in the event that it goes bad and his team loses, he won’t get the blame since he did what everybody always does. I think it more or less fits.

a eskpert
Guest
a eskpert

A risk is when there is increased variance in outcomes, not that the average outcome is any worse (sorry the economist in me is coming out). So if there was also a chance that Keuchel would be substantially better than the bullpen or pitch a CG, as well as totally melting down, Moral Hazard might be the appropriate term.

Mike B.
Guest
Mike B.

Correct. It could be considered a morale hazard, however.

Cat Latos
Guest
Cat Latos

I think AJ Hinch could literally roll out Roberto Hernandez tomorrow and not get fired.

chuckb
Guest
chuckb

That’s not what a moral hazard is. You’re describing people being risk-averse which is not at all the same thing as moral hazard.