A brief history of bullpenning

BySam Miller ESPN logo
Thursday, October 4, 2018

The weird thing the Oakland Athletics did Wednesday night wasn't that weird.

It seems weird, because for a century and a half baseball hasn't been played that way. The A's, using a "reliever" to start the winner-take-all wild-card game -- with no expectation that he would go past the first or second inning, and with an expectation that six or seven or eight pitchers would ultimately appear not by demand but by design -- were sharply breaking (as have the Tampa Bay Rays and other teams this year) with the established way we always did things.

But what's notable is baseball was never that far off from doing this weird thing. The history of The Weird Thing is that it was never that wild of an idea, never that far removed from baseballers' minds. Consider this brief, incomplete history of how we got here:


Here's a headline in the Topeka State News, April 10, 1903:

This was still the dawn of relief pitching: In early baseball, rosters had been limited and substitutions illegal, so relief pitchers would have to come from another position on the field. The rules were liberalized in the 1890s and expanded managers' options. In 1902, the Chicago White Sox manager Clark Griffith "was planning on using Virgil Garvin scientifically," which meant relieving him when he began to tire, according to an article quoted in Peter Morris' history book "A Game Of Inches." Traditionalists were "appalled" that starters weren't finishing their games, according to Morris.

McGraw's "radical and sensational" move took this notion further: Every game would have a starting pitcher who would go only five innings, instead of the default nine, to be followed by a second pitcher who would go the final four. McGraw wouldn't simply be willing to use multiple pitchers in a game, but would plan to. He started with the premise that multiple pitchers in a game would be better than one. The strategy shared some of the same benefits that A's strategy tried to capture: It would prevent the starting pitcher from tiring ("most games are lost on one or two innings") and it would keep opposing batters from getting comfortable seeing one pitcher a bunch of times ("any ball player knows that any change of delivery during a game is more or less puzzling"). It rejected the goal of trying to squeeze as much out of the starter as possible, favoring instead the aim of putting each pitcher in his strongest predetermined role.

But, Morris recounts, McGraw's plan was met by skeptical sportswriters, and inertia proved powerful. His Opening Day starter (Christy Mathewson) pitched well and McGraw let him go the distance. Then his Game 2 starter did the same, and McGraw let him go the distance, too. McGraw did it the old way and he won. In that Topeka State News story, there's a hint of what the biggest obstacle to doing it a new way would be: "McGraw says the only thing to fear in adopting the plan is the 'roasting' he will get" if the second pitcher pitches poorly. He "essentially abandoned his scheme," Morris writes.


One day in 1949, the St. Louis Browns started their young ace, Ned Garver. But Garver went only one (scoreless) inning, and was then replaced by a caravan of relievers: Nine pitchers in all went one inning apiece, including the rest of the Browns' usual starting rotation.

It was, to be clear, the last day of a 100-loss season, not the start of a bold new strategic direction by Browns management. But Garver -- who died last year --once explained to us that the plan wasn't arbitrary and wasn't a goof; in fact, it came from the pitchers themselves, who were merely following the same logic the A's (and Rays, et al.) are following: "We would talk about things like that and we would say, 'Well, maybe that would be a way to do it. You just come in to pitch one inning, the hitters don't see you the rest of the day -- they see another guy.' It was unusual to do that but I thought, well what the heck, it wouldn't be too bad of an idea. You could pitch one inning every day!"

The Browns didn't continue the experiment the next season. And they almost surely wouldn't have tried it at all if they weren't in seventh place on the last day of the season. But Garver's explanation shows this notion wasn't all that hard to understand, or even far from the players' minds.

The Browns lost that day.


In 1990, Pittsburgh Pirates pitcher Zane Smith went out and gave a media conference the day before he was to start against the Cincinnati Reds in Game 6 of the National League Championship Series. Smith knew, and his manager Jim Leyland knew, that Smith wouldn't actually start that game. Instead, Leyland was going to announce before the game that right-handed reliever Ted Power would start, instead.

The hope was the Reds would then set their lineup with a bunch of left-handed hitters to face the righty Power. Only then would Smith come into the game, as a long reliever, and he'd either get to face a bunch of left-handed batters or the Reds would have to burn their whole bench to regain the platoon advantage. "Look at it this way, all we're doing is starting the game in the seventh inning," explained Ray Miller, the Pirates' pitching coach. It didn't really work -- the Reds put only one extra lefty in the lineup -- and the Pirates lost, though they allowed only two runs.

"You can't manage in fear, worrying about what others are going to say," Leyland said after the fact, but it's worth noting what others were mostly saying was pretty harsh toward Leyland -- probably because Leyland had lied to reporters and sprung the change at the last minute. It was described not as a plan or a strategy or an interesting idea but as "chicanery," "a ruse," "subterfuge," "a ploy," "trickery," "a charade," "desperate," and, in the words of Reds right fielder Paul O'Neil, "a bunch of crap." Leyland "tarnished his reputation ... outsmarted himself. ... It smelled of panic," Tom Verducci wrote. "On the surface, it sure looks like they're not real confident in Zane Smith getting us out," Reds first baseman Todd Benzinger said before the game.

It would be interesting to see the response in a universe where Leyland hadn't lied about his starter -- there was little advantage to the deception, and Leyland said he just didn't want to have to answer a bunch of questions about it on his day off -- and had simply announced the strategy ahead of time. In that universe, the move might well have been seen as savvy, even prophetic.

"Why not keep going and turn all pitchers into short-order cooks?" George Vecsey asked in the New York Times that month. "Maybe Jim Leyland of the Pirates was setting a trend last Friday when he started a right-handed relief pitcher, Ted Power, and soon switched to a lefty starter, Zane Smith." Alas, Leyland's lie was the story; that and the fact the Pirates lost.


Less than three years later, Tony La Russa attempted what looks now like the prototype for bullpenning. (A quick word on the jargon here: When you hear somebody refer to "an opener," it means the pitcher who starts the game is not the one who is expected to pitch the longest. Rather, he's a traditional reliever who pitches the first inning and then gives way to somebody who'll pitch more bulk -- basically, the "starter" is coming in only after the first or second inning. When you hear somebody refer to "bullpenning," it means using a stream of relievers in one- or two-inning stints instead of asking any pitcher to stay in for very long. The two strategies both became more common this year, but aren't quite the same; they solve different problems using different resources.)

Ben Lindbergh wrote at length this summer about La Russa's midseason plan for the 1993 A's. The short story is: Instead of a starting rotation and a bullpen of relievers behind them, the A's had three "platoons" of three pitchers apiece, who would each pitch every third day, along with four ordinary relievers who might pitch whenever needed. No pitcher would stay in a game long enough to really get tired, or to face a lineup a third time. But "six days into the new rotation order, the A's were in last place and 1-5 since they started platooning pitchers, which signed the system's death warrant after two turns," Lindbergh writes.


By the time La Russa tried it, though, the sport had changed dramatically in other ways. Relievers were a part of almost every day's plan. Every team, and every manager, was engaging with the questions McGraw, Garver, Leyland and La Russa had all grappled with: How to use your best pitchers as much as possible, without tiring them out to the point they're not your best pitchers anymore; how to leverage the limited number of innings pitchers can throw, so better pitchers are available when the game is most on the line; how to balance the needs of a single game with the demands of a long season; and how to put every pitcher in a position where he will be his most effective self. While bullpenning and openers weren't yet common strategies, the conversation around pitcher usage had broadened.

This is especially apparent in baseball writing. In 2009, Baseball Prospectus' Ken Funck proposed a "SOMA" plan-- shorter outings, more appearances -- that looks a lot like bullpenning. In 2012, Dave Cameron wrote the Braves should "seriously consider the idea of skipping the starting pitcher entirely." It became almost an annual FanGraphs post: Site writers made the same proposal for the wild-card games in (at least) 2015, 2016 and 2017. In 2013, at Beyond The Box Score, Bryan Grosnick coined the term "opener" and laid out the benefits of starting a "reliever" for an inning before replacing him with a "starter." (Including: Getting the platoon advantage, avoiding a dangerous third trip through the middle of the opposing lineup.) It'd be easy to dismiss these -- and many, many other articles -- as just blog posts by just bloggers. But front offices are filled with alumni of Baseball Prospectus, FanGraphs and Beyond the Box Score. And in 2013, in a must-win ALDS Game, Joe Maddon really did go with an all-bullpen game, pulling his starter after one scoreless inning and running a parade of eight relievers after him.

The Rays lost 3-1.


Here's what I mean by "the weird thing wasn't that weird." Say you go get a doughnut every morning. You love maple bars. You get a maple bar every day. That's how doughnuts work: You have a favorite, you get it every day.

Now, is a bear claw weird? Not at all. It's a normal doughnut option that you're aware of, that you assume is probably also delicious, and that you maybe even half-consider from time to time. If they were out of maple bars one day, you might get a bear claw. It wouldn't shock your system. You might even find you like bear claws more. Or you might not, and then the next day you'd go back to maple bars.

"Sometimes the secret to survival for an unorthodox idea isn't how well it works, but how well it works quickly," Ben Lindbergh wrote in his history of Tony La Russa's 1993 experiment. Notice that, in each of these experiments, the plan failed -- the team trying it lost -- or, in the case of John McGraw, the alternative to the experiment happened to work too well. The league was open to brand new ideas, but brand new ideas are fragile things. They need to work quickly to get roots.

There's another game in this whole progression I think about. It was in 1914, the season that was, some historians have said, "the baseline year for platoons." Managers had just begun to really use the platoon advantage to organize the hitting half of their rosters.

So one April day, the Chicago White Sox were playing the Cleveland Naps. The Naps' leadoff hitter, Doc Johnston, batted left-handed. So did No. 3 hitter Shoeless Joe Jackson, and No. 5 hitter Jack Graney. The rest of the lineup was right-handed.

The White Sox started the game with Reb Russell on the mound, a left-hander. According to the box score, he retired the first five men he faced -- getting through the lefty-heavy stretch of the lineup -- and then was relieved by right-handed Jim Scott, who went the final 7 innings. Russell was basically the opener, facing the dangerous, left-handed part of the lineup -- just as an ace reliever would normally do later in the game, but here it was at the start. It was like Ray Miller said all those years later: starting the game in the seventh inning.

The full story, though, is this outcome was an accident. Russell wasn't supposed to come out early; he was hit by a line drive and had to leave, injured. It was only a fluke that the White Sox had perfectly demonstrated how an opener could work. The White Sox won, 7-0.

You could imagine, at almost any point in baseball's history, somebody attempting the opener, or attempting bullpenning. You could imagine it immediately working -- working as well as it accidentally worked for Reb Russell and Jim Scott. Working as well as it worked when the Rays started Sergio Romo on May 19 and May 20 this year, launching a trend that at least six other teams nibbled at this year. You could imagine us getting to exactly where the A's were on Wednesday night. It's easy to imagine, because, when you think about it, it's really not that weird a strategy.