The Baltimore Orioles are headed for an early offseason after losing the American League wild card game in Toronto on Tuesday night, and baseball Twitter had a field day over one of the biggest reasons why:
As the game went deeper into the night, eventually requiring extra frames, O’s manager Buck Showalter asked for bullpen innings from the second-, third-, eighth-, 10th-, 12th- and 16th-most valuable relievers on his active roster (according to their regular-season wins above replacement). Meanwhile, Zach Britton, far and away his top reliever by WAR, went unused even as the leverage index of the game — and, by extension, Baltimore’s season — reached absurdly high levels. Instead of focusing on how Britton could preserve the tie, Showalter kept waiting for a lead that his closer could protect, and in doing so illustrated a bullpen-management trap that too many skippers still fall into.
In the bottom of the ninth, Showalter used setup man Brad Brach (after already having had him pitch the eighth) to strike out Jose Bautista, with runners on first and second and nobody out — a moment that carried a leverage of 3.3 (i.e., 3.3 times as influential on the game’s outcome as the typical situation). He then brought in Darren O’Day — who notched a mere 0.1 WAR during the regular season — to face the next batter, Russell Martin, with a massive leverage index of 4.3. Mercifully for Baltimore, the Orioles escaped that situation after Martin grounded into a double play.
They also survived having O’Day pitch the 10th, which began with a leverage of 2.3. (He got three quick outs.) But in the 11th, Showalter started the inning with Brian Duensing, a league-average reliever, in another 2.3 leverage-index situation, then brought slightly-below-average starter Ubaldo Jimenez in to face Devon Travis — who promptly singled to bring the game’s leverage up to 2.9. Showalter let Jimenez continue against Josh Donaldson, who stroked another single, and then left Jimenez in against Edwin Encarnacion even as the leverage skyrocketed to a staggering 5.3, with runners on the corners with 1 out.
(All of this happened despite Britton being “available”.)
That’s when Encarnacion finally put Showalter and the Orioles out of their misery:
After the game, Showalter offered a few explanations, including the idea that Jimenez had the hot hand and that being tied on the road meant he should save his closer. (More on this later.) Although, as Yahoo’s Jeff Passan wrote this morning, Showalter’s justifications amounted to little more than the personification of a shrug emoji.
The common thread in all of the situations above was just how high Showalter allowed the game’s leverage index to get while still refusing to break the “in case of emergency” glass Britton was patiently waiting behind. Here’s every Toronto plate appearance from the ninth inning onward, along with its leverage index and the percentage of this year’s regular-season plays in the ninth inning (or later) that had a lower leverage index, according to Baseball-Reference.com’s data:
|Jimenez||Donaldson||1||1–||Single, runner to 3B||2.9||85|
|Jimenez||Encarnacion||1||1-3||3-run Home Run||5.3||97|
Eight of the 11 plays occurred in moments that ranked among the top 26 percent of all pressure-packed scenarios; five were in the top 20 percent and two were in the top 10 percent. In the most basic sense, these are the moments in which you probably want your best reliever on the mound instead of cycling through warmups and cooldowns in the bullpen.
But in fairness to Showalter, recognizing the threshold of the game’s most important moment isn’t always easy in an extra-inning game. You might think a situation like the one O’Day faced in the bottom of the ninth — runners on first and second, 1 out — would have been likely to be the most important remaining moment of the ballgame, based on what we knew to that point. But in the ninth inning or later in 2016, the odds of a 4.3-leverage moment being the most crucial remaining moment were only about 42 percent:
|LEVERAGE RANGE||NO. PLAYS||AVG. LEVERAGE||CHANCE OF BEING HIGHEST LEVERAGE FOR REST OF GAME|
|0.50 – 0.99||1,848||0.73||53.1%|
|1.00 – 1.49||1,598||1.30||41.2|
|1.50 – 1.99||1,656||1.76||36.4|
|2.00 – 2.49||1,814||2.26||25.3|
|2.50 – 2.99||975||2.77||45.1|
|3.50 – 3.99||437||3.74||46.7|
|4.00 – 4.49||327||4.32||42.2|
|4.50 – 4.99||456||4.72||49.3|
|5.00 – 5.49||199||5.28||57.3|
|5.50 – 5.99||146||5.79||58.9|
In actuality, there’s a complicated dynamic between a game’s current situation and its odds of eventually being the defining moment of the contest. The relationship isn’t constant — a low-leverage situation in the ninth inning (below a leverage index of 1.0) might be likelier to contain the game’s most crucial remaining moment than a medium-to-high leverage one (say, a leverage of 2.0) because the former can more easily taper down to a leverage of zero — and game’s end — while the latter means the game is already in a high-leverage state, and may continue to be so after a given play.
After dipping at leverages between 2.0 and 3.5, a trend does form between higher leverages and a higher likelihood of being the game’s most important remaining moment. But you have to approach a leverage of 4.7 or so to get around a 50-50 chance of the current play being the most important one left in the ballgame. Only about 5 percent of all plays in the ninth inning or later carry such weight, though they tend to come up often in nail-biters like the one Toronto and Baltimore played Tuesday.
Of course, many of these kinds of plays flow from one another, so you could potentially bring in your relief ace for a comparatively less-tense play and still eventually get to the Big Moment. Even more ideally, you could bring him in early enough to snuff out said Moment before it can ever happen. Seamheads are always complaining that teams don’t use closers before the ninth inning; Britton could have even come in before the ninth, and nobody would be tearing Showalter to shreds right now, even if the game’s eventual outcome was exactly the same.
Besides, Showalter probably wasn’t going through mental calculations anywhere near so involved. For one thing, it wasn’t a save situation, so the idea of using his closer there probably met some level of resistance in back of the manager’s mind. (It’s worth noting that Showalter ranked a mediocre 31st out of 45 qualified managers in Rob Arthur and Rian Watt’s measurement of bullpen-managing acumen.) And managers often refuse to bring their closers into tie games on the road anyway, under the theory that they should be saved in order to hold an eventual lead rather than be burned protecting a tie. (Never mind that said lead often never comes because an inferior reliever has already blown the tie.) A lot of orthodoxy played into Showalter’s decision to keep Britton on ice.
That choice will probably haunt Showalter forever, and it gave a truckload of ammunition to the internet’s never-ending horde of second-guessers. (Criticizing skippers for managing to the save and other forms of bullpen bungling will never get old.) Tough as it can be to manage a bullpen in the extra innings of an elimination game, though, the disparity between Britton’s numbers and those of the pitchers Showalter used — coupled with the game’s ever-escalating leverage index — make the criticisms fair. Losing is ultimately a team effort, but Showalter’s bullpen strategy didn’t do the Orioles any favors Tuesday night.
CORRECTION (Oct. 5, 3:40 p.m.): An earlier version of this article incorrectly described the chart showing the leverage of each play in the last three innings of Tuesday’s game. The last column of the chart shows the percentage of this year’s regular-season plays in the ninth inning (or later) that had a lower leverage index, not a higher leverage index.