On Sunday night, with Toronto and Texas tied at six heading into the bottom of the 10th inning of ALDS Game 3, Rangers manager Jeff Banister asked reliever Matt Bush to preserve his team’s season for a third consecutive inning. Bush hadn’t delivered an outing of more than two innings all year; with 22 pitches thrown through his first six outs, a third frame was likely to push him up to or over his season high of 33. And these were pressure-packed pitches: With the heart of Toronto’s order due up and (probably juiced) baseballs flying out of parks this postseason, Bush and Banister knew that one missed location could send Toronto to the ALCS.
Four batters later, the Rangers lost, on Bush’s 42nd pitch. In the aftermath of elimination, some Sunday night quarterbacks drew a connection to the previous week’s AL wild-card game, when the Blue Jays had beaten Baltimore while Orioles relief ace Zach Britton watched and waited in the pen. Yet again, the Blue Jays had advanced via an extra-innings win without facing their opponent’s closer — in this case, Sam Dyson, the American Leaguer with the highest ground ball rate behind Britton’s, and a workhorse with more appearances from 2015–16 than all but one other AL pitcher.
Banister’s decision wasn’t nearly as bad as Buck Showalter’s. For one thing, Bush was probably better than Dyson during the regular season, despite his lack of saves; it wasn’t as if Banister had deployed Martín Pérez, the Rangers’ closest Ubaldo Jiménez equivalent. For another, it was an inning earlier in the game, and Texas had already used six pitchers in relief of Colby Lewis; the Rangers were facing a greater risk than the Orioles of running out of arms. And from a second-guessing standpoint, it helps Banister that Bush did better than Ubaldo: He’d retired Toronto in order in his first two innings, and he would have escaped with an inning-ending double play in the 11th had the relay not been flubbed by Rougned Odor, the worst-fielding, most error-prone second baseman in baseball.
Still, the (defensible) decision to stretch Bush beyond his previous limit likely owed at least a little to many a manager’s October bugaboo: a reluctance to use the closer in a tie game on the road (i.e. a non-save situation) even if refusing to do so means risking losing before the closer can get into the game. Not every skipper is guilty of this sin against common sense and statistics. In the first game of the other ALDS, Indians manager Terry Francona shut down the Red Sox by bringing in fireman Andrew Miller in the fifth, with starter Trevor Bauer one out away from qualifying for a win. In Game 3, Francona did it again, using Miller in the sixth to help squeeze the last life out of the Sox. Maybe that’s the future. If so, then the Britton/Dyson scenario is the dark past we’re (maybe?) belatedly leaving behind.
So what are the greatest crimes against closer usage we’ve witnessed in recent Octobers? Managerial misuse of closers comes in a few flavors. We could criticize managers for not being more like Francona, or for not pushing their best bullpen guys even a couple of extra outs, but if we included every instance of a manager failing in one of those ways, we’d have a very long list. Let’s limit this to “tie game on the road,” since there’s a special spot in spectator hell reserved for managers who make us revisit the mind-blowingly backward debate about protecting a hypothetical lead vs. protecting an actual tie (which closers do just as well as other relievers). Or, as my Orioles fan editor put it:
Since 1988, when Tony La Russa and Dennis Eckersley helped solidify the one-inning, save-situation-oriented closer role, the Britton/Dyson scenario of a postseason road loss in which the closer (defined as the pitcher who led his team in regular-season saves in September/October) never pitched and the highest-leverage moment came in the ninth inning or later has arisen 17 times, including the two times this year. After filtering out some examples that didn’t fit the spirit of the search, we’re left with nine painful, non-Britton/Dyson examples. I’ve listed them below, in ascending order of Championship Leverage Index, a measure of how much the game affected each team’s odds of winning a World Series. The CLI values are calibrated to a baseline where 1.0 represents Game 1 of the World Series, and lower numbers equal lower stakes. For reference, the Dyson game had a relatively low CLI of 0.20 (because the Rangers were already in an 0–2 hole in an early round), and the Britton game had a CLI of 0.40 (because even winning the wild-card game would have left the O’s a long way away from a championship).
2003 ALDS, Game 3: Athletics at Red Sox
Manager: Ken Macha
Closer: Keith Foulke
CLI: 0.20
In fairness to Macha, Foulke — who that year led the AL in games finished and saves, with seventh- and 15th-place showings in the Cy Young and MVP voting, respectively — had pitched a combined four scoreless innings in the first two games of the series. Maybe Macha didn’t want to test him, even though there had been an off day after Game 2. Instead, he used Jim Mecir to pitch the 10th and rookie Rich Harden to pitch the 11th — or, as it turned out, the one out in the 11th that preceded this:
In further fairness to Macha, Foulke blew a lead and lost Game 4 the next day, as good a reminder as any that many “right” managerial moves backfire, just as many “wrong” managerial moves work out. In most cases, we’re talking about small swings in win expectancy that are infuriating not because they regularly make the difference between wins and losses, but because they’re unforced errors.
1998 NLDS, Game 2: Cubs at Braves
Manager: Jim Riggleman
Closer: Rod Beck
CLI: 0.31
Here’s how you know baseball has evolved in the past 18 years: Riggleman got eight scoreless, four-hit innings from the amazingly mediocre Kevin Tapani — a roughly league-average starter to that point in his career, and a below-average one that season — and actually let him try to finish the game, which would make Twitter apoplectic in this less complete-game-conscious time. Tapani regressed to garden-variety and surrendered a game-tying Javy López homer in the ninth. Rather than hand the ball to Beck, the horseshoe-mustached closer who’d saved 51 games and struck out a batter per inning that season (which was still saying something in 1998), Riggleman went to Terry Mulholland in the 10th. A lineout, walk, sac bunt, and Chipper Jones single later, it was over.
Riggleman got fired the following year.
2012 NLDS, Game 4: Cardinals at Nationals
Manager: Mike Matheny
Closer: Jason Motte
CLI: 0.43
If we weren’t already thinking about these as “Britton games,” we could call them “Mathenys.” The Cardinals manager is the most frequent offender, and his inaugural violation came during his first postseason series, when he used Lance Lynn in a tie game on the road instead of Jason Motte, who hadn’t pitched since Game 1. His reasoning ran along familiar lines.
Lynn lasted 13 pitches — 10 of them four-seam fastballs, as was his way — but he couldn’t record an out.
2014 NLCS, Game 5: Cardinals at Giants
Manager: Mike Matheny
Closer: Trevor Rosenthal
CLI: 0.47
This time, Matheny bypassed his closer in favor of long-sidelined starter Michael Wacha, who in addition to suffering from the aftereffects of a stress fracture in his shoulder, hadn’t gotten into a game in 20 days and hadn’t pitched in relief for more than a year. Matheny inexplicably chose Wacha to pitch with the Cardinals’ season on the line, and the Baseball Gods chose Travis Ishikawa to be the agent of their divine justice.
Sorry, did I say inexplicable? Actually, it was easily explained.
2014 NLCS, Game 2: Giants at Cardinals
Manager: Bruce Bochy
Closer: Santiago Casilla
CLI: 0.49
If there’s any consolation for Cardinals fans, it’s that future Hall of Famer and bullpen wizard Bruce Bochy made the same mistake in the same series, opting for deposed closer Sergio Romo instead of current closer Casilla. As Jonah Keri wrote in response, “the best practice for bullpen use that late in the game should be to make sure a team uses its best guys, instead of trying to get cute with lesser pitchers.” Kolten Wong won it.
2014 NLCS, Game 3: Cardinals at Giants
Manager: Mike Matheny
Closer: Trevor Rosenthal
CLI: 0.60
No, we’re not finished with this series yet: The 2014 NLCS featured Britton/Dyson scenarios in back-to-back-to-back games. Granted, Rosenthal had blown a lead in Game 2, throwing 28 pitches in the process, but he’d had an off day to rest and hadn’t pitched in Game 1. He’d also saved all three Cardinals wins in the NLDS and allowed only one run in his last 11 regular-season games, which made him a better option than Matheny’s actual choice, lefty specialist Randy Choate, who didn’t get an out.
1997 ALCS, Game 4: Orioles at Indians
Manager: Davey Johnson
Closer: Randy Myers
CLI: 0.62
Myers had lost Game 3 the previous day, but closers, the cliché goes, are supposed to have short memories. The lefty was the 1997 equivalent of Britton, coming off a season in which he’d led the majors with 45 saves and finished fourth in both Cy Young and MVP voting. With the game tied at seven, Johnson started the ninth inning with the immortal Alan Mills, who had already worked two innings and who had walked more men than he struck out during the regular season. You’ll never believe this, but Mills walked the leadoff man, who soon scored on Sandy Alomar’s walk-off single. Although the O’s lost with their closer still in the pen, this one wasn’t as perplexing as the wild-card debacle 19 years later. The pitchers who followed Mills, Jesse Orosco and Armando Brawlin’ Benítez, were almost as effective as Myers.
2010 NLCS, Game 4: Phillies at Giants
Manager: Charlie Manuel
Closer: Brad Lidge
CLI: 0.62
Roy Oswalt had started Game 2 three days earlier, and he’d thrown a bullpen session before Game 4. But with the score even at five and four relievers already removed from the lineup card, he volunteered to make his first relief appearance of the season (and his second in four years).
The Phillies took Oswalt up on his offer, even though Lidge had warmed up once in the game and had thrown only one inning in the previous 11 days. “As the closer, Lidge is not deployed in a road game until the Phillies have the chance to win,” the L.A. Times’ Bill Shaikin wrote in his game story. “That’s the way we’ve always done it,” Lidge said. Smart.
2003 World Series, Game 4: Yankees at Marlins
Manager: Joe Torre
Closer: Mariano Rivera
CLI: 1.16
According to a recent FiveThirtyEight analysis, Joe Torre is the best bullpen manager since 2000. That might surprise some Yankees and Dodgers fans who watched him work, although the most common complaint about Torre was that he wore out relievers, not that he didn’t use his best ones at the most important moments. With the exception, that is, of this moment, when he chose Jeff(!) Weaver(!) over the all-time top inning-per-inning closer.
Weaver was a sub-replacement pitcher in 2003, the Ubaldo of his day. He hadn’t pitched in the ALCS or the ALDS, or even in the regular season since a 12-pitch outing on September 24 — almost a month before the October 22 Game 4. Yet Torre, wanting to reserve Rivera for a save situation, went with Weaver over a future Hall of Famer in his perpetual prime. Rivera had pitched two innings the previous day (if only because Torre left him in to collect a save after the Yankees scored four times in the top of the ninth), but he’d had four days off before that.
The only surprise is that Weaver retired three hitters before Álex González took him deep.
“If he’s not in the game there, he shouldn’t be on the roster,” Torre told The New York Times, attempting to justify one bad decision by citing a second.
Although this list has loads of Matheny, who’ll never be confused for a great tactician, it also includes legends such as Showalter, Bochy, and Torre. It’s either reassuring or deeply disturbing to know that the “tie game on the road” error isn’t only the province of professional punch lines: Faced with career-making moments, even the most successful managers can buckle under postseason pressure and fall back on the book. Except, perhaps, for the ones with Andrew Miller.
Thanks to Dan Hirsch of The Baseball Gauge for research assistance.
