On Friday night, the Rays topped the Yankees 2-1, defeating their deep-pocketed division rivals three games to two in the American League Division Series and sentencing the Yanks to an almost intolerable 11th straight autumn without a ticker-tape parade. (Not that the pandemic permits parades.) The decisive, double-elimination nailbiter was a study in contrasting styles of roster construction. The Yankees used three high-profile pitchers acquired via free agency, whose combined 2020 salaries would have totaled $65 million in a non-shortened season, while the Rays rolled out four pitchers acquired via trade or amateur free agency, who collectively would have earned less than $4 million.
Notwithstanding the Yankees staff’s skew toward record contracts, the results on both sides were similar: All seven pitchers threw gas, producing a Statcast-reported average fastball velocity (including four-seamers and sinkers) of 97.2 mph—the second-highest figure in the 13-year pitch-tracking era. (According to Pitch Info’s park-corrected data, the average was only 96.7, which doesn’t rank quite as high but which I can confirm is still really fricking fast.) The slowest four-seamer thrown was an offering from Rays reliever Nick Anderson that zipped in between 94 and 95.
On a night of overpowering pitching, when 24 hitters struck out and each team tallied three hits—none of them with a runner on base—the scoring came exclusively from solo home runs that barely cleared the fence. (More than three quarters of the runs in the strikeout-centric series were produced on dingers.) The third and last of those blows was struck in the eighth by undrafted Rays utility man Mike Brosseau, who entered as a pinch-hitter in the sixth. In his second at-bat, Brosseau staged a masterful, 10-pitch morality play in which he learned to time Aroldis Chapman’s triple-digit heat and delivered the sweetest possible payback for nearly being beaned by Chapman in September. For the third time in the last five seasons, Chapman allowed a lead-changing home run in the eighth inning or later of a series-deciding postseason game, and for the second time in franchise history, the Rays advanced to the ALCS.
In the box scores and championship win probability tables, the credit for the series victory goes to Rays players: one-man wrecking crew Randy Arozarena, a supporting cast of unlikely offensive standouts (Brosseau, Ji-Man Choi, Kevin Kiermaier, Michael Pérez), and the pitchers (Charlie Morton, Anderson, Diego Castillo, Pete Fairbanks) who eluded drubbings by the Bombers’ bats. The Rays boasted the AL’s best regular-season record, beat the Yankees eight out of 10 times before the postseason, and were slightly favored by the stats before the series started, so it wasn’t surprising to see them squeak by, taking three out of five despite being outscored 24-21 in the series.
Where the Yankees are concerned, though, it never seems to be sufficient to accept that small-sample series can swing either way and that losses sometimes result from being outplayed by a talented team. Losing must mean something, and it has to be someone’s fault. Sure enough, in many minds, the Yankees handed the series to Tampa Bay on Tuesday, when they decided on a semi-unorthodox pitching plan for Game 2. In this version of events, the outcome was less a product of the Rays’ superior performance and the vagaries of short series than an indictment of the Yankees’ leadership structure and its deference to stats over soul (or something). Don’t believe this version of events.
Look around the rest of the league: The Braves, whose weakness was supposed to be their starting rotation, have shut out their opponents in four of their five postseason games (including one that went 13 innings). On the flip side, the vaunted A’s bullpen—baseball’s best during the regular season—allowed 17 runs (13 earned) in 18 2/3 ALDS innings against the Astros. That doesn’t mean that the Braves’ rotation wasn’t vulnerable, or that Oakland’s bullpen was bad. It means that one week of baseball (let alone a single game) is not enough to go on when judging a team’s true talent. Nor does one pitching plan that didn’t pan out prove that the Yankees’ process was unsound. To the extent that there’s a reason the Yankees lost—aside from facing a skilled opponent—what doomed them was depth, not decision-making.
Let’s review the Yankees’ supposed series-deciding mistake. After Gerrit Cole shut down the Rays in Game 1, setting him up for a short-rest reprise in Game 5, the Yankees were left with an array of less-than-stellar starting options for Games 2-4: righties Masahiro Tanaka and Deivi García, and lefties J.A. Happ and Jordan Montgomery. The most conventional choice would have been to start Tanaka—the longest-tenured Yankee and the one with the best postseason pedigree and prettiest regular-season ERA—in Game 2. The Yankees opted to hold Tanaka for Game 3. The second-most-conventional choice would have been to pick a single starter for Game 2—Happ, probably, given his experience and success down the stretch—and stick with him for a while. The Yankees didn’t do that either.
Instead, the Yankees’ collaborative brain trust—consisting of Aaron Boone, Brian Cashman, and a coterie of quants and coaches—decided to start the rookie right-hander García and replace him with Happ early in the game. The approach was part opener, part Curly Ogden/Wade Miley maneuver: In theory, García getting the start would prompt the Rays to load up their lineup with lefties. Then Happ could come in and pitch with the platoon advantage. (The Rays, who have started as many as nine left-handed hitters in the same lineup this season, went with five in Game 2.)
Because this wasn’t the conventional call, it was bound to be criticized if the players failed to perform. Unfortunately for the Yankees, neither hurler lasted long. García allowed a solo home run to Arozarena in the first, and Boone replaced him with Happ to start the second. The lefty went only 2 2/3, allowing four runs on two homers. The soft underbelly of the bullpen—Adam Ottavino, Jonathan Loaisiga, Jonathan Holder, and Nick Nelson—allowed two more runs over the final 4 1/3, and the Yankees lost 7-5.
Cue the tabloid condemnations, the dunking on the Yankees’ “Calculus Crew” and “Sultans of the Spreadsheet,” and the blistering take from bunting-obsessed broadcaster Alex Rodriguez that the Yankees should have acted “as an alpha” and “done it the old-fashioned way” instead of playing Jeopardy! and “trying to figure out how to outsmart the Ivy Leaguers.” (Jeopardy!, of course, being the game dominated in 2019 by a baseball-loving analytics nerd who adopted a sensible strategy that pissed people off.) According to A-Rod, alphas don’t pay attention to platoon advantages. Alphas, apparently, say “Screw the front office” and squint into the sun while taking bubble baths.
YES Network broadcaster Michael Kay joined the chorus excoriating the Yankees’ call. “If the Yankees end up losing this series, they will look at this gamble and say, ‘That’s the reason why we lost,’” Kay said after Game 2. Actually, the Yankees almost certainly aren’t saying that: Boone, who has expressed regret for mistakes in past postseasons, said after Game 5 that he didn’t regret the Game 2 decision, adding that it’s “kind of ridiculous” to assume that all would have been well if the Yankees had used a traditional starter instead of an opener. “All over the league, things like this are done and done really effectively, and we’ve done them really effectively,” Boone continued.
But plenty of Yankees fans are echoing Kay’s words. In the aftermath of Game 5, “Girardi” trended on Twitter as fans rehashed Game 2, blamed Boone and the brain trust, and lamented the departure of Boone’s predecessor, whose contract Cashman decided not to renew after the 2017 season. Girardi, the Boone haters said, would have stood up to Cashman and his quants. Just so we’re clear, this is the same Joe Girardi whose Phillies finished behind the COVID-decimated Marlins in the NL East this season and failed to qualify for a 16-team playoff field despite being blessed with the game’s fifth-highest payroll. Maybe Girardi isn’t a genius, or maybe a manager doesn’t make much of a difference when he doesn’t have good enough players or his players underperform. Both of those problems sabotaged Boone in the ALDS.
Let’s tackle the Tanaka complaints first. First, Tanaka wasn’t the savior some fans believed him to be. A fine pitcher in his first three pinstriped seasons, the right-hander has been barely better than league average from 2017-20, amassing a 97 ERA- and a 94 FIP- over that span (where 100 is average and lower is better). While he had a sterling October track record before the 2019 ALCS, a sample of seven successful postseason starts didn’t make him a magic elixir, as the combined damage from his last three playoff outings proved: 15 runs (14 earned) in 13 innings.
That Tanaka was the Yankees’ second-best starter says more about the rest of the rotation that it does about him. (More on that in a moment.) The important point, though, is that in a best-of-five series with no off days, Tanaka was good for a single start. Which start hardly mattered. By choosing Game 2, the Yankees could have made it more feasible for him to work out of the bullpen in Game 5, but it’s hard to construct a scenario in which the Yankees would have wanted to use Tanaka—who, again, isn’t great, and has almost no bullpen experience—in a make-or-break game. And by staggering their best bets to go deep into games (Cole and Tanaka), the Yankees did their best to build in a break for their bullpen. That didn’t work—as it turned out, Tanaka exited after four innings in Game 3—but it was worth trying. The only other argument for the importance of starting Tanaka in Game 2 over Game 3 is one based on “momentum,” which may feel real but doesn’t appear to exist, at least on a level that translates to wins and losses.
The past peripherals and projected performance of García and Happ were similar enough (and spotty enough) that neither inspired confidence against the Rays. Thus, the Yankees hoped that using the two in tandem—and disguising their strategy in a way that wouldn’t have been possible with a regular reliever in the opener role—would be a better option than relying on either alone. By lifting García early, the team would gain a temporary platoon advantage in Game 2 and be able to bring back the righty if needed in Game 4 or 5.
According to Boone, Happ was prepared for the plan (though it’s less clear that García was), but Happ’s postgame comments made it clear that he’d expressed a desire to start. In an all-hands-on-deck postseason and an era when the lines between pitcher roles have blurred, Happ’s insistence on a rigid role seems stubborn, but a player’s preference, however cumbersome, is still something a manager must take into account. Happ, however, had drawn this duty before. In his last regular-season outing of 2019, in a game against the Rays, Happ entered in relief to start the second (after a right-handed opener pitched the first) and allowed one run in five innings, with six strikeouts and one walk. Before Tuesday, he’d also allowed three runs in 10 innings in 10 career postseason relief appearances, three of which were last year. That’s pretty good for a guy who’s supposedly incapable of pitching well if he isn’t starting.
Boone may have concluded that Happ’s displeasure wouldn’t hamper his pitching enough to nullify the on-paper advantage—which Happ, for his part, backed up. “Ultimately, I pitch when I pitch,” Happ said, adding, “There was no hesitation and no dwelling on what was going on. I was focused on trying to perform.” His curt postgame comments about the opener plan might have been at least partly a product of a preexisting resentment. Happ was miffed at the Yankees for much of the year because of how he was handled; he implied that the Yankees were restricting his usage to avoid triggering a vesting option in his contract, while the Yankees countered that Happ’s pitching hadn’t merited more innings. Maybe his ego was still smarting from that, or maybe it was wounded because the start went to a 21-year-old rookie (or simply because he hadn’t pitched well).
According to Kay, the Game 2 gambit was “the New York Yankees trying to show everybody, ‘We’re smarter than you.’” Really, though, it was the Yankees trying to give themselves a slightly better chance to win the game by using essentially the same strategy the Rays did in Game 4 (when they started righty Ryan Thompson, pulled him after an inning, and inserted southpaw Ryan Yarbrough) and the Dodgers did in NLDS Game 3 (when they started righty Dustin May, pulled him after an inning, and looked to lefty Julio Urías for length). The difference is that Yarbrough and Urías pitched well for five innings, while Happ pitched poorly for 2 2/3.
It’s possible that Happ crashed and burned because he was used in relief, but it’s also quite likely that he had a bad day for unrelated reasons, and that he or García would have struggled in a normal starting role. If García and Happ had pitched well, Boone would have looked brilliant (even though Boone himself would likely concede that the tandem plan only lightly lifted the Yankees’ win expectancy). But some smart moves backfire: Cleveland was wise to summon James Karinchak with the bases loaded in the fourth inning of the team’s second wild card game against the Yankees last week, but Gio Urshela turned a tough pitch into a grand slam. Managers plan and players laugh.
The Yankees’ Game 2 strategy—which Boone believed in, as far as we know—reflects a larger and logical shift of power from the dugout to the front office, and the game’s accelerating pursuit of interdisciplinary decision-making. Teams used to silo their scouting and player development departments, preventing the former from drafting in ways that suited the latter’s strengths. They also used to separate their minor and major league staffs such that big league coaches knew next to nothing about rookies and weren’t well-equipped to aid their development. They don’t do those things anymore. Nor does the front office put the major league roster together and then hand it off to the manager to do with as he will, often to the team’s detriment.
The degree of managerial autonomy may vary from team to team, but in every organization, the manager contributes to and benefits from a two-way flow of far more information than one person could consider. Some of that information isn’t accessible to the outside observers who scapegoat skippers; managers base decisions on proprietary matchup stats and knowledge of their players that the public doesn’t have. (In the wild-card round, fans criticized Boone for pinch-running for Luke Voit; predictably, Voit revealed after Game 5 that he’s been hobbled by plantar fasciitis.) Baseball’s best team, the Dodgers, operates as a collective that distills all of that intel into in-game marching orders (albeit not always with the desired results). The Rays’ front office and field staff are so tightly integrated that their former director of analytics, Jonathan Erlichman—a math major at Princeton—has served since last season as a uniformed “process and analytics” coach who assists Kevin Cash and sometimes sits in the dugout. That hasn’t seemed to set them back.
Front-office feedback may have helped Boone speed up the slow hooks that plagued him in 2018: In the final three games of this ALDS, he handled his bullpen with appropriate urgency. A large and influential analytics group is nothing new in New York, and the Yankees have won the most games of any AL team since their last title. Never mind that those statheads were the ones who recommended that the Yankees acquire D.J. LeMahieu and Luke Voit, who won the batting title and home-run crown, respectively, this season. It’s pretty rich to blame the Yankees’ loss on computers when the team that beat them—not just this week, but throughout the season—was the Rays. It’s not as if the Yankees were vanquished by a bunch of stat-hating gunslingers who went with their gut.
Yet the fact that the victorious team is, if anything, even more closely associated with the very tendencies that the Cashman/Boone-bashers deride hasn’t stopped the pile-on. Sure, sabermetric experiments may work for the Rays, the argument goes, but it’s unbecoming for the big, bad, Bombers to search for the same small edges. As The New York Times’ Tyler Kepner wrote, “It sure seems as if the Yankees tried to outsmart the Rays, which is kind of like the Rays trying to outspend the Yankees. The Rays do unusual things better than anyone, because that is their only lifeline. The Yankees can be smart and edgy, too, but they separate themselves with $300 million players like Cole and Giancarlo Stanton. The Yankees can—and probably should—win this series. But they need to do it their way. Let the Rays be the Rays.”
Set aside the fact that the Yankees did “do it their way” in Game 5, entrusting the series to the well-compensated Cole, Zack Britton, and Chapman, and the Rays beat them anyway. The suggestion seems to be that because the Yankees are rich, they shouldn’t also try to do the smart things that the Rays do. It’s fine for the paupers to use openers or four-man outfields, because they have to be creative. The Yankees can spend, so they don’t have to be clever.
It’s true that the Yankees shouldn’t imitate Tampa Bay in every respect. A rich person wouldn’t waste their time picking up coins that a cash-strapped person might stop to pocket, and unlike the Rays, the Yankees don’t need to trade their young stars before they reach free agency or avoid bidding on big-ticket players like Cole. But the Yankees and Rays are playing the same sport, and they’re both trying to outscore their opponents. If a way of deploying their players helps the Rays win, it might also help the Yankees win—and if the Yankees use it too, they take away Tampa Bay’s advantage. They can be both rich and intelligent, like the Dodgers under ex-Ray Andrew Friedman.
If the Yankees had pulled their Game 2 gambit instead of signing or starting Cole, I’d be picking up my pitchfork too. But they did sign Cole, and it wasn’t enough, just as Cole wasn’t enough for the 2018 or 2019 Astros. Wealthy as the Yankees are, they didn’t have a $300 million pitcher for every game. They had one, and they used him twice. I know alphas don’t do math, but surely they can calculate that there are three other games in a five-game series. Unless there was a $300 million superstar sitting on New York’s taxi squad, Boone had to make do with what he had. And just like the Rays, the Yankees identified the opener as a means of ameliorating a short-handed staff. Their blueprint didn’t pay off, but that doesn’t mean it was the wrong move or that a slightly different one would have worked better. No matter how or when the Yankees used García and Happ, they weren’t going to be favored in a game against Tyler Glasnow started by anyone other than Cole.
That thin staff was the weakness Yankees fans should be bemoaning. Their undoing wasn’t the way they used García and Happ; it was having to use them at all to pitch important innings. The Yankees had the high-end talent to match any team, but they lacked the depth of the Rays or the Dodgers. Although the Yankees had the best starting pitcher in the series, the Rays arguably had the second-, third-, fourth-, and fifth-best. No Yankees starter except Cole projected for an ERA below 4.27, a bar beaten by Morton, Glasnow, Blake Snell, and Yarbrough.
That difference in depth was apparent in the late innings of Game 5. Even if Chapman, who got four outs in Game 4, had escaped the eighth unscathed and pitched a scoreless ninth (in what would have been his longest outing since 2016), only a potentially gassed Chad Green would have stood between the Yankees and dice rolls. The Rays, who had Morton, Snell, John Curtiss, and Aaron Slegers in reserve, could have gone another nine innings without using a subpar pitcher. (More arms are on the way: As FanGraphs’ Dan Szymborski noted on Monday, the Rays, who’ve built baseball’s best farm system, employed a major-league-leading 38 pitchers in their organization this spring who projected to be better than replacement level.)
The Yankees found themselves in that situation for reasons largely outside of their control. After being sidelined for most of last season with shoulder and lat injuries, 2017-18 ace Luis Severino underwent Tommy John surgery in February. Domingo Germán, who helped replace Severino’s innings in 2019, was ineligible for this year because of a domestic violence suspension. James Paxton was shelved in late August after suffering a flexor strain, and crucial setup man Tommy Kahnle was also lost to Tommy John surgery. Any one of those pitchers could have changed this series and the Game 2 equation considerably.
Granted, Cashman deserves some side-eye for not filling those holes at the trade deadline. By August 31, Severino, Germán, and Kahnle were all out for the year, but for the second consecutive deadline, the Yankees didn’t make any moves. (Neither did the Rays, but thanks to their depth, they replaced their losses easily.) At the time, the Yankees’ inactivity was somewhat understandable. Paxton was still expected back before the end of the regular season, and the compressed postseason schedule, which exacerbated the Yankees’ pitching problems, had not yet been determined. In hindsight, perhaps the Yankees could have anticipated a setback for the fragile Paxton (who experienced soreness and was shut down again on September 10) or that the postseason schedule (which was announced on September 15) would feature fewer off days.
Even so, only one impact starting pitcher, Mike Clevinger, changed teams at the deadline, and Clevinger got hurt and was no help to the Padres this month. Would Taijuan Walker, Mike Minor, Robbie Ray, or Ross Stripling have been notably better than García and Happ? And would it have been worth paying a high prospect price for, say, Lance Lynn, considering the shortened regular season, the steep odds in favor of the field in a 16-team playoff format, and the pandemic-driven uncertainty that the season would be completed at all?
Maybe so; it’s much more reasonable to blame Cashman for complacency in August than to blame him, Boone, or some combination of both for their actions in October. Lynn is under contract for 2021, so he wouldn’t have been a short-term rental (which is why he wouldn’t have come cheap). Barring a trade for Lynn, a relief reinforcement, at least, was in order. But before tarring and feathering Cashman, remember that winning a World Series in the era of four playoff rounds is way harder than it was when the Yankees cleaned up from the 1920s to the 1960s. It’s also more difficult than the dynasty teams of 1996-2000 made it look. To Cashman’s credit, the Yankees haven’t had a losing season since 1992. That 28-season winning streak is only the second-longest in franchise (and MLB) history, but it’s easily the most impressive stretch of success in the context of its hyper-competitive time.
That streak isn’t likely to stop in 2021. The Yankees can expect to get Germán back next year, as well as Severino at some point and a full sophomore season from García. Given those three, Cole, Montgomery, and any other offseason additions the Yankees might make, they’re bound to be back in the playoffs next October. Which means we may be only one year away from the next twist of fate that determines whether fans spend the winter celebrating Boone or blaming him.
