The Yankees and Twins, who’ll face off in the AL Division Series starting on Friday in the Bronx, produced nearly the same offensive stats during the regular season. Collectively, their hitters were separated by one home run, 3 points of batting average, 1 point of on-base percentage, and 4 points of slugging percentage. Their overall wRC+ marks were 1 point apart, and the gap between their hitter WAR totals, if we round up, was two wins, just like the gap between their team win-loss records. (The Yankees won 103 games, and the Twins won 101.)
Not quite as close: the total number of days their hitters spent on the injury list. Yankees hitters accumulated 1,149 days on the shelf, while Twins hitters spent only 374 days out of action. The gap in pitcher injury days was similarly lopsided: 1,171 vs. 255. Add those totals together, and the gulf looks like this:
No other team came within 600 days of the Yankees’ injury total this season, whereas only the Reds suffered fewer injuries than the Twins. The Yankees paid almost eight times as much money to injured players as the Twins did. (They also paid almost $20 million more to injured players than the Rays paid all their players combined.) The Yankees sent 30 players to the injured list, an MLB record, and Gleyber Torres, DJ LeMahieu, and Austin Romine were their only three position players to remain on the ever-revolving active roster all season. Even that trio nursed nagging injuries.
Some of the Yankees’ walking wounded got healthy just in time to earn a place on the ALDS roster. Luis Severino didn’t make his season debut until September 17, but he’s scheduled to start Game 3. Edwin Encarnación, acquired in June as an injury replacement, got hurt and missed most of August and September, but he’s in the lineup for Game 1 for the first time since September 1. So is Giancarlo Stanton, who played only nine games before returning to the active roster full time on September 18.
Even so, the byproduct of all of those regular-season injuries—along with some that have lingered into October—is that the Yankees’ division series squad reflects the team’s season-long look less closely than the rosters of the other seven division series survivors. Every division series roster comprises players who made at least 83.1 percent of the team’s regular-season plate appearances, except for the Yankees’ roster, which comes in at only 73.4 percent. In other words, more than a quarter of the Yankees’ regular-season plate appearances featured a player who isn’t active for the ALDS, whether because of an ongoing injury or because the big-name cavalry (Stanton, Severino, and so on) crowded out the stopgaps who had held the fort. Where have you gone, Mike Tauchman, Aaron Hicks, Clint Frazier, and Mike Ford, not to mention Opening Day starters Miguel Andújar and Greg Bird?
Relative to other teams’, the Yankees’ ALDS roster represents the regular-season club more faithfully on the pitching side, despite Severino’s late arrival and the absences of the injured CC Sabathia and the suspended Domingo Germán. Even so, the Yankees’ combined roster of ALDS hitters and pitchers accounts for only 71.9 percent of the team’s regular-season plate appearances and batters faced, which is still the lowest percentage in the playoff field. The Yankees of October resemble the Yankees of March through September less than the other October teams resemble their regular-season selves.
For the Yankees, that’s mostly a positive thing. With only a few exceptions, the players on the roster right now are the ones Aaron Boone and Brian Cashman would want if they had their pick of all the players under team control. A Yankees fan placed in suspended animation just before pitchers and catchers reported to spring training who awakened today would see a few surprises on the ALDS roster—no Dellin Betances, Gio Urshela and Encarnación starting—but wouldn’t suspect what strangeness had transpired in the intervening months.
The Yankees’ summer was extraordinarily strange. But the paradox of their regular season is that while in one way it looked nothing like it was supposed to—a ton of guys got hurt—in another, more important way, it looked exactly like it was supposed to. In February, before the injury cascade started, the ZiPS projection system foresaw the Yankees’ expected Opening Day roster accruing 51.1 FanGraphs wins above replacement. By the end of the season, dozens of IL stints later, the Yankees’ actual WAR came in under that preseason projection by all of 0.3 WAR. Essentially, the Yankees were exactly as good as the stats said they’d be, even though the people producing those stats turned out to be different.
That’s largely a testament to the Yankees’ depth and player development. If we update an exercise ESPN’s Sam Miller first tackled in August, we can see how the Yankees’ actual regular-season performance at each position (and overall) compared with both ZiPS’s published preseason projection and what the preseason projection would have been had the system somehow known how the Yankees’ plate appearances and batters faced would be distributed.
In the table below, the “Projection” column lists ZiPS’s estimate from February. The “Adjusted Projection” column uses ZiPS’s preseason player projections, but adjusts and weights them based on who actually received the playing time, so that Urshela makes a much larger impact on the projection than the injured Andújar. The “Actual” column shows the real result, and the last two columns show the differences between the actual and projected totals. The tiny difference between the team WAR total in the table (51.1) and the team total at FanGraphs (50.8) is attributable to rounding.
2019 Yankees Performance Compared With Projections (Actual and Adjusted)
The most salient stat here is the small difference between the initial projection for hitters (26.6 WAR) and the playing-time-adjusted projection (25.8), which suggests that the Yankees’ injury replacements on the offensive side were almost as good as the presumed first-stringers. Now, the actual difference should probably be a bit bigger than that, because some of the players who ended up seeing significant time at designated hitter were projected to play the field, and I didn’t account for the fact that their projected WARs would have been lower if ZiPS had classified them as DHs. Still, we’re talking about a projected difference of only a few wins, which seems almost miraculous in light of the extent of the attrition.
Some players whom ZiPS assigned playing time in February—including LeMahieu and Brett Gardner—weren’t expected to play or produce as much as they did, which explains some of that small gap. So does the presence of Encarnación, who didn’t start the season on the team. But that tiny difference is largely attributable to the depth the Yankees brought into the year. Every hitter who ended up playing for the team this season was projected to be better than replacement level before the season started, which is impressive considering how many replacements were required. Developing or acquiring superstars is an obvious way to win, and the Yankees stand out in that respect. But they also excel in avoiding the awful, a more subtle but still crucial roster-construction skill. The Yankees didn’t plan to rely on players like Luke Voit, Mike Tauchman, and Clint Frazier as much as they did, but those players projected to be good enough to start for some teams. This club was so stacked that LeMahieu, who ranked fourth in WAR among second basemen from 2015-18 and wound up within the top 20 in 2019 position-player WAR, was supposed to be a backup/utility type.
On top of that, of course, unexpected subs such as Tauchman, Urshela, and Cameron Maybin easily exceeded expectations, helping the Yankees vault over their projected position-player WAR total by several wins. Yes, the Yankees likely lucked out with those players to a certain extent, but they’ve also invested in the sport’s largest player-development department, and the information they provide to players helps them make improvements to their approaches and mechanics. Surpassing their position-player projection helped the Yankees make up for falling short of their pitching projection by an equivalent amount. Add some extra clutchness and fortuitous timing—the Yankees led all teams in cluster luck, and they beat their BaseRuns record by an MLB-best nine games—and you have a formula for an astonishing 103-win season.
Wins-wise, the Twins may be the biggest surprise of the season, but the way the Yankees reached their record isn’t far behind. The improbable path they took to October might have made them a postseason sweetheart if they weren’t, well, the Yankees. Boone has claimed that the gantlet the Yankees limped through to get to Game 1 made them stronger psychologically. Sabermetricians haven’t yet incorporated mental toughness into their projection systems, but with or without that purported edge, the Yankees’ current configuration makes them a formidable playoff opponent. The team’s turnover this summer makes it more difficult to discern its true talent, but according to the depth charts at FanGraphs, which parcel out playing time based on the remaining teams’ division series rosters, only the Astros’ current collection of players would project to be better over the course of a full season. But forget a full season. To bring their remarkable campaign to an equally exceptional conclusion, the Yankees need to be better—and, perhaps, healthier—than their rivals for only a few more weeks.