Over the first quarter of the season, the Yankees have struggled in a very specific, but remarkably consistent, subset of games: the final game of a series. So far this season, the Yankees have played in 14 series, winning eight, losing four, and tying two. In the final game of these series, however, the team is one of the worst in the league, going 4-10 in these games.
So what has caused this rather strange trend? After staring at box scores and stat lines for hours, I came down with two possible reasons for this trend that deserved further analysis: load management and pitching matchups.
The concept of load management is that teams ought to rest players — particularly those who recently recovered from an injury or are susceptible to injuries, such as Aaron Judge — on a regular basis in order to keep them in the lineup on a long-term basis. While I’m not going to get into a discussion here as to whether or not this is beneficial, the perception throughout this short season is that the Yankees tend to use the final game of a series as a “punt” game, resting multiple major contributors.
Does this theory hold water? Let’s take a look at the data:
Now, it’s hard to draw any conclusions from this data — after all, Mike Ford, Tyler Wade, and Brett Gardner, the three players that we would expect to see their playing time increase if starters were receiving scheduled rest on the final games of series, have had to fill in as the starter for a few days due to injuries to Luke Voit, Gleyber Torres, and Aaron Hicks; this skews the data. However, this does suggest two players to take a look at, DJ LeMahieu and Giancarlo Stanton.
So far this season, LeMahieu has sat twice, on April 14 against the Toronto Blue Jays and April 25 against Cleveland, both of which were the final game of a series. Stanton, on the other hand, sat four times before hitting the injured list, two of which were the final game of the series (April 4 and April 14 against Toronto), one of which was the first game of a series (April 22 against Cleveland), and one of which was mid-series (May 1 against Detroit). These two players suggest that there is some truth to the theory that the Yankees tend to prefer to sit players on the final day of a series. It is not such a statistically-significant trend that we can say that this is a definitive plan, and not merely how the schedule has fallen — not to mention how injuries have affected the planned rotations. At the very least, it’s unlikely that “resting players” would result in such a disparity in performance in these games.
When I first dove into this topic, I assumed that, once I looked at the pitching matchups, I would see that the Yankees drew the short straw and the other team’s aces tended to be lined up to start the final game of the series, matching up against Jameson Taillon. The reality, however, is a bit more complicated than that:
(Note: ERA is at start of game)
The first thing I see here is that there are two Gerrit Cole starts that the Yankees easily should have won based on the pitching matchup, but that an ineffective offense and a bullpen implosion caused to be losses. Winning these two games would increase their record in these games to 6-8, as well as put their record in the last six at 4-2, both of which would be close enough to their season records that it would be statistically insignificant. Furthermore, the game-time ERAs can be misleading, even beyond the fact that ERA is a very imperfect stat: Ian Anderson’s ERA has dropped a full 1.5 runs to 3.20, and Rich Hill was in the middle of a scoreless inning streak of more than 20 innings and that stretched over four starts.
In these games, the Yankees have been outscored 63-40, a stat all the more astounding when you consider it includes two games in which Yankees pitching shut out its opponents. Within these 14 games, Yankees pitching gave up seven runs twice and both nine and ten runs once apiece. These four games account for fully half the games that the Yankees have surrendered more than six runs in a game in 2021; additionally, 10 of the 21 games in which the Yankees surrendered four or more runs are found in this set, while only 4 of the 23 games in which they surrendered three runs or fewer occurred in the final game of a set. Unfortunately, there do not seem to be any similarities within the games to suggest a reason for these bad starts, beyond the fact that they’re bad.
All in all, despite already being two months into the season, it does not seem like there’s enough concrete data to make any statements about the Yankees’ performance in the final game of a series, for although the team has definitively played worse in these games than they have elsewhere, there does not seem to be any discernable cause for these trends. Perhaps, much like the team’s 5-10 start overall, a larger sample size will start to reveal some normalization.