clock menu more-arrow no yes

Filed under:

Predicting the 2020 Yankees through the lens of 2019

New, 27 comments

By analyzing how teamwide stats project future performance, we can use 2019 statistics to see how the Yankees stack up against the top of the AL.

MLB: All Star Game-National League at American League Ken Blaze-USA TODAY Sports

It’s Thanksgiving Day, which only means one thing for Yankees fans: the hot stove should finally be about to heat up, and with it, we begin to leave the 2019 season further in the mirror as we look more in earnest toward 2020. As we wait for more projections to come out, however, we can look towards this past season to get a rough idea as to where the Yankees stand going forward.

A number of statistics can be used to try and forecast a team’s true talent level and project its performance in the following season, ranging from team-wide values like a team’s actual and Pythagorean record to aggregates of individual stats like WAR. But which ones have been the best forecasters?

In order to find out, I ran some numbers on AL teams over the past two seasons, to see how accurate the prior season’s stats would have been at predicting a team’s record during the 2018 and 2019 season, utilizing a team’s actual record, Pythagorean record, bWAR, and fWAR from the previous year.

Overall, these stats have done a middling job at projecting team performance, although there is some correlation in each case. FanGraphs’ version of WAR has been the most accurate predictor, with fWAR accounting for 62% of the variation in team’s record the following season. Actual record has performed the worst, with an R-squared of .499, meaning less than half of the variation in the team’s record the following season can be explained by their record the previous season.

Admittedly, these relatively low values should not come as a surprise, as they fail to incorporate roster moves that occur during the offseason, and instead assume that the level of talent will remain approximately the same; nonetheless, they do give us a relative baseline for the team’s performance.

According to these four formulas, the 2020 Yankees would project to win 99.4, 97.2, 94.8, and 98.5 games, according to actual record, Pythagorean record, bWAR, and fWAR, respectively. That’s certainly not a bad baseline to open up the offseason with. Unfortunately, the same can be said for a number of other top teams in the American League:

  • Houston*: 102.9 by actual, 104.8 by Pythagorean, 112.4 by bWAR, 112.8 by fWAR
  • Tampa Bay: 93.2, 91.5, 93.5, 95.8
  • Minnesota: 97.6, 95.3, 94.4, 102.8
  • Oakland: 94.2, 95.3, 95.4, 94.8
  • Cleveland: 90.6, 91.5, 86.6, 89.0

*Astros projections do not incorporate any fallout from the electronic sign-stealing scandal, which may have skewed results.

Ultimately, this exercise doesn’t tells us all that much that we didn’t already know. The Yankees, based on any measure of their performance in 2019, look strong heading into 2020, but they will still have a mighty tough fight ahead of them in the American League in 2020. Nonetheless, while this does not tell us how the offseason will shape up, Yankees fans can be thankful that their team looks primed to continue sizing up well at the top of the league, and with a few more additions, perhaps this team could soon profile as the league’s best.