Baseball 2021

Tonight is the first time that the 2021 baseball season could end. The Braves have won three games, a fourth victory concludes the World Series. This is always a bittersweet day for me – exciting because there could be a new champion, sad because it will be a long wait until meaningful games are played next spring.

[The grand slam that Adam Duvall just hit makes it much more likely that the season will end tonight.]

I watched a lot of games this summer.[1] First, it was nice to have full summer of games after last year’s pandemic-induced truncated season. Second, the Red Sox played well, much better than last season, and a bit better than expected. The Red Sox were quite brutal last year, winning only 40% of their games, which was fourth worst in the majors (there are 30 teams). This year they won almost 57% of their games, which only six teams bettered.  

[The Braves 4-run lead lasted less than two innings, though they have regained the lead 5-4.]

But the turnaround was not entirely unexpected, even to non-fans of the Red Sox. There is a baseball website (indeed there are so many good ones that even I can’t read them all) called Fangraphs,[2] which predicts how each player will perform and, based on that, how many games each team will win. They predicted that the Red Sox would win 88 games; they finished the season with 92 wins.

Below is a chart that shows (according to Fangraphs) how each team fared against its expectation. You will see extreme variability. This is not necessarily a fault of the prediction system. It is common for teams that play well (in the first half of the season) to improve their team with trades and therefore win more games than would have been expected based on the season-opening roster. And – teams that aren’t playing well often make trades that weaken their roster this year, while (hopefully) strengthening their team in the future.

TeamPredictedActualWins over
WinsWinsPrediction
Giants7810729
Mariners739017
Rays8310017
Brewers799516
Cardinals799011
Dodgers9610610
Reds76837
Astros89956
Rockies68746
Athletics81865
White Sox88935
Red Sox88924
Tigers73774
Blue Jays89912
Phillies80822
Braves8988-1
Cleveland8180-1
Pirates6561-4
Royals7874-4
Yankees9692-4
Marlins7267-5
Cubs7771-6
Angels8477-7
Rangers7160-11
Mets9177-14
Orioles6752-15
Padres9479-15
Twins8873-15
Nationals8365-18
Diamondbacks7452-22

As you can see, the Giants massively exceeded expectations in winning the most games in the league. But the Dodgers, who were predicted to win more games than any other team also substantially exceeded expectations. And even teams expected to be bad can underperform. The Orioles and Pirates were predicted to have the fewest wins in the AL and NL respectively (seemingly perennially) and they both managed to underperform.

This chart is a proxy for how each fan base feels about their team. Those with an overperformance, like the Mariners and Reds, probably have pretty happy fans. Even though they didn’t make the playoffs, they were playing games that mattered deep into September. On the other hand, fans of the Padres and Nationals, who began the season with championship aspirations, are rather despondent.

Over or under performing is all well and good and might influence a fan’s psyche over the course of a season. But what really matters is making the playoffs, and eight of the teams that made the playoffs exceeded expectations. Only the Braves (one game fewer) and the Yankees (four fewer) made the playoffs while winning fewer games than expected.

I was quite happy with the Red Sox season. They led their division for 74 days in the middle of the season. They won 46 games against teams with a winning record, only three teams won more. They dominated teams with a losing record, going 46-22, but this highlights a huge inequity in baseball scheduling. The Giants and Dodgers, who led with majors with 107 and 106 wins respectively, played 99 and 100 games against teams with losing records, at least 30 more games against weak competition than the Red Sox.

[The Astros now lead 9-5 in the eighth inning, which almost certainly means there will be a game six in Houston.]

To top off the season, the Red Sox played their archrivals the Yankees in the AL wildcard game and prevailed. Then they played the top team in their division in a best of five series. The current playoff system is relatively stupid. It rewards teams who win a weak division more than teams, like the Dodgers, who won the second most games in the majors. And in this case, it forced the Red Sox to play the Rays, whom they had already played 19 times.

Quick fix for playoffs system, no wild card game. Only the top four teams make the playoffs, without regard to division, and all series are best of seven. (I may have written about this before. If I did, it bears repeating.)

The Red Sox beat the Rays in four games and then fell to the Astros in six games. Then I started rooting against the Astros, who are something of a pariah team based on a cheating scandal from a couple of years ago. I’m looking forward to seeing how the Red Sox tweak their roster during the off-season.

[Astros won 9-5.]

And I’m looking forward to at least one more World Series game this season.


[1] Major League Baseball sells a cable package for a bit under $200, which provides access to almost every game played every day of the season, and which I consider about the best money I spend every year.

[2] https://www.fangraphs.com/depthcharts.aspx?position=Standings

2 thoughts on “Baseball 2021”

  1. If the Fangraphs predictions really are based solely, or even mainly, on early-season performance, there’s also the statistical artifact of “regression toward the mean” to consider. Francis Galton discovered it while looking at the height of parents and their children.

    Here’s my favorite example: Parents often report that rewarding their kids for good behavior doesn’t seem to work but punishing them for bad behavior does. It could be that neither carrots nor sticks work and the explanation is this simple: On any given day, a kid’s behavior is most likely to be close to that kid’s average. That includes the days immediately following the extreme days on which the kid was a little angel or completely out of control.

    I used this phenomenon to my advantage in a fantasy basketball league Bob organized many years ago. Especially around Thanksgiving, when the early returns were in, and the trading started to get into full swing. The other owners were unduly impressed by the poor, or superb, performance of some of their players in the first month or so of the season — when the sample size of games played was small and the variability in performance was considerable.

    Here’s a nice Scientific American article on the subject from the mid-70s, using batting averages of the likes of Roberto Clemente, Thurman Munson, Del Unser, Ron Swoboda, Rico Petrocelli, Max Alvis and Frank Robinson:
    https://www.math.drexel.edu/~tolya/EfronMorris.pdf

    Hope everyone and your families are well.

    1. That was a fun league while it lasted.

      The Fangraphs predictions are based on career stats, age, and expected playing time. So lots of room for error, but typically not because of small sample size.

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.