This article has a few interesting tidbits, but overall it (inadvertently and unintentionally) mocks baseball analytics rather than praising them. It contains a whole lot that should make any analytically-minded baseball fan either laugh or cringe.
It furthers a host of misconceptions, with one of the biggest being the first paragraph: it was very clear that the analytics movement (broadly defined) had already won well before this World Series. The trend of front office hires over the past several years has made that clear: people like Jeff Luhnow in Houston, David Stearns in Milwaukee, Andrew Friedman moving to the Dodgers, and many others.
What’s particularly galling, and serves to further a host of other misconceptions, is that Jazayerli cites some facts and stats that any analytically-minded baseball fan knows to range between poor examples and garbage.
“An objective, data-driven view … can get up off the mat after a 3–1 series deficit in the championship round.” Anyone with an inkling of statistical understanding of baseball understands that the Cubs winning 3 straight games against another playoff-caliber team was an improbable outcome. That’s not to say that it can’t happen, and obviously it did happen. So good for the Cubs players going out and executing. I don’t want to diminish that they actually went out and did it.
After being down 3–1, though, an “objective, data-driven view” was that the Cubs’ chances to win the World Series were well under 50%. IIRC, Fangraphs detailed projections had the Cubs at about 20% odds to come back from down 3–1. That’s actually a testament to the strength of the Cubs team, because the odds if the three games were coin flips would only be 12.5%. And, yes, it was improbable for a less talented Indians team to get a 3–1 lead in the series, but that had already happened: those wins were banked.
He also cites season records — including partial season performance— as if it’s predictive; furthermore, he does so without noting any underlying stats (run differential, etc.). The Braves went 37–35 after the All-Star Break in 2016 (improbably, both because of the relative lack of talent on their roster and because they were outscored by 23 runs). What do you say, Dr. Jazayerli? Do you think that we should project the currently constructed Braves roster to be an over .500 team next year?
Also, the Cubs did have a “shutdown closer” when they traded for Chapman: Hector Rondon. This may be a case where the team had a lot more information than we did, because Rondon went on the disabled list with a triceps injury a few weeks later. And Chapman, in general, alternated between gas can and fireman in his post-season appearances: he preserved three key 1 run leads (highlighted by his 2+ inning outing in Game 5 of the World Series) but also coughed up two big leads (highlighted by Game 7 of the World Series). His performance highlights the unpredictability of short series baseball.
The key insight from understanding both baseball analytics and post-season baseball is this one: sometimes the objectively best (most talented) team wins the World Series, but more often than not it doesn’t. This year it happens that it did, even though the Cubs went into the postseason with something like a 20% to 25% probability of winning the World Series. But even a powerhouse team goes into each baseball series as a relatively narrow favorite — there simply aren’t NBA level favorites in baseball series — so the best way to win a World Series is to build really good teams year after year. There are edges to he had, but baseball is a game where a series of good decisions build up to the point where a team still only wins a bit over 60% of its games against all other MLB teams (and less than that against playoff-caliber teams).
This may sound harsh, but this article undermines rather than advances the understanding of baseball analytics for any reader not already familiar with the subject.