I defend analytics

 Something happened a month ago that I just didn’t feel like talking about, even though it was in my wheelhouse.  It was something where I would defend a decision that everyone in the world was attacking, but I guess the attacks were so ubiquitous I figured what would be the point?  I love tilting at windmills but it gets tiresome.  But the attacks have gone on for a month and at some point my natural contrariness gets backed into a corner and has to come out fighting.

The event I am speaking of was the decision of Tampa Bay Rays manager Kevin Cash to pull starter Blake Snell from Games Six of the 2020 World Series in the sixth inning while he was pitching a 2-hit shutout.  The relief pitcher, Nick Anderson, proved less capable and the Dodgers won the game and the Series.

Cash subsequently won the AL Manager of the Year award (for his work during the regular season).  Snell has said he was disappointed by the decision and he now is the subject of trade rumors.  Last week on ESPN one of their personalities opined that the award for “Turkey of the Year” should not only be given to Cash, but named after him in perpetuity for making the dumbest decision of all time.

The decision was motivated by analytics, namely the fact that Snell had gone through the Dodgers’ line-up twice and that he was not as effective when facing batters for a third time.  The most vociferous attacks on Cash have come from the anti-analytics community, who see the Rays’ loss as conclusive proof that analytics are stupid.

First of all, using analytics to make decisions does not me that the decision is always going to work out 100% of the time.  Analytics is about probability, that over time you will more often come out ahead if you make rational decisions based on past observation of outcomes.  Part of the allure of non-analytics is that practitioners remember when their “hunches” paid off and forget all the times that their gut led them astray.

Second, decisions have to be executed.  If Nick Anderson had come in and pitched two shut out innings, we might not be having this conversation.  But he gave up a double to the first batter he faced, Mookie Betts, and the Rays’ fate was sealed.  News flash—Mookie Betts is a pretty good hitter.  Maybe if he had faced Snell for a third time, he would have timed a fastball and hit a home run.  Anderson’s subsequent failure to execute does not impact the decision made to pull Snell before he started facing batters for a third time, which had historically proven to be a bad idea.

What might have happened if Snell had been left in?  Let’s go back to the 2015 World Series.  Mets starter Mat Harvey is pitching a gem in Game 5, cruising after eight innings.  Harvey, who had only pitched one complete game in his entire history (and that was a blowout, not a close game), convinced his manager to ignore what the numbers said and to leave him in.  Mets’ manager Terry Collins eventually gave in and . . . the results were not good.  Harvey blew the game in the ninth, the Mets lost the game in extra innings, and thus lost the Series 4-1.  Analytics doesn’t look so bad now, does it?

As far as I can tell, people who don’t like analytics have one thing in common; they aren’t good at math.  They hate what they are incapable of understanding.  Analytics are responsible for getting the low-payroll Rays into the Series in the first place; analytics is the only weapon teams like the Rays and the “Moneyball” A’s have to compete against teams who can afford to make mistakes and overpay players who don’t work out. 

I concede that analytics has made sports less interesting.  Analytics says home runs are the most efficient way to score runs, so we have to put up with most at bats ending in a strikeout, base on balls, or home run.  In basketball, the mid-range jumper is dying because the most efficient strategy is to combine dunks with 3-point baskets.  In football, the short passing game is preferred to a “ground and pound” rushing offence or unleashing a mad bomber at quarterback because it has proven most effective.

But it is evolution, not heresy.  Yes, I miss the days when Bob Gibson would have punched his manager in the nose before giving up the ball before the ninth inning of a World Series game.  Yes, I miss the days when the 1971 Orioles had four 20-game winners on their pitching staff, when now there aren’t four 20-game winners in all of the majors (in 2019 there were exactly two).  Yes, I wish modern hitters listened to Wee Willie Keeler who said the secret to success was to “hit it where they ain’t,” instead of hitting into a shift when a bunt down the third base line would be an easy double. 

I miss those days, but they are in the past, not the future.