In Game Six of the 2020 World Series—when Tampa Bay Rays manager Kevin Cash ran to the mound to remove dominant starter Blake Snell after 5.1 innings, 2 H, 9 Ks, and 73 pitches—my relationship with baseball analytics and the “old school” came full circle.
As a young baseball fan, I was fascinated by baseball history. I internalized all its players, myths, and numbers, focusing on the “classics” such as wins, home runs, batting average, and runs batted in.
Then, in the mid-2010s, I jumped aboard the analytics bandwagon (largely because the Twins seemed to have fallen so far behind the rest of the league and I wanted to know why). I read “Baseball Between The Numbers” and fell hard for concepts like shifting, launch angle, swing plane, and spin rate.
Much more recently—the last 1-2 years—I’ve begun to fear that analytics are turning the Great American Pastime into a homogeneous slog of walks, strikeouts, and an endless parade of pitchers from the pen to the mound. This came to a head in the now-infamous Cash/Snell conundrum.
The best way I can describe my current thinking about analytics within baseball is to compare it to playing poker:
When sitting down at the round table, it is 100% advisable to strictly play the odds. The feelings of the jacks, queens, or kings won’t be hurt if you slight them, and the general mood of the rest of the deck is moot. In other words, outside of their printed value the physical cards themselves—as inanimate objects—factor into the equation not one iota.
Contrast this with decisions made during a baseball game. When Cash hooked Snell, he was clearly playing the “pitcher is worse the third time through the batting order” card, if you will. An absolutely dispassionate evaluation of the numbers supports his decision (though even then, not by as much as one might think).
The problem? Unlike Vulcans, human beings are emotional creatures.
Taking one’s passion out of the equation is as impossible as it is inadvisable. While emotional decisions can often be poor ones, to be sure, the flow of causation can also be reversed. The entire tenor of that World Series game seemed to change as Snell cursed and walked towards the dugout. I texted my Dodger-fan brother right then and there saying “LA is going to win this game”. Even after the Dodgers did reign victorious, Mookie Betts flat-out stated that Snell’s quick hook felt like a boost to them.
The overall point here? It’s not that analytics are inherently bad. Far from it. After 100-some years of baseball being governed by a few basic principles, new thinking about certain topics has greatly improved the sport in both theoretical and tangible ways. But somewhere along the way, the human element has fallen by the wayside. Such a thing as, say, the boost an ace pitcher at his zenith can give a club.
Twins fans saw this play out with Baldelli and Berrios in the Houston series before it transpired on a much bigger stage a few weeks later. In the largest of possible moments, raw data-crunching prevailed over human intuition. Over a long sample size? This approach will pay off more times than not. In a singular moment that might decide the season? That could—and often has—turned out to be a very different story.