Sonny Gray, the Boston Red Sox pitcher who led the majors in wins this season, made his playoff début in 2013, with the Oakland A’s, a team that won ninety-six games despite having the fourth-lowest payroll in baseball. That team found sustained success in large part by making decisions based on data-driven analysis—including the decision to start Gray that day, rather than Bartolo Colón, a former Cy Young Award winner who, at forty years old, had won eighteen games that season. “We have a lot of smart people in our front office and in our baseball operations,” the team’s manager, Bob Melvin, said to reporters before the game, “and the short of it is it came down to Sonny and the last game he pitched in similar conditions in our ballpark. So that’s the route we’re going to go.” Gray rewarded his manager—and the smart people in the front office—with an old-fashioned pitchers’ duel, throwing a hundred and eleven pitches in eight scoreless innings, giving up only four hits and striking out nine. Justin Verlander threw a hundred and seventeen pitches for the Detroit Tigers and gave up four hits in seven scoreless innings. The A’s won it in the ninth.

The game was already changing when Gray entered the league; it’s changed even more dramatically since then. It’s been almost a decade since teams started using “openers”—pitchers who would stay on the mound for an inning or two before a reliever came in to eat up the lion’s share—or platoons of relievers, chosen to maximize well-studied matchups. Now what was once an interesting experiment has become routine. Late in the 2026 season, the Miami Marlins began bringing starters into games in relief on days when they’d more typically have a practice session, so as not to “waste” those extra pitches. (It’s worth noting that serious injuries have come to seem almost inevitable for big-time pitchers, as they throw harder and harder and put nastier spin on the ball.) In Game One of the A.L. Wild Card Series (the first game of the post-season!), the Houston Astros and Chicago White Sox both played so-called bullpen games, without any of their usual rotation pitchers on the mound. After the White Sox won that series, they started their Division Series against the Cleveland Guardians with another one.

Using relievers in this way isn’t a totally new idea, and there is some common sense in addition to data to support it. The more times a hitter sees a pitcher, the more comfortable he’s likely to be at the plate. And certain pitchers perform better against certain batters. Some teams, using the kind of analytical approach that the A’s pioneered, now focus on those sorts of matchups rather than relying on starters to go deep into games. (The team that really popularized using the bullpen aggressively, the Tampa Bay Rays, has managed repeated success despite a minuscule payroll.) The approach has been so effective—and so disruptive to the game’s flow—that, before the 2020 season, Major League Baseball instituted a new rule requiring that any pitcher brought into a game must face at least three batters and imposed a limit on the number of pitchers a roster can carry.

Some thought the rules didn’t go far enough—including Verlander, Gray’s matchup in his first playoff start. Verlander hated the trends diminishing the value of starting pitching. Before he retired, at the end of this season, he offered some ideas for how the league could preserve the art, including requiring a starting pitcher who has pitched fewer than ninety pitches to stay on the mound into the seventh inning. (He allowed some caveats for issues like injuries and blowouts.) It’s almost laughable how far from the norm that has become. When Verlander won the Cy Young in 2011, he threw a hundred pitches or more in all thirty-four of his starts. This season, the two locks for the Cy Young, the New York Yankees’ Cam Schlittler and the Milwaukee Brewers’ Jacob Misiorowski, threw at least a hundred pitches in a combined total of eight games, out of sixty-two starts.

At thirty-six, Gray, who started Game Two of the Wild Card Series between the Sox and the Yankees on Wednesday, is closer in age to Verlander than to Schlittler, who’d beaten the Red Sox in Game One. Still, Gray had one of the best seasons of his career in 2026 and played a central role in getting the Sox to the post-season. And partway through the fifth inning he was still pitching well, having thrown only seventy pitches and given up one run, on a solo shot out of the park. But with two outs, and runners on first and third, the Red Sox manager Chad Tracy pulled him. Gray walked back out to the dugout with his square jaw set in a line of displeasure.

In his place came Erik Miller. There was some logic to the switch. Gray had given up a few hits, and the next batter, Cody Bellinger, had already had a couple of threatening at-bats. The Yankees batters were moving through the order for the dreaded third time. And the Red Sox bullpen had been the best in the major leagues. Presumably, the Red Sox’s analytics department had detailed, proprietary analysis that informed Tracy’s decision to pull his starter when he did. There was a plan in place, Tracy explained afterward. “It just didn’t work out,” he added.

A central tenet of the sabermetrics revolution is that choices shouldn’t be judged simply by outcomes, which are determined by all sorts of factors, including random chance. They should be judged by “the process,” as the moneyballing general manager Billy Beane liked to say. Beane, as it happens, was the general manager who drafted Sonny Gray to the A’s. But never mind. Gray was furious with the process that got him pulled, and with the outcome. So were Red Sox fans, after Bellinger stroked a low slider into right field to put the Yankees up by two, triggering a 9–2 rout that sent the Yankees on to the Division Series and the Red Sox out of the playoffs. “I think I’m the best pitcher in the American League going through the order the third time. I think that’s a fact,” Gray said afterward. “Yeah, I mean, I would have liked to live or die with me in that moment, but that’s the competitor I am. And yeah, I think you can probably ask the guys, and they felt the same way.” He added, “I would have liked to have been trusted in that moment.”

A reporter did ask one of the guys, Willson Contreras. He mimed zipping his lips. But he didn’t keep his mouth shut entirely. “It’s not enough when you have talent, or you have data or you have computers telling you what to do. This is baseball, and the data is here to help us,” Contreras said. “But, at the end of the day, you’re going to need experience and veteran guys to win games.”

Ornery guys who think the nerds have gone too far say this kind of stuff all the time. So do competitive athletes who’ve just lost big games. And, of course, not even the most dogmatic stat-heads believe that stats should be treated as immutable laws. What’s more, there were data-oriented factors that might have argued against bringing in Miller, too. (Most obviously: Bellinger was especially good at hitting sliders with runners on base, and Miller depended on his slider; plus, Miller had been rocked in a similar situation the night before, when he also relieved the night’s starter in a tight game and surrendered an immediate home run to Ben Rice.) You could argue that Tracy should have relied on data more, not less.

Their opponent, the Yankees, in fact, went in the other direction—no doubt guided by data of their own. They carried ten pitchers on their roster against the Red Sox, instead of the usual twelve or thirteen. And in Game One, instead of pulling Schlittler, who was throwing a gem, the team’s manager, Aaron Boone, left him on the mound for a hundred and seventeen pitches—the most by any Yankee in more than four years. Then Max Fried threw ninety-eight in Game Two. The Yankees have had the best rotation this season, and we’re likely to see some of their starters in relief as the post-season goes on. Starting pitching has been “the biggest strength for us throughout the year,” Boone said. “It sustained us.”

But Gray, the Red Sox pitcher, was making an emotional appeal as well as a statistical one. It was live or die, as Gray had said. It was about trusting the human side.

And why not? Sure, the goal of the game is to win, not to satisfy some ancient sporting code, and the insights delivered by the data revolution in sports have helped teams do that—including teams that didn’t have big payrolls. They’ve also improved our understanding of the game in myriad ways. They have unlocked new pitches and new approaches, and have engendered fantastic creativity. But the experiences of the humans playing and watching sports should matter, too. Otherwise, what’s the point?

For me, the highlight of the season so far has been watching Misiorowski, the Brewers’ ace, blaze fastballs across the plate over and over. Misiorowski was throwing seemingly every pitch a hundred miles per hour, if not a hundred and four. When I wrote about him back in May, I did it with a sense of urgency. I figured he couldn’t keep it up. He’d dial back the velocity, or his team would use him more sparingly to protect him for the playoffs, or his elbow would explode. No one had ever done what he was doing. And no one, I thought, would be allowed to do what he was doing for long.

I am very glad to have been wrong. Misiorowski finished with the best earned-run average in baseball, the most strikeouts, and the lowest average number of base runners allowed per inning. He didn’t pitch as many innings as someone like Verlander used to, which might be our loss. But he pushed limits of another sort. He threw more than a thousand fastballs over a hundred miles an hour—nearly five hundred more than anyone else. He made being a starting pitcher look heroic. And the Brewers, as it happened, finished with the best record in baseball.

It wasn’t all due to their ace’s exploits. Misiorowski leads his team in wins above replacement (WAR), a stat that measures a player’s impact on winning, alongside the second baseman, Brice Turang, who edges ahead or falls behind depending on the formula. As Ben Lindbergh points out in the Ringer, even the best pitchers have become much less valuable than the top hitters, because of how little they play. But I’m not here to watch a WAR leaderboard. I am here to see what a human can do, and what kind of pressure a person can survive.

The Brewers are one of those small-market teams whose success has depended on its front office’s ability to pinpoint market inefficiencies and on its use of new player-development data to turn its draft picks into stars. They are, in the era of the Los Angeles Dodgers, the sort of team that’s not supposed to exist anymore. They are not swollen with talent. They don’t hit a lot of home runs. Instead, their batters make good contact. They don’t swing and miss. They’re fast on the base paths. They have a manager who emphasizes trust. They have a bona-fide star—a comet who might someday burn up. Let him light up the sky, I say. We need it all—the decisions grounded in data, and the shooting stars. ♦