Political journalism suffers from a central contradiction: elections are finicky things, but the best way for a commentator to make a name for himself is to project as much confidence as he can. The collection of confidence can take many forms: journalists can position themselves as monarchs of gossip; they can embed with campaigns and provide a look from the inside; they can simply plug their ears and yell louder than the next guy. The key to staying in the game is to never allow the actual outcome of an election to change the way you go about your business.

After the 2012 Presidential election, political media had a moment when it seemed like that confidence game might finally come to an end. If you worked in the news business in any capacity after Nate Silver correctly called all fifty states in 2012, you likely remember feeling desperate to catch up to the new paradigm. Breezy D.C. reportage seemed to be on the outs. The new political journalism, made in Silver’s image, would be empirically driven, anti-narrative, and downstream from the revolutions that were happening in the front offices of sports franchises across the country, where decades of scouting wisdom and eye-test impressions were being replaced by the hard logic of numbers and a proliferation of acronyms: war, xfip, P.E.R., and D.V.O.A. Now, the pollster revolution that Silver started faces an uncertain future: he recently announced that he will be leaving ABC News, which has been home to his forecasting operation FiveThirtyEight for the past decade, after his current contract expires. Silver told me that his last day at ABC News will be at the end of this month.

The mania for data-based political coverage—evidenced by the arms race to build FiveThirtyEight imitators like the Washington Post’s Monkey Cage or The Upshot at the Times, the newspaper with which FiveThirtyEight was affiliated until 2013—was driven, at least in some part, by an imbalanced understanding of what the numbers were telling us. “People kind of liked it for the wrong reasons,” Silver told me. FiveThirtyEight, he explained, came out of a “specific tradition of gambling and forecasting,” which, when done properly, was mostly a tool to measure odds. But the conditions of the 2012 election, which Silver described as “boring” when compared with 2016 and 2020, and the emotional investment that Democrats had in Barack Obama turned Silver into someone they saw as an oracle who bore only good news.

This was certainly great for Silver’s career, but it wasn’t exactly the point of all those election models, nor was it a sustainable way to manage expectations. It’s likely that much of the readership of the Times was sincerely interested in poll aggregation, but there was also an undeniable appeal to prognostications that assured readers that Obama was going to defeat John McCain, and that any story about debate performances or shifts in donor priorities could be summarily dismissed as a vestige of an old way of talking about elections. The problem, of course, is that when your liberal audience wants you to provide only the number that allays its fears, you actually need the Democrats to keep winning. “The minute you have a forecast where there’s less certainty, people don’t like that,” Silver said. “The minute you have a forecast that doesn’t have a Democrat winning, they don’t like that very much.”

Public fights between young insurgents and the old guard often play out in the media. In baseball, team ownership quickly capitulated to the new “Moneyball” regime. But team owners don’t control who gets to be a local columnist, nor do they get to cast the voices on talking-head shows. So, as every team in the league handed its operations over to a cast of younger, nerdier dudes with charts, the data-converted masses began a coup against the old guard of sports journalists who still believed in antiquated stats like pitcher wins and R.B.I.s. The divide felt alluringly political. The quants were cast as progressives who could see the sport’s objective truths, which, at the time, tended to favor highly productive Latino players who might have been ignored or even derided by the old baseball press. A choice was presented: You could either stick with the crusty columnists who still relied on racially coded adjectives such as “cerebral” and “blue-collar” to describe players, or you joined the enlightened new order that understood that David Eckstein, the slap-hitting second baseman for the St. Louis Cardinals who received undue plaudits for his grit and hustle, was actually terrible at baseball.

The analytics revolution in the sport happened fast, but the media infighting lasted much longer. By 2010, the nerds had already won, but, if you looked at the way the quant-converted fans were carrying on, you’d have thought that Murray Chass ran every baseball team in the world and told all his players to bunt and never, ever, take ball four. The disruptors, in other words, had become the establishment, but weren’t exactly acting like it.

Silver’s data revolution has followed some of those same patterns. During FiveThirtyEight’s rise, between 2008 and 2012, Silver became the hero for everyone who was tired of traditional, bloated election punditry. “We were the insurgents within the New York Times,” Silver said. “We were implicitly critiquing their media coverage of the Presidential race.” That enviable status, in which one is both part of a venerated institution and free to point out the fossils in the room, gave Silver the trust of the Times’ reading public and also allowed those readers to air their grievances. The quants have undoubtedly improved political coverage—just like the baseball-analytics journalists, who, despite being incredibly annoying at times, were a welcome reprieve from the fusty writers who paraded around press boxes for decades with their Baseball Writers’ Association of America lanyards around their necks. But, like them, Silver ultimately became an institution unto himself; this, in turn, made him a target for criticism.