Three weeks ago, on the final “Super Tuesday” of 2016, it looked like Bernie Sanders was cooked. That night, the Vermont senator lost all five contests, each of them in a key, delegate-rich state. Hillary Clinton’s delegate lead, which was already daunting, began to look insurmountable. While no one was exactly calling on Sanders to drop out of the race, he was urged by many to help reorient his massive fan base toward the general election and beyond—to spend his energy attacking Donald Trump and not Clinton, and to begin doing the work necessary to maintain the political movement he started despite his defeat. Vox’s Andrew Prokop referred to Sanders’s chance of winning the nominations as being “incredibly implausible,” while my colleague Jeet Heer urged Sanders to stay in the race—but to focus on remaking the Democratic Party, not winning its nomination.

But after tonight’s big win in Wisconsin—Sanders was topping Clinton by 12 points with two-thirds of the precincts reporting—the Vermont senator has won six of the last seven Democratic contests, including a commanding victory in the delegate-rich state of Washington. He should win again when voters in Wyoming caucus on Saturday, giving him victories in 15 states to Clinton’s 18. Which means that Sanders will head to the next big-state primary, in New York on April 19, with the wind at his back, having stolen the momentum from Clinton. Right?

Maybe—but probably not. In politics, as in sports, narrative momentum is often confused with something more substantial. And Sanders’s path to the nomination remains implausible (though possibly not incredibly so); he still has a much better chance of remaking the Democratic Party than he does of becoming president. 

Following his disastrous performance three weeks ago, Sanders has gone on a nice run and won a few states by commanding margins, but hasn’t been able to substantially cut into Clinton’s delegate lead, or to really make up for that disastrous March 15 performance. Clinton won that day, as she has for most of this campaign, in big states—Florida and Ohio being the most notable—whereas Sanders’s run of victories over the last few weeks has come (with the exception of Washington) in smaller ones. 

Because the Democratic Party awards all of its delegates proportionately, Sanders can win a lot more small states by big margins, as he has throughout this campaign, and still not really cut into Clinton’s lead. Here’s how stark the math looks: On March 15, Clinton won 472 delegates; Sanders, meanwhile, has accumulated just 155 in his last six victories. His win tonight in Wisconsin further closes the gap (a double-digit victory should mean that Sanders will win a solid chunk of Wisconsin’s 86 delegates), but from the perspective of delegate math, the race is not substantially different than it was on March 15. The only way for Sanders to overtake Clinton would be to win every upcoming state—especially New York, Pennsylvania, and California—by overwhelming margins. Sanders’s victory on Tuesday caps an impressive run, but nothing from it—neither the margin nor the demographics of Sanders’s supporters, still overwhelmingly white—suggests that Sanders is poised to do that.

Narrative momentum, of course, is harder to gauge. Sanders has won a lot of states in a row, but they were states he was expected to win, in many cases by very large margins—many of them were holding caucuses, for instance (ten of Sanders’s fifteen victories have come in caucus states). Sanders hasn’t won a state unexpectedly since his shocking victory in Michigan on March 8. The week after Michigan, in fact, was the last time that Sanders looked like he had the kind of momentum he has coming out of Wisconsin. And that ended with him striking out on March 15.

Tonight’s victory in Wisconsin looks like it will be slightly larger than expected—projections had Sanders winning by 2.6 points—but Sanders still has not shown any sign that he’s winning over black voters: In Wisconsin, exit polls suggest that Clinton won 74 percent of the black vote. In 2008, 30 percent of New York Democratic primary voters were nonwhite, suggesting Clinton could see a bigger boost there than she has in states like Wisconsin. New York is also a closed primary, meaning only registered Democrats can vote, cutting off one of Sanders’s biggest bases of support: independents

The Sanders campaign has been trapped in a vicious cycle: Hillary Clinton takes a commanding lead and her voters become less enthusiastic, sensing that she is the inevitable nominee. Sanders’s voters are always enthusiastic, so they hit the polls. Sanders wins states, after which his policies are scrutinized more intensely (see, for instance, his recent interview with the New York Daily News, in which Sanders struggled to explain how he would accomplish many of his policies). Clinton’s voters become more energized, and she wins states. Rinse and repeat. 

This pattern has dogged Sanders every time he’s seen a lift in this campaign, starting when he won New Hampshire by a massive margin, and then lost Nevada and South Carolina by massive margins. Sanders hasn’t been able to sustain momentum in this campaign, and if he’s to have any shot at winning the nomination, that’s exactly what he needs to do: He has to keep winning, by larger margins than he did on Tuesday, in states that mostly do not look like Wisconsin. Unfortunately for Sanders, that hasn’t happened yet, and while New York polls may be tilting in a favorable direction, there’s still no indication that Sanders is about to channel his inner Daenarys Targaryen and break the wheel.

But narrative momentum has its uses. While earlier New York polls had him trailing by nearly 50 points, a recent poll showed Sanders within something that could be described as striking distance, 12 points behind. And a massive, diverse rally in the South Bronx last week showed the Sanders campaign aggressively courting a more diverse electorate. 

The Wisconsin victory will boost Sanders’s case because the media craves something that looks like a close race: On CNN tonight, there was even giddy talk of there being two contested conventions. New Republic contributor David Dayen predicted shortly after March 15 that Sanders was headed for a string of victories that “will give the traditional media an opportunity to herald a ‘fundamental change’ in the Democratic primary, even though that change will be largely geographic. You will surely hear pundits talking about how Clinton ‘can’t close the deal,’ and hours of theater criticism about her stump speeches, when the truth is that we’re just hitting an air pocket of pro-Sanders states. The delegate math will remain relatively unforgiving.” 

That string of victories has happened now, and we will continue to hear that kind of punditry for the next two weeks. It may make the race seem closer, but it might very well also energize Clinton’s base: The cycle turns again.

Sanders will put everything he has into New York, because he has no choice: Either win it by an overwhelming margin, or he can’t plausibly continue to paint himself as a contender for the nomination. Sanders knows this. Speaking after his victory in Wisconsin, he said, “Please keep this a secret—do not tell Secretary Hillary Clinton, she’s getting a little nervous and I don’t want her to get more nervous!—but I think we have an excellent chance to win New York and a lot of delegates in that state!” And at a rally in Wisconsin on Monday, Sanders said “I think if we win here, we win New York State, we’re on our way to the White House.” 

Given the delegate math, Sanders would have been more accurate to say: “If we win here, we win New York State, and if we win Pennsylvania, California, and New Jersey by gargantuan, unexpected margins, and then also convince superdelegates to switch over, we’re on our way to the White House.” But none of those things is conceivable without that big surprise in Clinton’s home state. For Sanders, the win in Wisconsin does make it slightly more plausible. But no more than that.