Bernie Sanders has spent the past two years losing steam. He ended the 2016 primary with 43 percent of the national vote, but now he’s polling below 20 percent nationally. His preferred candidates generally performed poorly in down-ballot Democratic primaries in 2018. And, unlike Hillary Clinton, Mitt Romney and other past second-place finishers, he hasn’t been able to scare off politicians like Elizabeth Warren who might try to compete for his voters.

But Sanders just might have an advantage that some of his opponents don’t: He’s well suited to capitalize on chaos. Sanders might not be able to win a one-on-one fight with a well-known, establishment-backed, consensus-building candidate like Clinton, but his particular mix of strengths and weaknesses could make him a solid candidate in a completely bananas multiway brawl.

Sanders outperformed expectations in 2016, but he was, in some ways, always destined to lose.

Clinton started the race with a better organization, more backing from the establishment, a longer history with the Democratic Party and more experience running a presidential campaign. Sanders built a solid coalition of millennials, progressives and anti-Clinton hard-liners, but that’s not enough to get to 50 percent plus one in a somewhat ideologically diverse and increasingly racially diverse party. Clinton could have tried to bury him with negative ads or roll out more opposition research, but she never felt threatened enough to do that. She had the luxury of taking the high road because she was always winning. And if she had really started losing or was disqualified by something like a health event anytime before Super Tuesday, another consensus Democrat like Joe Biden probably could have jumped in, gotten the party’s support and won the nomination.

The 2020 primary could become that sort of clean two-person contest, but right now it doesn’t look like that. It looks more like the 2016 Republican primary, where everybody is running and nobody is the heir apparent. And Sanders has a few advantages in that sort of an environment.

Unlike most of his competitors, Sanders has a real and established base. He’s currently in second place in the polls, and it’s safe to assume that at least some of these people are part of the 43 percent he had last time around. And he could grow his 2020 base by simply targeting some of the people who voted for him last time around.

Having a base is a huge advantage in a crowded primary. These sorts of races sometimes have a boom-and-bust feel to them — an unknown candidate grabs some media coverage, catches fire and then withers under scrutiny or just flames out as voters move on. That happened to Rick Perry, Rick Santorum, Newt Gingrich and Herman Cain at various points in the 2012 Republican primary. But it didn’t happen to Donald Trump in 2016 because he, like Sanders, had followers who weren’t just jumping on a bandwagon.

And in a chaotic primary environment, Sanders’s base could give him a solid number of delegates and maybe even democratic legitimacy.

That’s because Democratic primary rules have traditionally been proportional with a floor of 15 percent — meaning that candidates who can’t get above 15 percent of the vote get basically zero delegates, and the candidates who do get over that hump divide the delegates up proportionally among themselves. So if there are only two candidates who get 60 percent and 40 percent of the vote, they’ll get about 60 percent and 40 percent of the delegates, respectively. And if two candidates get 30 percent and 20 percent while the rest of the field is under the threshold, those candidates will still divide the delegates roughly proportionally and get 60 percent and 40 percent each. (You can find a step-by-step guide to the math here.)

These rules make it hard for Sanders to win a one-on-one race with an establishment candidate like Clinton. But if the field splits up in a weird way — say, a solid chunk of the vote goes to candidates who don’t hit the threshold, a few candidates get above 15 percent but Sanders is leading with 30 to 40 percent of the vote — his delegate total could start to inflate. Democratic rules don’t reward front-runners in the same way that Republican rules do: The GOP often has winner-take-all states and other rules that give the winners disproportionate shares of the delegates. But circumstances like this could allow Sanders to get more than 30 percent of delegates with only 30 percent of the vote.

More importantly, a situation like this could give Sanders democratic legitimacy. If the primary field has more than two or three strong candidates with solid followings, it’ll be hard for any of them to get the 50-percent-plus-one delegates they’ll need to lock down the nomination. In that case, the candidate who won the most votes — which very well might be Sanders — could correctly claim some amount of small-d-democratic legitimacy. It’s hard to deny the nomination to the candidate who gets the most votes.

Sanders has other advantages, too — he has a huge small-dollar donor network, experience running in a presidential primary, and the ability to put together large rallies and get media attention. I think I’ve been underestimating him this cycle, and it’s definitely possible to imagine scenarios where he ends up with the nomination.

But that doesn’t make Sanders the front-runner. This is an important needle to thread when you’re thinking about any Democratic candidate’s odds of winning. Sanders might have a better shot in 2020 than he did in 2016, but his win probability is still low. No candidate (Sanders included) is anywhere close to a 50 percent win probability in my view (for what it’s worth, my educated guesstimate is that he has a 15 to 20 percent win probability). The race is chaotic — and that chaos has caused me to become more bullish on Sanders than I have been in the past. But the smart money is still on the rest of the field.