“This is, by far, the most interesting work I’ve done in years,” says John Aristotle Phillips, who once had the FBI come and take his blueprints for a nuclear bomb.

Phillips is now the CEO of PredictIt, a website where people are doing a pretty tremendous job of shielding themselves from the realities of this apocalyptic election by betting on it. When the world ends, some of them will die with a few thousand more American dollars literally burning holes in their pockets.

Some of the most dedicated of those PredictIt traders are all in this bar, The Grayson, on the Lower East Side of Manhattan on Super Tuesday, many of them trying not to spill their free Coors or Bud Lights on their MacBook Pros as results pour in.

They are traders, by the way, and not degenerate gamblers. At least not by definition. PredictIt is a prediction market with a hard cap, and not one of those DraftKings-style sites where robots steal money from poor people who watch a lot of SportsCenter. It’s technically operated out of Victoria University of Wellington in New Zealand, but who are they kidding: Their offices are in Washington, D.C.

PredictIt’s PR people probably do not want us to write that these people are here to make a bunch of cash on the outcome of this election, but these people are here to make a bunch of cash on the outcome of this election.

FiveThirtyEight’s Nate Silver swears by these things. In 2012, right before President Obama was reelected and Silver correctly predicted the outcome of every single state, he wrote, “I am far from an efficient-market hypothesis purist, but markets are tough to beat in most circumstances.” Back in 2012, it was a website called InTrade, which has since been shut down. InTrade nailed the right winner of all but one state in Barack Obama’s reelection.

Now, in the U.K., it’s BetFair, which is a straight-up gambling site that prohibits U.S. users from hopping in. PredictIt allows users—and, yes, Americans—to bet on the futures of a candidate on a dollar scale under the guise of research.

Here’s how it works: Think Donald Trump is going to win the whole thing? He’s at 29 cents at press time. Put 29 cents down today and you can win 71 more cents over the course of the election.

The potential problem comes when something like this becomes too popular. What if Nate Silver, knowing these things are often as pinpoint-accurate as he is, starts flouting them as a healthy barometer for who’s actually going to win this thing? Then what if Donald Trump, who has built his entire campaign on the phrase “Check out these polls” (at least until he made the race exclusively about his penis this week), uses those same stories to prove that he is, in fact, going to win this thing?

Does this not create a perfect feedback loop and/or Human Centipede by which gutless, angry pumpkin man Donald Trump cannot be stopped?

It got a little hairy when I brought up how this might, you know, actually affect the election with Phillips.

Phillips really did submit a plan for an A-bomb, by the way, as a term paper when he was a junior at Princeton in the ’70s. (He was also the school’s mascot.) Since then, he’s been working on the stuff that has redefined how candidates campaign. He and his brother created one of the first computerized voting lists in the ’80s, for example, and his companies have worked with every White House since Ronald Reagan’s.

He and the site’s spokesperson, Brandi Travis, both insist that the site isn’t really in it for the gamblin’ world domination stuff. They both stressed the community of the site. The comments section of each market contains pretty much no racial slurs, which is, in fact, a Nobel Peace Prize-worthy feat for an election website, and probably a sentiment that should be engraved on the tombstone of capitalism.

Also, each user can only lay $850 on each “market.” There are ways around this. You can bet on Trump for the nomination and against him in every state, for example. Those are all markets. But it would be hard to bet millions in an effort to mess with outcomes. You’d get caught.

That’s why the Commodity Futures Trading Commission served that university in New Zealand a no-action letter in October 2014, effectively allowing the place to form PredictIt as a nonprofit.

But here’s what has Phillips thinking this is better than the A-bomb: People turn out to these events. They drink the Coors Lights. They trade in public. And they don’t want to murder each other when they disagree about who should be the next president.

Seems impossible, but it’s true: People supporting all different kinds of candidates have had their vitriol neutered by the color of money. And it’s not just your standard Manhattanite. The Grayson featured a diverse group—Republican, Democrat, white, black, Hispanic. They were, of course, almost all young dudes, and they had collectively bought out the world’s supply of Ralph Lauren polo shirts, but they were not of all one stripe. It was kind of beautiful, really.

This capitalism tombstone inscription is getting awfully long, but I’m sure it can afford it.

“There’s real promise in this community,” says Phillips. They’re up to 17,000 active traders every month.

Here’s that promise personified in the corner of the bar: Ryann Williams. He’s headed to the University of Virginia in a few months to get his MBA. He’s on his MacBook in the corner of the bar telling me about the difference between limit orders and market orders. (The value is in limit orders, he says.)

Williams is looking at an absolutely indecipherable graph that must be very important and he’s telling me about how appalled he is by the concept of Donald Trump. Also, simultaneously, he is betting a bunch of money on Donald Trump.

“If he’s making you money, you can hedge against your heart,” he says.

Probably not a good one for the capitalism tombstone, but awfully telling.

“Donald Trump has won a lot of people a lot of money,” says Williams, and he includes himself in that group.

Williams has $5,000 spread out across a few different markets. One of them? Donald Trump, a mortal political enemy, winning the GOP nomination.

Had he kept the bet, he would’ve lost a little money between Tuesday and press day. Trump went all the way up to 84 cents on Super Tuesday, as the media narrative made it appear he had the whole thing locked up. Now, it looks like the RNC is trying to pull some funny business by tag-teaming this race into a brokered convention.

Mitt Romney, who could previously be had for a penny on the PredictIt market, is now up to 4 cents. Trump hit a low of 62 cents by Friday.

“I do not agree with him at all,” says Williams. “But the value is there.”

Don’t worry, he’s effectively shorting Trump after all. He doesn’t have him to win the whole thing, and he can sell him at anytime. He used to do this to Marco Rubio, for example. Buy him, because his stock tended to soar directly after debates, then sell afterwards when his stock was higher.

For robot reasons that have nothing to do with PredictIt’s computers, that backfired hard.

“That, uh, cost me some money,” he says.

Still, the victory will feel even more triumphant if Trump can be shorted one last time—one big, financial penis joke to end them all, The Biggest Short.

And that indecipherable graph in front of Williams might be showing Trump’s value or might be showing something else entirely—who knows? it looks like the world’s worst Jackson Pollock painting. But it was created by Jim Schmitz, and he’s here, too, on the other side of the table. It’s part of a Rube Goldberg machine he created in 2008—“a fully automated computer program that could make real money on its own by trading securities in a small market called the Iowa Electronic Markets.”

“If I’d figured it out 10 years earlier, I’d be a millionaire,” says Schmitz. “But I didn’t, and now I’m unemployed.”

Schmitz is here on a lark trying to figure out if he should switch from the Iowa markets. There’s not a lot of action there, and action is required in a place where someone has to be available to buy the shares you don’t want.

He thinks he’s about to come over and try it out. It’s not enough to make a living on, but he says it’s probably still worth it.

“It is. It’s for the love of the game,” he says.

This might be the only way to love this election, anyway: go full vice. Embrace the hell. A tiny-handed man who ran a failed Atlantic City casino might have suckered America’s profoundly stupid into letting him run the country, but you can still beat him at his own sketchy, barely legal game along the way.