There’s a lot of data to sift through when considering the race to serve the next term as the junior senator from the great state of Texas. There are the polls, which currently suggest that Ted Cruz is going to continue serving in the role he’s held since 2012—at least if the electorate they’re polling reflects the people who will be voting in 2018, and not just the people who voted in previous elections. There are the fundraising numbers, which suggest that Beto O’Rourke’s campaign is a kind of unprecedented juggernaut that it’s hard to properly predict. There’s the data crunched by professional prognosticators like Nate Silver, whose algorithmic approach to analyzing those numbers grants O’Rourke a roughly 1-in-4 chance of winning the seat, and the less scientific analysis from well-regarded sources like the Cook Political Report, which categorizes the race as a toss-up.

And then there are the people whose job it is to make money off of the predictions other people are making: the bookies.

In Texas’s last high-profile statewide race, between Wendy Davis and Greg Abbott, it was impossible to find a bookie anywhere who would give you odds on Davis winning a term as governor. Silver, by Election Day, staked Davis’s chances at winning at 1-in-166. But in the race between Cruz and O’Rourke, offshore gambling sites like Sportsbook.ag, Bovada, and Paddy Power are all in on taking money from people who want to bet on the outcome.

Oddsmakers, of course, have a financial incentive to get it right. If they offered even odds—where a $100 bet wins you $100—people who wanted to take advantage of Cruz’s lead in polling would load up on that bet, while people who believe in O’Rourke despite his underdog chances wouldn’t stand to have their faith proportionally rewarded.

The job of a bookie is to ensure that enough money gets placed on each side of a bet that, no matter the outcome, the house ends up making money. In order to do that, they set the odds in such a way that it’s enticing to people who believe in each outcome. In the case of the Senate race, all three sites put the odds roughly the same: Cruz is between -525 and -600 to win—which means that you have to bet $525-600 in order to win $100 back—while O’Rourke is between +350 and +400, so a bet of $100 will bring you back $350-400 should he pull off the upset.

That doesn’t mean the oddsmakers are always right. When I spoke to Paddy Power spokesman Lee Price in January around that site’s betting odds for Amazon’s HQ2, he told me, “We don’t know anything that you or your readers don’t.” Oddsmakers “sit in a room and crunch numbers, based on what’s on the news, and logic, rather than inside information.” For some perspective: before polls closed on Election Night 2016, placing a bet on Trump winning would have netted you a higher return than betting on O’Rourke right now. The biggest difference between a bookie and the person refreshing FiveThirtyEight every twenty minutes is that they’re financially incentivized to be dispassionate.

At Paddy Power and Sportsbook.ag, the only Senate race they’re taking odds on is the one in Texas. That reflects two things: one, there’s a very high level of interest in the race, and two, it is a legitimately competitive contest. (Bovada is taking odds on many of the seats up in November, most of which reflect closer races than Texas’s.) If Cruz’s victory were considered a sure thing, you wouldn’t see so much action on it—nobody is taking odds on whether Mitt Romney wins in Utah, for example.

The 2018 Texas Senate race is only one place where O’Rourke’s name shows up on gambling sites, though. The other place you’ll see it is in the field of potential 2020 presidential candidates.

While the Senate gambling odds mostly reflect the polls, the presidential field is significantly ahead of the polling in terms of enthusiasm for O’Rourke. And while it’s absurdly early to attempt to handicap the 2020 campaign, CNN attempted to do so over the weekend:

O’Rourke is polling at 4 percent in a poll that’s based, largely, on name identification—if the person who answers the phone doesn’t know who Kirsten Gillibrand or Eric Garcetti is, those potential candidates are unlikely to do well, while someone like John Kerry, who was the Democratic nominee in 2004, polls at 5 percent even though you’d struggle to find a passionate base of supporters for him. One would expect names like Joe Biden, Bernie Sanders, and John Kerry to do well in those polls, since they’ve all been involved in presidential politics in the past. When a figure who hasn’t played in that arena before does well—like California senator Kamala Harris, New Jersey senator Cory Booker, or O’Rourke—that’s slightly more instructive.

The oddsmakers get that, too. The specific odds that any site will give you on these candidates vary widely—Paddy Power has O’Rourke for president at +800, while Bovada has him at +2000—but the order that the candidates end up in is pretty similar. Harris leads the field on both sites, while Sanders, Biden, and Elizabeth Warren fill the next three spots. Paddy Power goes next to California lieutenant governor Gavin Newsom, and then to O’Rourke; on Bovada, Cory Booker, Kirsten Gillibrand, and Oprah Winfrey (!) come in before O’Rourke does. That tells us a few things: one, if you feel bullish that Beto O’Rourke is going to be the Democratic nominee for president in 2020, you should place that bet at Bovada, and two, O’Rourke is pretty high on the list of candidates people actually want to put money on.

None of this is to say that O’Rourke is likely to be nominated for president in 2020—as Texas Monthly‘s Eric Benson noted in August, O’Rourke’s not even a particularly likely candidate, and for him to even consider running, he’d have to lose in November. But while the people who lay the odds are dispassionate observers of the data, the people who place the bets are not, and in a wide-open field like this one, general enthusiasm tends to correspond with lower odds—the same way that even bad Dallas Cowboys teams tend to open the season with better odds to win the Super Bowl than good Jacksonville Jaguars teams. Oddsmakers use their heads; gamblers bet with their hearts. That’s pretty much the basis for the entire industry.

You can try to use that information to read anything you’d like into the odds that are out there. But the safest takeaways from the gambling odds are simpler. The bookmakers give O’Rourke roughly the same odds to win in November as the San Francisco 49ers have to win tonight against the Packers. Those numbers aren’t necessarily encouraging for O’Rourke’s supporters, but they’re infinitely better than the odds Wendy Davis faced at this point in 2014, when no reputable bookie would take your money on the race. Among the 2020 presidential field, meanwhile, the people who are passionate enough to bet on it seem to like O’Rourke a lot more than the people answering CNN’s polls. What any of this means for the future is anybody’s guess, but if you feel confident enough in your analysis to put your money where mouth is, there are a lot of places on the internet that will let you do it.

(Note: A previous version of this story had a paragraph where the + and – signs denoting the odds for each candidate were inverted.)