How Prediction Markets Tricked America Into Gambling

How Prediction Markets Tricked America Into Gambling

On February 28th, 2026, an anonymous account on a website called Poly Market placed a series of bets totaling over

half a million dollars that the United States would strike Iran within the next 24 hours. Hours later, US and Israeli

forces launched a coordinated attack on Thran, killing Iran's Supreme Leader.

That anonymous account cashed out $553,000 in profit, and nobody knows who they are. In 2025, prediction markets

processed over $44 billion in trading volume. And the platforms behind them are now valued at over $5 billion each.

But here's the part that doesn't sit right. These aren't called casinos.

They're called exchanges. So how did betting awards, elections, and basketball games become a Wall Street product instead of a legal sports book?

Poly Maros. The prediction mark. Prediction market.

Let's start with the actual bet on Poly market. You're not technically placing a wager. You're buying what's called an event contract. It works like this.

There's a question on the screen.

Something like, will the Fed cut interest rates in March? You can buy the yes side or the no side. The price of each side fluctuates between 0 and $1 based on what traders think will happen.

1:13

1 minute, 13 seconds

If you buy yes at 40 cents and you turn out to be right, you get a dollar back.

1:17

1 minute, 17 seconds

That's a 60 cit. If you're wrong, you lose your 40 cents. Simple enough. But when you zoom out and look at the

1:24

1 minute, 24 seconds

mechanics, it's functionally identical to sports betting. You pick an outcome,

1:29

1 minute, 29 seconds

you put money on it, you either win or lose. The only difference is the language. And that language is doing a lot of heavy lifting right now because

1:36

1 minute, 36 seconds

that's the entire reason this industry exists in a legal gray zone worth tens of billions of dollars. To understand how we got here, you actually have to go

1:45

1 minute, 45 seconds

back to 2001. After September 11th, the Pentagon's research agency, DARPA,

1:51

1 minute, 51 seconds

proposed something called the policy analysis market. The idea was to let people bet real money on geopolitical events like terrorist attacks and

1:59

1 minute, 59 seconds

assassinations. Because the theory was that crowds with money on the line would predict these things more accurately than intelligence agencies. The concept

2:07

2 minutes, 7 seconds

is built on something economists call the wisdom of crowds. When people have actual skin in the game, they tend to be more honest and more accurate than when

2:14

2 minutes, 14 seconds

they're just answering a poll. It makes sense in theory and in controlled academic settings, it actually works.

2:20

2 minutes, 20 seconds

Markets where participants wager real money have outperformed traditional polls in predicting election outcomes going back decades. The problem was the

2:28

2 minutes, 28 seconds

optics. The program lasted about 48 hours before Congress killed it. Two senators called it grotesque and unbelievably stupid. The guy who ran the

2:37

2 minutes, 37 seconds

program resigned and for the next decade, the idea of prediction market pretty much stayed locked away in academic world. But in 2014, a New

2:45

2 minutes, 45 seconds

Zealand university reached out to the Commodities Future Trading Commission,

2:49

2 minutes, 49 seconds

which is the federal agency that oversees futures and derivatives trading. They pitched something that sounded completely harmless. They wanted to run a small not for-p profofit

2:58

2 minutes, 58 seconds

prediction market called predicted as an academic research tool. The CFTC give predicted what's called a no action letter. Basically, the government

3:06

3 minutes, 6 seconds

saying, "We won't enforce regulations against you as long as you stay small."

3:10

3 minutes, 10 seconds

They set caps of 5,000 traders per market and $850 per person. The whole thing was treated like an academic experiment. Then in 2018, a pair of MIT

3:19

3 minutes, 19 seconds

graduates named Tarak Mansour and Luanna Lopez launched Kelchi with a much bigger ambition. They didn't just want an exemption, they wanted the full CFTC

3:28

3 minutes, 28 seconds

approval to operate as a regulated exchange, the same classification as the Chicago Merkantile Exchange, where institutional investors trade commodity

3:36

3 minutes, 36 seconds

futures worth trillions of dollars a year. And in 2021, they got it. That single regulatory decision is what cracked the whole thing open. Because

3:44

3 minutes, 44 seconds

once the CFTC said event contracts are derivatives and not gambling, it meant federal commodities law applied instead

3:52

3 minutes, 52 seconds

of state gambling law. And that meant Koshi could operate nationwide without needing a gambling license from any individual state. And this is where the

4:00

4 minutes

loophole kicks in. The same legal framework that lets farmers buy insurance against a bad harvest now lets you bet on whether your NBA team will

4:09

4 minutes, 9 seconds

cover the spread tonight. The CFTC treats both products the same way. A contract tied to an uncertain future outcome settled in cash. It doesn't

4:16

4 minutes, 16 seconds

matter if that outcome is the price of corn or the score of a basketball game.

4:20

4 minutes, 20 seconds

And because federal law preempts state law in this area, at least according to the CFTC, prediction markets argue they can ignore state gambling regulations

4:28

4 minutes, 28 seconds

entirely. Which is exactly why state like Nevada, Michigan, and Maryland are now actively suing these platforms.

4:34

4 minutes, 34 seconds

Nevada's gaming control board says Kelshi is running an unlicensed sports book. Kelshi says they are federally regulated exchange. And right now, the

4:42

4 minutes, 42 seconds

courts are split on who actually has authority here. Some states are winning,

4:46

4 minutes, 46 seconds

some are losing, and nobody has a definitive answer. But the legal question is almost secondary to the bigger problem that's already playing

4:54

4 minutes, 54 seconds

out in real time, which is insider trading. Because the Iran bet we talked about at the top of this video wasn't even the first time this happened. A few

5:02

5 minutes, 2 seconds

months earlier, an anonymous Poly Market user made over $400,000 by correctly predicting that the US would invade Venezuela and remove Maduro from power.

5:12

5 minutes, 12 seconds

A more than $400,000 mystery has been uncovered following the capture of former Venezuelan President Nicholas Maduro. Just hours before news actually

5:20

5 minutes, 20 seconds

broke of the military operation, an anonymous user on a prediction market placed a significant bet that Madura would be out of power by the end of the

5:28

5 minutes, 28 seconds

month. The perfectly timed wager is raising questions about possible insider knowledge with national security implications. The information about that

5:36

5 minutes, 36 seconds

invasion was so tightly held that Congress wasn't even notified in advance. A senator involved in new legislation to ban government officials

5:43

5 minutes, 43 seconds

from trading on these platforms pointed out that when something is that classified and someone bet six figures on the exact outcome, the math kind of

5:51

5 minutes, 51 seconds

speaks for itself. But finding the loophole was only half the battle. The other half was getting millions of people to actually use these platforms.

6:00

6 minutes

And that's where prediction markets borrowed a playbook straight from the sports betting industry. First, they went after distribution. In 2025, Robin

6:08

6 minutes, 8 seconds

Hood started offering Kelshi's contracts directly inside its app. That meant the same 24 million people already trading

6:15

6 minutes, 15 seconds

stocks and crypto on Robin Hood could now bet on NBA games and fed rate decisions without downloading anything new. Prediction markets went from

6:23

6 minutes, 23 seconds

something you had to seek out to something that showed up in the apps you were already using. Then came the content machine. Every time a major event happened, whether it was an

6:32

6 minutes, 32 seconds

election, a Fed meeting, or celebrity scandal, screenshots of parley market odds started flooding social media and cable news.

6:39

6 minutes, 39 seconds

While poll after poll showed the presidential election in a dead heat,

6:42

6 minutes, 42 seconds

look, this is our running average of the polling averages in the battleground states. What? They're all close.

6:48

6 minutes, 48 seconds

In the world of online prediction betting, a different picture emerged.

6:51

6 minutes, 51 seconds

Crypto gambling site Poly Market showing that a majority of users thought Donald Trump would win in the final weeks of the campaign. Supporters, including

6:59

6 minutes, 59 seconds

billionaire Elon Musk, citing those numbers, saying they were more accurate than polls.

7:04

7 minutes, 4 seconds

Either too early to call or too close to call.

7:06

7 minutes, 6 seconds

On election night, as our NBC News decision desk carefully analyzed the vote and was hours away from projecting Trump is the winner, Paulie Market

7:14

7 minutes, 14 seconds

declaring on X, the prophecy has been fulfilled.

7:17

7 minutes, 17 seconds

Reporters stopped saying poll show and started saying prediction market show.

7:22

7 minutes, 22 seconds

That kind of earned media is worth more than any ad campaign because it frames the platform as a source of truth rather than a place to gamble. And every time a

7:31

7 minutes, 31 seconds

news anchor puts a polymarket graphic on the screen, it is a free advertisement to millions of viewers who now associate the brand with being informed rather

7:39

7 minutes, 39 seconds

than being reckless. Finally, the onboarding. Koshi lets you sign up with just an email and start trading in minutes. No crypto wallet, no

7:46

7 minutes, 46 seconds

complicated setup. You deposit money the same way you'd fund a PayPal account.

7:51

7 minutes, 51 seconds

They made it as frictionless as possible because they know the hardest part of any gambling product is getting someone to place that first bet. After that,

7:58

7 minutes, 58 seconds

psychology takes over. And the psychology behind all of this is almost identical to what we've already seen in DraftKings and FanDuel. Prediction

8:07

8 minutes, 7 seconds

markets target the exact same demographic. Young, educated,

8:10

8 minutes, 10 seconds

overconfident men who believe they're smarter than the market. The framing is the key difference. Sports betting apps say bet. Prediction markets say trade.

8:19

8 minutes, 19 seconds

Sports betting apps have odds. Prediction markets have probabilities.

8:22

8 minutes, 22 seconds

And that reframing does something powerful to the brain. It gives the user psychological permission to wager more aggressively because it doesn't feel like gambling. It feels like investing.

8:32

8 minutes, 32 seconds

They're not a degenerate better losing money on parlays. They are sophisticated traders expressing a view on the market.

8:39

8 minutes, 39 seconds

But the dopamine loop is the exact same one. You have an opportunity to win something of value. You don't know when or how much you will win and you can

8:47

8 minutes, 47 seconds

immediately repeat the behavior. That's the same three-part scarcity loop that hooks slot machine players, Instagram scrollers, and day traders on Robin

8:55

8 minutes, 55 seconds

Hood. The packaging changed. The brain chemistry didn't. And honestly, [music]

8:58

8 minutes, 58 seconds

the prediction market version might be even more dangerous because you don't feel like you're gambling. You feel like you're being smart. And the data backs

9:06

9 minutes, 6 seconds

this up in a big way. Researchers analyzed over 86 million trades on Poly Market and found that only 0.51% of

9:13

9 minutes, 13 seconds

accounts have realized profit exceeding $1,000. 80% of participants are actually net lossers, which is almost the exact

9:22

9 minutes, 22 seconds

same win and loss ratio you see in traditional gambling. The house edge in prediction market isn't built into rigged odds. is built into the spread,

9:31

9 minutes, 31 seconds

the transaction fees, and the massive information is symmetry between professional traders and casual users.

9:37

9 minutes, 37 seconds

Because the person on the other side of your $50 bet on a Fed rate decision might be a hedge fund analyst who watches the bond market for 12 hours a

9:44

9 minutes, 44 seconds

day or it could be someone with the real insider information about a military strike as we saw in the Iran bets. And

9:52

9 minutes, 52 seconds

those people are going to win far more than you are. So, what does this mean for you if you're thinking about downloading one of these apps? First, understand what you're actually doing.

10:00

10 minutes

Buying an event contract on Koshi is functionally the same thing as placing a bet in a casino. The language is different, but the risk is identical. If you're wrong, your money is gone.

10:10

10 minutes, 10 seconds

There's no margin call, no stop-loss, no brokers calling you to discuss your portfolio. Your money just disappears.

10:17

10 minutes, 17 seconds

Second, be especially careful with anything tied to politics or geopolitics. The Iran and Venezuela incidents showed that people with

10:24

10 minutes, 24 seconds

insider information can and will trade on these platforms, and there are currently no safeguards to prevent this.

That bill introduced this week would only ban government officials themselves. It wouldn't stop their friends, their staffers, their adviserss, or anyone else who might be

in the loop. Third, if you do decide to participate, set a hard dollar limit before you start. Decide how much you're willing to lose, not how much you hope

to win. Treat that money as entertainment spending the same way you would a night at the casino because that's what it is. Every casino in

history was built on the gap between those two numbers. The fact that this one calls itself an exchange doesn't change the math. Prediction markets are

growing faster than any regulators can keep up with. The legal fights between state gambling commissions and federal commodities regulators will play out

over the next few years. And until those lines get clearly drawn, the people making the most money are the platforms collecting fees on every single trade,

whether you win or lose, which sounds a lot like the house always wins. They just found a smarter way to say it. Let me know what you think about prediction