How Is a Prediction Market Different from Betting?

In November 2024, a French trader placed a series of bets on Donald Trump winning the U.S. presidential election. By the time the results came in, he had reportedly made somewhere around $85 million in profit - on a crypto prediction market platform called Polymarket.

France's gambling regulator opened an investigation. The U.S. Commodity Futures Trading Commission had already fined Polymarket two years earlier for operating without a license. And yet Polymarket itself insisted it wasn't a gambling platform at all. It was a financial exchange.

So: was that $85 million a lucky bet, or a trade?

The honest answer is that the distinction between prediction markets and betting is real - but it's not the clean, bright line most explainers make it out to be. The difference is partly structural, partly legal, and partly depends on who's defining it.

Here's what actually separates them - and where the line gets genuinely blurry.

How are prediction markets different from betting? The short answer

A prediction market is a trading platform where you buy and sell contracts tied to real-world events. Each contract has a price between $0 and $1 (or 1¢ and 99¢). If the event happens and you hold a "Yes" contract, it pays out $1. If it doesn't, it pays $0. The price at any given moment - say, 65¢ - reflects what the collective pool of traders currently believes the probability of that outcome to be.

Traditional sports betting works differently. You accept odds posted by a bookmaker - odds the bookmaker sets, adjusts, and profits from. Your payout is locked in at the moment you place the bet. There's no trading, no price discovery, and no exit until the event resolves. The bookmaker builds in a margin - called the vig (short for vigorish), also known as the overround - so the odds always favor the house slightly.

The structural difference, in one sentence: in a prediction market, you trade with other participants; in a sportsbook, you bet against the house.

That difference has consequences for pricing, costs, flexibility, and how the law treats each product - which is what the rest of this article is about.

How sportsbooks work

When you place a bet at a sportsbook - DraftKings, FanDuel, a Vegas window - you're accepting a price the bookmaker has set. Take a standard NFL spread bet. You'll typically see both sides listed at -110, meaning you need to bet $110 to win $100. If both sides implied 50% probability each, the total would be 100%. But -110/-110 implies roughly 52.4% on each side, which adds up to 104.8%. That extra 4.8% is the bookmaker's margin - called the vig (short for vigorish) or overround. It's the built-in cost of using the sportsbook, embedded invisibly in the odds themselves.

Once you place a bet, you're locked in. Fixed-odds sportsbooks don't let you exit mid-event. The price you got is the price you get. Your counterparty is the bookmaker, whose goal is to balance the book and collect the vig regardless of outcome.

How prediction markets work

On a prediction market like Kalshi or Polymarket, you're not betting against a house. You're buying a contract - specifically, a Yes or No share on a question like "Will the Fed cut rates in June?" or "Will the Lakers win the title?"

The price of that share, say 42¢, represents the market's current implied probability that the answer is Yes. If you buy a Yes share at 42¢ and the event happens, the contract resolves at $1.00 - a gain of 58¢. If it doesn't happen, the contract resolves at $0 - a loss of 42¢.

Three things separate this from a sportsbook:

The price is formed by the market, not set by a firm. When you buy a Yes share at 42¢, someone else is selling it to you - another trader who believes the probability is lower than 42%. The price moves continuously as new information arrives and traders update their beliefs. Polymarket's documentation explains the mechanics: outcome shares are priced between $0.00 and $1.00, with trading itself setting the price.

You can exit before resolution. You bought Yes at 42¢ and the probability jumps to 70¢ because new information came in? You can sell - lock in your gain and walk away before the event resolves. This is fundamentally different from a one-shot wager.

The cost is transparent. Rather than embedding margin in the price (as sportsbooks do), prediction markets typically charge explicit exchange-style fees. Kalshi publishes a taker fee schedule that varies by contract price. Polymarket charges fees that vary by market type. PredictIt charges a 10% fee on profits when selling a position at a gain. None of these fee structures are hidden - they're disclosed separately from the market price. One useful way to frame it: sportsbooks hide the cost in the odds; prediction markets show the cost in a fee schedule. Transparent doesn't mean cheap, though - in thinly traded markets, the spread between Yes and No prices can exceed a sportsbook's vig.

Where the line gets blurry: betting exchanges

Here's the part most explainers skip - and it matters.

Not all "betting" looks like a sportsbook. Platforms like Betfair operate as betting exchanges: instead of betting against the house, users bet against each other. You can "back" an outcome (bet it will happen) or "lay" it (bet it won't). Prices are set by supply and demand. You can exit positions mid-event.

That structure is, mechanically, almost identical to a prediction market.

The difference between a Betfair-style exchange and a Polymarket-style prediction market is largely regulatory and cultural, not structural. Betfair is regulated as a gambling operator. Polymarket is registered as a derivatives exchange. The underlying trading mechanism is similar enough that the conceptual line between "prediction market" and "betting exchange" is a matter of legal classification more than product design.

This is the cleanest way to hold the distinction in your head: "betting" describes the intent - putting money on an uncertain outcome. "Prediction market" more often describes a trading mechanism and a probability-like price. The same underlying product can be classified as either depending on how it's structured, where it's registered, and what jurisdiction is doing the regulating.

What the price actually means - and when it lies

One of the most common statements about prediction markets is that "the price equals the probability." A contract trading at 65¢ means the market thinks there's a 65% chance the event happens.

That's a useful mental model. It's also not exactly true.

Academic economists have been careful about this for two decades. Justin Wolfers and Eric Zitzewitz, in their foundational 2004 paper "Prediction Markets" in the Journal of Economic Perspectives, argued that prediction market prices are often close to the average belief among participants - but they can be distorted by fees, risk aversion, budget constraints, and limited participation. Prices are useful but biased proxies for probability, not perfect meters.

Charles Manski, an economist at Northwestern, showed in formal models that the market price doesn't have to reflect what the average participant actually believes - budget differences, risk aversion, and who's even allowed to trade all pull the number away from a clean probability.

For sportsbooks, the math is slightly different but the problem is the same. Implied probabilities can be extracted from odds, but they don't add up to 100% - they add up to more, because the bookmaker's margin is embedded. Clarke (2017) and others have formalized methods to remove the overround when extracting clean probability estimates from odds.

The practical takeaway: prediction market prices are a money-weighted, real-time opinion signal - not a guaranteed forecast. Sportsbook odds are a product price: a probability-like number with a house edge built in. Both are useful. Neither is a truth machine.

The regulatory reality: same question, different answer

Here's where it gets strange: whether something is gambling or a financial contract in the U.S. has less to do with what it is than with who's regulating it.

The Commodity Futures Trading Commission (CFTC) regulates U.S. derivatives markets. It has treated many event contracts as swaps or derivatives under the Commodity Exchange Act framework. Kalshi was designated as a CFTC-registered contract market in 2020. Polymarket, which is offshore and crypto-based, was fined $1.4 million by the CFTC in January 2022 for operating unregistered event-based binary options and accepting U.S. customers without authorization.

Sports betting, by contrast, is regulated at the state level. After the U.S. Supreme Court's May 14, 2018 decision in Murphy v. NCAA struck down the federal law (PASPA) that had blocked most states from legalizing sports betting, states began licensing and regulating sportsbooks individually.

That split has produced real conflict. In the "election contracts" litigation, a D.C. district court opinion issued September 12, 2024 vacated the CFTC's prohibition of Kalshi's congressional-control election markets - ruling that the CFTC, not state gambling regulators, had jurisdiction. The D.C. Circuit declined to stay that ruling in October 2024. In May 2025, Reuters reported the CFTC dropped its appeal, leaving Kalshi's election markets intact.

Federal jurisdiction, for now, held.

But state-level resistance has intensified. In March 2026, Arizona filed criminal charges against Kalshi, alleging its sports contracts constitute illegal gambling under state law. Kalshi's position is that its event contracts fall under exclusive federal oversight as CFTC-regulated derivatives. The CFTC reaffirmed its claim of exclusive jurisdiction over prediction markets in a February 2026 filing. On March 12, 2026, the CFTC initiated a public comment process through an Advanced Notice of Proposed Rulemaking seeking input on how to regulate prediction markets going forward.

The upshot: the same Yes/No question - "Will the Chiefs win the Super Bowl?" - can be a legally traded financial contract on one platform and an illegal gambling wager on another, depending on the platform's registration and which law applies.

The bottom line

Prediction markets and traditional betting share the same DNA - you're putting money on an uncertain future outcome. But the structure underneath is different enough to matter.

In a sportsbook, you're a customer accepting a price. In a prediction market, you're a trader in a market. One gives you fixed odds and a locked-in bet. The other gives you a position you can build, hold, and exit as information changes.

Whether that makes prediction markets "not gambling" depends entirely on who's answering the question. Regulators are still arguing about it. Courts are still deciding. The $85 million French trader is still, presumably, watching how it all plays out.

What it means for you: the products work differently, they cost differently, and they're subject to different rules depending on where you live and which platform you use.