Polymarket: Financial market or gambling?

By Jack Liu and Aiona Gambucci

Source: Stock Picture

For centuries, people have tried to predict the future. Today, they can trade it. On Polymarket, users buy and sell shares in future events. From elections to weather forecasts, users can turn their beliefs into market prices. Prediction markets are not new: early versions emerged in academic settings in the late 1980s. Yet Polymarket brought them into the mainstream during the 2024 United States presidential election, when its prices pointed more clearly towards a Trump victory than many traditional polls. But if market odds can inform public expectations, they can also distort them. Is Polymarket a financial market or simply gambling in disguise?

Where it all started

In 1988, one of the first commercialised prediction markets was launched by the University of Iowa. Originally, users entered small trades on political outcomes (Wolfers et al., 2004). In the 2000s, the firm Intrade started offering bets on events other than politics, proving that prediction markets are accurate and self-correcting. Through the 2024 United States presidential election, newer platforms such as Kalshi, Polymarket, and PredictIt gained popularity among users (Montevirgen, 2026). While Kalshi and PredictIt are still in use, Polymarket is the most liquid prediction market based on traded volume. 

Trading mechanism behind Polymarket

Polymarket is a decentralised betting platform built on blockchain, where users can bet on whether an event will happen or not. However, Polymarket is not a traditional betting platform; it’s a prediction market. It is a stock market for future events where users can not only place their bets but also trade them. Instead of buying Apple stocks, users buy shares of whether it will rain tomorrow. Anyone with a crypto wallet can participate. 

Users can basically bet on anything, from sports to finance to politics. This is how it works: they buy shares of "Yes" or "No" of a bet with the price ranging from $0.00 to $1. The price reflects the probability of the occurrence of the event, given by the traders. If a share costs $0.80, wagerers believe that the event will happen with a likelihood of 80%. The payment is done with the stablecoin USDC.

Unlike traditional betting sites, Polymarket does not determine the odds of a bet; users do through trading. The price of a share climbs when there is a high demand. If users think the share is overpriced and the probability is overestimated, they can sell their shares. Price evolutions are visible in real time, allowing users to observe how market expectations change. This helps users to compare across different predictions of an event what others think the most likely outcome is. For example, for the weather forecast in Amsterdam on May 29, the “yes” share price of 27°C is higher than the one for 30°C. Polymarket charges taker fees in several market categories, while geopolitical and world event markets remain fee-free. Part of the collected fees is used to reward market makers for providing liquidity.

At the resolution date, trading halts: winners are awarded $1 per share, and losing shares become worthless. Traders of the two sides compete against each other. The value is simply transferred from losing shares to winning ones. So there is no house in this case, and the direction of the outcome does not financially affect Polymarket. Now who decides which side is right or wrong? When it comes to resolving markets, Polymarket has a unique way to do it that even differentiates it from modern competitors like Kalshi. Instead of the house announcing the correct results, Polymarket outsources this task to its users and to a service called UMA. First, any user can submit the result by placing a bond. If undisputed for 2 hours, this becomes official. However, if disputed by another user, then UMA token holders vote on the results. Nevertheless, Polymarket has stepped in once in a while and overridden UMA’s decisions.

What can trigger market price movements?

Polymarket’s market-based pricing is central to its appeal: probabilities are not set by a bookmaker, but emerge from the buying and selling decisions of users. However, this mechanism also creates a vulnerability. If a single trader or a small group of traders places a large enough bet, the displayed probability can change substantially, even without any new public information about the event itself. This became controversial during the 2024 United States presidential election, when four accounts linked to a French trader accumulated large positions on betting on a Trump victory. On 7 October 2024, Trump’s implied probability on Polymarket rose from around 50% to 53.3%, despite no clear new fundamental information being released that day (Morrow, 2024). Although the trader’s prediction ended up being correct, the case raised an important question: does a Polymarket price always reflect the judgement of a broad crowd, or can it be shaped disproportionately by a few wealthy participants? The controversy contributed to restrictions on access to Polymarket in France (Loumagne, 2026).

How herd behaviour affects the accuracy

The French case illustrates how prediction markets may be affected by herd behaviour, even when the final market prediction turns out to be correct. Prediction markets are often justified through the “wisdom of crowds”: prices are expected to aggregate the independent information and beliefs of many participants (Lin et al., 2013). However, this logic becomes weaker when traders no longer act independently but instead respond to the actions of others. On Polymarket, price movements are visible in real time. When a large trader places a substantial bet and shifts the probability of an event, other users may interpret this movement as a signal that the trader has more information than the average. They may then follow the price movement rather than relying on their own analysis.

Importantly, herd behaviour does not automatically mean that the market prediction is wrong. In the 2024 United States presidential election, Polymarket predicted the correct winner. Nevertheless, the market may have arrived at this prediction through a combination of independent information, large individual positions, speculation, and imitation. Less-informed participants may behave like noise traders, while a smaller number of informed or highly skilled traders move prices closer to the eventual outcome. Therefore, Polymarket’s accuracy should not automatically be interpreted as proof that its prices always represent the independent judgement of the crowd. Instead, the French case suggests that prediction markets can be informative while still being vulnerable to behavioural biases and disproportionate influence from a small number of traders (Wolfers et al., 2004; Lin et al., 2013).

Who are the winners

Most of the profits are actually concentrated in a very tiny group of people. Indeed, 67% of the gains are captured by 0.1% of the users, according to the Wall Street Journal (2026). This represents around 2,000 out of 1.6 million active traders who won roughly $500 million in total. The big factor is the skill gap among the traders in this prediction market: regular people like students compete against professional gamblers and trading firms. Obviously, most do not stand any chance against highly skilled and better-informed traders, so a small group profits off of the ignorance of the inexperienced majority of users. This high concentration contradicts the wisdom of the crowd theory. When a mere 0.1% dominates the market, then we are rather talking about the wisdom of experts.

Cases of insider trading and manipulation

For better-informed traders, there is nothing wrong with using their knowledge and research to gain a competitive edge. However, using confidential information to trade on certain prediction markets may be a form of fraud. On Polymarket, insider trading is made especially easy. When you open an account, no verification of your identity at all is required. And funding with cryptocurrency keeps your identity anonymous. Recently, there have been numerous reports of accounts with a suspicious trading pattern. Since Polymarket is built on the Polygon blockchain, all transactions are public. For example, a recently created account bet large sums only on US military operations related to the attack on Venezuela in late 2025. This did not go unnoticed on social media. It turned out that this trader was a US soldier involved in that war; therefore, he knew the results in advance. Leaks like this pose a threat to national security. The soldier basically revealed the secret war plans to the world, which could potentially jeopardize the whole operation.

Other famous insider trading cases also show the risk of market manipulation. Sometimes, traders can have a direct influence on the outcome. Early this year, a wagerer allegedly used a hairdryer to manipulate a temperature sensor in Paris in order to win the weather temperature bet. Thus, he won around €20,000 on Polymarket. Or in another instance, a wagerer invaded the pitch during the Super Bowl after placing the bet for that. These examples show the power the individual user has over this market.

When it comes to cheating, no method is too extreme. During the conflict in Iran in March 2026, gamblers allegedly threatened a journalist to edit his article to their favour. One trader bet $23 million that Iran would attack Israel on 10th March. Wagerers wanted the writer to tell that the Iranian missile fragments hit Israeli soil instead of the Iranian missile being intercepted.

At the same time, informed trading can improve the accuracy of prediction markets, creating a difficult distinction between valuable information aggregation and the unfair use of confidential knowledge.

How fair are market arbiters?

Unlike sports betting, where the outcomes are clear and undisputable, real-life events are not always black and white. Indeed, it is not always easy to think of outcomes only as “yes” or “no” due to their chaotic nature. When you can bet on basically anything, no amount of rules can cover every scenario and satisfy everyone. If Progrès Niederkorn wins 3:1 in a match, nobody will doubt this actually happened. On the other hand, when it comes to the bet that the United States would invade Venezuela, it is a matter of definition. What exactly counts as an “invasion”? Here comes UMA into play, whose role is to solve this conflict.

For this particular invasion case, Polymarket even changed the rule of a market after it had already started. The platform added that kidnapping the Venezuelan president Maduro did not count as an invasion. It argued that an invasion is the act of taking control of the country, so this kidnapping alone did not qualify. Therefore, this market remains unresolved and Polymarket has not proceeded to pay out any winners yet. Of course, wagerers who massively bet on an invasion were furious and the “yes” share price plunged. 

There are also some controversies around the UMA judges. For a decentralised platform, the voting power is heavily concentrated in a tiny minority of UMA token holders. The more tokens one owns, the more votes one has. There are around 6,400 UMA token holders, but nine of them hold more than half of the tokens according to Bloomberg. So most often, they dictate the “truth” as they account for more than half of the votes, especially when they collude and pick the same side. 

Additionally, UMA does not ban its voters from participating on Polymarket to prevent  conflicts of interest and relies on their good faith to stay neutral and unbiased. In one in five disputes, judges also traded the bet and therefore had a financial incentive to manipulate the votes to their favour. The trading activity even surges drastically during disputes. Connected on UMA’s public Discord debate channel, traders can place bets as they can sense in advance which side is more likely to win. At this point, it’s not the truth that matters, but betting on how a small group of token holders will vote. Still, some holders have been fired after being accused of market manipulation. Other holders argue that financial motivation is necessary for the success of UMA. Otherwise their engagement in solving disputes would drop, which could degrade the accuracy and quality of resolutions. In the end, only 0.2% of the millions of markets needed UMA’s help to solve disputes.

Prediction vs stock market

How does betting on events differ from betting on companies? Bets on Polymarket are somewhat like short-term stock option trades: in both cases, participants try to profit by correctly predicting what will happen in the future. If you are right, you can win big. But if you are wrong, you lose all of your stakes as the contract expires and becomes worthless. Time plays against you. Whereas for stocks, if you are not right today, you can just hold it. Unrealised losses can revert to potential gains. Stocks also often pay out dividends and give access to voting rights. 

Prior literature shows that short-term trading reduces long-term wealth, as short-term price movements are random and do not follow any predictable pattern (Fama, 1965). With a win rate of 51% but overall still in the red, the success of retail wagerers is a result of luck rather than skill, given that there are only two options available, yes or no (Akey et al., 2026). Less than 30% of traders are profitable. While the median Polymarket trader is slightly down up to $100, the bottom 10% has lost around $4,000. Compared to options trading, the level of profitable retail traders is similar at below 30%, and the typical retail trader is down $5,000 (Hu et al., 2021).

So despite most participants losing, why are people still attracted by prediction markets? The concentrated user base explains it. Most of the Polymarket traders are young men with a high risk appetite, believing they can become millionaires overnight. With their smaller net worth, they have less to lose. And with housing prices vastly outpacing income, stocks do not cut it anymore. To chase their dream and make a fortune, young adults are forced to turn to riskier investments like crypto and prediction markets. Foolishness and high tolerance of risk create a toxic combination, which explains why they buy bets at more extreme price levels and tend to lose money.

Ethical controversies

The different types of bets available on Polymarket raise ethical concerns. Betting on elections or financial outcomes may help gather information, but betting on military attacks or deaths means that users can profit from human suffering. A perspective is that traders have the incentive to make the outcome happen. Even if Polymarket does not support these events, it turns them into tradable opportunities. This can make war and violence appear like a game. This is an indication that the platform is not regulated enough, sparking political debate: in March 2026, US lawmakers proposed the DEATH BETS Act to ban prediction contracts linked to war, terrorism, assassination and an individual’s death (US Congress, 2026). 

Polymarket also operates in a regulatory grey zone. In the EU, there is no single law for online gambling. Each country applies its own rules. France classified Polymarket as unauthorised gambling and blocked access to the platform. Other European countries, including Belgium, Germany, Poland, Portugal and Romania, have also restricted or blocked Polymarket (Polymarket, n.d.). This shows the difficulty of regulating a platform that combines financial-market features with gambling. While Polymarket presents itself as a prediction market, many regulators treat it as an unlicensed betting platform.

A recent scandal highlights how Polymarket ads specifically target young people using deceptive tactics. The platform paid influencers to brag about massive fake gains on a cloned Polymarket website built to simulate bets. In reality, if you copied the bets, you would lose. So what explains this fraudulent strategy? We previously learned that 0.1% of users win the most on Polymarket, and money from losing shares is transferred to winners. However, this is only possible if there is enough liquidity on the losers' side. Therefore, Polymarket’s marketing campaign on social media is so aggressive because to keep the platform running, it needs to lure enough unskilled users to fund the gains of the winners. If only skilled ones traded against each other on Polymarket, they would not win nearly as much and would quit.

What’s next?

Polymarket is more than a traditional betting platform: it uses financial-market mechanisms to turn expectations about future events into visible prices. However, accurate predictions do not automatically make the platform safe or fair. Large traders, insiders, and herd behaviour can influence prices, while markets on war or death raise ethical concerns. As prediction markets continue to grow, regulators will need to decide where information gathering ends and gambling begins. Polymarket may provide insights into the future, but the future of the platform itself will depend on whether these risks can be controlled.