Have you been hearing a lot about prediction markets recently? You’re not alone. 

The concept of a prediction market has entered mainstream public consciousness in a big way lately. The viral success of sites like Polymarket during the recent US Presidential election has shined a spotlight on the concept, but prediction markets aren’t anything new. 

While there seems to be a much more widespread awareness of these systems, misunderstandings about them abound. In the article below we’ll take a look at exactly what a prediction market is, what it’s not and exactly how the process works. 

What Is a Prediction Market

A prediction market (sometimes referred to as a betting market or idea futures market) is a place where users can make predictions on the likely outcome of an unknown event. A correct prediction is typically incentivized with financial rewards, so that each participant is competing for correct predictions directly.

The structure of predictions are traditionally set to binary options, with each option being valued at its share of the “vote” placed on it. More complex options change the predicted likelihood in the same manner, just spread out across more possible predictions. 

How Prediction Markets Work

Prediction markets do not work like traditional betting pools on sports or other gaming. The value of any given prediction is only based on how many users relative to the total have predicted that specific outcome.

For a decentralized prediction market, it often works like this:

1. You want to predict on a question, so you purchase $10 of “Yes”, which is currently at 25%.
2. This means that you now have $10 of Yes tokens on that question, but there are 3x as many “No” tokens as there are Yes tokens. 
3. If new developments cause people to buy Yes tokens, No tokens become relatively less valuable and “Yes” tokens become relatively more valuable.
4. You may sell your “Yes” tokens before the issue is decided. If you bought when Yes was at 25% and sold when it was at 50%, the prediction you placed would double to $20.
5. Once an issue is decided, that correct prediction has all the value from the question. If you have not sold and the final objective answer turns out to be Yes, your Yes tokens would be valued at $40.

A crucial misunderstanding about these platforms is that percentage chances do not indicate “odds” in a prediction market, as they would in traditional gambling. The likelihood shown instead indicates the number of predictions placed on that option versus other available options. This means that prediction markets aggregate the predictions of the crowd, rather than any odds based on objective facts.

Prediction markets are intended to tap into the “Wisdom of the Crowd Effect”. This phenomenon was first observed by Aristotle in his 350 BCE book, Politics, but statistician Francis Galton is widely credited with bringing it into the modern day. He observed in the early 20th century that often the median prediction of a crowd would be closer to the final result of an event than even qualified opinions.

“It is possible that the many, though not individually good men, yet when they come together may be better, not individually but collectively, than those who are so, just as public dinners to which many contribute are better than those supplied at one man’s cost”

-Aristotle, Politics

This doesn’t mean that prediction markets are always right. When people are betting on roulette, for instance, they can be confident that they have a 2.7% chance to hit any individual number (2.63% with a 00 wheel). This is based on the physics of the wheel and the high level of predictable randomness in roulette.

Prediction markets are only based on what the users of the platform are saying. If a given option is listed as 70% (Yes) and 30% (No), that just means that 70% of the funds used to predict on that question have chosen yes. This isn’t inherently bad, but it is wise to consider that these predictions are often vulnerable to manipulation, misunderstandings or cognitive biases within the users of the market.

Why Prediction Markets?

While it’s easy to dismiss prediction markets as just another type of gambling, they have applications that are far different from traditional gambling. Prediction markets have been used to glean insights for generations — typically in politics, business and investing.

The first widely known modern electronic prediction market was the Iowa Electronic Markets, established by The University of Iowa. This is run as a non-profit, and users can make predictions about elections or key economic indicators. This has been run continuously since the 1988 presidential election for research and education, and at times it has been more indicative of election results than polling.

The 1990s saw several prediction markets rise and fall, speculating on incredibly diverse issues across science, world events and culture. As the age of data began to rise, however, more entities started utilizing the powerful predictions used by aggregating median opinion.

Pharmaceutical companies and investors have noted that they often use internal prediction markets to forecast which new developments are likely to get through clinical trials. Google announced in 2005 that it uses internal predictions to forecast likely launch dates. Even the US Department of Defense had a short-lived project to use prediction markets from the public to indicate likely threat occurrences.

Into the Decentralized World

In the world of web3, prediction markets have understandably risen back to public prominence. Due to the nature of decentralization, these prediction markets can create a very wide benchmark on opinion. The choice/token structure of a prediction market is perfect for a blockchain, and smart contracts ensure a trustless exchange between platforms and users.

The first on-chain prediction market was Augur, created on Ethereum in 2018. While it received some use early in its lifetime, it quickly stagnated due to the lack of community in the web3 world and due to its sidestepping of regulatory agencies. Augur v2 would later release and revamp the concept as a totally decentralized and open source platform, but it has since faded into obscurity due to Ethereum gas prices and developers moving on. 

Polymarket’s launch on the Polygon blockchain in 2020 pushed prediction markets into the modern world. While Polymarket had early troubles with the Commodity Future Trading Commission (CFTC), they were able to resolve the issues and cooperate with commodity exchange regulations. While volume has receded sharply on the platform over the past several weeks, more than $3bn of total predictions have been placed on the site so far in 2024… much of it on questions focusing on the US presidential election.

Crucially, prediction markets have become somewhat self-sustaining in the era of blockchain. Users can propose questions and interact with a wider variety of prediction events than ever before. This is both good and bad. More data can provide us with more insights, but with the question end of the process more highly democratized, it’s far more possible to encounter leading or misleading questions without realizing it. Perception also does not always equal reality. The unavailability of reliable information or predictors failing to inform themselves can easily shift results.

Prediction markets aren’t perfect, but they do show us important things about the opinions of their users.Think of them as a different kind of poll, rather than objective truth or “odds”. While traditional polling on issues has low response rates and low participation, prediction markets incentivize participation. Since they incentivize being correct as well, they accurately show us a picture of what participants think is more likely. That’s not truth, but it’s good data.

Hopefully we’ve given you a little more insight into how prediction markets work and what they’re used for. Decentralization isn’t intuitive for most of us, and it’s almost like prediction markets were just web3 ahead of their time.

Because prediction markets originally were more about data than wagering, it can be incredibly insightful just to browse through predictions on issues even for those who have no plans on actually placing predictions. No one has built a prediction market on GalaChain yet, but if you do choose to engage with these platforms elsewhere, we hope this article has given you some crucial knowledge to help you do so safely and wisely.