Written by Scott Duke Kominers, Research Partner at a16z crypto
Compiled by Chopper, Foresight News
Prediction markets allow users to trade on the outcomes of various events. These platforms began scaling up in the United States last year and now track events spanning from geopolitics to winners of entertainment awards. But what exactly are prediction markets?
As an economics researcher who has long studied market mechanisms and incentive systems, my answer is simple: prediction markets are essentially ordinary markets. Markets are the fundamental tool for allocating resources, directing goods and services to those who need them most. In this process, markets also possess information-aggregation capabilities: the process of supply and demand reaching equilibrium integrates all information held by participants and transforms it into signals such as prices.
Prediction market platforms and related products directly leverage this information aggregation capability to forecast the outcomes of specific future events. The platforms offer tradable assets tied to particular events, where holders receive payouts if the predetermined outcome occurs. Users trade these assets based on their assessment of the likelihood of each event. For decades, many companies have used prediction markets to uncover hidden insights from employees, helping them determine whether key products will launch on schedule. Researchers also employ this tool to evaluate the reproducibility of experimental findings. Today, numerous media organizations partner with prediction markets to supplement frontline reporting and traditional journalism with collective wisdom, enriching the depth and breadth of their content.
Prediction markets aggregate individual judgments from all participants and consolidate these perspectives into a trading market to estimate the probability of various events. Users bet on the outcomes of these events, following the same logic as predicting stock prices in the stock market or trading oil prices in commodity markets. The difference is that assets like oil are influenced by multiple complex factors, whereas the underlying asset in a prediction market only generates returns if the specified event occurs.
When oil prices rise, we can determine that demand exceeds supply, but we may not know the underlying cause: is it market concern over escalating tensions in the Middle East, or the emergence of new applications for oil? Prediction markets, however, can create separate trading instruments for individual possibilities, enabling precise segmentation of forecasts. For example, if a market is established for “Whether the Strait of Hormuz will remain open for navigation at a specified time,” the contract terms could stipulate: if the event occurs, each contract pays out $1. As users continuously buy and sell, the market price becomes a probability indicator, reflecting the collective judgment of all traders on the likelihood of the event occurring.
The mechanism works as follows: Suppose the current price of the contract is $0.50, indicating that the market perceives the probability of the event occurring as 50%. If you believe the probability of passage is higher than 50%, say 67%, you can buy the contract. If your judgment is correct, the contract you purchased at $0.50 will ultimately pay out $0.67. This purchase increases the market price and the implied probability, signaling that traders believe the market previously underestimated the likelihood of the event. Conversely, if someone believes the current price is too high, they may sell or short the contract, thereby lowering the market’s implied probability estimate.
Well-functioning prediction markets offer significant advantages over other forecasting methods. First, they directly produce quantitative probability outcomes, which is a key strength. Polls and surveys only measure the proportion of opinions; to derive event probabilities from them, additional statistical methods are required to analyze the relationship between sample data and the broader population. Moreover, poll results are typically static snapshots at a single point in time, whereas prediction markets continuously update judgments in real time as new participants join and new information emerges.
More importantly, prediction markets come with built-in incentive mechanisms. Both buyers and sellers put real money at stake, and incorrect predictions result in financial losses. This compels participants to carefully analyze the information they have and prioritize trading in areas where they possess greater familiarity or information advantages. Conversely, the desire to profit through information and expertise encourages individuals to proactively conduct research and dig deeper into relevant event details. A well-known example is that, ahead of the 2024 U.S. election, some prediction market participants employed unconventional polling methods to gather information inaccessible to traditional polling organizations.
Finally, the scope of prediction markets is extremely broad. In theory, traders with insights into the oil industry can express their views by going long or short on crude oil contracts, but in reality, many event outcomes cannot be predicted through mainstream commodity or stock markets—these are precisely the scenarios where prediction markets excel. For example, recently, several prediction markets have begun listing related assets to collectively evaluate the performance of various AI models across different tasks. Such niche trends are difficult to reflect in traditional commodity markets. Anyone can create and fund a prediction market to answer these specialized questions.
Prediction markets are not a new concept; their earliest forms date back to 16th-century Europe, where they were used to forecast the next pope. Modern prediction markets integrate knowledge from economics, statistics, market design, and computer science. In the 1980s, Charles Plott and Siamak Shand pioneered the formal academic framework for this mechanism. Shortly thereafter, the world’s first modern prediction market—the Iowa Electronic Markets—launched. Leveraging internet technology, this model has since aggregated fragmented information from around the globe and continued to grow and evolve.
However, several challenges still need to be addressed to fully unlock the potential of prediction markets. First, at the infrastructure level: how to determine the final outcome of events and reach consensus, how to ensure transparent market operations and traceable transactions; and when disputes arise over contract settlements—or even in cases of manipulation—how to implement scalable resolution mechanisms.
Second, there are challenges at the market design level. First, those with access to critical information must participate. If all participants are uninformed, market price signals become meaningless. Conversely, if informed parties refuse to engage, forecast outcomes will be biased. As early as 2016, I proposed that prediction markets underestimated the likelihood of events such as Brexit and Trump’s first election as U.S. president, because participants at the time failed to recognize the rising trend of populism.
In addition, if individuals with insider information enter the market to trade, especially those who have the ability to influence the outcome of events, it can trigger significant risks. Imagine if insiders at a papal conclave placed bets in advance on prediction markets for the next pope, using insider knowledge to trade ahead of the announcement—or worse, secretly manipulated the election results to benefit their own positions. The consequences would be obvious. Once participants generally believe that insider trading is prevalent, they will withdraw from the market, ultimately leading to its collapse.
Another risk is that someone might deliberately manipulate the predicted market price to influence public perception of event probabilities. In this way, prediction markets could shift from tools for aggregating opinions into instruments for manipulating public opinion. For example, a campaign team could use campaign funds to artificially inflate the market probability of their candidate winning, creating a false impression of leadership. However, prediction markets possess a degree of self-correction: whenever prices deviate significantly from reasonable levels, traders will place counter-bets to hedge against irrational pricing.
Various issues indicate that prediction markets need further refinement of rules, with clear standards for participant eligibility, contract design, and overall operations. However, if industry participants can systematically address these challenges, prediction markets will ultimately become an essential tool for humanity to anticipate the future and manage uncertainty.