Written by: Scott Duke Kominers, a16z crypto research partner

Translated by: Chopper, Foresight News

Prediction markets allow users to trade on the outcomes of various events. These platforms began to scale significantly in the United States last year, and now their tracking of events ranges from geopolitics to entertainment award winners, among many other areas. But what exactly are prediction markets?

As an economist who has long researched market mechanisms and incentive systems, my answer is simple: prediction markets are essentially just regular markets. Markets are the fundamental tool for allocating resources, enabling goods and services to flow to those who need them the most. In this process, the market also has the ability to aggregate information: the process of clearing supply and demand consolidates all the information held by participants and transforms it into signals like prices.

Prediction market platforms and related products directly leverage this information aggregation capability to predict the direction of specific future events. The platform will launch assets corresponding to specific events, and as long as the pre-set results come true, holders can earn profits, while users trade these assets based on their judgments of the likelihood of the events occurring. For a long time, many companies have used prediction markets to extract tacit information held by employees to judge whether key products can launch on schedule. Researchers also use this tool to assess which experimental conclusions are replicable. Nowadays, many media organizations also choose to cooperate with prediction markets, using collective intelligence to complement frontline interviews and traditional reporting, enriching content dimensions.

Prediction markets aggregate all participants' individual judgments on the future and integrate these viewpoints to form a trading market, thus calculating the probabilities of various events occurring. Users bet on the outcomes of events in such markets, with the logic being no different from predicting stock prices in the stock market or trading oil prices in the commodities market. The difference is that the prices of assets like oil are influenced by multiple complex factors, while the assets in prediction markets only generate profits if a designated event occurs.

When oil prices rise, we can ascertain that demand currently exceeds supply, yet we may not know the underlying reasons: is the market worried about an escalation in the Middle East situation, or has oil found new applications? Prediction markets can establish trading assets specifically for single possibilities, allowing for precise breakdowns of predictions. For example, if a market is established for "Will the Strait of Hormuz have normal navigation at a specific time?", the corresponding contract rules can be set as follows: if the event occurs, each contract pays out 1 dollar. As users continuously buy and sell, the market price becomes a probabilistic indicator, reflecting the aggregate judgment of all traders on the likelihood of the event occurring.

Its operational logic is as follows: suppose the current price of each asset is 0.5 dollars, indicating that the market believes the probability of the event occurring is fifty-fifty. If you judge the probability of navigation is higher than 50%, say 67%, you can buy that asset. Once your judgment is correct, the asset you bought at 0.5 dollars could ultimately yield 0.67 dollars in profit. This buying action will further raise the market price and probability estimate, indicating that some traders believe the market previously underestimated the likelihood of the event occurring. Conversely, if someone thinks the current price is too high, they will sell or short the asset, thereby lowering the market's probability valuation.

Compared to other prediction methods, well-functioning prediction markets have significant advantages. First, they can directly output quantifiable probability results, which is a core highlight. Polls and surveys can only count the percentage of opinions, and drawing conclusions about event probabilities requires statistical methods to analyze the correlation between sample data and the overall population. Additionally, poll results are often just static data for a particular timestamp, whereas prediction markets update judgments in real-time as new participants enter and new information emerges.

More critically, prediction markets come with an inherent incentive constraint mechanism. Both buyers and sellers invest real money, and any judgment mistake leads to losses. This compels participants to carefully sift through the information they hold and prioritize trading in areas where they are more familiar and have informational advantages. Conversely, the desire to profit from information and expertise also encourages people to actively research and delve into event-related clues. A well-known example occurred on the eve of the 2024 U.S. presidential election, where participants in prediction markets specifically used unconventional methods to conduct polls, aiming to obtain information that traditional polling organizations could not grasp.

Lastly, prediction markets have a very wide coverage. Theoretically, traders with information about the oil industry can express judgments by going long or short on crude oil contracts, but in reality, there are numerous event outcomes that cannot be predicted using mainstream commodities markets or stock markets; such scenarios are precisely where prediction markets can play a role. For instance, many prediction markets have recently started launching relevant assets to comprehensively evaluate the performance of various artificial intelligence models in different tasks. Trends in such niche areas are difficult to reflect in traditional commodity markets. Anyone can set up and fund a prediction market to answer such niche questions.

Prediction markets are not a new phenomenon; their early forms can be traced back to 16th century Europe, where they were used to predict the selection of the next pope. Modern prediction markets integrate knowledge from multiple fields including economics, statistics, market design, and computer science. In the 1980s, Charles Plott and Sushil Bikhchandani first established the formal academic framework of this mechanism. Shortly thereafter, the world's first modern prediction market – the Iowa Electronic Market – was officially launched. Enabled by internet technology, this model has been able to integrate scattered information from around the world and continue to grow.

However, to fully unleash the potential of prediction markets, many challenges still need to be addressed. First is the infrastructure level: how to determine the final outcome of an event and reach a consensus, how to ensure market operation is transparent and transactions are traceable; when disputes arise over contract payouts or even encounter human manipulation, how to implement a large-scale decision-making mechanism.

Secondly, there are challenges at the market design level. First, the groups possessing core information must participate. If all participants are uninformed, then market price signals hold no reference value. Conversely, if informed parties are unwilling to join, the prediction results will become biased. I pointed out back in 2016 that during Brexit and Donald Trump's first election as U.S. president, prediction markets underestimated the likelihood of those events occurring because the participants at that time failed to grasp the rising trend of populism.

In addition, if individuals with insider information trade in the market, it could trigger risks, especially if these individuals can influence the direction of the event. Imagine if internal members of the papal election meeting placed bets in the "next pope" prediction market ahead of time, using insider information to front-run trades; or even intervening covertly in the election's outcome for their own positions. The consequences would be dire. Once participants generally believe that the market has insider trading, everyone would choose to exit, ultimately leading to the collapse of the entire market.

There is also a risk that someone might deliberately manipulate prediction market prices to guide public perception regarding event probabilities. In this way, prediction markets would turn from tools for aggregating opinions into means of manipulating public opinion. For instance, a campaign team might use campaign funds to intentionally raise their side's probability of winning in the market, creating a false sense of leading. However, prediction markets have a certain self-correcting ability: as long as prices deviate significantly from reasonable ranges, traders will place reverse bets to hedge against unreasonable pricing.

All these issues indicate that prediction markets need to further refine their rules, clarifying participant access, contract design, and overall operational standards. But if industry practitioners can tackle these challenges one by one, prediction markets are bound to become important tools for humans to predict the future and cope with uncertainties.

免责声明：本文章仅代表作者个人观点，不代表本平台的立场和观点。本文章仅供信息分享，不构成对任何人的任何投资建议。用户与作者之间的任何争议，与本平台无关。如网页中刊载的文章或图片涉及侵权，请提供相关的权利证明和身份证明发送邮件到support@aicoin.com，本平台相关工作人员将会进行核查。