a16z: Why Prediction Markets Matter
TechFlow Selected TechFlow Selected
a16z: Why Prediction Markets Matter
Prediction markets are, at their core, markets.
Navigating Web3 tides with focused insightsBy Scott Duke Kominers, Research Partner at a16z crypto
Translated by Chopper, Foresight News
Prediction markets allow users to trade on the outcomes of various events. These platforms began scaling widely across the U.S. last year and now track everything from geopolitical developments to winners of entertainment awards. But what exactly are prediction markets?
As an economist who has long studied market mechanisms and incentive systems, my answer is simple: prediction markets are, fundamentally, ordinary markets. Markets are foundational tools for allocating resources—directing goods and services toward those who value them most. In doing so, markets also aggregate information: the process of supply-and-demand clearing synthesizes all participants’ knowledge and distills it into price signals.
Prediction market platforms—and their associated financial instruments—leverage this information-aggregation capability directly to forecast the likelihood of specific future events. Platforms introduce assets tied to concrete events; holders receive payouts if the predefined outcome occurs, and users trade these assets based on their personal estimates of the event’s probability. For years, many companies have used prediction markets to surface employees’ tacit knowledge—for example, to assess whether key products will launch on schedule. Researchers likewise use them to evaluate which experimental findings are likely replicable. Today, numerous media organizations partner with prediction markets to supplement frontline reporting and traditional journalism with collective intelligence—enriching the depth and dimensionality of their coverage.
Prediction markets gather individuals’ personal forecasts about the future, then consolidate those judgments into tradable markets that estimate the probabilities of various events. Users bet on event outcomes here much as they would predict a public company’s stock price in equity markets or trade oil prices in commodity markets. The key distinction is that while oil prices reflect a complex web of influences, prediction-market assets pay out only if a specific, pre-defined event occurs.
When oil prices rise, we infer that demand currently exceeds supply—but we often cannot pinpoint why: Is it anxiety over escalating Middle East tensions? Or the emergence of new applications for petroleum? Prediction markets, by contrast, can isolate and price individual possibilities precisely. For instance, a market could be created for “Will the Strait of Hormuz remain open for navigation at a specified time?” Its contract terms might stipulate: $1 per contract paid out if the event occurs. As users continuously buy and sell, the market price becomes a real-time probability indicator—reflecting the crowd’s aggregated judgment about the event’s likelihood.
Here’s how it works in practice: Suppose the current market price is $0.50 per contract—implying the market judges the event’s probability at 50%. If you believe the actual probability is higher—say, 67%—you can buy the contract. If correct, your $0.50 investment yields a $0.67 return. This purchase pushes up both the market price and the implied probability, signaling that some traders believe the market previously underestimated the event’s likelihood. Conversely, if others deem the current price too high, they may sell short or offload contracts, thereby lowering the market’s implied probability.
Compared with other forecasting methods, well-functioning prediction markets offer distinct advantages. First, they produce quantified probability estimates—a core strength. Polls and surveys merely tally opinion shares; converting such data into probabilistic forecasts requires statistical modeling and assumptions about sample representativeness. Moreover, poll results are static snapshots, whereas prediction markets dynamically update in real time as new participants join and new information emerges.
More crucially, prediction markets embed built-in incentive and accountability mechanisms. Buyers and sellers risk real money; misjudgments incur direct losses. This compels participants to rigorously vet their own information and focus trading where they possess domain expertise and informational advantage. Conversely, the opportunity to profit from insight and expertise motivates people to conduct original research and dig deeper into event-related evidence. A prominent example occurred ahead of the 2024 U.S. presidential election, when some prediction-market participants deployed unconventional polling methods to uncover information inaccessible to traditional survey firms.
Finally, prediction markets boast extraordinary breadth. While traders with oil-industry knowledge could express views via crude-oil futures, countless other events lack corresponding instruments in mainstream commodity or equity markets—precisely where prediction markets shine. Recently, several prediction markets have launched assets evaluating how different AI models perform across specific tasks—a granular trend unlikely to register meaningfully in conventional commodity markets. Anyone can build and fund a prediction market to address such niche questions.
Prediction markets are not new. Their earliest precursors date back to 16th-century Europe, where people used them to forecast papal elections. Modern prediction markets synthesize insights from economics, statistics, market design, and computer science. In the 1980s, Charles Plott and Shyam Sunder established the formal academic framework for these systems. Shortly thereafter, the Iowa Electronic Markets—the world’s first modern prediction market—launched. Powered by internet infrastructure, this model aggregated fragmented global information and steadily expanded.
Yet unlocking prediction markets’ full potential still faces significant hurdles. First, infrastructural challenges: How do we objectively determine final event outcomes and reach consensus? How do we ensure market transparency and transaction traceability? And how can large-scale adjudication mechanisms resolve disputes—or even deliberate manipulation—around payout determinations?
Second, market-design challenges. One critical issue is ensuring participation by informed actors. If no one possesses relevant knowledge, market prices become meaningless. Conversely, if knowledgeable parties choose not to participate, forecasts skew. As early as 2016, I argued that prediction markets underestimated the probabilities of Brexit and Donald Trump’s first U.S. presidential victory because participants failed to grasp the rising tide of populism.
Another risk arises when insiders—those with non-public information—trade in prediction markets, especially if they can influence the underlying event. Imagine Vatican insiders placing bets on “next pope” markets before the conclave concludes—exploiting privileged access to front-run trades, or even covertly manipulating the election itself to benefit their positions. If participants widely suspect insider trading, confidence collapses and the market disintegrates.
A third risk involves deliberate price manipulation to shape public perception of event probabilities. Prediction markets thus risk shifting from information-aggregation tools to instruments of narrative control. For example, political campaigns could deploy campaign funds to artificially inflate their candidate’s market-implied win probability—creating a false impression of momentum. Yet prediction markets retain some self-correcting capacity: if prices stray significantly from reasonable ranges, arbitrageurs will place counter-bets to correct mispricing.
All these issues underscore the need for refined rules—clarifying participant eligibility, contract design, and operational standards. But if industry practitioners systematically solve these challenges, prediction markets could ultimately become indispensable tools for humanity to anticipate the future and navigate uncertainty.
Join TechFlow official community to stay tuned
Telegram:https://t.me/TechFlowDaily
X (Twitter):https://x.com/TechFlowPost
X (Twitter) EN:https://x.com/BlockFlow_News
a16z
