Jan 29, 2026
On Prediction Markets and Privacy
Prediction markets consistently outperform traditional polls, but full transparency limits who participates and how honestly. Public bets introduce social pressure, legal risk, and market manipulation. Privacy can unlock better incentives, stronger signals, and more accurate forecasts by protecting participants and restoring genuine information discovery.
By Gonzalo Silman, Co-founder WakeUp Labs.
On Prediction Markets and Privacy
Prediction markets have gone from a niche experiment to a mainstream instrument in a remarkably short time.
They are regularly referenced by major media outlets to forecast elections, major sporting events, cultural awards, and even geopolitical conflicts. Platforms like Polymarket and Kalshi are no longer confined to crypto-native conversations; they are increasingly being treated as legitimate forecasting tools alongside traditional polls.
However, what matters most isn’t visibility but performance.
Across multiple high-stakes events, prediction markets have consistently produced forecasts that outperform traditional opinion polling conducted by well-funded consulting firms and expert-led surveys run by political scientists and economists.
This is not a coincidence, and it is not hype. Prediction markets work.
They are also one of the very few crypto-native products that have demonstrated real product–market fit. Not just in terms of TVL, transaction count, or volume, but also in the secondary ecosystem they enable. Real businesses are being built on top of them: explorers, analytics tools, arbitrage systems, and infrastructure connecting crypto markets with traditional betting platforms.
As concrete examples, some friends of mine built a Polymarket Explorer (@poly_data - https://x.com/poly_data) during a hackathon. With basic ad monetization, it generates meaningful revenue despite being a side project. Others are building arbitrage bots operating across Kalshi, Polymarket, and legacy gambling platforms.
At WakeUp Labs, we have been exploring a different dimension: the intersection between prediction markets and privacy. In a project called Aztec Markets, we have been testing privacy-preserving mechanisms using Aztec, validating hypotheses around incentives, information disclosure, and market behavior in a real-world setting.
But to understand where they can go next, we need to be precise about what they are, and where they currently fall short.
What a Prediction Market Actually Is
The term prediction market was formally introduced and developed by Robin Hanson in the early 1990s.
A prediction market is a market-based system in which participants trade contracts linked to future events. These contracts may represent binary outcomes, categorical outcomes, or scalar payoffs. Market prices aggregate participants’ beliefs and function as real-time forecasts of the probability of those events occurring.
A large body of empirical and experimental research shows that prediction markets tend to outperform traditional polls and surveys precisely because they embed direct monetary incentives. Participants are rewarded for revealing accurate beliefs and penalized for being wrong. Unlike surveys, there is a real cost to being careless, biased, or dishonest.
The economic intuition behind prediction markets can bring to mind Friedrich Hayek’s 1945 essay *The Use of Knowledge in Society, a paper I love*. Hayek argued that decentralized markets are uniquely effective at aggregating dispersed, local, and often tacit information that no centralized authority or survey mechanism can fully capture. Prediction markets operationalize this idea in a measurable and falsifiable way.
By making incentives programmable, settlement global, and participation permissionless, crypto creates a perfect environment for prediction markets.
In this context, Shayne Coplan founded Polymarket in 2020, not only validating the theory but packaging it into a product that is accessible, engaging, and scalable.
The Transparency Problem
Most prediction markets today operate with full on-chain transparency. While users are pseudonymous, trades, positions, timing, and size are all publicly observable.
This is often framed as a feature, but one could argue it's a limitation.
When every bet is visible, participation is no longer just about expressing a belief. It becomes a public statement with social, professional, legal, and political consequences.
At the same time, the bet itself can be tracked, copied, or even front-run.
As a result, some of the most informed participants choose not to participate at all.
The first goal to achieve related to privacy is simple: when a user places a bet, it should be practically impossible to determine who took that position. At the same time, the user should retain plausible deniability and be unable to credibly prove to a third party that they made that bet.
Governments, employers, regulators, traders or large institutions often have strong incentives to identify who is betting on elections, wars, corporate events, or product releases. Especially when insider or sensitive information is involved. These actors frequently have the resources required to deanonymize users.
There are already documented cases of individuals facing scrutiny or consequences related to prediction market activity.
Why Private Prediction Markets Make Sense
1. Social Pressure Is a Tax on Information
Betting carries stigma, regardless of how rational or information-driven it is. Many users prefer to describe their activity as “investing” to protect their social standing, professional relationships, or public image.
When positions are visible, users self-censor. They avoid controversial, politically sensitive, or reputationally costly markets. Privacy reduces the expected cost of being seen holding a particular belief and increases the willingness to contribute information.
This includes protection against retaliation, coercion, or doxxing of positions. In environments where beliefs can be linked to identity, privacy restores plausible deniability and lowers the barrier to honest participation.
From a market design perspective, this is not optional. It directly affects signal quality.
2. Censorship and Geopolitics Distort Participation:
In jurisdictions with strict regulatory regimes or aggressive enforcement, such as the United States or China, participation in prediction markets can expose users to legal or professional risk. This discourages usage precisely in regions that concentrate capital, expertise, and valuable information.
Without privacy, informed participants opt out. With privacy, participation becomes safer.
The literature on private prediction markets explicitly highlights this effect. When employees, insiders, or domain experts fear exposure, they remain silent, leading to systematic biases and over-optimistic forecasts.
3. Transparency Enables Market Exploitation
Prediction markets behave like DEXs, and they inherit the same failure modes.
Front-running, back-running, sandwich attacks, and transaction censorship are all possible when trades are visible. Bots can detect large bets, exploit slippage, manipulate odds, or extract value based on order flow rather than information.
Privacy significantly protects users from others profiting off them through these strategies.
4. Privacy Is a Precondition for Integrity
Just as elections can only be free and fair when votes are private, prediction markets that cover sensitive or controversial topics need strong privacy guarantees. Without them, some markets won’t be created, and users could face real-world consequences for expressing beliefs through economic action.
Financial privacy matters too. Public transaction histories leave users exposed to phishing, targeting, and even physical threats. Unsurprisingly, many cautious users simply choose not to participate.
Conclusion
Prediction markets work, but I’m fairly certain they could improve their accuracy with privacy.
When participation comes with social, legal, or professional costs, the most informed actors self-censor. When trades are fully observable, markets drift toward strategic behavior and away from genuine information discovery.
Privacy isn’t about hiding outcomes or avoiding accountability. It’s about protecting information contributors, reducing manipulation, and restoring the conditions under which markets can actually aggregate truth.
There’s a strong and natural fit between prediction markets and privacy-preserving mechanisms. The benefits are concrete: broader participation, stronger signals, better incentives, and more resilient market design.
Privacy doesn’t need to stop at concealing bettor identities. The design space is vast and still largely unexplored.
As privacy technologies mature and become faster and more affordable, these ideas will move from theory into practice. The next generation of prediction markets won’t just be more private. They’ll be more accurate, and we’ll be there to build them.
By Gonzalo Silman, Co-founder WakeUp Labs.
On Prediction Markets and Privacy
Prediction markets have gone from a niche experiment to a mainstream instrument in a remarkably short time.
They are regularly referenced by major media outlets to forecast elections, major sporting events, cultural awards, and even geopolitical conflicts. Platforms like Polymarket and Kalshi are no longer confined to crypto-native conversations; they are increasingly being treated as legitimate forecasting tools alongside traditional polls.
However, what matters most isn’t visibility but performance.
Across multiple high-stakes events, prediction markets have consistently produced forecasts that outperform traditional opinion polling conducted by well-funded consulting firms and expert-led surveys run by political scientists and economists.
This is not a coincidence, and it is not hype. Prediction markets work.
They are also one of the very few crypto-native products that have demonstrated real product–market fit. Not just in terms of TVL, transaction count, or volume, but also in the secondary ecosystem they enable. Real businesses are being built on top of them: explorers, analytics tools, arbitrage systems, and infrastructure connecting crypto markets with traditional betting platforms.
As concrete examples, some friends of mine built a Polymarket Explorer (@poly_data - https://x.com/poly_data) during a hackathon. With basic ad monetization, it generates meaningful revenue despite being a side project. Others are building arbitrage bots operating across Kalshi, Polymarket, and legacy gambling platforms.
At WakeUp Labs, we have been exploring a different dimension: the intersection between prediction markets and privacy. In a project called Aztec Markets, we have been testing privacy-preserving mechanisms using Aztec, validating hypotheses around incentives, information disclosure, and market behavior in a real-world setting.
But to understand where they can go next, we need to be precise about what they are, and where they currently fall short.
What a Prediction Market Actually Is
The term prediction market was formally introduced and developed by Robin Hanson in the early 1990s.
A prediction market is a market-based system in which participants trade contracts linked to future events. These contracts may represent binary outcomes, categorical outcomes, or scalar payoffs. Market prices aggregate participants’ beliefs and function as real-time forecasts of the probability of those events occurring.
A large body of empirical and experimental research shows that prediction markets tend to outperform traditional polls and surveys precisely because they embed direct monetary incentives. Participants are rewarded for revealing accurate beliefs and penalized for being wrong. Unlike surveys, there is a real cost to being careless, biased, or dishonest.
The economic intuition behind prediction markets can bring to mind Friedrich Hayek’s 1945 essay *The Use of Knowledge in Society, a paper I love*. Hayek argued that decentralized markets are uniquely effective at aggregating dispersed, local, and often tacit information that no centralized authority or survey mechanism can fully capture. Prediction markets operationalize this idea in a measurable and falsifiable way.
By making incentives programmable, settlement global, and participation permissionless, crypto creates a perfect environment for prediction markets.
In this context, Shayne Coplan founded Polymarket in 2020, not only validating the theory but packaging it into a product that is accessible, engaging, and scalable.
The Transparency Problem
Most prediction markets today operate with full on-chain transparency. While users are pseudonymous, trades, positions, timing, and size are all publicly observable.
This is often framed as a feature, but one could argue it's a limitation.
When every bet is visible, participation is no longer just about expressing a belief. It becomes a public statement with social, professional, legal, and political consequences.
At the same time, the bet itself can be tracked, copied, or even front-run.
As a result, some of the most informed participants choose not to participate at all.
The first goal to achieve related to privacy is simple: when a user places a bet, it should be practically impossible to determine who took that position. At the same time, the user should retain plausible deniability and be unable to credibly prove to a third party that they made that bet.
Governments, employers, regulators, traders or large institutions often have strong incentives to identify who is betting on elections, wars, corporate events, or product releases. Especially when insider or sensitive information is involved. These actors frequently have the resources required to deanonymize users.
There are already documented cases of individuals facing scrutiny or consequences related to prediction market activity.
Why Private Prediction Markets Make Sense
1. Social Pressure Is a Tax on Information
Betting carries stigma, regardless of how rational or information-driven it is. Many users prefer to describe their activity as “investing” to protect their social standing, professional relationships, or public image.
When positions are visible, users self-censor. They avoid controversial, politically sensitive, or reputationally costly markets. Privacy reduces the expected cost of being seen holding a particular belief and increases the willingness to contribute information.
This includes protection against retaliation, coercion, or doxxing of positions. In environments where beliefs can be linked to identity, privacy restores plausible deniability and lowers the barrier to honest participation.
From a market design perspective, this is not optional. It directly affects signal quality.
2. Censorship and Geopolitics Distort Participation:
In jurisdictions with strict regulatory regimes or aggressive enforcement, such as the United States or China, participation in prediction markets can expose users to legal or professional risk. This discourages usage precisely in regions that concentrate capital, expertise, and valuable information.
Without privacy, informed participants opt out. With privacy, participation becomes safer.
The literature on private prediction markets explicitly highlights this effect. When employees, insiders, or domain experts fear exposure, they remain silent, leading to systematic biases and over-optimistic forecasts.
3. Transparency Enables Market Exploitation
Prediction markets behave like DEXs, and they inherit the same failure modes.
Front-running, back-running, sandwich attacks, and transaction censorship are all possible when trades are visible. Bots can detect large bets, exploit slippage, manipulate odds, or extract value based on order flow rather than information.
Privacy significantly protects users from others profiting off them through these strategies.
4. Privacy Is a Precondition for Integrity
Just as elections can only be free and fair when votes are private, prediction markets that cover sensitive or controversial topics need strong privacy guarantees. Without them, some markets won’t be created, and users could face real-world consequences for expressing beliefs through economic action.
Financial privacy matters too. Public transaction histories leave users exposed to phishing, targeting, and even physical threats. Unsurprisingly, many cautious users simply choose not to participate.
Conclusion
Prediction markets work, but I’m fairly certain they could improve their accuracy with privacy.
When participation comes with social, legal, or professional costs, the most informed actors self-censor. When trades are fully observable, markets drift toward strategic behavior and away from genuine information discovery.
Privacy isn’t about hiding outcomes or avoiding accountability. It’s about protecting information contributors, reducing manipulation, and restoring the conditions under which markets can actually aggregate truth.
There’s a strong and natural fit between prediction markets and privacy-preserving mechanisms. The benefits are concrete: broader participation, stronger signals, better incentives, and more resilient market design.
Privacy doesn’t need to stop at concealing bettor identities. The design space is vast and still largely unexplored.
As privacy technologies mature and become faster and more affordable, these ideas will move from theory into practice. The next generation of prediction markets won’t just be more private. They’ll be more accurate, and we’ll be there to build them.
