Prediction Markets Speedrun for Power Users and Builders
Origins, theory, stories of Polymarket and Kalshi, awesome tools and opportunities for new products.
Intro
Questions to think about:
What do you think prediction markets are?
Why do you think they were developed?
Have these questions in mind as you read this article. At the end you should know the answer.
Prediction markets all participants to bet on the future outcome of an event. They are used in order to discover the best possible prediction about some future event in order to make better decisions in the present.
Origins:
Prediction markets in various forms have been around for centuries. In 1503 people were voting for the outcome of the Pope election. In late 1880’s betting percentages were estimated for presidential elections. The early research foundations on this topic was in 1945, a scholarly paper "The Use of Knowledge in Society" by economist Friedrich Hayek. It was one the most influential works on economy, it argues that centralized planned economy is limited in comparison to decentralized open markets, because it’s impossible to aggregate all the knowledge. That localized knowledge is more effective in making pricing decisions. That prices are signals important for facilitating coordination of individual actions. This paper was impactful in development of Wikipedia and Kalshi market.
In 1988 the University of Iowa’s Iowa Electronic Markets was formed, a research tool to predict election, with participants using their own funds to make prediction bets, the results were more accurate then traditional polls. More research continued in the 2000’s with Google using them internally to predict product releases, plus other corporations like Microsoft and Eli Lilly. 2006 CIA study “Using Prediction Markets to Enhance US Intelligence Capabilities: A ‘Standard & Poors 500 Index’ for Intelligence.” Congressmen were skeptical, criticized it as gambling, while academics were supportive. The Hollywood Stock Exchange (1996) correctly predicted 32 of 39 Oscar nominees in 2006. Robin Hanson was a researcher at George Mason University, ”Combinatorial information market design” in 2003 and “The promise of prediction markets” in 2008. His papers were very influential to the founder of Polymarket.
To chat with the research papers mentioned above and other topics on prediciton markets see this custom ChatGPT the “Prediciton Market Professor”
The theoretical research foundations show that predictions are made best when individuals have hard incentive of putting their money on their views. This is more effective then centralized planning, where individuals are biased, or do not have access to the aggregated information, as knowledge is by it’s nature decentralized. The individuals at the edges have the best information about the local process, the conditions and facts. The example used in “The Use of Knowledge in Society” - tin markets, the prices of tin are discovered by the supply and demand comparison, as buyers and sellers have options of where to make the trade. In essence prediction markets are there to incentives those with information asymmetry to profit from their knowledge. Prediction markets incentive those with insider information to profit from it. From the Polymarket official documentation:
Our markets reflect accurate, unbiased, and real-time probabilities for the events that matter most to you. Markets seek truth.
Coinbase CEO Brian Armstrong stated: “If your goal is for 99% of people to get signal about what’s going to happen, you actually want insider trading.” Polymarket CEO Shayne Coplan said: “People having an edge to the market is a good thing...it’s sort of an inevitability.”
At the same time Polymarket gets a lot of criticism for ‘insider trading’ something that is not allowed in traditional markets, but here they are essentially their central point.
Kalshi
First founded in 2018 by Tarek Mansour and Luana Lopes Lara. They met at MIT, and inspired by the experiences working at major financial institutions, observing that many financial decisions were driven by predictions about future events. They saw an inefficiency, that they could make it more efficient to trade directly on event outcomes. The essentially reduced the job they were performing at Goldman Sachs to be most efficient.
They joined YC in 2019 and received funding from major VC’s, their goal being to build with regulatory approval. By 2024 they were the first approved to operate in the USA. Notably they were not developed as a blockchain prediction market, but as a traditional centralized one, later they were built on Solana and Tron. Now you can onramp with a crypto wallet from Ethereum or others, but it is not as open and permissionless as Polymarket.
They do have a builder program with $2 million in grants for “developer, creator, researcher, or entrepreneurs“ they are on Solana and Base.
More info here: https://kalshi.com/builders
Quick start docs: https://docs.kalshi.com/getting_started/quick_start_market_data
Polymarket
What is Polymarket? From the source:
Markets seek truth.
This is a way for creating the most accurate predictions about future events. Participants trade contracts on future events, buy shares of yes and no votes. Shares are always priced at $1 every pair of outcomes is collateralized by $1 USDC. For example, a common topic is sports. You are not betting against the ‘house,’ only the counter party is another participant.
Prices are probabilities. Prices reflect supply and demand. You can sell your prices any time, prices get updated as news change. Economic incentives ensure market prices adjust to reflect true odds. More knowledgeable participants are incentivized. Polymarket is an opportunity for people to profit from their knowledge and people use Polymarket to make better informed decisions about future events.
On January 9, 2026 is an NHL match between the Minessota Wild and the Seattle Kraken. The image below shows the current bet for 60% chance of winning for Wild. Each ‘Yes’ share costs $0.60.
Adoption Growth of Polymarket
Milestones:
2020: Polymarket launched, by NYU dropout Shane Copeland. Founder finds traction through DM’ing people on crypto twitter and reddit.
2022: Fined US$1.4 million by the CFTC, agreeing not to operate in the U.S. - but fine was lowered because of ‘substantial cooperation’
2024: $3.6 billion volume during the U.S. elections. Correctly predicted Trump win over Harris while polls were ‘too close to call.’ Predicted that Biden would drop out of the race weeks before the announcement. In November founder’s home was raided, phone seized, searching for evidence investigating if allowed US users to make bets.
2025: October 2025, Intercontinental Exchange (ICE), the parent company of the New York Stock Exchange, announced a strategic investment of up to $2 billion in Polymarket, valuing the company at $9 billion post-money. 27-year-old founder Shayne Coplan becomes one of the world’s youngest self-made billionaires.
Early 2026: U.S. Re-entry: The platform is gradually relaunching in the U.S. with an invite-only beta for those on its waitlist. Dow Jones will feature Polymarket’s data, expanding its reach. Initial U.S. markets will include sports and politics, leveraging its global success
The recent controversy over a newly created Polymarket account won $436,000 betting on Venezuelan regime change just before the January 3rd US military strikes highlights that insider trading is continuing to be a part of the platform.
In contrast, internationally it has been banned in Switzerland, France, Poland, Singapore and Belgium for violating gambling laws. There are fundamental disagreements about whether prediction markets are sophisticated information aggregation tools or simply online gambling with a veneer of economic theory.
Polymarket Hybrid System
Polymarket order book is hybrid decentralized. This means it combines off-chain and on-chain parts. Users create orders off-chain, an operator matches them off-chain, and then triggers their execution via smart contracts. Shares are ERC-1155 tokens representing outcome positions, they are implemented using Gnosis’s Conditional Tokens Framework (CTF). USDC is used as collateral to mint and redeem shares.
USDC is deposited to mint shares. Shares are burned to redeem USDC after market resolution. Only winning outcome token holders receive payouts.
All trades are fully collateralized and the payout logic is handled non-custodially by the smart contracts.
Polymarket Conditional Tokens Smart Contract:
https://polygonscan.com/address/0x4D97DCd97eC945f40cF65F87097ACe5EA0476045#code
Highlighted Projects In the Ecosystem:
An directory of polymarket ecosystem.
Awesome Tools List
https://github.com/aarora4/Awesome-Prediction-Market-Tools
Existing Open Source Trading Bots
https://github.com/topics/polymarket-trading-bot
How to build on top of Polymarket
Get API keys:
Login to your account > go to profile > Builder Codes
Get Market data from WebSockets API.
https://docs.polymarket.com/developers/CLOB/websocket/wss-overview
Place your first order:
https://docs.polymarket.com/quickstart/first-order
New Opportunities for New Startups
This is very early days for prediction markets, it is like 1998 of the dot com bubble. There is an early mover advantage. Here are a couple areas in which there are opportunities:
Infrastructure & Trading Tools
Copy Trading Bots
Automating replication of whale accounts, or bets that are determined to have a high likelihood of being made with asymmetric knowledge.
Popular GitHub repositories are: Trust412’s copy trading bot and Joshbazz’s wallet tracker.
Arbitrage Detection Systems
Prediction markets spread across multiple platforms (Polymarket, Kalshi, Metaculus), price discrepancies create profit opportunities that can be detected and arbitraged fast.
Market Making Infrastructure
Automated market makers that provide liquidity across different prediction market platforms. Agents that do deep research and are able to price opening prices accurately, as currently that is done manually by domain experts.
Analytics & Intelligence
Whale Tracking Dashboards to identify large position movements.
Sentiment Analysis Engines
Combine multiple news and social media sources, track news sentiment, and track market data that can provide signals for market movements.
Performance Analytics Platforms
Tools that help traders track their prediction accuracy over time, identify successful strategies and benchmark against other market opportunities.
User Experience & Access
Mobile trading apps. Mini apps on Base, Farcaster and World.
API Aggregation Services
Normalize data and trading interfaces across multiple prediction markets, to lower barrier to entry for traders and other developers.
Educational Platforms
Gamify prediction market learning through simulated trading and information deep research.
Conclusion
Considering that we are so early in the world of prediction markets, called as the new ‘asset class’ that incentivize people to predict the future, what is the end game? It seems it is building better systems at aggregating and analyzing information. Better and better ‘deep research’ AI agent systems, who ever has the best AI prediction and data collection systems will be most rewarded on these markets.