Onit lets anyone create configurable prediction markets that incentivise sharing information. Existing prediction markets have proven themselves as effective mechanisms to avoid media biases and express the publics true belief on many events. Although they are an improvement on polls and traditional betting platforms, there are still major limitations on how expressive they can be.

Most prediction markets today operate by reducing questions into yes/no tokens and using a function to determine the price of each based on their relative demand. This is great for things with binary outcomes, but is insufficient for other questions with multiple choices, and even worse for events with continuous outcomes.

In the AI market below, betting on xAI winning implies that the other options will lose, yet you do not have access to any of the liquidity from the people betting on OpenAI. For the Bitcoin market, you are again constrained to pre-defined options rather than naturally answering the question with your own range.

A recent paper from Paradigm explored the idea of 'continuous outcome' prediction markets where traders can bet on exact values, and set a (normal) distribution to express their confidence around that number. We explored this idea and eventually built the first example of an 'infinite outcome' prediction market. This new market mechanism overcomes many of the limitations of existing prediction markets, and even unlocks entirely new styles of market.

Onit's infinite outcome prediction market unlocks:

• Adaptive markets: If a new AI player emerges you can dynamically add options to the market, without the need to create a new liquidity pool. This also enables more personalised markets for things like 'Who will get the most reactions on Farcaster this month?' where the network can surface suggestions for others to bet on.

• More expressive markets: Instead of betting on pre-defined options like B: <$95k you can now bet $90k +/-$6k.

• Flexible topics: These infinite outcomes can be used to encode other information, enabling date ranges, exact football scores etc.

This increased flexibility enables more targeted markets which can more optimally elicit different types of information from different types of audience. Want to uncover who is most up to date with social trends? Let anyone propose options and see who added the outcome with the highest demand. Want to get an estimate of how many products you will sell this year? Run a continuous market and only allow people with an onchain transaction tied to you platform to bet.

The Onit protocol consists of 4 parts:

• Smart Contracts: A set of immutable smart contracts that deploy our infinite outcome markets

• API: To create markets or bets, and get data related to any market

• SDK: For building prediction market components and accessing typed API data

• Onit App: The consumer entrypoint for the protocol is the currently launched as a Farcaster Frame where users can access a range of markets and discover where their friends stand on certain topics.

The full protocol can be used by any dapp, company, or individual to deploy markets to elicit information relevant to them. Building onchain uniquely enables the protocol to analyse bettors onchain transaction history and surface markets that are most relevant to them.

Over the coming weeks we will be rolling out our API to more teams, so if you are building an app that could benefit from decentralised prediction markets for sport, politics, crypto, or beyond, drop us a message.