We're building Prediction Markets as a permissionless hyperstructure primitive.

These aren't your typical DeFi primitives. They're unique because they harness the collective wisdom of crowds to predict outcomes—like a crystal ball that rewards you with yields when you bet on the right outcome.

We're solving two major problems that plague today's prediction market exchanges:

MARKET CREATION:

Currently, you can't create a prediction market programmatically in a definitive way. Popular marketplaces only let you propose the creation of PMs (rather than instantly creating a tradeable market). Then, a third party must judge and grade the market and assign a truth source.

Too primitive.

We believe creating a new prediction market should be as straightforward as launching a token or vault—and that's exactly what we're building.

MARKET SETTLEMENT:

Current prediction markets rely on third-party optimistic providers for settlement. While functional, this approach has flaws. Delegating such a crucial component to an off-chain party disrupts the flow and creates unnecessary dependencies.

The source of truth for settlement is clearly defined in listed prediction markets. However, the execution of settlement happens off-chain. What we see on-chain is merely a third party submitting the chosen option—far from optimal.

As prediction markets gain traction, third-party providers might face financial incentives to manipulate markets. We've already seen cases where providers have altered responses for political correctness, as evidenced in this prediction market. See these tweets

Squiggly Hair Shanks on Twitter / X

Conor McMenamin on Twitter / X

HOW WE SOLVE THIS PROBLEM?

We start by categorising prediction markets by the way they settle. One requiring a price feed to settle → Price Markets One requiring non-price data to settle → Twitter Markets

Screenshot 2025-01-13 at 7.26.54 PM.png

All the Price markets have a settling module contract associated with it holding the settling logic. When creating a price market over any ERC20/SPL token, we store the target price and the end deadline onchain. ( taken from user input ) During settlement, the module contract does the following things : → fetches market params : target price, end deadline, token address → fetches current price onchain from corresponding high liq pool of the token → compares them both amd sets the winning tokenId

All binary outcomes of a market are minted as an ERC1155 token with their tokenID calculated by uint256(keccak256(**abi**.encodePacked(conditionId, "YES"))) uint256(keccak256(**abi**.encodePacked(conditionId, "NO")))

All the twitter markets operate by the following way : → While creating prediction market over any arbitrary question, user puts the following things onchain : * Market question * Twitter username of a trusted individual * Market deadline → As soon as he does that, the market opens up for trading till the deadline. → During settlement, anyone can call the initSettlement(conditionId) function. → This emits a PnpInitSettlement(conditionId) event → Our offchain AI model listens for this events, grabs the twitter data from the given twitter username put onchain, makes a structured prompt with { market question, tweet data } and outputs “YES” or “NO” This response is then put onchain and the market is settled.

We plan to make a network of operators running this models and giving BLSSigned “YES” , “NO” response ( against staked value )

LIQUIDITY MECHANISM

We use Pythagorean Bonding curve to price our decision tokens.

Screenshot 2025-01-13 at 8.22.35 PM.png