Laytus - parlay prediction markets

Laytus is a structured prediction market protocol that solves a fundamental pricing flaw in multi-outcome positions, traditional parlays multiply probabilities assuming full independence, which systematically misprices correlated risk and exposes liquidity providers to adverse selection when expanded to prediction market events. We built a 5-layer NLP and ML correlation detection pipeline that identifies shared entities, logical implications, and category relationships between markets, feeding into a Gaussian copula model that correctly prices joint probability. Every position is packaged as an EIP-712 signed certificate and verified against 18 independent on-chain risk checks before a single dollar moves. Liquidity providers deposit USDC into an ERC-4626 vault with a 7-day withdrawal delay for bank-run protection, earning yield from defaulted positions and protocol fees. Investors receive on-chain NFT receipts for their positions, with future features including a P2P secondary marketplace, thematic Composites, cross-margin matching, and a junior/senior tranche model all on the roadmap. Laytus is built on Robinhood Chain with a Python FastAPI microservices backend, React frontend, and a continuously hardened ML pricing model that ingests real-time adversarial flow.

During the hackathon we executed a full UI redesign, overhauling the color scheme, branding, and component library to give Laytus a polished, production-ready feel. We brought the Earn and Portfolio pages fully online, wiring them to live on-chain data so LPs can deposit, initiate withdrawals, and track positions in real time. A major milestone was migrating the entire protocol stack to Robinhood Chain Testnet, which involved redeploying all smart contracts, updating EIP-712 domain binding to chain ID 46630, reconfiguring the backend RPC and settlement executor, and locking the frontend exclusively to Robinhood Chain via RainbowKit. We also did a comprehensive documentation refresh, updating  architecture diagrams, service descriptions, and deployment references to accurately reflect the current state of the protocol 

N/A, in talks with some teams rn