Mechanism
How Bayes Market Works Under the Hood
Bayes is not just about predicting outcomes — it’s about doing so with fairness, speed, and liquidity. Here's how the system operates.
🔄 Hybrid AMM/CLOB Trading Model
Bayes uses a hybrid trading architecture combining the strengths of Automated Market Makers (AMMs) and Central Limit Order Books (CLOBs):
Key Components:
CFMM (Constant Function Market Maker): Each prediction market uses a bonding curve to algorithmically price outcome tokens (e.g. YES / NO), backed by on-chain reserves. Prices update automatically based on supply and demand.
Order Book First, AMM Second: The system prioritizes matching user orders through a CLOB interface. If no match is found, the AMM fills the gap using the bonding curve to ensure trades can still execute without delay.
Core Benefits:
Always-on Liquidity: Even if no one else is trading, the AMM ensures you can buy or sell at fair prices.
Probabilistic Pricing Logic: Prices are normalized such that:
Price(YES) + Price(NO) = 1This keeps market odds intuitive and mathematically coherent.Smart Contract-Based Settlement: All trades and payouts are enforced transparently via smart contracts — no human intervention needed.
ERC-20 Composability: Outcome tokens are issued as standard ERC-20s, allowing integration into DeFi, wallets, and analytics tools.
🧠 Oracle and Dispute Resolution
Market outcomes must be trustworthy — here’s how Bayes ensures that.
Bayes Market uses a multi-layered oracle system that blends automation, optimism, and arbitration — with decentralization as its end goal.
🔍 Key Architecture Highlights:
Combines objective on-chain data (e.g. blockchain events) and subjective real-world outcomes (e.g. election results, sports)
Designed for speed, manipulation resistance, and transparent credibility
Backed by a DAO-governed fallback layer for edge cases
We believe market resolution is the foundation of trust — and have built a system flexible enough to handle a wide range of scenarios.
🧪 Phase 1: Semi-Centralized Reporting (Current)
In the current launch phase, outcomes are submitted via:
Trusted reporters, including the Bayes team or verified sources
A 2-hour dispute window post-expiry for anyone to challenge a result
This setup ensures fast resolution while allowing users to raise challenges in borderline or controversial cases.
In situations such as:
Event cancellations
Ambiguous rules
Unexpected edge conditions
Bayes activates a fallback arbitration flow, informed by internal checks and external data.
🛡️ Future: Optimistic + Decentralized Settlement
Bayes is actively evolving toward a more resilient and trustless model:
Adopting a UMA-style optimistic oracle where proposed results can be disputed and escalated
Introducing stake-based dispute resolution
Integrating community governance and delegated arbitration
Allowing a transparent appeals process tied to user reputation
This multi-layered system will eventually become community-driven, auditable, and highly scalable — suitable for both niche meme markets and serious macro predictions.
🔄 Continuous Improvement
Bayes’ oracle system is not static. We're iterating continuously based on:
Market feedback
Dispute outcomes
Reporter accuracy
Scaling and decentralization milestones
We’re building for the long run — with fairness, speed, and resilience at the core.
Last updated