Prediction Market V2

Document Version: 2.0
Last Updated: January 2026
Smart Contract Version: Solidity ^0.8.24
MiCA-Aligned Disclosure

Zetarium Prediction Market V2

Abstract

Zetarium Predictions is an on-chain outcome-based participation platform enabling users to engage in prediction markets with transparent settlement. Prediction markets allow participants to express expectations about future outcomes via market participation rather than opinion or speculation alone.

Disclaimer: Participation in prediction markets involves risk. This document is for informational purposes only. Zetarium does not provide investment advice, guarantees, or custodial services.

1. Overview

1.1 What are Prediction Markets?

Prediction markets are platforms where participants can express expectations about future outcomes through market mechanisms. Unlike traditional speculation, prediction markets utilise:

On-chain accounting
Public settlement logic
Deterministic or oracle-assisted resolution

This minimises trust assumptions and improves fairness.

1.2 Market Design

Each prediction market defines:

Component	Description
Event Definition	A specific event or outcome set
Participation Conditions	Rules for joining and trading
Settlement Criteria	How outcomes are determined
Resolution Logic	AI-based or price-based determination

Markets may include binary or multi-outcome structures and are designed for transparency and auditability.

1.3 Strategic Role

Predictions serve to:

Expand user participation beyond traders
Increase engagement and platform stickiness
Generate complementary liquidity flows
Support decentralised information discovery

Important: Participation does not imply guaranteed outcomes or returns.

2. System Architecture

Zetarium System Architecture

The Zetarium Predictions protocol consists of three main layers:

Layer	Description
Smart Contract Layer	Market management and asset control on the blockchain
Indexing Layer	Fast access and querying of blockchain data
Resolution Layer	AI-powered automatic market settlement system

Users
  |
  v
Frontend (Web Interface)
  |
  v
Application Layer
  +- Predictions Engine
  |     +- Market Logic
  |     +- Settlement Rules
  |     +- Oracle Layer
  |
  v
Smart Contracts (On-Chain Settlement)
  |
  v
Blockchain Infrastructure (BNB Chain)

3. Protocol Components

3.1 Conditional Token System

Zetarium Predictions uses the Gnosis Conditional Tokens Framework to manage tokens representing outcomes of future events. The system operates in binary (YES/NO) format.

How It Works

When a market is created, separate token classes are generated for each possible outcome. Users deposit stablecoins like USDC to receive outcome tokens.

Scenario	Outcome
Correct predictors	Can exchange tokens 1:1 for collateral
Wrong predictors	Tokens become worthless

This mechanism guarantees market settlement and ensures fair gains or losses.

3.2 Automated Market Maker (AMM)

Zetarium provides liquidity and pricing through an AMM system using the Constant Product Formula:

$$x \times y = k$$

Core Principles

Principle	Description
Liquidity Pools	Users can become LPs by depositing USDC
Automatic Pricing	Prices determined by token reserve ratios
LP Earnings	Fees from each trade added to pool value
Arbitrage Mechanism	Price imbalances corrected by arbitrageurs

3.3 Optimistic Oracle System

The security and verification layer uses an optimistic oracle model inspired by UMA Protocol.

Core Philosophy

"Correct information is provided most of the time; intervention is only required in case of errors."

Operating Mechanism

Step	Action
1	Bond Deposit: Reporter deposits USDC collateral
2	Outcome Reporting: Market outcome reported, challenge period begins
3	Dispute Process: Auto-approval if no dispute, voting if disputed
4	Voting: Governance token holders vote for correct outcome
5	Finalisation: Correct report → bond returned, wrong → slashing

Warning: The system is designed based on game theory principles. The cost of false reporting exceeds potential gains.

4. AI-Powered Resolution (V2 Feature)

4.1 Multi-AI Verification System

Zetarium AI Consensus System

The most innovative feature of Zetarium Predictions V2 is that multiple independent AI models form consensus to determine market outcomes.

Decentralised AI Consensus

Instead of trusting a single AI model, multiple models from different organisations are used:

Model	Organisation	Expertise
Gemini Pro	Google	Fact-checking expertise
GPT-4o	OpenAI	Advanced reasoning capabilities
Claude Sonnet	Anthropic	Nuanced understanding
Llama 3.1	Meta	Open-source verification
Mistral Large	Mistral AI	European perspective

4.2 Research and Reasoning Orchestration

The orchestration system disciplines web searching and logical inference:

Principle	Description
Topic-specific trust network	Source only trusted in relevant domain
Source scoring	Originality, verifiability, recency, independence
Search strategy	Multiple engines with operator-supported queries
Logical chain	Claim → Evidence → Contradiction detection → Conclusion
Weighting	Trusted sources receive higher impact coefficients

4.3 Consensus Algorithm

Weighted Voting System

Each AI model's decision is weighted by its confidence score. At least 60% of the total weight must converge on the same outcome.

Minimum Requirements

Requirement	Value
Minimum successful AI responses	3 models
Minimum confidence score per model	50%
Minimum source references	3 different sources
Time verification	Before resolution time

4.4 Evidence Report System

A detailed evidence report is created for each market resolution:

Component	Content
AI Model Results	Each model's decision and confidence score
Source Archiving	Web sources uploaded to Wayback Machine and IPFS
Consensus Details	Voting results and power distribution
Cryptographic Signature	Digital signature proving report authenticity
IPFS Storage	Permanent, decentralised storage

4.5 Price-Based Resolution

For objective data like cryptocurrency prices, a deterministic oracle system is used:

Oracle	Description
Chainlink	Decentralised price feeds
Pyth Network	High-frequency price data
Uniswap V3	On-chain liquidity pools
Binance	Centralised exchange prices

Type	Process
AI-Based Markets	Multiple AI models research → Consensus → Evidence report → Reporting
Price-Based Markets	Multi-oracle price fetch → Median calculation → Threshold comparison

5.5 Dispute Process

A 1-hour challenge period begins for reported outcomes:

Situation	Result
No Dispute	Automatically approved and finalised
Disputed	Community voting begins

5.6 Redemption

When the market is finalised:

Correct predictors: Exchange tokens 1:1 for USDC
LPs: Burn LP tokens to claim pool share and fees

6. Security Mechanisms

6.1 Economic Security

Mechanism	Description
Bond System	Cost of false reporting exceeds potential gains
Slashing Mechanism	False reporters lose 100% of bond
Whistleblower Reward	Disputers receive share of slashed bonds

6.2 AI Security

Protection	Description
Prompt Injection Protection	User inputs marked as "untrusted"
Multi-Model Tolerance	Single AI error caught by consensus
Evidence Archiving	All sources uploaded to IPFS and Wayback
Confidence Thresholds	Low confidence outcomes marked as "UNCERTAIN"

6.3 Smart Contract Security

Protection	Description
Reentrancy Protection	Protection against repeated call attacks
Safe Token Transfers	Secure libraries for ERC20
Access Control	Critical functions only callable by authorised addresses
Overflow Protection	Automatic protection with Solidity 0.8+

7. Risk Disclosure

Participation Risks

Participation in Zetarium Predictions involves risks including:

Risk Category	Description
Market Risk	Prediction outcomes may result in total loss of position
Smart Contract Risk	Despite audits, vulnerabilities may exist
Oracle Risk	AI or price oracle failures may affect resolution
Liquidity Risk	Insufficient liquidity may impact trading
Regulatory Risk	Regulatory treatment of prediction markets may change

Important Disclaimers

Zetarium provides no guarantees regarding market outcomes
Participation does not constitute investment advice
Users should only participate with funds they can afford to lose
Past performance does not indicate future results

8. Settlement & Transparency

Prediction markets utilise:

On-chain accounting
Public settlement logic
Deterministic or oracle-assisted resolution

This minimises trust assumptions and improves fairness.

Resources

Resource	Link
Website	prediction.zetarium.world
Documentation	docs.zetarium.world
Twitter	@Zetarium_
Telegram	t.me/Zetarium_World
GitHub	github.com/zetariumworld

This documentation is provided for informational purposes only. Zetarium does not provide investment advice, guarantees, or custodial services.

On this page