Skip to main content
Understanding how Kash’s automated market mechanics work helps you make better predictions and maximize returns.
Hero Light

​
How Market Creation Works

AI agents create all markets through a comprehensive validation process. While users can suggest markets, the actual creation involves rigorous AI evaluation and quality control.
Automated Market Detection5 Specialized AI Personas continuously monitor social signals:
  • Crypto-Degen Agent: Token listings, ETF approvals, DeFi milestones
  • Sports Agent: Match outcomes, player performance, championships
  • Tech Agent: Product launches, AI breakthroughs, earnings results
  • Macro Agent: Interest rate decisions, economic indicators, elections
  • Pop-Culture Agent: Award shows, album drops, celebrity drama
Creation Process:
  1. Signal Detection - AI identifies trending topics with prediction potential
  2. Quality Assessment - Validates verifiable outcomes and clear criteria
  3. Market Generation - Creates binary YES/NO markets with resolution parameters
  4. Instant Deployment - Markets go live with automatic liquidity bootstrap

​
Core Market Structure

Simple YES/NO OutcomesEvery Kash market has exactly two possible outcomes:
  • YES: The stated event will happen
  • NO: The stated event will not happen
  • Winner-take-all: Only one outcome can be correct at resolution
  • Clear criteria: Each market specifies exact resolution conditions
Example Market:
"Will Bitcoin reach $150,000 by March 1, 2025?"
YES: $0.65 (65% implied probability)
NO: $0.35 (35% implied probability)
Total: $1.00 (complete market)
Outcome Tokens as Tradeable Assets
  • Share Pricing: Ranges from 0.01to0.01 to 0.99 per share
  • Market Completeness: YES + NO prices always equal $1.00
  • Probability Reflection: Share price directly represents market consensus
  • Continuous Trading: Prices update with every transaction
Share Value at Resolution:
  • Winning outcome shares: $1.00 each
  • Losing outcome shares: $0.00 each
Revolutionary Social-Adapted Pricing TechnologyKash uses a proprietary bonding curve that revolutionizes prediction market liquidity:
  • Zero Bootstrap Time: Instant liquidity without waiting for providers
  • Whale-Resistant Design: Advanced anti-manipulation mechanisms
  • Social Intelligence: Adapts to viral content and engagement patterns
  • Built-in Leverage: Natural position amplification without borrowing
  • Short-term Memory: Context-aware pricing based on recent activity
Technical Implementation:
  • Built on Base blockchain for low fees and fast execution
  • Smart contracts with integrated social signal processing
  • Mathematical curve with time, volume, and engagement variables
  • Advanced memory functions for intelligent price discovery
Learn more about our Custom AMM Bonding Curve and its revolutionary advantages.

​
Pricing and Liquidity

Probability-Based PricingMarket prices directly reflect outcome probabilities:
  • YES at $0.70 = 70% implied probability
  • NO at $0.30 = 30% implied probability
  • Always sums to $1.00 for complete market
  • Prices adjust with each trade based on supply and demand
Price Movement Drivers:
  • Buying Pressure: Increases price of purchased outcome
  • Selling Pressure: Decreases price of sold outcome
  • Trade Size: Larger trades create bigger price impacts
  • Market Depth: More liquidity reduces price volatility

​
Complete Market Lifecycle

1

Market Creation & Validation

AI-Driven Process with Quality Control
  • AI agents detect trending topics or process user suggestions
  • Validation phase: Mira Network ensures verifiable outcomes and clear criteria
  • Content safety: AI checks prevent prohibited topics
  • Market parameters set: Question, resolution criteria, trusted sources, 30-day maximum deadline
  • Instant Liquidity Bootstrap: Passive LP Vault automatically deploys USDC
2

Active Trading Period

Open Price Discovery
  • Continuous trading available 24/7
  • Prices fluctuate based on new information and trader activity
  • Market consensus develops through collective prediction
  • Real-time odds visible in X conversations and web interface
  • Maximum duration: 30 days before automatic closure
3

Market Closure

Trading Freeze
  • Trading halts at predetermined closure time
  • All positions locked at final prices
  • No new predictions accepted after closure
  • Preparation for automated resolution begins
4

Automated Resolution

AI-Powered Settlement with Cryptographic Proofs
  • zkTLS Verification: Reclaim Protocol fetches data with cryptographic proof of authenticity
  • AI Analysis: EZKL zero-knowledge ML proves correct YES/NO determination
  • Combined Proof: zk-SNARK submitted to resolution contract on Base blockchain
  • Instant Settlement: Winners automatically receive $1.00 per winning share
  • Platform fee: 2% charged on all predictions

​
Fee Structure and Economics

2% Fee on All TradesFee Application:
  • 2% fee on all trades (buying or selling shares)
  • No additional fees on winnings, losses, or holding positions
  • Transparent calculation shown before and after each trade
Example Calculation:
Trade: $100 on YES at $0.40 = 250 shares
Platform Fee: $100 × 2% = $2.00
Net Cost: $102.00 total
If YES wins: 250 shares × $1.00 = $250
Net Profit: $250 - $102 = $148.00 (145% return)
Platform-Sponsored Transactions
  • All transaction fees covered by Kash protocol
  • ERC-4337 Integration: Account abstraction enables sponsored transactions
  • No ETH Required: Users don’t need to hold ETH for gas fees
  • Base Network: Low underlying costs due to Layer-2 efficiency
What’s Covered:
  • Market creation transactions
  • Prediction placement and position changes
  • Resolution and payout transactions
  • Wallet funding and withdrawal operations
Advanced Price Impact MitigationMulti-Layer Protection:
  • Automatic Slippage Limits: 5% maximum slippage on social media predictions
  • Dynamic Pricing: Real-time price updates prevent stale price execution
  • MEV Protection: Large trades routed through private mempools
  • Batch Execution: Multiple trades executed simultaneously at fair prices
How It Works:
Social Media Prediction: "YES $100"
1. System checks current market price
2. Calculates maximum shares at 5% slippage
3. Executes trade with price protection
4. Refunds excess if better price achieved
Protection Levels:
  • Small Predictions (under $50): 2% maximum slippage tolerance
  • Medium Predictions (over $50): 5% maximum slippage tolerance
  • Large Predictions (over $100): Automatic routing to RFQ system
  • Institutional Predictions (over $200): MEV protection with batch execution
User Benefits:
  • Predictable Costs: Know maximum cost before trade execution
  • Fair Pricing: Protection against price manipulation
  • Automatic Optimization: System finds best execution path
  • Refund Protection: Excess funds returned if better price achieved
Sustainable Protocol EconomicsRevenue Sources:
  • 2% platform fee on all trades
  • Interest earned on idle USDC deposits
  • Potential MEV capture from transaction ordering
Revenue Usage:
  • Monthly $KASH token buybacks
  • Protocol development and security improvements
  • User acquisition and community rewards
  • Automated liquidity provisioning for new markets

​
Advanced Features

Flexible Trading OptionsBuying and Selling:
  • Purchase YES or NO shares at current market price
  • Exit positions before market resolution
  • Sell into automated liquidity pool
  • Lock in profits or limit losses
Portfolio Tracking:
  • Real-time position values and P&L
  • Historical performance analytics
  • Export capabilities for tax reporting

​
Resolution and Settlement

Pre-Defined Resolution CriteriaEach market specifies exact resolution parameters:
  • Primary Sources: Official websites, APIs, verified data feeds
  • Secondary Sources: Backup verification from reputable outlets
  • Resolution Criteria: Specific language defining winning conditions
  • Dispute Process: Community governance for edge cases
Common Source Types:
  • Government agencies and official bodies
  • Sports leagues and official results pages
  • Financial data providers (Bloomberg, Reuters)
  • Company investor relations and SEC filings
Trustless Settlement with Zero-Knowledge ProofsHow It Works (Step-by-Step):
  1. Market Closure: Trading stops at predetermined time
  2. Secure Data Fetching: AI agent connects to pre-specified trusted source (e.g., official results page)
  3. zkTLS Authentication: Cryptographic proof generated that data actually came from the claimed website
  4. AI Analysis: Natural Language Processing model analyzes the fetched data to determine YES/NO outcome
  5. Zero-Knowledge Proof: Mathematical proof generated that the AI model made the correct decision on authentic data
  6. Blockchain Verification: Combined proof submitted to smart contract on Base blockchain
  7. Instant Settlement: Contract automatically pays winners $1.00 per winning share
Why This Matters:
  • No human bias - AI follows predetermined rules consistently
  • Cryptographically secure - impossible to fake data sources or manipulate AI decisions
  • Instant resolution - no waiting for committees or manual verification
  • Publicly verifiable - anyone can check that resolution was correct
  • Dispute-resistant - mathematical proofs eliminate most potential disagreements
Trust Through Technology: Instead of trusting a person or company to determine outcomes fairly, you’re trusting mathematics and cryptography - the same technology that secures Bitcoin and Ethereum. The AI model weights are public, the data sources are specified in advance, and the entire process generates cryptographic proofs that can be independently verified.Publicly Verifiable AI Models:
  • Open Model Weights: All AI model parameters publicly available for inspection
  • Deterministic Inference: Same inputs always produce identical outputs
  • Cryptographic Commitment: Model weights committed to blockchain before market creation
  • Version Control: All model updates tracked and governed by community
Auditability Features:
  • Reproducible Results: Anyone can verify AI decisions using public weights
  • Model Transparency: Complete neural network architecture publicly documented
  • Training Data Disclosure: Information about training datasets and methodologies
  • Bias Detection: Community can analyze models for potential biases or manipulation
How Model Verification Works:
Verification Process:
1. Download public model weights from IPFS/Arweave
2. Run same input data through model locally
3. Compare your result with on-chain resolution
4. Mathematical proof confirms identical computation
Privacy-Preserving Features:
  • Selective Disclosure: Only necessary data elements are revealed in proofs
  • Data Minimization: Personal information in source data is automatically redacted
  • Computation Privacy: AI inference happens without exposing sensitive intermediate values
  • Audit Without Exposure: Verification possible without revealing private data sources
Community Oversight: The combination of public model weights and zero-knowledge proofs creates unprecedented transparency. The community can verify that AI models are fair and unbiased, while the zkTLS system ensures the data being analyzed is authentic. This eliminates the need to trust any centralized authority for market resolution.Technical Proof Chain Architecture:
  • Layer 1 - Data Authenticity: zkTLS proves data came from claimed source
  • Layer 2 - AI Inference: EZKL proves correct model execution on authentic data
  • Layer 3 - Result Verification: Combined proof submitted to Base blockchain
  • Layer 4 - Community Audit: Public verification of entire proof chain
Proof Composition Process:
Proof Chain Generation:
1. zkTLS generates proof P1: "Data D came from source S"
2. EZKL generates proof P2: "Model M correctly processed D → Result R"
3. Proof combiner creates P3: "P1 ∧ P2 → R is valid"
4. P3 submitted to resolution contract on Base
5. Smart contract verifies P3 and executes payout
Cryptographic Guarantees:
  • Non-repudiation: Impossible to deny data source or AI decision
  • Immutability: Proofs cannot be modified after generation
  • Verifiability: Anyone can independently verify the proof chain
  • Privacy-Preservation: Sensitive data elements remain hidden
Comprehensive CoverageMarket Invalidation:
  • Event cancelled or significantly changed
  • All participants receive full refund of original stakes
  • No fees charged for invalid markets
Ambiguous Outcomes:
  • Community governance process activated
  • $KASH token holders vote on resolution
  • Majority decision determines final outcome
Technical Failures:
  • Backup resolution mechanisms available
  • User funds always protected
Understanding these market mechanics gives you an edge in prediction trading. The combination of AI validation, automated liquidity, and cryptographic resolution creates the most advanced prediction market system available.

Start Trading

Learn to make predictions

Find Markets

Discover opportunities