A decentralized prediction market protocol built on Solana that allows users to bet on price movements of cryptocurrencies using real-time price feeds from Pyth Network.

PythPredict enables users to create binary prediction markets where participants can bet YES or NO on whether an asset's price will move within a specified timeframe. The protocol uses Pyth Network's high-fidelity price oracles to ensure fair and transparent market resolution.

• Real-Time Price Feeds: Integration with Pyth Network for accurate BTC/USD and ETH/USD price data
• Binary Markets: Simple YES/NO betting on price movements
• Automated Resolution: Markets resolve automatically based on oracle prices at settlement time
• Fee System: 1% platform fee on all bets
• Fair Payout Distribution: Winners receive proportional shares of the total pool
• Token Conservation: Guaranteed token conservation with no loss of funds

• Node.js v16+ and Yarn
• Rust 1.70+
• Solana CLI tools 1.17+
• Anchor Framework 0.31.1

1. Clone the repository

git clone https://github.com/TrebuchetNetwork/PythPredict
cd PythPredict

2. Install dependencies

yarn install

3. Build the Solana program

anchor build

4. Set up local validator (optional for local testing)

./scripts/setup-local.sh

The project includes a comprehensive test suite (master-test.ts) that validates all core functionality.

anchor test

The master test suite (tests/master-test.ts) performs extensive testing across 5 main sections:

• Creates prediction markets with correct parameters
• Places bets and validates fee calculations (1% fee)
• Updates odds based on bet distribution
• Resolves markets after settlement time
• Distributes winnings proportionally to winners
• Prevents double-claiming of winnings

• Handles minimum bet amounts
• Maintains perfect token conservation (no tokens lost)
• Resolves empty markets with no bets

• Creates and manages multiple concurrent markets
• Places randomized bets across markets

• Fetches real BTC price from Pyth Network mainnet
• Creates market at actual BTC price (~$118,829 in example)
• Simulates different trading strategies
• Monitors price movement for 20 seconds
• Resolves based on actual price changes

📊 TEST SUITE FINAL REPORT
============================================================
📈 Test Results:
   Total Tests: 15
   Passed: 15 (100.0%)
   Failed: 0

💰 Token Conservation:
   Initial Supply: 3500.00 tokens
   User Balances: 2922.18 tokens

✅ EXCELLENT: All critical tests passed!

# Build the program
anchor build

# Run tests
anchor test

# Deploy to devnet
anchor deploy --provider.cluster devnet

# Monitor test coverage
./coverage.sh

programs/pythpredict/
├── src/
│   ├── lib.rs           # Program entry point
│   ├── instructions.rs  # Core instruction handlers
│   ├── state.rs         # Account structures
│   └── errors.rs        # Error definitions

1. Market Account: Stores market parameters, pools, and resolution data
2. Position Account: Tracks user bets and payouts
3. Vault Accounts: Hold YES/NO pools and fees
4. Oracle Integration: Pyth price feeds for resolution

• initialize_market: Create a new prediction market
• place_bet: Place a YES or NO bet
• resolve_market: Resolve using oracle price
• resolve_with_external_price: Resolve with manual price (testing)
• claim_winnings: Claim payouts after resolution

• Program ID: J7TLVPzbd47RpiHV8BBPLQuixU53P5qijkrwkvN4u98W
• Default cluster: Localnet
• Test script: yarn run ts-mocha -p ./tsconfig.json -t 1000000 tests/master-test.ts

Test wallets are stored in .wallets/ directory:

• alice, bob, charlie, dave, eve (test participants)
• Each funded with SOL and test tokens

1. Market Creation: Creator sets initial price, settlement time, and oracle feed
2. Betting Phase: Users bet YES (price will move) or NO (price stays same)
3. Price Monitoring: Oracle prices tracked in real-time
4. Resolution: At settlement, compare final vs initial price
5. Payout: Winners share the total pool proportionally

User Payout = (User Stake / Winning Pool) * Total Pool

Example: If YES pool has 100 tokens, NO pool has 50 tokens, and YES wins:

• YES bettors share 150 tokens proportionally
• Each receives their stake + proportional share of losing pool

• PDA-based account derivation
• Signer verification for all critical operations
• Overflow protection with checked math
• Time-based settlement enforcement
• Oracle confidence validation

MIT

Contributions are welcome! Please read our contributing guidelines before submitting PRs.

For issues, questions, or suggestions, please open an issue on GitHub.