HadiAlizade77/add-strategy

Add Trading Strategy

Create a new trading strategy: $ARGUMENTS[0] in category $ARGUMENTS[1].

Project Context

• Strategies live in packages/backend/src/services/strategy-engine/strategies/{category}/
• Every strategy implements the StrategyModule interface
• Strategies are registered in the strategy engine and configured via the strategies DB table
• All parameters are configurable from the UI via the strategy_configs table
• Signals are written to the signals table

Strategy Interface

interface StrategyModule {
  id: string;
  name: string;
  category: MarketCategory;
  strategyType: StrategyType;
  allowedRegimes: Regime[];
  evaluate(context: StrategyContext): Promise<Signal | null>;
  getRequiredData(): DataRequirement[];
  validateConfig(params: Record<string, any>): ValidationResult;
}

Scaffolding Steps

1. Create strategy file: packages/backend/src/services/strategy-engine/strategies/$1/$0.strategy.ts
2. Implement the StrategyModule interface with:
   • evaluate() — core logic returning a Signal or null
   • getRequiredData() — what external data this strategy needs
   • validateConfig() — Zod schema validation for parameters
3. Define a Zod schema for all configurable parameters with defaults, min, max, and descriptions
4. Register the strategy in the strategy engine's registry
5. Create a DB seed entry for the strategies table
6. Create unit test: packages/backend/tests/unit/strategies/$0.test.ts with:
   • Known input → expected signal output
   • No signal when conditions not met
   • Edge calculation accuracy
   • Confidence scoring edge cases
   • Regime filtering

Signal Output

{
  strategy_id: UUID,
  market_id: UUID,
  direction: 'buy_yes' | 'buy_no' | 'sell_yes' | 'sell_no' | 'no_trade',
  outcome_token: string,
  confidence: number,       // 0 to 1
  estimated_edge: number,   // fair_value - market_price - estimated_cost
  estimated_cost: number,   // fees + slippage estimate
  fair_value: number,       // model's fair probability
  market_price: number,     // current Polymarket price
  reasoning: string,        // human-readable explanation
  raw_data: object,         // all inputs that produced this signal
  regime: string,           // current regime
}

Testing Requirements

• Minimum 5 unit tests per strategy
• Test with fixed deterministic inputs for reproducibility
• Test edge cases: zero liquidity, stale data, extreme prices (0.01, 0.99)
• Verify estimated_edge calculation matches expected formula