{"data":{"kind":"file","path":"README.md","version_id":"kl8zm6x7hr1yepnkolq83j1m","entry":{"name":"README.md","path":"README.md","is_directory":false,"size":1142,"modified_at":"2026-08-02T11:52:06.005000","content_hash":"3e8f18d883bbdc7813dfe302eae69bc93b6660ef0c869cf16d0cd7cde0450c06"},"entries":[],"content":"# Trading Decision Environment\n\nEvaluates an LLM's ability to analyze financial market data and make sound trading decisions.\n\n## Task Format\n\nEach task presents:\n- A market scenario with price data, indicators, and context\n- The model must analyze the data and recommend a trading action (BUY, SELL, HOLD)\n- The model must provide risk management parameters (stop-loss, take-profit)\n\n## Evaluation Criteria\n\n1. **Decision Accuracy** (weight 0.35): Does the model's recommendation align with the ground-truth action given the market conditions?\n2. **Risk Management** (weight 0.25): Does the model set appropriate stop-loss and take-profit levels?\n3. **Reasoning Quality** (weight 0.25): Does the model correctly identify key indicators and patterns?\n4. **Position Sizing** (weight 0.15): Does the model suggest appropriate position sizes based on risk?\n\n## Dataset\n\n100+ trading scenarios across multiple market conditions:\n- Trending markets (bull/bear)\n- Range-bound markets\n- High volatility events\n- Breakout patterns\n- Reversal patterns\n\n## Usage\n\n```python\nimport verifiers as vf\nenv = vf.load_environment(\"vive/trading-decision\")\n```\n","encoding":"utf-8","truncated":false,"total_bytes":1142},"status":null}