From Paper to Paper Trading
QuantMar reads ArXiv papers, extracts trading strategies, backtests them against survivorship-bias-free data, and deploys to IB paper trading. All autonomous. All local.
What It Does
ArXiv Paper Search
Search quantitative finance papers, score implementability, and extract trading signals automatically.
Strategy Generation
LLM generates Python Strategy classes from paper methodology — AST-validated, sandbox-safe.
Backtest Engine
Event-driven backtester with market/limit/stop orders, slippage, commissions, and walk-forward validation.
Auto Mode (CLI)
Overnight loop: fetch papers, code strategies, backtest, mutate, promote. Wake up to candidates.
Survivorship-Bias Free
TimescaleDB universe with delisted tickers. No look-ahead. EODHD adjusted prices.
IB Paper Trading
Same strategy.rebalance() call runs against Interactive Brokers TWS for paper trading.
QuantMar vs QuantConnect
| Feature | QuantMar | QuantConnect |
|---|---|---|
| Daily OHLCV + fundamentals | Yes | Yes |
| Strategy SDK (Python) | Yes | Yes |
| Walk-forward validation | Yes | No |
| ArXiv paper search | Yes | No |
| LLM strategy generation | Yes | No |
| Autonomous overnight research | Yes | No |
| Data ownership (your DB) | Yes | No |
| Intraday / tick data | No | Yes |
| Options / futures | No | Yes |
| Short selling / margin | No | Yes |
Install the CLI
Works on Linux & macOS. Requires Python 3.11+, Docker, and Ollama.
pip install quantmar qf init qf doctor qf auto --universe spy_500 --hours 8