qback - Options Backtesting Engine

A parallelized options backtesting engine designed for brute-force parameter optimization and strategy discovery. It tests various parameter combinations to identify profitable, robust trading strategies based on a configurable set of rules.

Key Features

• High-Speed, Parallelized Core: Utilizes multiprocessing to run thousands of backtests per minute across all available CPU cores.
• Comprehensive Parameter Grid-Search: Tests all combinations of entry/exit times, strike offsets, stop-loss rules, and re-entry logic.
• Advanced Re-entry Strategies: Implements various re-entry mechanisms, including strategies that correctly reverse market opinion after a stop-loss.
• Configurable Stop-Loss Types: Supports multiple SL types, including percentage of premium, points-based, and index movement (points and percentage).
• Composite Score Ranking: Ranks viable strategies using a weighted composite score based on key performance metrics (ROI, Return/MDD, Win Rate, etc.).
• Detailed Performance Metrics: Calculates ROI, Max Drawdown, Return/MDD Ratio, Win Rate, Expectancy, and more for every strategy.
• Data-Driven: Uses the efficient Parquet file format for fast data loading and in-memory lookups.

Quick Start

1. Install Dependencies:

   pip install -r requirements.txt

2. Configure Your Strategy: All parameters, fees, and ranking criteria are controlled in src/config.yaml. Edit this file to define:

   • Entry and exit times.
   • Stop-loss types and values.
   • Re-entry strategies and max attempts.
   • Slippage and brokerage costs.
   • Filtration criteria for viable strategies.
   • Weights for the composite score.

3. Run the Backtest: Execute the backtest from the src directory:

   cd src
   python main.py

Data Format

The engine requires data in the Parquet format for performance.

Index Data (.parquet)

Column	Type	Description
timestamp	datetime	The timestamp of the index price
index_close	float	The closing price of the index

Options Data (.parquet)

Column	Type	Description
timestamp	datetime	The timestamp of the option price
expiry	datetime	The expiry date of the contract
strike	float	The strike price of the option
option_type	string	'CE' for Call, 'PE' for Put
close	float	The closing price of the premium

the aux/ directory has a program convert_to_parquet.py that helps convert csv(s) to parquet(s).

Re-entry Strategies Explained

The engine tests a set of canonical re-entry strategies after a stop-loss is hit.

• RE-ASAP: Re-enters a new trade immediately at the current at-the-money (ATM) strike.
• RE-COST: Waits for the premium of the original option to return to its original entry price before re-entering.
• RE-MOMENTUM: Waits for a confirmation of momentum (3 consecutive moves) in the underlying index before re-entering at the new ATM.

Reversal Logic (_REVERSE)

The _REVERSE strategies implement a reversal of market opinion. When a trade is stopped out, the engine assumes its initial market view was wrong and enters a new trade in the opposite direction.

• RE-MOMENTUM_REVERSE: If a bearish Short Call is stopped out by a rally, the engine waits for confirmation of upward momentum and then enters a bullish Long Call.
• RE-COST_REVERSE: If a bearish Short Call is stopped out by a rally, the engine waits for the first pullback (dip) in the index to enter a new bullish Long Call at a better price.

Output Files

All output files are saved to the project root directory:

• runs.csv: Contains the detailed performance metrics for every single parameter combination tested.
• viable_strategies.csv: A filtered subset of runs.csv containing only the strategies that met the performance criteria defined in config.yaml.
• trades_<strategy_id>.csv: A detailed, trade-by-trade log for the single best-performing strategy identified from the viable set (ranked by composite score).

Project Structure

.
├── aux/                      # Auxiliary scripts and notebooks
├── src/
│   ├── backtest_engine.py    # The core backtesting logic
│   ├── main.py               # Main execution script
│   ├── config.yaml           # ALL strategy parameters and fees
│   ├── Nifty-Index-Monthly-Data.parquet  # Sample index data
│   └── Nifty-Options-Weekly.parquet    # Sample options data
├── README.md                 # This file
└── requirements.txt          # Python dependencies

Flowchart of the Backtest Engine