optopsy

A nimble backtesting and statistics library for options strategies.

Optopsy is a Python backtesting engine that lets you go from "How do 45-DTE iron condors on SPX perform with a 50% profit target and 2x stop loss vs holding to expiration?" to detailed performance statistics in seconds, not spreadsheets.

Full Documentation | API Reference | Examples

🤖 Looking for AI/LLM integration?

optopsy-mcp provides a high-performance MCP server for strategy screening and backtesting. Powered by a complete Rust rewrite of the Optopsy engine, it is specifically built for seamless interaction with Large Language Models.

Features

• 38 Built-in Strategies - From simple calls/puts to iron condors, butterflies, condors, ratio spreads, collars, calendars, and diagonals
• Per-Leg Delta Targeting - Select strikes by delta with target, min, max per leg
• Trade Simulator - Chronological simulation with capital tracking, position limits, and equity curves via simulate()
• Portfolio Simulation - Weighted multi-strategy portfolio backtesting via simulate_portfolio()
• Early Exits - Stop-loss, take-profit, and max-hold-days rules for automatic position management
• Commissions - Model broker fees with per-contract, base fee, and min fee structures
• Risk Metrics - Sharpe, Sortino, VaR, CVaR, Calmar, Omega, tail ratio, and more via compute_risk_metrics()
• 80+ Entry Signals - Filter entries with TA indicators (RSI, MACD, Bollinger Bands, EMA, ATR, IV Rank) via pandas-ta-classic
• Custom Signals - Use custom_signal() to drive entries from any DataFrame with a boolean flag column
• Slippage Modeling - Realistic fills with mid, spread, liquidity-based, or per-leg slippage
• Live Data Providers - Fetch options chains and stock prices directly from supported data sources (e.g. EODHD)
• Smart Caching - Automatic local caching of fetched data with gap detection for efficient re-fetches
• Plugin System - Extend with custom strategies, signals, and data providers via entry points

Installation

# Core library only (latest stable release)
pip install optopsy

# With Data CLI (download & cache market data)
pip install optopsy[data]

Requirements: Python 3.12-3.13, Pandas 2.0+, NumPy 1.26+

Data Management

Optopsy includes a standalone data CLI for downloading and caching historical market data.

pip install optopsy[data]

# Download historical options data (requires EODHD_API_KEY)
optopsy-data download SPY
optopsy-data download SPY AAPL TSLA

# Download stock price history
optopsy-data download SPY --stocks

# List available symbols
optopsy-data symbols
optopsy-data symbols -q SPY

# Cache management
optopsy-data cache size
optopsy-data cache clear

Data is cached locally as Parquet files at ~/.optopsy/cache/. Re-running download only fetches new data since your last download. See the Data Management documentation for full details.

Core Library Quick Start

import optopsy as op

# Load your options data
data = op.csv_data(
    "options_data.csv",
    underlying_symbol=0,
    option_type=2,
    expiration=3,
    quote_date=4,
    strike=5,
    bid=6,
    ask=7,
    delta=8,
)

# Backtest long calls and get performance statistics
results = op.long_calls(data)
print(results)

Output:

   dte_range   delta_range  count   mean    std    min    25%    50%    75%    max  win_rate  profit_factor
0    (0, 7]   (0.2, 0.3]    505   0.64   1.03  -1.00   0.14   0.37   0.87   7.62       ...            ...
1    (0, 7]   (0.3, 0.4]    269   2.34   8.65  -1.00  -1.00  -0.89   1.16  68.00       ...            ...
2   (7, 14]   (0.2, 0.3]    404   1.02   0.68  -0.46   0.58   0.86   1.32   4.40       ...            ...
...

Results are grouped by DTE (days to expiration) and delta range, showing descriptive statistics for percentage returns. Every run also returns win_rate (the fraction of trades with a positive return) and profit_factor (gross wins over gross losses); the values above depend on your dataset and are elided here.

Simulator

Run a full trade-by-trade simulation with capital tracking, position limits, and performance metrics:

result = op.simulate(
    data,
    op.long_calls,
    capital=100_000,
    quantity=1,
    max_positions=1,
    selector="nearest",       # "nearest", "highest_premium", "lowest_premium", or custom callable
    max_entry_dte=45,
    exit_dte=14,
)

print(result.summary)         # win rate, profit factor, max drawdown, etc.
print(result.trade_log)       # per-trade P&L, entry/exit dates, equity
print(result.equity_curve)    # portfolio value over time

The simulator works with all 38 strategies. It selects one trade per entry date, enforces concurrent position limits, and computes a full equity curve with metrics like win rate, profit factor, max drawdown, and average days in trade.

여기까지가 README 앞부분입니다. 나머지는 원문에서 읽을 수 있습니다. README 전문 보기 ↗