Quick backtest a strategy on a symbol. Creates a complete .py script with data fetch, signals, backtest, stats, and plots.
Detected risks:
The backtest skill is a specialized tool for quantitative traders and algorithmic trading enthusiasts who need to quickly validate trading strategies against historical market data. It automates the entire backtesting workflow by generating complete Python scripts that fetch data, apply trading signals, run performance simulations, and produce comprehensive analytical reports. Instead of manually writing boilerplate code for data retrieval, signal generation, and portfolio simulation, users can invoke this skill with a simple command specifying their strategy, symbol, exchange, and timeframe.
This skill leverages VectorBT for high-performance vectorized backtesting, integrates with OpenAlgo for Indian market data access, and supports multiple technical analysis libraries including TA-Lib for standard indicators and OpenAlgo's custom indicators for specialized strategies. It handles the complexities of Indian equity and futures markets, including accurate fee structures for delivery trading and F&O contracts, lot size requirements for index futures, and benchmark comparisons against NIFTY 50. The generated scripts include signal cleaning with exrem to eliminate duplicate entries, proper configuration loading from environment files, and automatic generation of QuantStats tearsheets and Plotly visualizations.
The skill is designed for both retail traders conducting strategy research and professional quants building systematic trading systems. It supports popular strategies like EMA crossover, RSI mean reversion, Donchian breakouts, Supertrend trend following, MACD momentum, and custom algorithms like SDA2 and dual momentum. Whether you're testing a simple moving average strategy on a single stock or validating a complex multi-indicator system on index futures with 5-minute bars, this skill generates production-ready backtest code that handles data loading, signal generation, portfolio simulation, performance metrics, benchmark comparison, trade export, and visual reporting in a single automated workflow.
The skill primarily uses OpenAlgo API for fetching Indian market data (NSE, NFO exchanges). It also supports loading data directly from DuckDB databases, with automatic format detection for Historify-style schemas (market_data table with epoch timestamps) and custom OHLCV schemas. When using DuckDB, the skill can operate standalone without OpenAlgo dependencies.
The skill includes templates for 10+ strategies: EMA crossover, RSI, Donchian channels, Supertrend, MACD, SDA2, Momentum, Dual Momentum, Buy & Hold, and RSI Accumulation. It uses TA-Lib for standard indicators (EMA, SMA, RSI, MACD, Bollinger Bands, ATR, ADX) and OpenAlgo's ta library for specialized indicators (Supertrend, Donchian, Ichimoku, HMA, KAMA, ALMA).
The skill applies accurate fee structures for Indian markets: 0.111% fees plus ₹20 fixed charges for equity delivery, and 0.018% fees plus ₹20 for futures. It correctly handles lot sizing for index futures (NIFTY: 65 lots, BANKNIFTY: 30 lots effective Dec 31, 2025) and uses NIFTY 50 as the default benchmark with proper exchange specification (NSE_INDEX).
Each generated script produces: complete portfolio statistics from VectorBT, a comparison table showing strategy vs benchmark performance (returns, Sharpe ratio, Sortino ratio, max drawdown, win rate, trade count, profit factor), plain-language explanation of results, QuantStats HTML tearsheet (if available), Plotly charts showing equity curve and drawdown with dark theme, and CSV export of all trades with entry/exit details.
Yes, the generated Python scripts are fully editable. They're created in organized directories (backtesting/strategy_name/) and follow clear patterns based on templates. You can modify indicator parameters, adjust fee structures, change position sizing rules, add custom filters, or integrate additional analysis. The scripts use standard libraries (VectorBT, TA-Lib, pandas) making customization straightforward.
Quick Setup:
.claude/skills/Repository
marketcalls/vectorbt-backtesting-skills