AI trading: backtesting from scratch
Learn to test an idea with data before risking capital: backtesting, market data, and guided skills.
Direct answer: An entry pack for testing strategies with historical data and market data in the assistant. It combines Python backtesting frameworks with an exchange-connectivity library and data MCP servers, plus skills that guide the analysis and the verdict on your own idea. Designed to work first in simulation with risk management, not to copy third-party signals.
Included in the Esencial — $15
Plans are cumulative: the Pro Plan includes all Essential Plan packs, and the Complete Plan includes all packs (Essential + Pro). This pack unlocks with the Esencial ($15).
Buy with the Esencial ($15) Or build your custom plan
Use cases
- Test a trading idea with historical data before trading
- Compare strategies with Python backtesting frameworks
- Consult market data from the assistant via MCP
- Review the verdict on your own idea, with risk management
How to get started
- Start with the simplest resource in the pack and try it on a small real case.
- Connect the second resource only once the first one already works in your daily flow.
- Review the full list of 10 resources and keep the ones you use every week.
Suggested starter prompt
Act as my assistant using the tools from the pack «Essential Trading: Backtesting from Scratch». Learn to test an idea with data before risking capital: backtesting, market data, and guided skills. Start with the simplest resource, explain each step before running it, and ask for confirmation before changing data, settings, or permissions.
Included resources (10)
- Freqtrade (Software, freqtrade/freqtrade, ★ 55.9k) — Python cryptocurrency trading bot with backtesting, simulated mode (dry-run), parameter optimization, and the FreqAI machine learning module. GitHub
- VectorBT (Software, polakowo/vectorbt, ★ no data) — Python library for vectorized backtesting on pandas and NumPy, accelerated with Numba, for testing thousands of parameter combinations and analyzing portfolios. GitHub
- backtesting.py (Software, kernc/backtesting.py, ★ no data) — Lightweight Python framework for testing strategies with historical OHLCV data, with trade statistics and interactive charts. GitHub
- Backtrader (Software, mementum/backtrader, ★ no data) — Event-based Python library for strategy backtesting, with indicators, analyzers, and a readable bar-by-bar workflow. GitHub
- CCXT (Software, ccxt/ccxt, ★ no data) — Unified library for connecting to 100+ cryptocurrency exchanges and getting market data, balances, and histories through a single API. GitHub
- yfinance-mcp (ozp) (MCP Server, ozp/yfinance-mcp, ★ no data) — MCP server with Yahoo Finance data: prices, financial statements, options, holders, and news, with SQLite caching and multi-market support. GitHub
- Financial Datasets MCP Server (MCP Server, financial-datasets/mcp-server, ★ no data) — Financial Datasets MCP server with US stock and cryptocurrency fundamentals: financial statements, ratios, prices, and news. GitHub
- claude-trading-skills (agiprolabs) (Skill, agiprolabs/claude-trading-skills, ★ no data) — Collection of 68 trading, DeFi, and quantitative finance agent skills: market data, technical analysis, backtesting, risk management, and portfolio analysis. GitHub
- strategy-tester (Skill, apexstoa/strategy-tester, ★ 2) — Claude skill that tests your own trading idea with your own price data and delivers a numeric verdict (GO, NO-GO, or INCONCLUSIVE), without signals or recommendations. GitHub
- vectorbt-backtesting-skills (Skill, marketcalls/vectorbt-backtesting-skills, ★ no data) — Backtesting skill with vectorbt: configuration, execution, optimization, quick statistics, strategy comparison, and reusable templates. GitHub
Versions by tool
Related packs
- Pro Trading: AI Agents and Broker Connection
- Complete Trading: Bots, Analysis, and Portfolio
- Essential Sales Agent
- Pro Prospecting and Sales
Frequently asked questions
What is this pack?
A set of 10 tools for testing trading ideas with historical data before trading.
Who is it for?
For anyone starting in systematic trading who wants to validate an idea with data and risk management.
What does it include?
Python backtesting frameworks, an exchange-connectivity library, market-data MCP servers, and guided analysis skills.
Can I trade with real money using this pack?
This pack is presented for analysis and paper trading: testing strategies with historical data without risking capital. Trading with real money requires setting up a broker, supervising every trade, and accepting the risk of loss; the pack makes no promises of results.