backtest-runner

Backtest financial trading strategies against historical data with MLflow tracking.

Updated Apr 11, 2026

Search Tags:#backtest #mlflow #risk-metrics #walk-forward #vectorbt #pbo #pybroker

Ready to run

npx skills add https://github.com/bitandbytes/Argus --skill backtest-runner-bitandbytes

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: backtest-runner
Source: https://github.com/bitandbytes/Argus/tree/main/.claude/skills/backtest-runner
Command: npx skills add https://github.com/bitandbytes/Argus --skill backtest-runner-bitandbytes

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This skill guides users in running backtests, walk-forward optimization, and historical strategy validation across different backtesting engines. It helps ensure robust evaluation by enforcing lookahead-bias checks, logging with MLflow, and comparing parameter sets.

Core Features & Use Cases

Full-pipeline backtesting across PyBroker for end-to-end strategy evaluation.
Walk-forward optimization and PBO analysis to assess robustness and avoid overfitting.
Validation checks for lookahead biases and performance degradation across out-of-sample periods.
MLflow integration for experiment tracking and result reproducibility.
Parameter sweep support with vectorized backtesting for rapid experimentation.

Quick Start

Run a backtest by executing the backtest script with a ticker and a date range.

Frequently Asked Questions about backtest-runner

High-intent search queries and answers about installing and using this skill.

FAQPage Schema

How do I run a walk-forward optimization to backtest trading strategies without overfitting?▼

Walk-forward optimization backtests trading strategies by dividing historical data into sequential training and testing windows. It validates parameter robustness and calculates Probability of Backtest Overfitting (PBO) to prevent overfitted results.

Can I use vectorbt or pybroker for full-pipeline backtesting across multiple tickers?▼

Yes, full-pipeline backtesting supports both PyBroker and VectorBT engines to evaluate trading strategies across multiple tickers and time ranges. This enables end-to-end strategy assessment and rapid parameter sweeps.

What is lookahead bias in backtesting and how do I check for it?▼

Lookahead bias occurs when a backtest uses future data unavailable at the time of trading. The backtest pipeline includes validation checks to detect lookahead bias and performance degradation across out-of-sample periods.

Does MLflow integrate with backtesting workflows for experiment tracking?▼

MLflow integrates with the backtesting workflow to log experiments, track parameter sets, and ensure result reproducibility. This allows you to compare historical strategy validation runs and maintain consistent tracking.

How do I calculate risk metrics during a historical strategy validation?▼

Historical strategy validation calculates configurable risk metrics alongside backtest execution. These metrics are applied across multiple tickers and time ranges to assess strategy performance and robustness.

What is the best way to compare parameter sets when running vectorized backtests?▼

Vectorized backtesting supports parameter sweeps to rapidly compare different parameter sets. This approach accelerates experimentation and evaluates strategy robustness across varied configurations.