Backtesting Frameworks

• You are developing or refining a trading strategy and need a rigorous backtest from hypothesis through evaluation.
• You are building backtesting infrastructure, including point-in-time data pipelines, execution simulation, and cost models.
• You want to validate strategy robustness using train/validation/test splits or walk-forward analysis.
• You need to identify and eliminate common backtesting biases before drawing conclusions about a strategy.

• You need live trading execution or investment advice—this skill covers simulation only.
• Your historical data quality is unknown or incomplete, which would undermine any results.
• You only need a quick, high-level performance summary rather than a full backtesting system.

Guides construction of data pipelines that ensure only information available at each historical moment is used, eliminating look-ahead bias and producing trustworthy results.

Incorporates realistic transaction costs, slippage, and execution assumptions into simulations so that performance estimates reflect real-world trading conditions.

Supports building event-driven backtesting engines that accurately model order flow and fills, closely mirroring how strategies would behave in live markets.

Provides patterns for out-of-sample validation through walk-forward testing and proper dataset splits, reducing the risk of overfitting and inflated performance metrics.

Covers common backtesting pitfalls—such as survivorship bias, look-ahead bias, and overfitting—with concrete guidance on how to detect and eliminate them.

Includes a detailed resources/implementation-playbook.md with patterns and examples that can be referenced for in-depth, step-by-step implementation guidance.

A quant researcher defines a trading hypothesis and uses this skill to build a full backtest—from data pipeline to performance evaluation—ensuring results are free from common biases.

An engineering team constructing a reusable backtesting platform uses this skill to implement event-driven simulation, cost models, and validation frameworks that meet production standards.

A strategy that performed well in initial tests is subjected to walk-forward analysis and out-of-sample validation to confirm that its edge is genuine and not a result of overfitting.

A trader reviews an existing backtest using this skill's guidance to identify whether look-ahead bias, survivorship bias, or unrealistic cost assumptions are inflating the reported returns.