Quant Research Validation Toolkit

Quant Research Validation Toolkit is a public, synthetic-data demonstration of research-bias controls for systematic trading research.

The toolkit focuses on the failure modes that most commonly make backtests look better than they are:

• look-ahead bias
• label leakage
• survivorship bias
• missing point-in-time universe rules
• overlapping-label leakage
• inadequate embargo / purge windows
• missing data availability checks
• market-structure tradability assumptions

This repository is designed as a public methodology showcase. It does not contain proprietary strategy code, private parameters, live platform configuration, real trading logs, or paid vendor data.

What This Repository Demonstrates

• Point-in-time feature availability audit
• Label/feature leakage guard
• Purged and embargoed train/test split
• Point-in-time universe manifest
• Survivorship-bias audit
• Data availability audit
• Tradability / limit-state audit
• Reproducible synthetic audit report

Quick Start

python -m pip install -e ".[dev]"
python -m pytest -q
python examples/run_validation_audit_demo.py --config configs/sample_config.yaml

Expected Outputs

runs/<run_id>/
  audit_report.json
  audit_summary.csv
  universe_manifest.json
  split_report.csv

Boundary

This is not investment advice and not a trading system. Public examples are synthetic and intentionally simplified.