Quantitative Research with FactorBench
Build multi-factor models with custom formulas. Test hypotheses rigorously with transparent, auditable backtests on point-in-time data-no black boxes.
Free Screener · 7-Day Trial on Paid Plans · Cancel Anytime
Why Quants Choose FactorBench
🔬 Custom Factor DSL
Safe, AST-validated formula language for custom metrics-no eval(). Compose factors like (ROE + ROA) / 2 or EV/EBITDA * FCF_Yield.
📊 Multi-Factor Models
Combine value, momentum, quality, and custom factors with flexible ranking, weighting, and sort directions.
🔍 Point-in-Time Rigor
Enforces fundamental lag windows and pins backtests to common price dates-no look-ahead bias, realistic results.
📈 Long Backtests
Test strategies on up to 25 years of historical data to validate performance across different market cycles and regimes.
🔎 Full Diagnostics
Survivor counts, selected tickers, weights, per-factor values-complete transparency for auditable, reproducible research.
📋 Reproducible Results
Save screens and pin methodology-rerun anytime with consistent results. Share research with confidence.
How It Works
Hypothesis Formation
Define your research question: Does combining low P/E with high ROE outperform? Does momentum + quality beat pure momentum?
Build Factor Model
Use built-in factors or create custom formulas with the DSL. Combine multiple factors with ranking and weighting.
Rigorous Backtest
Run point-in-time backtests with equal or cap-weighting. See performance, drawdowns, and diagnostics.
Validate & Document
Review diagnostics, adjust methodology, and save reproducible screens for production or publication.
Example: Quality Value Multi-Factor Model
🎯 Hypothesis
Combining value (low P/E) with quality (high ROE, low debt) should outperform pure value strategies with lower drawdowns.
⚙️ Methodology
Custom composite score: (1/PE + ROE/100) / 2. Filter: Debt/Equity < 0.6. Rank by composite descending, select top 20%.
✅ Results
15-year backtest shows 12.3% CAGR vs 9.8% for pure value, with 18% lower max drawdown. Full diagnostics for auditability.
Custom Formula: quality_value = (1/PE + ROE/100) / 2 Filters: P/E > 0 (positive earnings) Debt/Equity < 0.6 Ranking: quality_value descending Selection: Top 20% by rank Portfolio: Cap-weight Backtest period: 15 years Rebalance: Annually Diagnostics: - Survivor count per period - Portfolio weights - Per-factor values for selected tickers
Built for Quantitative Researchers
Custom Formula Language
Safe DSL for composing metrics-no eval(). Build factors like (ROE + ROA)/2 or EV/EBITDA * FCF_Yield with full type safety.
Multi-Factor Ranking
Combine value, momentum, quality, and custom factors with flexible sort directions and weighting schemes.
Point-in-Time Data
Configurable fundamental lag windows, common price dates-no look-ahead bias. Realistic, trustworthy backtest results.
Full Diagnostics
Survivor counts, selected tickers, weights, per-factor values, TTM/MRQ basis-complete transparency for auditable research.
Research transparency: Every metric reports its source/basis and we handle NaNs defensively. Reproducible methodology for publication-grade research.
FAQ
What is the formula language for custom factors?
A safe, AST-validated DSL for composing metrics like (ROE + ROA)/2 or MarketCap * PriceToBook. No eval() or exec()-only allow-listed operations.
Can I test multi-factor models?
Yes. Combine value, momentum, quality, and custom factors with flexible ranking, weighting, and sort directions for each factor.
What diagnostics are available?
Every backtest shows survivor counts, selected tickers, portfolio weights, per-factor values, and data basis (TTM/MRQ) for full auditability.
What is point-in-time data?
Point-in-time data uses only information available on each date with configurable fundamental lag windows-no look-ahead bias for realistic results.
Is there a free trial?
Yes-the free plan includes full screener access. Paid plans with backtesting include a 7-day free trial with credit card required.
Research Rigorously. Validate Thoroughly. Invest Confidently.
Build multi-factor models with transparent data and auditable backtests.