Everything saleable is generated from one production capability: point-in-time data infrastructure, a validated factor library, and a documented simulation framework. This page describes how it is built and what must be true before anything is published.
U.S.-listed equities above $300M market capitalization and $2M average daily dollar volume, plus liquid sector, style, and broad-market ETFs. Approximately 2,200 securities at launch. Liquid ADRs added in Year 2.
| Phase | Domain | Content |
|---|---|---|
| Launch | Price and volume | OHLCV, split- and dividend-adjusted, daily |
| Launch | Corporate actions | Splits, dividends, mergers, delistings, symbol changes |
| Launch | Index membership | Historical constituents for survivorship-free universes |
| Launch | Macro | Rates, term structure, credit spreads, volatility indices |
| Year 2 | Fundamentals | Income statement, balance sheet, cash flow; as-reported and restated, point-in-time |
The most common failure in sub-institutional quantitative research is look-ahead bias: testing against data that was not available on the date the decision would have been made. Marnello preserves as-reported values with original publication timestamps, maintains survivorship-free universe snapshots, and constrains every computation to information available at the simulated decision date. Expensive to build, easy to get wrong, and rarely maintained below the institutional price tier.
The Company launches with a factor set requiring only price, volume, and corporate action data, and adds fundamental factors when point-in-time fundamentals are licensed. A rigorously constructed four-factor library is preferred to a six-factor library built on compromised data.
| Factor | Construction | Available |
|---|---|---|
| Momentum | 12-1 month total return; 6-1 month; risk-adjusted variants | Launch |
| Low volatility | Trailing realized volatility, beta, idiosyncratic volatility | Launch |
| Size | Market capitalization decile, float-adjusted | Launch |
| Short-horizon reversal | 1-week and 1-month reversal, volume-conditioned | Launch |
| Value | Earnings yield, book-to-price, FCF yield, EV/EBITDA; sector-neutralized | Year 2 |
| Quality | ROIC, accruals, gross profitability, leverage stability | Year 2 |
Daily-frequency portfolio simulation with configurable rebalancing. Transaction cost modeling using spread-based estimation scaled by volume participation. Position, sector, turnover, and liquidity constraints. Repeated-sampling simulation and walk-forward parameter estimation to limit overfitting.
Every simulation reports annualized return, volatility, Sharpe, Sortino, maximum drawdown and duration, Calmar, turnover, average holding period, hit rate, and factor exposure attribution.
An institutional buyer cannot use a model they cannot document. The documentation is not supporting material. It is the product.