Melquiades is a quantitative value stock screener built on the methodology from Wesley Gray and Tobias Carlisle's Quantitative Value (2012). It ranks global stocks across 10 regions on four fundamental factors — TEV/EBIT, Piotroski F-Score, ROIC, and Accruals — producing a single composite QV Score. No editorial judgement. No narrative. The numbers.

The name comes from García Márquez's One Hundred Years of Solitude. Melquiades is the wandering Gypsy alchemist who arrives in Macondo bearing strange instruments — magnets, telescopes, ice. He could decode patterns others dismissed as chaos. His parchments, written in a language no one understood, contained the entire history of the Buendía family before it happened.

That's the bet: cheap, financially healthy businesses will eventually be recognized as such, and a disciplined unemotional screen — applied consistently across thousands of stocks — will outperform human stock-picking over long horizons.

The composite ranking methodology — equal-weighted percentile ranks of TEV/EBIT, F-Score, ROIC, and Accruals — is derived directly from Gray and Carlisle's systematic backtests across US and global equity markets. Their research showed the four-factor composite beats any single factor in isolation and outperforms human stock-picking across every market studied.

Additional factor foundations: Piotroski (2000) for F-Score, Altman (1968) for Z-Score, Sloan (1996) for the accruals anomaly, and Greenblatt (2005) for the Magic Formula earnings yield concept.

Melquiades is open-source. The QV engine, factor formulas, and data pipeline are all readable and auditable — no black box. The academic methodology is public, the data is available, the engineering is not complicated. If you find a bug or a better factor formulation, open a pull request.