Quant Strategies Lab
Estrategias cuantitativas con validación rigurosa: Walk-Forward expandible, embargo de 1 día y sin look-ahead bias.
Who is this for?
For investors looking to trade correlated asset pairs (e.g. two ETFs in the same sector) by exploiting temporary deviations from their historical relationship. The strategy goes long the "cheap" asset and short the "expensive" one when the deviation exceeds a statistical threshold.
When does it act? When the spread Z-score exceeds ±2σ according to the ADF cointegration test.
Escáner market-neutral de pares cointegrados. Ejecuta test ADF (Augmented Dickey-Fuller) automatizado sobre todos los pares posibles del universo seleccionado. Los pares con
p-value < 0.05 se marcan como cointegrados y se calcula un Z-score rolling para señalar entradas cuando el spread diverge más de 2σ.
Who is this for?
For those who want to select the best stocks from a universe using multiple criteria simultaneously: momentum, value, quality and low volatility. The model scores and ranks each asset — you invest in the top ones.
When does it act? At the end of each rebalancing period, buying the top N% and selling the bottom N%.
Sistema cross-sectional de scoring compuesto que combina tres factores: Quality (ROIC), Value (EV/EBITDA) y Momentum. Cada factor se normaliza con Z-scores para generar un ranking QARP (Quality at a Reasonable Price) que elimina el sesgo subjetivo de la selección de acciones.
Estrategia long-short basada en la investigación de AQR (Asness, Frazzini, Pedersen 2019) que compra acciones de alta calidad y vende las de baja calidad. El quality score combina tres pilares: Profitability (ROIC, gross margin, FCF/Assets), Growth (crecimiento de profitabilidad 5Y) y Safety (baja volatilidad, bajo leverage). Top quintil = Long, Bottom quintil = Short.
Long-short strategy that exploits the low-volatility anomaly: low-beta stocks deliver higher risk-adjusted alpha than high-beta stocks. The portfolio goes long the lowest-beta quintile (leveraged to beta=1) and short the highest-beta quintile (de-leveraged to beta=1), producing a market-neutral portfolio. Betas are estimated with rolling OLS and Vasicek shrinkage toward the cross-sectional mean.
Stocks with low idiosyncratic volatility (residual after removing the systematic component) deliver better risk-adjusted returns. Choose from 4 volatility estimators: Realized (classic std), EWMA (RiskMetrics), GARCH(1,1) (conditional) or HAR-RV (Corsi 2009, daily/weekly/monthly components).
IVOL Configuration
Estrategia cuantitativa avanzada que combina detección de regímenes via HMM (Hidden Markov Model), XGBoost con Walk-Forward Analysis y sistema de doble embargo (train→val + val→test). Incluye sizing dinámico por volatilidad, régimen y probabilidad; stops Chandelier + CVaR; y optimización de histéresis. Todos los cálculos se ejecutan en el servidor — cero lógica expuesta en el frontend.