GJR-GARCH volatility modelling, Filtered Historical Simulation, and AI-powered news cross-analysis — engineered for investors who act on signal, not sentiment.
GJR-GARCH(1,1,1) captures the leverage effect — negative shocks amplify volatility more than positive ones. Calibrated on 10 years of price history using Skewed Student-t innovations. Not Gaussian assumptions. The real distribution.
Cross-references live news against price structure and volatility regime using Claude AI. Surfaces hidden catalysts, flags contradictions, outputs a directional verdict. Every output is a decision input, not a summary.
Filtered Historical Simulation on GJR-GARCH residuals generates 10,000 forward paths, condensed into a calibrated High/Low range with explicit probability. Each horizon recalibrated independently. A number you can size a position from.