Volatility Modeling & Regimes
Forecast volatility with the GARCH family - asymmetry and leverage effects, dynamic correlations, vol targeting - then detect market regimes with HMMs and trade adaptively.
GARCH(1,1): Forecasting Crypto Volatility
How the GARCH(1,1) model captures volatility clustering in crypto, how to fit it by maximum likelihood with the arch library, and how to turn conditional-variance forecasts into position sizing and dynamic stops.
Asymmetric and Heavy-Tailed GARCH: EGARCH, GJR, and Student-t
Plain GARCH(1,1) treats good and bad news the same and assumes Gaussian shocks. EGARCH, GJR-GARCH, and Student-t/skew-t innovations fix both - and give you honest VaR and Expected Shortfall for crypto.
DCC-GARCH: Dynamic Correlations for Pairs and Portfolio Risk
Crypto correlations are not constant - they spike toward 1 in every drawdown. DCC-GARCH models a time-varying correlation matrix, giving you dynamic hedge ratios for pairs trading and honest time-varying portfolio risk.
Volatility Targeting and Trading with GARCH Forecasts
A GARCH volatility forecast is only worth something if it improves a trading decision. We build a volatility-targeted crypto strategy, evaluate forecast quality honestly, and compare GARCH against realized-vol and EWMA baselines in a walk-forward backtest.
Hidden Markov Models in Trading: How to Adapt Your Strategy to Market Regimes
How to identify the current market regime (bull, bear, sideways) using Hidden Markov Models and automatically switch trading strategies. With Python code and backtests.
