Regime Detection
Regime detection is the quantitative process of identifying statistically distinct market environments (regimes) from on-chain data — periods where the relationships between volume, volatility, flows, and returns follow different structural dynamics. Market regimes in crypto include: trending bull markets (elevated correlation, declining dispersion), mean-reverting chop (high intra-range volatility, rangebound prices), liquidity crisis (spike in bid-ask spreads, withdrawal congestion), and volatility regime shifts (transition from low to high volatility clusters). On-chain data sources for regime detection include DEX swap volumes, liquidity provision/withdrawal waves, exchange net flows, stablecoin minting/redemption, and gas price dynamics. Hidden Markov Models, GARCH-family models, and structural break tests are common statistical tools. Regime detection is practically applied to dynamic LP management (width adjustment for volatility regime), algorithmic execution (aggressiveness calibration), and risk management (volatility targeting for structured products).