Volatility is not constant. It comes in clusters. High-volatility days tend to follow other high-volatility days, and low-volatility periods tend to persist until they do not. This is one of the most robust empirical findings in financial markets, and it has direct implications for how you should manage risk.
The statistical term for this is heteroskedasticity, and GARCH models (Generalized Autoregressive Conditional Heteroskedasticity) were developed specifically to capture it. A GARCH model estimates current volatility as a function of recent returns and recent volatility estimates. When a large move occurs, the model's volatility estimate jumps up and then gradually decays back toward its long-term average. This captures the clustering effect and produces more accurate risk estimates than assuming constant volatility.
For practical regime detection, you do not necessarily need a full GARCH model. Simpler approaches work well. Comparing the 20-day realized volatility to the 60-day or 200-day average can tell you whether you are in a higher or lower volatility regime relative to recent history. When short-term volatility exceeds long-term by a significant margin, you are in a high-vol regime. The reverse indicates low-vol.
Why regimes matter for trading: different strategies perform differently in different volatility environments. Trend-following strategies tend to perform well when volatility is rising or elevated, because trends tend to be more directional and persistent during turbulent periods. Mean reversion strategies tend to work better in low-volatility, range-bound environments where prices oscillate around a central value.
The VIX index provides a market-implied volatility regime indicator for equities. Readings below 15 suggest complacency and tend to be associated with low-volatility regimes that favor range-bound strategies. Readings above 25 indicate elevated fear and tend to coincide with trending markets where momentum and defensive strategies outperform. The transition zones are the trickiest to navigate.
In crypto, the absence of a widely followed implied volatility index makes regime detection more reliant on realized volatility measures. The 30-day annualized volatility of Bitcoin serves as a rough proxy. Historical ranges suggest that annualized volatility below 40% is relatively calm for crypto, while above 80% indicates a high-vol regime. These thresholds are much higher than equity market equivalents, reflecting crypto's fundamentally different volatility structure.
One actionable application: scale your position sizes inversely with volatility. When volatility doubles, halve your position sizes. This keeps your dollar risk roughly constant across regimes rather than letting it expand during exactly the periods when large losses are most likely. This approach, sometimes called volatility targeting, has been shown to improve risk-adjusted returns across multiple asset classes.
Regime changes tend to happen quickly. The shift from low to high volatility is typically abrupt (driven by a shock or surprise), while the shift from high to low is gradual (as fear slowly fades). This asymmetry means you need to be quicker in recognizing and responding to regime shifts toward higher volatility than in the other direction. The cost of being slow to reduce risk is higher than the cost of being slow to increase it.