GARCH and a volatility rank as market-regime classifiers

When the tape changes without warning

A market can leave a short-horizon reversal pattern and enter a prolonged directional run without a gradual warning. A strategy that stays in the old mode can absorb a sizable drawdown before it is adjusted.

Regime switching is presented as a way to retune a strategy when conditions move from a directional mode to a mean-reverting mode. A volatility measure sits at the center of most such algorithms.

Both the level of volatility and whether that volatility is rising or falling are used as a first indication of whether the market looks directional or mean-reverting.

Building the volatility-switch-indicator

The construction is described as needing at least one year of observations. Three to five years or more are said to give a more reliable view of how a security responds when volatility changes.

A 21-session historical-volatility series is built from successive mid-price percentage changes. Historical-volatility here is the standard deviation of those changes over that fixed window. The series is then converted to a zero-to-one rank by counting how many of the recent volatility readings are less than or equal to the latest reading. That rank is the volatility-switch-indicator: a score that crosses 0.5 to flag a change in market personality.

A switch value that rises through 0.5 is treated as increasing volatility and a choppy, mean-reverting setting. A value that declines through 0.5, especially from a recent reading near 1, is treated as falling volatility and the start of a directional episode.

Rising volatility is framed as higher uncertainty and a choppy tape. Falling volatility is framed as greater directional certainty. The 21-session window is an unoptimized starting point rather than a fitted lookback.

Which tools sit with each side of 0.5

Market-regime-classification in this workflow is the rule that labels the tape as trending or mean-reverting so a different indicator family can be applied.

Above the 0.5 threshold the setting is mean-reversion-mode: a choppy, higher-uncertainty state. Short-horizon reversal tools such as a stochastic oscillator or a relative-strength index are paired with the switch.

Below 0.5 the setting is trending-mode: a lower-volatility, higher-certainty state. Trend tools such as MACD or a moving-average crossover are paired instead.

The same decision beside a GARCH forecast

A simple historical-volatility rank is presented as a starter regime classifier that can sit beside more formal volatility-forecast constructions, including GARCH-based forecasts used to switch a trading system.

A GARCH model is the formal counterpart to that rank: a quantitative volatility model that turns ordered price observations and a defined lookback into an explicit forecast. The volatility forecast is a projected path used to anticipate whether conditions are likely to stay choppy or become more directional.