Realized Volatility

What is realized volatility?

Realized volatility is the annualized standard deviation of an underlying’s observed returns over a completed period, a measure of how much the asset actually moved. The standard close-to-close calculation takes a series of returns, measures their standard deviation, and annualizes the result. Historical volatility, actual volatility, and statistical volatility are common aliases, so check the method behind whichever label your platform uses.

The metric is backward-looking by construction. A 20-day realized volatility reading summarizes the last 20 return observations. Implied volatility instead comes from option prices and refers to a future horizon. Comparing them is central to volatility trading, and the comparison closes once the option’s horizon has passed and realized volatility covers the same period.

Realized volatility has no directional sign. Large gains and large losses both increase return dispersion. A smooth 10% rally can realize less volatility than a stock that finishes unchanged after repeated 3% swings.

How is realized volatility calculated?

The close-to-close method begins with daily prices and calculates simple or logarithmic returns. It then computes the sample standard deviation and multiplies by the square root of the assumed number of trading days in a year, commonly 252. The square-root rule follows from variance scaling with time under standard assumptions.

Match the method to the risk you carry:

• Close-to-close volatility includes overnight and regular-session movement between closes, the risk a position held overnight carries.
• Open-to-close volatility isolates the regular session and leaves out overnight gaps, the risk a day trader carries.
• Intraday estimators use shorter intervals and need microstructure-noise controls.
• Range-based estimators use high and low prices to extract more information from each day.
• Fixed-window measures drop one old observation as each new return enters.
• Exponentially weighted measures give recent returns more influence, so they react faster to a change in regime.

Variance is the square of volatility. Variance adds through time, which is why volatility indexes calculate in variance terms before converting back to volatility.

The annualized number is a pace. It puts the observed daily dispersion on a one-year scale, so a five-day window and a 60-day window read in the same units.

Realized vs implied volatility

Options traders compare implied volatility at entry with realized volatility over the option’s life. A delta-hedged long option makes money when the path delivers enough tradable movement to beat its premium, decay, and execution costs. A short option makes money when realized movement stays contained, and when a gap hits, it outweighs many quiet days, because short-option losses accelerate with the size of the move.

The horizons must align. Comparing a 30-day implied measure with a five-day realized spike mixes different windows. Track rolling realized volatility during the life of the option and compute the final value over the exact start and end observations.

Set the measurement rule before the trade. If an option is entered at Monday’s close and expires at Friday’s close, the realized series should span those same endpoints and include every intended overnight interval. Choosing the most favorable intraday sampling after seeing the outcome makes the implied-versus-realized comparison unreliable.

How traders use realized volatility

Realized volatility supports regime analysis. A rising 10-day measure above the 60-day measure shows recent movement accelerating. Comparing overnight and intraday components reveals where risk is occurring. Strip scheduled announcements from the series to estimate ordinary-day movement, and keep them in to measure the total risk a holder lived through.

Risk managers use realized covariance and correlation alongside individual volatilities to estimate portfolio movement. An index can realize less volatility than its components because gains and losses offset across stocks.

Traders also split realized variance by session. Earnings and other overnight news put much of an event-heavy stock’s variance into the gaps between sessions. A strategy that hedges only during market hours carries that overnight piece unhedged, even when two stocks show identical close-to-close volatility.

How to choose a realized volatility window

Choose the sampling interval and lookback that match the trade horizon. Close-to-close returns measure one path; intraday and high-low measures capture additional movement within the session.

Annualize with one day-count convention, such as 252 trading days, and keep comparisons on the same basis. Inspect large observations entering or leaving the window so a rolling change is connected to the returns that created it.

Compare observed movement with the IV paid or collected, then evaluate the actual payoff and hedge results. Direction, gaps, skew, timing, and execution all contribute to the position’s profit or loss.

Limitations of realized volatility

• Check the calendar before you trust a quiet reading. Realized volatility covers a completed window, so a calm 20-day lookback misses the earnings date inside the option’s life.
• Match the method before you compare. Lookback length, return interval, close source, annualization factor, and overnight treatment all change the result, so line up all five before you read one stock’s 16.46% against another’s.
• Read direction from the chart. Repeated reversals lift the chart’s 10-day reading above 40% while price finishes near where it started.
• Check what left the window. A fixed window jumps when a large return enters and drops when it leaves, so a sharp drop on a quiet day marks an old shock rolling off.
• Price the trade, not just the volatility gap. Realized beating implied favors the delta-hedged long option, and gaps, path, timing, and execution still decide the final profit or loss.

Bottom line

Realized volatility measures how much an underlying actually moved, expressed on a common annual scale. Fix the window, sampling, and annualization convention before the trade, then compare the result with the implied volatility paid or collected over the same horizon and read it beside the actual payoff and hedge results.