Volatility methods guide

Understanding Intraday and Hourly Volatility

A practical guide to building a time-of-day volatility profile from short-interval returns, comparing UTC hours and separating historical averages from forecasts.

Updated September 7, 2026 · Evergreen educational guide

What is volatility?

Volatility describes the dispersion of returns in a defined sample. Wider or larger short-interval changes produce a higher measured reading; more tightly grouped changes produce a lower one. Volatility measures variability, not whether the market moved up or down overall.

Every reading depends on the instrument, venue, data frequency, return definition, sample window and scaling convention. The annualized-volatility guide covers those broader foundations; this guide focuses on comparisons by hour of the day.

What is hourly realized volatility?

Hourly realized volatility summarizes variability observed inside an hour from shorter-interval data. A study can calculate one reading for each hour and then group observations that share the same clock hour across multiple days. The result is an intraday profile with one comparable bucket for each hour.

“Realized” means that the measure is built from price changes that occurred in the sample. It does not mean that the same value will be realized in a future hour. The exact result changes with the formula, sampling frequency, timezone and data-cleaning rules.

How can an hourly volatility profile be obtained?

1. Fix the marketDeclare the instrument, venue, provider and price field.

2. Fix timeSelect a timezone and assign every short-interval observation to a clock hour.

3. Form returnsUse one documented return formula and sampling interval consistently.

4. Measure each hourApply the declared realized-volatility convention to the eligible observations.

5. Group by clock hourCombine comparable readings for 00:00, 01:00 and every remaining hour.

6. Summarize and countReport the average or other statistic together with sample size and cutoff date.

This is a general reproducible framework, not a claim about one undisclosed implementation. A report should also state how it handles missing intervals, duplicates, outliers, partial hours and days with incomplete data.

What does high hourly volatility mean?

Under a fixed method, a high hourly reading means the included short-interval returns were more dispersed than those in a lower reading. It can mark a period in which price variability was historically concentrated, but it does not identify direction or establish that the clock caused the movement.

Abrupt repricing, concentrated order imbalance, reduced available liquidity or leveraged position adjustments can coincide with high variability. They are possible research hypotheses, not explanations that can be proven from a volatility bar alone. Testing them requires separate evidence for the same observations.

What does low hourly volatility mean?

A low hourly reading means the included short-interval returns were more tightly grouped under the same calculation. It describes quieter realized variation in the sample, not a guarantee that the next observation will remain quiet.

More balanced trading, greater available liquidity or the absence of abrupt repricing can coincide with lower variability. As with a high reading, the metric does not prove those conditions were responsible. A low historical average can still contain individual high-volatility sessions that are hidden by the summary.

What can hourly volatility be useful for?

Time-of-day context: identify hours that historically contained more or less variability.
Monitoring design: decide when a market process may require closer observation.
Sample comparison: compare the same UTC hour across compatible rolling windows.
Execution research: investigate whether separate liquidity, spread or order-flow data change by hour.
Risk communication: distinguish a historical activity profile from a directional forecast.

These are analytical uses, not automatic trading instructions. A ranking should be tested for stability before it is treated as a durable characteristic of the market.

Limitations and common mistakes

Ignoring timezone: the same instant belongs to different clock hours under different zones.
Confusing volatility with direction: variability does not say whether the net return was positive.
Reading rounded ties as exact rankings: displayed precision may not distinguish close estimates.
Hiding the distribution: an average does not show every day's result or the effect of outliers.
Mixing methods: different return, scaling or cleaning rules can change the profile.
Predicting repetition: a historical high-volatility hour is not guaranteed to remain the highest.

Minimum reporting standard

State the instrument, venue, provider, price field, timezone, sampling interval, return formula, realized-volatility formula, grouping statistic, sample size for every hour, missing-data treatment, units, window dates and cutoff timestamp.