What Is Volatility in Trading?
Volatility is a measure of how far price moves over a period, and it contains no information about direction. It rises in a selloff and in a rally alike, it arrives in clusters rather than evenly, and its main practical use is setting position size and stop distance.
Volatility gets talked about as though it were a mood — markets are “nervous”, or “calm”. It is a measurement, it has no opinion about direction, and the number it produces is the one that should be deciding how large your position is.
How it works
Volatility measures distance travelled, not direction. Take a period, measure how far price moved within it, average that over some number of periods. The answer is in the instrument’s own price units and it is always positive.
This is why “volatility is up” tells you nothing about what to do. A market rising violently and a market falling violently produce the same reading. Financial media almost always says it during a decline, which quietly teaches people that volatility means down. It does not.
It clusters. Large bars are followed by large bars far more often than chance would produce, which is the single most reliable thing anyone knows about volatility. It is also the reason a stop that was generous last week can be tight this week without you changing anything.
Quiet clusters too, which is what makes a long calm stretch feel like a permanent state right up until it ends.
A worked example
Here are the actual numbers from this site’s shared series.
The median bar range is 0.493. The ninetieth percentile is 1.101 — so one bar in ten is more than twice the size of a typical one. The largest single bar was 2.338, about 4.7 times the median.
Now put a stop on it. A fixed stop of 0.5 sits just above the median bar. That means roughly half of all ordinary bars are large enough to reach it on their own, before the market has done anything that resembles a move against you.
Sized against volatility instead, the stop moves with conditions. The median ATR14 on this series is 0.5994 and the ninetieth percentile is 0.7954 — so a 1.5-ATR stop is about 0.90 in a typical stretch and about 1.19 in a loud one. Same rule, different distance, because the market changed and the rule noticed.
And the position size follows from it. A wider stop with the same money at risk means a smaller position. That is the whole mechanism: volatility does not tell you whether to trade, it tells you how much room to leave and therefore how much to buy.
And it is measured backwards. Every one of these figures describes bars that have already printed. A volatility reading is a statement about the recent past being used as an estimate of the near future, which works because the clustering is real and fails at exactly the moments the clustering breaks.
What it is actually for
Two jobs, and neither is prediction. The first is stop distance: how far away a level has to sit before ordinary movement stops reaching it. The second is position size, which follows directly, because a wider stop with the same money at risk buys fewer units.
Everything else people do with it is optional. Volatility filters, regime switches and expansion signals are all built on top, and all of them inherit the same limitation — the number says how far, never which way.
It is not volume
Volume is how much traded. Volatility is how far price moved. They often rise together and they are different measurements, and a market can travel a long way on thin volume — which is exactly what a gap is.
A gap is distance covered with nothing changing hands. It counts fully toward volatility and it offered no opportunity to act, which is why it is the case that breaks stop-loss arithmetic.
The original data
On this site’s shared series the median bar range is 0.493 and the ninetieth percentile is 1.101. The largest single bar measured 2.338. Median ATR14 is 0.5994, ninetieth percentile 0.7954.
The spread between typical and extreme is the finding. A bar at the ninetieth percentile is more than double the median, and the largest is nearly five times it. Any rule expressed as a fixed distance is therefore a different rule in each of those conditions, while appearing to be the same rule to the person using it.
A round trip costs 0.0098 here, about 2% of the median bar range — and about 0.9% of that largest bar. Even trading costs mean something different depending on the volatility you are in.
When it fails
The characteristic failure is reading rising volatility as a direction. Bars get bigger, the chart looks dramatic, and it feels like information about what comes next. It is not — the same expansion precedes continuations and reversals, and it appears at the bottom of declines and the top of rallies alike. Acting on it as a signal means taking a position on the basis of a number that was never about direction, and the outcome is decided by something the measurement never contained.
A second failure is using a fixed percentage stop across conditions, which silently means risking a different multiple of ordinary movement in every regime.
A third is assuming a calm stretch will continue. Clustering makes quiet feel permanent, and the transition out of it is usually abrupt rather than gradual.
A fourth is trading the same size in every regime, which is the fastest way to turn an ordinary losing run into an account-ending one.
And a fifth is confusing implied with realised volatility. One is what the options market expects, the other is what actually happened. They disagree regularly, and the disagreement is itself a market that people trade — but they are not interchangeable inputs to a stop distance.
Related
ATR covers the standard way of measuring this and turning it into a stop distance. Risk management covers position sizing, which is where the number is actually spent. And stop loss covers where the level belongs once the distance is known.
For years I treated a quiet market and a wild one as the same problem and used the same stop distance in both. That is the mistake this page exists to describe. Volatility is not a signal and it will not tell you what happens next — it tells you how much room the market currently needs, and if you ignore that you get stopped out by perfectly ordinary movement.
— Michael Whitman
This page is educational, not financial advice. Test every idea on your own charts before risking money.