Volatility and Average Range

The short answer

Volatility measures how much price has moved around its own average over some past window. It is directionless, backward-looking, and says nothing about whether you personally are exposed to a loss.

Its genuine use is arithmetic: a typical range tells you how far price routinely travels, which tells you how far away a stop has to sit to survive ordinary movement, which — for a fixed risk percentage — tells you your position size.

What the measures actually describe

Two families of measure get called "volatility", and they are not the same thing.

Return-based volatility is the standard deviation of returns over a window, usually annualised. It answers: how dispersed have the period-to-period changes been?

Range-based volatility looks at the distance between high and low. Average true range is the common version — the average, over N periods, of the largest of: the current high minus the current low, the current high minus the previous close, and the previous close minus the current low. The "true" part exists to capture gaps, which a plain high-minus-low misses.

true range     = max(high − low, |high − prev close|, |prev close − low|)
average range  = mean of the last N true ranges

Both are computed from OHLC data you already have. Neither contains information about the future; they are summaries of what has already been recorded.

Range-based measures are usually the more useful of the two for a trader, because they are denominated in the same units as your stop. A standard deviation of returns expressed as a percentage has to be converted before it tells you anything about where to put an order. An average range is already in points.

Volatility is not risk

These get used interchangeably and they should not be.

Volatility is a property of the instrument. It exists whether or not you hold a position, and it is symmetric — a measure of movement, with no opinion about up or down.

Risk is a property of your position. It is your exposure to loss, and it depends on size, stop distance, correlation with everything else you hold, and whether you can actually exit at the price you assumed.

A market with low measured volatility can still ruin an account, if the position in it is large enough. The mechanism is not subtle: low volatility invites larger size, because the stop seems safe to place close by, and a larger position is more exposed to the day the volatility regime changes. The reading that made the position feel safe is the reading that made it big.

Regulators talk about volatility in terms of market-wide machinery — FINRA's description of the guardrails covers single-stock trading pauses under the Limit Up/Limit Down plan and market-wide circuit breakers tied to declines in the S&P 500, and the SEC's investor education material covers the same ground. Those exist to interrupt disorderly trading. They are not a floor under your position, and a halt is one of the conditions in which you cannot exit at all.

A more honest sentence than "this market is risky" is "my exposure in this market is X, and the market routinely moves Y".

Average range, stop distance, and size

Here is the chain that makes average range worth computing.

1. A stop placed inside the market's ordinary daily movement will be hit by ordinary daily movement. That is not a bad trade; it is a badly measured one.
2. So the stop distance has to be at least large enough to sit outside routine noise — a multiple of average range is one common way to express that, alongside a structural level from ranges, consolidation and breakouts.
3. Stop distance is risk per unit.
4. For a fixed risk percentage, position size is risk amount divided by risk per unit.

So average range does not tell you whether to trade. It tells you how big the trade can be, given what you already decided to risk.

The numbers below are invented for teaching. They describe no real instrument and are not a trade idea.

• Account equity: 20,000
• Risk per trade: 1% → 200
• Average daily range over the last 14 sessions: 4.00
• Stop distance chosen: 1.5 × average range = 6.00

Position size is 200 ÷ 6.00 = 33 units (rounding down).

Now the same account a month later, after conditions have changed. Average range is 8.00. Same 1.5 multiple gives a stop distance of 12.00. Position size is 200 ÷ 12.00 = 16 units.

The risk is identical at 200 in both cases. The position halved because the market's typical movement doubled. This is the same relationship set out in position sizing explained, and the position size calculator does the arithmetic if you would rather not.

The failure mode is what happens if you skip step 4. Keeping the stop at 6.00 in the second environment means being stopped out by movement the market now makes routinely. Keeping the size at 33 units while widening the stop to 12.00 means risking 396 — roughly 2% — while believing you are risking 1%.

What a regime change does to the measurement

Every volatility measure is an average of a lookback window. That has an unavoidable consequence: at the moment conditions change, the measure still describes the conditions that just ended.

Concretely, when a quiet market turns violent:

• The measure is too low, because most of its window is still made of quiet periods.
• It therefore implies a stop that is too tight for the new environment.
• Which implies a position that is too large, because size was derived from that stop.
• And it corrects slowly — a 14-period average needs a meaningful share of its window replaced before it reflects the new level.

It behaves the same way in the other direction, staying too high after a violent period ends and making positions smaller than they need to be. That error is cheaper.

A shorter lookback reacts faster and is noisier; a longer one is stabler and lags more. There is no setting that removes the lag, because the lag is the averaging. Anyone offering a volatility measure that anticipates regime changes is offering a forecast, not a measurement.

Two practical consequences follow. First, a range-based stop distance is a planned distance, not a guaranteed one: FINRA's guidance is explicit that a triggered stop becomes a market order and can fill at a price markedly different from the stop price, and that is most likely precisely during the transition your measure has not caught yet. Second, the per-trade risk percentage should be small enough that being one regime behind is survivable — which is the argument made at length in risk of ruin and drawdown math.

What to do with this

Four things, in order.

1. Compute the average range for the instrument and timeframe you actually trade, and write the number down. Not "it's volatile" — a number, with a date.
2. Fix a multiple and leave it fixed. The multiple is a decision about how much noise you will tolerate. Renegotiating it per trade turns it into a way of justifying the size you already wanted.
3. Recompute the range on a schedule — weekly is enough for most people — and re-derive size when it moves materially. A stop distance that was correct last month is not automatically correct now.
4. Never use a low reading as a reason to be comfortable. It is a reason to expect a tighter stop, and therefore a larger position, and therefore more exposure to the thing the reading cannot see coming.

Volatility is a ruler. It measures the market. It does not measure you.