This article explains volatility as a concept and how it is measured. It is educational and general, not personalised investment advice, and nothing here is a recommendation to buy, sell, size, or hedge any position. Where it describes statistical behaviour, that is a widely-observed tendency, not a precise claim about any single asset or period.
Volatility is the most quoted and most misread number in markets. People treat it as a synonym for risk, for danger, for losing money, when it is really just a measure of how much a price bounces around. It says nothing about which way. Here is what volatility actually is, exactly how it is calculated from returns and annualized, the difference between the volatility that already happened and the volatility options are pricing in, why the upside and downside halves are not the same thing, and why a high number does not automatically mean a risky asset.
The short version
Volatility is the dispersion of an asset’s returns: how widely its period-to-period returns spread around their average. It is usually measured as the standard deviation of those returns, then annualized by multiplying by the square root of the number of periods in a year. Volatility measures the magnitude of price moves, not their direction, and a high reading is not the same thing as high risk.
What volatility is (and is not)
Start with the thing being measured. Volatility is not computed on prices; it is computed on returns, the percentage change from one period to the next. A stock that grinds from 100 to 200 in a smooth straight line has huge price appreciation but low volatility, because each day’s return is small and similar. A stock that ends the year exactly where it started, having lurched violently up and down the whole way, has zero price change and very high volatility. Volatility is about the jaggedness of the path, not the destination.
The single most important thing volatility does not tell you is direction. It is a symmetric, sign-blind measure: a return of plus five percent and a return of minus five percent contribute identically to it, because the standard deviation squares the deviations and throws the sign away. That is why calling volatility “risk” is loose. It quantifies how far returns scatter, up or down together, and leaves entirely open the question of whether that scatter is mostly gains or mostly losses. A number that cannot distinguish a rally from a crash is a measure of turbulence, not of harm.
How it is actually measured
The standard recipe is deliberately mechanical. Take a series of returns over some period, daily returns are the common choice, and compute their standard deviation: the average distance of each return from the mean return, in the root-mean-square sense. That raw number is the volatility per period (per day, if you used daily returns), and on its own it is awkward to compare across assets or horizons.
So it is annualized. Because the variance of independent returns adds up over time, standard deviation scales with the square root of time, not with time itself. To turn a daily standard deviation into an annual one you multiply by the square root of the number of trading periods in a year (with roughly two hundred and fifty-two trading days, that factor is the square root of about two hundred and fifty-two). This “root-time” scaling is why annualized volatility figures are the ones you see quoted, and why the same underlying jitter can look small daily and large annually. The rule assumes returns are roughly independent from one period to the next, which is an approximation, real returns cluster and trend, so the annualized figure is a useful convention, not a law of nature.
One consequence worth internalising: volatility and compounding interact. Because a loss needs a larger gain to undo it, a jagged path compounds to less than a smooth one with the same average return. That gap is called volatility drag, and it is the reason two assets with the same headline average return can end up in very different places depending on how bumpy the ride was.
Realized versus implied volatility
There are two volatilities, and confusing them is a classic error. Realized (or historical) volatility is backward-looking: it is the standard deviation of returns that have already happened, computed from the actual price series over some trailing window. It is a fact about the past. Implied volatility is forward-looking: it is the volatility backed out of current option prices, the level of future turbulence the options market is effectively pricing in. It is an expectation about the future, and it can be wrong.
The two differ in kind, not just in timing. Realized volatility is measured; implied volatility is inferred from what people are willing to pay for optionality. Implied usually sits a little above recent realized (the gap is roughly the premium sellers charge for bearing uncertainty), and it tends to spike ahead of known events and during panics, when demand for protection surges. The best-known single number here is the VIX, an index of the implied volatility of the broad US market, often nicknamed the “fear gauge” because it rises when investors scramble to hedge. It is an implied-volatility reading, an expectation, not a record of what the market actually did.
Realized versus implied volatility, at a glance:
| Aspect | Realized (historical) | Implied |
|---|---|---|
| Direction in time | Backward-looking | Forward-looking |
| Source | Standard deviation of past returns | Backed out of current option prices |
| What it is | A measured fact about what happened | An expectation the market is pricing in |
| Can it be wrong? | No, it is what occurred | Yes, it is a forecast |
| Typical behaviour | Clusters: calm follows calm, storms follow storms | Spikes before events and in panics; usually above recent realized |
| Familiar example | A stock’s trailing annualized standard deviation | The VIX for the broad US market |
Upside volatility is not downside volatility
Because plain volatility is sign-blind, it lumps together two things an investor experiences very differently: the turbulence of a sharp run up and the turbulence of a sharp fall. To most people, a violent up-week and a violent down-week are not equally unwelcome, one is the whole point and the other is the fear. Yet standard deviation scores them the same.
This is why practitioners separate the two. Downside volatility (or downside deviation) measures dispersion using only the returns below some reference, ignoring the upside swings entirely, and it is the basis of risk-adjusted measures like the Sortino ratio, which, unlike the more familiar Sharpe ratio, refuses to penalise an asset for being volatile in the right direction. The distinction matters because a strategy can carry a high headline volatility that is overwhelmingly composed of upside surprises. Judged by total volatility it looks wild; judged by downside volatility it may be far tamer than its reputation. Which of the two you measure changes the story completely.
Why high volatility is not automatically high risk
The reflex to equate volatility with risk is understandable but wrong in an important way. Risk, in the sense that matters to a long-term investor, is the chance of a permanent, unrecoverable loss of capital, or of being forced to sell at the worst possible moment. Volatility is neither of those. It is a description of how much prices wobble along the way.
The two come apart in both directions. A quiet, slowly-declining asset can be low-volatility and genuinely dangerous, steadily destroying value with barely a tremor. A jumpy asset that swings hard but trends upward can be high-volatility and, for an investor who does not have to sell into the dips, not especially risky in the outcome that counts. Volatility only becomes risk under specific conditions: when it is skewed to the downside, when leverage turns a drawdown into a margin call, or when a fixed horizon forces a sale during a trough. Absent those, a big volatility number is a statement about the ride, not a verdict on the destination. This is also why position sizing is done on volatility rather than on a fixed dollar amount: sizing to how much a name typically moves is how a disciplined process keeps any single position’s wobble from dominating, the logic laid out in risk per trade.
How a systematic process treats volatility
A concentrated momentum book is, by its nature, a high-volatility thing to hold: it deliberately crowds into a small number of the strongest-trending names, and strong trends are precisely the names that move a lot. The useful question is not whether that volatility is high (it is) but which kind it is. Shishin’s published /track-record analysis is framed around exactly this distinction: the strategy’s volatility is high but mostly upside, the scatter is dominated by outsized winners rather than symmetric two-way churn, which is what a downside-aware read of the record is meant to show.
That framing also explains the “smoothed” overlay. The design clips the most violent up-days rather than the down-days: trimming into the sharpest upside spikes lowers the headline volatility number without trying to dodge losses, because the volatility being shed is the benign kind. The point is descriptive, not advisory: it is a worked example of treating upside and downside volatility as different quantities, and of judging a record by the shape of its dispersion rather than by a single sign-blind figure. The full breakdown lives in the public record at the track-record analysis.
So, what is volatility?
Volatility is the dispersion of returns, measured as their standard deviation and annualized by root-time, a clean number for how much a price moves and a deliberately silent one about which way. Realized volatility records what already happened; implied volatility is what options are pricing for the future, and the two are not interchangeable. Most of the misuse comes from three collapses: treating volatility as direction, treating upside swings as equivalent to downside ones, and treating a high number as proof of risk. Kept distinct, volatility is one of the most useful descriptions in finance. Blurred together, it is one of the most misleading.
Sources & further reading
- Hull, J. C. Options, Futures, and Other Derivatives, the standard reference for realized and implied volatility and root-time scaling.
- Sortino, F. A. & van der Meer, R. (1991). “Downside Risk.” Journal of Portfolio Management, 17(4), 27 to 31., the case for measuring dispersion below a reference rather than symmetrically.
- On how volatility erodes compound growth, see compounding and volatility drag; on the market’s implied-volatility “fear gauge,” the VIX; and on sizing to how much a name moves, risk per trade.
- For how a high-but-mostly-upside volatility profile reads in a real record, see Shishin’s track-record analysis.