# Volatility

> Volatility measures how much your portfolio moves around its own average. It is the raw material for most of the other risk measures on the Performance tab.

> These figures are derived from a return series that the app currently flags as being reworked.
> Values and holdings are unaffected. See [about these numbers](/metrics/about-these-numbers).

Volatility is how much your portfolio bounces around, expressed as a single
annual percentage. A portfolio at 8% volatility has drifted; one at 40% has
lurched.

It is the base ingredient for most of the rest of the risk grid. Sharpe divides
by it, Sortino replaces it with a one sided version, tracking error applies it
to your gap against a benchmark. Understanding this one makes the others
straightforward.

## How it is calculated

Gylder takes the daily percentage change in your portfolio's value, computes
the standard deviation of those changes, and scales it to a year:

```
volatility = standard deviation of daily returns × √252
```

The 252 is the number of trading days in a year. Multiplying by its square
root is the standard way of turning a daily figure into an annual one, and it
is why an 8% annual volatility corresponds to daily moves far smaller than 8%.

## How to read it

Higher means a wider spread of outcomes, in both directions. That last part is
the one people forget: volatility is symmetric, and a portfolio that jumps
sharply upward is measured as exactly as volatile as one that falls just as
sharply.

Gylder does not label a volatility figure as high or low. What is reasonable
depends entirely on what you hold. A portfolio of short dated government bonds
and one of small cap growth stocks are not on the same scale, and a single
threshold across both would be meaningless.

The comparison that does work is against yourself, over the same length of
window, before and after a change in what you hold.

## What it misses

Volatility says nothing about direction. Two portfolios can post identical
volatility while one doubled and the other halved.

It also assumes the spread of returns is reasonably well behaved, which real
portfolios are not. Concentrated positions and thinly traded assets both tend
to produce a figure that looks calmer than the position actually is, because
the price simply did not update on the days it would have moved most.

For the downside only version, see [Sortino ratio](/metrics/sortino-ratio).
For the worst actual loss rather than the average spread, see
[maximum drawdown](/metrics/max-drawdown).

## A note on precision

Deep in history, Gylder stores your value series weekly rather than daily to
keep the dashboard fast. Volatility computed across a long window is therefore
an approximation, and a slightly smoothed one, since weekly sampling cannot see
the daily moves inside each week. It is honest for reading your portfolio. It
is not a compliance grade measurement.