Drawdown, standard deviation and other risk measures

Risk measures turn the vague idea of "how bumpy is this investment" into numbers you can compare. No single measure captures everything, and each answers a slightly different question: how far did it fall, how widely do returns scatter, and how strongly does it follow the market. Knowing what each one counts, and what it ignores, keeps you from trusting a tidy number more than it deserves.

Drawdown

A drawdown is a decline from a peak to a later low, expressed as a percentage of the peak. The maximum drawdown is the largest such decline in the period you are looking at.

Drawdown = (trough value - peak value) / peak value
Gain needed to recover = 1 / (1 - loss) - 1

Drawdown is intuitive because it matches what investors feel: how much did I lose from the top? It also exposes the asymmetry of losses. A larger loss needs a disproportionately larger gain to get back to even:

Its limits: it depends entirely on the window you pick (a longer history almost always contains a bigger drawdown), it looks at only one path, and it says nothing about how long the recovery took unless you also track that duration.

Standard deviation (volatility)

Standard deviation measures how widely returns scatter around their average. A higher number means a wider range of typical outcomes. Return series are usually quoted on an annual basis. A common convention scales a monthly figure by the square root of 12, or a daily figure by the square root of 252, the usual count of trading days in a year.

Example (illustrative numbers)

A fund has a monthly standard deviation of 4%. Annualized, that is 4% x the square root of 12, or about 13.9%. If annual returns are roughly bell-shaped, about two-thirds of years would land within one standard deviation of the average, and about 95% within two. If the average annual return is 8%, a one-standard-deviation range would run from about -6% to +22%.

The bell-curve assumption is where this measure is weakest. Real stock returns have "fat tails": extreme days and months occur more often than a normal distribution implies. Standard deviation also treats a large gain as just as risky as a large loss, which is not how investors experience it. See volatility and the VIX for market-level volatility.

Downside deviation

Downside deviation fixes the second problem by counting only returns below a chosen target, often zero or the risk-free rate. It is the denominator in the Sortino ratio, a variation on the Sharpe ratio that penalizes only harmful volatility. It is a better match for how losses feel, but it needs enough data to estimate the lower tail, and it depends on the target you choose.

Beta

Beta measures how strongly a stock tends to move with the overall market. A beta above 1 means it has historically moved more than the market; below 1, less. It captures only market-related risk, not a company's own risks, and it changes with the period and index used. See beta explained.

Value at risk

Value at risk (VaR) estimates a loss threshold that should not be exceeded over a set period with a stated probability. For example, "one-day 95% VaR of $10,000" means that on 95% of days, losses are expected to be smaller than $10,000. What it does not tell you is how bad the other 5% of days can be. VaR is widely used by financial institutions, but it can understate risk when markets behave unlike the data used to build it.

What none of them capture

• They look backward. Past volatility and drawdowns describe what happened, not a limit on what can happen.
• They miss permanent loss. A stock can fall to zero, and a diversified index's history does not say that about any single company.
• They ignore liquidity. Holdings that are hard to sell can show smooth prices, then be hard to exit.
• They ignore your situation. A 30% drawdown is a different event for someone buying every month than for someone withdrawing. See risk tolerance and time horizon.
• They depend on correlation. Portfolio risk depends on how holdings move together, which is not stable. See correlation between stocks.

Where this shows up on StocksXray

The price chart on each stock page, for example Apple, lets you see past peak-to-trough declines directly. The site's valuation models draw betas and the equity risk premium from Damodaran datasets.

Key takeaways

• Drawdown measures peak-to-trough loss, and recovery needs a bigger percentage gain than the loss.
• Standard deviation measures scatter of returns, treating gains and losses alike.
• Downside deviation counts only returns below a target.
• Beta measures market sensitivity, not total risk; VaR gives a threshold, not the worst case.
• Use several measures together and treat all of them as descriptions of the past.