Historical volatility
How much did a market actually swing in the past, and why does the length of the look-back window change the answer?
In plain words
Historical volatility measures the size of a market's actual price swings over a chosen past window, usually converted to a yearly percentage so different windows and markets can be compared. It is also called realised volatility. Because the window is chosen, the same market can show very different figures depending on the period covered.
Why it matters
Volatility calibrates position size, stop distance and margin use. A quiet window can make a market look tame, so risk limits set from it may be too loose when swings widen. Short windows react quickly but jump around; long windows are steadier but hide recent change.
In a real week: ETH, week of 16 April 2018
For ETH in the week of 16 April 2018, realised volatility annualised from hourly prices was 82.4%. Over that week the price ranged 29.39%, from a low of 497.31 on Monday at 13:00 UTC to a high of 643.49 on Sunday at 17:00 UTC, with the largest one-hour move 3.02% on Friday at 22:00 UTC. Only 9 weeks of history were available, so any longer window could not yet be measured.
A common mistake
Quoting a volatility figure without its window. A weekly and a yearly figure for the same asset are different measurements, and a number built on very little history carries little weight.