What is volatility?
How much a price moves. Not which way it moves - and that distinction does more work than it first appears to.
The definition
Volatility is the standard deviation of returns over a period. It answers: how far do this asset's returns typically land from their own average?
Because returns are squared in the standard deviation, a rise and a fall of equal size contribute identically. Volatility has no sign.
The consequence of having no direction
This is the property that trips people up most often. An asset that rose 20 % and one that fell 20 % over the same window, with similar paths, have approximately the same volatility.
So volatility on its own can never tell you whether something good or bad happened. It tells you something eventful happened. Any directional reading has to come from a different measurement.
SDE compares sentiment against a stability term derived from volatility. Because volatility is unsigned, SDE cannot distinguish enthusiasm during a rally from enthusiasm during a crash — both look like "high sentiment, high volatility". Readers who assume SDE compares sentiment against price direction will misread it, which is why that page opens with the point.
The window changes the answer
Volatility is not a property of an asset. It is a property of an asset over a chosen period, and the choice is not neutral:
- Short windows react quickly and are noisy. A single unusual day dominates a 7-day figure.
- Long windows are stable and slow. A 90-day figure can still be elevated by an event two months past.
Neither is correct in general. When comparing volatility figures from different sources, the window is the first thing to check — it explains most discrepancies.
Realised, implied and the difference
| Type | Derived from | Describes |
|---|---|---|
| Realised | Historical prices | What already happened |
| Implied | Option prices | What the market is pricing for the future |
Moonboard uses realised volatility. There is no implied volatility in any metric on the platform, and no forward-looking volatility estimate. Everything is backward-looking by construction.
Units: the common mistake
In Moonboard's data model, volatility is stored as a fraction, not a percentage:
Other fields in the same record — sentiment, marketDominance, socialDominance — are already percentages. Several formulas multiply volatility by 100 before use for exactly that reason, and mixing the conventions is the most common way to reproduce a Moonboard number incorrectly.
How Moonboard uses volatility
| Metric | Use |
|---|---|
| Moonboard Score | Banded into points. 2–5 % scores highest; both calmer and wilder score lower. |
| OPS | Inverted into a stability term at 10 % weight. |
| SDE | Inverted into stability and subtracted from sentiment. |
| COR | Inverted, then standardised as one of five inputs. |
Three of the four invert it, so lower volatility scores better. The Moonboard Score is the exception: it prefers a middle band on the reasoning that a completely flat asset is not interesting either. That is a stated preference, not an empirical finding.
Volatility and liquidity
The two are related and not the same. Illiquid assets tend to be volatile, because a given order moves the price further. But a deep, liquid market can be volatile when information genuinely arrives, and a dormant asset can be calm simply because nobody is trading it.
Low volatility is therefore not automatically good — it can indicate stability or abandonment. Reading it alongside turnover separates the two cases.
Related
- Market liquidityThe related concept
- SDE methodologyWhere the unsigned property matters most
- Liquidity researchHow the two interact in practice
- Moonboard ScoreThe volatility bands in full