Skip to main content
A z-score expresses how unusual a current value is: the number of standard deviations it sits away from its historical average. A z-score of zero means the value equals the historical mean; positive means above average; negative means below.

Beginner

What It Means

Raw numbers lack context. Is a currency at 1.10 high or low? A z-score answers instantly: it restates the current level as a distance from that series’ own normal, measured in units of its own typical variability. A reading near zero is close to the historical mean; a reading of +2 or −2 is two standard deviations from it. How frequently such readings occur depends on the distribution and measurement window.

Example

Suppose a currency has averaged 1.05 over its history and now trades at 1.10. Whether that is a big move depends entirely on how much the currency usually swings. If its typical swing is half a cent, 1.10 sits ten standard deviations out and the reading is extreme. If its typical swing is five cents, the same level is one standard deviation out and entirely unremarkable.

Why It Matters

Z-scores put currencies, yields, and sentiment measures on a common scale of “how stretched is this?”, which makes readings comparable across completely different series, and turns “the currency is strong” into the more useful “the currency is unusually strong by its own history.”

Advanced

How to Read It

In the reports, large positive or negative z-scores identify levels far from their historical averages. Interpretation depends on the series: a high currency reading describes strength relative to the chosen window, while a high volatility reading describes elevated variability. Assessing mean reversion requires evidence that the series tends to return toward a stable reference level.

Common Misreadings

  • Treating an extreme z-score as a reversal signal: a z-score measures distance from a historical mean; reversal probabilities require a model of how the series behaves
  • Ignoring the lookback window: a z-score is always relative to the history it was measured against, the same level can look extreme against a short history and ordinary against a long one
  • Assuming symmetry of meaning: an extreme reading in a trending series (like a yield during a policy cycle) is less anomalous than the same reading in a range-bound series

In the Reports

The currency section leans on z-scores most heavily, placing the exchange rate against its own history, and the components of the composite market-sentiment score are each expressed as z-scores before being combined, which is what makes liquidity, momentum, and breadth comparable.

Standard Deviation

The unit a z-score is measured in

Mean Reversion

The pull back toward average

Composite Market-Sentiment Score

Built from z-scored components
Last modified on September 21, 2026