Z-score (standardized deviation)
Education only · our voice · free public data
Definition
A statistical normalization expressing how many standard deviations a current reading sits above or below its recent mean (e.g., over a trailing 90 days). It makes different indicators comparable on one scale.
How to read it
A z-score is computed as (current value minus mean) divided by standard deviation over a lookback window. It converts any raw indicator into a unit-free measure of 'how unusual is this right now.' A z of 0 is average; +2 means two standard deviations above the mean (statistically stretched high); -2 means stretched low. The '90d' tag means the mean and standard deviation use the trailing 90 days, so the score reflects recent-regime extremity rather than all-time levels. Z-scores let you compare a stretched VIX to a stretched sentiment model on the same footing and are the backbone of extremes-based, contrarian reads.
How practitioners use it
Used as context among multiple indicators — never as a standalone signal to act.
Less common professional uses
Financial data is fat-tailed and non-stationary, so a z of +3 is not as rare as the normal distribution implies — treat thresholds as heuristics, not probabilities. Window choice drives the signal: a rolling 90-day z resets its baseline as regimes shift, so a persistent trend can keep 'resetting to zero' and mask a genuine extreme — cross-check with a longer window or a percentile. Mean/variance are themselves noisy in short windows; a low-volatility stretch shrinks the denominator and manufactures large z-scores from small moves. Pair z-score (distance from mean) with percentile rank (position in distribution) — they disagree in skewed series, and the disagreement is informative.
Sources & provenance
Standard statistical normalization (educational overview); Portal desk education notes
This page is educational content published by Pachira Aquatica Global LLC. It is not investment advice and not a recommendation.