A z-score answers "how unusual is this?" in the universal currency of statistics: standard deviations from the mean. It strips away units and scales, which is exactly what makes scores from different tests comparable.
Reading a z-score
An IQ of 130 against mean 100, SD 15: z = (130 − 100)/15 = +2.0 — two SDs up, the 97.7th percentile, with only 2.3% of the population beyond. Negative z means below the mean: 85 on the same scale is z = −1, the 16th percentile. The red marker on the bell curve shows your position live.
The 68–95–99.7 mental map
Normal data obeys the empirical rule: ~68% of values within 1 SD, ~95% within 2, ~99.7% within 3. It converts any z-score into instant intuition — |z| under 1 is everyday, past 2 is genuinely uncommon, past 3 is a headline. The calculator's verdict line applies the rule for you.
Comparing across different scales
Is SAT 1350 better than ACT 30? Convert both: (1350−1050)/200 = 1.5 and (30−21)/6 = 1.5 — a dead tie in relative performance. The same trick standardizes heights vs weights, this year's exam vs last year's, anything with a mean and SD. Computing those two ingredients from raw data is the standard deviation calculator's job.
When not to trust it
Percentiles from z assume roughly bell-shaped data. Incomes, city sizes and wait times are skewed — their means are dragged by tails, and z-percentiles mislead; rank-based percentiles serve better. Check mean-vs-median with the average calculator: a big gap is the warning sign. Counting-based probability, the other half of the story, lives in the probability calculator.
Normal CDF via Abramowitz–Stegun approximation (error < 1.5 × 10⁻⁷).