What this calculator does
The normal distribution is the bell-shaped curve that describes many measurements: exam marks, heights, manufacturing tolerances, and the averages of almost anything once the sample is large. It is defined by two numbers, the mean μ (the centre) and the standard deviation σ (the spread).
This calculator works in both directions. Give it values and it returns the probability of falling below, above, between or outside them. Give it a probability and it returns the cut-off value, such as the mark needed to be in the top 10%. Every result comes with z-scores, a shaded curve and the matching Excel formula.
The key formulas
P(X > x) = 1 − Φ(z) · P(a < X < b) = Φ(zb) − Φ(za)
x = μ + z·σ, where z = Φ⁻¹(p)
Φ is the cumulative distribution function of the standard normal: the area under the curve to the left of z. A printed z-table lists Φ for z values to two decimal places. This calculator evaluates it to full precision for any z, including extreme tails such as z = 8 where a table runs out.
Worked example: university exam marks
Press Load example. Marks in a large university statistics exam are roughly normal with a mean of 62 and a standard deviation of 12.
- What share scored above 80? z = (80 − 62) / 12 = 1.5, and P(X > 80) = 0.0668, so about 6.7% of students, or 1 in 15.
- What share scored between 50 and 74? That is μ ± 1σ, so 68.27%, the first part of the 68–95–99.7 rule.
- What percentile is a mark of 75? P(X < 75) = 0.8607, the 86th percentile.
- What mark puts a student in the top 10%? Switch to Find a value, choose Right tail and enter 10%: the cut-off is 77.38 (z = 1.2816).
Where z-scores show up in analytics
- Outlier detection: values with |z| > 3 are a common first flag in data cleaning.
- Confidence intervals: the 1.96 in “mean ± 1.96 × SE” is the z that leaves 2.5% in each tail. Choose Central with 95% to see it.
- Standardising features before clustering or regularised regression puts every variable on the same z scale.
- Hypothesis tests such as the A/B test z-test turn a difference into a z and read its p-value from this same curve.
Getting the same result in Excel and Python
The Excel & Python tab shows the exact formulas for your inputs. The main Excel functions
are NORM.DIST(x, μ, σ, TRUE) for P(X < x), NORM.INV(p, μ, σ) for the
inverse, and STANDARDIZE(x, μ, σ) for z. In Python:
from scipy.stats import norm
norm.sf(80, loc=62, scale=12) # P(X > 80) = 0.0668
norm.cdf(74, 62, 12) - norm.cdf(50, 62, 12) # P(50 < X < 74) = 0.6827
norm.isf(0.10, 62, 12) # top 10% cut-off = 77.38
Common mistakes
- Using 1 − Φ(z) when the question asks for “below”. Sketch the curve and shade the area first; the chart above does this for you.
- Mixing up variance and standard deviation. σ is the square root of the variance. If a question gives a variance of 144, σ = 12.
- Applying the normal curve to skewed data. It badly underestimates how often extreme values occur in incomes, claim sizes or delivery times.