Normal Distribution Calculator: Curve, Area, and Inverse

Draw a normal distribution curve and read the area below, above, between, or outside your values, with an inverse mode that finds x from a probability or percentile.

At a glance

Computes
Normal-curve area below, above, between, or outside your values, and the inverse lookup.
You supply
Mean, standard deviation, and either a value x or a probability p.
Use when
Your quantity is normally distributed and you need a probability or a cutoff.
Not for
Small samples with an unknown population SD. Critical Value Calculator
What to find

Standard deviation (not variance), must be positive.

Region

The value to measure to.

Probability density; N(0, 1)

−4−2024N(0, 1)

Probability

Enter a value to see its area

Set μ, σ, and a bound, then calculate.

Supporting detail
DetailValue
Export

Φ(z) = 1/2 · [1 + erf(z / √2)], z = (x - μ) / σ, x = μ + σ · Φ⁻¹(p) How?

How this is calculated

Areas come from the cumulative distribution function Φ evaluated on the z-score z = (x - μ) / σ. Below is Φ(z), above is 1 - Φ(z), between is Φ(z_b) - Φ(z_a), and outside is 1 - [Φ(z_b) - Φ(z_a)]. The inverse solves x = μ + σ · Φ⁻¹(p) for a left-tail probability p in (0, 1).

Tail areas use the complementary incomplete-gamma form Q(1/2, z²/2), equivalent to erfc, so a small finite tail is calculated directly instead of by subtracting a rounded CDF from 1. The display keeps the page's conservative 1e-7 probability floor: values below it use a strict bound while the export records the numerical estimate and its precision state. The inverse uses Acklam's rational approximation (about 1e-9). To summarize a data set before assuming normality, start from descriptive statistics.

The chart requires a finite, ordered actual-x range. Rescale values that exceed drawable precision.

Everything on this page is continuous. A count is different. For successes in a fixed number of trials, the binomial distribution page computes each exact count probability. For events in an interval at a known rate, the Poisson distribution page does. Both print a normal approximation beside the exact answer and state the convention used to judge it.

Formula: Φ(z) = 1/2 · [1 + erf(z / √2)], z = (x - μ) / σ, x = μ + σ · Φ⁻¹(p)

The empirical rule (68-95-99.7)

For a normal distribution, the between mode returns the areas the empirical rule names: about 68.2689% of the area lies within ±1σ of the mean, 95.4499% within ±2σ, and 99.7300% within ±3σ. These are the exact values this tool computes for N(0, 1) between -1 and 1, -2 and 2, and -3 and 3.

Sources

  1. NIST SEMATECH e-Handbook, Normal distribution (1.3.6.6.1). NIST. Retrieved .
  2. Abramowitz and Stegun, Handbook of Mathematical Functions, 7.1.26 (p. 299). National Bureau of Standards. Retrieved .
  3. Peter Acklam, An algorithm for computing the inverse normal cumulative distribution function. Peter John Acklam. Retrieved .