What is the normal distribution?

The normal distribution (also called the Gaussian distribution) is the famous “bell curve.” It’s used to model many real-world quantities that are the result of lots of small, independent effects adding together. Think: measurement noise, manufacturing variation, exam scores (sometimes), or the distribution of sample means via the central limit theorem.

A normal distribution is defined by just two parameters: the mean (μ), which sets the center of the bell, and the standard deviation (σ), which sets the spread. Bigger σ means a wider bell; smaller σ means a tighter, narrower bell.

The bell curve itself is the PDF. For a normal distribution, the PDF at a value x is:

PDF: f(x) = (1 / (σ√(2π))) · exp( - (x - μ)² / (2σ²) )

Important: the PDF value is not a probability by itself. It’s a density — a “height” of the curve. The probability that X falls in an interval (like between a and b) is the area under the curve between those points.

The CDF answers “what fraction of values are ≤ x?”: CDF: F(x) = P(X ≤ x). If you’re grading a test and you want to know what percent scored below 72, that’s a CDF question.

For the normal distribution, the CDF has no simple elementary closed form, but it’s computed accurately using the error function (erf) or high-precision approximations. This calculator uses stable approximations that are widely used in practice.

A z-score converts a raw value x into “how many standard deviations from the mean”:

z = (x - μ) / σ

If z = 0, you’re exactly at the mean. If z = 1, you’re one standard deviation above the mean. Negative z means below the mean. Z-scores are powerful because they let you convert any normal distribution into the standard normal distribution N(0,1). Once you’re in z-space, you can use the same CDF logic everywhere.

A super common question is “what’s the probability that X is between a and b?” For any continuous distribution: P(a ≤ X ≤ b) = F(b) - F(a). For a normal distribution, it’s the same idea: compute the CDF at b and subtract the CDF at a.

A percentile is the inverse problem. Instead of asking “what probability is ≤ x?” you ask “what x value has probability p to the left?” In symbols, if p is between 0 and 1, the percentile cutoff is: x = F⁻¹(p).

This is how you find “top 10% cutoff” (p = 0.90), “bottom 1% cutoff” (p = 0.01), or the median (p = 0.50). In this calculator, you can set a percentile (0–100) and it will compute the cutoff x that corresponds to that percentile for your chosen μ and σ.

The normal distribution shows up in pop culture and everyday life more than people realize: “What’s the cutoff for the top 5%?”, “How rare is a score like this?”, “If the average is 70 with σ=10, what are my odds?” This page is designed for quick screenshots: you get a clean result grid plus a visual bell curve with markers for x and the interval [a, b]. Use it in study groups, tutoring sessions, or just to settle arguments in group chats.

Examples you can copy

Here are practical examples you can reproduce in the calculator. Try them and compare your intuition to the numbers.

Suppose exam scores are approximately normal with μ = 70 and σ = 10. What percentage scored below 85? Compute z = (85 − 70) / 10 = 1.5. The CDF at z = 1.5 is about 0.933, so roughly 93.3% scored below 85. That means 85 is around the 93rd percentile.

Using μ = 70 and σ = 10 again, what percentage scored between 60 and 80? That’s z values of -1 and +1. The area between -1 and +1 is about 0.6827, so roughly 68.27%. This matches the 68–95–99.7 rule.

With μ = 70 and σ = 10, what score is the cutoff for the top 5%? “Top 5%” means 95th percentile. Set percentile to 95 and you’ll get x ≈ 70 + 10·1.645 = 86.45.

A machine produces parts with a mean diameter of μ = 50 mm and σ = 0.2 mm. What fraction are within 49.8 and 50.2? That is μ ± 1σ, so about 68%. Tightening tolerance to μ ± 0.4 (2σ) would include about 95%.

• If your probability between a and b is near 0, your interval is very far in the tails (rare event).
• If it’s near 1, your interval covers most of the bell.
• If z is large (like 3 or -3), you’re deep in the tails and probabilities get tiny fast.

Behind-the-scenes math (in plain English)

This calculator does three core things:

• Standardize (convert to z) with z = (x − μ)/σ.
• Compute CDF using the error function relationship: Φ(z) = 0.5 · (1 + erf(z/√2)).
• Invert the CDF to get percentiles using a well-known rational approximation (fast and accurate for typical use).

Accuracy note: percentiles near 0.1% or 99.9% can be sensitive (small changes in p produce big changes in x). That’s not a bug — it’s a property of the tails. For most coursework and day-to-day decision-making, these approximations are more than sufficient.

• σ must be positive. If σ ≤ 0, the distribution isn’t valid.
• PDF ≠ probability. Probability comes from area under the curve, not the curve’s height.
• Between uses CDF difference. P(a ≤ X ≤ b) = F(b) − F(a).
• Top X% = (100 − X)th percentile. Example: top 5% cutoff is the 95th percentile.

Frequently Asked Questions

• Is the normal distribution always a good model?

  No. It’s a useful approximation when values cluster around a mean with symmetric variation. But many real datasets are skewed, heavy-tailed, or multimodal. Use the normal distribution when it fits the context (or when CLT-style approximations apply).

• What does “CDF at x” mean in simple terms?

  It’s the probability of getting a value less than or equal to x. Visually, it’s the area under the bell curve to the left of x.

• Why is the PDF sometimes greater than 1?

  Because the PDF is a density, not a probability. For tight distributions (small σ), the curve can be tall while still having total area 1.

• How do I find the cutoff for the top 10%?

  Set percentile to 90. The calculator returns the x value where 90% of outcomes are below it — so only 10% are above it.

• Does “between” include endpoints?

  For continuous distributions like the normal, it doesn’t matter. P(X=a) is 0, so including or excluding endpoints changes nothing.

• What if my range [a,b] is reversed?

  If a > b, the calculator will swap them automatically so you still get the correct probability.