Normal Distribution — The Bell Curve Explained
Why everything clusters around the average
Data Clusters Around a Center
Heights, test scores, measurement errors — when many small independent factors add up, their sum always piles up near a central value. Watch the scattered points converge.
The Bell Curve Appears
Smoothing those points into a continuous curve gives the famous bell shape — the normal distribution: f(x) = (1/σ√2π) · e^(−(x−μ)²/(2σ²)).
Mean (μ) Sets the Center
The mean μ is the peak of the bell. Move the 'μ (mean)' slider and the whole curve slides along the fixed number line, while the faint dashed N(0, 1) outline stays put for comparison. The shape stays exactly the same — only the location changes.
Spread (σ) Controls the Width
Standard deviation σ controls how wide the bell is. Move the 'σ (std dev)' slider: small σ gives a tall, narrow peak and large σ a short, wide bell. The total area always stays exactly 1, so doubling the width halves the peak height.
Area Equals Probability
The shaded region between the two yellow handles is P(a ≤ X ≤ b) — the probability a random value lands in that range. Drag the handles or move the 'Left edge a' and 'Right edge b' sliders, then press 'Draw 100 samples': the share of random draws landing between a and b closes in on the shaded area.
68–95–99.7 Rule
68% of values fall within 1σ of the mean, 95% within 2σ, and 99.7% within 3σ. These three numbers appear everywhere in statistics. Press 'Probability challenge' and try to shade a target area such as 34.1% or 13.6% using only these landmarks.
How It Works — Step by Step
-
1
Data Clusters Around a Center
Heights, test scores, measurement errors — when many small independent factors add up, their sum always piles up near a central value. Watch the scattered points converge.
-
2
The Bell Curve Appears
Smoothing those points into a continuous curve gives the famous bell shape — the normal distribution: f(x) = (1/σ√2π) · e^(−(x−μ)²/(2σ²)).
-
3
Mean (μ) Sets the Center
The mean μ is the peak of the bell. Move the 'μ (mean)' slider and the whole curve slides along the fixed number line, while the faint dashed N(0, 1) outline stays put for comparison. The shape stays exactly the same — only the location changes.
-
4
Spread (σ) Controls the Width
Standard deviation σ controls how wide the bell is. Move the 'σ (std dev)' slider: small σ gives a tall, narrow peak and large σ a short, wide bell. The total area always stays exactly 1, so doubling the width halves the peak height.
-
5
Area Equals Probability
The shaded region between the two yellow handles is P(a ≤ X ≤ b) — the probability a random value lands in that range. Drag the handles or move the 'Left edge a' and 'Right edge b' sliders, then press 'Draw 100 samples': the share of random draws landing between a and b closes in on the shaded area.
-
6
68–95–99.7 Rule
68% of values fall within 1σ of the mean, 95% within 2σ, and 99.7% within 3σ. These three numbers appear everywhere in statistics. Press 'Probability challenge' and try to shade a target area such as 34.1% or 13.6% using only these landmarks.
Key Takeaways
Bell Shape Is Inevitable
Whenever many small independent factors combine, the result is approximately normal. This is the Central Limit Theorem — the deepest theorem in statistics.
68–95–99.7 Rule
68% of data falls within 1σ, 95% within 2σ, 99.7% within 3σ. These numbers let you reason about spread instantly, without a calculator.
Area = Probability
The total area under the bell is exactly 1 (100%). Any shaded region equals the probability of a value landing there. P(μ−σ ≤ X ≤ μ+σ) = 0.683.
Standardizing with Z
Z = (X − μ)/σ converts any normal N(μ, σ) to the standard normal N(0, 1). This lets you compare measurements from completely different distributions.
Connected Concepts
Central Limit Theorem
Where bell curves come from — add enough random things and you always get a normal distribution.
Binomial Distribution
Flip a coin N times — for large N the result looks exactly like a bell curve.
Z-Score
Z = (X − μ)/σ converts any measurement into standard deviation units.
Standard Deviation
σ is the bell’s width — see how it is computed from real data points.
Hypothesis Testing
Tail areas under the bell become p-values: how surprising is your result?
Linear Regression
Residuals in a good regression model are normally distributed — the bell curve at work.
Correlation
Plot two normal variables against each other and the cloud is an ellipse whose tilt and thinness are set by r.
Updated
- — Added a link to Correlation, showing how two normal variables form a tilted cloud.
- — The bell now moves on a fixed number line; added edge sliders, random-sample histogram and a probability challenge