The Normal Distribution

Why does the bell curve appear everywhere?

Measure the heights of a thousand people. Plot them. You get a symmetric mound — most people near the middle, fewer toward the extremes. The same shape appears in test scores, measurement errors, and blood pressure readings.

This is not a coincidence. It tends to emerge whenever a quantity is the sum of many small, independent influences — a pattern the central limit theorem makes precise.

The intuition behind the shape

When a quantity is the result of many independent random factors added together, the extreme outcomes become increasingly rare. For height: you need every genetic and environmental factor to push in the same direction to produce an extreme value. The probability of that compounds quickly toward zero.

Most people end up in the middle because that's where the countless small factors roughly cancel out.

Variance explorer

50.6

Mean

85.6

Variance

9.25

Std dev

Spread (std dev)10

230

Outliers0

04

Adjust the Spread slider. You're changing the standard deviation — how tightly the values cluster. Notice how the shape stays symmetric but becomes wider or narrower.

Reading the bell curve

A normal distribution is fully described by two numbers: mean (where it's centered) and standard deviation (how spread out it is).

The 68-95-99.7 rule is the key fact to internalize:

• About 68% of values fall within 1 standard deviation of the mean
• About 95% fall within 2 standard deviations
• About 99.7% fall within 3 standard deviations

The formal definition

The probability density function of the normal distribution is:

$f(x) = \frac{1}{\sigma\sqrt{2\pi}} \exp\!\left(-\frac{(x-\mu)^2}{2\sigma^2}\right)$

Where $\mu$ is the mean and $\sigma$ is the standard deviation. The $\exp$ term says: the further you are from the mean, the probability drops off exponentially.

Why it matters in practice

• Quality control: manufacturing tolerances assume normal variation. A process is "in control" if output stays within ±3σ.
• Statistics: many statistical tests assume normally distributed data (or use the central limit theorem to justify this assumption).
• Finance: asset returns are often modeled as approximately normal — though real returns have "fat tails" that the normal underestimates.
• Grading: when people say grades are "curved," they often mean scaled to fit a normal distribution.

What the normal distribution is not

Not everything is normal. Income is heavily right-skewed. City populations follow a power law. Heights of adult men and women together are bimodal.

The normal distribution is a model, not a law of nature. Use it when the generating process involves many small independent additive effects. Be skeptical of it everywhere else.