From Coin Flips to the Bell Curve

By Victor Arce

Flip a fair coin a few times and count the heads. Any single flip is pure 50/50 chance — but the total count is far from random-looking. Some totals are much more likely than others, and as you add flips a familiar shape emerges: the bell curve. This is the gateway to the whole family of probability distributions, and to one of the deepest results in statistics.

Step up the number of flips below. Each bar is the probability of getting exactly that many heads; once there are a few flips, the smooth normal curve is drawn over the bars so you can watch the histogram grow into it.

What you're seeing

With n flips, the count of heads follows the binomial distribution: P(k heads) = C(n, k) / 2ⁿ.

The C(n, k) part counts the number of ways to get k heads out of n flips — exactly the rows of Pascal's triangle. There's only one way to get all tails (k = 0) or all heads (k = n), but many different orderings give a middling count, so the middle bars are tallest. The distribution is centered on the mean μ = n/2 and its spread is σ = √n / 2.

As n grows, the jagged bars settle into the smooth, symmetric normal distribution N(μ, σ²) — the classic bell curve. The match is no coincidence.

Why the bell curve is everywhere

The reason is the Central Limit Theorem: when you add up many independent random contributions, their sum tends toward a normal distribution — almost regardless of what each individual contribution looks like. A coin flip is the simplest possible random step (just 0 or 1), and summing n of them already pulls the shape toward the bell.

That universality is why the normal distribution shows up far beyond coin flips: measurement errors, heights, test scores, noise in signals. Whenever an outcome is the accumulation of many small independent effects, expect a bell. From here, the other distributions in this category — Bernoulli, binomial, Poisson, normal — are best understood by how they relate to this same story.

Back to Probability & Distributions