Why Averages Always Make a Bell
The bell shape comes from adding independent wiggles together, not from the numbers themselves being bell-shaped.
A complete interactive classroom, not just a preview.
Start when you are ready to enter this Stage's 8 scenes and explore, respond, and learn as you go.
Why does the distribution of the average of many random numbers always become bell-shaped, regardless of where the numbers come from?
Roll a single die and you get a flat spread from 1 to 6. Add two dice together and a bump appears in the middle. Keep averaging more numbers and the bump sharpens into a perfect bell — but why?
No individual roll looks bell-shaped; the bell only emerges from the combination. Something invisible must be shaping the result.
A live simulation that lets learners roll 1, 2, 5, or 50 dice, watch the histogram reshape with each setting, and see the bell curve snap into view as the sample size grows.
Every average is a sum of small, symmetric wobbles — and the Central Limit Theorem guarantees those wobbles stack into a bell.
The bell curve only appears when the underlying numbers are already normal, so averages must inherit that shape from the source.
- Formal proofs of the Central Limit Theorem
- Advanced statistics like confidence intervals or hypothesis testing
- Heavy-tailed or pathological distributions where the theorem fails
- 01Roll One Die. Now Roll Two. Now Roll Fifty.slideQuestion
Pose the driving question with a concrete contrast: a single die gives a flat 1–6 spread, but the average of many dice looks smooth and bell-shaped. Ask which is doing the shaping — the dice, or the averaging.
- A single die gives a flat, uniform spread.
- The average of many dice looks bell-shaped.
- Question: where does that bell come from?
- 02Where Does the Bell Come From?quizPrediction
Force one committed guess before any evidence, so the learner notices whether their initial intuition holds.
- Make a single, committed choice about the source of the bell shape.
- 03Average Any Source, Watch a Bell EmergeinteractiveEvidence
Let learners pick a wildly non-bell source (uniform, skewed, bimodal, or a die) and dial the sample size from 1 up to 50. Watch the histogram transform live and see the bell snap into focus as n grows.
- Try uniform, skewed, and bimodal sources.
- Increase the number of values being averaged.
- Watch the spread narrow and the shape round out.
- 04Build the Bell One Wiggle at a TimeinteractiveEvidence
Show the bell emerging additively: each new value added to the sum nudges the result up or down with a small, roughly symmetric wiggle. Watch those wiggles stack into a bell-shaped distribution.
- Each added value contributes a small up-or-down wiggle.
- The wiggles are roughly symmetric around zero.
- Stacking many symmetric wiggles produces a bell.
- 05The Central Limit Theorem, IntuitivelyslideExplanation
Explain why the theorem works in plain language: each value added is an independent tug, positive and negative tugs roughly cancel, extreme totals are built from many tugs all pointing the same way (unlikely), and middle totals have many ways to happen.
- Each value is an independent up-or-down tug.
- Extreme totals need many aligned tugs — rare.
- Middle totals have many possible combinations — common.
- This is the Central Limit Theorem in action.
- 06Heights, Test Scores, Measurement NoiseslideTransfer
Apply the idea to real-world examples: adult heights, SAT scores, and measurement error all look bell-shaped because each is the average of many small genetic, environmental, or random influences.
- Height = average of many genetic and nutritional factors.
- Test scores = average of many small skill components.
- Measurement error = average of many tiny disturbances.
- 07When the Bell Fails to AppearslideBoundary
Show the limits: if the underlying values have infinite variance (e.g., a Cauchy distribution) or strong correlations, the bell never forms. This keeps the claim honest.
- Heavy-tailed sources can break the theorem.
- Correlated values break independence.
- The theorem has real preconditions.
- 08The Bell Lives in the Averaging, Not the NumbersslideResolution
Close the loop with the direct answer: the bell shape comes from adding many small independent contributions, not from the source distribution. That is why averaging almost anything produces a bell.
- The shape is inherited from the act of averaging.
- Central Limit Theorem guarantees this for any reasonable source.
- This is why bell curves are everywhere in nature.
Discussion threads for a Stage aren't available yet.
