Why statistics matters

Most important questions can't be answered with certainty: Does this drug work? Will this model generalize? Is this difference real or luck? Statistics is the discipline of drawing honest, defensible conclusions from data that never tells the whole story.

Build the intuition

The two halves: probability and statistics

Probability reasons forward: given a fair coin, how often will it land heads? Statistics reasons backward: given the heads we observed, is the coin fair? Probability is the theory of randomness; statistics is the practice of learning from data through it. This course builds both, intuition first.

Why intuition fails at uncertainty

Humans see patterns in noise, trust small samples, and ignore base rates. We're sure a hot streak will continue and stunned when a 99%-accurate test is usually wrong. Statistics is the corrective lens: a set of habits that keep randomness from fooling you, and keep you from fooling yourself.

What good statistical thinking buys you

It separates signal from noise, quantifies how sure you're allowed to be, and reveals when data simply can't answer a question. In an age where decisions are increasingly made by models trained on data, statistical literacy isn't optional — it's the difference between using evidence and being used by it.

See it move

InteractiveThe law of large numbers

Flips40

After 40 flips: 23 heads = 57.5%. Single flips are pure chance — but the running average is drawn, inevitably, toward 50%. Randomness has long-run structure.

Statistics in one picture: a single flip is unknowable, but the long-run average is lawful. The whole field lives in the gap between the chaos and the order.

A worked example

Same number, opposite conclusions

A new treatment cured 7 of 10 patients (70%). Impressive?
With only 10 patients, 70% could easily arise from a useless treatment plus luck — the uncertainty is enormous.
But 7,000 of 10,000? The same 70% is now rock-solid. The percentage didn't change; the evidence did.
Statistics is what tells those two identical-looking 70%s apart — and most of this course is learning how.

Out in the world

Replication crisis

Whole fields have discovered that famous results don't reproduce — often because of statistical missteps: tiny samples, p-hacking, ignored uncertainty. Understanding statistics isn't just academic; it's why some 'findings' deserve trust and others don't.

Common confusion, cleared

“Statistics is just computing averages and making charts.”

Those are tools. The discipline is reasoning: quantifying uncertainty, designing fair comparisons, and knowing the limits of what data can say.

“With enough data, statistics gives certainty.”

It gives calibrated uncertainty — ranges and probabilities, not guarantees. Honest 'we're 95% confident it's between X and Y' beats false certainty every time.

Recap

Probability reasons forward from model to data; statistics reasons backward.
Human intuition is unreliable under uncertainty — statistics is the corrective.
The goal is calibrated honesty: conclusions with their uncertainty attached.

Progress saves in this browser.