Frequentist Statistics

Frequentist statistics defines probability as the long-run frequency of events in repeated experiments. A coin has a 50% probability of heads not because of any belief about this particular flip, but because heads appears half the time over thousands of flips. This framework produces the tools most founders encounter without knowing the philosophy underneath: p-values, confidence intervals, hypothesis testing. Its strength is objectivity — conclusions depend on data, not prior beliefs. Its weakness is that it answers a question most people don't realize they're asking: "If the null hypothesis were true, how surprising is this data?" rather than "Given this data, what should I believe?" Understanding the frequentist lens helps you interpret experimental results correctly and recognize when its assumptions break down — especially with small samples or one-off decisions where long-run frequency has no meaning.

Use frequentist methods where they're strong: large-sample experiments, A/B tests with high traffic, and quality-control processes with repeatable measurements. Recognize their limits for one-off strategic decisions where there's no population to sample from. When results depend on p-values, understand what the p-value actually claims — and what it doesn't. Pair frequentist analysis with Bayesian reasoning when you have strong prior information.

Frequentist statistics defines probability as long-run frequency and produces the hypothesis tests, p-values, and confidence intervals most businesses rely on. Use it for repeatable experiments with large samples; recognize its limits for unique, one-off decisions.