Definition
A base rate is the unconditional, background frequency with which an event, trait, or category occurs in a defined population. It is the number that answers the question "how common is this in general?" before any case-specific information is brought to bear. When five percent of a population carries a disease, five percent is the base rate; when one in two hundred drivers passing a checkpoint is over the legal alcohol limit, that proportion is the base rate.
Base rates are the foundation of statistical reasoning under uncertainty. Any conditional probability — the chance of being ill given a positive test, the chance of guilt given a matching fingerprint — depends jointly on the strength of the new evidence and the base rate that preceded it. Ignore the base rate and you will systematically overweight the evidence.
Why it matters
How it works
Combining a base rate with new evidence requires holding two numbers in mind at once. Imagine a disease that affects one person in a thousand and a test that is ninety-nine percent accurate in both directions. Out of ten thousand people screened, ten will truly have the disease and the test will correctly flag roughly ten of them; the remaining nine thousand nine hundred and ninety are healthy, and the test will wrongly flag about one hundred of them. A positive result therefore corresponds to about a ten-out-of-one-hundred-and-ten chance of true disease — roughly nine percent, not the ninety-nine percent the test accuracy alone might suggest.
The discrepancy comes entirely from the base rate. When the underlying condition is rare, even a highly accurate test produces more false positives than true positives in absolute terms. The correct inference always asks first how common the condition is, then how strong the evidence is, and finally how the two interact. This is the structure that Bayes theorem encodes, and it is the discipline that protects against jumping from a single striking fact to an unwarranted conclusion.
The same logic powers the outside view. Before forecasting a specific case — will this start-up survive, will this project ship on time — find the reference class it genuinely belongs to and start from how often the outcome occurs across that class. That frequency is your anchor; the case-specific details should adjust it, not replace it. Forecasting from the reference class beats forecasting from the inside story, because the vivid particulars of a single case are exactly what tempt the mind to ignore how common the outcome already is.