What is base rates?
- Base rates tell you the probability from history for a specific reference class.
- Your prior odds start from base rates, not from gut feel.
- Bayesian updating combines base rates with new evidence to refine your guess.
- Ignoring base rates leads to the base rate fallacy and costly mistakes.
What is base rates?
Base rates are the probability from history that an outcome occurs in a given reference class. They show you what usually happens before you add new evidence. Your prior odds start from these base rates, and Bayesian math updates them.
Imagine 1% of people have a rare disease. That 1% is your base rate. A test is 80% accurate. If you test positive, your chance is still under 5% because the base rate is low.
Base rate has two unrelated meanings
Searching base rates can show two different ideas. In statistics and decision making, a base rate is a probability from history. In central banking, the base rate is the policy interest rate a central bank sets, and the phrase carries no probability weight at all.
Bank of England and Bank of Canada each call their main interest rate the base rate. It sets the floor for short-term credit across the economy, and lenders pass its direction into mortgages, savings accounts, and business loans. This financial sense is a price, not an odds figure.
The two uses never mix. A central bank rate tells you the cost of money. A statistical base rate tells you the chance an outcome occurs. Know which sense a source means before you apply it, because quoting the wrong number sends you down a false path.
Most investing writing uses the statistical sense. When an analyst says the base rate favors the market's long-run return, they mean the odds history hands you. Keep the central-bank meaning in mind only when the subject is monetary policy and interest rates.
Find your reference class
Your reference class is the group you belong to. Use the same group's history to get your base rate. Pick a tight class, like all 30-year-old drivers, not all drivers.
The class you pick shifts your base rate. Classify as all drivers and the accident odds run high. Tighten it to your age, region, and car, and the number turns personal and useful.
Probability from history beats gut feel
Your gut overweights vivid stories. Probability from history cuts through that. If 90% of startups fail, that is your base rate, and the same ratio applies across similar ventures. Anchoring your bet in what these situations usually deliver.
Your prior odds are your starting point
Prior odds come straight from the base rate. If 1 in 100 has the trait, your prior odds are 1 to 99. Update them only with strong evidence; a single data point rarely justifies a big shift.
Update like a Bayesian
Bayesian math combines your prior odds with new data. Start with the base rate, then adjust each piece of evidence by how strong and reliable it is, weighting solid facts over weak signals.
A positive test on a rare disease still leaves you likely healthy. That is the base rate doing its work: the 1 in 100 odds dominate even an 80 percent accurate test.
Your posterior odds are what you act on
Your posterior odds are the base rate updated by the evidence. Start from the prior odds, then fold in how likely the new fact is under each possibility. The result is the number you actually bet on.
Featural evidence is the one-off detail
Your base rate is what usually happens. Featural evidence is what marks this exact case apart, the details that set it off from the crowd. Bayes' rule pairs the two, so the one-off facts update the odds history gave you.
Screen with the cost of errors in mind
A false positive flags a trait you do not have. A false negative misses the trait entirely. When one error costs far more than the other, adjust your threshold, because the base rate alone does not set it.
Prevalence is the medical name
In medicine the base rate goes by another name: prevalence. It is the share of a population carrying a condition right now. Before a single test, prevalence is your prior, the honest starting point every result gets weighed against.
Base rates power clinical trials
Medical trials lean on base rates. Measure the control group's recovery rate first, then compare the treatment group against that number to see what the drug truly adds.
The comparison cuts both ways. If 1 in 20 recover on their own but only 1 in 100 improve with the drug, the treatment is doing active harm, despite the prescription pad. The base rate difference exposes it.
The base rate fallacy can cost you
People ignore base rates and overreact to new evidence. That mistake leads to bad bets and missed opportunities. Always ask yourself what the base rate is for this reference class.
Base rates in investing
In investing, history is your base rate. Before you bet a stock doubles, ask what share of similar businesses ever have. The market's base rate for the average firm is ordinary growth, not ten-baggers, and anchoring on that keeps hope honest.