Bayes' Theorem

What Is Bayes' Theorem?

Bayes' Theorem is a formula in probability theory that describes how to revise the probability of an event when new evidence appears. It connects the conditional probability of one event given another with the reverse conditional probability, so an analyst can start with an initial estimate and adjust it as information arrives. It is named after the English mathematician and minister Thomas Bayes.

How It Works

The theorem is usually written as:

P(A|B) = P(B|A) × P(A) / P(B)

Each part has a plain meaning. P(A) is the prior probability, the belief about A before the new evidence is considered. P(B|A) is the likelihood, which measures how probable the evidence is if A is true. P(B) is the overall chance of seeing the evidence under all possibilities. The result, P(A|B), is the posterior probability, the updated belief after the evidence is taken into account.

An Illustrative Example

Suppose a screening test correctly identifies 90% of people who have a certain condition and correctly clears 90% of people who do not. If only 1% of the people tested actually have the condition, then a positive result implies a probability of roughly 8%, not 90%. The low base rate of the condition means that false positives from the large healthy group outnumber true positives from the small affected group. The same reasoning applies to investing: a signal that looks strong can still be unreliable when the event it predicts is rare.

Uses in Investing

Analysts apply Bayesian reasoning when they revise estimates of a company's prospects after a new report, update risk models as market data arrives, or assess whether an apparent pattern in historical data is likely to be meaningful rather than the product of chance. Methods that apply the theorem repeatedly to update beliefs fall under Bayesian inference, a broader approach to statistics.

Limitations

The result depends heavily on the prior probability, which is often a subjective judgment. Different starting assumptions can lead to very different conclusions from the same evidence. Estimating likelihoods accurately can also be difficult, and complex problems may require substantial computation. Bayes' Theorem is a tool for consistent reasoning under uncertainty, not a guarantee that a conclusion is correct.