Odds Form of Bayes

This formula rewrites Bayes’ theorem in odds form: instead of updating a probability directly, you update the prior odds of a hypothesis by multiplying by a likelihood ratio. The factor P(B|A) / P(B|¬A) tells you how much more likely the observed evidence B is if A is true than if A is false; that is why it acts as an evidence multiplier. It works because Bayes’ theorem can be algebraically rearranged so the prior-to-posterior change separates cleanly into a prior term and an evidence term.

Use this when you want to update a yes/no hypothesis from evidence, especially in diagnostic testing, classification, or decision analysis. It is especially helpful when you already know or can estimate odds and likelihoods more naturally than raw probabilities.

• Using P(A|B) in place of O(A|B) and forgetting to convert between probability and odds
• Swapping P(B|A) and P(B|¬A), which reverses the evidence update
• Forgetting that the denominator is the probability of the evidence under the alternative hypothesis, not 1 - P(B|A) unless that is explicitly the false positive rate