Motivation for Bayes' Theorem
You'll learn why we need a systematic way to reverse conditional probabilities and update our beliefs when new evidence arrives.
The Reversal Problem
You've learned how to compute conditional probabilities like P(B|A): "Given that A happened, what's the probability of B?" But real-world problems often ask the reverse question.
Imagine you know:
- The probability a medical test is positive given you have a disease: P(Positive|Disease)
- But what you actually want is: the probability you have the disease given the test is positive: P(Disease|Positive)
These are not the same probability, and one doesn't immediately tell you the other!
Why We Can't Just Flip It
Consider this: almost everyone with a rare disease tests positive (high P(Positive|Disease)), but if the disease affects only 0.01% of people, most positive tests might still come from healthy people who got false positives. So P(Disease|Positive) could be quite low even when P(Positive|Disease) is high.
You can't simply swap the events in a conditional probability.
The Need for Systematic Reversal
What you've built so far gives you the tools:
- The Multiplication Rule: P(A ∩ B) = P(A|B)·P(B) = P(B|A)·P(A)
- The Law of Total Probability: breaking down P(B) across different scenarios
Bayes' Theorem combines these to give you an explicit formula for reversing conditional probabilities. It lets you start with P(B|A) and the "base rates" P(A) and P(B), then systematically compute P(A|B).
Key Takeaway: Bayes' Theorem solves the practical problem of reversing conditional probabilities, allowing you to update beliefs about causes given observed effects—essential when evidence flows in the opposite direction from your initial knowledge.