Course contentsShow
← Mathematics
Lesson 2820 of 3,12155. Probability — FoundationsFree lesson

Sequential Updating with Bayes' Theorem

Updating probabilities step-by-step as multiple pieces of evidence are observed sequentially.

Sequential Updating with Bayes' Theorem

What You'll Learn

You'll learn how to refine probability estimates step-by-step as you observe multiple pieces of evidence, using each updated probability as the starting point for the next observation.

The Core Idea

When you receive multiple pieces of evidence one after another, you don't need to process them all at once. Instead, you can apply Bayes' Theorem repeatedly:

  1. Start with your prior probability for a hypothesis
  2. Observe the first piece of evidence and compute a posterior probability
  3. Treat that posterior as the new prior for the next piece of evidence
  4. Repeat for each subsequent observation

This is called sequential updating or Bayesian updating.

Why This Works

Each application of Bayes' Theorem takes your current belief (prior) and adjusts it based on new data (likelihood). The mathematical beauty is that the order doesn't matter—if you have two pieces of evidence A and B, updating with A then B gives the same final result as updating with B then A, or even processing both simultaneously.

A Practical Example

Imagine you're diagnosing whether a patient has a disease (D):

  • Step 0: Prior probability P(D) = 0.01 (1% prevalence)
  • Step 1: First test comes back positive. Apply Bayes' Theorem → posterior P(D|Test₁⁺) = 0.15
  • Step 2: Use 0.15 as your new prior. Second test positive → P(D|Test₁⁺, Test₂⁺) = 0.70
  • Step 3: Third test positive → P(D|Test₁⁺, Test₂⁺, Test₃⁺) = 0.95

Each test refines your belief, compounding the evidence.

When to Use Sequential Updating

This approach is essential when evidence arrives over time, in medical diagnosis with repeated tests, machine learning algorithms, or any scenario where you continuously update beliefs as new information emerges.

Key Takeaway: Sequential updating lets you process evidence piece by piece, using each posterior probability as the prior for the next step—efficiently accumulating information without recomputing everything from scratch.