Course contentsShow
← Mathematics
Lesson 2821 of 3,12155. Probability — FoundationsFree lesson

Combined Applications of Law of Total Probability and Bayes

Solving complex problems requiring both forward probability calculation and backward inference.

Combined Applications of Law of Total Probability and Bayes' Theorem

What you'll learn: How to solve complex probability problems that require both calculating forward probabilities through partitions and then reversing direction to infer causes from observed effects.

The Integration of Two Powerful Tools

Many real-world problems require you to first break down a probability using the Law of Total Probability, then reverse engineer what caused an observed outcome using Bayes' Theorem. Think of it like a detective story: first you map out all possible ways a crime could have occurred (forward), then you use the evidence to determine which scenario actually happened (backward).

The Two-Step Strategy

Step 1 - Forward (Law of Total Probability): When you need to find P(B), partition the sample space using events A₁, A₂, ..., Aₙ and compute:

P(B) = P(B|A₁)P(A₁) + P(B|A₂)P(A₂) + ... + P(B|Aₙ)P(Aₙ)

Step 2 - Backward (Bayes' Theorem): Once you observe B has occurred, update your beliefs about which partition caused it:

P(Aᵢ|B) = P(B|Aᵢ)P(Aᵢ) / P(B)

Notice that the denominator P(B) comes directly from Step 1!

When to Use Both

This combined approach appears when:

  • You first need to find the probability of an observable event
  • Then given that event occurred, you want to identify the most likely underlying cause
  • Problems involving multiple stages of manufacturing, disease diagnosis with prevalence calculations, or multi-source system failures

Key Takeaway: Complex probability problems often require the Law of Total Probability to compute marginal probabilities of observable events, which then serve as the denominator in Bayes' Theorem to perform backward inference about hidden causes—master the interplay between these tools to solve realistic multi-stage problems efficiently.