Theorem
Probability • Addition and Multiplication Theorems
Statement
For any events, , because any overlap is counted more than once in the sum. Applied to the complements, it gives .
How to use it
- 1To test a claimed inequality between such probabilities, try extreme events: every event empty, or every event the whole sample space.
Worked example
Show that .
Step-by-step
- 1
Why: Union bound on the complements.
- 2
Why: De Morgan: .
- 3
Answer
Watch out
The inequalities point one way only. The union is at most the sum; the intersection is at least the sum minus (for three events).
Builds on
Also known as
Boole's inequality, Bonferroni inequality