Let $A_n, n\ge1,$ be events in a probability space $(\Omega,\mathcal{F}, P)$ that satisfy the property $$P(A_n\cap A_m)=\emptyset, m\not=n$$ Prove $P(\bigcup_{n\ge 1} A_n)=\sum_{n\ge 1}P(A_n).$

I can prove the result for finite unions by using the measure-theoretic version of the Inclusion-Exclusion Principal, but this is useless since $\lim P(\bigcup _{i=1} ^n A_i)$ isn't necessarily equal to $P(\bigcup _{i\ge 1}A_i).$ Does the Inclusion-Exclusion Principle hold for infinite unions? I believe there's an easier way to solve than using that principal, but I am not sure how to do it.

I tried constructing a new sequence $B_n=A_n-\bigcup_{i=1}^{n-1}A_i$, but this doesn't allow me to use the assumption. The sequence $B_n=A_n-\bigcap_{i=1} ^nA_i$ doesn't suffice since it's not a disjoint sequence.


The given condition says that the sets $A_i, A_j$ only intersect in some tiny sets that nobody cares about. So throw it all away. Then show that putting any amount of it back doesn't change anything.

  1. Let $N = \bigcup_{n \ne m} (A_n \cap A_m)$. Show that $P(N)=0$.

  2. Let $B_n = A_n \setminus N$. Check that $P(B_n) = P(A_n)$. Check that the $B_n$ are disjoint. Let $B = \bigcup_n B_n$ and show that $P(B) = \sum_n P(A_n)$.

  3. Show that $B \subset \bigcup_n A_n \subset (B \cup N)$. Conclude that $P(\bigcup_n A_n) = P(B)$.

| cite | improve this answer | |

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.