Theorem Statement
Assumptions
- A_n are measurable events in a probability space
Proof Sketch
The event that A_n occurs infinitely often is the intersection over N of the union of all A_n with n at least N. Apply the countable union bound to each tail union. Since the probability series is summable, the tail sums go to zero, so the limsup event has probability zero. No independence is used.
Proof Obligations
Summable probabilities for the event sequence
assumptionVerify the series of event probabilities is finite before taking tail sums to zero in the first Borel-Cantelli argument.
Without summability the first lemma gives no finite-occurrence conclusion.
- source
- durrett-2019:thm231
Limsup as shrinking tail unions
proof stepThe proof represents infinitely-often occurrence as an intersection of future tail unions and bounds each tail by the remaining probability sum.
Confusing a finite union with the tail limsup loses the almost-sure conclusion.
- source
- durrett-2019:thm231
No independence and no converse in this direction
boundaryThe first lemma does not need independence, but it also does not prove the divergent-series converse supplied by the second Borel-Cantelli lemma.
Adding independence to the first lemma or inferring the converse misstates the theorem's scope.
Tail sums must actually vanish
boundaryAfter applying the countable union bound to future tails, verify the remaining probability sums go to zero from summability.
Quoting Borel-Cantelli without the tail-sum limit loses the almost-sure finite-occurrence conclusion.
- source
- durrett-2019:thm231
Source Locators
Probability: Theory and Examples (2019)
proofChapter 2 — Laws of large numbers and independence tools
durrett-2019:ch2
- kind
- chapter
- review
- reviewed / B
- extract
- manual
Chapter-level support for Borel-Cantelli lemmas.
Probability and Measure (1995)
backgroundChapter 1 — Probability measure foundations
billingsley-1995:ch1
- kind
- chapter
- review
- reviewed / B
- extract
- manual
Probability-space foundation context.
Probability: Theory and Examples (2019)
proofTheorem 2.3.1 — Borel-Cantelli lemma
durrett-2019:thm231
- kind
- theorem
- review
- reviewed / A
- extract
- manual
- number
- 2.3.1
Anchor: Borel-Cantelli lemma
Reviewed theorem-level support for the first Borel-Cantelli lemma.
Curated Trace
- 1Tail union boundinput theorem
Each future tail of bad events is controlled by the sum of its probabilities.
Checkpoint
Use the countable union bound on the tail, not only on the first N events.
- 2First Borel-Cantellilemma
Summable probabilities force the limsup event to have probability zero.
Checkpoint
No independence assumption appears in this direction.
- 3Infinitely often eventinterpretation
The lemma translates a tail-union measure calculation into finitely many occurrences almost surely.
Checkpoint
A_n i.o. means every tail contains some occurring event.
- 4Almost-sure convergence patterndownstream
Many strong-law proofs reduce convergence failures to a summable sequence of tail probabilities.
Checkpoint
The lemma proves a sufficient condition, not a converse without independence.
Failure Modes
No independence and no converse in this direction
The first lemma does not need independence, but it also does not prove the divergent-series converse supplied by the second Borel-Cantelli lemma.
Risk: Adding independence to the first lemma or inferring the converse misstates the theorem's scope.
Check: The first lemma does not need independence, but it also does not prove the divergent-series converse supplied by the second Borel-Cantelli lemma.
- source
- durrett-2019:ch2
Tail sums must actually vanish
After applying the countable union bound to future tails, verify the remaining probability sums go to zero from summability.
Risk: Quoting Borel-Cantelli without the tail-sum limit loses the almost-sure finite-occurrence conclusion.
Check: After applying the countable union bound to future tails, verify the remaining probability sums go to zero from summability.
- source
- durrett-2019:thm231