NICEFA Logo
nicefa.
Library
←Module

Proof: Borel-Cantelli Lemma 2 (Independence, Divergent Sum)

The second Borel-Cantelli Lemma for independent events with a divergent sum of probabilities.

The Formal Theorem

\sum P(A_n) = \infty \text{ and independence } \implies P(A_n \text{ i.o.}) = 1

Analytical Intuition.

The second lemma is the guarantee of certainty. If events are independent and their total probability is infinite, the event isn't just possible—it is inevitable.
CAUTION

Institutional Warning.

Requires independence. Without it, the events could "overlap" in a way that the sum is infinite but the event never happens twice.

Institutional Deep Dive.

01
The second Borel-Cantelli Lemma is a statement of profound inevitability under specific conditions, a cornerstone in the theory of convergence of random variables and stochastic processes. It offers a stark contrast to its first iteration, which merely provides an upper bound on probability. Here, we encounter a guarantee of certainty.
02
Core Analytical Logic
03
The lemma states: given a sequence of independent events {An}n=1∞ \{A_n\}_{n=1}^\infty , if the sum of their probabilities diverges, i.e., ∑n=1∞P(An)=∞ \sum_{n=1}^\infty P(A_n) = \infty , then the probability that infinitely many of these events occur is unity. Formally, P(lim sup⁡n→∞An)=1 P(\limsup_{n \to \infty} A_n) = 1 . The term lim sup⁡An \limsup A_n denotes the set of outcomes belonging to infinitely many An A_n .
04
The logical scaffolding supporting this assertion is built upon the premise of independence. Without this condition, the lemma collapses; consider An=A A_n = A for all n n , where P(A)=p∈(0,1) P(A) = p \in (0, 1) . Then ∑P(An)=∑p=∞ \sum P(A_n) = \sum p = \infty , but lim sup⁡An=A \limsup A_n = A , and P(A)=p≠1 P(A) = p \neq 1 . Independence ensures that the occurrence of one event An A_n provides no information about the occurrence of any other Am A_m (m≠n m \neq n ), thus preventing a form of probabilistic redundancy or exhaustion. Each event An A_n truly offers a new, independent opportunity for the phenomenon to manifest.
05
The proof pivots on analyzing the complement. If P(lim sup⁡An)<1 P(\limsup A_n) < 1 , it implies a non-zero probability that only a finite number of events An A_n occur. This means there exists some integer N N such that for all n>N n > N , the event An A_n does *not* occur. In other words, for all n>N n > N , the outcome lies in Anc A_n^c . Consequently, the probability of finitely many occurrences is equivalent to P(∪N=1∞∩n=N∞Anc) P(\cup_{N=1}^\infty \cap_{n=N}^\infty A_n^c) . By monotonicity and continuity of probability, this is lim⁡N→∞P(∩n=N∞Anc) \lim_{N \to \infty} P(\cap_{n=N}^\infty A_n^c) .
06
Exploiting the independence of {An} \{A_n\} , the events {Anc} \{A_n^c\} are also independent. Thus, for any N N ,
P(⋂n=N∞Anc)=∏n=N∞P(Anc)=∏n=N∞(1−P(An)) P\left(\bigcap_{n=N}^\infty A_n^c\right) = \prod_{n=N}^\infty P(A_n^c) = \prod_{n=N}^\infty (1 - P(A_n))
A standard analytical inequality, 1−x≤e−x 1 - x \le e^{-x} for x∈R x \in \mathbb{R} , is now applied to each term:
∏n=N∞(1−P(An))≤∏n=N∞e−P(An)=e−∑n=N∞P(An) \prod_{n=N}^\infty (1 - P(A_n)) \le \prod_{n=N}^\infty e^{-P(A_n)} = e^{-\sum_{n=N}^\infty P(A_n)}
Given the hypothesis that ∑n=1∞P(An)=∞ \sum_{n=1}^\infty P(A_n) = \infty , it follows that ∑n=N∞P(An) \sum_{n=N}^\infty P(A_n) also diverges to infinity for any finite N N . Therefore, e−∑n=N∞P(An)→e−∞=0 e^{-\sum_{n=N}^\infty P(A_n)} \to e^{-\infty} = 0 . This demonstrates that P(∩n=N∞Anc)→0 P(\cap_{n=N}^\infty A_n^c) \to 0 as N→∞ N \to \infty . Consequently, the probability of only finitely many events occurring is 0. This necessitates that the probability of infinitely many events occurring is 1. The infinite sum, when converted to an exponential of a negative sum, inexorably forces the probability of the complement to zero.
07
Geometric Mechanics
08
Visualize the sample space Ω \Omega as a unit hypercube, and each event An A_n as a measurable subset within it. The condition ∑P(An)=∞ \sum P(A_n) = \infty means that the 'total measure' or 'volume' attributed to the individual opportunities An A_n is unbounded. However, this is not merely an additive measure. When considering the complement Anc A_n^c , we are looking at the 'space outside' each event.
09
The product ∏n=N∞(1−P(An)) \prod_{n=N}^\infty (1 - P(A_n)) represents the cumulative probability of *avoiding* all events An A_n for n>N n > N . If these events were overlapping in a dependent manner, this product could remain non-zero. However, independence dictates that each Anc A_n^c represents an independent 'region of non-occurrence'.
10
Consider the transformation: an infinite sum ∑P(An) \sum P(A_n) corresponds to an infinite exponent in e−∑P(An) e^{-\sum P(A_n)} . As this exponent goes to negative infinity, the entire expression collapses to zero. Geometrically, this means that the intersection ∩n=N∞Anc \cap_{n=N}^\infty A_n^c – the region where *none* of the events An A_n occur beyond a certain point N N – must have its measure shrink to zero as N N increases. It is squeezed out of existence.
11
The 'infinitely often' condition, lim sup⁡An \limsup A_n , can be thought of as a region that is visited by an infinite sequence of An A_n . If the probability of being *outside* this region (i.e., being in only finitely many An A_n ) is zero, then the region itself must encompass almost the entire sample space. The sum diverging is not merely a quantitative statement; it is a qualitative statement about the persistent 'presence' of event opportunities, which, when independent, cannot be circumvented.
12
Institutional Pitfalls
13
Students frequently stumble on the precise interpretation and application of this lemma due to several common conceptual deficiencies.
14
Firstly, the indispensable requirement of **independence** is often conflated with weaker notions or simply overlooked. A cursory reading might lead one to believe that a sufficiently large aggregate probability is enough. This gross error disregards the lemma's core mechanism: the ability to multiply probabilities of complements. Without independence, one cannot decompose P(∩Anc) P(\cap A_n^c) into a product, thus invalidating the entire analytical chain. The failure to rigorously check this premise is a fundamental flaw, indicative of a superficial engagement with the material.
15
Secondly, students often misinterpret ∑P(An)=∞ \sum P(A_n) = \infty . They might erroneously assume this implies that individual P(An) P(A_n) must be large, or even bounded away from zero. This is patently false. Consider P(An)=1/n P(A_n) = 1/n , for instance. Each term tends to zero, yet the harmonic series diverges. The power of the lemma lies in its ability to guarantee inevitability even when each individual chance of occurrence is vanishingly small, provided these chances accumulate unboundedly and are independent. The intuition must be calibrated to understand that an infinite accumulation of even infinitesimally small, yet distinct, opportunities leads to certainty.
16
Finally, the abstract nature of "infinitely often" and "almost surely" often eludes those without a firm grasp of measure theory. The phrase "almost surely" is not a weak statement; it signifies that the exceptional set, where the event does not occur, is negligible—a set of measure zero. This is a profound distinction from merely "with probability 1," which, without the underlying measure-theoretic framework, can be misinterpreted. The brutal clarity required here demands that students understand that an event of probability zero is, for all practical and theoretical purposes within this context, an impossibility. To neglect these nuances is to miss the profound implications of the lemma.

Academic Inquiries.

01

Why independence?

Independence prevents the probability mass from clustering on the same outcomes.

Standardized References.

  • Definitive Institutional SourceInstitutional Reference (nicefa v1)

Related Proofs Cluster.

Advanced

Borel-Cantelli

Infinite predictors.

Advanced

Martingale Convergence

Fair game fate.

Advanced

Ergodic Theorem

Time-space unity.

Institutional Citation

Reference this proof in your academic research or publications.

NICEFA Visual Mathematics. (2026). Proof: Borel-Cantelli Lemma 2 (Independence, Divergent Sum): Visual Proof & Intuition. Retrieved from https://nicefa.org/library/advanced-probability-theory/borel-cantelli-lemma-2-independence

Dominate the Logic.

"Abstract theory is just a movement we haven't seen yet."

Master the ProofEarly Access