I want to show the following:
Let X1,X2,… i.i.d. random variables on (Ω,A,P) with infinite expectation. Let Sn=X1+…+Xn. Then
P(limn→∞Snn exists in R)=0.
Using the second Borel-Cantelli-Lemma, I already showed that
P(|Xn|≥n infinitely often)=1.
Now I'd like to show that
P(limn→∞Snn exists in R)≤P(|Xn|≥n at most finitely often).
Here's what I did:
Suppose that there exists ω∈Ω s.t. Xn(ω)≥n infinitely often and Sn(ω)n→L∈R.
Note that
limn→∞Sn(ω)n+1=limn→∞nn+1Sn(ω)n=L
Then
limsupn→∞|Sn+1(ω)n+1−L|=limsupn→∞|Xn+1(ω)n+1+Sn(ω)n−L|≥limsupn→∞|Xn+1(ω)n+1|−|Sn(ω)n+1−L|=limsupn→∞|Xn+1(ω)n+1|−liminfn→∞|Sn(ω)n+1−L|=1−0=1,
which is a contradiction to limn→∞Sn(ω)n=L. Therefore
{limn→∞Snn exists in R}⊂{|Xn|≥n at most finitely often},
which implies (1) and therefore the claim.
Is this correct? I'm mostly unsure about equation (2). Is it always true that the limsup of a difference of two positive sequences is the limsup of the first sequence minus the liminf of the second one?
Thank you for your help!
Answer
As @KaviRamaMurthy correctly pointed out, (2) is incorrect. However, the proof can be fixed since for any real sequences an and bn, if bn is convergent, it holds that
limsupn→∞(an+bn)=limsupn→∞an+limn→∞bn.
So I can write
limsupn→∞|Sn+1(ω)n+1−L|=limsupn→∞|Xn+1(ω)n+1+Sn(ω)n−L|≥limsupn→∞|Xn+1(ω)n+1|−|Sn(ω)n+1−L|=limsupn→∞|Xn+1(ω)n+1|−limn→∞|Sn(ω)n+1−L|=1−0=1,
No comments:
Post a Comment