Probability with Measure
2.2 Liminf and limsup
The main difficulty with limits is that, in general, limits do not exist. Most sequences do not converge to anything. However there are two closely related concepts that always exist. They are much easier to
work with but they take more care to define.
Let be a sequence of real numbers. Note that the sequence is monotone decreasing, because as gets larger the set contains less terms. Lemma
2.1.1 implies that this sequence has a limit, with the caveat that the limit might be infinite. The sequence is monotone decreasing, so its
limit is equal to . With this is mind we make the following definition:
Heuristically, is the smallest value that the tail of the sequence stays below.
We can do the same construction the other way up, which gives
Heuristically, is the largest value that the tail of the sequence stays above. Note that (2.3)
and (2.4) are always well defined, as extended real numbers, and that .
-
It is helpful to see a picture:
The sequence displayed is a sample of , where are i.i.d. uniform random variables on . This is chosen to make a clear picture. As the term will oscillate within and the second term will tend to zero. Note that the dotted lines converge downwards to (in red) and upwards to
(in green).
We will spend the rest of this section making connections between , and .
Proof: Suppose first that converges, say . We will consider the case here; the case is similar and we will omit it, as discussed in Remark 2.1.2. For all there exists
such that for all . Hence for all , which implies that
Without loss of generality we may choose . Letting , upon which , gives that
We have therefore proved both part 2 and the forwards implication of part 1.
We need to prove the reverse implication from part 1. Suppose that (and we don’t yet know that converges). For all we have
Note also that
Combining (2.5) with (2.6) and using the
sandwich rule, we have
We can write , and we know that both of these terms have a limit as . The first tends to zero by
(2.7) and the second converges to . From the algebra of limits we thus obtain that converges and
. Since was our assumption, this completes the proof. ∎
Proof: You already know from real analysis that for real .
This equation also holds for extended reals, but we’ll omit checking the extra cases involving infinities here. The result follows from this along with (2.3) for , and with (2.4) for .
∎