2
$\begingroup$

In a paper in Lecture Notes in Mathematics vol. 1874, Yan states the Kochen-Stone theorem in the following form, where $A_n$ is a sequence of events such that $\sum_{n=1}^\infty P(A_n) = \infty$:

$$ P(A_n~\mbox{i.o.}) \geq \limsup_{n \to \infty} \frac{(\sum_{k=1}^n P(A_k))^2}{\sum_{i,k=1}^n P(A_i A_k)} = \limsup_{n \to \infty} \frac{\sum_{1 \le i < k \le n} P(A_i)P(A_k)}{\sum_{1 \le i < k \le n}P(A_iA_k)} \tag*{$(2)$} $$

((2) is his numbering).

I understand his proof of the inequality in (2), but I can't make sense of his proof of the equation (which states that the diagonal terms in the sums in the first fraction in (2) are negligible). He says "Since $\sum_{k=1}^\infty P(A_k) = \infty$ and $$ \left(\sum_{k=1}^n P(A_k)\right)^2 \le 2\sum_{1 \le i < k \le n}P(A_i)P(A_k) + \sum_{k=1}^n P(A_n) \tag*{$(3)$} $$

we have $$ \lim_{n \to \infty}\frac{\sum_{k=1}^n P(A_k)}{\sum_{1 \le i < k \le n}P(A_i)P(A_k)} = 0 \tag*{$(4)$} $$ Thus equality (2) holds." ((3) and (4) are my numbering.)

I see how to get from (3) to (4) (by dividing by $\sum_{k=1}^n P(A_k)$ and noting that the left-hand side of the resulting inequality diverges), but I can't see how to get the equality in (2) from (4). Any suggestions on how to see this or any alternative proofs will be gratefully received as I am doing some work that relies heavily on the equation and am reluctant to use a result that I don't understand. Any relevant references would also be much appreciated.

$\endgroup$

1 Answer 1

2
$\begingroup$

Let $$S_n=\sum_{k=1}^n P(A_k),\quad T_n:=\sum_{1\le i<k\le n}P(A_i)P(A_k),$$ $$R_n=\sum_{i,k=1}^n P(A_iA_k),\quad U_n:=\sum_{1\le i<k\le n}P(A_iA_k).$$ Then $2T_n\le S_n^2\le2T_n+S_n$, and $S_n<<S_n^2$ (because $S_n\to\infty$); we write $a<<b$ or, equivalently, $b>>a$ to mean $a=o(b)$; all the limits are taken as $n\to\infty$. Thus, $$S_n^2\sim2T_n.\tag{*}$$ Also, $2U_n\le R_n\le2U_n+S_n$, $S_n<<S_n^2$, and $S_n^2\le (1+o(1))R_n$ (by the inequality in (2)). Thus, $$R_n\sim2U_n.\tag{**}$$ From (*) and (**), we get $$\frac{S_n^2}{R_n}\sim\frac{T_n}{U_n},$$ which immediately yields the equality in (2).

$\endgroup$
1
  • $\begingroup$ Thanks. So Yan was begin rather coy! What I have called (4) says $S_n/T_n \to 0$, which leads to your $(*)$. But Yan was leaving your $(**)$ entirely to the reader. $\endgroup$
    – Rob Arthan
    Sep 18, 2020 at 15:11

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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