I'm having some trouble verifying the following line from Extremes and Related Properties of Random Sequences and Processes by Leadbetter:

If (1.2.1)' and (1.2.3)' hold, then obviously so does (1.2.2)' enter image description here

With $F'_n(x)$ converging weakly to $G(x)$ and $\alpha_n' \to a$ and $\beta'_n \to b$, it's clear that $F'_n(ax+b) \to G(ax+b)$ weakly. However, why should $F_n'(\alpha'_n x+ \beta'_n)$ converge weakly to $G(ax+b)$? Since $F_n'$ may not be continuous, it seems we cannot guarantee the convergence even though $\alpha'_n x + \beta'_n \to ax+b$.

  • 2
    $\begingroup$ Phil, could you cite the source you took the snapshot of? It's appreciable if you mention the source alongside the image/snap. $\endgroup$ Jan 13, 2023 at 7:41
  • $\begingroup$ @User1865345 sure it's added now $\endgroup$
    – Phil
    Jan 13, 2023 at 15:14
  • $\begingroup$ Thanks for the citation. $\endgroup$ Jan 13, 2023 at 15:15
  • 1
    $\begingroup$ See math.stackexchange.com/questions/285133/… $\endgroup$ Jan 13, 2023 at 16:46

1 Answer 1


If $X_n \to X$ in distribution and $b_n \to b$, then we may pick $Y_n = X_n$ in distribution for $1 \leq n \leq \infty$ with $Y_n \to Y$ almost surely.

Clearly ,$Y_n + b_n \to Y + b$ almost surely, so $X_n + b_n \to X+b$ in distribution.

We can prove the same thing for $a_n X_n \to aX$, which means $\frac{X_n - \alpha '}{\beta_n'} \to \frac{X-a}{b}$ in distribution.


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.