Issues with Exercise 9.6.2 about all-of-statistics HOT 2 OPEN

I got roughly the same coverage.
I came here thinking I have done something wrong.

from all-of-statistics.

I agree with the comments above.

@Pipe-Vash If you want to verify that you get the correct asymptotic coverage without having to generate millions of samples ($N \approx 10^6$), you can sample $Y$ from a Gaussian with smaller variance $\sigma^2 < 1$. Then you check for coverage by testing whether the bootstrap intervals contain the true skewness of a general $\text{LogNormal}(\mu=0, \sigma^2)$.

from all-of-statistics.