= Solution
The <empirical distribution function> is $\widehat F_n(x)=n^{-1}\sum_i\mathbf1_{X_i\leq x}$. For continuous F, the <probability integral transform> gives independent uniform variables $U_i=F(X_i)$. Strict increase of F is unnecessary. More explicitly, generate the observations by $X_i=F^{-1}(U_i)$ with the generalized quantile inverse. Then $\mathbf1_{X_i\leq x}=\mathbf1_{U_i\leq F(x)}$, so $\widehat F_n(x)=G_n(F(x))$, where $G_n$ is the uniform empirical distribution.
Continuity and the limits of F ensure that its range contains $(0,1)$. The endpoint discrepancies for $G_n$ are zero. Hence
$$
\boxed{D_n\overset d=\sup_{0\leq u\leq1}|G_n(u)-u|,}
$$
which depends only on n. Equivalently it is $\max_{1\leq i\leq n}\{i/n-U_{(i)},U_{(i)}-(i-1)/n\}$, a function of uniform <order statistics>. This proves the distribution-free property of the <Kolmogorov-Smirnov statistic> even when F has flat portions.
Choose $c_{n,\alpha}$ from this common distribution so that $\mathbb P(D_n\leq c_{n,\alpha})=1-\alpha$. The <Kolmogorov-Smirnov confidence band> is
$$
\boxed{\max(0,\widehat F_n(x)-c_{n,\alpha})\leq F(x)\leq\min(1,\widehat F_n(x)+c_{n,\alpha})\quad\text{for all }x.}
$$
The event that all these inequalities hold is exactly the event used to choose the critical value, giving simultaneous coverage $1-\alpha$. The <Glivenko-Cantelli theorem> states that for an independent identically distributed sample from any <distribution function>, continuous or not, $\sup_x|\widehat F_n(x)-F(x)|\to0$ almost surely.
For an integrable r-argument <U-statistic kernel>, replace it by $h_s(x_1,\ldots,x_r)=(r!)^{-1}\sum_\pi h(x_{\pi(1)},\ldots,x_{\pi(r)})$. Every permutation has the same expectation by <exchangeability>, so this <symmetrization of a U-statistic kernel> preserves the target functional. For a symmetric <U-statistic kernel> h and $n\geq r$, the <U-statistic> is
$$
U_n={\binom nr}^{-1}\sum_{i_1<\cdots<i_r}h(X_{i_1},\ldots,X_{i_r}).
$$
Each summand has expectation theta, hence the statistic is unbiased.
For the population <variance>, use $h(x,y)=(x-y)^2/2$. <Independence> gives $\mathbb E(X_1-X_2)^2=2\operatorname{Var}(X_1)$ when the second moment is finite. The corresponding statistic simplifies using $\sum_{i<j}(X_i-X_j)^2=n\sum_i(X_i-\overline X)^2$:
$$
\boxed{U_n=\frac1{n(n-1)}\sum_{i<j}(X_i-X_j)^2=\frac1{n-1}\sum_i(X_i-\overline X)^2.}
$$
Thus <variance as a U-statistic> is precisely the usual unbiased <sample variance>, for $n\geq2$.
Back to article page