\section{The compound-Poisson benchmark} Let $U,V$ be independent $\Pois(\lambda/2)$ variables, and put $K=U+V$, $D=U-V$. Define \begin{equation}\label{eq:F} F(\lambda)=\frac{\Var(\E[K\mid D])}{\lambda} =\lambda\sum_{d\in\mathbb Z}\frac{(p_\lambda'(d))^2}{p_\lambda(d)}, \qquad p_\lambda(d)=e^{-\lambda}I_{|d|}(\lambda). \end{equation} This is the symmetric compound-Poisson intensity experiment of \cite[Eqs.~(18), (22)--(23) and (29)]{DuvalHoffmann2011}, normalized by complete-jump information. We record elementary estimates needed for the finite-population reduction, independently of a Gaussian limit theorem. \begin{lemma}\label{lem:Fbounds} For every $\lambda>0$, \begin{equation}\label{eq:Fbounds} 00}|R_\Omega(\lambda)-F(\lambda)| \le C\Omega^{-1/4}\sqrt{\log(e\Omega)}. \end{equation} Consequently, uniformly for $0<\Delta\le T$ and $\theta\in[\theta_-,\theta_+]$, \begin{equation}\label{eq:uniformhorizon} \left|\frac{\FI_{\Omega,\Delta,T}(\theta)}{\Omega T/\theta} -F(\Omega\theta\Delta)\right| \le C\Omega^{-1/4}\sqrt{\log(e\Omega)}+ \frac2{\Omega\theta_-T}. \end{equation} \end{theorem} The clipped-score and tail estimates, and the complete proof of Theorem~\ref{thm:uniform}, are given in Appendix~\ref{app:population-proofs}. Theorem~\ref{thm:uniform} gives additive error even when the information fraction vanishes. The relative approximation in Section~\ref{sec:window} identifies its exact window of validity. At finite population, the fraction is not a function independent of $\Omega$: for example, \begin{equation}\label{eq:omegaone} R_1(\lambda)=\frac{4\lambda}{e^{4\lambda}-1}\ne F(\lambda). \end{equation}