\section{Relative agreement and the sharp sampling window}\label{sec:window} The stationary particle representation permits control of the observed score at intensities where additive convergence does not determine relative error. \begin{theorem}[Relative crossover and sharp window]\label{thm:relative} Let $R_\Omega$ and $F$ be as in \eqref{eq:R} and \eqref{eq:F}, and put $t_0=(\log2)/2$. For every integer $\Omega\ge16$ and real $1\le L\le t_0\Omega$, \begin{equation}\label{eq:relativewindow} \sup_{1\le\lambda\le L} \left|\frac{R_\Omega(\lambda)}{F(\lambda)}-1\right| \le\frac{9L}{\Omega} +2000\Omega^{-1/6}\log(e\Omega)^{1/3}. \end{equation} For every deterministic sequence $L_\Omega\ge1$, \begin{equation}\label{eq:relativeiff} \sup_{1\le\lambda\le L_\Omega} \left|\frac{R_\Omega(\lambda)}{F(\lambda)}-1\right|\longrightarrow0 \quad\Longleftrightarrow\quad L_\Omega/\Omega\longrightarrow0. \end{equation} More precisely, for fixed $0