Kepler planets around a given star have similar sizes to each other and regular orbital spacing, like "peas in a pod." Several studies have tested whether detection bias could produce this apparent pattern by resampling planet radii at random and applying a sensitivity function analogous to that of the Kepler spacecraft. However, Zhu argues that this pattern is not astrophysical but an artifact of Kepler's discovery efficiency at the detection threshold. To support this claim, their new analysis samples the transit signal-to-noise ratio (S/N) to derive a synthetic population of bootstrapped planet radii. Here, we examine the procedure of sampling transit S/N and demonstrate it is not applicable. Sampling transit S/N does not set up random, independent planet radii, and so it is unsuitable for corroborating (or falsifying) detection bias as the origin of apparent patterns in planet radius. By sampling the planet radii directly and using a simple model for Kepler's sensitivity, we rule out detection bias as the source of the peas-in-a-pod pattern with >10σ confidence.