Substructure detection rate as a function of SDSS/WISE positional offset for mergers with high-resolution NIR imaging, including the Primary Keck AO Merger Sample. The sample is divided into eight bins in offset, where the detection rate in each bin is defined as the fraction of systems in which Keck AO imaging reveals unresolved NIR substructure. The error bars represent 90% binomial CIs. This figure evaluates whether the SDSS/WISE offset selection method remains reliable at small offsets, where astrometric noise could potentially dominate. The absence of a systematic decline in detection rate toward smaller offsets may suggest that the observed SDSS/WISE offsets are associated with genuine nuclear substructure rather than being solely driven by astrometric uncertainties.