The first secure detection of a gravitationally lensed gravitational (GW) will be a watershed moment, as it will bring together these two pillars of General Relativity for the first time. Accurate selection and interpretation of candidate lensed GWs is challenging for numerous reasons, including large sky localization uncertainties for most GW detections, the broad range of gravitational lenses spanning galaxy/group/cluster-scales in the dark matter halo mass function, and uncertainty in the intrinsic mass function of compact object remnants of stellar evolution. We introduce a new magnification-based approach to predicting the rates of lensed GWs that is agnostic to the mass and structure of the lenses and combine it with expressions for arrival time difference for representative lenses and their catastrophes to delineate the range of expected arrival time differences. We also predict the lightcurves of lensed kilonova counterparts to lensed binary neutron star (NS-NS) mergers and assess the feasibility of detection with the Vera Rubin Observatory. Our main conclusions are: (1) selection of candidate lensed NS-NS mergers from the mass gap between NS and black holes in low latency is an efficient approach with a rate approaching one per year in the mid-2020s, (2) the arrival time differences of lensed NS-NS/kilonovae are typically $\lesssim1\,$year, and thus well-matched to the operations of GW detectors and optical telescopes, and (3) detection of lensed kilonovae is feasible with the Vera Rubin Observatory. Whilst our predictions are motivated by lensing, they provide a physically well-understood approach to exploring the mass gap electromagnetically as the number of detections in this exciting region of parameter space grows.