The retention of floating matter within the surf zone on a rip-channeled beach is examined with a combination of detailed field observations obtained during the Rip Current Experiment and a three-dimensional (3-D) wave and flow model. The acoustic Doppler current profiler–observed hourly vertical cross-shore velocity structure variability over a period of 3 days with normally incident swell is well reproduced by the computations, although the strong vertical attenuation of the subsurface rip current velocities at the most offshore location outside the surf zone in 4 m water depth is not well predicted. Corresponding mean alongshore velocities are less well predicted with errors on the order of 10 cm/s for the most offshore sensors. Model calculations of very low frequency motions (VLFs) with O(10) min timescales typically explain over 60% of the observed variability, both inside and outside of the surf zone. The model calculations also match the mean rip-current surface flow field inferred from GPS-equipped drifter trajectories. Seeding the surf zone with a large number of equally spaced virtual drifters, the computed instantaneous surface velocity fields are used to calculate the hourly drifter trajectories. Collecting the hourly drifter exits, good agreement with the observed surf zone retention is obtained provided that both Stokes drift and VLF motions are accounted for in the modeling of the computed drifter trajectories. Without Stokes drift, the estimated number of virtual drifter exits is O(80)%, almost an order of magnitude larger than the O(20)% of observed exits during the drifter deployments. Conversely, when excluding the VLF motions instead, the number of calculated drifter exits is less than 5%, thus significantly underestimating the number of observed exits.