Simulating the impact of the fraction of missing v_r measurements on the relative error of inferred radial velocities. The top panel shows relative errors of inferred v_r on the open cluster sample as a function of missing-ness. The bottom panel shows the result of this study using the compact cluster sample. We find twice larger errors on the compact cluster sample. Additionally, the estimation procedure seems almost independent of the number of missing radial velocity measurements in both samples, except in a small region of the compact cluster sample, where the error increases if fewer than 5% of sources have v_r measurements.