Souza (2022) replicates the findings of Kelly and Pruitt (2013), then critiques their findings with alternative empirical choices. We challenge this critique, and argue that the choices of Kelly and Pruitt (2013, 2015) are the natural economic and statistical choices. The attenuation of predictability in Souza’s (2022) empirical analysis is due primarily to a number of ill-advised implementation choices. We conclude that the results of Kelly and Pruitt (2013) are notably robust in the sample following publication of their paper.