Abstract Dascher-Cousineau et al. (2020) apply the so-called foreshock traffic-light system (FTLS) model proposed by Gulia and Wiemer (2019) to two earthquake sequences that occurred after the submission of the model: the 2019 Ridgecrest (Mw 7.1) and the 2020 Mw 6.4 Puerto Rico earthquakes. We show in this comment that the method applied by Kelian Dascher-Cousineau et al. (2020) deviates in at least six substantial and not well-documented aspects from the original FTLS method. As a consequence, they used for example in the Ridgecrest case only 1% of the data available to estimate b-values and from a small subvolume of the relevant mainshock fault. In the Puerto Rico case, we document here substantial issues with the homogeneity of the magnitude scale that in our assessment make a meaningful analysis of b-values impossible. We conclude that the evaluation by Dascher-Cousineau et al. (2020) is misrepresentative and a not a fair test of the FTLS hypothesis.