Same as Fig. 2, but for GJ 436. Panels a and b show that the CARMENES, HARPS and HIRES data sets used for the construction of the best fit are fully consistent with a planetary companion with a period of 2.64 d. In panel a: the 1.60 d GLS peak is a 1-day alias of the 2.64 d periodicity induced by the planetary companion. Panel c: shows the GLS periodogram of the best fit residuals, which yields several peaks above the FAP = 0.1%, most likely due to the observational window, stellar activity and their aliases.