Fade slope analysis is important for the deployment and operation of satellite systems at high frequencies such as Ka/Q bands and above. In order to mitigate the attenuation of the signal due to atmospheric impairments, rain in particular, such systems have to implement appropriate fade mitigation techniques (FMT). The determination of the required tracking speed of FMT for following the signal variation and to enable the short-term prediction of the propagation conditions, is dependent on fade slopes. In order to characterize and model the fade slope statistics, long periods of measurement data are needed. In this paper, fade slope statistical distributions are presented, derived from the excess attenuation of three years' worth beacon measurements from the Alphasat satellite at 19.7 GHz and 39.4 GHz in Ljubljana, Slovenia, and compared to the fade slopes predicted by the ITU-R P.1623-1 model.