Context. In a previous paper we reported a planetary companion to the young and very active K5Ve star BD+20 1790. We found that this star has a high level of stellar activity (log R HK = −3.7) that manifests in a plethora of phenomena (starspots, prominences, plages, large flares). Based on a careful study of these activity features and a deep discussion and analysis of the effects of the stellar activity on the radial velocity measurements, we demonstrated that the presence of a planet provided the best explanation for the radial velocity variations and all the peculiarities of this star. The orbital solution resulted in a close-in massive planet with a period of 7.78 days. However, a paper by Figueira et al. (2010, A&A, 513, L8) questioned the evidence for the planetary companion. Aims. This paper aims to more rigorously assess the nature of the radial velocity measurements with an expanded data set and new methods of analysis. Methods. We have employed Bayesian methods to simultaneously analyse the radial velocity and activity measurements based on a combined data set that includes new and previously published observations. Results. We conclude that the Bayesian analysis and the new activity study support the presence of a planetary companion to BD+20 1790. A new orbital solution is presented, after removing the two main contributions of stellar jitter, one that varies with the photometric period (2.8 days) and another that varies with the synodic period of the star-planet system (4.36 days). We present a new method to determine these jitter components, considering them as second and third signals in the system. A discussion on possible star-planet interaction is included, based on the Bayesian analysis of the activity indices, which indicates that they modulate with the synodic period. We propose two different sources for flare events in this system: one related to the geometry of the system and the relative movement of the star and planet, and a second one purely stochastic source that is related to the evolution of stellar surface active regions. Also, we observe for the first time the magnetic field of the star, from spectropolarimetric data.