The paper is a short version of the plenary lecture at XV International School-Conference “Problems of Geocosmos” hold in Peterhof, April 22–26, 2024. The lecture was aimed at an audience with broad scope of interests in Earth sciences, but macroseismic studies and historical earthquakes are not exactly professional research field of the conference attendees. Therefore, possibly some points in the paper are too general. Though for seismic hazard assessment long-term parametric earthquake catalogue covering also historical times is needed, transferring macroseismic information into earthquake parameters may be not justified by the intensity scale properties, which is the basis of such transferring. I am not calling to throw away historical earthquake catalogues because their theoretical background has certain drawbacks; we just ask to be aware on “deep waters” behind such catalogues. On contrary to quantitative macroseismology, the classical one mostly works on finding original documentation of earthquake effects and on verifying its reliability. Some of problems on this way are illustrated in the presentation by addressing to studies of historical earthquakes in Eastern Siberia. An analysis of historical and cultural context of the region give an idea what can be expected in terms of completeness and homogeneity of earthquake history.