Probabilistic methods can provide more insight on risk allocation and can be used in geotechnical baseline reports (GBR). A novel, probabilistic approach of the spatial distribu-tion of karstic cavities is presented to assess the risk of liner instabilities when tunnelling through karstic limestone. The presented research work is based on the following geotechnical baseline statement: “There is a 10% chance of encountering a sediment-filled cavity, with a maximum size of 2m3”, which was assumed to be correct for the analysis. A Matlab script gen-erates a random distribution of karst cavities, modelled as a circular shape, which is transferred into FEM 2D and 3D numerical analysis. The measured variable is the index for capacity utili-zation of linings in tunnels (CULT-I), which act as proxy for liner stability assessment. The methodology presented provides a framework for further investigations to assess the risk from karst cavities to tunnelling works.