Run-off road and head-on crashes together constitute around 38% of all casualty crashes and a higher proportion (closer to 50%) of all fatality crashes in Queensland, Australia. These statistics are a fair reflection of the national condition. Vehicles leaving the travelled way are a significant contributor to Australian road trauma. The Australian National Road Safety Strategy proposes a number of infrastructure treatments for tackling these two crash types, including the use of an appropriate road safety barrier. Road authorities deploy longitudinal road safety barriers primarily to prevent errant vehicles from impacting with hazardous roadside objects that could cause an adverse outcome for either the occupants of the errant vehicle or third parties. However, road safety barriers are not equal and are differentiated in the first instance by their capacity to contain impacts of different speed, mass and angle of incidence. While roadway departure speeds and departure angles are well-addressed in contemporary academic literature and methodologies for road safety barrier selection, the mass-distribution of the in-service vehicle fleet is less well represented. This study proposes the use of data obtained from weigh-in-motion technology to represent the mass-frequency distribution of the in-service vehicle fleet. Combined with roadway departures conditions reported by others, a methodology is presented for calculating the likelihood of vehicle-barrier impact exceeding the road safety barrier capacities prescribed by the predominant global test protocols for road safety barriers. The methodology is used to consider how different roadway configurations and traffic compositions might influence the likelihood of barrier capacity exceedance. The results from modelling of various scenarios are reported. The results suggest that the relative likelihood of barrier capacity exceedance varies as a function of cross-sectional geometry as well as traffic composition, so suggesting that a “one-size fits all” approach to road safety barrier selection is not appropriate.