Changes in mobility influence road safety. The effects of safety measures may even be overshadowed by the effects of temporary mobility fluctuations. Usually mobility is corrected for by defining risk as the ratio between fatalities and mobility. Due to lack of sufficient data, mobility is often approximated by car mobility. In this paper we will show that the resulting "general" risk is misleading. Stratification by traffic mode shows that the risk for car drivers, motorcyclists and truck drivers is roughly constant between 1950 and 1970 and that it has been decreasing exponentially afterwards. This contradicts the development of general risk, which has been decreasing the whole period between 1950 and today. Further stratification shows that stratification by traffic mode of the fatality alone is still insufficient. Changes in mobility of the other party in a crash are also important. The development of risk of car-car accidents differs significantly from that of single vehicle car accidents. A comparison between single vehicle motorcycle accidents and motorcycle-car accidents shows a similar discrepancy. Stratification of mobility by traffic mode, and of fatality data by the relevant traffic modes involved, can enhance the understanding of the influence of mobility on safety.