Due to the safety-critical nature of intelligent crossroads, ensuring communication reliability between vehicles and the corresponding road side unit (RSU) is of utmost importance. This requires knowledge of the maximum number of vehicles in the system to be able to assess interference. However, due to the open-ended nature of the application, i.e., vehicles can enter and leave at arbitrary points in time, it becomes difficult and inefficient to perform an analysis based on deterministic methods. For this reason, this paper instead proposes a probabilistic technique to estimate the maximum vehicle count at crossroads. To this end, we investigate how different factors like traffic protocol, vehicle density, and vehicle length influence probabilistic estimates of the maximum vehicle count. We then highlight how these estimates impact communication reliability by deriving guarantees towards packet loss in an exemplary crossroad vehicular ad-hoc network (VANET). As shown in a detailed case study and simulations using OMNeT++, pessimism and overdesign can significantly be reduced compared to deterministic approaches, while still maintaining high safety levels.