Unmanned aerial vehicles (UAVs) are well-suited for various short-distance missions in urban environments. However, the path planner of such UAV is constantly challenged with the choice between avoiding obstacles horizontally or vertically. If the path planner relies on sensor information only, i.e. the path planner is a local planner, usually predefined manoeuvres or preferences are used to find a possible way. However, this method is stiff and inflexible. This work proposes a probabilistic decision-maker to set the control parameters of a classic local path planner during a flight mission. The decision-maker defines whether performing horizontal or vertical avoidance is preferable based on the probability of encountering a given number of obstacles. Here, the decision-maker considers predictions of possible future avoidance manoeuvres. It also defines an ideal flight altitude based on the probability of encountering obstacles. This work analyses the building height of all European capital cities and the probability of encountering obstacles at different altitudes to feed the decision-maker. We tested the feasibility of the proposed decision-maker with the 3DVFH*, a commonly used local path planner, in multiple simulations. The proposed probabilistic decision-maker allows the local path planner to reach the goal point significantly more often than the standard version of the 3DVFH*.