Limited Preemptive Scheduling is an attractive paradigm that enables controlling preemption related overheads, by appropriate preemption point selection. The selection of preemption points is essential to ensure schedulability and the associated analysis accounts for upper bounded preemption overheads, thus introducing a potentially high level of pessimism in the results. In this paper we propose a probabilistic distribution model of preemption related overhead and an accompanying method for preemption point selection based on quantiles, which provides controllable probabilistic relaxations. An experimental evaluation demonstrates the improvement of the extent to which this new approach facilitates finding solutions to the preemption point selection problem.