Optimal assessment tools should measure in a limited time the knowledge of students in a correct and unbiased way. A method for automating the scoring is multiple choice scoring. This article compares scoring methods from a probabilistic point of view by modelling the probability to pass: the number right scoring, the initial correction (IC) and the negative marking (NM) method. We will compare the probabilities for students to pass when their assessment is translated into a score by means of the NM and the IC method. Moreover, given a knowledge level of the student, the variance of this probability will be discussed for both methods.