The paper presents the results from a simulation model made for a process of making bicycles, comparing the difference in process performance measurements for deterministic time operating times are used in the first instance and probabilistic times in a second stage. The number of units produced, average production time per unit of output and utilization rates for three areas of assembly to vary the nature of the operation times as controlled factors having other elements are set. In each of the proposed scenarios 20 repetitions are performed, resulting in a production of 35 units in the first instance, while considering variables times the number of units obtained varies between 32 and 37; the other two performance measures analyzed have similar behaviors, presenting a specific value in the first stage and a range of results in the second. Based on these results we conclude that in the simulation of processes for decision-making, requires a proper use of statistics in all phases, so that the model considers precisely the natural variations of the production processes.