Occupancy information is commonly regarded as an important ingredient for energy efficient building climate control. In a recent study of the authors, the energy savings potential of using long-term occupancy information (in the range of days) in the control of office buildings was investigated, assuming a geometric distribution of occupancy and vacancy intervals. In this work, the findings of the previous study are validated by considering real occupancy data from an office building in Switzerland. The study is based on a Model Predictive Control (MPC) framework, which enables a direct comparability between different controllers. Assuming a perfect model, MPC can be thought of as an upper bound on the performance of any controller and hence directly yields the value of occupancy information. The study compares the energy use of an MPC controller having three variants of occupancy information but the same occupancy realization. The variants are: (a) the use of a standard fixed occupancy schedule; (b) the use of the schedule, but adjustment of lighting and ventilation based on instantaneous occupancy measurements; and (c) the use of a perfect prediction of occupancy.