The state-of-the-art in the practice of probabilistic network planning for complex construction projects PERT (Program Evaluation Review Technique) is evaluated along the lines of thought of the reality test that was introduced in the workshop “Error in the Sciences”, Lorentz Center, Leiden University, 2011. The conclusion is that the application of the PERT-methodology during project execution can be improved by: including in the planning software the actual durations of the activities as soon as these are finished; monitoring progress by keeping the probability of timely completion just over 50%; and adopting risk ranking for prioritising managerial attention.