This research examines television programme rescheduling to maximise total ratings for one network across a week. Respondents read these re scheduled programme schedules and choose the network to watch during prime time. Only a small but informative fraction of over seven million potential programme rearrangements are selected for the experiment to get meaningful information without introducing respondent fatigue. Models are used to predict ratings as a function of the Programmes offered at a particular time. The next stage uses the models to predict ratings for all the possible programme schedules and determine which rearrangement maximises the total weekly ratings for one network. For the data used in this study, the proposed optimal television scheduling method increased the predicted total weekly ratings during prime time for one network by 24% while average channel share increased by 19%. To validate the method, comparisons between actual and predicted ratings are made for a network that actually rescheduled some of its programmes.