This paper traces the development of Predictive Models for establishing the coverage and frequency of viewing Advertising campaigns on Television. Most of these models were designed with the need for accuracy taking precedence over the need for flexibility, with a result that it was necessary to take into account many short term variable factors. In 1971 JWT finished work on their 'PREFACE' model by using the Negative Binomial Distribution predicting the full frequency patterns using the sole criterion of total ratings. The paper then proceeds to demonstrate how this simple concept is applicable in many other countries - irrespective of the method of time buying, the number of channels and the availability of advertising. This leads the authors to suggest that in order to use such predictive formulae in a country it may no longer be necessary to incur the considerable expenditure of continuous panel and meter data collection.