STATISTICS as a scientific method goes back to the seventeenth century when William Petty and others computed mortality ratios on the basis of church records of births and deaths. This "political arithmetic" as it was then called became a method for discovering statistical regularities in human affairs. It was used for such practical purposes as the computation of life annuity tables. It was also used by theologians, as Peter Suessmilch, to prove the wisdom of God by demonstrating the astonishing regularities in the creation. Statistics as a scientific method and academic discipline has spread in these two hundred or two hundred and fifty years of history from application to political and social affairs to all fields of human endeavor. There is today no field of inquiry or art in which statistics as a method does not find some use or some occasional misuse. Every use of statistics assumes that certain magnitudes or relationships, found to exist at the time when the counting took place, will continue to exist. Therefore, the use of statistics is based on the assumption of some continuing regularity in human affairs. But how can we assume that any magnitudes or relationships ascertained for the past have any validity for the present or the future in a world in which explosive changes appear to drown out what regularity there may be? Rapid change and political turmoil appear to make questionable any conclusions for the future drawn from measurements of the past. In this paper I will deal with this problem in the field of economic statistics and, more specifically, in the field of business cycle statistics. Here is a question of great importance for modern life. Next to problems of individual freedom and responsibility and foreign policy there is hardly a question more important than that of maintaining the continued growth and stability of the economy with a minimum of government regulation. The importance of these objectives was officially recognized in the Employment Act of I946. The determination of policies designed to promote economic growth and stability depends largely on our ability to appraise the economic outlook. The economic outlook cannot be appraised simply by looking at the most recent statistics alone. Recent statistics show, for instance, that business activity has been contracting, and that the contraction has slowed down in April or May of this year (I954). These facts themselves do not tell us whether, henceforth, economic activity is likely to expand again or to continue to move sidewise or whether the contraction has only slowed down temporarily and is likely to continue later this year or next. The question is: How can we derive from statistics of the past conclusions for the future? Inventory oscillations of a few months' periodicity move around a cyclical curve of several years' periodicity which in turn moves along a secular trend line. For an effective anti-cyclical policy we need to know where we stand in the cycle and on the trend. We look to business cycle analysis and business cycle statistics for an appraisal of probabilities in economic development this is the minimum we need to have as guidance for a policy of economic stabilization.