Prediction, Butterfly Effect, and Decision-Making

Abstract

Prediction has become an integral part of management. Managerial decisions cannot be done without them. Any investment decision of relevance requires forecasting of crucial parameters, like interest rates and demand. Any spending on research and development requires an estimate of future needs. Or so it is believed. Yet all scientific evidence points in the same direction: long-term, science-based predictions are impossible in both, natural science and social science. Of course, “long-term” means different time spans for different systems depending on the speed of change. The speed of deviation from a perfect cycle of the orbits of sun, Earth, and moon together is so slow that long-term means thousands of years. Climate change and its impact on weather and on sea level are much faster and long-term means a couple of decades at most. Weather and an epidemic like COVID-19 change so rapidly that long-term means just a couple of days. Thus, forecasting cannot be a prerequisite of good decisions anymore. Rather it is the other way round. Understanding a decision’s impact on a system’s future behavior enables good decisions.