Heuristics associated with forecasting chaotic events: a rare cognitive ability

Abstract

Chaos is a mathematical phenomenon in which seemingly random events are actually predictable by simple deterministic equations. Chaos has been identified in numerous situations requiring humans’ situation awareness, sense-making, and control. The management of chaos could be a rare skill, however, and the heuristics for doing so are not well understood. These hiatuses thus motivated a new theoretical issue in ergonomics science concerning the distribution of this ability across different chaotic attractors and some of the heuristics that might be used to forecast events. Untrained undergraduates (N = 147) forecasted number series from four chaotic attractors of varying levels of computational complexity. Performance was measured as the correlation between forecasted numbers and real numbers. Participants’ performance varied by type of attractor and whether the forecasts were made for one to four steps into the future. The less capable participants used moving averages strategies, whereas the best forecasters matched the real numbers more closely.