Comparative forecasting: Hindsighting the Czechoslovakian invasion

Two dissimilar information-processing techniques may be used to forecast the same phenomenon. Event-pattern recognition, objective in nature, uses predetermined indicators and is dependent upon selected data flows; the Bayesian method rests more heavily upon subjective interpretation. These two methods may be used to forecast a variety of phenomena—here they are retrospectively tested for their ability to forecast the August 1968 Soviet invasion of Czechoslovakia. Objective and subjective methods may be used to generate combined, or at least cross-validated, forecasts; and the author suggests how—and at what costs—comparative forecasting may be used productively for international political analysis.