Objective-statistical forecasting instruments, according to the classification criterion “context of justification”, form a further essential group of forecasting instruments in addition to the subjective-intuitive forecasting instruments and are used to set up limited-general period forecasts (progress forecasts). A functional and possibly also causally interpreted relationship between dependent and independent model variables is assumed. The prediction of event values of future points in time depends on the forecast result values of the previous periods for this or another type of event.
These interdependencies can lead to partial compensation and accumulation of forecast errors over time. The methods belonging to the objective statistical forecasting instruments (Product life cycle-Concept, GAP analysis, Cross-sectional analysis, indicator method, dwell time distribution, simulation models) are in some cases able to generate the forecast of an event interval and thus increase the overall probability of an event occurring, which decreases as the forecast period increases. In some cases, however, the forecasting instruments do not allow any statements to be made about empirical event probabilities.
Based on the validity of the time stability hypothesis, according to which the regularities of the past can essentially also be assumed for the future (empirical regularity) and the functional, possibly also causal, relationship between the dependent (endogenous) and the independent (exogenous) model variables , the objective statistical forecasting instruments can be classified as follows:
⦁ Forecasting instruments that only show a functional relationship between the event value (not also the event type) and the time as an independent variable,
⦁ forecasting instruments that have no functional-causal relationship with other types of events,
⦁ Forecasting instruments that show a functional and possibly also a causal relationship between different types of events (on the basis of hypotheses) and
⦁ Forecasting instruments with invariant transition distribution that determine the temporal-quantitative transition between event types on the basis of random samples.