Table 4.

Guidance on when to rely on predictions (Pielke et al. 2000)

When to rely on predictions:When not to rely on predictions:
• Predictive skill is known• Skill is low or unknown
• Decision-makers have experience with understanding and using predictions• Little experience exists with using the predictions or with the phenomena in question
• The characteristic time of the predicted event is short• The characteristic time is long
• There are limited alternatives• Alternatives are available
• The outcomes of various courses of action are understood in terms of well-constrained uncertainties (i.e. the likelihood of false positives and false negatives)• The outcomes of alternative decisions are highly uncertain