Within intelligence systems, it is important to assess the probability that certain events will occur and to determine the reliability of information on them. Any degree of uncertainty over these events or the information can be quantified into numerical probabilities, which can then be readily and accurately communicated within the system. Four advantages stem from using numerical probabilities: (1) they can be given to a decisionmaker in a variety of forms (e.g., charts, graphs, tables); (2) they are more precise and less wordy than verbal equivalents; (3) they can be scored; (4) they focus attention on confirmable events, which can be definitely judged as true or false, given that all the relevant facts are known. For implementation, it would be required that people be motivated and trained to express degrees of certainty and uncertainty in terms of numerical probabilities.
Brown, Thomas A. and Emir H. Shuford, Quantifying Uncertainty into Numerical Probabilities for the Reporting of Intelligence. Santa Monica, CA: RAND Corporation, 1973. https://www.rand.org/pubs/reports/R1185.html.
Brown, Thomas A. and Emir H. Shuford, Quantifying Uncertainty into Numerical Probabilities for the Reporting of Intelligence, RAND Corporation, R-1185-ARPA, 1973. As of November 30, 2022: https://www.rand.org/pubs/reports/R1185.html