*In scoring a forecast, one of two purposes should be considered: either to determine the utility of the forecast or to determine the skill of the forecaster. If the purpose is utility, then the score for each forecast should be directly proportional to the value of the forecast in meeting specified operational requirements. If skill is the purpose, then the total score for a series of forecasts should be a reflection of the forecaster's ability to analyze and classify a weather situation for forecasting purposes; within a well-defined class of antecedent weather the probability of a subsequent event, like rain, is increased above (or decreased below) the relative frequency of that event in the total of all weather situations, as established in a long period of time.
[1]
C S Peirce,et al.
The numerical measure of the success of predictions.
,
1884,
Science.
[2]
P. Galle.
Étude critique sur la méthode de prévision du temps de Guilbert
,
.
[3]
R. H. Muller,et al.
Verification of Short-Range Weather Forecasts (A Survey of the Literature)† I.
,
1944
.
[4]
F. Green.
Meteorological Forecasts for Special Purposes
,
1946,
Nature.
[5]
Irving I. Gringorten,et al.
Forecasting by Statistical Inferences.
,
1950
.
[6]
G. Brier.
VERIFICATION OF FORECASTS EXPRESSED IN TERMS OF PROBABILITY
,
1950
.
[7]
Glenn W. Brier,et al.
Verification of Weather Forecasts
,
1951
.