Two types of measures describe the accuracy of estimation models. The first deals with the characteristics of a set of estimates and uses statistics to describe the population. The second deals with the time history of estimates. You can apply these two types to any estimated quantity. All you need is a set of paired values. Each pair consists of the estimated value, E, and the actual (true) value, A. For the first type (set of values), these are the estimated and true values. For the second type (time series), these are the estimated value and the time that it was estimated, plus the (final) true value.
The residual error, R, is defined as
Ri = Ai – Ei
where i denotes ...