The error of an observed value is the deviation of the observed value from the (unobservable) true function value. Do NOT use this tag for SOFTWARE ERROR messages.
I am unhappy about the "unobservable". Y'all may have noticed that I have been working on tags related to point forecast accuracy. Most questions in this area also carry the error tag, and the "error" in question is the one between the point forecast and the eminently observable future realization.
I propose changing the tag wiki excerpt as follows:
The error of an estimate or prediction is the deviation between this value and the "true" value. The "true" value may be unobservable (as in estimating regression parameters), or observable (as in forecasting or predicting future or yet-unknown realizations). Do NOT use this tag for SOFTWARE ERROR messages.
In addition, the tag wiki currently reads as follows:
In statistics and optimization, statistical errors and residuals are two closely related and easily confused measures of the deviation of an observed value of an element of a statistical sample from its "theoretical value". The error of an observed value is the deviation of the observed value from the (unobservable) true function value, while the residual of an observed value is the difference between the observed value and the estimated function value.
The distinction is most important in regression analysis, where it leads to the concept of studentized residuals.
To be honest, I don't see how this can be salvaged to (a) give more information than the excerpt above, but at the same time (b) address estimating / predicting both unobservables and observables.
Therefore, I propose deleting the tag wiki altogether.
(I'm open to this confusion implying that we should have two different error tags, e.g., estimation-error and forecast-error or prediction-error. However, I don't really want to retag 443 questions.)
This is related: Error tag is ambiguous
What do you think? I didn't want to change a tag with 443 questions without a discussion.