Withhold a calibration that is an identity rather than an estimate

Scoring an unpenalized maximum-likelihood fit on the very rows that produced it forces the calibration intercept to zero and the slope to one: they are the fit's own score equations rearranged, not evidence about calibration. Printed in a column beside a held-out slope they read as the model being perfectly calibrated, which is the one thing they cannot show.

Usage

mm_withhold_identity_calibration(row, tol = 1e-06)

Arguments

row

A one-row metrics frame from mm_metrics().

tol

How close to 0 and 1 counts as the identity.

Value

The row, with the calibration withheld and the reason appended when it is an identity.

Details

The identity is detected rather than assumed, so a fitting method that does not force it, a penalized fit for instance, has its apparent calibration reported as the estimate it would then be.