Fitting with the efficiency uncertainty without template morphing

[SphericalCowww]

I just thought I'd mention the idea here. The implementation of efficiency on a spectrum is different from that of bias and smearing, in that it does not need rebinning. For this reason, if the efficiency is interpolated for a bin size, we would only need the mean and +/- sigma for the efficiency, then use a single parameter to A to shift the efficiency to (mean + A*sigma). I imagine this can be implemented like the typical amplitude scaling factor, but I am not sure how easy/difficult it would be.

[kdund]

In my opinion the best way to implement this would be allowing some variables to be functions of other variables-- that way the signal_rate_multiplier could be the product of cross_section and efficiency