How does Sherpa apply the BACKSCAL value recorded in the header of a spectrum file?

Like XSPEC, Sherpa recognizes the BACKSCAL header keyword in a spectrum file as having the definition put forth by HEASARC:

BACKSCAL = (THE DETECTOR AREA FROM WHICH THE SPECTRUM IS EXTRACTED) / (TOTAL DETECTOR AREA)

Sherpa uses the backscale value in creating a background-subtracted source spectrum:

Background_subtracted_spectrum (cts/s) = SOURCE_COUNTS/SOURCE_EXPOSURE - BGD_COUNTS/BGD_EXPOSURE/BGD_BACKSCAL*SOURCE_BACKSCAL ,

as well as in computing χ² statistics. Also, when source and background spectra are simultaneously modeled in Sherpa, the backscale value is used in calculating a fraction of the background model-predicted counts that contribute to the source model-predicted counts.

Note that Sherpa provides functionality for setting and retrieving a backscale value within a Sherpa session, with the set_backscal and get_backscal commands.