ben_calgary said:Following this thread I went back and re-read the extension data properly.
This is where I part company with the consensus forming above. The objection nobody has made: this only works if the measurement is doing what we assume it is doing, and that assumption has not been tested here.