The Consequences of Assuming m=0 for Global Model-Fitting