On the veracity and explanatory value of two-systems models