The best way to look at predictive value of any non-interpretive metric is to go back on a large number of stocks and see if holds true in the large majority of cases. You'll find that "fill the gap" is nonsense.