Look at the author's own note on forecast accuracy, its less than 15% at the maximum. Does any one want to bet their money on such a small number ?
Combine this with the 'revisions' over the years, I have a sincere doubt about the reliability of this program.