Current analytics seem to be strictly based on data from past play. So when this data is used as a tool for deciding which players to draft, it by definition FAILS TO ACCOUNT for drivers of FUTURE success, including a player's work ethic, desire, development potential/aptitude, demeanor and surrounding resources/environment.
Josh Allen's recent play appears to be revealing the limitations of current analytic methods.
Obviously, you would have to do a lot more extensive psychological testing to begin addressing these limitations.