Scouts and staff aren't meant to be perfect judges of players, even with 20/20 in the relevant attributes. Their perceptions of players change over time and are impacted, to an extent, by player performances.
In this case its very likely that you'll be getting unrealistic outputs, based on unrealistic inputs. The game systems aren't designed to accommodate such a severe change in a player, as it cannot organically arise during the game. From our side, the question would be, how do we go about testing any changes and judging whether they're truly accurate or not given it will never happen in one of our internal tests?