I'm sure extension necessarily causes fastballs to be "perceived" as having higher velocity or not (in all or most cases), but it could certainly be a contributing factor to his success.
Folks are assuming batters don't start tracking a ball until it leaves the pitcher's hand, which is probably not completely accurate in all (or most) cases. What if batters track/time a pitch much earlier than the time of release? If that's the case, then where the pitcher releases the ball would have very little impact on "perceived" velocity.
Does work exist that shows the effects of perceived velocity vs actual velocity?