Ah, but that's name recognition at work. We see "Tom Brady" make a connection to the real player with that name and then we know (or at least assume with high probability) that he'll be awesome in the game.
Likewise with fictional players - "Tremendous Maxwell" doesn't mean anything to me. So I can't say "hey that guy plays for X team and had a good year so he's probably at least decent in the game." like I could with, say, Revis or D-R Cromartie.
Rookies, though, don't have pro history, by definition. We also know the college game doesn't necessarily translate to the pro game - so college history can't always be a true indicator. That's great on our end for those who want more mystery - but how do you do that on the developer end that's not cheap, cheesy, done for just the sake of "making it harder" (i.e. too many guys flop after having quality college numbers), and yet won't have us coming back complaining about how easy it is to min-max/game the system and ignore the depth?
Just to take one of the most basic measurables - the 40yd dash. Let's say a player ran a 4.6 40. That could be because he's on the slow side, but still is explosive and reaches his top speed in a few steps (which could still make him an effective HB if he has power, vision, etc) or it could be because he's fast, but takes too long to reach top speed (not explosive, not aggressive enough in his running style, too tense, etc).
Both are 4.6 for different reason, but one is useful, the other might just be a KR specialist. That's just one measurable - how would we want EA to do this? All 4.6 guys are just slow? Would the game use another measurable to decide which of these guys to generate?
Fun indeed - brings up another question - who all of these numbers (college stats, combine results, pro day results) all be determined almost at random (since the game likely won't actually have these things and HC used it just to "unlock" ratings as time goes on) and just may or may not match with the hidden ratings? Would the hidden ratings "roll up" these results with some fudge factor (say 95 SPD is a base 4.3 40 time, but he could run at 4.6 at combine and 4.2 at pro day)?