The overall rating I will agree is about as stupid of a rating as one could have. I actually kinda liked how head coach did it, especially since you need some type of overall for the CPU teams to use to judge their players. I would like to see this hidden from users though.
As far as me saying "ratings are opinions" this is one of the big problems with ratings in any sports game. It really comes into play with things like progression and leads to crazy things like people wanting stat based progression. I for one realize that once you are in the game the ratings are actual abilities. However, what I was saying is those ratings are determined by opinions same as a learning ratings would be.
I do see the problem with a learning rating giving too much knowledge to a player in terms of how good a player can become the same way as potential can or any development based rating. I have actually described a much more realistic way of doing the ratings that would solve a lot of those problems.
The basic idea is each of a players abilities is shown as a range of possible ratings and as a coach scouts that player more, plays against that player, or has that player on his team the ranges would become more and more accurate zeroing in on the actual rating.
Physical ratings would be zeroed down pretty quickly especially when scouting like at the combine or when having the player on your team. Skill and knowledge ratings would be zeroed in on at a slower rate, and potential, learning, and other development type ratings will be zeroed in on very slowly taking probably a couple years with the player on your team to get an accurate reading on them.
That way, when you draft a guy you would know pretty much just his physical abilities and maybe have an idea of his skills. His potential, while you might know a little, you won't know much. If the player is actually playing after a few weeks, probably by mid season, you will probably know pretty much how good he is. If he isn't playing it would take probably a couple years of practice to learn that. His potential and learning however would not be probably known until he has played for 2 or 3 seasons. If he isn't playing you might never get zeroed in on his potential.
With a system like that you would actually have to guess some on free agents or stick to guys you have seen play in games against you. Top players you might learn more about from news somewhat like those of us in here know about the top players. Just through general knowledge you might learn about some guys. Also, you have stats and awards to go off of. Still, there is some real guess work that could go into picking up players and this could make trades a little tougher to pull off to make the system more realistic.
What needs to happen here is there should be an actual potential rating that actually caps potential and shows the player's ceiling. However, this is not what should be shown. If they are not going to go with a system like I described earlier they should at least make some room for error in the shown potential rating.