I understand the real world logic perspective but there are issues with that.
In real life we deal with unknowns. But there is no unknown in Madden. So questions like will a guy be able to recover from injury, or is a guy just in a slump or is he regressing (e.g. Thomas and V. Davis) mean nothing in Madden. I don't think there is a quick fix for this. In other threads there is talk of hiding/blurring ratings. That could make roster decisions more lifelike, but I doubt EA is willing to do that.
With regards to the hoarding of positions, there are multiple things contributing to this, yet the good news is that quick fixes would work.
1. Seems there is still a under the hood importance rating that each team has for each position. EA foolishly sets it high for positions where a team has a good player. So FS is marked high importance for the Seahawks, DE is high for the Texans, etc. That is dumb because now those teams think it is more important to upgrade those positions.
2. Deciding whether or not to upgrade a position is constant. May seem like it contradicts #1, but it doesn't. Lets say Seattle did not sign Smith and somehow still got rid of Thomas. So their current FS could be a 75 ovr. Then imagine a 80 ovr FS is available. So they upgrade the FS position by acquiring that guy. That makes sense right? Going from 75 to 80 is a solid move (depending on age, more on that later). The problem is that the 5 point difference is constant at all levels, so just as going from 75 to 80, teams will try to go from 85 to 90, or even worse 90-95. So when you pair this with #1, Seattle saw that 99 ovr Harrison SMith was available. Smith is an upgrade over Earl Thomas, and because that position is flagged high importance, Seattle though it was the right move. That is wrong.
3. Drafting in bunches. This really frustrates me more than the above. If a team has a need for a position going into the draft, you can almost be certain that they will draft multiple guys at that position. I've seen teams draft 3 WRs, 3 CBs, etc. This is caused by one of two things:
a)Team needs don't update throughout the draft. E.g. teams "forget" they spent a 1st round pick on WR so they use a 2nd round on that position as well
b)Needs do update, but since we immediately know the type of player we draft we immediately know whether or not the need is filled. So lets say Baltimore needs a WR and drafts one in the 1st round. Well turns out he is a bust at 72 overall so they still need a WR. So they draft another in the second. So basically what we see is teams drafting WRs until they find one they like (Matt Millen syndrome)
4. Not sure how much age factors into AI roster decisions. Here is what the Pats did as an example:
a)Brady retired, need a QB
b)drafted one in the 1st round. ~73 ovr. Well that is no good
c)drafted one in the 2nd round. 77 ovr. That's good right?
D)Wrong. Sign Fitzpatrick. 78 ovr
e)Start Fitzpatrick all year. Rookie(s) don't develop. Nor does Fitzpatrick
f)Due to high QB importance for the Pats, next offseason they want to upgrade the QB position.
g)go back to b and repeat.
Teams don't make much of an effort to develop rookies. If he is not good enough to start the day he is drafted, the rookie pretty much has no career unless the starter gets injured.
I don't think it ruins CFM though. I'm still having fun with it. Half of the time the AI happens to draft a guy they like, so they stop. But it does seem like each team has one position that is troublesome.