[Caveman24]

Quote: Originally Posted by CM Hooe

Not super-hard in concept. How I'd do it - which may not be the best way, but is a way - for each position group, set a mean, standard deviation, min, and max for each attribute (possibly by finding this data from past Madden rookie classes created by Donny Moore) and then for each rating find the inverse normal distribution based on an intelligently-generated random input. By "intelligently-generated random input", I mean that the entire process shouldn't be truly random. A wide receiver's SPD and ACC should remain relatively close together, a 4-3 OLB should tend faster / smaller / better coverage ratings / worse pass rushing ratings than a 3-4 OLB, and so on. I'm not sure if this is how it was done in the past, as I obviously don't have access to Madden's source code. In the past, however, the problem was past Madden games, however they did it, did not intelligently randomly generate players. It resulted in some really weird player builds. Between that and a complete lack of control over how players progressed - the progression logic was one-size-fits-all for all players at a given position - long-term roster management was pretty difficult. Thus why they went to hand-made draft classes instead.

The progression system what was exactly wrong with it? Was it 7th rounders being as good as the 1st rounders?