The fighter depends more on his inherent traits than the rogue or magic-user. Encounter an enemy? Use your Strength to hit him, Dexterity to avoid being hit, and Constitution to make it hurt less when you do get hit. Encounter a locked door? Roll Strength to break it down, or to climb around it, or roll Intelligence to find a key.
The rogue depends on specific skills. Enemy? Stealth to avoid him, or Diplomacy to make him non-hostile, or Sleight of Hand to tie the lanyard on his sword to something so he can't swing it. Locked door? Lock-picking, or Bluff to convince someone to let you through.
Wizards are based on neither core abilities nor skills. Instead they have... What? Spell slots? But they still have skills and core traits, while neither fighters nor rogues have slots. Plus spell slots become exhausted over the course of a day, while skills and abilities can be used over and over.
So this theory is incomplete. But even though it's incomplete, we can take several things away from it: first that it's okay to give the rogue more skill proficiencies/skill ranks than other classes, plus shit like 5e's expertise--which we already knew, but using this theory to reaffirm it, helps prove the theory to be at least partly valid.
Another thing we can establish though, is that fighters should get more bonuses to their inherent abilities than rogues or magic users. 5e seems to be going down this road, giving the fighter a straight-up two bonus ASIs (and the barbarian his weird capstone, plus allowing barbarians, paladins, and monks new ways to use their ability scores, plus fighting styles giving flat bonuses to attack and damage in defiance of the modifier/proficiency/advantage guidelines).