I suppose that depends on what you're striking him with. A wielder of a pick or mace designed to injure an opponent in heavy armour might well expect decent results with a solid blow, but you could bang your sword (especially a sword wielded in one hand) all day on a man in full plate, and still not do him any appreciable harm.
In the real world, yes. D&D is not trying to simulate the real world, so none of this applies. If you're looking for accurate representations of medieval warfare, D&D 4E is not remotely the game for you, even before you add in the wizards and dragons and such.
And that being so, I argue that the character in heavy armour should still have a higher AC than any high-dex character in no/light armour.
What about a guy in light/no armor with a high Dex who is blessed with the divine protection of a wrathful god? Pretty sure there weren't any of those at Bannockburn.
90% of the time the heavy armor guy
will have a better AC. What we're looking at here is an edge case where a player has deliberately gimped his ability to do his primary job to give himself a higher AC. The reason Avengers get a bigger AC boost when wearing light armor is simply because they aren't expected to max out their Dexterity. Even then, at most levels the difference is going to be miniscule: compare a 1st-level ACvenger (10 base + 3 hide armor + 3 armor of faith + 5 Dex = 21) to a paladin (10 base + 8 plate armor + 2 heavy shield = 20). Now remember that the paladin expended exactly
zero resources to get that AC, while the avenger has blown both of his starting feats
and 3/4 of his starting attribute points on an ability score that doesn't benefit him at all.
Also, bear in mind that the
only light-armor wearers who can match or even slightly exceed heavy armor ACs are those who have some form of innate magical defense: swordmage warding, armor of faith, etc. Show me a rogue who can, without magic, equal the AC of a fighter without magic, and then maybe there'll be a verisimilitude problem.