Well as simulation, AD&D was pretty bad. But as a role playing game it was fairly good.
Realism is an illusion in tabletop gaming. What produces that illusion is having to make choices that have consequences that play out. There's a certain *rhythm* to a game that's working well. It goes like this: decision (attack the creature), immediate result (creature is not surprised), string of action rounds, second decision (run away), result (party gets through the door) then problem (how to secure the door?).
Adding detail to a system in terms of a broader selection of alternatives at each point does add something to the game, but until you master all that detail it bogs the rhythm of the game down. Later editions of D&D seemed to me to be fine for people who'd played continually since the original AD&D, but bogged down the game for people who wanted to play casually or were coming to it new. I think from a *design* standpoint the subsequent changes narrowed the appeal of the game.
That's not to say I'm against making things more complex. For example played under house rules that added a decision after the initiative role; you could take the initiative or you could cede it for a bonus on a counter hit. It didn't slow the subjective pace of the game because it was a simple decision with immediate consequences.
