I don't find this things unrealistic. AC glosses over details, but isn't terribly unrealistic in my opinion (just simple).
To me, combining the concepts of dodging and blocking is actually unrealistic, as opposed to many other rules that are indeed questions of detail versus abstraction.
For example, a well-armored fighter might have an AC of 30 and a touch AC of 12. A huge cloud giant power attacks and gets a 25 attack roll. Clearly, he misses. But by the rules, the attack does physically touch the fighter. It's rather hard to imagine how a huge-sized creature could take a full swing and have his blow be rendered meaningless by armor. No matter how good the armor is, simple intertia suggests the fighter should be thrown across the room.
Whereas if the same giant attacks a halfling monk with an AC of 30 and the same touch AC, the rules dictate that a 25 misses completely. This is entirely reasonable; the giant could miss a creature that is the size of a bug to him, and a true miss would deal no damage no matter how hard he swung.
But by the rules, both characters have the same AC. Don't get me wrong, that mechanic is a D&D-ism, and it many less extreme cases it works well enough that people are okay with it. But realistic it ain't. Its virtue (as with hp) is simplicity.
Spell slots reflect the phsyics of how magic actually works in the setting. Granted a lot of people take issue with those physics, but the mechanic is connected to the flavor of how spell use operates in the game. The reason people crtiicize it, is this is not usually how magic works in movies and books (normally, but not always, it is closer to something like magic points). But that doesn't mean taking the vancian approach that D&D did is less realistic than how it was done in other games.
Of course realism isn't the issue with magic. But I think the Vancian style doesn't match with what newbies expect from a magic system, so my case is that if we were writing D&D from scratch, we'd write something simpler, more flexible, and more generic.
Understand when people complain about mechanics disrupting immersion on believability grounds they are not usually talking about a lack of detailed realism, or calling for complexity. A lot of the time they are talking about things which create a gap between what their character is doing and the mechanic. This is going to be subjective to a degree. But something like Vancian magic doesn't cause a problem for most of them, because there is a 1-1 relationship between what you are doing and what your character is doing and a connection between the mechanic and the flavor.
Indeed. What I was saying is that Vancian magic (among other things) was not a great approach to begin with, not because of anything immersion related, but because it's too complicated, too hard to balance, and doesn't model fiction well. But it's a D&D-ism, and we're used to it. So when a new game comes along and presents what is supposedly the same class, but which doesn't have the mechanics we're used to, it's hard to accept.
This would be true regardless of whether the new mechanic was better, worse, or not clearly either.
So my belief about 5e is that if they're going to kill a sacred cow and piss off some of the fanbase, they should at least create a system that has merits that the previous approach didn't have.
I can also be a question of volume. Just because there are mechanics in the game that are less realistic, that isn't a license to completely throw realism out of the window.
The goal of virtually any fiction is to create an emotional response in the audience. I think it's entirely possible to create a sense of immersion and involvement without being 100% realistic.
The Nolan Batman movies are a wonderful example of this. We get a few shots of him prepping batgear and talking about where it comes from. Is the gear really technically feasible? No. Is the scenario by which he gets it and keeps it secret plausible? No. So in the next movie, one guy finds out his secret. Is he really the only one who would figure it out? No. The key here is that some issues that speak to reality are assertively used and competently executed.
The same mentality works for D&D. The game merely needs to provide tools to acknowledge a variety of realistic considerations, and let DMs choose when it's worth implementing them and how to do so for maximum impact.