Really? None of your hyper-intelligent wizards are at all curious about the way arcane magic, which they've dedicated their lives to studying, actually works?
They are. But I view magic in 4e(actually in every edition) as billions upon billions of possible variations based on any number of things. It's an art rather than a science. Moving your pinky finger an extra millimeter might make a Fireball double the radius but discovering that fact and having the skill to do it without blowing yourself up are things that are hard to do. Most Wizards stick to what they know and can rely on. So, although Wizards are curious about the abilities other Wizards have, it's like saying that I'm curious about how Quantum Mechanics works. I'd like to understand it completely and I took Physics courses in University. However, I don't want to spend the time and dedication it would take to fully understand it(which would probably take years).
I assume, like science, there are hundreds of different competing theories as to how the intricacies of magic work and while all of them seem to work, in theory, many are incompatible with one another. So, when most Wizards see other Wizards doing weird things, they say, "Weird, their magic must work under a different principle than mine. That's interesting but without one of them to teach me it and a couple of years to understand where it comes from, I guess it'll just be something they can do and something I can't."
Okay, maybe I missed something in the DMG somewhere about using the same Encounter power more than once during an encounter. I was pretty sure that the rule was "don't do it", but if you actually can and people just agree not to.... then I'm not sure what to make of that. I can't imagine that a wizard would choose to let someone die, rather than risk being boring by casting an impressive spell twice. (And un-impressive spells, reflected mechanically as At-will abilities, can be spammed indefinitely without being boring?)
It makes decent sense as a narrative construct, I'll grant, but from any sort of in-game explanation it just doesn't pass muster.
It doesn't try to pass muster as an in-game explanation. It doesn't WANT to. It ISN'T an in game explanation. It doesn't need one.
The rule is that you can't cast two of the same encounter power in the same combat. The in game rule is whatever fits your world best. Either they can only absorb so much energy before needing to rest(and some Wizards due to a quirk of birth or focus in a different discipline can do it faster) or the Wizard in question simply decides to use a different spell that round because he doesn't WANT to use the same spell again. Despite the player wanting to use the spell again.
Either way, the system is designed around at-wills being spammable but encouraging people NOT to use them unless they have no other choice.
The entire system is a metagame construct to create balance and interesting combats within the game. Because, you'll find that any in game explanation of why things work the way they do will be abuseable by players who will look at the rules and say "This one spell is the best spell in the game, I'm going to do nothing but cast it over and over. When I run out, we need to rest, since I refuse to use any spell that isn't the best." People are boring and they'll attempt to abuse in game explanations of things. Like, if someone finds out that a spell does fire damage suddenly their mind goes racing into the things they know about fire. It burns, it is hot. What benefit could you get from making something hot? Then suddenly PCs are asking if their fireballs also heat up the opponents armor enough to do damage for the next couple of rounds as it burns them from being red hot. They are asking if enemy Wizard's robes are on fire. They want to cook their food using it. They want to use a fireball to burn down a whole village of orcs without having to step inside.
If you have a metagame construct, you can still have a fun game without people attempting to get more benefit than they are supposed to out of a spell. When you accept that the game IS a game with players playing it then you can say "I understand that a fireball in real life might set his cloak on fire, but we are going to assume that for game purposes it doesn't. It would make the spell too powerful for your level and it would require more time and effort to keep track of it." Which isn't an in-game explanation of anything. It is instead an agreement that the game is more fun when it is balanced and easy to keep track of. You simply don't have your PCs constantly trying to figure out in character why their fireballs don't set things on fireball because then you are purposefully trying to ruin the agreement that you've come to.
That would make more sense if combats were resolved by the first hit, but if you hit 35% of the time while your comrade hits 55% of the time, then that's going to become extremely apparent over the course of a dozen encounters. The NPC is just better, and the PC would be forced to acknowledge that, which doesn't jive with the narrative idea that PCs are supposed to be better than other people. How can someone be better if they require magic items just to be on par?
Is it? NPCs should never use monster stats, first of all, unless you are fighting them. Most NPCs should have 1 hitpoints and an AC of 10. If they are to accompany the PCs, it's best to use the rules for making up allies in the DMG(or DMG2, I always forget). That way their bonuses are in line with the PCs.
As for the discrepancy, I don't think I'd notice the difference between my hit rate on a tennis ball and my partners if one was 35% and the other 55%. Maybe, over time, I'd think "Wow...he's better than me." I don't think I'd ever say "He can't be better than me! That's impossible! I'm supposed to be the best at this game! It's unfair that I'm going to have to buy a better racket to compete!"
The narrative idea that PCs are better than other people means that IN GENERAL they will be better. There will always be exceptions. Like most of the monsters they fight. Some special NPCs and so on.