Here's why I don't like it when a game attempts to balance things across some mythical "average adventuring day": it reminds me of
Britannia.
Britannia is a boardgame in which you will be playing a sequence of various tribes/peoples invading Britain, from 43 AD (Romans) to 1066 (Normans). It is meant for four players, each of which plays several invaders over the course of the game. At various points in the game you will get points depending on how much you've conquered. One of the players will, of course, play the Romans. The Romans score a buttload of points in the early game, and then that player will score very few points over the rest of the game until the last round or two when they also get to play the Norwegians. This causes an effect where it's hard to tell how you're doing, because just because the guy playing the Romans has way more points than anyone else at the end of round 5 it doesn't mean that he's actually doing well. This, to me, is bad game design.
Similarly, in an attrition-based system the main impact of taking a crit or something is not so much "Damn, now I'm in trouble" as it is "Ouch, that's going to cost a big spell to heal." Most fights are not dangerous in themselves, as they're only there to sandbag you into spending resources. This, to me, is boring. I want fighting encounters to actually pose a danger in themselves, not just as a matter of "Oh, this encounter went poorly, it cost you 20% of your resources instead of the 15% it was supposed to."
I want the encounter itself to be the main unit of adventure design, and for that to work you need encounter-based resources. It's fine if there are some daily resources to represent digging deep, but the majority of the cool stuff you can do should either be at will or refresh on an encounter basis.