As far as extended rests are concerned, we haven't adopted any official house rule - but in practice I tend to end up regulating them, in my capacity as GM, because I am the one who controls the pacing of events. This becomes particularly straightforward at paragon and epic, where the PCs spend a fair bit of time in places like the Underdark or the Abyss where it's easy to assert that, absent special circumstances, the environment is not conducive to resting.
For me, the guiding principle in pacing challenges vs rests is always to keep the pressure and the players sweating - if the PCs have fewer surges left in the party than there are characters, and are facing an above-level challenge so that the players are thinking "How are we going to get ourselves out of this one?", then it's working well. To me, it's not really much like classic AD&D attrition/resource management. It's more like: (i) players have the job of managing resources
within encounters, to try and achieve victory; (ii) I have the job of framing them into enough interesting and challenging encounters that they will feel pressured to use their daily resources while not wanting to be profligate with them; and (iii) if I'm doing (ii) properly, than (i) will become more intricate because the players will care not just about encounter but also daily resources.
The overall incentive that makes this structure work, at our table, is pretty informal: if the players blow all their resources and then can't handle another encounter or two before resting, they'll look like squibs and I'll make fun of the them! In other words, it's a set of "soft" social dynamics and expectations. Playing with strangers, or in a tournament-style environment, probably some more formal way of rationing extended rests would be necessary - I've never used the 13th Age approach (rest before making it through 4 level-appropriate encounters and you suffer a story setback), but it looks to me like one good way of introducing that formal rationing.
Be slightly careful with YES AND. It can easily be taken too far and lead to a game that is almost entirely mush with very poor storytelling. Individual scenes can work amazingly with improv, but longer stories take a lot more skill because to make a story
<sni>
[indent[ou come up with an idea and it’s like ‘okay, this happens’ and then ‘THIS happens.’ No no no. It should be ‘this happens’ and THEREFORE ‘this happens.’ BUT ‘this happens’ THEREFORE ‘this happens.’[/indent]
This is an interesting point.
I think the "therefore" or the "but" can be pretty loose, as long is everyone is on the same page. From the GM's point of view, though, it's easy to see (or, perhaps better, to
project) a "therefore" which means nothing to the players because they're not privy to the backstory. Whereas I think if you anchor stuff that happens, and developments that occur, to the PCs - backgrounds, activities, goals, preferences, etc then the players can get into it even if it doesn't really make much sense.
That is, I think in a RPG - where the players are able to provide a lot of the drive and direction for events - the "therefore" can be a sort of "meta" therefore and the players will still respond to it. Like
in my most recent Dark Sun session, when the Templar inquisitor tracked down the PCs with a psychic hound (modelled on the X-Men's Rachel Summers): if you look at the events through the lens of a Le Carre-style novel, or even a Raymond Chandler style novel, it's a bit obscure why the inquisitor wants to hunt down the PCs. But from the point of view of the players, (i) they're on the run from the Templars, and (ii) this inquisitor is the (ex-)handler of one of them who (in the PC background) was an assassin in the service of the Templars up until play started at the moment of Sorcerer King Kalek's death, at which point the character decided to break free of the hold his masters had on him.
So the "therefore" to the players is much more about the events speaking to the logic and rationale of their characters, then making a fully sensible plot that would work in a mystery story. I think this is a point where the requirements for GMing become a bit different from the requirements for writing; and also where you can get player-driven story-type action without needing to worry too much about "the story" in a bigger picture sense.