In programming, we call it "shaving the yak". You start with a simple task: move the login button five pixels to the right, say. But each task requires you to complete another, stranger, harder, task first. To move the button you need to recompile the stylesheet. To do that, you need to update your Gemfile. But that leads to a dependency conflict that requires you to update Ruby to a version that's not installed, so you have to reinstall that, but the Chef recipe to do it is out of date, so you need to fix that first. The escalation never stops, and pretty quickly you find yourself in a yurt on the Mongolian steppes, shaving a yak, and you can't remember why.
I just played a game like that.
Folly is a street game containing a number of quests. But where a simple quest structure would be A -> B -> C, with each completed objective setting a new one, Folly uses A(B(C(D(E)))), each destination saying "Sure, I can help you. But first I need you to.."
It gets even harder, because there are two types of interruption - being dead, and having the wrong type of card. If both have just happened to you, then you insert two extra tasks ("Visit Death to get revived" and "Find a player with the right card, intercept them, challenge them, complete the challenge, and swap cards with them") before whatever else you were doing.
That's not a quest. That's a to-do list.
And yet - and yet - it works. Folly was a lot of fun! I got slightly lost, but usually had a reasonable idea of what to try next, and quickly adapted my style from "run! run fast!" to "listen to everything and make extensive notes".
There's a lesson there. I'd never have attempted to run a game that hideously complex. But the kind of games Fire Hazard makes - short briefings, high energy, limited story - are not the only kinds of game that can be made.
I have a theory about why it worked: I had a spirit guide.
Every game is a battle to stay in the 'magic circle'; to suspend disbelief and see the imaginary world of the game, rather than the real world. The more story-driven the game, the more different the characters are to the players, and the more looking-silly-in-public is required, the harder it is to stay there. Blink, and that "power cell" turns back into a coke bottle full of glowsticks, and there's no point in anything any more. Fairies really do only exist if people believe in them.
People are strongly influenced by the people around them, so it's the crew and the players themselves that maintain the integrity of the circle. Lock them in a room, and you've contained the energy; disengaged players will get brought back in by the intensity around them. But street games have an energy leak - the real world is right there, and it's full of disbelievers.
Folly worked for me because, although it's normally played alone, we bent the rules to play as a team. Amy knew that she'd have to leave halfway to run her own game, so rather than risk breaking the game, she became my invisible-to-the-game-world partner. She could advise and encourage, but all in-game challenges, I had to complete myself.
This is an interesting mechanic to start with, because often players will want different levels of involvement in a game. The Spirit Guide isn't quite a spectactor but isn't quite a player, is free to leave at any time, and can't be addressed directly by in-game crew or players. But, more importantly, this mechanic closed the loop: having someone around for whom the game world was indisputably real and the (quite silly) tasks were absolutely necessary kept me in the circle as well.
And this is why Citydash enforces a stick-with-your-team rule. We generate intensity anyway in a number of ways, but keeping players in small squads makes a closed loop that amplifies it. Even Undercover keeps players in pairs, even though that undermines the spy theme a little. It's worth it.
If you're going out into the real world, don't go alone.