For some reason this game only ran when there were two sessions running simultaneously. We're going through the game with a GM, in a small university lecture hall (because the convention was held at a uni). The GM wore a headset, and every now and then a second GM with a headset would wander into the room, chat with our GM and then wander back out again. Getting quick glimpses through the door, we could see a stragetic map with figures all over it that wasn't really meant for our eyes... I think in retrospect that we were meant to have seen the map, but not necessarily the fine details on it.
Roughly halfway through the game, the second GM comes in with someone we all know, someone who was playing that second game. He looks a bit surprised as he enters the room, then with a toy machine gun he makes noises as though he is shoiting us all down in a spray of bullets. Dice are rolled, some of us are injured, and more dice are rolled as he "drives off". In that minute, we realise that the game situation hasn't been changing because the GM is just being nasty...it's been changing because a second team running simultaneously with us has been making their own changes to the scenario, just like we had been changing theirs. The game instantly shifted up a gear.
We didn't know what the characters of those players had done, but we played the metagame, knowing what the players were likely to do, and refining our strategies according to that.
This massive LARP idea that I'm working on could work in a similar vein. The players are generally told from the outset that the area used for the game has been hired for a specific timeframe, and that it will have to be split in half because two different groups want to run games at the same time. One group will be running a fantasy LARP, the other will be running a sci-fi LARP. Once certain criteria are reached in the two games, a portal between the worlds opens up and the two games (with 50+ players each) merge into one game (with 100+ players).
This adds an instant pacing mechanism to the game, and while I don't see it causing too much of an issue it could add a degree of complication that pushes the game away from the elegant simplicity I'm aiming for.
More to think about...it's probably time to start nailing down some of these concepts.