One of my chief concerns with the Dark Before Dawn LARP was whether there would be enough for the players to do. In a 6 1/2 hour game you really want to make sure that no one's bored, or at least, that they're not bored for long. Naturally a particularly social bunch eager to explore their character's psychological vulnerabilities could take any LARP pitch and roll with it for even ten or more hours, but most players need a little more structure than that.
My first consideration involved the characters' goals. Give them goals that tie in with each other and suggest alliances, rivalries, and enemies, which were straightforward enough to understand but complex enough to intrigue. The larger the web, the more entertaining these goals can be, but with a small group of 11 players I guessed that these goals would give them 2 - 3 hours worth of entertainment, tops.
In reality these goals became a subtle backdrop to the character's actions rather than holding star position - likely due to the possessions and mysterious location. The goals did entertain, at least, but they didn't chew up time for long. This was always a possibility, just because you give someone an enemy doesn't mean they won't put it aside for now and focus on what's in front of them, especially if there's no obvious and subtle way to deal with them otherwise.
My second consideration was to involve a mystery for the characters to solve. They were in an unusual place. They all disagreed on the exact date. They didn't remember arriving at this location and had expected to walk in elsewhere. This consideration took up an hour or two of actual game-time, interwoven throughout the six, as people queried each other for further knowledge and poked the various props about the place. I discovered that mysteries were pretty unbalanced ... they gained weight as the game progressed and the little hints started tumbling together into an avalanche. It certainly worked a treat and coloured everything else but it would have been pretty bare on its own.
My third consideration was to throw in some antagonists for the characters to butt heads against. Initially I was going to draft in some cast but since the game attracted far fewer players than expected, I went back to basics and instead asked a few of the players if they wouldn't mind being possessed for a bit. Since I didn't want them to lose their characters, and since relying on combat would be a bit of a drag, I threw in several exorcisms about the place that they could theoretically use and made it so that each exorcism targeted one person and that if the wrong person was picked, the Strix would become immune to that technique.
I also made it so that the first two possession victims would be the ones with the necessary skills to use two of those forms of exorcism. This ate up the lion's share of four hours, especially when the first possession victim took so long to dispel. The third, ironically enough, was immediately exorcised on a whim by a character who didn't realise that the technique would lose its potency if performed on the wrong person.
My final consideration was the capstone plot ... everyone needed to cast their vote for one entity or another, or abstain completely. This took up the last 30 - 45 minutes.
All told, we had just enough structure and plot to keep 11 people occupied for all but 15 minutes of the game (the bit shortly before the ending plot struck).
Naturally this was in part due to the low number of players, the mystery set up which tweaked characters to focus on solving puzzles (i.e. possessions) and the way external threats tend to put a dampener of internal politics and intrigue. Part of it is doubtless due to local style and the desires of the various players involved, but it does make you think.