R.E.A.L.M.S. - What Went Right
1. Open to new ideas
At the time we ran REALMS, we really had very few ideas about larp outside of the one I had gone to (the two partners who ended up running it with me had no experience to speak of at all). This caused us to try very different things with both the plot and the rules, especially over time. Additionally, we were willing to try things that a "professional" larp wouldn't have done -- for example, our weekend events were held at a member's vacation house that he graciously let us use. The site was not perfect, but without any money it worked great for us. Instead of stretching to try to accomodate use of a better site, we did the best we could with what we had (and it ended up working out well).
2. Kept plot notes on individual characters
Whenever we worked on a plot, we made sure to note any "factions" which the plots involved. As the event went on, any NPCs who had significant experiences with the PCs would jot them down on plot's copy of that character's character sheet. After the game, this information would be compiled and would affect further events. For example, one plot had the Red Hand tribe of goblins pitted against the Blue Hand tribe. Most of the players chose to ally with the Blue Hand tribe, who offered rewards for Red Hand badges. After the event, we included notes on the players who were seen by Red Hand goblins to help the Blue Hand tribe, giving them negative points with the Red Hand tribe (and of course they got positive points with the Blue Hand tribe). Additionally, we noted the specific event when one character managed to kill the leader of the Red Hand tribe. By keeping written record of this, it was much easier to bring back the Red Hand tribe a few months later and remember who they wanted to kill and how far they would go to get them -- they rushed the character who killed their leader and went after them much more than the others who had just killed a few Red Tribe members in general.
It did not take long at all for returning NPCs to make these notes, and it was definitely worthwhile to do. Most of our later plots were influenced by flipping through the character sheets and thinking, "oh yeah -- so-and-so helped this big time farmer in the area they're going to be in" or "hmm, the King's guards are looking for this guy".
3. Extensive preparation for running games
Before we held any events, we made sure we had a good supply of props. No monster would go out without at least a reasonably-colored tabard and decent mask. This was surprisingly easy to achieve, especially by going to a costume store the day after halloween. Latex masks that normally cost $20 could be got for $5. I spent about $200 in one day and got far more costuming props than I had hoped for.
After the brief playtesting period, we made sure to have *all* the rules ready. Even though players wouldn't attain the most powerful spells and enchantments for a long time, we created all the rules for them up front. This allowed us to give the players exposure to the highest end of things and gave them a definite goal to go for by seeing them in action. Since running REALMS, I have experienced several new games where there's always some things that are undefined since it will be a while before the players can achieve them. My experience with REALMS shows that it is definitely worth it to work all of these things out before you start to run real games with your rules. The players will feel better about it and the game world will feel more consistent as a whole.
Again before the first event, we had major background material available. Extensive lists of monsters and their abilities; a series of in-game documents talking about the world; systems to work with the Essences (powerful beings that could grant powers based on player initiation to their societies), etc. etc. The monster lists were especially useful, since we made sure to stick to the descriptions given -- once a player saw a wolf mask with a white tabard and got an idea of its abilities, they felt good that the wolf mask with black and orange stripings meant a very different monster. We had less calls of "What do I see?" than I've experienced in any other larp I've played because the monster stats were so detailed up front and we stuck to them well.