Every LARP-running experience features some lessons to learn. This one was no exception. The lessons included:
If someone says they really, really, really want to play a Jaegermonster, let them. The Jaegers in both runs were *extraordinary*, a major highlight of the game. (The trio in the second run were reportedly novice LARPers, and were better than most veterans I know.)
Do not try to move eight-foot-long tables of uncertain stability by yourself. Especially, do not do so by shoving them. Dropping the end of a table on the arch of your foot is not an ideal way to start a day of GM'ing. (Ow, ow, ow...)
Dumb casting luck can strike twice. I never thought I'd have a re-run of the Ozma case, but it did almost repeat. The player who was going to play my favorite character of the game (a high-angst, high-romance character with a 15 page character sheet) had to drop on the evening before game run. I almost just ran without the character, but dervishspin stepped up to the challenge. From getting that 15-page sheet 11 hours before the game and knowing nothing about the comic, she came in the next morning, costumed just right, and *nailed* the role. It was a delight to watch.
Rocket-powered golf clubs do *not* make a wise demonstration example for Spark mechanics. The universe is listening, and has a wicked sense of humor.
Mostly, I determined that not only are adaptation games a bit harder than normal ones, adapting an ongoing, non-episodic story is quite a bit harder still. Oz might have been using other peoples' characters, but at least we had the entire L. Frank Baum corpus in front of us, and knew exactly how much freedom we had to embroider. (Quite a bit, given how internally inconsistent Oz is to start with.)
But you have to fit a Girl Genius game inside an ongoing story, one where only the Foglios really understand the details. Worse, all evidence is that they *do* know many of those details, and just haven't told us yet. So I had to start with three months of simply evaluating everything that we knew, to figure out where my opportunities to invent were. And it's still likely that at least 80% of the guesswork in the story is just plain dead-wrong. (Although I still hold out hope that my Skifander backstory is at least partly correct -- that was pulling together lots of hints, so I think it's plausible.)
Overall, a good experience, and the game will probably get cleaned up and re-run at some point down the road. But it's good to get to the end of the process, and let out all the steam that's been building up and driving me forward for the past couple of months...