For your style of "sandbox play" (which I find very intriguing but have never executed as a referee), can you give us a sense of how much of the information above you had "planned out" versus how much you made up based on the actions of your players? I know your players read the blog so I'm not looking for you to give anything away, but I'm also fascinated by comments you make that seem as though you make up really interesting campaign plot points "on the spot", but then toward the end of your summary above you mention "The additional details about Areon, as well as the Stranger from "Earth" were ideas I'd wanted to work into the campaign at some point", so clearly you had thought about those ahead of time. I'm just trying to understand your process a little more. I have definitely fallen into a pattern of preparing way too much background information for my games, most of which never gets used but which takes time away from me working on the actual things that my players are dealing with.I thought this was a question worth addressing at some length, since I get asked it (or variations of it) a lot.
I've often described my refereeing style as very "seat of the pants" and talked about how little preparation I do before each session. That's true to a great extent, but I think I may have also misled people somewhat. For example, I have maps of Dwimmermount, which I made well in advance. I've also keyed these maps with a few special encounters, but the rest I leave empty or at least undescribed. As the players explore the dungeon, I may or may not add encounters or items to some of these empty or undescribed rooms. Whether I do so or not depends greatly on how the session is unfolding at that time or if I have any ideas I'd like to insert into the campaign. The same is true of details about the wider world. I have a (very vague) map of the world outside Dwimmermount, Muntburg, and Adamas and some areas on that map have been detailed in advance, but much of it is left blank, to be filled in as needed/desired.
Before each session, I don't spend hours poring over my maps or working out details of what the characters will encounter. However, I do spend a lot of time thinking, often at odd hours of the night or while I'm out walking around or engaged in some daily, boring activity. I come up with lots of ideas and I store them all away for use later. Some of these ideas are just inklings of thoughts and others are more elaborate. Some are (mostly) original and others are variations on stuff I've read in books or gaming products or seen in some other context. Some I think about how I might include them in the campaign and others I'm not even sure would work very well in it. But they're all useful, because they're the "raw materials" on which I draw when I'm in need of something to include in a session.
When I say I "don't plan beforehand," what I mean is that I don't (generally) decide in advance a lot of details about what's in this room or that one or what's beyond this hill, etc. This applies on every level of the campaign, from a shopkeeper in Adamas to wider cosmos beyond the world on which the campaign takes place. The advantage of this approach is twofold. First, I have a lot of flexibility. I can bend and shape the nature of things to suit both the players' tastes and interests and those unexpected "in the moment" events that good gaming often generates. Second, it helps to foster a sense of mystery, because, quite literally, no one -- not even I -- know the answers to certain questions. So, when I throw out a piece of information or have an NPC say "X is true" about the world, it's not rarely cut and dried. Until I absolutely need to establish something about the campaign, I don't, preferring instead to keep things vague and/or contradictory, which, so far, has worked very well.
Of course, this approach has some potential pitfalls, not least of which being that, unless you're quick on your feet and have a good memory, you're liable to wind up with a very muddled and inconsistent campaign. Fortunately, those are my two main talents as a referee, which means I am able to make this barely controlled chaos seem a great deal more intentionally coherent than it really is. I suspect that another issue inherent in this approach is that it's a lot harder to game when you're not "on." That is, unless you're really feeling up to it, you're more likely to stall and delay and not actually play. In retrospect, I realize that many of my "off" sessions have been as much as result of my own lack of energy for the demands of thinking nimbly on any given night than external factors. What I do is very demanding on the referee and, if he's not feeling up to the task, sessions tends to be less than exciting.
All that said, I'm quite happy with the approach I've adopted. It certainly suits my temperament these days and the campaign has survived and prospered despite its ups and downs. I also feel that I've been a lot more creative and open-minded in this campaign than I have been in many of many of my past ones, precisely because I have to be. I can't fall back on predetermined answers to questions about the setting, so my responses to player queries often invite even more questions rather than being definitive. That's to the good, I think, but it also means that there are now dozens of dangling threads, any one of which could, if pulled too hard, unravel the campaign. It's a juggling act, to be sure, but then that's why I enjoy being a referee and, by most accounts, I do a pretty good job at it, so I plan to stick with it and see where this craziness leads.