For another example of how one dude handles wilderness, I just make a short table (12 or 16 entries). The players tell me the direction they are heading (I do have terrain figured out beforehand, but nothing else) and I check each hex for a random encounter. Most of these encounters are lifted from Carcosa, so they have a lot of underlying depth in a sentence or two, and are also easily modified to suit near any campaign.
This is a really interesting approach. It lies somewhere between having nothing other than terrain and rolling wandering monsters and keying up hexes statically. In computer science terms, this method is somewhat like late binding. I see several advantages: one, less material is needed; two, the referee can be surprised along with the players; three, you end up building a setting through play gradually rather than all at once prior to the game (compare to the character build versus development through play game styles).
There are several things that I want in a hexcrawl that are not supported by the Aplus method though, assuming I am understanding it correctly. I would like the direction chosen by the players to matter regarding more than the terrain type. Assuming that there is only one list of encounters in play, it seems like you would have the same die roll no matter which direction was chosen. It’s not exactly the same thing as a quantum ogre, but it does seem to preclude information gathering beforehand.
Unless information gathering, in addition to actual travel, is grounds for determining hex contents. In other words, things start to exist only when you look at them, and researching rumors and travel are both ways of “looking at” hexes. That sounds promising, but I suspect it would fluster me at the table, so I still think I would prefer to precompile. Also, I’m really bad at taking notes during play, so I fear that I would end up losing much of the richness created at the table. (I really think good session note taking is one of the most valuable referee skills, and I’m terrible at it.)
There are two other little subsystems that I have been working on (for future posts) which also require having some hexes set down beforehand. The first is autogenerating rumor lists based on the contents of adjacent hexes. Yes, it’s about as simple as it sounds, but I added a few complications to decrease the predictability somewhat. The second is creating relationships between the contents of different hexes. For example, the wizard in the tower in hex A might be interested in capturing the creature in hex B or taking vengeance on the fighter is settlement C. I don’t really have a system for that yet, but I’m working on it.
All that being said, I like the Aplus method and think it is very practical, especially for people like me who probably tend to make the perfect the enemy of the good. I may try it the next time I want to get a game going with minimal prep. It did also make me step back from the systems that I had been working on and ask myself what I was gaining from the amount of work I was doing, which is useful to do periodically.