Recall: 1-2 monster, 3 trap, 4 special, 5-6 empty. Yes, differing chances for treasure too, but ignore that for now. Translating these possibilities into wilderness terms:
1-2 lair, 3 hazard, 4 special, 5-6 empty.
What are these things? They are subtly different than the dungeon equivalents, so here are some definitions.
- Lair: a place where monsters reside. Probably counts as a small dungeon, and may sometimes connect to the underworld. Examples: goblin caves, bandit fort, zombie graveyard.
- Hazard: something that is potentially dangerous, but only if PCs interact with it. Examples: lava flow, quicksand, time-stopped wizards mid-duel, town where everyone was killed by a disease, magical radiation.
- Special: something with interesting utility, unlikely to be directly dangerous (though don’t underestimate the power of creative players to make anything dangerous). 50% chance that a special is a settlement. I wish I had a better word than special for this category. Non-settlement examples: wizard’s tower, oasis, dimensional gateway, healing spring, antigravity zone.
If followed rigorously, this system will result in hexes with at most one interesting feature. Frequency is variant, since one third of all areas will probably be empty. A six mile hex is a big area though, so I would sort of like the system to provide the possibility of more than one result per hex. So, new rule: Keep rolling per hex until you roll a 5 or a 6. Each result rolled will be progressively more hidden or off the beaten path. Basically, I see a hierarchy of obviousness regarding hex contents:
- Impossible to miss: if you enter the hex, you are aware of it. If you are following roads, then anything on the road is of this category. Examples: a mile high tower on a flat plain, the smoke from an army’s campfires, New York City (probably doesn’t fit in a 6 mile hex, but you get the idea).
- Standard: some chance in 6 of discovery. If you’re just trying to move through a hex, you still get a passive chance of discovery.
- Less: some chance in 6 of discovery, and you must proactively spend time exploring the hex.
- Least: like above, but lower chance of discovery.
This “obviousness” hierarchy is still something I’m working on. The similarity to secret door systems is not lost on me, but I haven’t decided exactly how it should best be systematized yet. I also feel like I read something similar to this somewhere, but I’m not sure where. Also, in case it’s not clear, the default level of obviousness is standard, not impossible to miss.
If there are dangerous things (hazards) and interactive things (specials) there should also be a category of things that are interesting but not so interactive, right? Otherwise, I suspect the hazards and specials will start to feel too common, even if you try to improvise other details at the table. Dungeon stocking systems often have the concept of dungeon dressing; thus, we need wilderness dressing too. Since fluff is crunch, the dressing might still be of use to creative PCs too. My first thought for this was an independent roll per hex, with a 2 in 6 or 3 in 6 chance of added dressing. Like the general hex stocking step, keep rolling until you get a negative result. So, the majority of hexes will not have a dressing element, but a few will have more than one. Hopefully this distribution will feel organic and keep players guessing. Examples of wilderness dressing: abandoned farmhouse, small canyon, half-buried dinosaur skeleton, crater from a past explosion.
To summarize the system as I see it working now: for each hex, roll 3d6 (each die being a different color would probably be convenient). Die 1 determines the main result (lair, hazard, etc), die 2 determines the treasure, die 3 determines if there is dressing. Die 1 and die 3 are each re-rolled until they come up negative, determining ever more hidden hex features. For each result, the relevant subtable is consulted for details about type. I see these tables as a mixture of generic elements that can repeat and unique Vornheim-like entries that are crossed off and replaced with something else once rolled. Monsters are determined by rolling on a terrain-specific table. Treasures are determined by monster hoard type (if part of a lair result) or the standard dungeon “unattended treasure” table (maybe? I haven’t thought this through yet). In the future, I might do some table consolidation, so that fewer die rolls are needed, but I want to leave the various rolls separate for now to preserve the probabilities and make it clear where results are coming from. Also, I don’t think the number of dice required can by much reduced if I want to support a variable number of features and dressing elements per hex.
Latching on to the wrong part of your post:
New York City is a lot bigger than 1 hex, but interestingly Manhattan is less than 1 hex by area (it is about 2 hexes long).
Manhattan is about 23 square miles, a 6 mile hex is about 32 square miles.
Manhattan is about 2 hexes in length but only about a third of a hex wide at its widest.
A lot happens in Manhattan!
All of NYC would occupy almost 10 hexes.
New new rule:
Rolling until I get a 5 or 6 is generating too much content. So, instead, on a result other than empty, there is a 1 in 6 chance of rolling again.
Same thing for dressing.
This is probably needlessly complicated, but I’ll do a full review and edit of the entire system once it is done.