Applied Wilderness Theory

Due in part to the recent Hexographer 2 beta releases, I've been playing around with building a microsandbox - 10 hexes by 10.  I'm not happy with it yet, but here are some notes that have come out of it:

• Taking the "wilderness as dungeon" model, 100 hexes is comparable to a hundred-room dungeon.  That's on the upper end of anything I've ever actually built and stocked, so this might be bigger than I really wanted.  On the other hand, I do still have a fair number of "empty" hexes.  Honestly might be a bit too big for the "micro-sandbox" label.
• It's about the size of a duchy, I think?  A 24-mile hex is ~16 6-mile hexes, so a hundred six mile hexes is about 6 24-mile hexes, which is about right for a duke with 5 vassal counties.
• Borderlands are the correct civilization level for "wilderness" adventuring.  They're perfect.  Previous hexcrawls were largely wilderness, which per Lairs and Encounters have many (2-8) lairs per hex, which is just impractical to stock (sure, sure, dynamic lairs, maybe some year).  Borderlands tend to average between 0.33 and 1 expected lairs per hex, which is great (gee, sounds familiar).  So in a 100-hex area, I have around 40 lairs - quite a few, but waaay more workable than the Shieldlands campaign, where I probably had 40 lairs within 12 miles of the town the PCs were running.
• Corollary: About half of borderlands six-mile hexes require zero work to clear.  No lairs, no problem.
• Assuming civilization more-or-less surrounding the sandbox area, the 25-mile borderlands radius covers almost all of a 10x10 hexmap (close enough for me).  We also know that it's 50 miles from the edge of the sandbox to a class IV or better market (probably about 6 miles from the edge of the sandbox to a class VI market, and maybe 24 miles from the edge to a class V market).  So that lets you track time spent getting to and from markets in civilized areas without actually having to roll random encounters with dirt-farmers or track hex-by-hex movement and rations (presumably you can buy them off of hamlets you're passing through daily).
• Again thinking of wilderness-as-dungeon, I decided to steal a few pages from the dungeon stocking rules, which note that the random encounter tables generate roughly 33% stupid enemies, 33% beastmen / "factious" enemies, and 33% high-intelligence / high-power "men and monsters".  I tweaked these numbers a bit, towards 50% critters, 33% beastmen, and 17% men-and-monsters.  This yields about 20 animal lairs, 13 beastman lairs, and 7 other lairs.
• Animal lairs are super low-effort to stock, and present a more natural-feeling wilderness (whether or not predator densities that high are actually sustainable is another question).  I also cut out the really boring animals (goats, normal-size hawks, a lone rattlesnake, ...) in favor of a more...  folklorish, Northern European / Tolkeinesque carnivorous animal selections (boars, bears, wolves, giant spiders, giant bats, giant weasels, ...). 
• 13 beastman lairs is at the upper edge of the reasonable, I think.  They have a lot of moving parts, and I'll probably need to differentiate them so that players can keep them straight.  I wonder if that was secretly the point of having all of those cookie-cutter beastman races that are mechanically almost-the-same; minimal viable differentiation (which is easier to keep track of: blue orcs vs green orcs, or orcs vs guys with hyena heads?).
• I've also been putting some thought into "alert / alarmed" mechanics for reactions to PC actions (and other nonmechanical difficulty factors), and detection of the party when it is nearby.
• Strongly considering a Shadow of Mordor-style orc name generator, because 13 is a lot of chieftains (never mind subchieftains, witch-doctors, and shamans...).
• Seven "men and monsters" were actually pretty easy.
• One day is almost certainly the analog for a 10-minute turn in dungeon-time here.  This was less true when I ran the Bjornaborg wilderness game on 1.5-mile hexes, but on six-mile hexes at 1-2 hexes per day, that's definitely the correct time to track (and this is an important thing to figure out, because that lets you start honing in on that central play-loop which has historically been poorly-defined and vexing for us [1][2], and also helps figure out the resource model).  Rations start to look sort of like torches in this accounting, in that a torch lasts six turns and non-iron rations spoil in seven days.  Removing water requirements by assuming fresh, drinkable water from abundant streams also greatly eases the logistical burden of rations (down to 1lb/man*day, or 10 man-days per stone, instead of 1 stone per man-day when you have to carry water).  An even simpler abstraction would be 1st per man-week, assuming that you're carrying some backup water or wine or whatever.
• At 10x10 hexes with 1-day "turns", your upper-bound on expedition length is about a month (in 30 days you could get to most any point on the map and back, I think).  This does tie back to one complaint that my players had about wilderness adventures - if you have a choice between one adventure a month in the wilderness, or one adventure a week in the dungeon, wilderness needs to have a much higher treasure yield per expedition to make sense in terms of gp/game-time (which matters because of monthly expenses).  I'm not sure that the microsandbox I'm working on has that high-yield property (yet / currently), but that's also less important if the game is explicitly structured/pitched as "a wilderness game" with no / minimal dungeons and all characters starting at like 5th+ level.
• An orc wilderness lair has an average treasure value of 14kgp, which is like a decent but not outstanding treasure map.  They're also a source of liberated human prisoners, though, which sounds to me like a pretty good way to pick up some mercs or a replacement PC if things went poorly.
• I'm not sure my players will fall for "legend tells of a great dragon's hoard somewhere in the highlands" again...  maybe I just need to use a bigger dragon, because bigger dragons have bigger treasure.  Right?
• In terms of calibrating difficulty by analogy with the dungeon, wild animals are low-threat, beastman warbands should be medium-threat, and beastman lairs and special monsters should be higher-threat.  At one encounter per day, a 5th-level party with some merc backup could probably take a warband, in much the same way that a 1st-level party could probably take a gang of 5 orcs in the dungeon (depending some on surprise, tactics, armament, and luck; if the wizard gets sleep/fireball off, you're probably fine, but if he gets interrupted and pincushioned, then it's going to be a rough day).
• I still think the right way to run such a combat is probably on 30' hexes.  Wilderness movement and ranges are in 30' intervals already, it's about the area of a ground-burst fireball, it translates nicely to DaW on platoon scale, and it's sort of analogous to fitting the whole party into a single 10' square (as happens for small parties in dungeons).
• Still haven't really worked out local weather patterns for this map yet.  If adventures take weeks-to-a-month, seasonal weather rules might be necessary.  I do have some stuff for temperature (mostly as it related to the choice to have a fire or not when camping, choosing between mercenary morale penalty and no natural healing or an extra random encounter roll - worth noting, of course, that most beastman villages are also going to have fires, so that's a nice way to locate lairs from a distance).
• Two big differences with dungeons that I'm noticing practically are that you can be in the same hex/room as monsters and not realize it, and that visibility of the structure of the terrain (ie, "that hex is forested") is much easier to see from a distance than the structure of the dungeon.  Most wilderness features are hidden in a hex, while most dungeon-room features are pretty obvious when you're in that room.  I recall being nonplussed with L&E's system for finding hex features, though I forget why; will probably reread those and then work something out.