In Defense of Empty Hexes
In this post, I argue that hexcrawls don't need to have a key for every hex, and might even be better if they don't.1
My perspective on this is influenced by my experience running Antarctica jam.2 Per discussions on Discord and what people have written, there's a common (but not universal) perception that hexcrawls should be dense, with, say, at least 50% of the hexes filled.3 I won't pretend I'm not tilting at windmills here, but I do think there's a common perception that a hexcrawl is incomplete unless it's wall-to-wall with hexes.
In the context of AAJ I try not to push my ideas on others -- the ethos of the jam is about presenting people with an impossible task and enjoying the diversity of approaches that ensue. But on my personal weblog, I won't shy away from expressing my annoying, contrarian opinions. The hexcrawls are too damn dense!
Some of the first hexes I crawled were in Hot Springs Island. It's a rich, complex scenario, but in most sessions we would move exactly one hex and then spend the session poking around the 1-3 features that were there for us to explore. There wasn't really much agency to how we were exploring the island itself, because everywhere we went we knew to expect 1-3 interesting discoveries.
Conversely, in EvilTables' Northern Strata game the party is currently exploring a sparsely-stocked tundra. Points of interest are often spaced out by ten or more hexes, and in that emptiness a whole new dimension of the game unfolds. Keyed hexes become something to seek out rather than blunder across -- we actively ask for maps and directions when we meet an NPC. When we embark, we pay attention to logistics -- firewood, food, pack animals -- because it may take a week or more to get anywhere. Random encounters become a bigger part of the emergent story. We use scouts and vantage points to scope out neighboring hexes. We follow the course of rivers, and we make lots of notes on the VTT. All of this is really fun, and much of it would be sidelined if Tables had stocked his map with back-to-back hexes.4
Empty hexes are analogous to empty rooms in a megadungeon. Some of the reluctance to use empty hexes/rooms comes, I think, from an assumed passivity on the part of players, a sense that every move must be rewarded or validated by an opportunity for interaction.5 But in a megadungon, that empty room is interactive -- it has doors to choose from! And likewise with an overland sandbox: there's a horizon to scan, rivers to follow, roads to diverge from, misremembered directions to argue over, opportunities for random encounters, etc.
You can trust players to pass through empty places without getting bored. I promise it's true. Whether underworld or overworld, the game itself is more than just glue holding together a string of pre-written micro-scenarios.
Sparser stocking also feels more grounded from the player's side. Agency in an exploration game is directly proportional to mappability. When hexes (or grid squares) are used as a "minimum mapping unit" (i.e. rather than inflated to the maximum possible size such that there are not empty spaces), the terrain becomes tractable and concrete. If everything is exactly one unit away from six other interesting things, the world feels like an abstraction without space.
It's even worse when hexes have more than one feature -- to pick on HSI again, most of the gameplay happened within a hex, and hex interiors aren't mapped, so either the GM has to do the legwork of mapping these features (in which case, why not just use smaller hexes in the first place?), or players are left with a weird low-agency limbo. Saying "I want to go north-east" is an easy, unambiguous statement of intent, navigating the quantum flux of an unmapped multi-feature hex is not.6
Finally, an unkeyed hex is not truly 'empty' -- it contains a terrain type and a horizon on which distant features might be visible. It might have paths or waterways. It has a (terrain-dependent) chance of incurring a wandering encounter. It might have a risk of getting lost. It has proximity to other features on the map.
If you've read this far, thanks for indulging my need to opine. I should note that this is mostly aimed at module design trends -- I suspect that most people end up stocking pretty sparsely when throwing together hex maps for their home campaign. And obviously, navigation does not need to be a prominent part of your game. But I personally love it, and accordingly want to at least make the case that a sparsely stocked hex map is not a failure or cop-out.
That is, there's a strong and weak version of this post. The weak version is that empty hexes are fine & legitimate regardless of the merits of dense hexcrawls. I personally think that hexcrawls with empty hexes are better than those that fill every hex, but I acknowledge that it's a personal preference. Also, all of this is separate from the issue of the geographic size of a hex. For a related discussion, see my post There Is No Optimal Hex Size.↩
But also, my experiences as a player. My most recent hexcrawls have been a Hot Springs Island campaign (very dense multi-feature hexes), EvilTables' Northern Strata game (density ranging from medium to low), my Castle Mordengard game (low density), and Scribble's REDUX campaign (extremely low density).↩
The most popular tool by far for filling hexes are Luke Gearing's Wolves Hexfill tables which puts something in every other hex on average. The tables are great, but I think that 50% density is too high -- when I use them, I prefer replace the first die roll with a d8 or d10.↩
Another good example is REDUX session 29 -- we traversed dozens of hexes over a month of in-game time, and only encountered two keyed PoIs. But the random encounters and terrain were enough for a compelling session.↩
As an aside, I have some beef with the tendency to describe OD&D-era dungeons as "funhouses". Yes, they tended to be quirky and not 'ecologically coherent', but if anything evokes the experience of being at an amusement park, it's the 'something to interact with in every room' OSE house style.↩
A good example is the recent Dead Letters episode on Luka Rejec's Holy Mountain Shaker, where the hosts express frustration that the nodes in the pointcrawl are unmapped lists of indeterminate features. Obviously, it's possible to negotiate these things, but something is lost when you don't have a map.↩