OSR: How Big Should Your Open World Be, and How Empty Your Dungeons?

I got introduced to RPGs through the Elder Scrolls series before moving to tabletop roleplaying. While I loved the latter-engine games, from Skyrim's icy tundras to Morrowind's blighted wastes, no game in the series is as big as Daggerfall. The exact size is up for debate, but the procedurally generated wilderness is comparable to the area of Great Britain. Some 4,000 dungeons, thousands more towns, temples and farmhouses, and you can visit them all through seamless horse travel! Do you know what this means?

It means that the game is effectively a pointcrawl.

Oh boy, I wonder what's behind those trees?
I know! MORE TREES!

While it's technically possible to travel all over the Iliac Bay on your horse (or, Divines forbid, on foot!) you'd certainly be bored to tears before getting anywhere. Realistic distances combined with 'realistic' travel speed means endless hours pressing the 'W' key.

That is why the intended way to play the game is to fast travel. Open your map, click on a destination, pay travel expenses and watch the days count down over a few seconds before you load in. Oh, and if you're sick then you die on the way. Sucks to be you. 

All that wilderness gets no interaction. Maybe you decide to try riding from one town to another for novelty one time, or you run away from the guards into the forest after you massacre the townsfolk and manage to get away, but then you just open your map and fast-travel to your next destination. In theory, you've got a truly massive open world dotted with adventure just waiting to be found. In practice, you've got a whole lot of distance between anything interesting which always gets skipped over. The wilderness could be a small area around each location plus a prompt to fast travel elsewhere, and you'd lose very little. It's just too big.

So if you're building an open world, how big should it really be?

I'm thinking of hexcrawls like Rick Stump's Seaward campaign, which contain a jaw-droppingly large number of hexes, many just terrain, but with hundreds of labeled locations built up over decades of play. That gives me some of the same feeling as Daggerfall's map, with the distinction that travel in such a hexcrawl has some meaningful choices without taking up much table time. 

These worlds are characterized both by the amount of content in them, and the empty space between that content. For tabletop games, the 'size' of a world amounts to those two factors. How much is there to do, and what resource costs are imposed for getting from one to another? Supplies, in-game time and table-time all count as resource costs, especially if the party is going hex-by-hex. If there are no costs associated with travel, then any sense of distance is a transparent fiction. All your encounters might as well be themed areas in an amusement park.

When the party travels hex-by-hex, choosing the next direction to travel in, rolling for random encounters and getting lost, you can really sell the sense of scale, and the sense of real exploration as well. Check out the Briar Woods session reports in the link above, which get across the grueling nature of exploration in a hostile environment, and how rewarding it can be to discover a secret or to map a whole region. But it also sells how much table-time is sacrificed to build that tension.

Oh Dear

Pointcrawls lose some of that, especially if you can only reliably travel to places you already know, but they really expedite the process at the table. Want to find the adventure? Here are the locations, pick one, pay the resource costs and roll for wanderers. Quicker, simpler, easier gratification. You don't get to feel much like rugged explorers mastering the world around you, but you get to the interesting content faster.

I recognize that some GMs, Stump included, would protest that the unpredictable events and the intense process of exploration that their games create are as much a part of the interesting content any prepped dungeon adventure or town intrigue. Most days, I even agree, and lean towards that sort of play myself. But if I had started the first session of my Castle Xyntillan campaign by making the players travel hex-by-hex and find the right fucking mountain pass in order to get to the massive dungeon filled with exotic treasures, ancient mysteries and extravagant NPCs, there probably wouldn't have been a second session.

Of course, this talk of open worlds and hex-vs-pointcrawls does miss the point that the same dynamic is present in the smaller scale. D&D and its descendants are fractal like that. Some of the same concern over space between content is present in how many empty rooms are in your dungeon.

'Empty' is, of course, a relative term. I'd like to think most GMs today know an empty room doesn't have to be literally bare. It can (and should) have some flavor text, some hints, some humor, even some treasure every now and then. It can still interact with the rest of the dungeon. Just like how an empty terrain hex can still have flavor text and provide a challenge, by getting the players lost, being a site for weather changes and wandering encounters, and providing context for the non-empty hexes nearby.

Given that virtually all my experience with tabletop RPGs has been in the OSR, I've had it quite firmly implanted in my mind that some number of empty rooms is good for a dungeon. I tend to go by Moldvay's 1/3 standard myself, though I know some people have way more. As a result, I do sometimes forget this isn't the default assumption. I recall the first (and so far only) 5e game I played, where a friend of mine was trying his hand GMing for the first time. In the process of making his very own dungeon, he didn't try to create a realistic space (or even a fully representational map) as much as a diagram showing the flow from one encounter location to another. The flow happened to be a straight line in this case, but the principle applies.

Encounter 1, Gate. Encounter 2, Spike Hallway. Encounter 3, Chocolate River, and so on, until the eleventh and final encounter, a big boss fight. 

There was variety, a mix of puzzles with combat encounters, and one point with two paths to choose from. But it was transparently a path from one setpiece encounter to another. No wanderers, no resources consumed from travel, no time records kept (as the specter of Gygax howls from the infinite space between worlds), and minimal interaction between encounter areas. To someone accustomed to the OSR way of things, this feels odd. But it fulfills the same goals as choosing a pointcrawl over a hexcrawl often does. It cuts empty space between the interesting stuff. Exploration is encounter-by-encounter, not room-by-room. If the GM has put some work into giving the space some verisimilitude, then empty spaces are glossed over in narration. If not, disbelief is suspended and you get right into the next encounter.

On the micro level, I have very strong preferences for empty rooms over the alternative, but on the macro-scale, I don't have nearly as strong a preference, and may even lean to the other side. I can't think of a principled reason for this, rather a parochial sense that dungeon design with empty rooms is The True and Right Way, while hexcrawls seem a bit outdated and harder to do for a novice GM. After all, filling out terrain hexes with the odd flavor text doesn't take any more effort than doing the same for a dungeon room.

It may be a matter of ratio between 'empty' and 'full.' Every couple months when I get swept up in the notion of making a big world to hexcrawl in, I run into a big conceptual wall which tells me I'm going to fill ten empty hexes for each one with real play content, whereas a dungeon will have more rooms with meaningful choices in them than not. 

Sarkomand's Fault, first played draft. Still too big.

Looking at DMiurgy's post on his Sarkomand's Fault hexcrawl, it looks like I wasn't the only one with that impression. He actually dug down and made a hexcrawl that big, and downsized later when it was plainly too much. My takeaways from that post are: 

1. You don't need that many encounters.

2. You don't need that much space between them either.

Even the heavily edited version of that map had more unkeyed than keyed hexes, although those unkeyed hexes were often the site of improvised encounters. I haven't exactly done an exhaustive search of the genre, but I haven't found a hexcrawl with more than half the hexes keyed. In contrast, I don't know of many dungeons where more than a third of the rooms are unkeyed or 'empty'.

Is there a particular reason for this difference of the micro and macro scale, when the cost of stocking rooms and hexes seems similar? I don't know. I leave that to your comments below.

------

If you liked what you read, please be sure to subscribe to the blog. Until next time, have an excellent week!