How fine-grained does one like one's terrain types, sir? And is it possible to come up with some principles regarding how many such types there should ideally be?
I ask because I happen to have been paging idly through Oriental Adventures just now, and came across in it a page of Encounter Tables by Terrain Type. These are:
- City/Town
- Tropical/Sub-tropical mountains
- Tropical/Sub-tropical ocean
- Temperate swamp
- Rural
- Tropical/Sub-tropical forest
- Temperate mountain
- Temperate freshwater
- Steppe
- Tropical/Sub-tropical swamp
- Temperate forest
- Temperate ocean
- Aerial
- Tropical/Sub-tropical freshwater
- Temperate plain
- Cold regions (all terrain types)
Clearly the designers got bored by the end - or perhaps just didn't have enough monster varieties - and decided not to bother with distinct tables for 'cold' forests, mountains, swamps and so on. There are also some oddities and lacunae in their approach. One is that there is no distinction drawn between tropical, temperate, and sub-tropical cities and rural areas. Another is that there is no concept of hills: everything is either mountain or forest, or open flatland.
Otherwise their usage is pretty consistent with what one would expect, and what I imagine most DMs resort to when populating a hexmap and writing up wilderness encounter tables.
But this does raise the (to me, at any rate) curious fact that, for all that there are more blogposts and print and digital products about how to create, populate and run hexmaps than there are grains of sand on all of the beaches in the universe, I don't believe I have ever read anything by way of principles guiding selection of terrain types for hexes. All I've ever seen are rough and ready preconceived ideas put into practice. People just do it 'by eye' and don't, I think, give it a great deal of thought. Nobody really seems to think through, in a reasoned way, why the division is between temperate forest, temperate mountain, etc., but not, for example, temperate deciduous versus coniferous forest, or dense versus light forest, or mountains above and below the treeline, or whatever. Likewise, nobody really seems to consider whether there are meaningful differences in terms of movement rates or the population of wilderness encounter tables between, say, a temperate plain versus a steppe, or tundra versus arctic, or whatever.
Without having thought the matter through a great deal, and therefore offering this post as a conversation starter more than anything else, I will suggest there are probably three main approaches to the matter:
- The realistic approach: the DM simply tries as hard as he can to come up with a list of terrain types which reflects actual real-world complexity as he sees it. If he thinks a forest flatland is different to forested hills, or a rural piedmont is different to a rural plain, then he duly deploys different hex styles to reflect those differences.
- The game management approach. The DM uses as many terrain types as is practically useful given the nature of the campaign. If the hexmap is mostly jungle, then he may divide the terrain into dense versus moderate versus light jungle; big and small rivers; bogs versus marshes versus swamps; forested hills; etc., so as to maintain variety. On the other hand, if there is already natural variety within the hexmap (because one area is deserted, another mountainous, another more temperate, etc.) then there is no need to get too fine-grained.
- The monster-availability-led approach. Rather than starting off with the map, the DM makes a list of all the monsters he can think of and, from that, decide how many terrain types are meaningful. Why draw a distinction between arctic and tundra if the number of cold-place monster varieties is very small and the wilderness encounter table will therefore largely look the same? Why draw a distinction between dense versus light forest if there are only so many forest-dwelling monsters in the bestiary? And so on.
Have any readers given it any more thought than this? How do you decide how many terrain types your campaign map will have, and how fine-grained or abstract do you tend to get?
Why yes, we readers have thought about this, as it happens! Here's what I've thought, scribbled into notebooks, and not put properly into practice.
ReplyDeleteFirstly, there's a whole science of ecology that has produced biome classification systems, so if realism of some kind is the goal, it would be foolish not to start with one of those, and then adapt it for use at the table by lumping and splitting where the needs of running a game so dictate.
Secondly, I didn't do that, and approached the question from much more of a game designy perspective: just like we define monsters in large part by numerical values on a set of orthogonal dimensions, we can do the same with terrain. I had numbers from 0 to 3 for vegetation (desert, sparse, moderate and abundant) and ruggedness (flat, undulating, rough, and impassable), and then modifiers of +-1 for aridity and elevation, those being so closely correlated with vegetation and ruggedness that treating them as separate factors seemed inelegant to me. Rivers, lakes and coasts were to be treated separately but I never got around to codifying that. This scheme was for a fairly small campaign setting with no significant latitude differences, so I didn't have a latitude/temperature variable but adding one would be pretty obvious.
Those rating were combined in different ways to get the difficulties of rolls for travelling and foraging, and contributed to probabilities for random encounters, starting distances and surprise.
In practice what happened was what usually happens with these highly rational complex systems: I used it for a while in its unfinished form, never got around to tying up the loose ends, and the players kept puttering around in the same 4 hexes that once they finally decided to move on , I couldn't be bothered to use the system when expanding the map, and just winged it. And now the campaign just came back from a one-year-break and the party have been exploring the same dungeon for almost two years in real life, and I'm not sure they're ever going to see more than a couple of hexes!
Haha, great comment. Yeah, it's funny how these things work out. Or never do... I really like the vegetation/ruggedness/aridity/elevation thing.
DeleteI'm currently half-wrestling with this in sketching out a West Marches map. With the idea that every region is likely to have a unique random encounter table, and regions should have clearly distinguishable edges so that players can get some sense of difficulty the way they would from levels in dungeons, it's hard to have a "west half of forest / east half of forest" or "north half of forest / south half of forest" split; either there's a clear terrain change, or an ecosystem change even though it's "one forest".
ReplyDeleteBut it's also a problem I don't have to solve entirely up-front, and shouldn't let analysis paralysis stop development. In both of these forest-near-starting-town cases there is a trail or a waterway that marks a hard border between the two halves, so I *could* skip the biome change, but that doesn't sit right with me. Or just have an invisible gradient and see how well it works - seeding rumours can forewarn the players. Also:
4. The tool-based approach: The terrain types that are available in mappers like Worldographer or HexKit give you a base set; you usually want only a subset of them, anyhow.
5. The Tool UX-based approach: In HexKit you can have layered terrain types in a hex - put down a mountain, then a forest, then a settlement - but this turns out to be very hard to read for some of the tilesets. The players have been going through a forest recently and we keep getting confused about where the edges are - is this bit of forest flat, hilly, or mountainous?
Yes, good point about Worldographer, etc. I suppose the obvious way to split geography with difficulty levels is to have conveniently placed rivers and mountain ranges. Although I guess there's nothing wrong with rumours like, 'As one goes deeper into the forest, things get more dangerous...'
DeleteIn Chanbara, I used very basic terrain types, but included different lists depending on the season. Japanese culture puts a lot of emphasis on seasonal distinctions, so I thought it was appropriate. We never ended up doing much hex crawling, though, so the players never got to experience the changes.
ReplyDeleteLists by seasons -- that's something really rare to see.
DeleteI also create encounter tables based on specific regions of my campaign map, not really grouping them by a terrain type. West Mountains have different encounters than Grey Mountains in the south.
ReplyDeleteYes, unless your hexes are very large having encounters solely by terrain type seems counterproductive. First, creatures in a region are not going to avoid one particular hex because it happens to be light deciduous forest vs. heavy coniferous. Second, the possibility of different terrain types gives another element of variety to the encounter. Third, if all terrain of a particular type on your map is populated from the same encounter table then there's no distinction between regions and no difficulty curve; it's popular to call it "West Marches" style now but the idea that the deeper you push into the wilderness the more difficult things get is an old one and a good one.
DeleteIn my most recent campaign of this type I had ten or so regions each with five or ten terrain types each with a weighted encounter table for daytime and night time, and had a devil of a time thinking up enough "natural"-seeming monsters to populate them all, even with the overlap.
I chose the actual terrain of the hexes for navigational purposes: light deciduous vs. heavy coniferous doesn't matter for encounter purposes, but it DOES matter for when the players are mapping and trying to find their way. noisms doesn't seem to have addressed this in his post. A forest peppered with glades, hills, valleys, tarns, thickets and streams is an entirely different navigational prospect than a solid block of fifty "forest" hexes.
There's nothing wrong with sharing some monsters between the tables, and having others that are not shared. Orcs would live in the forest of the nearby plains, for example, but forest giants wouldn't.
DeleteFor wilderness encounters one is mostly concerned with animal life, so the plants are of interest only to the extent they influence what animals live there. In North America (the only ecosystems I know much about) there isn't any difference between plains and hills in terms of what animals live there -- bears, panthers, bison, deer, etc. So a more useful nonfictional starting point would be what animal ranges in the real world match your fictional setting.
ReplyDeleteSee comment above. I think clearly there are some animals that are more adaptable than others, and this would likely be true of monsters too. Leopards live in a very wide variety of habitats. But okapi don't. Similarly, while orcs might live in a wide variety of habitats, desert trolls don't.
DeleteOD&D doesn't even have a Hills terrain, so neither does my game. (And actually it makes sense to me; most "clear" terrain is liable to be rolling to some extent unless your game is set in Kansas, and my sense is that moderate hills don't meaningfully slow down travel on a days and weeks scale.)
ReplyDeleteSpoken like a man who has never been to the Lake District or the Grampians. ;)
Delete