Variations on a Theme: Adapting Polyhedral Pantheons

A-Z 2015 "V"

I’ve been doing a lot of work on Polyhedral Pantheons this month. The processes and mechanisms were devised around creating a pantheon, but they can be adapted to other purposes.

Primary elements of the process and its results are:

  • Attributes are placed on sites (points and/or faces) of a polyhedron.
  • Entities are described based on the attributes assigned to each site.
  • Entity attributes are ‘coherent’, in that nearby entities will share some attributes.

I can think of several things this can be used for.

The In Cörpathium Mechanism

In Cörpathium
In Cörpathium, click to read the excellent article

The polyhedral allocation mechanism used in Polyhedral Pantheons can be combined with Last Gasp Grimoire‘s In Cörpathium dice drop mechanism.

  • Assign twelve attributes to the points of a d20, and each face represents an entity of interest. In In Cörpathium these are ‘boroughs’ (‘wards’ or ‘neighborhoods’ if you prefer).
  • Build a d20 table with the resulting entities.
  • Roll dice (In Cörpathium uses a ‘seven dice set’ — d4, d6, d8, 2d10, d12, d20 — plus more d20 depending on how big the city is), keeping them reasonably close together.
  • The number on each die indicate which entity was rolled.
  • The physical placement of each die indicates the relative location of the entity.
  • The orientation, the ‘corners of the face’ of each die, is used to determine what each entity is related to.

Last Gasp has rules for resolving duplicate rolls and ‘mandatory selections’ (some wards are always present in Cörpathium, and some in particular locations, and as I recall some specific relationships).

This mechanism is referenced fairly heavily below. It could also be used in Polyhedral Pantheons, and I expect I’ll include that when I write ‘Third Pass’ development (building the mythology), but I haven’t discussed it yet.

Adaptations of Polyhedral Pantheons

Here are some ways the Polyhedral Pantheons mechanism could be repurposed.

Physical Geography

Assign terrain types, major inhabitants, and major geographical modifiers to each point of a polyhedron. The faces could then represent geographical regions with those attributes.

A face with (forest, elf, magic) is so easily recognized it’s cliche. The  adjacent faces (forest, elf, ruins), (swamp, elf, magic), and (forest, goblin, magic) are perhaps not as common.

I would use this mechanism to identify what regions are present, but I would probably would not try to reflect the physical relationships suggested by the polyhedron. That is, I would not try to keep all forest faces together, all elven faces together, and so on. The only way to do that completely is to use the polyhedron as the geography. It would probably be more useful to populate the polyhedron and use a subset of the faces, placing judiciously.

Given the icosahedral (d20) point allocations below,


we get the following faces.


Between them I have identified twenty region types that might be present on this map. I might decide that Face 1 is the elven homeland in this geography, and Face 3 is the hilljack homeland. Face 6 is overrun by jhesiri, and the kouzelnik are rooting around Face 19 while groups of them are fighting over something in Face 9.

I haven’t explored this idea much yet, but I suspect that perhaps a third each ‘terrain’, ‘major inhabitant’, and ‘modifier’ would work pretty well. The example I through together has only two races (elf and goblin), two terrain types (forest and hills), and seven modifiers (magic, city, resource, war, holy, unholy, haunted, and ruins).

The In Cörpathium mechanism could work well here. It might not, because the In Cörpathium mechanism probably works better when there can be more artificial divisions, but I’d be prepared to give it a try.

Planar Geography

Planar geography is in some ways even simpler because the combinations don’t need to really make sense, and it can be appropriate to have the resulting geography directly reflect the polyhedron used. In fact, I have long used a planar cosmology that has the elements arranged ‘tetrahedrally’ (that is, equidistant in a notional three-dimensional space), much as with the Elemental Tetratheon. I can easily imagine using that very arrangement (for the elemental and alignment planar traits) as a starting point for a planar cosmology. Three planes of dominant fire, three of dominant earth, and so on, and three each of good, law, evil, and chaos. I still wouldn’t have an evil plane of fire here, just as there are no evil fire deities in the Elemental Tetratheon, but that’s fine.

Come to that, the ‘least deities’ of each subpantheon of the tetratheon, the ones on the ‘points’ touching two faces of that element (i.e. the last three listed in each subpantheon) could be comparable to the para-elemental planes. This gives me twelve more planes, one each being the path from one specific element to another specific element, without being a route back.

The In Cörpathium mechanism could work well here. In some ways there are even fewer restrictions on connections than with geography, but there are also fewer rules about ‘sensible placement’, and almost nothing about what links make sense and what links don’t.

City Wards

This might be a narrowly-focused application of physical geography, above. The attributes might focus more on the nature of the wards of a city: trade, religion, wealth or nobility, military, etc. Or the local population, by race (and the attribute highlights difference from the local norm: you might have a halfling enclave in a city otherwise full of humans, so you only mention the halfling attribute because it’s different and you don’t have to have a ‘human attribute’).

In this case you might have some that retain the physical relationship suggested by the polyhedron. If you have a small enclave of halflings they might all be clustered together, while the market areas are scattered around the town.

I suspect this could work well, perhaps very well, when you have settlements with common characteristics. For instance, within a particular kingdom you might use the same attribute polyhedron (and thus the same ‘d20 table’) for all cities in the kingdom. You might vary the number of dice based on city size, any particular ward type might or might not be present, but there will be a strong commonality to the cities that could make then quite distinct from those of a neighboring kingdom.

The In Cörpathium mechanism can be expected to work well here. This is what it was made for.


This could be a powerful tool for creating NPCs with similar nature. Use attributes based on personalities, affiliations, goals, locations, and perhaps conditions (such as ‘secret’, likely indicating that the NPC either is hiding something, or has learned something). Again this gives some coherency to the result by allowing for multiple NPCs with similar nature to be present, giving a pattern of sorts for PCs to recognize.

The In Cörpathium mechanism might work here, but I expect that a GM would really only use it when looking for ideas for a scenario. If the scenario is understood and the GM is just looking to expand the cast then the primary relationships would likely already be known, but some secondary relationships, or relationships between secondary characters, might be randomly generated. Physical location of the dice is likely of lesser importance as well.

13th Age-Style Icons

You might use this mechanism to build a series of icons similar to those in 13th Age. Assign the attributes to each point and then build around that. This is much like the Characters section above, but the resulting entities have much broader scope and influence than you would normally see in NPCs.

The In Cörpathium mechanism could work better here than would be expected with normal NPCs because of the expanded scope they have. The NPC relationships are fairly constrained by their relative lack of scope and relevance apart from the scenario itself, but icons have a much longer reach. Still, I expect I would probably want to make most of these decisions myself.

Or perhaps not. Icons are not as powerful as gods, but they take a similar place in a 13th Age-style setting. Perhaps it would be a good approach.

  • I used polyhedral process in building my Icons for the Sandbox series and it worked every bit as well as I could have hoped. Icons stand as focal points of a setting and I feel must share some themes or be somehow connected. Polyhedral process works wonderfully here.


This mechanism probably isn’t so useful for a regular dungeon, mostly because a regular dungeon is likely to be too internally consistent to really benefit. That is, if the dungeon is a goblin warren you’re likely to find ‘just goblin warren stuff there (mostly)’, and a white dragon’s lair is probably going to be similarly homogeneous.

A megadungeon, though, is almost always broken down into regions that differ, sometimes wildly. The node-based megadungeon has a dwarven safeholddark temple, mad wizard’s laboratory, dragon’s lair, aboleth enclave, and another half dozen regions more. I can see the polyhedral mechanism being a very powerful tool for an initial outline. One point might be ‘outdoors’, another ‘underdark’ (indicating entrances and links to the underdark respectively), then various region types or characteristics on the points.

The In Cörpathium mechanism would be a wonderful fit here for determining relationships between the regions. The physical placement might or might not be relevant, but in a megadungeon almost anything can connect to almost anything.


I define a campaign as a series of related scenarios that has a beginning and an end. Open-ended adventuring might happen in a ‘setting’, but a campaign tells a story. I suspect many people might call this a ‘story arc within a campaign’. As a result, I would expect a campaign to include repeating and shared elements from adventure to adventure. Not all attributes need to be present in each adventure, and not all adventures will even get used (but each costs me maybe 10-15 minutes if I don’t use it, just enough to make notes about what it is and move on).

This sounds like it should be a good fit. Put attributes relevant to the theme or purpose of the campaign on the points of the polyhedron, then the faces are the ‘adventures’. For the Donnerkonig Heirs campaign ‘artifact’ might be one of the attributes (there are four artifacts associated with the Donnerkonig, and they need to be retrieved), ‘gateway’ (adventure that potentially leads to another campaign), and so on. This identifies something of the nature of each adventure that I can build on.

The In Cörpathium mechanism fits quite well here, I think. I try to design campaigns (and each adventure) to be as nonlinear as possible, so the node-based graph that comes out of this can be a good fit. It also encourages me to get some variety in the adventures while still allowing me to constrain them to certain characteristics. I can still do my high-level design in an abstract manner and ensure all required adventures are present and accessible, without forcing them into an ‘adventure path’.

Closing Comments

I suspect that the d12 would likely see less use than normal, mostly because making five disparate attributes work together can be hard. However, having twenty attributes available to work with might be useful.

In all the discussion above I had attributes always on points, with the faces being the result. Because the points are often descriptive attributes I don’t expect to make them entities, so assigning other attributes to faces is pointless nonsensical. However, I can see times when putting attributes on faces could be useful. Any attribute assigned to a point will be associated with multiple faces. Assigning an attribute to a face means it applies to only that face (because you are not creating entities for the points). You might put a race attribute on a face, for the one ward of the city where that race lives. You might put a haunted attribute on a face for a region full of ghosts.

For that matter, the ‘attributes on the points’ don’t need to be descriptive. They might themselves represent entities relevant to the process. If I had a more developed setting for the Donnerkonig Chronicle I might create a polyhedron with the various entities in the setting on the points and use it to decide which of them (and it would be more than one much of the time). Similarly, I can easily, easily imagine in a 13th Age-style setting having an ‘icon polyhedron’ with the twelve icons on the points (the Prince of Shadows can appear anywhere, so leave him off) that might be used to indicate which ones might be interested in any particular event.

I have spent a lot of time working on Polyhedral Pantheons, but the underlying mechanism could be used in many other ways.

One comment

  1. Pingback: 13th Age-Style Icons in the Sandbox, Part 2: Definition and Planning | In My Campaign - Thoughts on RPG design and play

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.

Back to Top