There are two different ways to design these kinds of games and rules.
You can have the relevant central data-structure to be the set of diegetical entities that have been shared by at least two participants. (I.e. the Forge definition of the SIS.) You create rules that inject entities into that set, alter entities in that set, or remove entities from that set. AW is set up that way. This leads to rules such as
Whenever there’s a pause in the conversation and everyone looks to you to say something, choose one of these things and say it.
All well and good as long as you only need to make rules that directly operate on that particular set.
However, when you want to regulate things that are more… offscreen… this design architecture starts to show its limitations. Your rules become indirect and multi-level. Threats in AW have their own list of moves operating directly on that primary data-structure and setting them up means creating new operands on that primary data-structure.
As a concrete example, here’s AW2e on page 118 creating a new operand on the primary data-structure:
When you go into Dremmer’s territory, roll+sharp. On a 10+, you can spot and avoid ambush. On a 7–9, you spot the ambush in time to prepare or flee. On a miss, you blunder into it.
Instead of creating a territory and an ambush team and injecting it into an offscreen canon set of diegetical entities, in AW you set up & prepare & canonize a new rule that operates on the onscreen uttered set of diegetical entities, the SIS.
The threat map is a way to create and organize such rules; rules that are direct operands on the primary data-structure. The secondary data-structure in AW is the list of such rules. This includes the agenda, principles, moves for both MC and players, and also the custom moves created & organized by the MC during prep. And by “created”, some are just selected from pp 107–115 and then injected into the secondary data-structure.
I don’t mean to be picking on AW. It’s a very disciplined design, much more so than many PbtA games, which makes it an excellent and very clear example for the point I’m making. I hope you see what I mean that almost every rule in AW is a direct operand on the primary data-structure, and the secondary data-structure is the list of such currently applicable rules.
In our model (blorb), we want to have rules that operate directly on the (much larger) data-structure of “the set of diegetical entities that are considered canon for this campaign” so that’s why we make that
our primary data-structure instead.
An example of operating on that data structure would be
(For those that are unfamiliar with D&D: they have stealth +6 and multiattack with longbow. They love hiding and shooting.)
In Dremmer’s territory there are three Scouts (Monster Manual p 349).
This fact can be added to the set of canon entities directly and straightforwardly (moderated by rules such as “no Paper after seeing Rock” and “separate prepping from running”).