I’m writing a grand technique sport that’s typically performed from a zoomed-out, “strategic” perspective. With battle encounters, I need gamers to have the ability to zoom in, in the event that they select to, letting them see a “tactical” map with particular person models transferring about on a 2D battlefield.

Think Total War, for argument’s sake.

Now there could also be a number of battles occurring in parallel in numerous areas of the world. The tactical view, furthermore, will probably be fairly heavy on AI (for pathfinding wants and different concerns), in addition to operating at a extra granular clock pace (minutes within the tactical view vs hours within the strategic view). It is deliberate to be comparatively compute intensive.

The query is learn how to resolve unseen battles – these battles that the participant doesn’t select to view, or can’t for FoW causes. One possibility is to create a second, much less advanced, summary mechanism that resolves unseen battles utilizing fewer computations (akin to Paradox video games), after which to tune each mechanisms in order that they provide comparable statistics. Is this a viable technique?

My unique thought was to run the tactical AI within the background for each battle that happens, however I think that won’t scale up very properly.

(Since I discussed them already, does anyone know the way the Total War sequence does it?)