Long-run issue: modelling a spherical Earth for global games?

[Maklodes]

I'm wondering how a global strategic map would work with an (approximately) spherical world.

The traditional approach from global strategy games, like Civilization, was to model the Earth, implicitly, as a cylinder. The poles are effectively absent. Traversing the southern "edge" of the map 360 degrees is just as long as traversing the equator 360 degrees. Other options included a plainly flat world, and an odd toroidal world, where, effectively, you could cross from the northern edge of the map to the southern edge in much the same way that you can cross from the eastern edge to the western edge.

A spherical (or even ellipsoidal) world is possible, but might lead to some (mildly) complicated computational issues, as far as Great circle arcs go (making pathfinding buggier and more complicated than before, when combined with pre-existing factors like mountains and usch). It would also effectively require abandoning any type of traditional "grid" or "hex"-based system in favor of continuous locations, which might make questions like "when do two armies actually encounter each other" more tricky. Adding "skin" to a spherical world is also an issue. One might use a sort of plate carée projection, then project that back onto the sphere, but that would necessarily make more polar latitudes more detailed than more equatorial ones.

One approach I'm partial to, which I think captures (imperfectly but fairly well) most of the features of a spherical world while preserving the possibility of a hex or square grid system is an icosohedral world. It would have twenty triangular flat faces. On each face, the traditional grid systems would work normally. At the intersections of the faces, there would be some mildly challenging issues of how crossings would work, but I would guess nothing insurmountable.

Thoughts?

[Razor]

It's an interesting point you raise.

Despite its cons I prefer a cylindrical approach to the spherical approach. Political maps usually make use of a rectangular projection which works best for this kind of stuff. A spherical projection of the world would obviously be more realistic, but no one really looks at the world as a sphere but rather as a rectangular map. Personally I favor the Miller projection.

Also, both the North and South Pole don't really play a decisive role in Total War games and wars, unless you'd want to have a Total War that covers scenarios set in the present and the future that incorporate global warming and Intercontinental Ballistic Missiles etc. But even then a cylindrical approach would still be somewhat possible although it could lead to awkward situations. Personally I don't see a battle being played out on the North Pole or South Pole due to their difficult terrain and environment.