My game, has the map divided into the following types:
1 Massive Continent (islands are all so close by, they're included as one continent.)
8 Kingdoms (All with entirely different Biomes- Mountains, Plains, Tundra, Jungle, etc.)
20x16 Regions, which are composed of 100x100 miles of grid. Although there is 3.2 million square feet of map, there is actually closer to 2.6 million square feet of land.
10x10 Sections inside each region, composed of 10x10 miles of grid. So each region, has a 10x10 grid inside of it.
Each section, contains a specific type of area, such as: Plains, Mountains, Farmland, Grassland, Hills, Swamp, Forest, Jungle, Volcanic, etc. The type of terrain which consumes the section the most, in case of terrain type transitioning, is what the section is considered for the sake of the game's simplicity.
I assign population density to every single section type. Since sections are tied to regions, the regions can then tell me the population, such as if a City or Town should be there. Villages would not be drawn, as there are thousands of them. They are just assumed to be all over the place.
In a simple equation, based on the section's data, I can translate a world map into a realistic population, and then take that population to create City Maps and Town Maps, based on population of the town and number of specific jobs in each city/town.
Then I just hope and pray that the number of towns is manageable. If not, make them randomize too, or use a set of various "town templates" from a pre-created handful of towns and some random generation for jobs.
The MDME, although giving me crazy formula populations, is still a great source for me to create a living populous inside what is already my created World-Map.
Is it safe to assume that the facts are still legit, even if the suggested population density numbers are too high?