The Great Antarctic *crawl pt. 12 - The Dreaded Demography Diversion
How many people live in Antibor? The usual D&D approach is to either not think about it (100% valid) or to use some kind of generic table with settlement population ranges. But I've been having so much fun generating polities that I wanted, as a final flourish, to get some quasi-realistic population numbers down.
There's some evidence that populations obey a power-law distribution at multiple scales. I decided to to run with this and distribute populations to the settlements such that they obey a power law overall and (roughly) within each polity (at least for those with the largest populations).
This function allows one to map a set of uniformly distributed random values to a power law distribution. It requires bounding the range with a maximum & minimum value and an exponent that determines how dramatic the falloff from small-common to big-rare is.
I put the minimum value arbitrarily at 33 -- low enough include things like manor houses, monasteries, & tiny hamlets in the 'settlement' category, but big enough to be worth mapping.
Picking an upper value and an exponent was a headache and I spent too much time down the maddening rabbit hole of ancient/medieval/early-modern demographics. Obviously, everything hinges on assumptions about technology and social organization that I haven't clearly defined and probably never will. It's further complicated by the fact that Antibor isn't uniformly inhabited, so average population density is harder to interpret (are we talking about population density over the whole continent, or population density in the inhabited regions?).
For ruinous, Dying-Earth vibes, I opted to err on the very low side. A max value of 333,333 and an exponent of the golden ratio 1.618 (just for fun) gave me a total population of 9 million. This puts it somewhat below the density of the former Soviet Union (e.g. European Russia + Central Asia + Siberia) during the early modern period. Since Antibor is mostly wilderness & steppe, this feels acceptable.
Here's the breakdown in terms of size and population share:
| Population Range | # Settlements | # People |
|---|---|---|
| 33-999 | 4,901 | 0.8 mil |
| 1,000-9,999 | 510 | 1.5 mil |
| 10,000-99,999 | 124 | 3.6 mil |
| 100,000+ | 20 | 3.4 mil |
It's too urbanized to be a proper ancient/medieval/early-modern setting. (The issue stems from simply not having enough settlements on the map -- to get a reasonable population density while also keeping urbanization low, I'd need to either place way more tiny farming villages or assume that there are unmapped hamlets scattered around.) But this being a sci-fantasy elfgame setting, I think I'm ok with it -- and I have a few ideas for justifying the stark urban/wilderness divide.
To assign the population values to cities, I did a weighted random shuffle using each settlement's arable land score. Polity capitals got a ranking boost that scaled by the number of subject cities under their yoke.
Here's the resulting map:
Despite the issues mentioned above, I'm very happy with how believable it looks. High-population cities are clustered near arable land, but the biggest cities are fairly evenly-distributed (and surrounded by smaller subject cities).
Addendum: Using this Info
As alluded above, I'm inclined to ignore population figures unless it's critically relevant for some aspect of gameplay. Not only are 'realistic' medieval demographics a bit of a mindfuck, they're also a form of information only available to the local elites and their tax collectors. I'll be storing the populations in my Antibor .csv file, but in the city descriptions I intend to stay vague: 'a town of several thousand', 'a teeming city said to be home to many tens of thousands'. The exception will be smaller sites, where knowing the exact number of bandits is useful because the players might fight them all.