Why not start by placing players in map generation?

[BoredCitizen]

I'm not a fan of the map generation algorithm. My understanding is that it works by first generating the map, and then giving score to plots in order to decide where to place the players.

I was wondering why that is, as many of the problems (I have) with the resulting maps can be solved by first choosing locations for the major & minor civs, and then generating the map and features around them.

I started working on an algorithm (barely got anywhere so far, as I'm reading up on some interesting theory first), and I expect that in practice, this approach has its own drawbacks compared to the current method.

Just wondering if anyone more experienced/familiar has any insight which might save me from wasting my time, or help me in my endeavors.

 

[Gedemon]

not the first time that I see this idea mentioned, IMO it's not a bad one, but if I had to work on it, I'd still start by generating the landmasses, then start positions applying some bias (anticipating terrain/features bias using latitude), then everything else.

 

[BoredCitizen]

not the first time that I see this idea mentioned, IMO it's not a bad one, but if I had to work on it, I'd still start by generating the landmasses, then start positions applying some bias (anticipating terrain/features bias using latitude), then everything else.

I figured I'd need to look a bit into how some features are generated if I wanted to make something usable.

"Not a bad one" is encouraging for me, guess I'll go for it