davidlallen
Deity
You know what I'm not a big fan of - the starting spot being pre-settled, and the maps not generating *usable* starting locations. With the amount of plains and desert, sometimes you are stuck with a capital with no food resources and lots of plains hills. If you gave a starting settler instead of a automatic start location, you could move to a more hosiptable place. I would imagine that in a future scenario like this, people would congregate at the few "gardens of eden" with desolate wastes in between..
I thought about this a little. The problem with giving a settler is that people can choose to explore a little and then pop down the first city next to some ruins. Even if you don't *know* there is a depot/airfield there, it is highly likely. That gives you a big leg up on getting rebuilt units.
While working on 0.7, I added some print statements into the mapscript to try to make a solid rule of what is a "bad start location". Or the opposite, what is a "reasonable start location". I did not come to any conclusion, but I agree it is a problem which needs to be solved.
Each plot in the BFC has an "unimproved" food value and an "improved" food value which is probably higher. Any kind of improvement takes a tech, which means it's not available right at the beginning.
I tried making a total of the unimproved food in the BFC. But, that's a high number and bad locations don't stand out. I tried counting if there is any surplus, unimproved food -- without a farm, only grass plus some bonus like cow counts. Those are actually pretty rare.
When I turn on the vanilla routines to help this, I get a big smear of grassland around the start position, which *really* stands out. Refar pointed out, and I agree, that if there are exactly N gardens of eden for the N players, and every other location stinks by comparison, then a rush becomes the best alternative. So I want to avoid making the starting locations *too* superior.
Can we work out the exact rule, to determine if a start position needs help?