Bay 12 Games Forum

Please login or register.

Login with username, password and session length
Advanced search  
Pages: 1 [2]

Author Topic: WorldGen Science part 2 - beasts and civs  (Read 12984 times)

SAFry

  • Bay Watcher
  • Call me Seb
    • View Profile
Re: WorldGen Science part 2 - beasts and civs
« Reply #15 on: July 02, 2012, 01:06:08 am »

Gurkie, glad the research has been of some use  :)

I've been using the Legends Viewer, I was using the one in the Masterwork utilities pack but I assume it's the same one as this http://www.bay12forums.com/smf/index.php?topic=72702.0

I was looking at the 'alive' section which I must admit I didn't realise until halfway through this series of experiments only relates to living historical figures. However at 10 years, which is when history was set to end in these experiments, it seemed to work fine.

Couple of things; I forgot the parameter set I was using I had reduced all the good/bad biomes to 0 in my world recipe and also note in region 2 I reduced the caves to 0 and still got megabeasts and semimegabeasts. So I'm not sure how 'caves' relates to megabeasts, it may just be a leftover from the previous version. In all my experiments the numbers of beasts generated were in equal proportions.

So what have you found? 5000 beasts? How long are you letting history run? Even though they breed it seems that their numbers usually take a downward trajectory due to fighting. Also how are you finding lairs, in adventure mode?
« Last Edit: July 02, 2012, 01:14:29 am by SAFry »
Logged

Gurkie

  • Escaped Lunatic
    • View Profile
Re: WorldGen Science part 2 - beasts and civs
« Reply #16 on: July 02, 2012, 10:51:37 am »

Here's the worldgen I was using:

Spoiler (click to show/hide)

Basically, I run this worldgen with 2 years of history and then find a nice site I want to plant a fortress. Then I'll export the worldgen of the one I like and recreate it with as many megabeasts/evil regions/civs I can get away with. Then I'll go back to the embark spot and see if any lairs showed up, if not I'll regenerate the world leaving the terrain seed, but all other seeds are set to random. My current fort has a nice female dragon in a cave, an upright spoiler, 3 biomes, and a brook. I'd like to add some evil terrain to it, but the lairs always disappear when I do that, even generating it 10+ times. When I embark on a non-evil site and my map has 4000+ beasts, I usually get a lair within 1-3 world generations.

Thanks for telling me about Legends Viewer. Taking a look at my current world which has a lair on my embark site and no evil regions, I have 4936 alive creatures, 4000 of which are megabeasts. This is a 17x17 map set to make those 4000 beasts. But I was hoping it would make more lairs, Legends says I have 1644 lairs.

When I generate the same world with 100,000 megabeasts I get around 7000 of them and 3400 lairs.
Also, just did a 129x129 world with similar settings, and it had 12000 megabeasts and 6000 lairs.
« Last Edit: July 02, 2012, 12:54:43 pm by Gurkie »
Logged

SAFry

  • Bay Watcher
  • Call me Seb
    • View Profile
Re: WorldGen Science part 2 - beasts and civs
« Reply #17 on: July 02, 2012, 01:33:21 pm »

Wow, a pocket world with 4000 megabeasts!

Well you've given me some ideas for new experiments, looks like lairs have nothing to do with [SITE_CAP:18] or [MOUNTAIN_CAVE_MIN:5] [NON_MOUNTAIN_CAVE_MIN:1] and considering a 17x17 map actually only has 289 squares I'm a bit puzzled as to where 1644 let alone 3400 lairs would even go!?

By the way, do you do full region embarks when scumming for lairs? Takes a while to load but handy zooming into the right spot. When you find a really good one do us a favour and post it in the worldgen cookbook http://www.bay12forums.com/smf/index.php?topic=101280.0, I'd like to see it.

Gurkie

  • Escaped Lunatic
    • View Profile
Re: WorldGen Science part 2 - beasts and civs
« Reply #18 on: July 02, 2012, 02:58:20 pm »

A 17x17 map has 289 squares, but then you zoom into one square for embarking and that square contains 16x16, so it's 73,984 squares you can embark on. I'm a little disappointed with the pocket world, yeah I can get 400 megabeasts, but I can't increase forgotten beasts (I have 12). I'd do a larger map, but it already takes me 10-20 minutes to generate a pocket world with max megabeasts.

I don't look for lairs right away, I first find that perfect spot with lots of biomes, upright spoilers, river, flat surface, and good minerals. Then I'll export the world map gen parameters to get the seed, and re-gen that world to get different critters/evil/megabeasts on that same embark. I've given up on volcanos, since they always make the terrain too sloppy, and the flat ones are too rare to get everything else I want.

Can't do full region embarks, I always crash, haha.
I'll try to post up my current embark with the dragon. It's a sweet spot.

« Last Edit: July 02, 2012, 03:11:19 pm by Gurkie »
Logged

lorb

  • Bay Watcher
    • View Profile
Re: WorldGen Science part 2 - beasts and civs
« Reply #19 on: July 02, 2012, 03:39:43 pm »

Just posting to let you guys, who are making this thread (especially SAFry) and also the first part know that this is a great source of information for me. You make the worlds i gen fit my wishes better :)
Logged
Please be gracious in judging my english. (I am not a native speaker/writer.)
"This tile is supported by that wall."

bored german guy

  • Escaped Lunatic
    • View Profile
Re: WorldGen Science part 2 - beasts and civs
« Reply #20 on: March 16, 2017, 01:39:14 pm »

I did some rather tedious science on civ distribution and survival during world gen. Not sure that this is the right forum for that, but I wanted to share the results.
If there are no objections to my conclusions, I would probably put some of them up on the wiki as well.

Here's what I did:
-Let several world creations run on same world for 100 years
-Vary number of civilizations
-Vary number of demon types
-Vary number and size of caves


Here's what I discovered in worldgen:
-Yearly Beast attacks do increase smaller than proportionally to number of civilizations. To clarify: More Civs means more beast attacks, but fewer beast attacks per given civilization.
-Therefore survival rate of individual civilizations increases with number of civilizations (In my example: 24% of civs survive out of 15 starting civs, but 82% survive when 40 starting civs are set)
-The yearly number of beast attacks does NOT increase as Civs grow.
-The yearly number of beast attacks does decrease as civs fall (again less than proportionally)
-So it is possible, to create an exceptionally FUN world with loads of dragons and titans, and still have civilizations thrive
-Number of Megabeasts is not suitable to limit Civ-growth. Civs are either wiped out almost immediately or grow almost undisturbed(since civs grow exponentially, but beast attacks stay constant).
-Goblins seem to have the highest survival rate of any civilization (more than twice the survival rate of any other race, even though in my setting, there were only a few starting Goblin civs), Kobolds have the lowest survival rate by a small but consistent margin
-Number of demon types directly influences the starting number of Goblin Civs.
-No demons means no Goblins (that much is known already)
-Even a low number of demon types will give you one goblin civ
-Higher number of demon types set a limit on number of Goblin Civs of roughly one civ per four demon types for low demon numbers, other limits (like maximum number of civs in total) still apply.
-It could be possible that only certain types of Demons allow for Goblin Civs, since Civ number was not definite for a given Demon number
-Kobold civ numbers do not depend on maximum cave size

Suggestions:
-To get beast riddled world (medium size but mostly water in my case), with long history, surviving civs but no (or weaker) goblin dominance:
1. Set high number of megabeasts, semi-megabeasts and titans(71, 144, 44) → fun world
2. Set high number of Civs (40) → high survival rate of civs
3. Set high number of caves (61, 79) and high savagery → most Civs will be harmless kobolds
4. Set low demon type number ( 8 ) → decreases goblin civs, weakens their dominance
5. High savagery → limits civ growth without hitting maximum site cap or having a spammed world, which creates a more natural history feel
6. The rest is fine tuning of beast- and demon-numbers

Raw Data
"Vary Civ Number" worldgen results
Spoiler (click to show/hide)

"Vary Demon Number" worldgen results
Spoiler (click to show/hide)

World Gen parameters; incomplete because too long
Spoiler (click to show/hide)

Comments and Disclaimers
-The used world has rather strange geographical settings. Tendencies of tests are clear but absolute values are untypical of DF game.
-The used world is very savage in defined areas, severely limiting the maximum number of dwarven Civs.
-High number of caves allowed for many Kobold Civs.
Logged
Pages: 1 [2]