Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[Biopass]

Here's some weird news. The large world with no werebeasts is already in year 790, spitting out a new year every 10 seconds or so. Still in the age of myths, and population only just hit 300K. The gen is the fastest I've seen on a large world so far. As another interesting note, there are only 1.4 million events. So that should give you a good idea of exactly how much of an effect they have, and how much of history they make up.

Whatever the case, I think a nerf is definitely in order. Maybe the longer a werebeast is around, the less chance bites have of carrying infection? And the longer the werebeast generation, the weaker the bite is, so a 4th-generation werebeast has much less chance to infect? That would stop the exponential growth and make werebeasts a much less common thing. Vampires and lycanthropes are almost mundane now simply because there are so many. The vampire kill numbers people have been putting out there also point towards a needed nerf in the mechanics.

[xaldin]

Here's some weird news. The large world with no werebeasts is already in year 790, spitting out a new year every 10 seconds or so. Still in the age of myths, and population only just hit 300K. The gen is the fastest I've seen on a large world so far. As another interesting note, there are only 1.4 million events. So that should give you a good idea of exactly how much of an effect they have, and how much of history they make up.

Whatever the case, I think a nerf is definitely in order. Maybe the longer a werebeast is around, the less chance bites have of carrying infection? And the longer the werebeast generation, the weaker the bite is, so a 4th-generation werebeast has much less chance to infect? That would stop the exponential growth and make werebeasts a much less common thing. Vampires and lycanthropes are almost mundane now simply because there are so many. The vampire kill numbers people have been putting out there also point towards a needed nerf in the mechanics.

Well you're confirming what I'd posted that a large is certainly doable though my system stayed perfectly stable at 1.25-1.5 sec per year after about 500 until end the idea is the same, it can be generated, without the were critters. Trade routes have slowed it compared to prior DF versions but really, it is all about the were critters making a mess of the world.

[NW_Kohaku]

Here's some weird news. The large world with no werebeasts is already in year 790, spitting out a new year every 10 seconds or so. Still in the age of myths, and population only just hit 300K. The gen is the fastest I've seen on a large world so far. As another interesting note, there are only 1.4 million events. So that should give you a good idea of exactly how much of an effect they have, and how much of history they make up.

Whatever the case, I think a nerf is definitely in order. Maybe the longer a werebeast is around, the less chance bites have of carrying infection? And the longer the werebeast generation, the weaker the bite is, so a 4th-generation werebeast has much less chance to infect? That would stop the exponential growth and make werebeasts a much less common thing. Vampires and lycanthropes are almost mundane now simply because there are so many. The vampire kill numbers people have been putting out there also point towards a needed nerf in the mechanics.

The problems with werecritters could also be simplified by making werecritter rampages less frequent, while upping the carnage when they do rampage.  We don't need 50,000,000 instances of "the civilian ran away". 

Toady could also do the most sane and reasonable way to control werecritter populations from an in-universe perspective: When over a certain number of rampages occurs per year, humanoids band together and go werecritter-hunting, and send out heroes more often with the specific goal of finding werecritter lairs and killing them when in their weak naked peasant form. 

Many of these towns will have several dozen werecritter lairs dotting the landscape around the town.  In fact, just playing adventurer, it's almost impossible to fast travel without ramming headlong into another freakin' lair. 

[Biopass]

That's also a good idea.

So the 0-werebeast world is here: http://www.mediafire.com/?pec0t08z44glych

It has the lowest population I've seen in a large world, along with the lowest number of events. This is also the longest I've genned a large world for, the full 1050 years. I'm guessing it's the geography again, breaking up civs that would otherwise grow and expand at ridiculous rates. These two resulted in the relatively short worldgen, compared to the 16 hours on my previous large world that only made it to 400ish before I had to kill it.

Any other world requests or experiments people want done?

[NW_Kohaku]

Any other world requests or experiments people want done?

Actually, yes, if it isn't too much trouble.

Could you try restricting the amount of items that are available to civs to trade?  For example, try restricting the number of raw items that they can use (like the number of clothing items), but then, also restricting what materials they can use (by editing down the raws so that there are far less creatures to make creature-leather or food items out of, and less stone or minerals to trade in).

I wonder how much simply having 500 different KINDS beasts to make into leather, and over a dozen different body part stocks to turn into prepared grasshopperman brain as separate from prepared mosquitoman brain as different from chopped mosquitoman liver bloats the trade process...

[NW_Kohaku]

There is now a suggestion thread to try to suggest solutions to the problems in worldgen, so if you have any sorts of insights into how to make a faster worldgen, please participate in the thread in the link.

[thvaz]

I hope Toady notices it, he didn't mention the werewolves problem in his devlog today...

[KillzEmAllGod]

the more werecreatures the more that are made.. basicly.

[dirty foot]

the more werecreatures the more that are made.. basicly.

It mathematically follows an epidemic progression.

[knutor]

Is there anyway to forecast fullmoons?

[Proteus]

Some maths with my current world generation

I just got wrote down a couple of datapoints (minutes progressed/year)  of my current world generation (large world with a few changes, planned to go till 300) in excel and let it approximate a logarithmic function with them ...
the result is (with 35 datapoints picked, between 224 and 278)
year = 33 ln(minute) + 114,02
(which gives us
minute = exp((year-114,02)/33) )

with other words, my computation according to this equation will be complete ~280 minutes after I started it, with one year, at this time, getting calculated in 8 minutes

If we assume that world generation would follow the same approximation one might conclude that I would need 21 hours to calculate 350 years and 4 days to get world generation to calculate 400 years and we would need 3.5 million years to compute a world which is 1050 years old

(one might however add that
[list=od

• the inclusion of further data points seems to suggest that the approximation formula tends to get shifted, with the factor gaining weight and the offset losing some, resulting in the approximation of needed minutes getting corrected downwards
• with more years getting calculated some of the limits entered into the parameter file might kick in, very probably resulting in [minutes needed] / [year calculated] getting maxed out at some point, making the formula more and more unreliable as world generation progresses
• every world generation is individual ... the night before for example I had a 350 year old world getting generated in 6 hours ... unfortunately I wasn´t around when it finished and windows, due to DFs inactivity assumed, that the program had crashed, giving me no chance to save the generated world I also had worlds with an age of 300 years getting generated within 2 hours, so my current world generation can definitely be seen as some kind or worst case scenario
• Other computers with more GHz per core (mine has 3.4) might get slightly better results, even if calculating exactly the same world

[Vherid]

Currently genning a Default settings with 34.03, 257x257, 1050 year history, all medium except ores set to frequent and it hasn't been taking that long at all, almost done. It's been ~30 minutes I think and I'm at year 939 with 1,414,568 events.

Done.


It crashed on accept. Retrying with LAA.

Second try, this time with LAA on, program just instantly closed on accept.
Third try with LAA on, successful. 257x257, 1050 years, everything medium, frequent ore.

[simonthedwarf]

I'm able to consistently generate worlds, but 1000+ moves along very slow.

[Garath]

Toady updated the game a lot since this thread started (and since someone last posted in it) and world gen is handled a lot better already. In 34.03 I had troubles getting past year 300 in a medium world, but in 34.07 I made a 1050 years large region with mostly land, so lot's of civs spreading.

[PTTG??]

I reached 2000 with little problem.