Some perspective

[Surly]

If you don't want to make things change too suddenly maybe you could give the animals a vicious dose of the sneezes (preferably a strain that didn't spread to characters too easily)?

Yay, let's introduce incurable disease for the animals as well.

[Doug R.]

I like the idea of animals dying off if they don't attack enough, but it is actually rather problematic. I tried to run some simple simulations - a small area with two types of animal, one eating plants, and one eating the other animal. They survived for a while, but always one of the two would eventually die out.

That's because you need to cheat. Nature cheats too. Some species of frog can actually change sexes in times of reproductive stress.

Simulations like this interest me, being a biologist and all. I'm willing to talk to you more about it off-list if you want.

[Drakulya]

Well, if you also implemented say... the concept that you can't hurt someone if your weak, that would be very bad itself, but you reverse the numbers, and now a population can't be killed off by each other WITHOUT Cantrian help. So, bears could grow and grow... but kill off the deer to the point where they can't be killed any more, then the bears start starving out and their numbers drop and the deer rise, then it goes back and forth. Maybe it will even settle on a good balance, making it very realistic in comparison to some of the things I've heard about.

This is just in my mind though, it may work out differently else where.

[N-Aldwitch]

If we had Draggable Animals implemented, we could definately implement this in populated areas- take a couple of each animal and put them in a room and if the Canivores wipe out their food source, that's their problem, they starve to death.

[Jos Elkink]

Drakulya, the problem is that indeed it is very likely that one of the two will die out entirely ... and from that they can't recover ...

Doug, I'd certainly be happy to talk about this ... jelkink on YIM or something like that?

[SekoETC]

It should hardly be possible that beasts eat the whole population of herbivores. The population of both will fluxuate but naturally it shouldn't reach zero for neither one. You see, the smaller the herbivore population, the harder it is for carnivores to find food. So they will gradually start dying, some may eat people or resort to cannibalism. Some will always find a rabbit here or there. But I think the problem lies in the small populations Cantr has. There is more room for random chance in small populations. If there were thousands of animals like in real life, then it wouldn't trip over the extremes. Hmm... I could try to simulate it in Flash. But I do have lots of school stuff to do.

[Jos Elkink]

As I said, I tried a super simple simulation - carnivors and herbivors and herbivors die if they are eaten or can't find food and carnivors die if they can't find food ... after a few waves up and down, one of the two always died out.

[Cdls]

As I said, I tried a super simple simulation - carnivors and herbivors and herbivors die if they are eaten or can't find food and carnivors die if they can't find food ... after a few waves up and down, one of the two always died out.

Wouldn't it work better if you raised the reproduction rate of the herbivores while getting rid of the condition that they need to eat as well? The reasoning for that is as follows:

The carnivores only need to eat to survive.

Herbivores have the double negative of needing to eat and survive carnivore attacks.

So getting rid of the herbivores needing to eat should let them last longer. And with the increase in reproduction, there should be enough to be eaten by the carnivores while still maintaining a balance.

[Drakulya]

I personally believe this could be solved with some fairly complex mathematics... I'll put some thought into it, but I have no idea how to simulate such things. I'll see if I can get some numbers to look right though...

[Sho]

Predator-prey simulations are usually fairly stable - I was in a dynamics class where we coded some things like that, and they usually work out. And the simpler the system the easier it is to debug. Jos, post code?

[SekoETC]

Each location has a roof for how many animals it can feed. Once animals go above that limit, some of them will start going hungry and die. Deaths will bring them below the limit and there will be enough food again. So herbivores shouldn't go through the roof even if there weren't any carnivores to limit them. Since if they do, the population will subsequently take a plunge.

The amount of suitable food available depends on what the animal eats, if people are destroying original plants by farming, mining and building and land moisture (rainfall lately).

You could link it to gathering spots. Every gathering spot being occupied takes away from the animals. And also buildings will push the animals away.

As I suggested once, animals should also eat food left on the ground.

[Jos Elkink]

I totally agree with your ideas, Seko, but it's a rather large change of the whole animal system. About every five words in your post is a lot of programming ... We don't even have the herbivore-carnivore distinction in Cantr yet. We don't have natural growth of plants they could eat. Etc.

And simulating a proper ecological system is really quite complicated.

Sho, if I can find the code, I'll definitely post it. Doubt it, though ...

[SekoETC]

Yeah, having an existing world to base things on does pose some limitations. If I made a game, I would include humidity distribution from the start. That along with temperature zones would define the amount of vegetation per area, and amount of vegetation defines the amount of herbivores it can support, which would define the amount of carnivores. At first the amount of humidity per area would remain the same, but weather scripts could be added later on, which would alter the distribution and affect everything else. It may sound difficult but in fact in the end it might be easier than inventing every location one by one.

But back to Cantr, you don't need to define life conditions for every location separately - you can base it on terrain type. Naturally different animals prefer different terrains, at least I heard the lions don't travel over mountains so you must have it coded in some form. We could borrow a bit from this game called... Evolution, was it? I played a demo once. Or twice... Each of the animal types (and mind you, there's far more of them than in Cantr) has preferred terrains and temperature. If they're in a good environment, they multiply. If they're in a descent environment, the population remains the same (births cover for the deaths). If they're in a bad environment, they start to die off.

[Jos Elkink]

When you design games, the real key is to figure out what real life things you can implement in a really abstract sense and still get a good game effect. Many of the things you are saying, you can totally abstract from, and still have a nice game. Digging slots is a good example of this.

And yes, if I were to rewrite the game from scratch, animals would be implemented very differently ...