Evolution in Gameplay

making them random mutate isnt bad, since any flawed creature would die out before passing the flaw. There should be a option to "split species" that'd spit into different species or consider a new evolutionary step individuals which had something like 98% of differences from the original species ( thats a lot actually since 2% makes us different from chimps)


many species stay hundreds of millions of years without major changes till there's something like a boom. That boom is change. heating is nothing but a increased molecule movement; at -273.15 C theres basically none.

in this case the trigger for movement is change (and not energy like in heat) many ways for that to happen are

slow (or fast like in the actuallity) change in temperature and composition of atmosphere (volcanoes,meteors, life itself)
isolation (continental drift)
changes in the biosphere (new species, extinction and evolution of old)
latitude , longitude and altitude (which are environmental factors)
chance

any of those can be done in separate and influence evolution in it's own way. It's the merge of all those factors that cause it.

eumesmo wrote:

making them random mutate isnt bad, since any flawed creature would die out before passing the flaw. There should be a option to "split species" that'd spit into different species or consider a new evolutionary step individuals which had something like 98% of differences from the original species ( thats a lot actually since 2% makes us different from chimps)


many species stay hundreds of millions of years without major changes till there's something like a boom. That boom is change. heating is nothing but a increased molecule movement; at -273.15 C theres basically none.

in this case the trigger for movement is change (and not energy like in heat) many ways for that to happen are

slow (or fast like in the actuallity) change in temperature and composition of atmosphere (volcanoes,meteors, life itself)
isolation (continental drift)
changes in the biosphere (new species, extinction and evolution of old)
latitude , longitude and altitude (which are environmental factors)
chance

any of those can be done in separate and influence evolution in it's own way. It's the merge of all those factors that cause it.

What i'm saying here is that flawed creatures WON'T die out because there are not enough factors in the game environment to kill it.
Huge disasters like you are suggesting won't solve the problem, because they will kill animals that are fit, too.
And yes, evolution happens in spurts, but it does NOT happen because of natural disasters. Punctuated equilibrium happens because most mutations don't really do anything, or are detrimental, but one in a million will be helpful and the creature with that mutation will pass it on to more offspring because it is more fit.

We simply can't simulate this perfectly.

the species need to hide from predators and eat (and maybe hunt). If a species has a bigger leg compared to the other that keeps him from escaping as fast becoming a meal or makes him harder to get resources. And even if it reproduces theres only 25% change of an offspring getting the bb gene set and not a BB or Bb set. though a Bb is a host....

65 million years ago the meteor caused the conditions to change making dinosaurs unfit for survival. that caused a boom in mammal evolution

and everyone knows we can simulate it in perfection , but it's still better then the Lamarkian way ( or the spore way which is to throw Darwin's work to the garbage)

Hello. I'm a graduate student in applied mathematics. Ever since I learned of evolution, I have been fascinated by its complexity. It is my life's passion to explore and understand how evolution works from a physical perspective. I wish to complete this task by studying and applying mathematics to explain how systems evolve. So when I stumbled across this project, I could not help but add my two cents.

You do not need to account for every variable to replicate pure evolution. most variables only affect the outcome very slightly. so many of the variables could be grouped into a generic noise term. Brownian Motion is a good candidate for a noise term.

The act of evolution can be simplified by reducing a system down to a probability distribution which represents a diverse population, and the environment for which population exists in. The environment does not have to be a geographically location, but simply a given resource needed for survival. evolution can then be measured by how much the population influences, or is influenced by its environment. i.e. a predator-prey model. You can even use this way of modeling evolution, to evaluate non biological evolution. For instance, if government was a species, what is its resource? Taxation is the first thing that comes to mind. Using resources as a measure for viable evolutionary changes is an ideal way of incorporating natural evolution into gameplay. If the species you create is too good or too poor at gathering resource, it will be reflected by how stable its population growth is. furthermore, since the population is represented by a probability distribution, there will be a given amount of failure and success within the population. It is the successes which drive the process of evolution. with this method, you can insure that the creature you released into the environment at one stage of evolution, is not the creature you will end that stage of evolution with.

The editor for such a game should not be used only to create something cute or vicious to then dominate your resources with. rather, it should be a save point by which you can prepare for the next evolutionary challenge. editing your creation should be limited by how well your creation performed in the last stage of evolution. some kind of experience point system could be viable here. the point being that you must take care in editing your creation, before releasing it upon the world.

While I did understand most of it, I stopped at "some kind of experience point system".

And when I say most, I mean I didn't understand this part in particular.

Quote :

You do not need to account for every variable to replicate pure evolution. most variables only affect the outcome very slightly. so many of the variables could be grouped into a generic noise term. Brownian Motion is a good candidate for a noise term.

The act of evolution can be simplified by reducing a system down to a probability distribution which represents a diverse population

I'd like it if you could explain better. I'll admit; math isn't my favorite subject when it comes to learning...

well, most influences on evolution come from completely random events which alter the environment a species inhabits. so rather than focus on every facet of every kind of event, you can approximate it. this approximation could be handled by constructing a noise term which embodies a type of random process, like weather. that noise term would then simulate every random change in the weather over time and how it affects the evolution of species. a species could then have certain attributes which could be beneficial or detrimental given whatever weather conditions the noise term produces. another example of a noise term would climate.

the best kind of noise utilizes brownian motion. brownian motion is a stochastic process. a stochastic process is a set of events which occur one after another. each event could have multiple outcomes depending on some probability. for example, rolling a die over and over is a stochastic process.

its how the noise interacts with the species which could model minute changes in a species evolution. this can happen because the species you create in an editor may only represent the average of the population, where in reality there is diversity within a species. its the diversity which allows a species to survive random processes. this is why a population must be represented by a probability distribution. the best distribution to use is a normal distribution(bell curve). while it is an embellishment on reality, its a close enough approximation for the distribution of characteristics throughout a single species.

let me know if you need further clarification. i would be happy to help!

El_Noumo, that is my idea, even if we cannot simulate everything, getting the major variables, adding change over time and some chance of success.....

anaerobic cells were outgrown by aerobic ones because the last ones evolved changing the atmosphere with the add of O3 and O2, and reducing the amount of CO2, while producing more ATP from glucose

since the anaerobic cells did not adapt to face this....

this is a really interesting way to look at the problem, El_Noumo, i didn't really think at all of using probabilities and distributions to simulate evolution, although i knew we would have to have a relatively high degree of randomness. My worry here is that in the game we will be often looking at individual organisms and their specific mutations. My question to you is: can we combine your generalized evolution with specific chains of evolution that we will see in the game. Will the sort of noise patterns affect individuals, or only the general evolution of the population. Will it account for species splitting off due to environmental factors? Will it do this accurately enough to be included in the game, yet simply enough to not taxe a processor too much?

this method could diversify a species rather easily. furthermore, it does not rely on what we know in biology, but rather reinforces what we know in biology. so you could easily use the same method to evolve not only the biology but the technology, trade and other social abstracts.

lets take for example a certain species A with a normal distribution representing each of the that species characteristics. certain aspects of its environment will be influenced by the species, and other aspects will influence the species. lets focus on the environmental influences on species-A.

one characteristic which could force evolution on our species is climate. since the species responsiveness to climate is dictated by the normal distribution, you will have essentially three different outcomes. Those members who fail, those who succeed and the average of the population. The average represent the identity of the population with respect to climate. let say species-A lives in a non-seasonal climate to make things easier. this means that species-A is subject to essentially the same climate all year round. those who succeed are capable of living in a wider range of climates. the average are capable of living in a smaller range of climates, and those who fail, are only capable of living in the strictest climate conditions. if the climate changes outside of normal living conditions according to some random process, the species will have to adapt to survive.

now, adaptation is built into this scenario simply by have a specific species characteristic with a given probability distribution that is subject to some random event. in our scenario, species-A is subject to some change in the climate. regardless if the climate goes up or down, there will be pending consequences on some portion of the population. the ones with the strictest living conditions will be hit the hardest. depending on how long the change in climate lasts, this portion of species-A will be reduced. the average of the population may not encounter as much of a problem. two things could happen here. either they can outlast the change in climate, or the change in the climate still falls within their livable range. At this point we still have the same species, with slightly augmented characteristics. while some small portion of the population has died off, the species has survived overall. under more extreme situations, like an ice age or something, the species could be forever altered. this would happen because only a small portion of the population of species-A can withstand harsher climates with respect to its normal climate. that small portion will then be the future of the species.


now lets take a look at a different scenario. lets assume our species-A has been very successful in its home environment. So successful, that the species needs more resources to sustain its population growth. this means that the species will have to seek out new environments in which to extract resources. But new environments may have different average climates than species-A's home environment. as the species spreads into these new environments, not all members will be successful. for the small portion which may be successful, each generation will reinforce that portion of the population with more successful members. eventually, the ability to withstand the home climate will become more of a recessive trait. this will create a new average amongst those successful members of species-A. it is at this time, we could call this successful portion of species-A, now species-B. This is an example of diversification by adaptability to new environments.


furthermore as portions of the population die off, the distribution is altered. to accommodate for this, each successive generation that follows could reintroduce the failed portion of the population as the the species rebounds.

while this all might sound complex, it is pretty simple. we used just a few variables here. population size, population climate characteristic distribution, and some markers for failure and success.

El_Noumo wrote:

this method could diversify a species rather easily. furthermore, it does not rely on what we know in biology, but rather reinforces what we know in biology. so you could easily use the same method to evolve not only the biology but the technology, trade and other social abstracts.

lets take for example a certain species A with a normal distribution representing each of the that species characteristics. certain aspects of its environment will be influenced by the species, and other aspects will influence the species. lets focus on the environmental influences on species-A.

one characteristic which could force evolution on our species is climate. since the species responsiveness to climate is dictated by the normal distribution, you will have essentially three different outcomes. Those members who fail, those who succeed and the average of the population. The average represent the identity of the population with respect to climate. let say species-A lives in a non-seasonal climate to make things easier. this means that species-A is subject to essentially the same climate all year round. those who succeed are capable of living in a wider range of climates. the average are capable of living in a smaller range of climates, and those who fail, are only capable of living in the strictest climate conditions. if the climate changes outside of normal living conditions according to some random process, the species will have to adapt to survive.

now, adaptation is built into this scenario simply by have a specific species characteristic with a given probability distribution that is subject to some random event. in our scenario, species-A is subject to some change in the climate. regardless if the climate goes up or down, there will be pending consequences on some portion of the population. the ones with the strictest living conditions will be hit the hardest. depending on how long the change in climate lasts, this portion of species-A will be reduced. the average of the population may not encounter as much of a problem. two things could happen here. either they can outlast the change in climate, or the change in the climate still falls within their livable range. At this point we still have the same species, with slightly augmented characteristics. while some small portion of the population has died off, the species has survived overall. under more extreme situations, like an ice age or something, the species could be forever altered. this would happen because only a small portion of the population of species-A can withstand harsher climates with respect to its normal climate. that small portion will then be the future of the species.


now lets take a look at a different scenario. lets assume our species-A has been very successful in its home environment. So successful, that the species needs more resources to sustain its population growth. this means that the species will have to seek out new environments in which to extract resources. But new environments may have different average climates than species-A's home environment. as the species spreads into these new environments, not all members will be successful. for the small portion which may be successful, each generation will reinforce that portion of the population with more successful members. eventually, the ability to withstand the home climate will become more of a recessive trait. this will create a new average amongst those successful members of species-A. it is at this time, we could call this successful portion of species-A, now species-B. This is an example of diversification by adaptability to new environments.


furthermore as portions of the population die off, the distribution is altered. to accommodate for this, each successive generation that follows could reintroduce the failed portion of the population as the the species rebounds.

while this all might sound complex, it is pretty simple. we used just a few variables here. population size, population climate characteristic distribution, and some markers for failure and success.

I'd like this in the game rather than the current idea we're getting at. People, let's talk about this please.

YourBreakfast wrote:

El_Noumo wrote:

Snip

I'd like this in the game rather than the current idea we're getting at. People, let's talk about this please.

This is great for population characteristics and general evolution, and in fact, this is an almost perfect way to run Auto-Evo in parts of the planet the player does not see. However, it doesn't answer the question of evolution in individuals. We're going to be dealing with individual animals in the game, with individual characteristics. We need to discuss the best way to simulate evolutio on a scale which is easily visible to the player as an organism.

So to the idea above, yes, and no.

eumesmo wrote:

Most of this can be selected with a random math expression only the last one is harder but it's only needed till the space stage and only in your homeworld.

Sorry about digging up an old post but... Why will this only be on your homeworld? Aren't we simulating evolution the same way on other planets?

if it can be done without lagging the whole thing, then yes i think so.... i was just assuming that it'd be processor heavy to make all that in tens of planets at once.....

The Uteen wrote:

Aren't we simulating evolution the same way on other planets?

After all, why would it be needed? Organisms on different planets will be generated only after they will be seen, and that's only possible at space stage when evolution is so slow it's practically invisible.

Commander Keen wrote:

The Uteen wrote:

Aren't we simulating evolution the same way on other planets?

After all, why would it be needed? Organisms on different planets will be generated only after they will be seen, and that's only possible at space stage when evolution is so slow it's practically invisible.

Good point. We need to look into procedural generation of planets from a given base. I'll set up a thread for it.

What about the procedural evolution of organisms? It will have to no doubt happen very quickly, as the player enters a planet's atmosphere. Maybe this is where El_Noumo's population vs evolution algorithms could come in handy?

US_of_Alaska wrote:

What about the procedural evolution of organisms? It will have to no doubt happen very quickly, as the player enters a planet's atmosphere. Maybe this is where El_Noumo's population vs evolution algorithms could come in handy?

Actually, when you enter a planet's atmosphere, you're going to be seeing everything for the first time, and it's going to be there for the first time. Organisms can be populated there procedurally, but some quick auto-evo could help differentiate the species there.

~sciocont wrote:

US_of_Alaska wrote:

What about the procedural evolution of organisms? It will have to no doubt happen very quickly, as the player enters a planet's atmosphere. Maybe this is where El_Noumo's population vs evolution algorithms could come in handy?

Actually, when you enter a planet's atmosphere, you're going to be seeing everything for the first time, and it's going to be there for the first time. Organisms can be populated there procedurally, but some quick auto-evo could help differentiate the species there.

These organisms I hope will be very, very, simple, and not specialised? It's okay making specialised creatures for the game but super auto-evoing them would not get them to drive to change.

Simulating continental drift, natural disasters, random UFO activity, volcanoes, earthquakes, plagues, meteor impacts, as well as just the usual auto-evo in super form will make the most realistic outcome. That shouldn't be too hard to do in under a second. *cough*With a supercomputer from 2100*cough*

Off topic: Yes, I've crawled back onto the top ten posters... I'm tenth!

The Uteen wrote:

~sciocont wrote:

US_of_Alaska wrote:

What about the procedural evolution of organisms? It will have to no doubt happen very quickly, as the player enters a planet's atmosphere. Maybe this is where El_Noumo's population vs evolution algorithms could come in handy?

Actually, when you enter a planet's atmosphere, you're going to be seeing everything for the first time, and it's going to be there for the first time. Organisms can be populated there procedurally, but some quick auto-evo could help differentiate the species there.

These organisms I hope will be very, very, simple, and not specialised? It's okay making specialised creatures for the game but super auto-evoing them would not get them to drive to change.

Simulating continental drift, natural disasters, random UFO activity, volcanoes, earthquakes, plagues, meteor impacts, as well as just the usual auto-evo in super form will make the most realistic outcome. That shouldn't be too hard to do in under a second. *cough*With a supercomputer from 2100*cough*

Off topic: Yes, I've crawled back onto the top ten posters... I'm tenth!

I'm sayin that when you go to a new planet, the game is going to select content from a content library and populate it there (the orgs here will be selected from the content library, then put on the planet in environments that suit their attributes. We can add in code that keeps a given org/tech from being put in two places in your game, so there won't be zebras on every planet you see.

OT reply: 9 behind me!

~sciocont wrote:

I'm sayin that when you go to a new planet, the game is going to select content from a content library and populate it there (the orgs here will be selected from the content library, then put on the planet in environments that suit their attributes. We can add in code that keeps a given org/tech from being put in two places in your game, so there won't be zebras on every planet you see.

OT reply: 9 behind me!

Dislike. I don't want content libraries to be how my planet is populated. I want all the organisms to be evolved to suit the planet, not have flimsy spaghetti monsters running around on a 20 Gs planet. I vote no to spore-population.

US_of_Alaska wrote:

~sciocont wrote:

I'm sayin that when you go to a new planet, the game is going to select content from a content library and populate it there (the orgs here will be selected from the content library, then put on the planet in environments that suit their attributes. We can add in code that keeps a given org/tech from being put in two places in your game, so there won't be zebras on every planet you see.

OT reply: 9 behind me!

Dislike. I don't want content libraries to be how my planet is populated. I want all the organisms to be evolved to suit the planet, not have flimsy spaghetti monsters running around on a 20 Gs planet. I vote no to spore-population.

We're talking here about random planets out in space that can hold life that you might explore. We can't run all of those at the same timme, so the easiest way is to populate them procedurally with organisms with many similar tags.

~sciocont wrote:

We're talking here about random planets out in space that can hold life that you might explore. We can't run all of those at the same timme, so the easiest way is to populate them procedurally with organisms with many similar tags.

So as long as they match the conditions? That's getting better, but i'd still like them to run through a SuperAutoEvo algorithm just to give them some uniqueness and really lock them into that planet, you know?

US_of_Alaska wrote:

~sciocont wrote:

We're talking here about random planets out in space that can hold life that you might explore. We can't run all of those at the same timme, so the easiest way is to populate them procedurally with organisms with many similar tags.

So as long as they match the conditions? That's getting better, but i'd still like them to run through a SuperAutoEvo algorithm just to give them some uniqueness and really lock them into that planet, you know?

It's a posibbility and we can definitely try it when we get to that stage.

~sciocont wrote:

US_of_Alaska wrote:

~sciocont wrote:

We're talking here about random planets out in space that can hold life that you might explore. We can't run all of those at the same timme, so the easiest way is to populate them procedurally with organisms with many similar tags.

So as long as they match the conditions? That's getting better, but i'd still like them to run through a SuperAutoEvo algorithm just to give them some uniqueness and really lock them into that planet, you know?

It's a posibbility and we can definitely try it when we get to that stage.

That's all i ask.

Are you saying they will be complicated creatures like zebras and chosen to be already adapted to he environment?

The Uteen wrote:

These organisms I hope will be very, very, simple, and not specialised? It's okay making specialised creatures for the game but super auto-evoing them would not get them to drive to change.

They need a need to change, people.