Population Model v. 0.1

[Leland]

That's a matter of definitions; in this context I've been using the word "region" to mean a player defined area, and that's the word used in the design doc. Perhaps we need new definitions?

How about calling a player defined areas territories or provinces, and an area with distinct geographical properties a region? (then again, we've used the word region so long that maybe we just stick with it. Definitions of term should be explicitly stated somewhere in any case, IMHO)

It's a valid point that geographical conditions can be used to division populations; however, there are also plenty of other boundaries between people. Urbanisation, religious and ethnic differences, social class, closed national borders and whatnot. It's a good point that geography should be used as one such factor, but it's similarly important to remember that there are other considerations and to come up with a model which treats them equally is a little bit tricky (or maybe I just want to make it tricky ).

[amjayee]

About the definition of the region, it has been given - "somewhere". Onve again, there are so many threads that only the original members have somewhat good idea of what has been discussed. But yes, I was the one who invented the word region for us, and I've become aware that it's not the clearest word to use here. The idea is, that region is a supergroup for many kinds of political entities that make up a nation (or civ); states, provinces, protectorates, colonies etc. are all regions. What is the subgroup is decided by the government system; all those subgroups are different from each other.

So, region is a political thing rather than geographical.

About population; the reason why we decided to use regional population was the fact that we simply have to store quite much information for each population, so the amount of data would become too large if population units are used. Besides, most of the population units would be much like each other. So, we became to think of a way to compress the data; each region has one, and only one population. But, the population will not be at all uniform. We will store the amounts of people belonging to different nationalities, different religions etc. and all these can have their own effect in the game. So, in this system we would have much the same information as when using population units, but in a "statistically compressed" form. After all, the group of people will work much like one large individual with the average properties of the individuals making up the group. All this has been discussed long ago, but it's quite impossible to digest all the discussion we have had in such a short time. Just keep reading and asking.

[amjayee]

You have a good point, but the flaw in it is, that people would not emigrate from region to region, but across land, from tile to tile. And also, the region the people will move into, must be quite lucrative so they will abandon their earlier life. So, there two effects combined would cause, that since mountains are not very good place to live, people are reluctant to move there. Also, the mountains slow down the emigration just as earlier.

All in all, I don't find any problem for using the region and population combination. After all, people are the target of politiocs, so it would be quite silly to have a political entity without people to exercise politics, and to be the target of politics. I don't quite agree that population is not a political entity, quite the contrary. We don't understand population as sciences usually - it is just a group of people with some geographical area they live in, and some political properties and affiliations the players try to rule.

[Leland]

Sorry, the word "political" was misleading... I meant that regions are *administrative* areas whereas populatons are not. I stand corrected about the migration plan, but I still don't think that populatiosn and regions should be mixed. It just sounds wrong! Let me try yet another example... (the idea is that if you can shoot down all my example situations, the model can be trusted to work )

Suppose there is a industrial small region adjacent to a larger rural one. In the small region there are about million people, most of which are factory workers; in the larger the population is around 3 million. To be exact, let's say that 60% of small region inhabitants are factory workers and something like 80% of the rural people are peasants (or field workers, doesn't make any difference here), rest of the classes being something else. The living conditions of the industrial area are getting bad: there is a danger that the workers will rebel. Since they are a mojority of people there it's likely that they are a powerful group. Now, the intelligent player sees the opporunity of riots and conveniently combines the two regions. In the resulting big region (actually, not much bigger than the previous rural region because of the relatively small land area of the industrial zone) there would be 15% factory workers and 60% peasants. Now the workers are a clear minority and have no political power to riot. Problem solved?

Obviously something like that would not happen in the real world because workers and peasants live in different areas defined by the available resources (factories and land). How would the region-tied population model solve this? Hmm...

I'm still opting for my solution: don't let the player arbitrarily define boundaries for populations. The government can impose laws forbidding migration or use other means to indirectly control migration, but in the end it should be people who decide to move or not. I'm not sure how much memory/processing power regions will require (in purely geographic level), but giving the populations a chance to have their own "regions" (meaning, the area where the population lives) would only double this and have no cost on the populational memory requirements. Regions would then have truly statistical information: a weighted average of all populations (also partial) on their area.

Conclusion: as you said, population "is just a group of people with some geographical area they live in". Why should that area be *arbitrarily* defined by the player by just drawing a line on the map?!? Makes no sense to me.

(I'm having this gnawing feeling that I may be so utterly wrong that once I realize it the paradigm shift will tear me in half... until then, prove me wrong. Anybody. Or prove me right, that's also welcome. )

Leland
[This message has been edited by Leland (edited January 04, 2001).]

[Leland]

I'd just like to apologize that I sound like an old record... I can be vey stubborn at times. It just occurred to me that even if the final decision will be to use regions as the section where populations live, there should be a back door left in implementation level: use an abstract notion of "area" and have both region and population refer to it instead of the population referring directly to the region. So if the populations are from the beginning built to be associated with geographical areas, they can still be implemented according to your plan and be connected to regions, yes?

Not that I'm giving up a fight or anything...

[Leland]

Thanks for the info, I am slowly beginning to buy the idea.

However, as you said regions are a political thing whereas populations are not. That does nto sound good. I can think of all kinds of little "cheats" you can do when populations are tied to regions. What if you've got two regions and a mountain range between them to block migration, and you then set up a third region that covers the mountains? Let's denote the areas A, B and C and the mountains an m:

Before:
AAAAmCCCC
After:
AAABmBCCC

Now, let's assume that there is a pressure to migrate from A to C, but the mountains slow it down or stop it entirely. After setting up region B in between the people are free to migrate from A to B, and because B only has information abotu the statistics there is no way to stop people moving from B to C. Sure, there might be other things about the mountains that make region B impractical, but the player migth be willing to make those tradeoffs to momentarily boost migration.

I seriously don't think populations and regions should be mixed. Perhaps a better option would be to use some sort of virtual regions, that is regions which the player has no power over, and couple the "statistically compressed" populations with those. This would double the memory needed for the region info, but the populations would not be much different from your proposal.

Well?

[amjayee]

I don't know how to put this, but I simply can't see the point in your idea. In real history, there has been boundaries and borders for regions, that have been independent states or chiefdoms, or parts of larger empires. Similarly in the game, we would have restricted areas on the map, whose borders are drawn by player or decided by computer. That area has a population, that lives inside that region, thus being the population of that region. I can't possibly see, where else the population could live, without being a population of another region. It's all in the definition of the population - it is a group of people, that live in a certain region! I can't see any other sensible way of handling this system. Having population and region separated would result in a system where population doesn't mean anything sensible. Region is about administration, but that requires people to administrate, and also administrative people.

Also the situation you described would be no problem; first, only very rarely we would have entirely industrial region, and besides, even inside regions, rural and urban areas are considered different, for various reasons, largest being that cities have been hugely different from countryside through all history. So, in your scenario, industrial people would be concentrated in cities, and we could have a system, where certain cities could have riots because of those industry workers.

All in all, I think a better way to solve the problems in a good and realistic system is to simply make a rule or system that actually solves those problems, instead of creating an entirely new system, which doesn't have that particular problem, but causes new ones and is perhaps also less realistic and good. And the latter seems to be the case in this particular idea.

[This message has been edited by amjayee (edited January 05, 2001).]

[VetLegion]

amjayee I suggest you make Population Model v 0.2.

I would like to get more involved in the discussion, but many things have been discussed and I am not sure what "state" is the Model in right now, with modifications and all.

[cavebear]

"Province" seems the right term for a player-defined area. It suggests an artificial construct, whereas "region" seems geographically independent of player intentions.

[Leland]

Agreed, but everyone has consistently used the word "region" to mean a user defined area and I think it would be too much of a hassle to change the definition. As an alternative solution to disperse confusion I suggest that someone gathers a definition of the game terminology and puts it on the website. As a newcomer I've bumped into this sort of confusion, so I could probably do it. But not yet, I have a few things to deal with.

Anyway, I think that the player should not see all the regions as the same. Depending on their status and the government model, they could be called "provinces", "colonies", "states", "territories", "zones", "districts", "chiefdoms" or whatever. Some of these would have distinct properties, some could be added only to spice up the game.

Leland

[Leland]

Oookay... I think I have to bend in this matter since I'm being the only one against the idea of regional definition of populations. However, I think that your original model had some minor flaws in addition to that (the definition of cities, the relationship between cities and regions, is there a capital city, refined migration model... and all sorts of things that would be better commented if you made another draft. Besides, I'm afraid I scared all other posters away with my endless gripe about regions, so a new thread is in place anyway) so you need another version of the population model anyway.

But let me make one last commentary on this topic (okay, I'm not promising its the last one, but I'm starting to accept the idea, at least until I see it implemented):

quote: Originally posted by amjayee on 01-05-2001 10:07 AM In real history, there has been boundaries and borders for regions, that have been independent states or chiefdoms, or parts of larger empires. Similarly in the game, we would have restricted areas on the map, whose borders are drawn by player or decided by computer. That area has a population, that lives inside that region, thus being the population of that region.

After WW2 Germany was split into east and west. This is an instance of forced and arbitrary setting of regions; however, I don't think that this action made the two populations as two distinct populations. In your model, the populations would become two populations but with exactly the same statistical properties, right? Okay, after fifty or so turns this regional division is broken. Now the populations have become differentiated by fifty years of different government styles and laws; but what is is exactly what happens to the population when they are again united? Does the new unified population just have the weighted average of the two previous populations? In real world, there are still great differences between east and west germany in populational level, in social structure, in economy, infrastructure (okay, not part of population model I guess), education, ... but the way I see it, the game cannot handle these differences because the only information stored is either statistical or some sort of ad-hoc rule.

Of course, East and West Germany both consisted of several regions of their own, but it need not be so in a similar game situation with smaller areas.

quote: I can't possibly see, where else the population could live, without being a population of another region. It's all in the definition of the population - it is a group of people, that live in a certain region! I can't see any other sensible way of handling this system. Having population and region separated would result in a system where population doesn't mean anything sensible. Region is about administration, but that requires people to administrate, and also administrative people.

I've said it many times: drawing lines on a map does not change populations in real world. In Scandinavia there is a big population of "lapplanders" (or whatevery they are called, can't find the english word rigth now) in northern parts of Finland, Sweden and Norway, as well as on Russian ground. You can't make that population disappear by redifining the territorial borders within any of those countries.

I am not advocating that there would be populations that are not withing any region, that would of course be impossible. It's conceivable to have uninhabited land that is not part of a region, but once a population moves in it automatically becomes part of a civ, and if civs are divided into regions then the population by necessity exists within regions.

quote: Also the situation you described would be no problem; first, only very rarely we would have entirely industrial region, and besides, even inside regions, rural and urban areas are considered different, for various reasons, largest being that cities have been hugely different from countryside through all history. So, in your scenario, industrial people would be concentrated in cities, and we could have a system, where certain cities could have riots because of those industry workers.

Oops, forgot about the distinction between cities and regions... that cities have populations too. Well, the difference need not be between urban and rural populations: there could be a dispute between farmers and hunters for instance. Think of the North American Indians. Or religions and nationalities. To make special rules for all these things would in my opinion result to a very incoherent system.

quote: All in all, I think a better way to solve the problems in a good and realistic system is to simply make a rule or system that actually solves those problems, instead of creating an entirely new system, which doesn't have that particular problem, but causes new ones and is perhaps also less realistic and good. And the latter seems to be the case in this particular idea.

I have a tendency to rant about minor details and not see the big picture. To counter that, I'll try to revise the differences between the original pop model and the changes I proposed, that way I will clarify it to myself as well as for the rest of you, and open myself up for comments and criticism.

First, each population would have a fundamental property: the area they live in. This would be stored in the same way as the area of a region. Within this area, each tile has a number telling the amount of people living in there, just as in the current model. One problem I see is that this way there would be a possibility of areas within a population that have exactly 0 people; this is certainly counter-intuitive, but it should be possible because a) defining areas with one boundary is easier to implement than more general areas and b) this could be used to allow the distinction between rural and urban areas. Cities would be populations within larger rural populations, and the amount of rural population would be zero in the city tiles.

Now, each population would have rest of the properties as in the old model. Goupings, age structure and all that. Regions, however, would not contain this information directly: instead the statistics for a region would be obtained by calculating the weighted average of the populations residing within it's borders. Thus, regions would truly have only statistical information and changing the regional limits would make absolute no difference to the populational structure.

If a population happens to be located in two regions, and the borderline of these have certain limiting properties (for instance, the government decides to close the borders of it's regions to stop unfovourable migration, or that the taxes in two regions differ), this would of course have to be taken account in the population's properties somehow. There would also have to be a way to simulate the separation of a population to two different populations and a the assimilation of populations. However, I think that the region simulation has to cope with similar algorithm's as well when calculating the economy, production and the assimilation of conquered areas. Still, this is probably the biggest fault in my proposal... I'd like to see some scenarios or examples on specific problems associated with this line of thought.

Migration would happen tile by tile as in regional models, with the exception of population borders: the border's might either change (when a tribe moves to an uninhabited area, for instance) or there will be diffusion between two adjacent populations. I don't see any problems bigger than those associated with regions in this respect.

That was maybe not very clear... question's are welcome. Anyway, I thougth about one of my previous posts and the example of two regions separated by mountains. I don't think tile-by-tile migration solves it really... allow me to re-explain.

At the beginning there are two populations, A ja B, with different properties (say, religion X is predominant in A and religion Y is predominant is B) and there would be a natural diffusion between the two populations if there weren't mountains between them:

AAAmBBB (m meaning the mountains)

Now, the player sets up region C between A and B so that the mountains fall into that region:

AACmCBB

Now the properties of Region C would be an average of A and B; hence the religion Y would have greater following in it than in A, similarly religion X would have greater following in it than in B. If the game attempts to be realistic, there should be diffusion between A and B as well as B and C. This does not mean that the population (that is, the number of people) in C would change: the followers of X would simply move to C and the followers of Y to A until there is an equilibirium. Same thing should happen to B. The only migration woudl happen around the borders; there is no need to cross the mountains!

Perhaps the current model can solve this somehow... maybe there is no diffusion of this kind. But certainly there would be diffusion of other kinds: in technology level (though I'm not quite sure how you are going to arrange that.. not my specialty), social structure (say, A need farmers and B needs craftsmen) or economy (product R is cheaper in A, but P is cheaper in B).

There would actually be quite a simple rule to work around these problems: forbid the player from defining regions. But this is out of the question. I cannot foresee a simple rule to make the region based population to work without compromising with diffusion and realistic social interaction. Of course, now that I've been writing this rant for some time, my point of view is terribly biased and unrealistic itself; that's why I urge all readers to give comments. Am I nuts or what?

Amjayee, I'd like to see some examples of specific problems with the system I proposed. I am too involved in this to see them myself.

Leland
[This message has been edited by Leland (edited January 05, 2001).]

[amjayee]

First an amusing out-of-topic note: you used once "ja" instead of "and".

Ok, now I start to see what you are meaning with your idea. But there are some points I have to say about it. First and most critical is, that in your system it would be really hard to find out the populational properties of the people living in particular region. Region population is a totally essential thing so that economy and government models could work properly. We have always to remember that all models need to work also together. Some compromises have to be made to allow that.

I admit now that I understand your ideas better, I see how that could make certain things more realistic and better, but as I said, its largest problem is the fact that it's mostly "populational" and ignores the other requirements of our population model.

So, in your model, would each population consist of only one nationality, or what? And that group would have an area of land? I can see how that would allow a more realistic system of immigration and nationalistic clashes. But also I can't help seeing how difficult it would be to handle; there could propable be dozens of populations in some hexes. And is it really so important to know in what hexes excactly each nationality lives?

We could of course have two populations - region population and then those kinds of populations you suggest. But wouldn't that be over-complicated? Would it really be so helpful? I mean, in any case we need the region population for economy and government. Combining that from the other populations, that could be very many, and each of which could span several regions at once. How on earth we could find out the amount of each nationality in each region unless we save the amount of people in each hex of their area for every population? That would mean lots of memory. I'm not so convinced the system would be so good this would be acceptable.

I really wouldn't like to make a bad impression here. Your system has its potentials, and I don't have time to argue my points very well right now. It has brought us to see some problems that might arise from our model, and that's good. We just have to see now, that we don't have the resources to get a perfect model which could model everything. What we noe need to do is to sit down and think how to settle this issue.

I'm willing to admit that our model is not perfect, and that your model is good, though might be hard to implement and execute as you proposed. Also I will stick to the fact, that region population is absolutely necessary for economy model, government model, and for the situations when regions change possession; until recently, people in different countries have not had the possibility to travel freely between countries, if those two countries are hostile. So, in most cases the populations would be split in two or more parts. This is handled easily and adequately well with regional populations, in my opinion.

What we now need to do is to find a compromise. Arguing is rarely a good option in these situations, when we know that neither one of this models is perfect. We need to decide now, what things the population model needs to model, and what kinds of situations it needs to be able to handle. We cannot achieve everything - we will need to find out what are the most essential properties for the model, and the rest just needs to be dropped out, if they prevent us from achieving those essential properties.

I can make the new version of the model. I'd just like now some discussion about what is needed. I will myself come back to it later. Then, I will design the new model, hopefully that's done soon.

[Leland]

quote: Originally posted by amjayee on 01-08-2001 05:29 AM First an amusing out-of-topic note: you used once "ja" instead of "and".

Even more out-of-topic is that I had had already corrected more than one "ja" before submitting the thing... awwww.

Anyway, my "model" has its faults. I don't mean that each nationality would necessarily have one population or that there would have to be several populations in one tile. Actually, now that I have reconsidered my position, only thing that is needed is to leave out the user definable regions for populations. That's what makes the regional system so unrealistic: if there are two regions with different properties, combinign them into one by just calculating an average destroys information. And vice versa, if there is one region with distinct properties, partitioning it to two regions increases the amount of information stored about that geographical area. This kind of decision should not be up to the player: can you imagine a first person shooter where you could "round up" your health from 36% to 40% at will?

I can't contribute much more to this conversation for a few days, but if you are thinking about the model you might want to try solve it by limiting the player's ability to define the regions. For instance, there could be "primitive regions" which would consist, say, one city and it's surroundings or a geographically distinct area, and the player dedined provinces and districts would be aggregates of these primitive regions.

Leland

[S. Kroeze]

Since in this rather ancient thread many issues associated with the economy, like consumption and health of the population, are discussed I thought it might be useful to bring it to everyone's attention again.