Announcement

**Mark_Everson** · May 18, 2000, 19:37

LGJ:

At least my take on how it should be done is there in the Econ HTML document on the web page. A little more than one-third of the way down. Provinces can't be arbitrarily resized, but there is some scope for action in that area by the player. Like I say in the HTML document, I don't want to give people any reason to micromanage the province size. I picture resizing your provinces as something that might be done three or four times over the entire game.

There is further, more detailed information at the top of the demo 5 Econ thread. Some of it is of course just a first try for demo 5, but I think it gets the general idea of what I'm trying to do with provinces across.

I guess give it a read, and then let me know what you think.

My responses to some of your points...

Provinces, or something like them are Required for the current economic system to be feasible. Basically it's because a lot of things done in the economic model are too computationally intensive to do for every square. If we use provinces that are of order or 20-50 squares large, then the calculational load becomes much less of a problem. So that is kind of the minimum province size I have been shooting for. This also matches in general administration-type provinces in lots of empires throughout history.

I realize that models other than the Econ one are going to use provinces, so clearly there is a reason for there to be some compromise. I think your ideas of allowing the player to automatically organize provinces along cultural or other guidelines are quite reasonable. However, there should be a functional maximum province size that depends on technology IMO. This maximum province size would help add some realism in the economic model, the government model (based on what I have seen from Rodrigo) and probably in other areas such as the characters model.

[This message has been edited by Mark_Everson (edited May 18, 2000).]

**shimmin** · May 19, 2000, 08:37

The way I see it (I may be misunderstanding something) if it makes any difference whether a square is in province A or province B, then the player has reason to micromanage it.

But...

If it makes no difference whether a square is in province A or province B; if provinces have no material effect on the game, why are they there at all?

--

One possibility is to let the computer continually resize the provinces as infrastructure / travel speed improve, optimizing the player's empire without forcing them to check each turn to see if they could do something better.

**Richard Bruns** · May 19, 2000, 08:48

My Thoughts:

I like the province system a lot. (I would like it a lot more if military management was province-based rather than unit-based, but that is not important now.) I think that provinces should, if possible, be a coherent cultural, geographic, and political unit defined by the computer, and that they should be fixed in place so the player cannot change them.

I disagree with LGJ about not having provinces at the beginning. Defining all minor tribes as a province would be a good thing IMO. It would make economic and political interactions with that tribe a lot easier for both the players and programmers to deal with. I would prefer that minor tribes be treated as little civs, and having them form a coherent province structure would let this happen.

Geography should play a major role in generating provinces, tribes, and cultures. A path algorithm could define areas that are similar and easy to get around within. These areas could then be assigned culture groups. This is reasonable to do because unique culture groups are more likely to form in these areas. This combination of similar land and similar people makes the province a natural political and administrative unit.

As technology improves and people become more culturally homogenized, the province barriers should erode somewhat. As the civ grows, the provinces combine so the player doesn't have too many provinces to deal with.

**Lord God Jinnai** · May 19, 2000, 13:53

I like Rich's idea since I see how provinces are being used, except that there will have to be one exception for small provinces. This is the case when an attacking empire conquers part, but not all of another province. Say it was split in half, 15 to 15. As it stands now, 15 would be too small, but the area and people may be so differnt that it wouldn't be right to group them together with another province. My suggestion is that we have a minimum combining level for small provinces that might get attacked or something and loose land. These provinces should not be combined with other ones until the original one is something like 5 or less.

Also I read the provinces part in the economy section and the way the diminshing returns work isn't going to work because the provinces continually get larger over time. I think they should pretty much be set, only major events like civil wars, conquest of one small province by another small province (ie Feudalistic periods), drastic change in culture(s), etc. These shouldn't change as technology increases. There is one exception:

At the beginning of the game we use Rich's method to determine where provinces are set up initially. The AI uses these to manage events and whatnot, but the player cannot micromange provinces until he reaches a certain level of gov/managment tech, at which time he can. This number will be low so that players can start doing it early on, but allow the player to experince the growth of control of the government as well as giving more realism etc. The player wouldn't need too many provinces anyway before this because they should only have a few citites by the time they reach this level, whatever it is.

Final note on provinces: I think we should allow city-state provinces. They are common enough throughout history and in modern times to allow it. These city-state provinces would only become so after reaching a certain level of population (based on say the known world population). It would also be limited to only a few cities (10-20). These city states would include the city and the 4 or 8 squares surrounding it. If the city falls below a certain level it is reabsorbed into the parent province if possible, if not, the most culturally suited province of the parent civ ajacent to it. If for some reason there is none next to it (Such as Hong Kong was), this doesn't occur and it becomes a "special" city-state province, where it is not counted on the city-state number level. If at any time it is below that point and the parent civ comes ajacent to it (assuming its not conquered), it automattically gets absorbed.

The reason for haveing city-state provinces is because once they become so large, they act very differntly then the rest of the province, ecconomically, politically, culturally, etc. This is mearly a way to represent this.

**Mark_Everson** · May 19, 2000, 19:40

Thanks everybody for your input. There are clearly a lot of tough issues involving province formation and size, so kicking these ideas around is useful. As I say below, I think many of these issues can actually wait until playtesting before we agree on a final implementation. But hopefully this discussion can save us some time by preventing us going down a clearly bad path at the start.

Shimmin:

Well, the idea is to have it not be critical At the Margins which square a province is in. So, it shouldn't make a big enough difference for most people to micromanage whether a particular square at the edge is in province A or B. However, there would be places where it would make a difference. I just would like these instances to come up fairly infrequently. I think we are going to have to mess around with several designs before we get this right since there are a lot of trade-offs. If we can get something automatic that would work reasonably well I would be very happy. But I'm not sure we can do a good enough job to satisfy most players under most circumstances.

Richard:

I really don't think provinces can be fixed in place. For one thing, the typical slow incremental conquests that will happen when one is starting with a very small state wouldn't make any sense with fixed provinces. For instance, let's say I am playing the nascent Roman republic that controls something like two squares. If I take two adjacent squares from other Latins, and one adjacent square from a Celtic tribe, to my mind I should still be able to have a single province for my five squares, even though it goes across what used to be the border between provinces. But maybe I don't understand what your point is...

BTW the barbarian cultures that you saw in demo 4 Are each a single province. Handling them that way is what seems sensible to me also. I also agree with your and others' points about using geography to set up things reasonably. It's just that after the set up I see things as being somewhat more chaotic, and thus needing to be more flexible, than I guess you do. I think we can just let the playtesting be the final judge on that...

LGJ:

Yeah, clearly under active conquests situations you need to be able to allow smaller-than-usual provinces. I really think you need to get used to the idea of having multiethnic provinces however. IMO frequently they're going to be completely unavoidable. And frankly, lots of despotic rulers in history didn't care one bit about the peoples they stuck together and moved about to create provinces.

I think province size Must be able to change with technology for the game to be playable. Otherwise there will be modern states with 42 provinces, which is way too many IMO. Do you recall what civilization gets like with 42 cities, not to mention 100! Now I think because of the generally smaller province size for ancient civilizations that sometimes a large ancient civilization could have 30 or so provinces. But one this large would typically be hard to hold together, and would tend to only last so long before being broken up by revolution or other factors. But this is really another point where we can wait until the game matures more before we need make a final decision.

I'm not sure I get it on your city-state provinces... You just seem to be wanting a smaller than usual province that has a very powerful central city. As far as I'm concerned people can do that right now without the need for any special rules. I just wouldn't want a lot of AIs also building these, because that would require more clock cycles for handling the economy. I really really want to avoid having lots of special rules for things unless they're necessary for game balance, or some other very important factor.

**Richard Bruns** · May 19, 2000, 23:04

I don't think it would be good to start with only two squares. That is a tiny amount of land. I thought that civs would start out with at least two provinces. I simply don't see how a 2-square civ could survive, let alone wage a multi-front war.

In ancient times, typical conquests were not slow or incremental. For almost all of history, one battle would determine the fate of an entire province. My view of ancient conquest is that winning a battle should turn the entire enemy province into a satellite or subject state. This is assuming that the entire defending army is a single task force, which it almost always was.

The entire province, being a coherent political unit, goes to your state when you defeat the defending army. Square-by-square type attrition campaigns almost never happened in ancient times.

**Mark_Everson** · May 20, 2000, 10:37

I think the "generic" starting position will probably be something like 10-30 squares of land if you want to start in a time period when sizable empires have already formed. However, some players will want to start earlier than that, or start small but with an edge, for instance, in military technology. To support these types of games I think we need to have a versatile approach to provinces, and province size.

What you say about ancient conquests changing the status of large chunks of land at a time is, I think, generally true when two fairly large states fought. But some provinces will represent collections of many small states that only have a common culture. These are the "barbarian cultures" in demo 4. The province in fact Does Not represent a coherent political unit. In this case winning one battle even decisively, will only represent having defeated one of the many small kingdoms within the province. So the squares representing that kingdom only should have their status changed. That could well be a few squares at a time.

IMO there should be a dynamic pushing toward the larger province sizes of 20-50 squares to improve game speed. But I think we need to recognize that the path to the stable provinces that we desire will sometimes take quite a bit of time.

**shimmin** · May 21, 2000, 19:31

As a player, there's really only one use I would have for provinces if I were allowed to manually resize them, and that's joint administration.

Initially, I would have my core be a province, and as I expanded, I would move more developed regions into a second province, keeping one region of my empire fine-tuned for growth and expansion, and the more developed core regions aimed at becoming an economic powerhouse. Special circumstances (border with aggressive neighbor -- tune this province for military readiness; plentiful local resource -- tune the region to most efficiently exploit it) would create other regions.

If the goal of the province is to speed gameplay, then the player should be able to make region A and region B part of the same province regardless of geography if they wish to administer A and B in the same fashion.

I realize, however, that this really doesn't fit the historical notion of what a province was. If the goal of subdividing the empire into a few "management strategy" blocks is to avoid micromanagement of individual cities, maybe we need a different name for the administrative blocks?

**Richard Bruns** · May 28, 2000, 10:40

I just realised that my misunderstanding caused a lot of unnecessary arguments.

I did not understand how much land was in a square. At first I thought that there would be three or four provinces and about sixty squares in a place the size of Italy. I saw province sizes as being about the size of the territories on the Castle Risk board, the one that only includes Europe and a few surrounding areas.

Then I realised that each square would have a width of one degree of longitude and a height of one degree of latitude. Looking at a map, I saw that this would mean Italy would be about ten squares, or a half to a fifth of one province as it is currently defined.

The land area I was calling a province is equivalent to the land area everyone else was calling "two squares."

I apologise for all of the foul-ups this caused.

**Richard Bruns** · May 28, 2000, 14:06

The recent discussion about province sizes got me thinking about the overall Clash map and the map of our world.

The Arabian desert is about the came size as Mesopotamia. The Sahara desert is about the same size as Europe. Siberia and the wastelands in Asia have the same land area as all of the Asian civilizations combined.

Does a parched wasteland deserve as many clock cycles as the "cradle of civilization"? Does a bigger collection of sand deserve as much processor time as all of western civilization? Do a few barren, ininhabited regions need to be given the same status as all of eastern civilization? I don't think so. Yet any system that treats all land squares equally will inevitably have this result.

If we make the squares and provinces smaller to try to give enough detail to the civilized areas, we will waste a lot of processor time on the wastelands. If we use larger squares and larger provinces to avoud wasting time, we will turn some entire countries into a single square.

I think that it is very important that we consider this problem of detail. I have an idea that could help solve the problem. Whatever you think of my idea, please give some thought to this issue of map detail.

In Mark's new economic system, all calculations are done by province. The inputs from all of the squares are added and then run through the production equations. This is a good time saver, but the province calculation speed will still depend on the number of squares in the province. The squares must be considered individually and then added together.

What would happen if all squares in the province were the same? If they all had the same properties, then all of the province equations could be found simply by calculationg one square and multiplying by the number of squares. Everything from the cost of roads to the returns from farms could be quickly calculated this way. Many long summations would become a few multiplications.

This would have two important results, The first would be that the quicker calculations would allow us to have more provinces without slowing things down. The second, and more important, would be that the calculation time for provinces would be independent of the number of squares in the province.

The calculations for a province with three identical squares would take just as much time as the calculations for a province with seventy identical squares. The computer only has to calculate one square and multiply.

This means that the number of provinces, rather than the number of squares, is the thing that determines how much time the computer takes. A map with thirty provinces and a hundred squares would take just as much time to compute as a map with thirty provinces and a thousand squares.

I think that such a system would be a good solution to the detail problem. We simply make the provinces in good land smaller and the provinces in bad land larger. If the number of provinces is the same, the processing time will be the same. Civilizations are more likely to form in good land, so the inhabited and civilized areas can have small, detailed provinces while the wastelands can have large provinces that are calculated quickly.

This system would also allow the map to become much more detailed. The area that used to be one square can become four, nine, or even sixteen squares with no increase in the time needed to calculate things. A more detailed map will allow more detailed and accurate province sizes. A map of Earth would no longer have to represent Switzerland with two squares. The country could instead become two provinces and have a realistic shape. Meanwhile, the huge expanses of wasteland in the Sahara would be combined into only a few provinces, so no calculation time is wasted on them.

Such a system would require certain sacrifices. Players could not resize provinces. Provinces would have to be taken whole in military combat. Ecological and economic provinces would have to have the same borders. I think it would be worth it, and in the absence of a better plan I think this should be considered.

I have thought about the military consequences of this, and I will put them in the Military Model thread.

**Mark_Everson** · May 28, 2000, 15:02

I think your comment about making provinces really big for things like the Sahara Desert is a good one. It would indeed save us quite a bit of time, cutting the province-related calculations by perhaps as much as 50%.

quote:

In Mark's new economic system, all calculations are done by province. The inputs from all of the squares are added and then run through the production equations. This is a good time saver, but the province calculation speed will still depend on the number of squares in the province. The squares must be considered individually and then added together.

Actually, that's not true. Once the province is formed, the individual squares are only considered occasionally. The only time penalties that are on a per-square basis is when the province's squares are added or removed, or when the number of sites in a square contained in the province changes. That's why I was so anxious to maintain the scalability in the utility function. It allows me to basically ignore the squares for long periods of time.

When a province is formed, the various sites in all the squares are simply added up. After that, additions of labor or capital are just handled at the whole-province level. Since we know how many people are working in each economic sector (like agriculture), we can come up with a good estimate of the population in each square when necessary by distributing the people working in the sector proportionally with the number of sites of that type in each square. And even when the number of sites in the square changes its not particularly expensive, since the square will just inform the province of the change, and the total number of sites in the province will be modified accordingly.

The most expensive part of this is when you need information on individual squares' population and what they contain in terms of capital. That penalty will come in for two cases. For the case of small states were a few squares make a difference, both the player in the AI will need to know about the properties of the individual squares that are nearby. The other cases when the player wants to examine individual squares for whatever reason. My guess at this point is that neither of these exceptions will be particularly expensive computationally since they will generally apply to only a very small fraction of the squares on the map. I also have some ideas for minimizing even the small penalty.

Anyways, that's why I like my current model, despite its defects. It gives pretty much the advantages you are looking for, without the extreme limitations (at least I consider them extreme ) of static provinces or coincident ecological and economic provinces.

**Richard Bruns** · May 28, 2000, 15:28

Mark:

So there would be no problem with multiplying the number of squares by sixteen? If they is the case, I think we should decrease the size of the squares. The dimensions of each square could be one fourth of a degree of longitude or latitude at the equator. This detail would make the map look and play a lot better.

I don't think that static province sizes are such a limitation. I think that resizable provinces would cause too many problems, but we won't be able to come to a conclusion about that until the game has been playtested.

**Lord God Jinnai** · May 28, 2000, 22:26

Rich:

I like your proposal, but not the fact of conquering whole provinces at a time. I don't mind those for purely war games like RISK and Genghas Khan 2, but Clash isn't purely a war game and I also really really want to be able atleast sometimes to only take a few square only out a large province w/o making it a warzone province.

I do have a proposal for this which hopefully will only increase the COU usage a minimal amount. For this we have 2 (or optional 3) layers of provinces. The first layer is for the type of provinces you talked about above. The second is based on that, but say it is devided by 2 nations, well here's how it works. Anything that doesn't have to do with whoever controls a section or not will be calculated on the first level for the entire province. Such things would be for disease, famine, climate changes, etc. The second level would take that same area and calculate each section seperatly for each level 2 province for both contries so say the province had 10 squares and Civ A controlled 5 and Civ B controlled 5. Lets say a drought hits the province. No matter which side your on, reguardless of how you deal with it, the intial effects of the drought will be calculated for the entire province. Now after that things like what your civ does in response are calculated after that. Also how it effects the economy might in some cases be done province wide and in some cases not. FE a drought will affect storages for both sides. But if storgaes run out in Civ A, prices there will go up, but not ness. in civ B.

Anyway the first could be the ecological provinces you talked about and the second could be sub-ecological/governmental provines.

The 3rd i said *maybe* to would be done by the player solely to ease micromanagment, IE have all things in area A, reguardless of ecological, political, econimic situation there have there policiies. They wouldn't have nothing to do with the ecology.

**Mark_Everson** · May 29, 2000, 00:15

Shimmin:

To cover your "management strategy" blocks of provinces I had envisioned the player to be able to issue orders on the level of Province "types". The types would be something like Core, New Conquests, Colonies, Front-line or whatever arbitrary ways that the player would like to divide provinces into classes for administrative/economic functions. So although the individual provinces would be limited in their size, most management could occur at the province-type level when the player gets more than a few provinces.

Richard:

Well, to make you feel better, Italy is something like 60 squares in demo 4. That may have been where you got your impression from! While I would like to put some of these discussions behind us, I don't want to do it prematurely either! By my calculations modern Italy actually should have something like 25 squares (60 miles per side), putting it solidly in the realm of a single province, but not by as much of a range as you stated. A large ancient province would be something about the size of France at the current scale. Of course, we may also decide that at this scale there are just too many squares, and enlarge the squares by up to a factor of 4 in area for the "standard" game. That would indeed make it more like your 10 squares.

I picked the current province size scale because it seemed a reasonable compromise between governmental provincial functions, reducing micromanagement as much as possible, reducing the number of calculations, and economic realities.

All:

On the general comments that micromanagement of province sizes would be the inevitable result if we let the player manage province size... I think you all are clearly right to some extent. I guess I'd characterize it as static provinces or purely computer-generated ones, are an even greater evil IMO. But I guess we will see how this argument plays out as we go forward into playtesting these particular aspects of Clash.

Announcement

Provincial Sizes

Provincial Sizes

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment