Provincial Sizes

[Richard Bruns]

That is a possibility. We'll have to see what Mark says. His current econ model might already be good enough for what I want.

[Mark_Everson]

Richard:

I think multiplying the number of squares by 16 would be pushing it. Actually, I'm almost certain it is prohibitively expensive. That's because we already have 64 K. squares. Another factor of 16 would give us one million. Even if you thrift on the sea squares in terms of memory usage, that is an enormous amount of memory to tie up. Even the incidental amount of processing that needs to go with each square would be prohibitively expensive if there were one million squares IMO. Personally I think 64 K. is already probably too many...

LGJ:

I'd thought of something like that also when Richard made his suggestion... Let's see how this stuff works out in playtesting since the existing Econ model may give us good enough results.

[Richard Bruns]

If no per-square calclations were needed, then the number of squares would not matter. And if we have the ability to make squares and provinces smaller, then a lot of the objections to my proposals would go away.

Almost all wars in history have been fought in a surprisingly small percentage of the land area on the planet. Everything else just isn't suited for military action. So we can make those provinces smaller; in some cases they could be the size of two or three squares on the current map. Especially valuable of strategic places could be even smaller. A system that had those provinces being conquered whole would not be appreciably different than the current system of square-by-square conquests. Meanwhile, the squares no one cares about can be lumped into provinces of thousands and thousands of land tiles that require no extra calculation effort at all.

The military aspect seems to be the biggest objection to the uniform province proposal. But if we make the computer smart enough to make small provinces in strategic areas, then the province based fighting loses its military restraints. IMO programming the computer to do a good job of province assignment would be a lot easier than programming a good enough AI to handle square based troop-micromanagement fighting.

The static provinces idea is an all-or-nothing proposal. If provinces are static at the current square size, there will be problems. If ecological and economic provinces are not combined at a smaller square size, there will be problems. But I think that if provinces can be homogenous, static, and small, then the game would be improved.

[Lord God Jinnai]

Well i might see it as okay, maybe. I'm not saying for sure now, if the max size for a province were 9-10 squares. But then also, wars have also been fought over seemingly insignifigant pecies of land also, only wanting small peices.

Another point for those larger provinces is the fact that say my civ and another civ wanted to carve out that province between us, we couldn't. We'd be at war with each other then, even if we were allies.

[Richard Bruns]

I think I have a compromise for the province structure issue. I will first outline what I think the basic issues are:

The Big Idea: Homogenous econ/ecology provinces could improve the game by making many calculations depend solely on the number of provinces, not the number of squares. This should make the game go faster and allow the map to contain more squares.

The Big Objection: Wargamers want tile-based combat and flexible provinces.

After pausing to consider these two points of view, I think that they do not have to be at odds. It should be possible to create a system with small and homogenous, yet flexible, provinces.

The splitting of provinces due to military action is easy enough to do. The problem is that if there is not some mechanism to reduce the number of provinces, any military action will create extra provinces and the number of provinces will become unmanageable.

However, this splitting would not be a problem if provinces could combine and grow. This was proposed sometime before, and even if the ecological provinces are the same as the economic ones, it makes sense for provinces to combine.

I am defining a province as a region with an infrastructure, society, and environment similar enough to be treated as identical on average and administered as a single unit.

As people continue to inhabit a plot of land, the original natural characteristics of the land become less important. The natural vegetation is slowly replaced with human dwellings and farms. Also, the original characteristics of the land will have less of an impact on farming as people change and improve the farmland. If the two original ecological provinces were similar, the differences between them will become minimal and they can reasonably be combined.

Infrastructure will also tend to become similar if two similar, adjacent places are being managed by the same civ. If the provinces are not radically different, we can assume that their infrastructures can be averaged and the provinces combined.

Society and culture also tend to become more similar as people are managed by the same government and live in the same area. No one can tell the difference between Goths and Lombards anymore.

Given this, there should be no problems with combining stable, similar provinces within the same civilization. If two regions have reasonably similar infrastructure and environmental conditions, they can be combined into a single province that is the average of the two original provinces. History supports this province expansion; the size of the divisions and administrative regions of countries has increased over time.

The mechanism for province combination will require play balancing, but we should be able to fix it so it counters the province increase due to military action.

In addition to the benefits of the homogenous provinces, this system should make the player's job easier. By combining provinces, the player's management tasks are simplified. As civilizations grow in size, the provinces will grow with them. The provinces in the interior of the civ that have been a part of teh civ for a long time will require less management, while any war zones on the borders will have smaller provinces that allow for detailed military actions.

On the wargaming note, I think that homogenous provinces would allow us to increase the number of squares on the map by a factor of four or possibly even nine. This would mean vastly increased opportunities for small scale strategic movement. As provinces could now be split, this proposal should make wargamers happy.

[Lord God Jinnai]

I generally like this idea. I have only 1 thing to add of anything major to this.

1. Technology should also play a part in province combining, basically if it exceeds maximum size for the current technology levels, it cannot be combined. We should also have a maximum size, reguardless of technological level.

There is only 1 problem i forsee with this, and it would be with any province model based solely on economic, ecological and infrasture. The problem comes with player admistation for those who want to do some micromanagment.

Lets take modern day Eastern Russia (Basically what's beyond Europe) as an entire Civ. Based on technological, economic, infrastural (virtually none), cultural and ecological information there would be at the most 2-3 provinces, if not 1. I'll say the most, 3. 1 for west section (Province A), one for Manchuria and where there's more of Chinese/Mongolian Culture (Province B) and the rest (which is most of the Civ) (Province C). Now that's all well and good, except if I, the player, wish to tell Province C to increase its mining capasity around Sibera, but only there. The rest I want to increase farming. Under the current system I can't because it would be averaged out across the whole province (Note that the mountainous regions would prob not be enough to put that section in another province since many provinces with differnt geological features are tied together). What I'm saying is that there should be a level of administrative provinces which normally as set by the base level provinces (rhw ones we've been discussing), however the player can change these. The only thing this does, is change how each section is adminstated as far as governmental affairs go, nothing else. There should be a limit on these as far as number and minimum size (not maximum) so the player doesn't micromanage every square.

[Mark_Everson]

Richard (and some applies to LGJ):

It seems there is no important detail in the game that you're not willing to change so that you don't have to modify your ecology system. Sorry, I shouldn't joke like that, I know you have the best interests of Clash at heart.

But Most of the things you would like to change were locked in long ago. Those particular details, like the size of a square, or the size of a province, or how many provinces there should be in a civ, have gone into numerous discussions in determining how a variety of models should work. If you are interested, you can use the search capability of Apolyton to find some of those old discussions. We just Can't go back and rework models from their foundations unless they are shown clearly not to work in playtesting. You can certainly discuss your ideas all you want, but we Cannot afford to turn the project upside-down to implement them. And I don't have time to discuss at length all your suggestions for ripping things up and starting from scratch. Most of our ideas will either be shown to be right or wrong only in the light of playtesting. Until we are at that point I really need to focus on implementing the existing models that have been more or less agreed-upon.

The only reason I'm posting in response to your proposal (which I estimate is going to take an hour of time out of pushing the coding forward) is I'm afraid you will take my silence for approval. And I also wouldn't want you to waste your Clash energy, which is quite valuable to the project, in a direction that isn't likely to do any good. I will certainly concede that your ideas might possibly result in a better game than the ones a bunch of us agreed to a long time ago. But nonetheless, those old assumptions have been discussed by many people, and locked into the project for some time. Given the project's painfully limited coding resources, I really don't think we can explore any of the alternatives at this point.

Just to fill you in on some of the thoughts that went into why things are the way they are, I will review as tersely as I can, why I think the provinces and map square core ideas should not be modified. Again, in general I think it would be of some value to re-address all these issues... but from the perspective of what we have available, I think the best approach is to get to playtesting as soon as possible. A few non-ideal models won't kill the project. Endlessly debating everything just might...

Some thoughts on the "standard" models vs. your recent proposal(s):

Provinces:
These have a multifaceted role in the game. The most important purpose of the province is as an administrative unit for the civilization. If provinces are made so small that you need 50 or more of them to have a decent sized empire than they have failed this purpose. I would vastly preferred if most civilizations could be run with 10 or less provinces. Your proposal violates this exceptionally important (IMO) design criterion.

The second purpose of provinces is to make the economic model, which is rather calculation-intensive, feasible to use in the game. I would actually prefer that it took significantly less time than it does, but I think we can get away with the strategy that has been mentioned earlier in this thread.

You seem confident that we will have enormous amounts of processor power to throw around just because we can make big provinces over much of the globe. (And in fact it Is a Very useful insight) But, considering only "militarily active" regions you are still left with something like 30% of the land area. (Figure taken from John Keegan's "A History of Warfare") So in fact your proposal is only gaining something more like a factor of two in efficiency of provinces given that even the "nonmilitary" land has to have some provinces. Additionally, given our ambitious targets for AI and realism, a factor of two gain in province economy in efficiency probably should just be reserved in case something else takes longer than we hope.

Map Squares:
First, as I said above, you are not really gaining much processor time, so you don't really have more to spend in putting yet more squares on the board. In addition to the marginal cost of extra squares to the economic model, the graphics cost of extra squares is not negligible. Plotting tiled graphics takes much more time for lots of little tiles than it does for the same screen area of larger tiles. I am already concerned that our graphics may be too slow once the number of tile overlays for roads, military units, and tokens to give the player various information are performed.

Second, the 100 kilometer tiles are designed with strategic movement in mind. It was thought that anything smaller would make the military part of the game too tactical. This would conflict with the overall strategic "leader-oriented" nature of Clash. If it makes you feel any better, I was once on the small-tile side of this argument. However, I was convinced otherwise.

Lastly, if the additional number of tiles really gives more military options, then doing the AI will become yet more difficult. The clock cycles required for the map AI clearly increases as the number of tiles does. Unfortunately the clock cycles increase not linearly with the number of tiles, but as a power law. For that reason, I have considered at times reducing the number of tiles. However, I decided to wait for playtesting to see where we come down on this issue.

Anyway, I'm really sorry if I'm being harsh. But there is still an enormous amount of design work to be done in terms of looking for fatal "gotchas" in the models, the interfaces for the models, and looking at how the major models interact... We need to move forward, building upon what we already have. If we are continuously worrying at the foundation elements of Clash, I really fear we will never get significantly far beyond them. It may be that some of your and LGJ's suggestions in this thread Can be grafted onto the existing models to improve them. I'll try to look at that tomorrow. But for tonight my Clash time is all used up.

I am asking you as a personal favor to please just figure out how to integrate the ecology model into the rest of the framework as that framework exists...

Mark

[Lord God Jinnai]

Please do not take this post as being negative or anything. I'm not always the most diplomatic.

Can you post basically what was decided because I did look b4 this came out and what I saw was about provinces that A> they should reduce porccessor time by reducing the calculations needed for each square (The whole reason for them) B> They should grow over time dependant on technology C> Based almost completely on ease of adminstration.

tiles: (which BTW i was basically standing neutral on since as far as I see it, if a player wants more tiles, he can create a bigger world and not use the standard one, but he should know that it might eat up proccessor time) Anyway, on tiles, I seen that A> the STANDARD would be 100km, but for scereio purposes the size can change. B> The number of tiles can change to suit differnt worlds/scerios, but that there will be a STANDARD number C> There will also be a STANDARD number aloted to ocean to help reduce proccessor time (though once you start having the ability to put cities in oceans, it may be diminished somwhat).

Finnaly, I am open to compromise with the current system if it seems to work out well. It just seems that the current system may not be able to hold up to all these new models as well as it does the military and economics model, which seems to be what it was based around.

[Richard Bruns]

Mark:

My recent province proposals were entirely incompatible with the current ecology model, and I knew it. Most of that model is designed around dynamic border-shifting provinces. If province shapes were made static, I would have to completely redo all of the terrain change aspects of the model. Doing things your way means less work for me. As soon as I know how many farm sites should be on mediocre land, bad land, great land, etc. I can easily output the farm sites per tile.

I didn't know that so many things were set in stone. I guess my view of the Clash design system was skewed by my work in the technology model. LGJ and I entirely restructured and rebuilt a model that everyone had already agreed on and done some work on, and the response was positive.

I won't waste any more of your time by trying to come up with a compromise on something that apparently will not change.

Thank you for your response. I prefer frank discussions to diplomatic niceties. In your earlier posts I thought you were discussing the pros and cons of the new system, so I was trying to work out compromises. In the future, don't be afraid to shoot down at first sight something I propose that has no chance of being implemented. That said, I am trusting you not to shoot down something that could change but that you don't like in its current form.

[Mark_Everson]

Richard and LGJ:

Sorry for the misunderstandings. We appear to have sorted things out in a reasonably efficient fashion.

I try not to be too dogmatic when wearing my hat as "lead designer". After all, everyone wants to get their own fun ideas into Clash, not just flesh out mine! I guess the way I'd characterize our design process is that if something is "set in stone" but all the stakeholders agree that there's a cool new idea they'd like to explore, then that's fine. (I need to add the caveat here that it shouldn't hold up existing progress, or throw away a bunch of existing work either) All these criteria were fulfilled with the technology model work. In this case the potential changes involved a lot more things that were already firmed up in a bunch of different models, and at least one stakeholder (me) who thought the new proposals were actually going in the wrong direction.

I'll try to be quicker next time if the proposals appear to be going against something that's locked in. And I won't abuse my awesome power ;-) to shut down a line of inquiry simply because I just don't like it.

[This message has been edited by Mark_Everson (edited June 01, 2000).]

[Lord God Jinnai]

So then, what, if any, of the proposals we've maid could reasonably be put into clash as far as the ecology model and provinces go.

[Mark_Everson]

LGJ:

As I'm sure I've demonstrated I'm not the most diplomatic person either... I'm at work right now, but had a few quick thoughts I can add that might help on your post. (And I still haven't had time to look over possibilities in the new proposals that could be appended to the existing system...)

Specs:
I think I have presented all of them above. In order of importance

Provinces:
1. Give player and AI reasonably small number of empire subdivisions to work with (ideally 10 or less, Max 20 over the long haul) Primary use of subdivisions is for govt and econ.
2. your A,B, C above.
3. Ideally 20-50 squares. (Size of historical large provinces)
4. Since empires are expected to grow, with conquest and technology, provinces should be dynamic to satisfy 1.

Map Squares:
100km
1. military: some tactical flavor remains esp for modern warfare;
2. AI: can't have too many squares to deal with;
3. graphics: get slow above this with target comp (may be a bit too much anyway when we get all the map levels going)

I should have stated it better in my post... I am talking about the square size that the game is tuned for, assuming something like a 600MHz P3. We need to use something as a standard, and what I didn't like was the suggestion that that be greater than the already large number (64K!) of squares we have. I in no way meant to throw doubt on all the discussions we've had about giving the player options in square sizes. The player should be able to go to a million squares if they have the desire and hardware for it. But remember we are scaling the default tileset to go with a certain number of tiles per screen when it looks best...

My main point is that if a new model doesn't fit within this framework then the new model should be the one to change. A year of discussion has already gone into these things! I'm sorry you guys weren't here for all of it. I realize its somewhat unfair in the cosmic scheme of things that you've been presented with a fait accompli, but that's life.

And, specifically, I don't think the ecology model would have to change much at all to fit in perfectly. I just don't see how the govt/econ and the ecol provinces can be made to match, at least as I understand the ecol model. Players will want to have all of forests, grasslands, and mountains in one administrative province some of the time. FE, all of modern Italy. And I think we need to let them do so.

[The Diamond]

Just for a moment, return to a core issue here. Provinces are primarily administrative in their founding. Right - hold on to that thought.

What we're looking at is maybe not defining provinces so strictly early on. In the very beginning, govts. won't be large enough to consider the idea of a province. It won't be until the civilisation starts to get a little too large in either population or size that separate provinces will need to occur.

To clarify, you need 2 separate administrations when the population is too big for one to handle or for when the geographical area is too great to convey orders, etc. to the far borders.

However, I think the decision of when and how to divide provinces should be left with the player (with suggestion and leading from the AI).

As an example, let's say I have 10 cities strategically placed in a circle around my capital. In total, I cover about a diameter of 20 map squares and an area of 314. Forget about whether this is realistic or not for now.

Anyway, I decide to make this all one province. My problems? Orders sent to the peremiter of my realm take ages, corruption is high the further away from the provinical capital you go, taxes take a while to come in (probably reduced as well) and military production is harder to co-ordinate. Also, my provincial policies have to take into account 11 cities and a huge geographical area...not easy to do the best for everyone.

So, on that note, I separate my realm into 11 provinces. Now I have the opposite problem. Orders are quicker, corruption is reduced, inefficiencies in policing and admin are reduced, but their costs rise. Military production is much easier to co-ordinate and I'm more flexible. However, if you incur a running cost for every province, then I'm paying through the nose to run this. Probably I'm taxing like hell and people are moving out of the current provinces, making me even larger and making the problem worse.

I think the player has to find a trade-off that works for them. Maybe the whole Sahara desert, where population is thin and the terrain, etc. doesn't differ much, could be covered by one province. However, in the US, where terrain varies enormously very quickly and population is high, maybe you'd need up to 52 provinces?

Also, there's nothing to stop a disgruntled outlying city declaring itself independent if you refuse to make it a province...

Comments?

[Mark_Everson]

LGJ:

I think the ecology model can be used pretty much as designed. As Richard said, he just needs estimates from me as to how many sites different types of squares should have. I have posted that stuff in the ecology model thread.

What I've taken away on provinces is that we should use really big provinces on fairly useless land, work hard to not give players the incentive to excessively micromanage their provinces, and try to encourage provinces based on geographic and ethnographic realities. Just about everything else that was brought up in this thread, I think will be better addressed when we can actually playtest some of this. Ideas regarding the economic nature of provinces, we can start to explore with Demo 5. For the government ramifications we will have to wait for demo 6 unless someone dying to code the government model shows up very soon.

The Diamond:

I think your ideas are pretty reasonable in general, although I might quibble with some of the specifics. The only thing I'd stress is that we need to make sure that these trade-offs are fairly gradual so there is no incentive at the margin to micromanage provinces. So specifically in your example when the player's civ started to get near 100 squares, they would start paying a serious net penalty for having everything crammed into one province. However, I'm pretty sure that we definitely don't want the United States to have 52 provinces, although I guess if the player really wanted it that way they could have it so. But there should be no net advantage for the player of 52 vs. 10 given modern technology. And we would not want the AI to have such small provinces, because the calculation load would get too large if everyone had that many provinces.

[Kanzid Stonebreath]

One quick question how many provinces would people say the modern U.S. would be approximatly where would the province borders be, and is the U.S. your idea of an average sized player nation in the end game?

------------------
What does this box do I wonder?