The pitfalls of expanding city-areas in Civ-3

[The diplomat]

quote: Originally posted by Ralf on 11-30-2000 04:53 PM I have read your and Tiberius posts a couple of times, and i think must redraw some of my early entusiasm Also, i think both yours and Tiberius calculation-examples suffers the same lengthy read-a-couple-of-times mathematical hard-to-digest problems, as in my critized example from the CTP-2 post. What i DONT want is a game that forces the player to understand some multi/complex-related, multi-benefiting, percentage/fractions-numbered AND/OR/ELSE boolean mathematical god-damn formula. Sorry [This message has been edited by Ralf (edited November 30, 2000).]

I guess I should have stopped while I was ahead

Well, I completely understand your point. I think you might have misundertood me. My explanations were meant to explain the inner workings, how the model works "under the hood". The player would not have to play with complicated %'s or anything. The city screen could be just as user-friendly and graphical as civ2!

If you don't mind, I'll try to explain my city idea again, this time with no fancy numbers. The idea is as simple as this:
-a city collects a fraction of the total city radius based on number of workers. More workers, more ressources.
-Exactly like in civ2, the player would see amount of food,shields and gold the city is collecting.
-food would allow pop growth. shields allow city improvements to be built. gold goes in taxes and science "beacons".

As far as the player would see, it is just as simple as civ2 but with less micromanagement.

I agree with the KISS principle. But I have many problems with the civ2 model. It is not a good model.


------------------
No permanent enemies, no permanent friends.

[airdrik]

Back to how to implement my idea of the city expanding at a 1 square-at-a-time basis, Forget the entire civ being one big conglomeration, and forget the merging of cities. instead, if a city gets big enough, you can choose to separate it into two different secters, kind of like the production centers I suggested. You don't have to.

As for how the ai would do the expanding, it could do it one of two ways: one is that it looks at all the adjacent tiles and picks the best one (or randomly selects between a tie). The other way is that it looks for the closest resourse tiles within x tiles and aims for it, adding squares until it reaches it.

Another way it could be implemented is that in order to expand into a certain adjacet tile, the tiles around it had to exceed a certain cumulative pop: x to expand into fertil ground, 2x for forest and hills, 3x for desert, swamp, and jungle, and 4x for mountains(if you expand onto them at all)

[Ralf]

quote: Originally posted by The diplomat on 11-30-2000 06:56 PM My explanations were meant to explain the inner workings, how the model works "under the hood". The player would not have to play with complicated %'s or anything.

Well, that i understood.

But, what the Civ-2/SMAC city-area model does so brilliantly (and that civers dont seem to miss, before its actually is streamlined away) is to display the exact inner relationship between each used city-area tile, and then graphically explaining by how much each of these tiles actually contributes to the city´s overal food-, recourse- and treasure-storages.
With a blink of an eye you can see that this developed tile contributs with exactly this amount of foods, shields or coins.

To have that kind of information about the "inner workings" of the city-area, in such a direct and intuitively easy-to-overview graphic format is really an example of KISS-philosophy in its purest form!


In your model, Diplomat, the player can ONLY see the total ADDED UP AMOUNTS. He cannot directly see and understand the exact inner relationships between the used tiles - and by how much each of these tiles actually contributes to the city´s overal food-, recourse- and treasure-storages.

This is going to raise questions amongst future civers (as it has over at the CTP-2 section).

Questions that that only can answer with some "number-crunching exercises" and lengthy "read several times before you understand" explanations (as it has over at the CTP-2 section).

[This message has been edited by Ralf (edited December 02, 2000).]

[The diplomat]

Ralf: Now I think I understand what you are getting at. You are saying that the "total city radius" system is less intuitive to the player because it simply gives the player the end result and hides the way it got there. Whereas, civ2 by graphically showing how much each tile contributes, ends up being more intuitive to the player. For example, the civ player can easily see that, "that forest tile is giving me a bunch of shields and that other tile over there is giving me lots of food!" Am I correct? If so, you make a very strong argument and I agree with you completely.

Do you feel that civ2's city model (specifically, how it collects ressources and uses them) has any MAJOR flaws? If so, what are there? If there were a model that was different from civ2 but just as intuitive, would you like it?

Here is a new city model that I have come up with that tries to adress your concerns. You should like it:

Only certain tiles would produce ressources. For example, forest, mountains, river, beach, and special etc... (note how important good city placement would be!) When you settle a city, the tiles that produce ressources would show up in the city screen. Just like in civ2, the player would see little ressource icons on the tile itself thus showing the player what tile is producing ressources and how much. However, the player would not need to actually place workers on any tiles. All tiles that are allowed to produce ressources would produce ressources no matter how many workers a city has. But, the number of workers would determine how much ressources a single tile produces. So, the number of resources each tile can produce would increase with the population of the city. The player would see graphically how a tile is producing more ressources with more workers: a tile would have say 2 little "shield" icons on it at first, and 4 little "shield" icons later when the city was bigger.
The player could make the tiles that don't produce anything, start to produce ressources by placing tile improvements on them. For example, a grasslands tile would originally not produce anything but if you built a farm improvement on it, then it would produce food, and the player would see the tile with little "food" icons on it to show that it is now producing food.
So, this models would show what each tile is producing and how much in the same way as civ2 (ie with "food", "shield" and "gold" icons on them) but at the same time the player would not need to actually place any workers on them. Workers would simply increase how much each tile produces.

What do you think? I tried to combine the best of both models.
The reason I am so adamantly against physically placing workers on tiles to produce stuff, is because I don't like that system. I believe there has got to be a better way that reduces micromanagement while still remaining intuitive for the player.

I want to finish by saying that this thread is turning into a very intelligent debate. I am really enjoying our discussion.

------------------
No permanent enemies, no permanent friends.

[Ralf]

Diplomat quote: "Do you feel that civ2's city model (specifically, how it collects ressources and uses them) has any MAJOR flaws?"

I dont think the Civ-2 "Place workers on 21-tile city-area" model is a BIG problem at all. But that i think originates from my own personal playingstyle: founding max 18-25 cities, and only expand/conquer after the Alpha Centauri launch - and then the hassle of micro-managing a really large 40-60+ cities empire was an overcome-able problem, because it was relatively few turns left before the game was over.
I realize however that Firaxis cant develop a game only on basis of this specific playing-style. There are different needs out there, and its upto Firaxis to come up with effective city-mayor solutions, so that these micro-manage hassles can be minimized.

Its another thing however, that this model could be tweaked a little better. I always thought that AI-civs over-emphasized foods and under-emphasized shields a little. But as i said in some earlier post: That should be completely tweakable trough the txt-files in Civ-3.

Diplomat quote: "If there were a model that was different from civ2 but just as intuitive, would you like it?"

Im not principally against changes in the city-area model, so rest assure: your are not talking to a stonewall here. (as if i had any kind of influence over it whatsoever)

Its just that designing a complete overhaul of the Civ-2/SMAC city-area model, simply cannot be compared with just another feature-update. If a strong AI is the "engine", then the city-area model is the "gearbox" of the TBS Civ-game concept. This game-mechanical feature is really that important.
Unbalanced units, city- or tile-improvements can be tweaked in the txt-files; unbalanced governments likewise; added small features with bad side-effects can be toggled off/patched and many other small bugs or minor less succesful features can be fixed/tuned successively with patches.

But a flawed city-area model?

Given all the other complexities and demands on the game, i would be rather cautious to change it - at least change it radically. The problem is that i cannot see how that intuitive "inner workings" graphic tile-information can be displayed in any other, and better ways. You have mentioned added tile-info to your idea in your last post - but im having trouble with the idea itself.

quote: Originally posted by The diplomat on 12-01-2000 07:52 PM [quote]However, the player would not need to actually place workers on any tiles. All tiles that are allowed to produce ressources would produce ressources no matter how many workers a city has. But, the number of workers would determine how much ressources a single tile produces.

quote: The player could make the tiles that don't produce anything, start to produce ressources by placing tile improvements on them. For example, a grasslands tile would originally not produce anything but if you built a farm improvement on it, then it would produce food.

The problem is that with above model you have replaced city-area micro-managing with a new frustration-factor, and also probably increased the workload of the AI.

With the great Civ-2/SMAC model the player can choose freely what to emphasize:

- more workers on grassland to ensure a faster population-growth,
- more workers on forrest and mined hills to ensure faster resource-gathering, or
- more workers on sea/ocean to ensure a higher rate of trade-earnings

With your model the player cannot choose what to emphasize, and thats likely to get some people (me included) rather frustrated.

You can argue that one could put in emphazise-sliders in the city-screen there players can tweak what kind of output they want.

But, in this case, we are back to square one: You have replaced micro-managing the city-area view, with micro-managing those city-area sliders.

What is the principal difference here? (This seems to be a Catch-22 type of dilemma)

Also, in the city-area view the AI only does what you command it to do by point-an-click. Working with sliders the programmers is forced to "translate" complex changes from the sliders to the even more complex variations of city-areas. And that is likely to produce a rather "fuzzy" relationship between sliders-input and city-area view output.

Totally anti-KISS if you ask me.

quote: Only certain tiles would produce ressources. For example, forest, mountains, river, beach, and special etc... (note how important good city placement would be!)

Yes, but the human player can easily overview and easily estimate were to best place his cities.

For the poor AI-programmer however, working in hes black-box of infinite combinations of possible terrain-tiles mixtures that the map-generator can produce, the task have become even harder:
How the heck can he program the AI to effectively find those important good city placements, on an totally unknown map, amongst big quantitys of non-producing tiles?

[This message has been edited by Ralf (edited December 03, 2000).]

[Ralf]

quote: Originally posted by Tiberius on 11-30-2000 03:13 AM make all buildings in a city population dependant! What do I mean? First of all, in the Civ2 system, you can build in a 3 or 4 size city all the CIs that are available. That's stupid!

But, isnt that a rather unlikely and hypothetical scenario?

WHY would the player want to build all available CI:s in a 3-4 sized city?
WHY invest so much resources in any small city?

It doesnt make sense.

The whole "problem" seems totally academic to me. Remember that the effectiveness of advanced late-game city-improvements IS often heavily dependent on population-sizes:

- The effects of universitys and science-labs is almost negligible in 3-4 sized cities.
- The effects of stock exchanges and super-highways is almost negligible in 3-4 sized cities.
- The effects of factorys and manufacturing-plants is almost negligible in 3-4 sized cities.
- The effects of chatedrals and police stations is simply not needed in 3-4 sized cities.

A full set of trade-routes can of course boost a small 3-4 sized city´s economy - but, by the time these trade-routes are established in those cities, they have already growt far beyond the size of 3-4.

[This message has been edited by Ralf (edited December 03, 2000).]

[Tiberius]

Sorry for not participating to the discussion, but I have Internet only at work, so in weekends I can't post.

Back to Civ3:
I quote Diplomat: "certain CI's should require a small amount of pop, to prevent small cities from building everything". That's the idea. Ralf, it's not that somebody wants to build everything in 3 size cities, but players should be forced to decide what they want: a lot of smaller cities close to each other or fewer, but better developed large cities. If you can't build factory+hydro_plant+mfg_plant (or library+university+research_center)in a 12 size city, that's a loss. So why shouldn't we encourage the development of big cities? And I ask you once again: Why only resource gathering is linked to the city size and resource processing (goods production) or other activities not ???

I agree that some CIs won't need any pop-support: barracks, aqueduct, granary. (Btw, Diplomat, the numbers were just an example, based on a first-look-common-sense. Important is the concept, at least now; details later)

quote: A good aspect about this idea of "pop support for city improvements" is that it forces the player to make a decision as to what city improvement to build. The player would have to make a choice because a city could not have everything right away. This adds more strategy.

That's a very good and valid point. Maybe Ralf didn't read it

Ralf, you are worried that players won't understand the system. Now, if you look at the city screen, and see that 40% of your pop is working on fields, 40% are working in CIs and 20% are unemployed (and they are unhappy), what's so hard to understand here? It's just like real life.

We started this (the thread) talking about the expanding city area concept. We all agreed that those 5-circle-layers huge cities from CTP2 are not the best anti ICS sollution. Linking pop to CIs it's an effective and easy to understand concept that helps big cities to become more important.

One more idea: let bigger cities to have more than 3 trade-routes. The trade-bonus will be one more arguement in the favor of big, well developed cities.

About Diplomat's tile-system: I don't know if I like it or not. I must think on it.

[Ralf]

Well, Tiberius - maybe i got myself too hooked up on that 3-4 size-example of yours. I didnt look any further. I realize that you and Diplomat have a point about half-max sized cities (and thats what a city of 12 population-points basically is, more or less) having virtually each and every city-improvement on the tech-tree, within its walls. So, should it be possible to build ALL city-improvements there is, in a 10-12 point city, in civ-3 as well? Answer: NO

On the other hand: I woudnt like it if i was forced to build "improved farmland" on 80-90% of the city-area, just to have that 24-28 city, in order to build those final end-tech CI:s, either. The reason is simply that such city-areas are to monotonous and darn right ugly to look at. I want a middle-ground all-CI:s size solution here.

If Firaxis ever decides to implement the "field/CI/unemployment-idea", i want it succesfully merged with the input-able city-area view in Civ-2/SMAC. With input-able city-area view the player can:

- See and understand the inner-workings between all used tiles within a blink of eye.
- Let the field-workers emphasize either food-, shield- or coin-generating tiles.

If it CANNOT be succesfully merged (= easy to overview and understand), i rather stick with the old Civ-2/SMAC input-able city-area view only. Any ideas?

Finally: Remember that there are other suggested game-concepts that fights over city-screen space as well. I have some hopes for a brand new stand alone game-parameter: Health.
I like it, because the health-parameter generates many nice spinn-offs. The basic idea (i think) is that full health means that the foods-, shields-, coins- and lightbulbes-production continues problem-free, like in Civ-2. Beneath full health means that all those four areas suffers gradually and accordingly.
Another reason for good city-development: Granary, Bathhouse, Sewersystem, Farmacy, Hospital

[This message has been edited by Ralf (edited December 04, 2000).]

[Tiberius]

I just can't stop thinking that the major decisions about Civ3 are already taken, so it's a bit useless to debate here field/CI/city radius/.... But, anyway, good ideas never hurt.

I agree with you about health (just don't see the connection with our debate so far ). I hope that a missing aqueduct won't limit the size of a city to 8, but along with the lack of other health improvements (bathhouse, pharmacy, hospital) increases the chances for diseases, plagues, etc.

Oh, I see now a connection. It's about (again) strategic decisions: I must create big cities for production and trade bonuses but I have to take care about the people: health, employment, happiness.
I like it

[Deathwalker]

Personally I like the idea of a fixed max city area size, like civ's 21 squares. I think if you make it expanded like I hear CTP 2 the placement of cities can be ruined?

------------------
I have walked since the dawn of time and were ever I walk, death is sure to follow

[Grumbold]

It seems to me that Civ-2 controlled excessive improvement building in small towns by maintenance cost. Until a civ has grown into a sprawling Republic with several well developed cities it cannot afford to support loss-making improvements like banks or factories in a size 3 city. The CtP games have run into a few problems because they don't seem to tread cash as such a hard to obtain commodity. Requiring population to support an improvement is another quite elegant solution but I instinctively dislike fractional costs. I would be tempted to leave no pop cost on early construction then add them in integer units to banks, factories etc.

The map scale of even the largest game world is unlikely to get anything but the biggest city to occupy more than one tile. However I like the concept of expansion. I would suggest that a city could exploit:
1 square with no improvements
2 squares with connecting road improvements and a depot in the city
3 squares with connecting rail plus railyard
4 squares with motorways/maglev and advanced depot.

That should allow cities to grow in more sensible ways, i.e. not a rigid radius but drawing resources along communication routes. Only the completist would bother pushing motorways out through appalling terrain just to make every tile within the maximum radius exploitable. Meanwhile two large cities could grow up close to each other but on opposite sides of a mountain range.

Slightly off topic, but I also like the idea of terraforming and motorway/maglev being impossible through mountains. Man's ability to adapt his environment to himself has some limits!

[Ralf]

quote: Originally posted by Grumbold on 12-05-2000 11:16 AM Requiring population to support an improvement is another quite elegant solution but I instinctively dislike fractional costs. I would be tempted to leave no pop cost on early construction then add them in integer units to banks, factories etc.

I agree about no pop-cost on early CI:s, and also the integer thing. Lets keep it simple in those future Civ-3 game-manuals, shall we.

quote: However I like the concept of expansion. I would suggest that a city could exploit: 1 square with no improvements 2 squares with connecting road improvements and a depot in the city 3 squares with connecting rail plus railyard 4 squares with motorways/maglev and advanced depot. That should allow cities to grow in more sensible ways, i.e. not a rigid radius but drawing resources along communication routes.

The fixed 21-square area is fully acceptable for me, but i think also your idea have some nice points (i reckon you mean 4 squares = 4 squares away, of course). However, under three non-negotiable conditions:

- Apart from the city-tile itself - max 20 area-tiles (to avoid "city-area tile-inflation", like in CTP-2).
- Powerful non-city-area related anti-ICS preventions (otherwise ICS gets worse, by this approach).
- I must have that input-able city-area view - without it, this idea is 100% dead - at least in my eyes.

Under those conditions, i could just about accept this idea. But what about ocean-tiles? Also, maybe 3 tiles away is enough? - otherwise that city-area view gets too big (or tile collect-symbols too small).
My fear is also that the AI perhaps cannot exploit this idea nearly as good as the human player. If this approach is more "AI-unfriendly" then the old one, i would rather stick with the old trusted 21-square fixed radius model.

quote: Slightly off topic, but I also like the idea of terraforming and motorway/maglev being impossible through mountains. Man's ability to adapt his environment to himself has some limits!

Non-transformable and mines/roads-only mountains is not just a question of realism. Also, the map looks nicer then not each and every square is railroaded or maglev-ed.

[This message has been edited by Ralf (edited December 05, 2000).]