The core of 4X games "boring endgame problem" - a short essay about Rapidly Increasing Complexity

[DeckerdJames]

I've given up on Humankind because Civ VI is so vastly superior, but especially late game.

Not bad for a six year-old game!!

The recent fixes make late game wars much more difficult, AI units harder to kill, cities harder to take, and this with Jet bombers and modern armor? Makes it much more interesting.

I may be in a minority but I don't automate - and wouldn't necessarily use the function. Yes its tedious very late game to scroll thru endlessly renewing "campus grants" or whatever, but by then you're only a few turns from winning, (usually!) right?

The build queue could get an additional check box (Boolean variable). When it is unchecked, it works like it does now. When it is checked, the queue loops, and anything that can be built more than once repeats. So if you had a queue with a warrior, a barracks building, and a city project, it would produce them in that order on the first iteration, but on the following iterations, it would repeatedly produce a warrior and a city project.

Essentially, they just have to have a modified version of the function that removes the item, when it is finished building, from the queue. After it removes it, it tries to add it to the end. It should not be able to do so if it can only be built once.

 

[DeckerdJames]

I agree despite disliking ro being different about almost every design decision for civ6, because Humankind was just such a miserable experience



See, that's the problem: if you don't like 1UPT combat system and unit management for whatever reasons (I do), and/or you are generally more interested about peacetime mechanics in 4X games than conquest (I do), and/or you don't have a lot of time to spend on video games (I don't), and/or your laptop can fluently handle Witcher 3 on high but civ6 next turn loading takes it a lot of time (somehow it is the case for me)...

If you have one of those problems, civ6 endgame becomes much less attractive. If you have many of those problems at once, like I do, civ6 endgame because completely unplayable. If I find 1UPT warfare system to be completely boring, lacking non - logistical challenge, and in general even besides that I think peaceful mechanics in civ tend to be much more interesting (especially as I like to "watch simulated living world of AI empires", not destroying it). And in this case civ6 endgame for me largely consists of hours of either joyless tedium (need to micromanage cities, or God forbid units if AI declares some stupid war and I have to manage army) or nothingness (waiting 25, 50 or 100 turns to fulfill obligatory conditions for victory you have already clearly achieved because the race in civ6 is already determined halfway through it). If you don't enjoy 1UPT conquest of AIs, civ6 loses a LOT of allure.

By "lacking non - logistical challenge" what I was trying to say is that the primary source of challenge in 1UPT in civ6 is not AI activity in itself, especially not AI offensive movements which are very rarely capable of threatening you. Nor do you have to actually think on the level of countering any sort of AI movements (I know AIs suck in many strategy games but in 1UPT they suck somehow much more). The main challenge of civ6 combat system is the gigantic pain in the *** constituting of the way movement, terrain and territorial control work. A lot of individual tokens you have to move one after another and which all create traffic jams with another, which often move at agonizing speed of one tile per turn because of a crapton of terrain obstacles, struggling with line of sight and range limitations, which you constantly stress about losing because individual experienced token can die in one turn after one wrong move, setting up siege engines only to realize you have no LOS or city one shotting them with not much you can do to prevent that - yes it is difficult, but it's because of horribly painfulin interaction of topography, logistics and interface, not because of AI armies being threatening by themselves and offering a glorious battle of wits and valor.

Don't anybody dare to say this is how military wargames and difficulty work by default and I just have to become better player, I have greatly enjoyed XCOM games on classic, I've finished famously "hardcore" Darkest Dungeon having pure fun, played tactical rpgs such as Original Sin 2 on quite brital difficulty levels, played many Total War games, won campaigns as Wallachia and Mysore and Ichma in eu4, I had fun with them, I don't have much of fun with 1UPT, so I criticise it knowing the context.

In my view, 1UPT could be greatly enhanced by some automation. Think of how in a real time strategy game you can group individual units together to form a larger company. When you select the company and reposition it, there is programmed logic that controls the movement for each individual unit. Otherwise, real time games would be drastically more difficult. Civ 7 could copy that and perhaps do better than merely assigning them to hotkeys. The main idea is to allow macro management. When you select a destination, they have position themselves in the target region and they have to move there smartly. Hopefully you can see a preview of where and how the company will deploy. The idea would probably require a good "unit organization" UI and a control panel for player created companies.

Secondly, automating the basic commands for a company should also have some automation. For example, once positioned, they are anchored, but some controls would allow for changing the parameters of their behavior. A slider bar could be used to control the size of a region that they will try to control and contest, though the company's knowledge of enemy units are subject to standard rules.. So then, putting them into a region and setting the slider would allow them to defend an area with out micro management. However, like any game does, overriding the actions of an individual unit should be possible. Turning off automation for a company should also be a selectable option in the company's control panel. The ideas though, "control", "contest", "defend", "assault", "recon", "patrol", and so on, are the automated behaviors. These are the sorts of buttons that would be on the company control panel. The control panel might be the same no matter how many are in the group, one or many, but how the computer automates the group depends on what the group consists of.

This kind of macro management makes big picture strategical decisions more of the player's focus. The units have to perform their functions smartly though.

 

[DeckerdJames]

Having to solve a sliding tile puzzle every time you move your units is absolutly awful as a mechanic and hilariously stupid as history. I’m not surprised it’s beyond the AI.

Cutting that Gordian Knot would solve so many problems with so many aspects of this game

Civ6 is so close to greatness, and it’s just a few small things that ruin that

Maybe they should test the idea of allowing units of the same kind to occupy the same hex. Instead of having corps and armies that are permanently formed, just allow the player to stack and unstack them. Not 1UPT, but rather 1UTPT (1 unit type per tile). Maybe there is a stack size limit? Not an actual stack of individual units, but they act as one unit with greater strength, but they can also be dismantled.

Personally, I don't want to go back to the "stack of doom", though. It just made the game less strategic and more logistical. You just needed "more".

 

[aieeegrunt]

Maybe they should test the idea of allowing units of the same kind to occupy the same hex. Instead of having corps and armies that are permanently formed, just allow the player to stack and unstack them. Not 1UPT, but rather 1UTPT (1 unit type per tile). Maybe there is a stack size limit? Not an actual stack of individual units, but they act as one unit with greater strength, but they can also be dismantled.

Personally, I don't want to go back to the "stack of doom", though. It just made the game less strategic and more logistical. You just needed "more".

There is a mod that does exactly that (one type per hex). If you are going to do that, why not make it a straight 3UPT and at the same time reduce range of ranged units to one and solve two major problems at a time.

All you need to do is keep unit and maintenance costs high and doomstacks solve themselves

 

[Noble Zarkon]

I don't want to go back to the "stack of doom", though. It just made the game less strategic and more logistical. You just needed "more".

I would argue the exact opposite! Stacks are strategic challenges because you are thinking about the timing (do I go now with X units or wait 20 turns and go with Y by which time I might be facing longbows) and composition (e.g. balance between siege and melee) and not the logistics of manouevering them all into the right place (because they're a stack!). The notion that stacks were all about size is a fallacy that somehow persists to this day.

 

[DeckerdJames]

I would argue the exact opposite! Stacks are strategic challenges because you are thinking about the timing (do I go now with X units or wait 20 turns and go with Y by which time I might be facing longbows) and composition (e.g. balance between siege and melee) and not the logistics of manouevering them all into the right place (because they're a stack!). The notion that stacks were all about size is a fallacy that somehow persists to this day.

Certainly you had to make decisions about the stacks, but it is different than needing to consider geographical relationships. It is a different kind of strategy. The problem I hear with 1UPT is AI incompetence and limited board space.

 

[Boris Gudenuf]

Certainly you had to make decisions about the stacks, but it is different than needing to consider geographical relationships. It is a different kind of strategy. The problem I hear with 1UPT is AI incompetence and limited board space.

What you don't hear about, and should, is the Time Space with 1UPT: that a single battle is not only spread out a hundred or more times larger in space on the game terrain than it should be (archers that can fire from one side of a large city to the other?) but also in Time. When in the Ancient Era it takes 80 - 160 years (2 - 4 turns) to kill a single Barbarian Spearman, What in God's Name is the 'battle' supposed to represent? My answer would be multiple campaigns spanning generations, which rather makes the graphic of 2 - 3 'units' fighting each other all that time look Silly.

In a Grand Strategy game like Civ in which the shortest possible 'turn' is One Year, no field battle should take more than one turn. Period.
Even that is a stretch, but it is acceptable to get a 6000+ year-long game compressed into playing time that takes less than a week or more of 'real time': no game combining Tactical and Grand Strategic decisions can be other than a bunch of compromises of all kinds, but the game needs to do better than the current Tactical Battle = Decades, War = Centuries that Civ gives us now.

What a 1 Year Minimum Turn indicates is that only the 'decisive' battles are being shown: the Borodino, Gaugamela, Sedan (1870 and 1940) that fundamentally decide where the war and the armies and possibly your Civ will go from that point. Gravelotte (1870), Bull Run (1861 or 1862) and Yukhnov (1941) are simply not important enough at the Grand Scale of a 4x game. You want to play them, go to a tactical computer game or start painting up soldiers in blue and gray uniforms (Bull Run) of Czech 38t tanks (for Yukhnov) in German markings.

And let's find a combat system for a 4x computer game that actually focuses on the time and distance scale of the 4x game, which is far above that of the individual unit of archers or riflemen . . .

 

[BuchiTaton]

I would like to have one tile personalized armies formed of a few units in a chosen formation, some examples:
* 3 frontal (Axemen), 2 lateral (Chariots), 1 behind (Bowmen)
* 2 frontal (Siegetowers), 2 lateral (Pikemen), 2 behind (Trebuchets)
* 4 frontal (Militias), 2 behind (Fieldguns)
This should come with the option to save our favorite layouts. This way you can design "shock armies", "siege armies", "raider armies", "desert armies", "jungle armies", "elite armies", "auxilar armies", etc.

The variables of units, formations and their positions (plus experience, equipment, etc.) would give such army a set of base value that would save us a lot of annoying micro, time and map saturation, since most pre-WW1 could be represented by 1vs1 to 3vs3 armies(occupied tiles) pitched battles instead of anachronistic long front lines.

 

[Civinator]

And let's find a combat system for a 4x computer game that actually focuses on the time and distance scale of the 4x game, which is far above that of the individual unit of archers or riflemen . . .

A good solution would be a separate battle screen, as it is used in Age of Wonders 2 SM. This would also solve the problem between the speed and the range of ships (and other units) as it was done in Imperialism by SSI. Against the argument, that such a setting would last a game for ever, an additional "short combat modus" as it is done in Age of Wonders 2 SM, where the result of a combat can be alternatively be calculated by simply pressing a button, can be introduced.

 

[DeckerdJames]

What you don't hear about, and should, is the Time Space with 1UPT: that a single battle is not only spread out a hundred or more times larger in space on the game terrain than it should be (archers that can fire from one side of a large city to the other?) but also in Time. When in the Ancient Era it takes 80 - 160 years (2 - 4 turns) to kill a single Barbarian Spearman, What in God's Name is the 'battle' supposed to represent? My answer would be multiple campaigns spanning generations, which rather makes the graphic of 2 - 3 'units' fighting each other all that time look Silly.

In a Grand Strategy game like Civ in which the shortest possible 'turn' is One Year, no field battle should take more than one turn. Period.
Even that is a stretch, but it is acceptable to get a 6000+ year-long game compressed into playing time that takes less than a week or more of 'real time': no game combining Tactical and Grand Strategic decisions can be other than a bunch of compromises of all kinds, but the game needs to do better than the current Tactical Battle = Decades, War = Centuries that Civ gives us now.

What a 1 Year Minimum Turn indicates is that only the 'decisive' battles are being shown: the Borodino, Gaugamela, Sedan (1870 and 1940) that fundamentally decide where the war and the armies and possibly your Civ will go from that point. Gravelotte (1870), Bull Run (1861 or 1862) and Yukhnov (1941) are simply not important enough at the Grand Scale of a 4x game. You want to play them, go to a tactical computer game or start painting up soldiers in blue and gray uniforms (Bull Run) of Czech 38t tanks (for Yukhnov) in German markings.

And let's find a combat system for a 4x computer game that actually focuses on the time and distance scale of the 4x game, which is far above that of the individual unit of archers or riflemen . . .

Does it take a hundred years to walk around an ancient city? What about a trader? How long does it take for a round trip between cities? The space-time relationship is probably not sensible for movement either, even if they move one hex at the time. I have to look at the Civilization series as a computerized board game. I want it to make sense in it's own universe, but it doesn't have to represent reality on a 1 to 1 scale. The unit layer is an idea. It doesn't match the geographical scale. Moving a worker around doesn't have to mean he actually took 100 years to move. The worker is an abstraction. All units, in Civ, are abstract ideas. It should make sense though, but with Civ 6 there are scales for units on the world map vs cities on the world map. We look at the distances differently depending on which of the two we are considering. I don't think of that is a flaw. It is just the kind of game Civilization is.

 

[DeckerdJames]

They could consider not making cities impassible. Any unit can occupy a city hex. If you are at war, the goal is to occupy the city hex for a period of turns. If you can do it, the city can begin to go through stages of conversion. Building walls might just lengthen the number of turns it takes to begin that conversion process.

I am not saying this is the right idea, it just one idea to alleviate some unit congestion. This makes the production of units more important, though, and that may be the opposite of what many civ players want. If the game was played this way, starting with a few warriors and slingers on turn 1 would probably be the way to go.

 

[UWHabs]

Does it take a hundred years to walk around an ancient city? What about a trader? How long does it take for a round trip between cities? The space-time relationship is probably not sensible for movement either, even if they move one hex at the time. I have to look at the Civilization series as a computerized board game. I want it to make sense in it's own universe, but it doesn't have to represent reality on a 1 to 1 scale. The unit layer is an idea. It doesn't match the geographical scale. Moving a worker around doesn't have to mean he actually took 100 years to move. The worker is an abstraction. All units, in Civ, are abstract ideas. It should make sense though, but with Civ 6 there are scales for units on the world map vs cities on the world map. We look at the distances differently depending on which of the two we are considering. I don't think of that is a flaw. It is just the kind of game Civilization is.

Yeah, as much as I really want things to be reasonable, there's so much of how everything works that makes absolutely no sense on whatever scale Civ is on. I mean, they can't even figure out what scale civ is on. Even on the largest maps, all of England can only support one or maybe 2 cities. So nothing really has anything reasonable for a scale, why I really don't think of a single battle between units as truly lasting for 500 years.

Now, granted, I do feel that does gloss over a lot because of that. In actual history, even cities in the middle of your empire still often needed castles and fortresses and city walls. Mostly because as much as you may have owned a land, there were always skirmishes and rebellions, and you always needed to protect things. And it's still a real annoyance that I can have a modern city that's size 15 (whatever population that is), but doesn't have a library or University. Or maybe is just starting to build a university but on the current timeline sounds like it will take about 100 years to finish. That is one piece that still irks me a little bit about the districts, I do still sort of feel that stuff like libraries and markets should be on the same level and granaries and monuments, but then districts would be more about upper level specialization.

 

[DeckerdJames]

Another way to go with the series, is to make the world much, much larger, meaning many, many more hexes. Consider the 7-hex diameter to be the "city" and increase the maximum territory it can control and work to a 21-hex diameter. Then multiply the movement rates of the units by 3 and make their line of site to 12. Drop the idea of single tile "forests". That would be a patch. Instead, a "forest" would be many hexes deep. Perhaps 20 or 30.

Expand the types of terrain. "Woods", for example, could come in several different densities. The idea is to make wheeled vehicles more limited to roads and less rough terrain. Sparse wood might be passable by tracked vehicles, but the thickest woods might only be passable by infantry and horse calvary and the like.

Along these ideas. Drop the idea of walls. Instead, fortifications are produced like buildings are in civ 6, and they add defensive value to the hex where it is built. They benefit the occupying unit.

If they go 1UPT, then go much bigger.

This might solve the congestion problem, but the micro management has to be solved with automation. The player needs to be able to sit high up the command chain and give orders to large groups. Micro individual units when preferred. Like tweaking their anchor spots once they reach a region.

4x is hard to design, I imagine.

Complexity issues can be solved by giving the player the ability to delegate to the AI. Want to move a giant army across the world? Assign it to the AI. Want to occupy a city hex? Tell the AI. Let the AI micro while the player macros.

I have a bad habit of perpetual editing. I will conclude this post.

 

[KayAU]

I agree despite disliking ro being different about almost every design decision for civ6, because Humankind was just such a miserable experience

I'm surprised so many feel so much worse about Humankind than I do, as I expected to be among the more critical voices. I based this assumption on Endless Legend, which I had a feeling everyone loved, and which I wanted to love, but couldn't. My biggest issue with it, unfortunately, is the region system, which is so deeply embedded into the game design, it is certainly staying. The regions feel like giant tiles, and it just takes a lot of joy out of the map for me, both when exploring and building. I also have mixed feelings about how cultures are implemented. That said, I do find a lot more enjoyment here than I did with Endless Legend, and it seems like Amplitude are committed to listening to feedback and continuing to improve the game. The recent update was certainly a step or two in the right direction. I don't see myself putting in 2K+ hours like I have in Civ games, but perhaps it could be a 300-600 hour game for me, like Stellaris and Fallen Enchantress.

As for Civ 6, I have recently returned to play a few games after a more than 1 year hiatus, and probably more like 2-3 years since playing frequently. Coming back, I am reminded of why I felt disappointed with it, as there was so much potential, but the devs stubbornly refused to significantly refine or rework any systems beyond their first implementation. The World Congress is probably the prime example for me. It could have been the thing which lifted the late game and gave you something new to focus on besides tedious micro-management, but as it stands, I would have turned it off completely if I could. On the other hand though, I am also reminded of how enjoyable the Civ 6 early game can be, and how good many of the civ designs are, such as Maya, Inca, Norway, Mali and Kongo.

Finally, on the topic of 1 UPT and combat, my preference remains an army system. I've commented on this many times, so I'll just quote myself:

My preferred system would be armies - several units grouped together which move across the map as one. The number of units which could be grouped like this would be limited, and the limit could be modified by a number of things, such as tech/civic development, leader/civ abilities, or the abilities of the army's leader. Many games have systems like this, so I will not go into great detail, but some good examples would be Fallen Enchantress, Age of Wonders 3, Stellaris, and Conquest of the New World. I think this solution is a good middle ground which resolves both unit congestion and stacks of doom, while at the same time making the military aspect of the game a great deal deeper and more enjoyable.

Battles could be resolved either through a tactical combat layer, or a good autoresolve function which takes things like unit synergies into account. I personally enjoy tactical combat, although it may shift the focus of the game too far away from empire management. I think Fallen Enchantress and AoW3 are good examples of somewhat complex tactical battle systems, while oldie CotNW had a simple but enjoyable one.

• Vastly improved tactical possibilities
• Possibility of making the AI an actual military threat
• Battles can resolve in a single turn
• Solves movement issues caused by 1 UPT
• Avoids stacks or carpets of doom

 

[dagriggstar]

The time thing can be designed around

Caravans could just appear on the map, constantly moving even during your turn. The unit art is just so you are aware of where the trade route is. To pillage, a unit merely needs to be anywhere along the trade route (rather than a specific tile on the route the unit happens to be on).
We don't need worker units. Tile improvements like farms could be passively gained over time as a citizen works a tile. Or they could even be built by the city itself.

I would like for wars to be shorter and more decisive (Without resulting in total defeat/victory situations).

 

[DeckerdJames]

The time thing can be designed around

Caravans could just appear on the map, constantly moving even during your turn. The unit art is just so you are aware of where the trade route is. To pillage, a unit merely needs to be anywhere along the trade route (rather than a specific tile on the route the unit happens to be on).
We don't need worker units. Tile improvements like farms could be passively gained over time as a citizen works a tile. Or they could even be built by the city itself.

I would like for wars to be shorter and more decisive (Without resulting in total defeat/victory situations).

I have had similar thoughts of whether the worker unit is needed. As you mention, a trader unit could be done away with too. However, outside realism, both of the units add gameplay value. Figuring out how and when to improve tiles is a huge part of the early game. Taking them away might make the game even more boring. Instead of doing away with the trader, they should make a better game out of trader units or the trading game. That's my opinion.

I have thought about the idea of a blueprinting system where you can blueprint a city layout and dictate in what order things will be built, far into the future. Instead of a queue, like a list, the queue is directly on the grid, with unbuilt things displayed semi-transparently. Perhaps there is a friendly data view of it that displays it like an easy to understand ordered list of builds, but the queue is built right on the grid. Of course, to see the blueprint is a UI option that can be cycled on or off. Then, you could lose the worker unit. However, the vulnerability of the worker unit adds gameplay value.

 

[DeckerdJames]

Road making isn't end game, gameplay, but I think Civ 7 could make it a little bit more prominent than civ 6. I think previous editions of the Civ series did more for the road making game. Honestly, I think laying down the transportation infrastructure can be fun, but I don't think it should be done with direct worker management. I think you should be able to blueprint the roads and a road crew unit will work on it automatically and complete sections, faster or slower, depending on the terrain, with the option to have it return to safe harbor, like in a defended city. Perhaps the road crew stops at each intersection and the player chooses the direction it goes, then it goes until the next intersection. If it is moving over road that is already built, it moves at its normal movement rate rather than its construction movement rate. It basically only moves on roads (or rails).

Civ 6 does get better when military engineers become available, but they are very micro intensive.

 

[aieeegrunt]

I have had similar thoughts of whether the worker unit is needed. As you mention, a trader unit could be done away with too. However, outside realism, both of the units add gameplay value. Figuring out how and when to improve tiles is a huge part of the early game. Taking them away might make the game even more boring. Instead of doing away with the trader, they should make a better game out of trader units or the trading game. That's my opinion.

I have thought about the idea of a blueprinting system where you can blueprint a city layout and dictate in what order things will be built, far into the future. Instead of a queue, like a list, the queue is directly on the grid, with unbuilt things displayed semi-transparently. Perhaps there is a friendly data view of it that displays it like an easy to understand ordered list of builds, but the queue is built right on the grid. Of course, to see the blueprint is a UI option that can be cycled on or off. Then, you could lose the worker unit. However, the vulnerability of the worker unit adds gameplay value.

You could leave the workers in like we have now, but also make tile inprovements and district buyable

This would help a LOT with late game tedium

 

[HorseshoeHermit]

Anything in a game, that is delegated to an A.I. to pick the details of, should not have those details actually in the game at all. Move up a step of abstraction to where you're just telling the game to do the thing the A.I. would be accomplishing. Having things in the game are either in player decisions or not in player decisions. If there are environmental factors then calculate them; if there are contested factors then resolve the contest.

The part of making workers that doesn't make sense to me is this part about eternal employment by the state. The parts that do make sense are where I invest "city development" and wealth into gaining the labour force, and where this force can be raided by enemies and barbarians. If they have charges, then yeah, I should just pick a tile improvement as a city-build and this can be pillaged; if they don't have charges, then what is really the significance of having so many that you outpace worked tiles, or that they can be captured and then they can live eternally in ANOTHER side's camp.

The loss, though, of an undefended worker is too totalizing. There is basically no game where I'm willing to steal a worker from another player because they will vendetta me into *both* of our oblivions just to be a rattlesnake about it. Similarly for pillaging improvements. This goes to what Boris was griping about for scale. If we did move to strategic layer for war only, then the timeline for conflict is messed up, and the only wars that could exist would be wars of genocide. I'd like if there was a system that let , for example, Aztec garland wars actually exist. If there was a system that created a tension of not being sure you could send trade, or build improvements, because you can't guard the lands from a pushy neighbour - instead of a game where people don't care because they know they will just war you into irrelevance (i.e. give up) instantly for the audacity.

 

[DeckerdJames]

How about getting rid of the worker unit and the idea of putting down improvements through a unit.

Instead, when you earn a city population point and get a new citizen, whatever tile the citizen is assigned to will begin to improve automatically. It improves over a number of turns until it reaches the maximum improved state per the technologies that have been learned and apply to that type of tile. If the citizen is reassigned to a different tile, the new tile begins to improve, but the old tile falls into a state of disuse over a period of turns. If the citizen is assigned to a tile that is in a state of disuse, it takes a number of turns to bring the tile back into an optimal state. Less turns than improving it from scratch though.

A citizen can also be made a specialized citizen. In that case, the citizen only works a particular type of tile but the yields from that tile are greater than if worked by an unspecialized citizen. The citizen also improves the tiles, that the citizen is specialized in, faster and brings a tile from a state of disuse into optimal condition faster than an unspecialized citizen. It also repairs pillaged tiles faster if they match the citizen's specialization. All of this happens automatically by assigning citizens to tiles, which can be done manually, or can be managed by the city's automatic management tool.

Eliminating the worker unit does remove the gameplay value of having to protect the unit and otherwise considering its safety and the timing of moving it around. On the positive side, it eliminates a lot of unit-based micromanagement when the empire is larger, and the AI can focus on damaging improvements during a war which is probably more important than chasing a worker around.

In the late game, the player can place facilities that can harvest yields from a region around it. So, for example, you would place a lumber mill in the middle of a forest, and it would take advantage of a region of the forest. All forest tiles within range of the lumbermill would automatically yield as though they are improved. Citizens would be assigned to the region through the lumber mill and their effective harvest rate would be greatly increased because, practically speaking, they harvest at a faster speed because they are in a mill and can harvest a larger area with same number of citizens. The rate of improved effectiveness and the range of the lumbermill would change based on the acquisition of relevant technologies. Citizens working in special facilities, like a lumbermill, are automatically considered specialists.

Therefore, in the late game, the idea is to gravitate away from citizens working all over the land, but instead are usually found in the various types of facilities. The AI, in the late game would have far fewer important targets to consider, just as the player would. For example, disabling the lumber mill with a bomb, could kill several citizens, and the citizens that survive could begin repairs automatically or, the citizens could work the forest tiles directly, but the overall yields would be greatly diminished without the lumbermill. Also, losing control over a forest region and the production value of the facilities there could be significant to the whole empire's efforts. That makes a forest region more strategically important. So perhaps it might be important to be able to occupy a facility like a lumbermill and capture control a forest region without having to capture the urban area.

This might all work a bit better, if the "city center" was an area instead of a single tile. Then you can blueprint out the urban area from the founding of the city even placing buildings and/or districts that are not yet available but will build as some later time when they become available. Then, a much larger area around the urban area, perhaps 21-hex in diameter or more, would be available for the placement of appropriate facilities, like lumber mills. Regions of terrain like forests, deserts, lakes, and other types could be enlarged. Military facilities could or would also be placed in the landscape around the urban area.

The whole idea is to continue the trend of having the player make fewer decisions but decisions which are very significant. In the early stage of a city's development, which tiles are being worked, and whether to change where they are working would be significant. In the later stage of a city's development, where to build facilities, and how many citizens to assign to them would be significant.

I don't know what would be built in the urban area, but things like universities and other things found within cities would probably be there. However, an observatory would probably be placed away from the city in mountains or some appropriate place. Both would be of scientific value, but instead of having a generic science point, you might progress faster from an observatory in the field of astronomy than you would in a facility intended to study the ocean, which might be built on the coast.

What do you think of these ideas? IMO, it might be an interesting game, but I don't know if the map would be too overwhelming large or if it is the right thing for Civilization.