The cities are probably the most important individual piece of any of the Civ games and they are exceedingly well done. Here, I'm going to try to break them down into their constituent pieces and see if I can isolate the things they do well.
This is probably the most important part of the piece. The cities are the source for everything else in the game and so are very valuable to the player. This is emphasized by the length of the game as producers are naturally more valuable if given a lot of time to produce in.
Additionally, this is the source of the majority of the interesting decisions of the game. Deciding what to build next makes for a great decision point as it offers the players a selection of things that they want, but of which they can only have one at a time. This is a positive feeling as the player is going to get something good, but the decision of what good thing to get is still hard.
Additionally, they work well as the primary sources of other currencies such as science, faith and gold. Having all of these be produced simultaneously makes the city feel more productive and less one-dimensional.
It also makes the value of a city harder to quantify. If the value were directly calculable, the decision of whether or not to build a new city would become rote and uninteresting. Adding this complexity to the calculation lets players choose whether or not to build new cities by feel instead of arithmetic.
Cities grow given time and food. Each new citizen tends to have diminishing returns as the more productive tiles each get worked. This results in a natural cut-off point for city growth.
This growth makes the cities become naturally more productive, but that productivity boost tends to be offset by the increasing costs that come with progression down the tech tree. So, the question from a design standpoint is whether the game would be better without both systems given that they cancel each other out.
These dueling systems also have the side effect of nerfing building cities in the late game. If a city starts with one population, but all of the production costs are scaled for cities of populations of 7 or 10, the city feels useless.
Instead of having one citizen per tile, I can imagine a system where the city just gets flat bonuses from all of the tiles and improvements in its borders. You can even still keep the idea of focusing on production or science with flat bonuses. The growth system does allow for more discrete shifts in the city's production however. Being able to start working a tile with a mine or a farm makes a big difference to how much the city makes every turn.
Additionally, this growth is highly analogous to leveling in an RPG where the party and the enemies both scale in health and damage. It feels good for players to watch numbers go up, even if the challenges scale with them. Theoretically, this should also allow for the waveform difficulty curves of RPGs where you start an area underpowered and end it overpowered. The structure of Civ is too noisy to get that point across however.
Another thing that is very satisfying for players is to be able to use something that was previously lying fallow. The ability to use tiles that were earlier unused naturally feels good for players, especially when the tile has a special quality such as a bonus resource or a natural wonder. This is of course tempered by the fact that players are going to be much less excited about being able to finally use that one tundra tile on the edge of the city's borders.
Constructing new buildings in cities also feels great, in part because of the previous points. It feels good to increase the powers of each of your cities and the buildings do that in a very direct manner.
They also contribute significantly to the fiction of the city, which I will cover below.
Cities are a projection of power. It feels good to watch your empire grow. This is emphasized by the tangible borders of the later Civ games, but even in something like Civ2, it felt like your country grew with every new city in a way larger than just the tile now occupied.
The fiction of a city in Civ representing a real world city is a lot of what makes the above points work so well. I like the idea of putting a library or an amphitheater in a city because I like libraries and amphitheaters. When a city grows, it feels like more than just a number going up, it feels like that city moves from being a village to being a metropolis.
The above points all break down why a city is so valuable to a player and every one that you take from an opponent makes for a major shift in both your strengths. The defense and annexation of cities hence become a critical part of the game.
Additionally, the game is built to support this in how mildly it treats warfare. While players have the option to pillage a city and the surrounding area, it seems more sensible to leave it untouched so that you can use it once the city is captured. Returning a city to full productivity after its capture is a remarkably quick and painless process in Civ. It's much more direct than something like Stellaris for instance.
It's also worth noting that combat is an integral part of the game and without it, Civ would be much less compelling. I'm going to go deeper into this in my upcoming article about units.
None of the above points are actually significantly affected by the unstacking of a city done with Civ6, but I'm still a big fan of the feature. It makes the map much more meaningful. I think that it's very easy for a game like Civ to feel like a spreadsheet and integrating the map works directly to alleviate that feeling.
Additionally, it does help with the fiction and it makes the buildings feel more meaningful.
Cities have always been the lifeblood of the Civ series, and I think that I've completely broken down all of the parts of this crucial piece. I'm going to go over the rest of the major components of the series over the next few months as part of my work. Please tell me if there's anything I missed or should have covered here.