Why should I raze cities?

[Vulcans]

Why do people always advocate razing cities so much?

So many people say their mistake was that they kept too many cities. But I don’t understand the problem.

I generally build infrastructure in new cities, markets, banks etc. and they always end out being profitable cities, if not instantly then within a very short period of time, and pay off more over the entire game.

These cities also provide resources. Even when you have all the resources you need, extra ones can be sold to AI for GPT.


Costs:

1. Upkeep: NOC+distance -50% with courthouse. So a 20gpt maintenance city instantly becomes 10gpt. Captured cities normally if using slavery you can instantly whip courthouse(organized trait helps) and other essential buildings. Actually i think there is also some cap on the upkeep for the number of cities, so the upkeep can't get that big.
   
2. In the end game state property becomes available, ripping the upkeep to a very minimum level.
   

Benefits offsetting upkeep of extra city:

1. Trade(6+ gpt): Cities have trade routs, and resource income so you can easily get over 6gpt simply from the trade routs (more later with harbor)
   
2. Unit support(1-4 Gpt): Cities also support population which support military. Eg a size 12 city can support 3 military units, and in the central region of my territory I only post 1mp. Meaning the city saves an extra 2gpt on military upkeep from extra population, so it supports your army.
   
3. Resources (1-10 Gpt): some resources give gold directly. Also picking up extra resources (even when you have them all) can be beneficial as you can sell extra to friendly neighbors. Also future resources could later appear in that area.
   
4. Religion(1-7 gpt): Chuck in a few religions if you have shrines, e.g. 3gpt for 3 religions. Naturally you need productive areas of territory producing missionaries, this becomes more useful later when you have captured a few shrines. Spreading religion becomes harder when there are more religions, so there is a limit to how far you can take this practically.

Income

1. Production: all those extra cities can be used to pump out military units, meaning that when you finally get MT etc you can suddenly turn 30 cities onto Calvary production (in the late game this can be multiplied with factories etc.).
   
2. City income (easily 30+ gpy/BPT): this is an increasing revenue, the longer you have a extra city it’ll earn more, with cottaged land, growing city, printing press.
   
3. infrastructure: a any decent location city can produce some revenue, this is increased over time the longer you have it, as you build libraries, banks, monastrys, markets, grocers, uni, observatory etc. just line up the list of buildings and watch the revenues multiply over time.

Side benefits:

1. Culture: you stop opponents from settling in the area, it can be very pesky on a crowded map, when you wipe out an AI, only to have one of your AI “friends” settling in the gaps between the cities you kept. having more cities also helps eat away at previous civs left over culture quicker.
   
2. Transport: only you can make use of those railway lines, you can get around, but the opponent can’t. also you can chose who can pass the land, with who you allow open borders, so blocking off some AI form some areas.
   
3. Fog busting: no barbs to worry about, so minimum MP is needed in the area.
   
4. Situational bonuses: eg merchantalism, statue of liberty, and representation gives a bonus of 12bpt per city. lots of other wonders give bonuses for any city just for being there. This alone can offset upkeep costs.

Calculation:

As far as I can see, the benefits from trade routs, unit support, resources and religion offser the courthouse reduced upkeep. Meaning that upkeep is effectively not an issue.

Upkeep: Eg 30gpt upkeep (you must have a HUGE empire to get 30gpt upkeep with State Property, actually i think there is a cap on the maximum)
-50% courthouse=15gpt upkeep.
-6gpt trade, -2gpt religion, -2gpt resources, -2gpt free unit support=3gpt effective upkeep

Actually 20gpt upkeep is more reasonalbe number, which would result in a 2gpt profit after trade etc is considered.

(note, these numbers are just an example, and the numbers would vary depending on the situation)

In my calculations, the side benefits offset the upkeep costs.

So I tested it razing anyway:

I’ve previously supported absolutely huge empires without razing a single city, naturally as the empire grows more Tecs become available to enable larger economies, printing press, democracy, state property etc, the whole time making cities more efficient, and allowing for even further expansion without ever having any problems (ok, this was only at empror, so maybe diety is a different story). I normally have all resources available, allowing me to rapidly grow my newly captured cities to size 10+, making them useful in a short period of time after large scale assimilation, and back to normal after a few turns with courthouses, markets and banks. only very short periods of slider drop, but still out-teching AI, money from selling technology, plunder etc keeps my slider at maximum even during periods of large scale assimilation. no real economic problems, and a very strong economy with all the assimilated cities.

After listening to everyone always complaining about having too many cities I decided to try it and raise about 50% of the cities in captured opponent’s territory, still leaving some cities for my domination territory. But then my other AI “friends” instantly came in and settled in the gaps. This only created more troubles as I later had to come back and clean out the pesky AI that wanted to settle in the gaps. What a pain, and only for supposedly saving a little upkeep in the short period before the city becomes profitable.

The hassle of going back to clean out the gaps delayed my domination victory date(and re-declaring affected reputations), and I wished I had never listened to the forum telling me to raze cities.


So what’s the point of razing cities and leaving gaps in my territory?

Sure, sometimes in very rapid expansion it can temporarily slow down your research rate. But the research is the total beakers produced, a empire twice the size at 40% research rate produces more beakers then a small empire at 70% rate.
And this is only a temporary problem, in the meantime you can trade to keep even with research. this economical challenging time is only very temporary, especially if you're building courthouses, banks etc, then you have twice the size empire, and then can out-research the AI when the infrastructure is up (at least on empror i find it easy to out-research the AI).

And it’s all an investment in the late game, when you get villages/towns then all this area will be producing a lot. And then replicable parts makes it all more worthwhile. and all those cities are also making production, so when you want a military push you can turn a whole pile of cities into military production. all those early extra cities you didn't raze become massive powerhouses in the industrial era, and make a HUGE difference.

Naturally cities in totally useless locations should be razed, but I think about 90% of cities captured are useful.

There are only two points in the game when I normally wouldn’t be interested in new cities, before COL, and then in the final few turns of conquest where you just want to kill fast for the final moves(and even in those situations, having a strong empire means a couple of gpt for those cities isn’t a big price, especially if it saves you from hassling about the conquest victim sending out settlers).


So please tell me why I should raze most captured cities!


.

 

[cabert]

It's not a raze or not raze question, really.
It's a question of when and what.

When :
- early in the game, when you don't know currency or CoL, you're better off with only 3 cities or really excellent ones above this.
- in the mid game, you can afford some more cities, but you don't want to win the liberalism race for one too many city, do you? To be able to whip a courthouse as soon as it comes out of revolt, a city needs to be size 9 (it will starve to size 8 the moment it comes out of revolt). Everything smaller than that can be a burden to your expansion.
- late in the game, money flows easily, you can afford to keep every inch of land. If it isn't profitable yet, you can make it so it becomes profitable later + it gives you more power, a better score and less diplomatic demerits.
- in the final turns, you can raze a good, large city to settle your own and win by domination earlier (no need to wait for the revolt to end).

What:
- sometimes, the AI just builds cities you don't understand the purpose of. No food, no good resources, one tile from the coast... Why would you keep those crappy cities?
- sometimes, the AI just builds cities blocking off good resources. The city may be good but it prevents working a fish or a gold mine. Better raze this one, build it where it gives the best options.
- sometimes, the cities come with wonders or are holy cities. Usually you don't want to raze those.
- sometimes, you need a foothold somewhere. Even a crappy city can be this foothold.

 

[rabidveggie]

Mainly raze early in the game before col or crappy city position. Online you raze so that a city can't be recaptured by a human. Sometimes I'll even raze there capital just to stick it to them.

 

[Solo4114]

It's pretty much been covered here. The key question, to me, is WHEN you are fighting. Early game, you cannot afford to keep all the cities (we're talking axe-rush era and before). When you hit the Renaissance or possibly during the classical era, you should have Code of Laws and Currency researched. Once you have these, your core cities (usually your capital and several other cities) can provide enough cash to support the newly captured cities.

One caveat to this is that, if you've founded an early religion (I'll often go for Polytheism rather than Buddhism, mostly because the AI seems to go for buddhism and a spiritual AI can grab that faster than I can), then you can finance more cities once you pop a Great Prophet and found the Holy City's building. Spreading the religion further will generate even more income, allowing for more cities to be built. This can help offset the costs to holding more cities before you have courthouses/income-generating structures.

As for when to raze cities, early in the game especially. The AI is often quite stupid in where it recommends placement for cities and where it places them itself. One tile from the coast (= no lighthouse so coastal tiles will only give 1 food). Overlapping fat crosses (the AI does this ALL the time). I think the AI does the overlapping crosses often for cities it never intends to grow very much. These are usually the "intra-empire" cities, rather than border cities. And, of course, the "can't make anything" cities that were clearly founded either (a) to grab territory (rather than grow cultural borders into the area over time), or (b) to get a single resource. These are the cities that have six desert squares, three mountain squares, and some tundra, but happen to also have beaver on the tundra or oil or incense in the desert.

Unless they have a wonder or are a holy city, these cities are often better burned to the ground to free up space that you eventually fill in yourself, either by settling more wisely than the AI did, or by expanding your cultural borders.

 

[seitz]

If you won't be able to defeat the entire civ, it's also better to raze cities than have them be taken over again by the culture of the enemy civ. Even if the culture doesn't take it over, all the good tiles may be taken up by the enemy culture.

 

[Kev]

One item to also consider when razing might be the diplomatic negative for doing so. It's very situational, but if I am planning on making the rival civ my vassal with whom I would eventually like to trade with again (and set them to researching what I tell them), then I might keep even some "iffy" cities to avoid the permanent negative.

Other than that, I agree with others as to the timing and purposes of keeping vs. razing.

 

[Gozert]

Benefits offsetting upkeep of extra city:

1. Trade(6+ gpt): Cities have trade routs, and resource income so you can easily get over 6gpt simply from the trade routs (more later with harbor)
   
2. Unit support(1-4 Gpt): Cities also support population which support military. Eg a size 12 city can support 3 military units, and in the central region of my territory I only post 1mp. Meaning the city saves an extra 2gpt on military upkeep from extra population, so it supports your army.
   
3. Resources (1-10 Gpt): some resources give gold directly. Also picking up extra resources (even when you have them all) can be beneficial as you can sell extra to friendly neighbors. Also future resources could later appear in that area.
   
4. Religion(0-4 gpt): Chuck in a few religions if you have shrines, e.g. 3gpt for 3 religions. Naturally you need productive areas of territory producing missionaries, this becomes more useful later when you have captured a few shrines. Spreading religion becomes harder when there are more religions, so there is a limit to how far you can take this practically.

Trade: This is not /turn, but /turn. Yes, you *can* convert that into , but then you will be sacrificing your research. You'll only get that kind of amounts of /turn at a very low research rates. The extra commerce one city brings isn't always good enough to warrant this drop. Unless, of course, you mainly depend on your specialists for research.

Unit Support: You also need troops to defend the new cities, especially since you are at war and vulnerable for counterattack. Also, you need at least a decent navy to prevent sneak attacks from the sea if you only put one defender in your core cities. However, a navy also has support costs...

Resources: Again, resources give you , not :wealth:. See Trade.

Religions: Yeah, if you HAVE the shrines it can offset a lot. I usually don't have more than one, perhaps two shrines unless I conquer a holy city. It's a matter of playstyle. Besides, you'll need to build missionaries, while you could build units/infrastructure. I'll agree that religion can give a lot of moolah through shrines and, especially, the Spinal Minaret (+4 /turn for a city with a temple and monastary of your state religion; +8 with Market, Grocer and Bank ). However, you do need to found the religions or capture holy cities first, and get Great Prophets for the shrines or hope one is already build in a holy city you take, and build a unique wonder. In the meantime, someone else can tech like crazy (not going after religious techs and generating Great Scientists instead of Great Prophets for academies or more value from lightbulbing) and build more than a handful of units to fuel the war effort instead of the temples, monastary and World Wonder.



Cities I keep need to be really good (i.e. lots of food, production and/or commerce, have a useful wonder, holy cities, strategic position, etc.). Yes, you can take a lot of decent cities, but I usually do not have the finance to keep merely decent cities without getting my research to rock bottom.

 

[Solo4114]

So, again, it all boils down to a balancing act and the details of your game. Even a hard-and-fast, this-way-is-always-ideal rule is likely to have many many exceptions to it.

Basically, the way I see it is this: if I can afford to keep 'em and it's not a pain in the butt to do so, I'll keep 'em. Why not, after all? If they're going to cost me more than I'm willing to spend in that specific game, or they're going to be too difficult to defend, OR they have overlapping fat crosses and such, then forget it. They're toast.


There is, of course, another method -- namely culture bombing. You can always try out-culturing the enemy civ. That takes a LOT more time, though.

 

[johncross21]

The main reason I raze cities is because I want to eliminate or cripple a rival quickly but don't have time to assimilate their cities (it takes time to build the courthouse, shrines etc)

 

[CapnKill]

Would you ever use a great artist to instantly end the resistance in a newly captured city or is that generally considered missuse?

 

[johncross21]

that called a culture bomb - it works - others do it

however i never seem to have the artists when i need them

 

[Mango]

that called a culture bomb - it works - others do it

however i never seem to have the artists when i need them

It's the only thing a GA is good for most of the time. They can save you 10-11 turns of resistence late in the game and keep your cities from flipping back or starving to death.

 

[DNK]

When going for a quick conquest or domination victory = raze happy. Keep half on the latter, so I can immediately start making culture in them to expand for the land size requirement. Otherwise, there's no point.

In non-end-game cases, a lot of cities just stink. Like there will be one great city of mine and a great AI city with a useless little one shoved in-between, taking up tiles. Or just a useless city to begin with.

 

[AnitaGaribaldi]

Why do people always advocate razing cities so much?

So many people say their mistake was that they kept too many cities. But I don’t understand the problem.

.

What level does you play? It makes a lot of difference. I used to raze very few cities and hardly could beat Monarch...

 

[Mango]

When going for a quick conquest or domination victory = raze happy. Keep half on the latter, so I can immediately start making culture in them to expand for the land size requirement. Otherwise, there's no point.

In non-end-game cases, a lot of cities just stink. Like there will be one great city of mine and a great AI city with a useless little one shoved in-between, taking up tiles. Or just a useless city to begin with.

The AI does tend to stick cities on marginal land if a resource can be had. By the time you're running through their cities you probably already have 3-4 sources of iron or coal, so there's no need to keep that fringe city that can only work a couple plains and water tiles. The AI will also stick cities on ice if they can grab a single fish, crab or whale. It's useful if you don't already have the resource, but for the most part I pass on those cities.

 

[sylvanllewelyn]

I'll tell you why you raze cities.

When you capture a city, you receive no defensive bonuses for many turns in anarchy, and many more once you leave anarchy. You need to split up your forces to guard it, which allows your enemy to take on and defeat either your guard or your attack force, or heck: one at a time. And once they capture cities back, they get all the culture and defensive bonuses again. Your CR3 maceman don't fare very well INSIDE a foreign city.

Simply put, you need to be very superior in military numbers to both keep a city and a force to keep up an advance. And you don't have that once you go beyond monarch (assuming you want to, by your tone, AnitaGaribaldi).

What happens is, you're outnumbered and outteched, just look at the powergraph. There are many war strategies, but it revolves around either
- pillaging them to death to "slow down" their research and production. Basically empty it. OR
- group all your armies in one area, and win by local superiority (often involving catapult cheese). Then raze their cities one by one.

The key concept: You are fighting for land, not cities!! When cultural borders touch, your AI enemies hemm you in and prevent you from expanding. As I-forgot-who said, "it's about keeping your neighbours in the dark ages". You can always found cities on the empty(ied ) land with settlers afterwards, and make sure those cities are culturally yours. It'll develop a lot faster than you think because the improvements are still all there. Once you have more land than your opponants, you can start thinking about victory conditions.

That, or you can try one of those crazy conquest-style strategies, but if you are, you won't be asking these basic questions.

 

[Vulcans]

Lots of good comments from everyone!
Thanks for the interesting discussion!

I agree with most of the points above, raze the cities in useless positions. Rush for COL just in time for mass assimilation of your nearest neighbor. And then keep what is useful. The sooner you have COL the sooner mass assimilation can start.
Especially with organized + slavery it becomes much easier to get courthouses quickly, making more cities attractive.

It's pretty much been covered here. The key question, to me, is WHEN you are fighting. Early game, you cannot afford to keep all the cities (we're talking axe-rush era and before). When you hit the Renaissance or possibly during the classical era, you should have Code of Laws and Currency researched. Once you have these, your core cities (usually your capital and several other cities) can provide enough cash to support the newly captured cities.

I tend to find it good to keep everything, but only assimilate one empire at a time, take several cities, then have a short break in the war-mongering while building courthouses, granary, barracks (and libraries in sci cities) then when reaching the next military Tec switch the empire to building the latest units, have another war, assimilate everything, and then courthouses in the new cities, some of the oldest cities get markets etc, and repeat for each war. So I keep everything, but keep a steady pace of progress, compared to a hard-core warmonger who wants a record breaking conquest date, who doesn’t want small breaks between wars, so probably has to raze more.

If you won't be able to defeat the entire civ, it's also better to raze cities than have them be taken over again by the culture of the enemy civ. Even if the culture doesn't take it over, all the good tiles may be taken up by the enemy culture.

I normally try to wipe out a civ in one sweep, at least to a stage where they are harmless. Often only leaving one city if they offer tec for a peace treaty, then finish it off later when I’ve got nothing to do between serious wars. Yes, their culture can be a pain for those 10 turns(sometimes a spare artist can solve culture problems).

One item to also consider when razing might be the diplomatic negative for doing so. It's very situational, but if I am planning on making the rival civ my vassal with whom I would eventually like to trade with again (and set them to researching what I tell them), then I might keep even some "iffy" cities to avoid the permanent negative.

Yes, diplomatic negative is one thing to consider, although I normally couldn’t care less of what they think of me after I am finished with them, as they are too small to keep up with tec pace, so are useless for trading, and are wimpy, so useless for bribing into war. Normally I would like to finish them off as soon as possible, so my people don’t long for their mother land (I play vanilla, so there is no vassal etc).

Trade: This is not /turn, but /turn. Yes, you *can* convert that into , but then you will be sacrificing your research. You'll only get that kind of amounts of /turn at a very low research rates. The extra commerce one city brings isn't always good enough to warrant this drop. Unless, of course, you mainly depend on your specialists for research.

Good point

In the end beakers and gold are interchangable, i often fluctuate my slider, getting money when needed, or fast research when needed. actually i think commerce is better then gold, as it gets multiplied by buildings.

Unit Support: You also need troops to defend the new cities, especially since you are at war and vulnerable for counterattack. Also, you need at least a decent navy to prevent sneak attacks from the sea if you only put one defender in your core cities. However, a navy also has support costs...

I normally know where the enemy will attack from, have decent defense in boarder cities, while cities 3+ steps from the boarder don’t need more then 1MP, they are far enough from the boarder to prevent sneak attacks, and if someone moves in I will see them coming, and can move units into place. Also in the end game(if not conquest with cavalry) when railroad is in existence I normally drop most military police, as I have a mobile force that can defend the entire empire.

Shrines and resources become available as you expand, controlling a large portion of the world gives you all the resources you need. And the odd resource you don’t have can normally be tradef for with the resources you have abundance. (although I never give strategic resources to anyone that could be a military threat)

It's the only thing a GA is good for most of the time. They can save you 10-11 turns of resistence late in the game and keep your cities from flipping back or starving to death.

This can be worth a lot! i used it in my last game, where there was a Buddhist shrine, 43 Buddhist cities, so coming out of anarchy 11 turns earlier gave me 473 gold, this is on top of the other benefits, like saving the pop from starving, revolting, earlier production in the city etc. yep, I think that’s the best use of a great artist in the late game, much better then light bulb the useless divine right.

What level does you play? It makes a lot of difference. I used to raze very few cities and hardly could beat Monarch...

I play Emperor, I generally keep most cities, and win the liberalism race etc. I think the reason I can win the liberalism race is because I get a strong economy from all the cities I took and kept. The way I figure it is that if the AI gets an X bonus, then I need to grow my empire to more then X times his size to be on par with research/production rates. Therefore I am assimilating as fast as possible, trading to keep on par with the AI during the medieval ages, then having 2*X his size I can pull ahead and out-research him(and his group of trading buddies).

I'll tell you why you raze cities.

When you capture a city, you receive no defensive bonuses for many turns in anarchy, and many more once you leave anarchy. You need to split up your forces to guard it, which allows your enemy to take on and defeat either your guard or your attack force, or heck: one at a time. And once they capture cities back, they get all the culture and defensive bonuses again. Your CR3 maceman don't fare very well INSIDE a foreign city.
.

Yes, indeed, I normally keep a line of good defense units on the front line of cities, in the interior area I sometimes don’t even bother with MP, as the AI can’t get there(I have all resources so happiness isn’t an issue), and if they try I have mobility to catch them. Actually I like it when they attempt to get to send a token force to one of my undefended cities, only to have it taken out by my mobile forces on my roads/railroads.

Although I normally don’t let the AI even try to get in, I bring the fight right to their doorstep, not letting their units past. Maybe one boat might land somewhere, rush a unit, move a few across the road/railroad network from adjacent cities and fight off the pitiful invasion force.

It’s also good if you can capture chicken Itza, giving 25% on newly captured cities.

What happens is, you're outnumbered and outteched, just look at the powergraph. There are many war strategies, but it revolves around either
- pillaging them to death to "slow down" their research and production. Basically empty it. OR
- group all your armies in one area, and win by local superiority (often involving catapult cheese). Then raze their cities one by one.

The AI is dumb in attacking, you can normally mop up the incoming forces by grouping yourt units and using your road network for mobility (fighting in your land also reduces WW) then on to attacking. You need to get some leverage, you need X times their land ASAP to compensate for their AI advantage.

The key concept: You are fighting for land, not cities!! When cultural borders touch, your AI enemies hemm you in and prevent you from expanding. As I-forgot-who said, "it's about keeping your neighbours in the dark ages".

The pain is if you raze cities, then there will be gaps, it takes a while for the 2nd, and 3rd cultural expansion (even if you whip theaters etc) so while you just wiped out AI1, then AI2 comes in between the gaps and settles. What a pain! Not worth it!

 

[Phrederick]

Costs:

1. Upkeep: NOC+distance -50% with courthouse. So a 20gpt maintenance city instantly becomes 10gpt. Captured cities normally if using slavery you can instantly whip courthouse(organized trait helps) and other essential buildings. Actually i think there is also some cap on the upkeep for the number of cities, so the upkeep can't get that big.
   
2. In the end game state property becomes available, ripping the upkeep to a very minimum level.
   

State property doesn't affect the situation at all, since by that time you should be able to absorb any built-up city you might capture. People talking about razing cities generally mean during attacks up through macemen at the latest. Even courthouses can't be taken for granted. If you're losing money at 0% research, you're going to have a hard time researching code of laws. I've been there before, and if you have a military advantage at that point, it's still worth your while to be warmongering, but your economy is going to have a hard time absorbing much.

Benefits offsetting upkeep of extra city:
Trade(6+ gpt): Cities have trade routs, and resource income so you can easily get over 6gpt simply from the trade routs (more later with harbor)

How do you figure that? Pre-currency, you'll have one trade route per turn. If you've captured a significant amount of territory already, odds are that city will have to have an internal trade route, so it's going to be 1 commerce/turn.

Unit support(1-4 Gpt): Cities also support population which support military. Eg a size 12 city can support 3 military units, and in the central region of my territory I only post 1mp. Meaning the city saves an extra 2gpt on military upkeep from extra population, so it supports your army.

I don't think people often raze size 12 cities. If I raze a city it's probably size 4-8, since the period when your economy can't absorb so many new cities is in the earliest stages, and it's rare to see size 10+ before hereditary rule.

Resources (1-10 Gpt): some resources give gold directly. Also picking up extra resources (even when you have them all) can be beneficial as you can sell extra to friendly neighbors. Also future resources could later appear in that area.

What about cities with no resources? What about pre-currency, when you can't trade them? What about when no one has more than 2 gpt to trade you? Obviously if a city has gold/gems/silver it'll have a better chance to pay for itself, but I'd put the trade value at 3-4 gpt max.

Religion(1-7 gpt): Chuck in a few religions if you have shrines, e.g. 3gpt for 3 religions. Naturally you need productive areas of territory producing missionaries, this becomes more useful later when you have captured a few shrines. Spreading religion becomes harder when there are more religions, so there is a limit to how far you can take this practically.

You don't always have a shrine, and it can be extremely expensive to spread them extensively. You can build almost one axe per missionary, and usually I'd rather have 7 axes to invade someone than 8 missionaries with 8 gpt. In other words, if you're in a situation where you're maximizing religion benefits, you're unlikely to be in a situation where your army is capturing cities faster than your economy can support, simply because military hammers are being diverted to religious hammers.

Income
Production: all those extra cities can be used to pump out military units, meaning that when you finally get MT etc you can suddenly turn 30 cities onto Calvary production (in the late game this can be multiplied with factories etc.).

Talking about lategame is irrelevant. If you raze an area, it's going to be resettled, and it's not hard to retake it when you can afford to add those cities into your empire.

City income (easily 30+ gpy/BPT): this is an increasing revenue, the longer you have a extra city it’ll earn more, with cottaged land, growing city, printing press.

When you've captured a significant amount of cities with axes, you'll be lucky to get 6 commerce from worked tiles per city. Obviously there's potential later on down the road, but in the short term it's going to take a while to develop commerce, especially if you're whipping in the needed production like courthouses/libraries/barracks/units/temples or what-have-you.

infrastructure: a any decent location city can produce some revenue, this is increased over time the longer you have it, as you build libraries, banks, monastrys, markets, grocers, uni, observatory etc. just line up the list of buildings and watch the revenues multiply over time.

Again, any city in a decent location will pay for itself over that timeframe. What matters is how much of an impact it will have before you have currency, code of laws, etc.

Side benefits:
Culture: you stop opponents from settling in the area, it can be very pesky on a crowded map, when you wipe out an AI, only to have one of your AI “friends” settling in the gaps between the cities you kept. having more cities also helps eat away at previous civs left over culture quicker.

It's not hard to control this. If you're such a warmonger that you've captured more cities than you can afford economically, you're probably going to be attacking your other neighbors sooner rather than later. Let them resettle and develop those cities, and by the time they've come on line and are productive again, you'll be able to afford to take them. Keep far away civs from settling there by avoiding Open Borders with them, and in all likelihood it'll be a close neighbor you'll want to invade anyway.

[*]Transport: only you can make use of those railway lines, you can get around, but the opponent can’t. also you can chose who can pass the land, with who you allow open borders, so blocking off some AI form some areas.

Railways are again irrelevant, since the timeframe where you might raze something is in the early game.

Fog busting: no barbs to worry about, so minimum MP is needed in the area.

An extra few warriors to fogbust is a lot cheaper than paying upkeep on a city. Plus, if you have enough of an army to capture so many cities that razing becomes an issue, you'll be able to rotate a few units away from the front to kill barbs and get promoted up to 10 exp.

[*]Situational bonuses: eg merchantalism, statue of liberty, and representation gives a bonus of 12bpt per city. lots of other wonders give bonuses for any city just for being there. This alone can offset upkeep costs.

Again, by that point it's pretty rare to raze anything. The timeframe for razing is before Civil Service at the latest, and more likely before Code of Laws.

Calculation:
As far as I can see, the benefits from trade routs, unit support, resources and religion offser the courthouse reduced upkeep. Meaning that upkeep is effectively not an issue.

I don't think your numbers are very accurate. In the scenarios where I can envision razing decent cities, here's what it would look like:

Upkeep: 15gpt. This includes the direct maintenance of that city plus the increase in maintenance in the rest of your city. Right before you invade a city, check your income. After you invade it, it'll decrease by up to 10 or more, and that's not counting the city maintenance of that city, since it's still in revolt. By that point I'd guess that the city in question will have a maintenance of up to 8, so it could cost up to 20gpt to add that city to your empire. I'm rounding down to 15 for the sake of argument, though.

Benefits: 2gpt trade, 2 gpt resources, 2 gpt free unit support. Now you're at a net loss of 9gpt. Plus, by the time this becomes an issue, you'll already have a handful of cities that are costing you that much, and it won't be long before your productive inner empire is overburdened supporting your new cities.

I tend to find it good to keep everything, but only assimilate one empire at a time, take several cities, then have a short break in the war-mongering while building courthouses, granary, barracks (and libraries in sci cities) then when reaching the next military Tec switch the empire to building the latest units, have another war, assimilate everything, and then courthouses in the new cities, some of the oldest cities get markets etc, and repeat for each war. So I keep everything, but keep a steady pace of progress, compared to a hard-core warmonger who wants a record breaking conquest date, who doesn’t want small breaks between wars, so probably has to raze more.

That's exactly the reason why razing has a place. If you're attacking at a breakneck place, you don't want cities that will take too long to become profitable. While you are digesting one empire, the others might be outteching you. The solution when that happens is not to wait to digest the first empire, but to only keep the best cities, and immediately move on to the next empire. It doesn't matter how crappy your economy is if everyone else has crappier economies due to you taking/razing their best cities.

I think overall the best approach is to keep some and raze some, where the balance will depend on the speed of your advance, and how good the cities are. If you're attacking slowly, only raze the cities with horribly crappy placement. If you're attacking very fast, only keep the cities with wonders, towns, or otherwise excellent resources. If you're between those extremes (like most games), strike a balance.