How long before a culturally overwhelmed city finally caves?

[phooey73]

He's got a lot of junk in the city actually. My impression was that its about 10 units. I checked last night and its actually only 5 land military, plus a couple galleys and some workers.

I guess I'll just look forward to more of this, as I crowd out Louis, Peter, and Saladin on the smaller landmasses where I've got 3 cities to their 1.

Cracking that 56% landmass really takes a lot of cities.

[johnmcd]

There is no rule on how many soldiers are needed. 10 is usually enough to keep it safe, but I've had 20 plus before and still had a % to flip.

[BigFree]

See if you can't buy it from him...

[Locutus]

Based on the code from the SDK, city flipping works as follows:

Algorithm
At the start of every turn, the game will look at every city in the game that's not occupied and of which the player with the strongest cultural presence (the 'attacker') is not the same as the physical owner (the 'defender'). For each of these cities, there is a (by default) 10% chance that the game will try to revolt it. If this happens and the city is Barbarian, the city flip immediately takes place. If the city is not Barbarian, it compares the culture strength of the attacker (see below) with the sum of the culture garrison (see below) of all units in the city (whether they belong to the defender or someone else). If the total culture garrison is higher than the culture strength of the attacker, the city will never revolt. If OTOH the cultural strength of the attacker is higher, then the bigger the difference between the total culture garrison and the culture strength, the bigger the chance the city will revolt (to be precise, the chance is culture garrison divided by culture strength). Of course, as mentioned, the second time a city revolts it flips.

The culture garrison (CG) is a property every unit has. For non-combat, air and naval units this is 0, for land combat units it depends on how powerful the unit is. Warrior and Quechua have the lowest value of at 3, Modern Armor and Mechanized Infantry have the highest at 16. As mentioned, the sum of these values of all units (regardless of their owner) in a city determines if a city can revolt and what the chance of it happening is. Note though that if the defender and attacker are at war with each other, the defender gets a 100% bonus on the culture garrison.

The culture strength that the attacker has in a disputed city is calculated using a rather complicated formula. It's influenced by the highest size a city ever had, the amount of plots bordering the city that the attacker owns, the average era (ancient, medieval, modern) of all surviving players, the culture value of the attacker has in the plot compared to the defender, and if the state religion(s) of both the attacker and the physical owner is/are present in the city. Obviously, the more of everything, the stronger the culture strength.

Strength Formula
The exact culture strength formula is the following:

City Strength = 1 + ((1 + Highest Size + (Owned Plots * Era Index)) * Culture Percent) / 100

where:

- Highest Size is the highest size the city ever had;
- Owned Plots is the number of plots neighbouring the city that the attacker owns;
- Era Index is the average value of the era indices of all surviving players, with 1 being the lowest (so ancient = 1, future = 7);
- Culture Percent is roughly how much culture the defender has compared to the attacker, as a percentage of the attacker's value (e.g. if the attacker has 60% culture, then defender has 40%, which is a third less (20/60 = 1/3), so Culture Percent becomes roughly 33).

Finally, the presence of the attacker's state religion in the city doubles the strength, the presence of the defender's state religion halves it.

Example 1
Let's say a city is size 8 (and was never larger than that), 4 plots surrounding it belong to the attacker, the average era is medieval (index 3) and the attacker has 67% culture, the defender 33% (so the attacker has 50% more culture). Now the culture strength can be calculated:

City Strength = 1 + ((1 + 8 + 4*3) * 50) / 100 = 1 + (21 * 50) / 100 = 1 + (1050 / 100) = 11.5

That means 2 medieval units at 6 CG each (e.g. Longbowmen) would keep this city safe from revolting, even if no state religion is present (or both civs have their state religion present). Let's say that at this point a culture bomb is dropped and gives the attacker 100% culture (effectively 100% more than the defender) and swallows up all neighbouring plots. All else remains unchanged. Now the situation is the following:

City Strength = 1 + ((1 + 8 + 8*3) * 100) / 100 = 1 + 33 * 100 / 100 = 34

At this point it would take 6 medieval units to defend it. If it only has two, there's a (34-2*6)/34 = 65% chance that IF the game tries to revolt the city, it will indeed revolt. That measn that on any given turn the chance that the city will revolt is 10% of 65% = 6.5%, so it should in theory take about 15 turns for the city to revolt for the first time and another 15 for it to revolt and second time and flip (assuming everything else stays the same all this time).

Example 2
Let's say a city is size 3 (and was never larger than that) with 6 surrounding plots belonging to the attacker, the average era is medieval (index 3) and the attacker has 85% culture (that means about 80% more than the defender). A situation like this may well occur if the AI founds a city in a stupid place and the human tries to get it to flip. This seems like a shoe-in for the attacker, right?

City Strength = 1 + ((1 + 3 + 6*3) * 80) / 100 = 1 + (22 * 80) / 100 = 1 + (1760 / 100) = 18.6

If the defender has 3 medieval units in this city (6 CG each), the odds that this city will revolt in any given turn are 1/10 * (18.6-3*6)/18.6 = 1/31. So despite the seemingly overwhelming odds for the attacker, it still takes in theory 31 turns before the city revolts even once, 62 turns for it to actually flip (in theory, in reality it might go faster or it might take even longer). And if the defender also has the state religion bonus, it takes twice as long: 124 turns! And if the defender adds one more unit, even a lowly Warrior, the city will even be completely impervious to flipping.

Strategy
So, what does this mean for strategy? Well, to 'cuturally attack' a Barbarian city, all you have to do is get more than 50% of your culture in the tile (or at least more than anyone else, if there are other players with cultural influence in the city as well) and wait. Taking over Barb cities is pretty easy to do, a lot easier than an AI or human city, so if you want take over a Barb city, make sure you get there first.

To attack a civilized city, make sure you own as many neighbouring tiles as possible, maximise your culture value in the city itself and also very important but perhaps not very widely known: spread your state religion there. It doubles the chance of a revolt! In case it's a close call, what can also help a bit is advancing to the highest era you can on the tech tree and either killing off primitive civs or giving them some of your more advanced techs (to increase the average era of all surviving players). More indirectly, you'll want try to lure away units from the city, as they form the main obstacle once you get your own culture rating over 60% or so. Requesting the defending civ to switch to a state religion that's not present in the city could be a big help and obviously preventing it from boosting its culture value on the plot on which the city is located (which will mostly come from the city in question, but might come from nearby cities as well) is also very helpful, if not usually something you have a great deal of control over (most you can do is keep them out of culture-friendly civics and make sure they're always at war with someone so they don't build too much).

Finally, flip any cities you want to flip as early in the game as possible, as the culture garrison values rise swiftly with time (more swiftly than era indices).

[phooey73]

Wow, thanks for the formula there!

Owned Plots == the ones inside the city radius (?)

So max is 20 then?

In that case, the city would probably flip if its tried. It needs about 45 to keep it Greek, and his garrison is worth about 30.

Does that mean 60% if its checked? So 6% chance each turn? What about on Marathon does it still check 10% of the time?

[Locutus]

Originally posted by phooey73
Owned Plots == the ones inside the city radius (?)

No, immediate neighbours only, so max is 8.

In that case, the city would probably flip if its tried. It needs about 45 to keep it Greek, and his garrison is worth about 30. Does that mean 60% if its checked? So 6% chance each turn?

No, not quite, if I understand you correctly. If the culture strength is 45 and culture garrison is 30, then the chance of a revolt is (45-30)/45 = 1/3 or about 30% -- or 3% per turn.

Of course, if you based culture strength on 20 neighbouring tiles you'll have to go back and do the math again, the chance of a flip will be even smaller.

What about on Marathon does it still check 10% of the time?

As far as I can tell, yes. Might be missing something but I doubt it.

[phooey73]

I'm looking at your culture ratio figures. Shouldn't 67% Attacker and 33% defender be a +100% more culture?

For +50% culture, it would be 60% attacker and 40% defender, right?

And then 85/15 is more like +600%.

[Locutus]

Originally posted by phooey73
I'm looking at your culture ratio figures. Shouldn't 67% Attacker and 33% defender be a +100% more culture?

No, you need to use the attacker value (67%) as basis, so 67-33/67 = 50%.

It's a bit confusing I admit...

[phooey73]

Okay.

So if all that is right, then 10 era-appropriate units should always, always, always be more than enough to eliminate any possible chance of the most overwhelmed and underfed city flipping.

And if there are *TWENTY* units, and a city somehow still flips.... we must be talking about warriors guarding a large city in the industrial age or beyond.

In my case, he has only 5 units, for a total of 30 points. Which is more than enough to make the chance ZERO.

Even if I wipe his culture down to zero and bring modern technology to all, he can either upgrade his units or double the size of his garrison, and the chance will still remain ZERO.

I suppose this is not really all that far fetched.

[Locutus]

Aargh, I found out I botched up my Great Artist example pretty bad, I fixed it now.



Yeah, the best possible culture strength in the game is pretty much a huge city (say size 30, although bigger is possible) in the late game (era is 7) with an attacker controlling all 8 neighbouring tiles and having 100% culture in the city itself (so 100% over the defender), and where the city also has the state religion of the attacker but not that of the defender:

City Strength = 1 + ((1 + 30 + (8 * 7)) * 100) / 100 = 1 + (87*100)/100 = 88

Double it for state religion and you have a culture strength of 176.

That would take 11 Modern Armors to defend -- but that's a pretty exceptional situation (a size 30 city with basically 0 culture and no state religion). For the modern era the situation is relatively worse: you lose 16 points on the culture strength (160 total) and the culture garrison per units goes down 4 (to 12), meaning you need 14-15 Tanks to defend in the worst-case scenario. For every earlier era you lose again 16 points per era on the culture strength (probably more because cities will be smaller), while the culture garrison decreases with only 2 points per age. Realistically, 10 units should in almost all situations be enough (if the defender only has state religion in his own city it can never take more than 8 units to defend it, even if the city is size 40).

[phooey73]

I had forgotten about the state religion factor.

So if the city has MY state religion in it (I am the "attacker"), I can double the "city strength" reading (???)

In that case there is hope immediately. He's only got 30 points of police and he needs 40. So I've got about a 2% chance per turn ?

[MPower]

Why is it better to flip a city, it is much easear to take it over (I'm sory I am new to the game and don't know a lot)

Thanks, Apriciate it.

[Locutus]

Originally posted by phooey73
I had forgotten about the state religion factor.

So if the city has MY state religion in it (I am the "attacker"), I can double the "city strength" reading (???)

I think I mistakenly used 'city strength' a couple of times where I meant to say 'culture' strength. But yes. But at the same time, if it has HIS state religion, that halves your city strength again. So if you're Taoism and he's got Islam, there are 4 options:

1) City has neither religion. No effect.
2) City has both religions. No effect (canceled out).
3) City has Taoism but not Islam. You get double culture strength -- so it's twice as hard to defend.
4) City has Islam but not Taoism. You get half culture strength -- so it's twice as easy to defend.

[phooey73]

Because Alexander is my boytoy? He's really "friendly" with me and I don't want to jeopardize that.

He also whipping Isabella for me.

[phooey73]

Well we're both Hindus, so its canceled out regardless.