Cities reverting to AI... (RANT)

[Grim Legacy]

The point:

A) My cities, conquered or not, revert to the AI too easily. The rate of reversion is, in light of my efforts, frustratingly high.

B) Actions to prevent reversion do *not* have any effect. Reversion seems to be pre-determined at a given turn, despite alleged 'anti-reversion' actions by the human player.

The case:

Game 1) - Emperor, Large.

I am second in the Histograph, the Russians surpass me by 40%. I engage in a massive war, destroying 4/5 of the Russian civ. The war is on 1 continent (mine) only. Between me and the remaining Russian cities, a wasteland is filled by small temple-bought cities of mine. Below that lies Moscow and Kiev (both have wonders).

I take these cities. Moscow is size 43, 20-odd resisters. I rush in 25 units. I starve the population, railroad is connected to my lands, I rush a settler to reduce the population. End of turn. I rush another settler, move in 10 more units. End of turn. City reverts.

I travel back in time and this time land a lenghty and painful barrage on Moscow to kill the offending population. I take the city at size 19. This time I manage to retain the city longer. I starve it down to size 4 (2 russian, 1 iroquois, 1 persian, LOL don't ask the AI did that) and built a rushed temple. I rush a library after about 7 turns. In the size 4 city, 12 troops are stationed. The city reverts.

Once more I travel back in time (only 5 years this time). I rush the library 1 turn earlier. I move in another 10 troops. The city reverts.

Note that at the same time, the lame persians and babylonians manage to retain 7 russian cities, while having a much lower culture than I have. Doh!

Game 2) - Deity, Large

I am in the earliest stages of the game, all my energies focused on building settlers as efficiently as possible so that I can colonize the land before the AI (Russians and Germans) do. I manage to push my cities out to the Germans, at the cost of Russian territorial success. The Germans have higher culture than I do, but that can't be avoided at this stage.

My building orders are simple at this stage: 2 warriors, settler, temple, settler, worker, library. What more can I do to gain culture and expansion? I can't go any faster. In the time I build a warrior, the AI founds 2 cities...

My city of Avignon borders on the German lands, but not closely -the distance between German cities and Avignon is 4 squares at minimum. Avignon is right below Orleans, which in turn is right below Paris (capitol doh). East and West of Avignon I have my own cities as well.
Avignon has a temple and 2 warriors stationed at size 4. It has 110+ culture.

It reverts to the Germans. !?! I carefully pronounce WTF? and undertake time travel once more.

I rush a library 5 turns before Day Zero. I move in 2 extra troops (4 troops total, size 4 city). I raise luxuries, making ALL citizens happy.

It reverts at Day Zero.

---

I think this is unreasonable. I can't even hold on to my own cities, and I'm *not* neglecting culture -beelining for temples and libraries and rushbuilding them.

I am keen to hear some word of advice, but until I see the Light, I'll whine about it.

[Capt Dizle]

Good post. Nothing I can say however, except to observe that its very strange that Firaxis expects us to pay money and then play this game for enjoyment.

I quit playing when I had to take a break to go to the dentist and realized I had less stress in his chair than in front of my computer.

[Grim Legacy]

Heheh. Well I do admit it is frustrating to lose cities this way. I'm cool with the notion that you have to *work* to hold on to your cities, but when my desperate efforts fail to have *any* effect...

But I found a little comfort in my latest Deity, Large game: I am alone on a sizeable and LUSH continent!!! I've found Eden! I am so gonna be an early Republic in this game, the AI will cry mercy!

So it *is* possible to get a good starting position in Civ3! Don't abandon hope.

[Peets]

I suggest to lower it in the editor!
I saw that on the very high levels it reverts back 90% change
to high if you ask me...

[Grim Legacy]

Originally posted by Peets
I suggest to lower it in the editor!
I saw that on the very high levels it reverts back 90% change
to high if you ask me...

Hrmpf that does sound high.

However, the editor is off-limits. I want to play the game as is. The editor feels like giving up.

I actually thought up a strategy for conquest to be able to grab the wonders after all... hehehehee.

[Ignorance]

Yes this is very frustrating.
Between this and the high corruption rates, I rarely bother trying to keep distant enemy cities.
This is not how it should be.
I don't want to raze every enemy city I capture, but there just doesn't seem to be a way around it.
I have tried it. But when the desert even when I have more culture, all citizens are happy and city is filled with troops, thats pretty much the only option.

[rah]

I have done a little time travel myself to test the mechanics of this and have always been able to Change History and keep the city from reverting.
I took out the French capital in the middle of all their culture. The entire city was surrounded by French cities 20+ in size. I figured this would be a good experiment. (difficulty level emporer, pre-patch)

Paris was about sized 25. When I took it, it had about 16 resisters. 3 turns later it reverted. Went back in time and airlifted in a few more troops, and it didn't revert.

A few more turns later it reverted, extra troops didn't stop it. But then (using the wayback machine only one turn) I switched every citizen that wasn't resisting into an entertainer every turn. THE CITY DID NOT REVERT. Rushed a temple, cath, settler, and starved it down to about size 6 and repopped with my own until I had 50 % my citizens. It never reverted after that, ending my testing.

I have never had one of my own cities revert after the city borders expand the first time. (always play religious as one trait)

RAH

[Skeletal Dragon]

City Conversion seems to be dependant on a few things:

1) City Size (effect is small)
2) Number of Culture Generating Buildings (effect is small)
3) Culture surrounding city (effect is big)
4) Resistors and/or civil disorder(?)
5) Total Culture/Score(?) of Civ (effect is very big!)

Note that these are from trying to convert enemy cities, but they should follow the same mechanics when applying them to your own cities converting. I've found that the city will usually convert in 1-5 turns if at least 2 of the 4 sides of its culture borders are touching another civ's culture borders.

From the looks of the things on this thread, it seems like that Resistors (and maybe Civil Disorder, I don't know since they have not been mentioned) play a huge role in cities converting back. And logically, it would be realistic since a Resisting population would have more of a chance of converting back. I've never had this happen to me since I always make turn the population into all Entertainers to get out of Civil Disorder so I can start rushing Temples and Cathedrals. Maybe its a bug with the check that the game makes with conversion? Maybe if the entire population was made Entertainers, it would find only Entertainers and not Resistors?

The strangest thing happened the other day as I was playing Civ 3. Me (Greeks), China, and the Zulus were under a MPP and my allies just wiped out the Babylonians on another continent and there was a lot of empty space in the territory. Since I was short on Rubber, I decide to make a city by a source of Rubber that was on the empty space. To the south was a new Zulu city and to the north was the former Babylonian capital. The new Zulu city had about 100 culture or so (since it had like border 2 or so), and my city was positioned so that my culture border would touch two of the Zulu city's culture borders. I rushed Temple, Library, and Cathedral (I had a Sistine Chapel ). My city had too much corruption (75% or so?), so I forgot about it for most of the game as I was trying to kill off the Japanese. Then suddenly, after my new city's culture borders increased (I think it was at about 500 or 1000 or so), the Zulu city converts! And the Zulu population was at 8 or so! And very far from my own continent! Wierd, but I'm not complaining I'm guessing that it was partly due to the surrounding Culture and partly due to my high culture as well. (I was leading in both Culture and Score)

I hope this helps.

[Ogie Oglethorpe]

Maybe I'm imagining things but it seems to me that nationality of citizens plays a role as well. Despite having quelled all resistors and keeping citizens as content, entertainers and/or happies, I think nationals from a given civ are more likely to revert back to the orignal civ than citizens of your flavor of nationality. This naturally means that captured cities are more likely to revert than ones of your own making. Yet another reason for the raze and replace phenomena that seems to be teh most effective means to deal with this. Kinna stinks tho when your aiming for cities with juicy wonders.

Just my thoughts

Og

[MarshalN]

Grim -- try declaring peace. I think it makes a difference there too... if you're at war with a civ where your citizens are originally from, they're more likely to revert than if you're at peace with that civ. I've found that to be true on occasions, but of course, peace is not always possible.

[Chris Woods]

It really needs to be changed that cities won't 'despose' (or whatever) if [city_size] < [number_of_garrisoned_troops]. This would still reflect the difficulty in integrating enemy cities near the heart of their empire by requireing a large amount of troops to be dedicated to maintaining the city while allowing the standard cultural takeovers to be effective. (It's not reasonable to garrison that many troops in every settlement you have.)

As it stands, taking cities is basically absurd. I've reached the point where my settlers almost outnumber my infantry in an offencive engagement and I raze anything without a useful wonder. You can always add the workers you get from razeing the city back into it once things have settled down.

The alternative (with what we have currently) is to genocide the enemy (take or raze all his cities very quickly) thus eliminating him as an option for the cities to revert to. This can be extremly difficult sometimes, and I have often had cities revert back when the AI's nearest city is leagues away and their cultural borders are no where near eachother.

Chris Woods

[Grim Legacy]

Originally posted by rah
I have done a little time travel myself to test the mechanics of this and have always been able to Change History and keep the city from reverting.
I took out the French capital in the middle of all their culture. The entire city was surrounded by French cities 20+ in size. I figured this would be a good experiment. (difficulty level emporer, pre-patch)

Paris was about sized 25. When I took it, it had about 16 resisters. 3 turns later it reverted. Went back in time and airlifted in a few more troops, and it didn't revert.

A few more turns later it reverted, extra troops didn't stop it. But then (using the wayback machine only one turn) I switched every citizen that wasn't resisting into an entertainer every turn. THE CITY DID NOT REVERT. Rushed a temple, cath, settler, and starved it down to about size 6 and repopped with my own until I had 50 % my citizens. It never reverted after that, ending my testing.

I have never had one of my own cities revert after the city borders expand the first time. (always play religious as one trait)

RAH

In your testing, you had turned every laborer into an entertainer. And this helped.
To my displeasure, I already do this by default, to starve the city. There is a remote chance I left 1 worker to gather shields, though, so I'll check that out.
I may try population migration too, never tried that.

As for my own city that reverted: it had expanded its borders twice already (110 culture)... and there was no disorder or foreign nationals.

I never play religious civs though, maybe that's a big difference there.

[Grim Legacy]

Originally posted by Skeletal Dragon

From the looks of the things on this thread, it seems like that Resistors (and maybe Civil Disorder, I don't know since they have not been mentioned) play a huge role in cities converting back. And logically, it would be realistic since a Resisting population would have more of a chance of converting back. I've never had this happen to me since I always make turn the population into all Entertainers to get out of Civil Disorder so I can start rushing Temples and Cathedrals. Maybe its a bug with the check that the game makes with conversion? Maybe if the entire population was made Entertainers, it would find only Entertainers and not Resistors?

Yes resitors play a major role. That's why I always roll in as many troops as quickly as possible. Only after resistance ends can one rush buildings.
Resistors can't be turned into entertainers though.

[Grim Legacy]

Originally posted by Ogie Oglethorpe
Maybe I'm imagining things but it seems to me that nationality of citizens plays a role as well. Despite having quelled all resistors and keeping citizens as content, entertainers and/or happies, I think nationals from a given civ are more likely to revert back to the orignal civ than citizens of your flavor of nationality. This naturally means that captured cities are more likely to revert than ones of your own making. Yet another reason for the raze and replace phenomena that seems to be teh most effective means to deal with this. Kinna stinks tho when your aiming for cities with juicy wonders.

Yup -so that's why I was so surprised to see Moscow revert back with 2 russians, 1 persian and 1 iroquois as a population.

I took Moscow cuz it had 3 (!) wonders.

[Grim Legacy]

Originally posted by MarshalN
Grim -- try declaring peace. I think it makes a difference there too... if you're at war with a civ where your citizens are originally from, they're more likely to revert than if you're at peace with that civ. I've found that to be true on occasions, but of course, peace is not always possible.

Yes it helps -makes them happier at least. Thing is, in my examples I was at peace with the civs in question.

[Grim Legacy]

Originally posted by Chris Woods
It really needs to be changed that cities won't 'despose' (or whatever) if [city_size] < [number_of_garrisoned_troops]. This would still reflect the difficulty in integrating enemy cities near the heart of their empire by requireing a large amount of troops to be dedicated to maintaining the city while allowing the standard cultural takeovers to be effective. (It's not reasonable to garrison that many troops in every settlement you have.)

As it stands, taking cities is basically absurd. I've reached the point where my settlers almost outnumber my infantry in an offencive engagement and I raze anything without a useful wonder. You can always add the workers you get from razeing the city back into it once things have settled down.

The alternative (with what we have currently) is to genocide the enemy (take or raze all his cities very quickly) thus eliminating him as an option for the cities to revert to. This can be extremly difficult sometimes, and I have often had cities revert back when the AI's nearest city is leagues away and their cultural borders are no where near eachother.

Chris Woods

I agree with you.

As for your last paragraph: I was thinking of 'removing' the other cities first, before taking the wonder cities, but with your far-away reversion story... I may be looking for eradication instead. Quite a hassle.

[rah]

Originally posted by Grim Legacy


In your testing, you had turned every laborer into an entertainer. And this helped.
To my displeasure, I already do this by default, to starve the city. There is a remote chance I left 1 worker to gather shields, though, so I'll check that out.
I may try population migration too, never tried that.

As for my own city that reverted: it had expanded its borders twice already (110 culture)... and there was no disorder or foreign nationals.

I never play religious civs though, maybe that's a big difference there.

Yes you have to check every turn for those resisters that stop resisting, I forgot to change them, one turn,to entertainers and the city did revert. But when I went back and changed it, it didn't revert. So please verify if you can. But the point was that is was not automatice, it can be influenced. I don't think the random number generator has any effect because i would use the way back machine and try some artillary barrages to change the seed number it would have no impact on whether the city reverted or not.

Your city changing is disturbing. I've never seen one go once it had over 100 culture. The only difference I can see religious having an effect is that with cheap temps and cath, your overall civ culture is probably higher then non religious civsl.

Was that city part of your core cities, or was it really out there?
Was it pre or after patch?
What difficulty level.

I have put new cities on enemy continents and not lost one yet.
Hmmm, as always more testing will be needed to truely understand the mechanics involved.


RAH

And i keep all entertainers even after resistance has ended until the culture has gone up and I've attained at least 50% of the pop as mine. I do also practice the +1 garrison principle.

[Padmewan]

As people noted earlier, starving the city by turning everyone into entertainers works. (I've also prevented reversions by turning everyone into tax collectors, which is more beneficial).

You have to make sure you switch them every turn. The AI will automatically put them back on production next turn. Whenever I had cities revert, it was because I stopped paying attention and let the AI manage the citizens.

[Blaupanzer]

Factors HAVE to include some random measure. I had a Zulu city jump to me at size six. (Me: Babs, then at regent, more culture than any two other civs combined. Zulus last on culture histograph.) I filled it with culture, held it for centuries. 100 or more turns later with 400+ culture and only 4 Zulu citizens left out of 12, it reverted back. Garrison = 3. Waybacked 10 turns. Made garrison 4. Two turns past the previous reversion date, it reverted again. By then my culture was three times theirs. Zulu capitol 3 squares closer than mine. Can't explain this; don't know how to prevent it; not clear whom to ask.

[rah]

I have never been able to change the outcome of a revert or not revert by going back and changing the seed number. Only by changing other things have I been able to influence it. These two statements tell me it's NOT random.

UNLESS the probability of the event is so high that changing the seed is unlikely to change the outcome. And since I've tested this over 30 times the probability of the event would have had to be close to 97-99% for me not to have experienced it.

The changing garrison to 4 is interesting. I usually try to keep no more than 3 foreign nationals in a city since I'm only likely to have a garrison of no larger than 3 on cities unlikely to be attacked. (carryover from old CIVII days) Hmmm. But I've always played Religious and never had a problem.

[inca911]

Hiya! A couple quick notes and clarifications.

According to the Firaxians (aka Soren), the following factors affect reversion:

Distance to capitals.
Total ratio of your culture to the other civ's culture.
Total ratio of the culture generated within the individual city under your reign vs. culture generated in the city under the AI's reign.
Garrisoning of troops.
Happiness (WLT*D). (not well quantified or reliable)

As a rule of thumb, Soren stated that you need one offensively capable unit for each potential defector to prevent reversion. Bombard-only capable units do not count (I don't know if naval units count, but I act as if they do not). So if the city is pop 23, you better have 23 units to spare to guarantee reversion does not occur. I have followed this guideline and I have *never* had a city revert back to the AI. Bombardment down to pop 1-3 is my mode of operation unless I want the entire city (in which case I use outdated units like warriors to provide a garrison even in the modern age!).

Note that rushing cultural improvements does not really affect the total ratio of your *generated* culture vs. what the AI has amassed in that city over the years so rushing cultural improvements is not a reliable method to use. In your example, you take Moscow--likely with a very high culture generated from the AI (strike one), probably near his new capital as well(strike two). It has a pop of 43 and you put in 25 units (only count as 25 if they are all offensive capable) making you 18 units short of guaranteed control. You starve the population (any actual population decline occur yet) then rush a settler so the city is at 41. Next turn rush another settler so you are at 39, and add 10 more units bringing your unit total to 35 (still 4 short!). I cannot rationalize your other examples, but based on Soren's rough guidelines, I would not expect to keep such a large city as pop 43 Moscow with only 35 units. Even without consideration of only offensive capable units making a difference, every factor is against you. Bombardment or massive unit commitment is the only way to go. Your cultural contributions in rush-building temples, libraries, etc are meaningless in the relatively short term and certainly a capital like Moscow will have enough culture to attempt a reversion even when the Russians are nearly dead.

Some of your examples are indeed a bit fishy, but it all is governed by cold analytical code and Soren's tip has kept me completely immune to cultural reversion, even though I tend to play the militaristic and industrious Chinese (and I don't pop-rush in Despotism)!

[rah]

"Total ratio of the culture generated within the individual city under your reign vs. culture generated in the city under the AI's reign."

Then that would tell me to rush those improvement (temps/caths) imediately. Getting it off 0 has to be big. Doing a simple compare 0 divided by any number (0) vs. any other ratio has to make a difference. And pop rushing them even makes the most sense. Eliminate negatives to increase positives. Gotta help. And your overall culture for that ratio comparison will also be improved. Every little bit helps.

RAH

[Skanky Burns]

Size 43?!? Arrrgh

If i were you, i would bombard that city to a nice managable size, like 2 or 3. If you simply must have a massive population in that new city, bring in some loyal settlers and have them increase the population after you have captured the cities. Rushing cultural buildings will always help, but in the case of a previous capital with 3 wonders no less, the only effect will be long-term.

[Master Marcus]

I don't know how random the defection engine is, but perhaps the citizens happiness - WLTKD or not - is more important sometimes than culture facilities and garrisons. In a game post-patch one of my well-garrisoned and cultured city ( superior in culture points than nearest Babs cities ) defected and the ONLY LOGICAL REASON I can think of is that I had at that time only 2 luxuries, not enough happy citizens and experienced a short civil disorder. I don't remember having lost a single non-conquered city to the AI in previous pre-patch games ( Regent, Monarch ).

[inca911]

Originally posted by rah
"Total ratio of the culture generated within the individual city under your reign vs. culture generated in the city under the AI's reign."

Then that would tell me to rush those improvement (temps/caths) imediately. Getting it off 0 has to be big. Doing a simple compare 0 divided by any number (0) vs. any other ratio has to make a difference. And pop rushing them even makes the most sense. Eliminate negatives to increase positives. Gotta help. And your overall culture for that ratio comparison will also be improved. Every little bit helps.

RAH

Likely it is not a true ratio, or if it is a true ratio, then they certainly have code to account for possible divide by 0 errors. It could be a simple differential determination. The point is that rushing cultural improvements will only help in the long term, but not in the short term. A Temple that has been in the city for several thousand years has generated a ton of culture that a rush-built temple and library will not even come close to approaching for a long, long time.

Master Marcus:
Happiness does indeed play a factor, but nobody (sans Soren) really has a handle on that. Certainly civil disorder will increase the chance of defection.

[Grim Legacy]

Originally posted by rah

Yes you have to check every turn for those resisters that stop resisting, I forgot to change them, one turn,to entertainers and the city did revert. But when I went back and changed it, it didn't revert. So please verify if you can. But the point was that is was not automatice, it can be influenced. I don't think the random number generator has any effect because i would use the way back machine and try some artillary barrages to change the seed number it would have no impact on whether the city reverted or not.

I double-checked. There were no laborers, only entertainers in the city. Resistance had been ended already (at peace with victim's nation). Garrison of 8 units, size 4 city. Temple present, libary rushed... too late.

Was that city part of your core cities, or was it really out there?
Was it pre or after patch?
What difficulty level.

Yes part of my core. 9 tiles from my capital. 1 city between reverted city and my capital (my own city Orleans).

After patch.

Deity.

[Grim Legacy]

Originally posted by inca911
As a rule of thumb, Soren stated that you need one offensively capable unit for each potential defector to prevent reversion. Bombard-only capable units do not count (I don't know if naval units count, but I act as if they do not). So if the city is pop 23, you better have 23 units to spare to guarantee reversion does not occur. I have followed this guideline and I have *never* had a city revert back to the AI. Bombardment down to pop 1-3 is my mode of operation unless I want the entire city (in which case I use outdated units like warriors to provide a garrison even in the modern age!).

Yes I am aware of this. I do not count artillery as troops. I was speaking of Tanks and Infantry units only.

Note that rushing cultural improvements does not really affect the total ratio of your *generated* culture vs. what the AI has amassed in that city over the years so rushing cultural improvements is not a reliable method to use. In your example, you take Moscow--likely with a very high culture generated from the AI (strike one), probably near his new capital as well(strike two). It has a pop of 43 and you put in 25 units (only count as 25 if they are all offensive capable) making you 18 units short of guaranteed control. You starve the population (any actual population decline occur yet) then rush a settler so the city is at 41. Next turn rush another settler so you are at 39, and add 10 more units bringing your unit total to 35 (still 4 short!). I cannot rationalize your other examples, but based on Soren's rough guidelines, I would not expect to keep such a large city as pop 43 Moscow with only 35 units. Even without consideration of only offensive capable units making a difference, every factor is against you. Bombardment or massive unit commitment is the only way to go. Your cultural contributions in rush-building temples, libraries, etc are meaningless in the relatively short term and certainly a capital like Moscow will have enough culture to attempt a reversion even when the Russians are nearly dead.

Well the first part of my first example *is* understandable, no argument there. What baffles me is what happens when I go to the extreme to prevent reversion... and get next to zero results. Note the drastic measures I took in the second attempt at Moscow.

[Grim Legacy]

Originally posted by Skanky Burns
Size 43?!? Arrrgh

If i were you, i would bombard that city to a nice managable size, like 2 or 3. If you simply must have a massive population in that new city, bring in some loyal settlers and have them increase the population after you have captured the cities. Rushing cultural buildings will always help, but in the case of a previous capital with 3 wonders no less, the only effect will be long-term.

Heh. It was size 68 !!! 5 cattle, pure grassland...

Note that when it reverted in my final attempt, it was size 4, and had 1 persian and 1 iroquois citizen (in other words: not even native russians!).

[Barchan]

My favorite part about this whole affair is that the game essentially sets you up for the fall. "There are resistors," says your scowling advisor as he then proceeds to suggest you place a heavy garrison there to quell them. Then your heavy garrison is swallowed whole when the city promptly reverts back to its original owner. Your advisor is then strangely quiet as to what you should do next….

[Grim Legacy]

Originally posted by Barchan
My favorite part about this whole affair is that the game essentially sets you up for the fall. "There are resistors," says your scowling advisor as he then proceeds to suggest you place a heavy garrison there to quell them. Then your heavy garrison is swallowed whole when the city promptly reverts back to its original owner. Your advisor is then strangely quiet as to what you should do next….

Well, one of the cooler hidden Espionage options also allows you to bribe the opponent's Military Advisor -so this should level the playing field.