City flips - some tests and a look at the rules

[Polonius]

Clearly, a lot of players are both mystified and irritated by the ways that captured cities sometimes flip back to their original civ. On several recent threads about this I had the temerity to suggest that a careful reading of the manual and Civilopedia might go a long way towards providing an answer. Predictably enough I was pelted with rotten fruit by posters who felt certain that the AI was striking unfair low blows at them, and that I didn’t know what I was talking about.

Perhaps there was some justification in their annoyance. However, I’ve seen this sort of baffled outrage spring up about several other issues in this game, and it then generally it dies down again as a better understanding of the game mechanics spreads. So I thought I’d try running some tests on a selected city to see if I could throw a bit more light on this by deliberately manipulating the situation.

In the posts below I will firstly spell out what the known rules DO say, and then give some results of what I found (and don’t hold your breath – it wasn’t a magic formula).
Please excuse all the waffle but, believe it or not, I am actually trying to help here.

 

[Polonius]

According to the Civolopedia, manual, etc. these are the factors that affect defections (in no particular order). Note that the suggestion is AFFECT and NOT that each factor ALONE will either CAUSE or SOLVE flips:

 Proximity of the foreign capital swings it one way, proximity of yours swings it the other. Proximity of large amounts of foreign territory can also contribute. Note: a Forbidden Palace is not a capital.

 Civil disorder or “We Love the King” days. These are pretty much the opposite ends of the happiness scale. So happiness apparently does have an affect.

 Garrisoning troops and Resistance. Citizens who are “Resisting” are said to be more likely to flip back. Garrisoned troops can quell resisters. Whether troops have a deeper benefit than this is unclear (to me anyway).

 Presence of “Foreign Nationals” who will be less happy, and more easily swayed by propaganda while you are still at war with their original civ (see other points on happiness and propaganda).

 Propaganda. That is to say spies who can flip cities. Unless you are a Democracy, hidden spies (who you can unmask if you have the Intelligence Agency and plant spies of your own) can flip cities with propaganda.

 Culture. Culture. Culture. The entry about City Defection begins with a paragraph about culture. When the citizens do flip they don’t say “we fancied our chances against your garrison” but the reason given is that they found the pull of their original culture too strong.

I have no reason to believe that the reasons given by Firaxis are either wrong or don’t work. Certainly, other posters seem to have found all of these factors to have some affect, just that none of them appear to be clinchers individually. But it does seem reasonable to believe that the most significant weighting is given to relative overall cultures. That is how the phenomenon is presented in the manual, and that is how it seems to work. However, like most things in this game it also appears to contain a random element.

It would seem to me that the more you meet the above conditions the higher your rate of protection against defections – but you will probably never reach a guaranteed 100% safety factor. Just as an archer can occasionally beat a tank (but don’t build your attacks round that fact).

 

[Polonius]

So how did my tests go? Firstly, I had never had a city defect back after capture (smug, smug) so I didn’t have a handy test save ready to use. Instead, I picked what looked like a reasonable candidate and tried to make it flip. In this game I had captured around 50 cities and kept about half of them (none flipped – but some I razed as being too far away and too late in the game to bother with). The one I chose to try and let flip was the first one that I had captured when I invaded a new continent. The city was only two tiles from the (Aztec) capital and was surrounded by a flourishing empire (fairly late in the game). It started with 9 citizens. It should have been relatively tough to hold.

So I emptied out my garrison, threw it into disorder, and encouraged resistance. Unfortunately however, I could not manipulate my high culture. I outranked the Aztecs by almost 4 to 1. So it was still extremely difficult to deliberately get a flip. (In many, many runs over hundreds of turns I managed only two or three, and these were when I restarted from pre-capture).

So it might be worth emphasising what the culture ratings actually mean.

In awe of………………3:1 ratio of culture points
Admirers………………2:1
Impressed….…………..1:1
Unimpressed...………...3:4
Dismissive……………..1:2
Disdainful……………...1:3

I think that some players are misled by “Impressed” and expect it to be more powerful that it is. It just means level pegging.

 

[Polonius]

So what are the answers? I believe that if you follow the guidelines in the Civilopedia (as outlined above) - particularly high culture - you will greatly reduce flipping back. As I said, I have followed that premise over many, many captures and never yet had a flip back in a game. If Resistance really bugs you, you can manipulate it using the editor (as long as the rule change is made on a save before you made the capture), but I don’t believe it will give anything more than a modest contribution against flips. Certainly, by setting the chance for to make it happen at 100%, I was able to get a city with all 9 citizens resisting (interestingly resisters don’t eat your food so can’t be starved out). But even in a city with no garrison they could still go turn after turn after turn without flipping. So the reverse is unlikely to be certain to work either.

Others have posted workable strategies that involve razing and rebuilding, starving out the original nationals, and parking your troops outside and just retaking flipped cities (watch out for enemies sneaking in, as empty cities seem to become a magnet for wandering troops). All would appear to work, but I prefer to cheer those new citizens up and put them to work for me immediately. Whatever style you like best.

It looks likely that you can probably never completely overcome a small chance of a random flip even if you have everything else right. But like other aspects of the game, you can get used to how it works and learn to expect varying degrees of it depending how well you’ve met the other conditions. And if you still can’t stand it – then make sure you save just before you capture a city. If it flips you can probably reload and beat that random seed - if you can be bothered repeating the necessary number of turns.

Remember though that exactly the same design idea runs through every combat action in Civ3 – if you know the rules you can plan around the likely outcome, but without ever being 100% certain of the result. Do you really want a game where it’s all predictable and A always beats B? Where it’s just a matter of learning the rules and patterns and then watching it unfold in a set manner? I don’t, but if you do there are games like that on the market. Civ3 just isn’t one of them – and I say Hooray for that!

This will be one of my last posts as I have other commitments calling, but I hope this can help someone, and that you all continue to enjoy your Civ3. Many thanks to Thunderfall and the crew for a great site, and to the players who have posted helpful and entertaining posts here over the past few weeks.

Cheers. Chris, aka Polonius.
I may be old, but there are a few obscure parts of me that are still not completely in ruins.

 

[HisDivineShadow]

Very interesting, and it ties in with what I have observed.

However, it would be worth testing it in a scenario where your culture was on more of a parity with the opposition. It could be that your test civ's high culture is swamping the other effects due to garrison size, proximity to capital and so on.

Another thing I've found is that if I'm having a problem with a city flipping, I can take it and immediately give it to a weak civ (preferably one with which I have a right of passage). The city will flip back after a few turns, whereupon I retake it, and it generally stays mine after that.

 

[Mack the Knife]

Polonius, i know you guys have nice beaches, pretty girls and weird animals over there. you need to go outside more!

 

[Polonius]

Originally posted by Mack the Knife
Polonius, i know you guys have nice beaches, pretty girls and weird animals over there. you need to go outside more!

Mack, I'm one of the weird animals.

But you're right. And I'm off out with my family this very afternoon - just as soon as I can locate my dark glasses and full "reality repelling" space suit. It's a new year here now, so I might as well get my annual token visit to the outside world over with early.:crazyeyes

To HisDivineShadow:

I think your ideas of further testing are great, but as a relentless high culture player I have no suitable saves to test with. I think you're exactly right about my culture swamping theirs (it's why I do it, it has several benefits).

Would another obsessive loony with the right type of save please step forward and volunteer to waste a day or two?

Cheers to you all, and I really do have to go outside now...

 

[Greadius]

Stationing a lot of troops in a city DEFINATELY presents a flip. I captured Berlin and signed a peace treaty, a few turns later it flipped. So I did the natural thing and reloaded. I tried turn again and it flipped again (I love those random number generators that generate the same random occurance everytime). Next time, I stationed a large army into the city and it didn't flip.
The next time a city flipped, I did the same thing (stationed a huge army) and it didn't flip the next try either.

 

[Mack the Knife]

DEFINATELY presents a flip

by the context of your explaination i take it you meant prevents. and i have a question. you say large then huge armies garrisoned. what numbers are we talking here? thanks

 

[Polonius]

Alas, stationing a large army may not always definitely prevent a flip (see heaps of other posts on this forum). It may well have been either the reloading that worked (if early enough) or perhaps the fact that the larger army tipped the balance just far enough to change the equation in your favour.

All the factors quoted seem to help the equation - which appears to be a cumulative one - but none have yet been proven to be cast iron certainties. At least that's my experience, and apparently that of many other players too.

But it should be testable. If you have some figures about the required troop numbers, perhaps others could test it out on other saves where apparently stable cities have flipped (particular useful would be ones that took quite a few turns to flip). If you're right you are guaranteed instant hero status!

 

[J Fred Muggs]

I'm wondering if game difficulty has anything to do with the flipping.

 

[Greadius]

Originally posted by Mack the Knife

by the context of your explaination i take it you meant prevents. and i have a question. you say large then huge armies garrisoned. what numbers are we talking here? thanks

I had a feeling when I hit post I misspelled something....

That game was a long time ago, so I can't retest it. I know that that it is ground units only (or according to the instruction manual). Initially I had about 5 troops in there, and when it didn't flip I had about 20. The population of the city was 6.

I've found that certain 'random numbers' reroll identically when I reload... in combat for examples. I've had 'amazing' circumstance fights occur identically in two consecutive reloads. Ditto with bombing attacks; the same structures gets damaged everytime, in the same order. I'd bet city flipping works the same way.

 

[Readylander]

Polonius,

I have not lost many cities to culture flips. I could only think of one additional factor that I consider when capturing cities:

The government types of the opposing civs. For instance, if you are under despotism and capture a city of a civ under democracy I believe you are more likely to lose that city to a culture flip than if you were under democracy yourself.

 

[Polonius]

Originally posted by Greadius


I've found that certain 'random numbers' reroll identically when I reload... in combat for examples.

Hi Greadius,
I have read that the game design uses a "seed' technique specifically to prevent players just reloading and re-doing fights. So if you only reload one turn back the outcome will be the same. However, if you go back further - "pre-seed" - the outcome can be different.

With city flips my guess is that a major "seed" is sown when the city is taken. Certainly, the few flips I got all required me to re-start pre-capture. If I reloaded after I'd set everything up I couldn't seem to get any noticeable change (as with your fights). It's hard to be certain though without doing thousands of turns worth of testing.

 

[Polonius]

Originally posted by Readylander
Polonius,

I have not lost many cities to culture flips. I could only think of one additional factor that I consider when capturing cities:

The government types of the opposing civs. For instance, if you are under despotism and capture a city of a civ under democracy I believe you are more likely to lose that city to a culture flip than if you were under democracy yourself.

Hi Readylander,

I think that you might well be right about this, but I couldn't find any support for the notion when I ploughed through the manual, etc. It was on my "list of suspects" but I couldn't dredge up any evidence for a conviction.

I think I'll have to give up on this one for now - before they lock me up. I'm happy with the way it's working for me, so I really do need to shut up and tiptoe away now.... All the best with your battles.

Cheers.

 

[Greadius]

Originally posted by Polonius
I have read that the game design uses a "seed' technique specifically to prevent players just reloading and re-doing fights. So if you only reload one turn back the outcome will be the same. However, if you go back further - "pre-seed" - the outcome can be different.

I figured that out when I attacked a galley with an Ironclad, pounded the galley down to one health without getting hit, then watched in amazement as it made a comeback and won. I think I reloaded 3 times before I realized the fight was pre-determined.

If I feel I'm getting the major shaft I'll just change the situation in the previous turn.

If flipping is "pre-seeded" then Berlin in my game should have flipped despite the giant garrison. I reloaded directly before I hit the turn button and woke & moved all the troops to stop it from flipping.

 

[Rhandom]

I appreciate that you are trying to make this have some sort of logic to it. But again, if ALL of those factors list above are in your favor, you still get flips. If all of those factors are strongly in your favor, you still get flips. If you overwhelm every single one of those factors by 5:1, you still get flips.

And if anything, in my experience (and many others have posted similarly) having a huge military in a city actually encourages flipping. I've had several flips which took 8+ units out of size 4 or less cities, reloaded, moved them out, and no immediate flip.

However... considering the bass ackwards way the game does random numbers, it is possible that changing the garison used up (or more likely, never got to - ) a "bad" "random" number on quelling resister checks, and the game then did the culture flip test after the resistance quelling check.

Polonius, insulting the people who have experienced really stupid flips by saying "Predictably enough I was pelted with rotten fruit by posters who felt certain that the AI was striking unfair low blows at them", like we were conspiracy nuts or crybabies, adds nothing. If you haven't experienced dumb flips, good for you. I have. Many of us have. We have done everything the guides have said to prevent them, and they don't get prevented. The formula for checking culture flips is quite likely broken or bugged.

Its not exactly a secret that the game shipped witha bunch of broken features. This is likely one of them. I have another that has ruined two games now when the last unit of a civilization is destroyed, completely crashing the game, on every reload. Others have seen the game crashing when trying to make peace. Its not every game, I've crushed many civilizations in various games, and it almost never happens. But it does happen, and it totally wipes the hours you had previously spent playing that game.

Likewise for bad strings of "random" numbers. There are definately places in the pregenerated strings where the game locks up into an all bad or all good series of numbers. Not just killing a single unit, but literally strings of 5 or more units that win every hit or lose every hit, when the odds were plainly against that occurance. Like rolling 15 or 20 1's in a row on a d6. It doesn't happen often, but when it does, it can completely ruin a game.

There is an unfortunate sentiment of "blame the victim, the designers are flawless" common on game fan sites. It can get very absurd, I remember in Anarchy Online, a game where 99% of the players experienced awful bugs, abyssmal framerates, etc, there would be a couple of new posters everyweek that would come on and say they had no problems with the game, it ran perfectly, and all of the "whiners" just had crappy computers/didn't know how to maintain their computers/ didn't know how to play the game/ did read the manual and so on. However, the AO nightmare was so bad that several times a patch would come out that would upset the delicate balance that their computers were in to make the game playable, and then they'd be back *****ing like the rest of us. Some even apologized for insulting everyone else, an utterly amazing event for a web board.

Now, even if the bad flips people rail against here are due to a 1/100 or whatever chance that the game may leave for a flip, no matter what (and I don't think this is the case), you have to ask - is it good gameplay to flip a city randomly without any warning and zero chance of preventing it? Yes, other games have random events that can be quite catastrophic, but... they also happen to the enemy. Many times, there are things you can do to prevent them - like the novas or meteors from MOOII. I've never recieved the gift of a size 15 city with 2 wonders in it in the heart of the enemy empire. I've never recieved a warning that a city was considering leaving. It just happens.

And thats what we want changed. Either a clear set of rules about how flips occur, a warning and a chance to prevent flips, or a fixing of whatever bug allows a size 7 city to flip with 5 military units and a 6:1 ratio of my citizens:captured citizens with 100+ culture generated from rush built temple, library, cathedral and colliseum, 3 spaces away from my forbidden palace, on the other side of the continent from the lone remaining enemy city on an island off the continent, in a WLTKD from all happy citizens, when I have a 3:1 advantage in culture and a 100:1 advantage in power, and living in a republic. Not just once, but 5 times in 3 turns cities with similar circumstances do this.

 

[Peteus]

 

[Lt. 'Killer' M.]

POLONIUS: Thanx a lot! I was high time someone actually tried it out. I did, too, and I`ve found one thing that really pisses me off though:

You can sometimes station 60 (!) troops in a city and still it will flip - how? The garrison shouöd reduce it to rubble!

The story is as follows: I`m beating the Babs up. Badly! They loos cities faster then they can reform their troops. Then I take their capital (with the Pyramids). I station lots of troops and singn a peace.

Now since I`ve had a few flips before I was cautious this time: I razed about half the cities I took to thin out the culural influence and also rushbuilt temples and libraries. Still: the city flips. I reload, put in more troops (from 30 to 62 units, all tanks, Mech Inf and bombers!) - it flips. A size 4 citiy takes 62 (!) units along!

I think this sucks. let the city be destroyed by unrest or let the troops out, but not this way. Oh, and another thing: When I moved the troops to a neighbouring city, that one flipped - someone trying to save the Babs?????

 

[muppet]

Originally posted by Greadius

If flipping is "pre-seeded" then Berlin in my game should have flipped despite the giant garrison. I reloaded directly before I hit the turn button and woke & moved all the troops to stop it from flipping.

I think what Polonius is trying to say is that:
Preseeding probably occurs for flipping as well as combat, and stationing of troops is a mitigating factor, not an absolute factor. Consider it a modifier to the seeded number result, not an overrider of the seeded result.

Just because a number is seeded does not mean that modifiers are not applied otherwise outcomes would be determined without accessing terrain modifiers in combat.

FOR SIMPLICITY SAKE: consider a scale of 1-10
where 10 = 100% chance of flip, and 1-5 = 0% chance of flip

seeded number = 6
Resistor + unhapiness = +4
Result = 10, therefore city flip this turn

Reload, add 5 garrison = -5
Result = 6+4-5 = 5, therefore no flip.

Without access to source code, as Polonius and others have said, we need a large representative sample set to generate the statistics in order to arrive at the code's approximate formula. That is a lot of work for anybody to do on their spare time!