Re-capturing past glories

[peter450]

I noticed that when you take a city off the AI it does not make any kind of concerted effort to get it back unless your relations are terrible or just through proximity, in other words it's no more likely to attack back if that city was one you took or founded yourself and thats fine for crap cities and the AI can spam a lot of cities too so it's probably good it does not try and get all it'c captured stuff back

However i think a lost capital or shrine city should most definately be a high priority recapture target, even after peace is signed, the AI should be looking at potencially preparing a war plan to get it back at a future point.

In a recent game an AI who signed peace after loosing a shrine city bringing in no less than 55 gold, which was very near there capital, decided a few turns later to go to war vs some obscure not very powerful AI civ on another island, why? having just been attacked, having just lost a very valuable city economically it's priority should have been on preparing to recapture that city ASAP

I don't think the AI should be doing this with all cities, but in the following cases i think the AI would be doing itself a big favour by looking at a strong counter attack as soon as they can, the AI's failure to focus on getting back obviosuly high value cities tends to lead to the trickle effect particularly vs humans of the AI loosing more and more high quality early founded cities and being left with the dregs of what it spammed later on, as it simply does not focus on getting back the cities that give it it's strong econonny and production.
It looses focus on whats important and a human then gradually through war and peace cherry picks there best assets leaving the AI very vunrable to a final push after it's fallen behind in power or teck as a result of loosing all these key cities, the AI should not settle for being nibbled to death and should make getting back cities of high value a priority, if it can't do it straight away for lack of units, it should build up a longer term plan via a big unit build up to try and recapture key lost cities as sioon as it can


I think the following cities should see the AI making a strong effort to get them back either straight away or as soon as it possibly can

A shrine city bringing in loads of gold

A lost capital (prime city location usally so they should want this back)

A city with an active (effects not yet obsolete) wonder or a city that had a lot of wonders built in it prev (high culture city), regardless of if they are now still working or not

Cities near the capital, (as loss of these makes the captial vunrable and also these are usally very big and old cities that contributed a lot)

 

[BobTheTerrible]

It makes sense to recapture lost cities, since the population will be welcoming of you. However, I don't think the AI has any sort of memory we can use yet, so it may not "remember" that they used to own the city that was just captured. I think that the AI should factor in the % of their culture in enemy cities and factor that into their calculations accordingly when choosing a city to attack.

 

[peter450]

Yeah that sounds like a good idea, if it's got a high % of there culture and is close to there borders it should definately factor that info into it's choice of which city to attack

 

[Cybah]

My opinion is completely different. The AI is ALWAYS re-capturing bad defended cities in my games. It will immediately attack my new cities if there are only some defending units while my attacking forces stack is on the way to the next city.

 

[jdog5000]

The game definitely keeps track of who originally founded each city, as well as who the previous owner was. Beyond that, culture % also gives a very easy measure.

The AI city targeting logic is a hybrid system ... the AI player will pick a target in a centralized way, but often the attack stacks decide to go after a different city. In the centralized logic, extra weight is given to cities the AI has owned before and where it has higher culture.

 

[peter450]

Say if you capture a ciuple of key cities and then get peace, does the AI have a higher likely hood to want to go back to war with you to get those cities?

I find the AI at war pretty good in this mod, a lot better than standard, the one area were i do think it's not much differant though is when it looses high value cities like a shrine city/cities around the capital, or the capital and then you get peace it seems to forget about the fact it's just lost some very important cities, were as a human player would either not make peace if it had the means to get that stuff back or would say yeah yeah peace and immediately start a massive build up to get those key cites back

Things like captials/wonder cities/shrine cities can mean the differance between a productive empire with a good discovery rate and cash flow for upgrades or a much weaker situation, a human player would see this and realise those cities must be re-captured to avoid a potencial big fall down the powergraph or falling behind teck wise over the long term

In my experiance the first few cities the AI founds tend to be the best as they are around the capitial, have lower maint costs and form the core of the civ's empire it's heart so to speak, while i don't think any changes are needed on relation to more far flung cities or ones built later on.
If the AI knows when cities are founded this should place a very high point of regaining these founding cities if they are lost, even after a peace treaty it should i think look to build up a stronger force in it's remaining cities to prevent loss and actively look to get those cities back, by either offering money, bribes, teck or building a large force and going to war

 

[jdog5000]

That's a good point, aside from the attitude effects from the prior war (which may be none if they declared on you and it backfired), the AI does not to my knowledge explicitly weight future war targets based on past loses of potentially key cities. It also certainly doesn't ever mount a coherent effort to recover them after a peace deal.

 

[Minor Annoyance]

...the AI does not to my knowledge explicitly weight future war targets based on past loses of potentially key cities. It also certainly doesn't ever mount a coherent effort to recover them after a peace deal.

If the AI starts making cities they used to own preferred targets then aren't we just teaching the AI to "take things personally" which is a human flaw? If they loose a lousy city it shouldn't care very much. Now if they are using the culture contained in the city as a factor for it being a target along with the value of the city then it would still more highly value its old cites without taking things personally.

 

[peter450]

I certainly don't think the AI should be focusing much on recapturing far flung small cities, or cities that don't carry much importance, but capital cities/wonder cities/shrine cities and formerly big important cities (usually the ones directly around the capital with lots of old culture) should factor in it's plans, i played a game a recently were a civ got really big and powerful, he had loads of big cities everywere!

while i don't think he should be making a special point of getting back a lost city just because he once owned it, if i were to say do a blitz on the area around his capital and take all those big old cities that have been standing since the dawn of time then he should be making a very serious attempt to get those back if possible, even if we go to peace, as without them he's going to suffer in the long term quite badly

With a human it's not just about taking things personally it's about knowing the value of something, your core starting cities are the ones that give you the vast majority of your teck science/culture/production etc thats what drives a human player to get important cities back, i've lost plenty of crap cities and not bothered going back for them, but everytime i lost a capital or core city near the capital i moved heaven and earth to reclaim it, because it's on those cities that your both short and long term strength depends a AI that looses key cities will wilt away in the longer term, the ai should be aware that some cities are worth a lot more than others and that loosing key cities is very bad for them long/medium term.

Don't get me wrong, i think the current city targetting logic works more or less fine, it's only in the situation were they loose capitals and key cities around it that there logic goes a bit awry,
cites that are captured tend to drop to like 1 or 2 pop if there surrounded by hostile culture, and may look not very good to the AI's routines, but if they were once prize cities in that AI's empire recapture will likely see them back up to there former glories quite quickly, as the main reason they became crap was unhappyness and hostile culture pressing right up against them when they were in enemy hands, back in friendly hands these cities can very quickly regain there former lustre and the AI should ideally factor that in

I agree that the AI should not be taking things personally, but i think in cases were there current high culture cities are inferior to cities that they no longer posses, but once did, they should go all out to reclaim those, as the only time that situation (of owned cities having less of your own culture than non owned ones) would arrise is if you lost your capital and big core cities around it and were left with newer founded cities that while perhaps big and productive would not be culturally as strong as the original founding ones

personally if they loose cities that are very near there capital, that should be high reclaim targets for security reasons and to keep the core around there capital intact and if there current cities have less of that civs culture than cities they once owned but no longer do, those high culture former cities should be targetted heavily aswell as they will be th former capital and surrounding cities no doubt and without them the AI will suffer badly, i think that would help the AI recapture past glories without making it all sulky about loosing just any old city as the ai spams lots of cities and i agree it should only make a special focus on the recapture of it's really important early cities

 

[Minor Annoyance]

Ah, I hadn't considered that a captured city would quickly go down in quality because of culture. So it may be that the AI is thinking too much about what the city means to the opponent now owning it rather than what it is them the original owner. I wonder if it's possible or valuable to have the AI consider two potential points of view regarding the value of a city. If the AI was trying to gain from a war it would target cities useful to themselves, but if they wanted to end the war they might take cities more valuable to the enemy to force them to quit. Although maybe there isn't enough difference it value to warrant having two calculation on city value (other than recapturing) for different war plans, assuming the plans I described are even something the AI does.