Better Ai? If you're 14...

[Roland Johansen]

It will have a bigger power advantage if power is calculated as strength^2. This larger power advantage is more accurate. This makes a difference when a warmonger is trying to decide whether to attack a small civilisation with good tech, or a large civilisation with bad tech. As a player, I will often choose the large civilisation, as I know that I can stomp over it, gaining experience, while taking few casualties. The power graph and AI decisions should reflect that reality.

This has far more to do with the human players abilities versus the AI. As a player, I can even get 5 to 1 ratio's against an AI with equal technology units. What does that mean? Are the AI units of a certain strength worth only 1/5th of the power of my units? We can't go that road because not all human players are equal and some even have problems with basically capturing cities.

Basically, you can't base a power to strength formula on your ingame experience fighting the AI's units of equal and lower technology because you're a far more capable tactician than the AI. The power to strength ratio should however work for every player versus AI and for AI versus AI conflicts.

The ideal exponent is presumably somewhere between 1.3 and 4, as I think we both agree. However, I think 1.5 is too low. I might go for something like 1.8, based on strength alone. When we add factors such as experience, war weariness, and upkeep, I still think strength^2 is fair.

There was a reason that I asked you to calculate the power ratio between strength 3 attackers and strength 3 defenders in your example. In your example, the strength 3 attackers only lose half the units of the strength 3 defenders which should lead to a different power valuation which is of course ridiculous. The reason that this illogical result occurs is because in your calculations, you completely ignored damage to units which weakens them. If 11 strength 3 units attackers attack 10 strength 2 units defenders and average results occur, then the result will be close to 10 wounded strength 3 units and 0 defenders. However that wounded aspect is very important. A counterattack against those wounded units will be far more effective. A strength 3 unit at 50% hitpoints is easily beaten by a strength 2 unit because of the large effect of hitpoints on the combat formula. The value of the exponent of 4 is therefore not reasonable, even in this matchup.

I do agree that 1.3 it a bit low, but not that 1.5 is a bit low.

Using goal seeking on making the average of 100% to 5% in 5% increments as close in ratio to a power curve, I get 1.32. Dividing the naive curve of ^2 by 1.32 gives us a 1.52 power curve once we take into account the fact that damaged units are weaker than the naive approximation assumed.

This last part wasn't clear to me.

Another thing to take into account is the power flux of each side. You could calculate this based on, say, X years of imperial production producing the highest power efficiency unit, where X varies based on game speed. This misses, however, being close to a new technology that will allow massive upgrades...

I suggested a similar approach to approximating the 'power' of producing units during the war.

It's a bit hard to estimate the power of upgrading due to a new technology. You could calculate the power of the army when it would be fully upgraded with the new technology and multiply that value by 1.5^-t where t is the number of turns until the technology is available (as long as t<=10). Then you take a weighted average of the current power and the future power (weights 1 and 1.5^-t).
In the case that upgrades are available with current technologies this leads to an average of the current power and the power while fully upgraded.
The final formula becomes a bit more complicated as an army can easily not be fully upgraded and be close to a new breakthrough military technology for even more upgrades.

The AI at some point decides your getting way to advanced and knows that it can't keep up, so it decides it needs to invade and at least destroy some cities and try to get as many other AI's to declare on you. Orginal BTS AI would do the same thing, but instead of 150 units it would send 20 and get maybe one far off AI to declare on you that meant nothing. That's the difference in Better AI.

This is to me, where Blake and others started down a road many others disagreed with. The ai in this case, isn't even playing "to win" it's basically just playing in a spoiler role. I still have problems, understanding why anyone would want the ai to do this. By mindlessly attacking the strongest player (and ok they might temporarily get a city or two, at the cost of their entire army), and thus almost guaranteing their own ultimate destruction, they are basically giving up there and then.

The AI has no tendencies to attack the strongest player or the player who is going to win. It will let your spaceship peacefully fly to Alpha Centauri even if it has a huge stack of units next to your undefended capital as long as you have friendly relations or your army is large enough to scare them from declaring war.

If you however have a small army and relations aren't good, then it might attack but it is unrelated to how close you are to any of the victory conditions or how many points you have on the scoreboard.

There has been some discussion however to maybe change this so that the AI is more aware of the victory conditions and how close someone is to them. This divides players into two groups: those who see it as a game with one goal 'victory' and those who want the AI to behave only according to diplomatic relations.
In the real world, there are actually two responses of countries towards strong countries:
A) Kiss their but and become friends
B) Create another group of nations that can compete with the strong country and its friends.
I would enjoy similar behaviour in civ4.

 

[DrewBledsoe]

. In fact, I would argue that it's rather the unintended consequences of improvements made in particular areas.

I'll go with that bit Dom, but then it raises the question of whether the improvements outweigh the negatives, in each particular individual's opinion. Has the "cure" poisoned the patient, as it were..

 

[Dom Pedro II]

I'll go with that bit Dom, but then it raises the question of whether the improvements outweigh the negatives, in each particular individual's opinion. Has the "cure" poisoned the patient, as it were..

I would say no, but that's also because I've never actually seen the AI do this. The most I ever saw was the AI come at me with a 60-unit stack, but that was actually pre-Better AI. I don't know what you're doing, but it's obviously causing this particular set of events to occur more often. As said before, a simple way to prevent the AI player from declaring war on someone with a strong, high-tech force is to increase the difference in power values between the units.

 

[MartinHarper]

Basically, you can't base a power to strength formula on your ingame experience fighting the AI's units of equal and lower technology because you're a far more capable tactician than the AI.

My experience is that the Better AI is entirely capable of putting together a stack of doom to gain local dominance, which is all it needs to take advantage of situations such as the one we're discussing.

In your example, the strength 3 attackers only lose half the units of the strength 3 defenders which should lead to a different power valuation which is of course ridiculous.

The attacker has gained local dominance, and therefore the effective power of her units has doubled. This is the value of good tactics. Of course, that shouldn't modify the power graph.

That wounded aspect is very important. A counterattack against those wounded units will be far more effective.

It is true that 10xS2 defenders will do more damage in total than 5xS3 defenders, despite causing fewer outright fatalities. However, I disagree that the wounded aspect is always as important as you make it out to be. If the attackers are in friendly territory, for example, then they can easily get to a friendly city to heal up, and the AI does a good job of doing this.

 

[Roland Johansen]

The attacker has gained local dominance, and therefore the effective power of her units has doubled. This is the value of good tactics. Of course, that shouldn't modify the power graph.

The point that I'm trying to make is that the way you're calculating a power exponent can result in any power exponent that you want. By using an instance of local dominance, I can make a large group of low strength units defeat a smaller group of high strength units where less low strength units are lost than high strength units. In this way, you can never create a model for unit power based as a function of unit strength that will fit all of these cases.

A good way to see when a group of weak units is of equal power to a group of strong units is by finding numbers of both groups where on average they will annihilate each other. Then you know that the groups are equally matched on average.
When you're looking at cases where one group is utterly victorious annihilating the other group and the stronger group remains with lots of wounded members than you can't expect to get a good formula that predicts cases where they're equally matched especially not when you're completely disregarding the wounded members of the victorious group.

 

[Minor Annoyance]

The AI has no tendencies to attack the strongest player or the player who is going to win. It will let your spaceship peacefully fly to Alpha Centauri even if it has a huge stack of units next to your undefended capital as long as you have friendly relations or your army is large enough to scare them from declaring war.

I was once up against a huge empire, far bigger than me that had before used the Apostolic palace unite the world against me forging me to buy people off, and had more units and more production then I, and the launched a massive sea attack taking out all the ships I had in the water and nuked several cities. I was pretty much finished after just one turn. All this happened one turn after my spaceship had made it to Alpha Centari.

There has been some discussion however to maybe change this so that the AI is more aware of the victory conditions and how close someone is to them. This divides players into two groups: those who see it as a game with one goal 'victory' and those who want the AI to behave only according to diplomatic relations.
In the real world, there are actually two responses of countries towards strong countries:
A) Kiss their but and become friends
B) Create another group of nations that can compete with the strong country and its friends.
I would enjoy similar behaviour in civ4.

I wouldn't want the AI to do things regardless of diplomatic relations (unless that's their personality). Going back and playing Master of Orion 1 and 2 I quickly remembered that non-aggression pacts are really nothing to count on, and alliances are just as temporary, and I missed having those things actually mean something. So I would like them to act in character, but not be oblivious to their being a competition going on.

 

[Roland Johansen]

I was once up against a huge empire, far bigger than me that had before used the Apostolic palace unite the world against me forging me to buy people off, and had more units and more production then I, and the launched a massive sea attack taking out all the ships I had in the water and nuked several cities. I was pretty much finished after just one turn. All this happened one turn after my spaceship had made it to Alpha Centari.

Heh, I didn't see that last remark coming. It's a remarkable story. I guess that while you were constructing the space ship, you must have recognised the threat of this strong civilisation so it must have felt as a tight victory. An AI that would have been more aware of the victory conditions wouldn't have let the game end like it did (which was of course the point of your story).

I wouldn't want the AI to do things regardless of diplomatic relations (unless that's their personality). Going back and playing Master of Orion 1 and 2 I quickly remembered that non-aggression pacts are really nothing to count on, and alliances are just as temporary, and I missed having those things actually mean something. So I would like them to act in character, but not be oblivious to their being a competition going on.

That's the middle ground and I thing I hold the same opinion.

 

[MartinHarper]

Roland, it seems like we both acknowledge that the existence of unfair fights and healing shows that Pow > Str ^ 1.3, and we merely disagree about how important those factors should be. I can't think of any mathematics that will tell us. Can you?

We could run an AI vs AI test to find out. We could give one AI +50% combat strength on all units, and give the other AI +100% military production. If the AI with the combat strength bonus tends to win, then we can conclude that Pow > Str ^ 1.7. If the AI with the production bonus tends to win, then we can conclude that Pow < Str ^ 1.7. Do you think this would be a fair test?

 

[Dom Pedro II]

I think we can all agree that EITHER an increase of ^1.3 or ^2 would be a hell of a lot better than the status quo

 

[Roland Johansen]

Roland, it seems like we both acknowledge that the existence of unfair fights and healing shows that Pow > Str ^ 1.3, and we merely disagree about how important those factors should be. I can't think of any mathematics that will tell us. Can you?

No, that's likely impossible (and yes, I'd also be in favour of increasing the exponent a bit over 1.3). You'd need to somehow approximate the average behaviour of the AI in an average game or something like that. Some way to estimate the size of the stacks that will meet one another. The game is way to unpredictable to do that.

I would like to point out a final thing. Power = Stength ^ 1.3 isn't the optimal condition for the weaker units to beat the stronger ones. It's just the ratio at which they will annihilate oneanother when they meet in numbers in which their power is balanced according to that formula. If one of them has local dominance in power, then it will be able to heal its units after the fight and thus get a better ratio at beating the other than the one given by that formula. It's just that when a situation of local dominance arises for the stronger units, then most will survive and heal back to full strength while if a situation of local dominance arises for the weaker units, then only a fraction will survive and heal back to full strength. So in a typical battle between AI civilizations that are balanced according to that formula, the one with the stronger units will benefit more from meeting the others forces in an unbalanced fashion.

We could run an AI vs AI test to find out. We could give one AI +50% combat strength on all units, and give the other AI +100% military production. If the AI with the combat strength bonus tends to win, then we can conclude that Pow > Str ^ 1.7. If the AI with the production bonus tends to win, then we can conclude that Pow < Str ^ 1.7. Do you think this would be a fair test?

In theory it's fair and actually a good suggestion. But in practice, I don't think it's that feasible. You'd need variations in every sense that they're possible in civ to get a fair representation of an average situation. So varying land shapes, varying city sizes, varying empire sizes, varying numbers of starting units (still balanced according to your above suggestion) and other things that I'm presently not thinking about. It would require many tests and each would take quite some time to set up in the world builder.

It's also rather unpredictable how future improvements to the combat AI will change the balance.

By the way, there's an element in this setup that I hadn't thought about at first and which will favour the weaker units due to the present AI (or at least I expect it will favour the weaker units). The AI always uses a certain number of units to defend its cities. This means that for the one with the stronger units, that a larger fraction of their units is used to defend (if we assume both have a similar number of cities and the weaker one more units according to your suggestion). And thus the one with the weaker units will have an attack stack ready before the one with the stronger units. The attack stacks formed by the AI are usually capable of capturing cities as they gain local superiority. So I guess that due to numbers, the initiative will be with the AI with weaker units. It will capture cities first and it will get the advantage of local superiority first.
I do admit that this is a weakness of the AI and one that ideally would be fixed in a future version of BetterAI. Still, this stupidity will harm the AI with strong units but numerical inferiority.

Good idea about the simulation, but I think it's too much work.

By the way, I did a search for the post where I arrived at figure of 1.3 and after some time found it: power to strength exponent of 1.3. The calculation assumes that hitpoints are lost in a continuous manner in battle while they are actually lost in big chunks. This discrete manner in which hitpoints are lost causes big jumps in the combat effectiveness of units and the continuous approximation is useful to get an average of these big jumps in combat effectiveness.
My post is just one post in a longer discussion about power in that thread. The present discussion is of course not the first one about this issue.

 

[Grave]

I have to agree with Drew on this one... the AI puts TOO much emphasis on building military units. I've played about 5 games in a row so far, and the AI seems to have a virtually unlimited amount of military units at it's disposal.

As a player who emphasises more on building city improvements and having a small, technologically advanced, highly specialized military, playing against the AI like this is near impossible. The player simply cannot build city improvements and build enough military units to stave off the AI, who if you're at war with, seems to continually assault you with wave after wave of 6+ stacks of military units. And you have to keep in mind that just about EVERY AI player in the game behaves the same. So if you go to war with one, his allies will go to war with you. Which is fine... but then you have even more stacks to defend against. It's too much.

And the thing is, eventually the AI will catch up with you in technology, because you'll switch your focus on trying to build up your military and neglect your science research and city improvements. So now you'll have an AI who is now technologically equal to you, and still has an army 10x your size!

What can be done about this? I'm experimenting on adjusting the iPower values right now. It seems to help a little, but only with the decision the AI makes in going to war with you. They still build disgusting amounts of units. And if you do capture one of their cities, they practically have NO city improvements. Just enough to build more military units.

So time after time, leaders like Gandhi and other peaceniks always die early in the game, and leaders like Ghenghis Khan and Alexander almost always dominate.


Right now, the formula I'm using is taking the iCombat value and squaring it. So a Musketman with an iCombat value of 9 has an iPower value of 81.
What can be used to make the AI put more emphasis on building city improvements?

 

[MartinHarper]

If you do capture one of their cities, they practically have NO city improvements. Just enough to build more military units.

I find that AIs do a reasonable job building city improvements. However, all cultural city improvements are destroyed on capture, along with a proportion of non-cultural improvements, so I don't often get to see them.

What can be used to make the AI put more emphasis on building city improvements?

Modify the LeaderHeadInfos.xml to alter the iBuildUnitProb value. Gandhi has 15, for example, whereas Ghengis Khan has 35.

 

[r_rolo1]

@ Grave

Sorry, but with all due respect, you are complaining in the wrong place. Civ IV, for all that matters, is a game where in average being agressive pays up, in the sense that in average you gain more with x hammers of units than with the same hammers invested in buildings in wonders or buildings ( or like I seen written by a SG player " If they build wonders and you build units, you will end with units and wonders" ). OTOH the combat system definitely allows zerging as a viable strategy ( as you say yourself ). So, consciously or not, you are saying that you want a AI that refuses to play by a viable strategy just because you want to roleplay a little in a way that the game rules will shred apart, pretty much like refusing to make three-point shots in basket by whatever reason and complaining that the other team does them .... Sorry, but that is at most rule issue, not a AI caracther issue, atlest in what concerns a mod that explicitly states as objective making a AI that plays better this game with the current rules.

About your complaints: It is definitely possible to play against a warring AI in disavantage of production. There are even posted victories in Deity OCC-AW. OTOH, it is definitely possible to play a game with 10-20% of the weakest AI in game in Immortal and never be attacked ( I have one posted ), so I find hard to believe that the AI is that unmanagable as you state. In fact, your post suggests that you don't have allies in the game you are describing.... maybe the biggest issue there is not the AI agressiveness, but the way that the diplomacy network ended to be.

 

[Grave]

@ Grave
Sorry, but with all due respect, you are complaining in the wrong place. Civ IV, for all that matters, is a game where in average being agressive pays up, in the sense that in average you gain more with x hammers of units than with the same hammers invested in buildings in wonders or buildings ( or like I seen written by a SG player " If they build wonders and you build units, you will end with units and wonders" ). OTOH the combat system definitely allows zerging as a viable strategy ( as you say yourself ). So, consciously or not, you are saying that you want a AI that refuses to play by a viable strategy just because you want to roleplay a little in a way that the game rules will shred apart, pretty much like refusing to make three-point shots in basket by whatever reason and complaining that the other team does them .... Sorry, but that is at most rule issue, not a AI caracther issue, atlest in what concerns a mod that explicitly states as objective making a AI that plays better this game with the current rules.

About your complaints: It is definitely possible to play against a warring AI in disavantage of production. There are even posted victories in Deity OCC-AW. OTOH, it is definitely possible to play a game with 10-20% of the weakest AI in game in Immortal and never be attacked ( I have one posted ), so I find hard to believe that the AI is that unmanagable as you state. In fact, your post suggests that you don't have allies in the game you are describing.... maybe the biggest issue there is not the AI agressiveness, but the way that the diplomacy network ended to be.

I've been around these forums a lot longer than you have, and I think I'm OK at modding, so I think I know my way around... with all due respect, of course.

Now, with that said... to me, Civ 4 is a strategy game. Not a warmonger game. The purpose of the game is to build a civilization that will stand the test of time. It's not a "who can build the biggest military first" game. If I wanted to play that, I'd go play Command and Conquer or something.

I do not need Better AI nutswingers to come to it's defense on reflex against me. Not that I need to justify myself, but I LOVE Better AI. It is probably the single most essensial "mod" out there for us ModPack makers.

Stop overanalyzing what I'm asking. All I am asking is what do I need to do to round out the AI a little better, adding more emphasis on city improvements? I'm not bashing on jdog, or anything about Better AI at all. I'm just asking a question, and if you do not have the answers, then do not talk down to me like I'm some troll.

Now, if we can please be a bit constructive and try and make valuable info out if this thread...

I see the Techs and City Improvements also have iPower tags. If I raise the values on these tags, could that add emphasis to the AI to build city improvements? Also, what does iAsset do?

Modify the LeaderHeadInfos.xml to alter the iBuildUnitProb value. Gandhi has 15, for example, whereas Ghengis Khan has 35.

I'll check this out... THANKS.

 

[r_rolo1]

I've been around these forums a lot longer than you have, and I think I'm OK at modding, so I think I know my way around... with all due respect, of course.

Now, with that said... to me, Civ 4 is a strategy game. Not a warmonger game. The purpose of the game is to build a civilization that will stand the test of time. It's not a "who can build the biggest military first" game. If I wanted to play that, I'd go play Command and Conquer or something.

I do not need Better AI nutswingers to come to it's defense on reflex against me. Not that I need to justify myself, but I LOVE Better AI. It is probably the single most essensial "mod" out there for us ModPack makers.

Stop overanalyzing what I'm asking. All I am asking is what do I need to do to round out the AI a little better, adding more emphasis on city improvements? I'm not bashing on jdog, or anything about Better AI at all. I'm just asking a question, and if you do not have the answers, then do not talk down to me like I'm some troll.

Now, if we can please be a bit constructive and try and make valuable info out if this thread...

I see the Techs and City Improvements also have iPower tags. If I raise the values on these tags, could that add emphasis to the AI to build city improvements? Also, what does iAsset do?

Ok, sorry if you thinked i was calling you a troll. I was just stating the same thing I said to the OP: the game rules allow and even reward sometimes zerging, so saying to the AI to not being agressive is pretty much nerfing it, that is the exact oposite of what is the objective of this project. I definitely agree with you that the game should not be a wargame, but the fact is that it is a wargame most of the times because of the added value that agression has and the AI should be coded to live in it. So the complaints about this subject should be directly focused to the persons that made the game rules and not the ones that code the program to play it.

Well, as you asked for something constructive ....

The hack and slash way of making the AI to fear more your less unit number-focused high tech army is definitely edit the ipower tags in the CIV4UnitInfos.xml ( for units ) and CIV4BuildingInfos.xml ( for buildings ) for values that reflect more what you have in mind ( remember that one unit there means 1000 soldiers ). This is a huge trial and error work to get it minimally right, as you, with your modding history should understand The more elegant way would be what is being discussed by others, that is making the soldier count of a unit being dependent in some way of their actual fight ability ( I would definitely prefer something more refined than a crude exponential, but processing times probably would not allow it ), but that probably gives even more work than the first option I stated.

Now I seriously doubt that the AI would be much more prone to build buildings just because of the enhaced power they give. Most likely editing the iBuildUnitProb in the CIV4LeaderHeadInfos.xml for every leader to a lower level would be more adequate to that objective ( less units built => more hammers for buildingis in average ).

Oh, and I'm not a fanboy of the better BtS AI project I definitely disagree with some of the options that the coders did so far and I would probably had liked that some aspects of the game that need to be reviewed had a bigger priority ( AI diplo needs a big cleanup for a example , especially in what concerns the UN and the AP ). I simply think that the extensive complaint you did about the AI agressiveness was made to the wrong persons, as i stated above.

 

[Roland Johansen]

I have to agree with r_rolo here. He wasn't talking down to you like you say or at least I can't read it in his post. He is explaining that the AI is using a strategy that is more efficient in civ than the one you are using. Of course, the AI is likely a far worse player than you are so its execution is far worse but the basic idea to focus a large share of production on units is a good choice given the game mechanics. The game mechanics in civilization IV make focussing a large share of production on units (instead of buildings) more efficient in many cases, especially since the AI isn't that good at war. There's no reason to get angry at r_rolo for pointing out that this is not an AI issue but a game mechanics issue.

If you which to change all that, then you need to change the game mechanics, not the AI choices in this game. A way to do that is by making units a less efficient investment by making them more expensive (in building cost and upkeep) compared to buildings. Of course, then the AI needs to be made aware of this change in game mechanics so that it builds less units because otherwise it will over invest in units. The iBuildUnitProb is probably a good way to change that.

Changing the power rating of units to a square of the strength won't cause the AI to build less units. It just changes the way that the AI perceives the power balance between weak and strong units. This might make the AI's with weaker units focus on building even more of these weak units to be able to somewhat compete in power with the civs with higher strength units. The AI just tries to keep up with its competition in military power.

edit: posted too late. r_rolo responded himself with a similar post.

 

[MartinHarper]

A way to do that is by making units a less efficient investment by making them more expensive (in building cost and upkeep) compared to buildings.

I'm not sure that will help. If units are twice as expensive, then that also means they are twice as expensive for my enemy. If they are twice as expensive for my enemy, that means that my enemy will have fewer units. If they have fewer units, they are easier to conquer. If they are easier to conquer, each of my units is more valuable.

Grave - I think you should instead focus on making wars of aggression less effective as a strategy:
* Allow city flips via culture (even after conquest). It should be possible to mod the game to make culture flips more likely. I don't know what factors set how likely a city is to flip.
* Increase the effect of War Weariness, and remove or modify things that negate the effect of War Weariness. Currently AIs get a handicap to reduce their war weariness - iAIWarWearinessPercent - you could increase this value to 100.
* Increase the unhappiness penalty from going to war against others of the same religion.
* Bring back Partisans/Guerillas generated from taking an enemy city.
* Increase the MAD deterrence effect of nukes by getting rid of the Non-proliferation treaty and the Manhattan Project prereq.
* Increase the power of the UN against warmongers by adding an "International Criminal Court Warrant" resolution that gives a civ that declares war a "-5 The world considers you a criminal" happiness penalty.
* Play on Quick (by the time you walk your army to the enemy, it's already obsolete).

 

[Roland Johansen]

I'm not sure that will help. If units are twice as expensive, then that also means they are twice as expensive for my enemy. If they are twice as expensive for my enemy, that means that my enemy will have fewer units. If they have fewer units, they are easier to conquer. If they are easier to conquer, each of my units is more valuable.

You're right. Everyone will still invest the same number of hammers in units, only these hammers will result in less units. This just means that wars are fought with less units but the same number of hammers.

By the way, I disagree with removing the war weariness bonus for the AI. It really needs that bonus. You're not changing the game mechanics with that rule, you're just removing an AI bonus that it needs to compensate for its artificial stupidity.

The other things could work.

My suggestions to make war less effective would be to change a few combat mechanics so as to help the defender.

1) A game rule change that makes it impossible to fully remove the city defence bonus. It can only be halved and nothing more.
2) Halve the effect of siege engines collateral damage on units in cities.
3) Very important: limit the amount of damage that a siege unit (or bomber) can inflict to something like 25% both normal and collateral damage (so a unit cannot go under 75% of max hitpoints due to attacks from siege units)
4) Give units in their own culture zone a combat bonus for all fights within that culture zone.

This will make defending cities a lot easier and taking cities costly.

All of the above of course has nothing to do with the BetterAI project. It's more a war mechanics mod.

 

[Dom Pedro II]

I would say that if you want to make war less preferable, the real trick is to make the product of war less valuable. Why do people go to war in civ? To capture cities. Cities give more gold, more research, more units, etc. So one possible solution is to increase distance and number cities maintenance. This will make huge, far-flung civilizations less valuable than small to mid-size civilizations. I've added something I call "security maintenance" to my mod which increases a city's maintenance costs from unhappiness and foreign populations.

Higher unit production costs would certainly mean smaller armies, but it would also mean less time that could be dedicated to building construction. On the other hand, increasing unit upkeep costs would decrease the army sizes. Increasing the supply costs for units fighting outside your borders would also help too. It would require players to amass a big war chest before going to war.

 

[Minor Annoyance]

City flipping after conquest has significant implications. It increases the number of unit needed to attack because more have to be left behind, and they often have to be left behind for a significant portion of the game, and you have to pay maintenance on those units that can never leave