Apart from small tweaks that are unlikely to break anything, I'm not developing this mod further.
Oh. That's unfortunate! I guess it doesn't make sense to make involved suggestions then.
Such games are still winnable more often than one tends to think. But of course starting over is understandable when a major setback appears entirely avoidable
If the AI is nice and is amenable to making peace after just taking this big city (instead of crushing me by taking my capital before I can prepare armies), that game was indeed likely still winnable.
Having some setbacks can make for more interesting games, but in this instance I felt my whole approach had underestimated the AI.
I'm attaching a screenshot from a 16-civ all-AI game I just ran. The "care for our brothers" modifiers among Hindus (and also among Taoists) don't exceed +5 here. This doesn't seem like a quite typical example; these are pretty irreligious leaders, but those in your game may have happened to care a lot about religion. Of course +5 is still a lot when there is scarcely any rational reason for liking civs of the same religion.
I checked and I actually have two +3, two +5 and one +6 (this one took a long time). But these modifiers provide an early boost that allows to consistently have open borders and trade because even if you anger a civ by refusing to help them in a war or trade-wise, they won't dip too low relation wise, and by themselves are quite significant. As you say, these boosts are big compared to the rational backing behind them.
The way shared religion allowed me to wage war without fear of my neighbours intervention, even though the target is of the same religion, is the most egregious issue in my opinion.
Ideas:
- Reduce the shared religion boost
- Add a diplo penalty among civ of the same religion when at war with a non-vassal civ of the same religion, to make involvement of other civs more likely
- Add a small diplo boost among civ of the same religion when at war with a non-vassal civ of a different religion (but that would be exploitable without other conditions)
- Change the aggressive expansion penalty so that a neighbour conquering a lot of lands is seen as threatening even if these conquered lands are not directly adjacent. My neighbours are way too complacent regarding my expansion.
One idea I had was to escalate the diplomatic reaction as the war wears on – by cancelling OB after a while (which might even restrict the aggressor's unit movement) or by adding a penalty "this war of yours spoils our relationship" (or "you keep brutalizing our friends").
I actually like this idea. This way, a small "border war" that doesn't last very long and mainly see some culturally disputed cities change hands is not going to instantly cause a gang-up if relations were good previously. But a long war with many cities being captured would increase diplomatic resentment significantly.
It would also be nice if the AI wasn't so stubborn when furious. The culturally contested city I want is weakly defended and undefensible but it seems the AI will rather have me capturing all their other cities than giving it up in a peace deal.
The expansionism penalty might've been a bit higher if the other civs had been more powerful militarily. I didn't want civs to get upset whose opinion isn't going to be taken seriously anyway. Makes it seem like everyone's ganging up on the leader – without accomplishing much.
Excessive ganging up might not be very fun. But the military power ratios were not such that the AI should have been deathly afraid of doing anything. Soon before the war started, the power ratios were 0.9 (my target), and 0.8 for each of my two main neighbours. One of these neighbour had to the north another neighbour that might have joined me but even cancelling these out, that still leaves an AI with a decent power ratio, that was in friendly term with my target, had no other threat to worry about (the sea to the east, my territory to the west, plus a small strip of border with a non-threat). The two neighbours at 0.8 ratio had already given up two cities each to me from cultural pressure against money/technologies (and I had revolt-flipped another one from each).
Your religion spreading so effortlessly is worrisome to me, but, then, this ...
My civilization was located in the centre of the landmass. I got some borders with 6 civilizations before wars started. So this location probably helped too.
Open borders + early shrine + no concurrent religion early on did the rest. The AIs usually convert early after one or two of their cities receives the religion, so even though the religion hasn't had the time to spread to most of the cities (in 600 AD the shrine revenue was 22gpt and 5 AI leaders had the religion). Confucianism got founded later on my continent but it could only get two civilizations. Taoism was founded later on my continent too but too late to matter.
In particularly, I wanted to adjust the Shrine income to the map size and to halve the Shrine income when running a different state religion.
Yes, I think some adjustment to the Shrine income depending on the map size is a good idea.
I tried the "Isabella Island" spawn and a holy shrine doesn't feel too powerful when it's only 9-10 cities that benefit from the religion. In huge, it can easily be 30 or 40 cities however. Changing the income depending on the state religion is also an interesting idea.
... is kind of rare when half(?) of the civs start on your continent.
It's not rare. If you assume two continents with each half of the civs, and that which civ discovers a religion is random you have a 12.5% probability of a 3-0 in early religions for one continent and a 87.5% probability of a 2-1 for one continent.
Now let's assume the player creates a religion and the other two religions are random. Then you have 25% of 3 religions on the player's continent, 50% of 2 religions on the player's continent (one concurrent religion) and 25% of 1 religion on the player's continent (no concurrence so the early religion can really dominate). If you count the fact that the player won't try to get a second religion and that without the player there is one more civ on the other continent, the odds are even higher than 25% to have the player's religion be the only early one.
It doesn't happen every game of course, but I wouldn't call better than 25% odds rare. That's rather common.
If you have three continents, the odds of an early religion having no concurrence and dominating your whole continent are even better, although the continental spread is smaller.
Temples and Cathedrals aren't generally very powerful buildings, so I don't see this as a big issue. Culture isn't universally valuable and cities only need so much happiness (and religious buildings provide it only at a substantial production cost). Wonders that boost all religious buildings change the math but apply only to the state religion.
Making additional missionaries to have multiple religions spread in many cities might not be too powerful of a play (if one consider early Culture wins aren't a dominating strategy). I'm biased by the fact that in the real world, cramming a lot of religions in one place is a bad idea, so the game encouraging this is strange.
It's kind of a shame that "excess happiness" is completely pointless in Civ4. Either you have enough and all citizens work at 100% efficiency or you lack and additional citizens immediately refuse to do any work. But that's too deep of a design choice to address.
The +2 production from Apostolic Palace nevertheless is way overpowered; perhaps I'll reduce it to +1. I was going to do that along with the other religion changes, but since those aren't going to happen ... Or perhaps it should be +2 production for the AP owners and +1 for other civs with the AP religion.
My early build for new cities in conquered territory was monastery+church after getting a missionary to spread the religion because of how the AP bonus would help to increase the hammer output early. It dropped off when the monastery became obsolete (this provoked a significant economic and to a lesser degree production regression because wonder-boosted monasteries are very powerful and scientific method doesn't really do enough to compensate immediately)..
I think reducing the AP bonus from +2 to +1 would be a good idea. At 400 hammers the AP isn't too costly at this stage of the game and it pays off easily even with +1. If many rivals have the religion, the hammer bonus is less helpful (it helps them too) but it's still a boon.
I can see how it might be easier to establish a new religious community in a big city than in some backwater, and the population factor also means that spreading multiple religions tends to become a bit easier as the game progresses, but, on the other hand, converting a big city is more powerful, so one wouldn't expect it to be easier.
The game having no religion % bar like it has for culture makes religious spread very black-and-white.
I think that Musketman largely being skipped is the main reason why the impact of Rifling on the revolt chance is so noticeable. I had floated the idea of giving Musketman a first strike (it's initially contemporary with Maceman, Knight, Pikeman, which have a shorter range), +25% vs. Melee, reducing its production cost by 5 and allowing an upgrade from Crossbow, but I don't think anyone voiced support for that. Liberalism is part of the problem too – gets players on the Paper-Education path to Gunpowder (which is then just 3 techs away from Rifling - via Replaceable Parts and Printing Press, which can be chosen as the free tech) instead of the shorter path via Guilds.
I still think that revolts are too easy to suppress post-Rifling. But regarding Musketman you are right it appears skipped. None of the AI civilization around me seem to have made them. With barely more strength than a Maceman and not any kind of bonus for special matchups (the civilopedia talks about city defense but it is only competent if using the right promotions), it is not very attractive.
It should arguably still lose to knights 1-on-1 when no defensive bonus is involved, but some kind of buff would make sense.
I've never tested that with the mod. Good to know that it works. I guess it ought to because cities on different continents can be founded just two tiles apart in any case. Generally, I'd be worried that Creative leaders could settle aggressively to steal resources even from the inner ring of rival cities.
My own reasoning for enabling this 2-tiles apart city founding is mostly for contested areas, when two civs are vying to control the same area. Civ4 already prevents founding a city in another civilization's cultural borders, even when at war, so this opportunity for more cities usually happens if there is a huge cultural advantage from behind or if the rival city still has no culture.This leads to a culture war (and if the other city defects after revolts it can be removed from the map). It's true that this may be quite powerful when a Creative leader is involved.
I've noticed though with this rule change that sometimes the AI would create a second city only two tiles apart from another it owns.
I looked through your manual and you mention a number of changes regarding how the AI picks spots to found cities on, improving it significantly over vanilla/K-Mod. It was interesting.
Looking at how the AI chooses city spots to settle, I have the following impressions:
- It only considers how good this one spot is. Frequently, it makes sense to sacrifice one ideal spot if it allows to have two good cities instead of one good and one terrible. Obviously considering how one spot might affect future spots is more involved code-wise.
- It has no concept of "unused good tiles". For example in my game I can see a place where the AI could create a city to exploit 9 more tiles (6 hills and 3 flat). With biology-boosted farms and taking two or three tiles from a neighbouring city it could make a great additional production city. This is a case where the overlapping tiles and the lack of bonus resources might be weighted too highly.
- It still has some trouble with properly evaluating overlapping tiles. Some of the cases where cities are founded two tiles apart are more sensible than they seem at first glance, but others not so much , and some 3-tiles apart placements are also not so satisfying
- Cities just one tile away from the coast are rare and some are justified, but there are still instances where the choice is puzzling.
- The AI doesn't take advantage of rivers as much as a human would. Levee only happen later in the game, but it really pays off, provided the city is on the river and has several river tiles to use.
Have you tried having old-AI vs new-AI autoplay tests to evaluate if an AI-related change makes it play stronger?
I could try doing code changes for this but I don't have the build environment for it and it would require validation that it actually produces stronger play.
It's not just a cap, I look at it as the probability of a revolt attempt – because it gets multiplied by another probability which is based on foreign culture strength and garrison strength. This second roll is the revolt suppression roll in my mind.
Yes, which is why it also increases revolt probability even with some garrisoned troops (which in the riflemen era was positive).
However I tested it later in the early game, and when you've just got archers the revolt probabilities are massive with this tweak I made. I'm convinced increasing the revolt probabilities with techs that improve the available defensive units up to rifles is the best way to balance revolt across different time period.
For some familiar point of reference in the real world, in the Ukranian oblast of Kharkiv, in which about half of the inhabitants are native speakers of Ukranian, an independent state was proclaimed in 2014, but "the uprising was quelled in less than two days due to rapid reaction of the Ukrainian security forces."
Wikipedia
In contrast, parts of Donetsk and Luhansk, with some 25-30% native speakers did revolt (with foreign military support). Well, this doesn't indicate what should happen at 40%.
Interesting point of reference.
I also had spies in that city for a long time doing some missions, but I got the impression that instigating a revolt only prevents the city from working until the revolt is quelled, without contributing to the "flip city" counter.
I think it would be interesting that, if you have a cultural majority in the city, instigating a revolt with a spy could contribute to the "flip city" counter. Although that might be overpowered without other tweaks.
It also seems that an ongoing revolt prevents a new revolt from happening, and so that failing to repress the first revolt will prevent a second one from happening. While a city in revolt is not productive, purposefully not repressing a revolt might sometimes prevent an otherwise unavoidable flip.
The calculation of the foreign culture strength is unfortunately pretty complicated, specifically how it combines the era (of the city owner and the cultural owner), ownership duration, current population, highest past population and tile culture in the city and in the surrounding tiles. It was already complicated in BtS; I've made it worse, but managed to simplify the garrison strength calculation so that players at least know how many units they need for revolt suppression
I have the impression from my games that foreign cultures reach a higher percentage in city tiles than in neighbouring non-city tiles. Is that just a coincidence from very-long-distance influence from other cities?
I know the cultural reach of high culture cities goes beyond their borders but not how much beyond and if it matters here.
Considering that the increased maintenance doesn't become a major factor until the midgame (and, depending on how much the player is able to expand, perhaps not even then), I don't think those two buildings are, say, breaking the game. At a point when the game is still undecided, the total city maintenance cost of a large civ on a Standard-size map seems to be something like 100 or 150. On a Huge map, it might be twice as much.
On a huge map you tend to have more cities (higher "number of cities" maintenance) and by necessity they are more distant from the capital (higher "distance to capital" maintenance). I'd estimate the cost difference between standard and huge to be quite more than double.
Also, if the game is still undecided by the time corporations kick in, the ability to get extremely cheap corps is nothing to overlook.
I think my takeaway on these is tied to my preference for huge maps
I agree that it would be better to weaken those two buildings a bit, but that seems difficult to do without making them look uninteresting.
Changing their design would be too difficult, but just tweaking the numbers to say 15%/70% instead of 20%/75% should make them still attractive yet less dominating?
Generally, I agree that one won't go too wrong by just opening borders with everyone. And, on a Huge map with multiple continents, I also agree that Mercantilism isn't going to do much good. It's better when there are few civs to trade with. That could also happen in the endgame when most civs have either become vassals (vassal-master trade is exempt from Mercantilism) or have been eliminated or marginalized – and when just two hostile super powers remain. Imo giving the civic such a powerful drawback was a bad idea. In the full balance overhaul that I've been musing about (without seriously considering to implement it), my idea for Mercantilism was to apply the trade restriction only to rivals, i.e. "rivals that trade with us receive no extra yield for sustained peace" – and replace the specialist ability with something less weird (who are these people who don't consume food?) and less powerful.
K-Mod has already reduced the upkeep class of Mercantilism; I don't think there's some small, elegant tweak to make Mercantilism into more than a niche civic. Nerfing Free Market would be easier, but, as you say, Mercantilism often can't even compete with Decentralization.
Mercantilism hurting foreign trade makes some sense considering what the real-world policy involved, but it should have been something like reducing the foreign trade gold bonus instead of forbidding all foreign trade. It seems hard indeed to balance it properly.
You're right, this wasn't a well-considered assessment on my part. The consensus among the veteran players on the forum is that founding an early religion is virtually never a good play at a challenging difficulty level. Ideally, founding an early religion should have a higher expected reward and a lower variance than it currently does. (The mod does make small improvements toward these goals: The AI is a bit slower to found the early religions, so there's a lower risk of being beaten to one, and the increased importance of culture makes the early culture boost from a Holy City more worthwhile.)
Maybe if the AI gets to Biology pretty early despite having rather few cities (low maintenance), such an unremarkable city could pay off for the AI and even make a significant contribution toward a Space victory.
