The Eternal Peace

Fins

Warlord
Joined
Jun 25, 2014
Messages
191
For many years, i was trying to play a Civilization game of


"Eternal Peace": a game which could last "forever" and remain peaceful - the opposite of the famous Eternal War of Civ2.


I.e., to set up a civilization, and see it live on and on, without hostilities with other nations. Hopefully, eventually with no hostilities between any AI nations, too. A game with no "winner", but equally, with no "losers" as well.

//Edit December 1st, 2024: I think, i finally managed to have it. I did it! :D So, i'm adding this game's final report under 1st spoiler, here. There are some graphs, screenshots and a few stories. The remainder of the original post - is now under the 2nd spoiler. //end of edit
Spoiler Eternal Peace - final report :
The year is now 12777 AD. There has been no war for the last 10,000 years - ever since last russian city was captured by India, in 2777 AD. And the situation is such that I am completely sure that this peace - would indeed last forever, if I'd continue to play. Things... "have settled". :)

A lot have happened in this game. In-game counter says that I played this game for 526h 34m, when I made the last save in 12,777 AD. So, 1st, here's a summary of the most significant events of this game.

The game started on a huge-sized "Ice Age" map, which generated two huge continents and one tiny single-tile island. Early in the game, I developed my american nation as usual, settling out much of the middle part of the eastern continent. I kept my army large enough for no AIs to try doing a war against me, and I happily kept minding my own peaceful business, prioritizing sciences and culture, but not forgetting religions, too.

The game was played on Warlord difficulty, which is way easier than I usually prefer - but Warlord has one deciding difference from any other difficulty mode: it has no decay of diplomatic relations with AIs. Higher difficulties - do. This means, for any extremely long game, all AIs could end up hating me for absolutely no in-game reason. I didn't want that to happen, just because such a mechanic looks rather silly, to me. It's not silly for a normal-length game, when it's slow and reasonable enough, but for this one, it had to be "Warlord". And so, naturally, I developed quite ahead of "historical" pace; for example, I built United Nations in 1812 AD - some ~140sh years before it happened in real world. 1st resolution it made - signed by all nations - was banning nuclear weapons. It stayed in power forever since. Nobody even had time to build Manhattan Project. Perhaps, our real world would be in better shape if that's how it'd happen in reality? Just wondering...

Anyway. By the game's 20th century, I was far ahead of all AIs, tech-wise. When my army was already fully switched to gunships and modern armor, the AIs were still busy upgrading from catapults to artillery and from cavalry to 1st few primitive tanks. And so, it was a big-time surprise when Montezuma declared war on me in 2004 AD. He was on the other continent, too - but he sent a small army to try and attack Los Angeles - my only port city on the western shore, at the time. He was still using some sail ships, though - which my subs and destroyers sank with utter ease. I easily destroyed his invading force as well, and then sent my own, across the ocean, to pacify him by force. In 2015 AD, I captured 1st of his cities. He kept throwing legions of infantry, riflemen, granadiers and cavalry at my invading force, but technological superiority was too high: stealth bombers, jets, some artillery, groups of modern armor units and small swarms of gunships rolled through his territories and cities with ease. Which were big territories - his empire occupied much of the middle portion of this 2nd continent. He was distinctly 2nd-largest nation (after me) at the time, by land mass.

While waging that one-and-only war I did in this game, I had two reasons to capture Montezuma's cities: 1st, it's more human to capture than to burn 'em; and 2nd, since I had to pacify him anyway to prevent future aggression, by reducing his empire to small enough size - why not to take the opportunity to establish my presense right in the center of the 2nd continent? Before this war, I was fully expecting to let the other continent go on all by itself, without my presense - to see how it goes. But such an opportunity - was just stupid to let go.

In 2072 AD, when Montezuma had only two small cities left - I accepted his peace offer. Now I was sure he won't try to break peace again. His neighbours soon made sure he won't try anything at all, too, by completely wiping out his nation. Oh well. He started it himself, so he can blame himself for all eternity while he burns in hell for his sins, i guess...

Then, other nations kept duking it out against each other for good 6 more centuries. Some died, some survived despite some heavy losses, and some greatly gained from those wars - but most of the time, it was useless waste of lives and efforts. I took my lesson from Montezuma's attack: built lots more military units pronto, to increase my empire's Power way further and scare away any wanna-be-next-agressor. It worked: nobody else declared war on me, ever since. Alas, for a few centuries, I had quite an arms race with the most militaristic of remaining nations - Rome. They kept building lots of army, but so did I. In the end, I won that race, and remained the largest military power ever since.

Meanwhile, in 2155 AD, Global Warming have started. Even while my americans didn't build a single Coal Plant, nor any Nuclear Plant, - other nations did, and it caused Global Warming. For over 6 thousands years, Global Warming was raging through the planet, desertifying and killing all in its path, until very last tile of fertile soil succumbed to it in 8238 AD. From one very lush world, it have become one complete sand and ice desert, with a bit of occasional tundra here and there. Grim place.

Yet in 2155 AD, when Global Warming was only starting - all the other nations kept fighting among themselves. I was focusing most of my empire to spreading culture and five religions, monasteries of which I had built in some of my cities. I kept going this way, with my empire gaining more cities peacefully via cultural expansion, ever since. Other empires eventually killed many enough of themselves for the survivors to be so far from each other, and in one case two of them so friendly to each other, that none of them saw any point in further war. This happened in 2077 AD, when Russia was destroyed by India; the last nation to be defeated. And then, 10 millenia of peace - was this game's world next chapter.

There were six surviving nations other than my empire, by then.

Two of those - Egypt and Incan empire - by 2777 AD were so weakened by wars and so pressed by my cultural expansion that they had only 3 cities remaining, each, and very little territory. Soon enough, their cities got completely surrounded by my empire's territory, becoming 1-tile outposts. Some of their own populations later revolted, pushing three of these six cities into my empire. For many thousands years since that, Egypt remained single-city state, and Incs survived in just two of their remaining single-tile cities. They keep large enough armies in each city to prevent any chance of revolts, though. Pretty nowhere else to go for all these units, that is. Overall, those two nations became the "insignificant outsiders".

Then, there were two "middlemen" of those 6 survivors: England and Persia. Big-time friends (defense pacts and all) to each other, they shared a border in the northern part of the eastern continent. Relatively small, but still significant, they resisted my cultural advancements for roughly 10 thousands years. Huge garrisons to prevent revolts, firm religious stance of England, relatively powerful culture of Persia, and some of their cities happened to be just perfectly placed to make my cultural advancement problematic - were obstacles. But more than those, they remained significant force for many millenia simply because I had to prioritize most of my cultural-expansion efforts to the couple of largest remaining AI nations: India and Rome.

Rome was the largest survivor of AI wars, occupying roughly 1/8th of world's land area at its peak. And India was just a little smaller than that. I knew that the longer these nations keep producing their own culture in most of those lands, the longer it'll take later on to change those lands' culture into prevailingly mine. So i had to hurry and culturally push as far into those lands as i could, ASAP. That's why i focused on those two 1st.

Long story short - this plan worked perfectly. Indeed, now in 12,777 AD, both India and Rome are practically consumed by my culture. Many of their weaker-garrison cities revolted and joined my empire, and almost all remaining cities - have zero workable tiles, being surrounded by my territory completely. Yet quite before it happened, I started to push more and more into Persia and England, too - devoting more and more Great Artists into advancements there, building culture-boosting things like Cathedrals in cities near their borders, etc. And so, by 12,777 AD, those two also ended up having just a few 1-tile cities, each: just two cities Persia, and four cities England.

The "end result" - and the reason why i count 12,777 AD as this game's practical "completion" - is that all six surviving AI nations are so outmatched militarily by my army, so separated and blockaded in terms of territory, plus so utterly lacking resources, that I see no chance whatsoever that any of them would ever try to make war - not with me, nor between any of themselves.

And this is how this game have indeed created a state of Eternal Peace. Mission - complete! :)

And here are some screenshots from 12,777 AD, to complete the report.

Score graph + Egypt's only surviving city:
12777 - score, Egypt.jpg

Note: I played most of this game while using Unversal Suffrage civic and 0% science expenses (after getting to Future Tech 1) - while Rome was doing Representation and 80% science expenses. So, when I switched to Representation myself, some time in 11,700sh AD - 1st major increase of score gaining speed occured, for me; and when I switched from 100% culture to 100% science expenses in 12,000 AD, plus changed from Artists and Merchants to Scientists in most cities - 2nd huge score gaining speed boost happened. I could leave all the AIs in the complete dust score-wise, and equally, I would, if I'd keep playing for few more millenia - because now, in 12,777 AD, I don't need utterly maxed-out culture anymore (other than in very few cities which could still capture few more tiles from little remains of AI territories). Thus, the game's "score" - is not a good indicator of how "good" any very-long-lasting game is going, sadly. Too biased towards science. :(

Demographics + both surviving Incan and one Indian cities:
12777 - Demographics, both incan and one indian cities.jpg

Note: funny thing how Egypt "population" is just 1000, but smallest army of all AIs - is 1,182,000 soldiers! :D I even see some egyptian troops in other AIs' cities, ones which Egypt maintained Open Borders with - yep, for 10k years. Curiously, I found that other-nations' troops in a city (like, Egypt troops in an Indian city) - have no effect on revolt chance in such a city, when my culture seeps in. Which means that it helps to keep close borders during cultural expansions - separate AIs' troops into each other's cities, make those troops not-a-factor for revolt chances. Anyhow, ever since 2777 AD, all the huge armies of previously-fighting nations - remain pretty much unchanged. Basically, their territory shrinked, but all the soldiers - remained as numerous as they were back when those AIs had the territories, civilian population and food supply to support them. Also, Life Expectancy - is capped in the game: it can't get above 100, no matter what. Got pretty close to that, eh. :)

Crop Yield + two english cities:
12777 - crop yield, two english cities.jpg

Note: after initial (before 2000 AD) expansion, I did two big further expansions: 1st after 4200 AD, when I built about a dozen small cities in various "gaps" to intensify cultural pressure along my borders; and 2nd expansion was done after 7500 AD, when, with Global Warming nearly complete, I settled most of previously-unsettled parts of shore lines of my empire (within my own borders). Those two expansions - are the two steep "bumps up" on the graph. Other than that, things evolved pretty gradualy, bit by bit. This graph directly shows Global Warming impact - but not full scale of it. Because during ~6 thousands years it took Global Warming to do full scale of its damage - my empire massively expanded both in terms of its territory and in terms of city density within its borders. Also, Coventry city on this screenshot - was the 2nd-last (till 12777 AD) AI city which joined my empire, which is why some of lands near it are still english lands. Which no english city can work anymore, and which lands would soon - several more centuries - also flip into my territory.

Power graph + both surviving persian cities:
12777 - Power, both persian cities.jpg

Note: this is how 10,000 years of complete peace look like, yep. Very stable! :) All the wars before 2777 AD, with all those near-vertical drops and gains - were but a tiny early accident, in comparison to sheer length of 10 further millenia. Gradual increase of my own power (blue line) - was happening from free mechanized infantry units which I was getting when AI cities were joining my empire, and from building more barracks and such. I.e., I built no new military units for 10,000 years, yep! And it seems that AIs did the same. Armies big enough as it is, no wars going on, nobody wants to start one - so why build more army, right? And so, nobody did. :) Also, you probably noticed one unusual thing on this graph - all those small but sudden, near-vertical drops on the pink (India), light-blue (Persia) and once on orange (Incs) lines: those are, I'm 99.9% sure, nothing else than Nuclear Meltdown events. Each time one of those happened in any heavily-garrisoned city (and some AIs had well over 50 units in some cities, in this game), - good bunch of units died, and that Power line of that civ got that relatively little, but distinct and instant dip down few notches. I've observed AI meltdowns during the game a few times, directly (there's no message about any meltdown in an AI city, so it's easy to miss), and also, I know that those are meltdowns because it's exactly India and Persia who had access to Uranium for any long time (required to build Nuclear Plants), and it's equally India and Persia both who built quite many Nuclear Plants in their cities, too: my spies saw every last one of 'em when visiting their cities. And one more pretty grim thing: even after India lost its only source of Uranium to my cultural expansion, for many thousands years more, Nuclear Plants in indian cities - kept popping up with those meltdowns. I think, one or maybe even two of indian cities - still have Nuclear Plants even now, in 12777 AD. So those are hella long-term "ticking bombs" for sure.

My army, and largest remaining AI-controlled single area (Delhi with 4 workable coast tiles):
12777 - my army, and largest remaining single area of an AI player.jpg

Note: vast majority of these units - never seen any combat. Especially the fleet: the war with Montezuma had no significant sea battles at all. Also, don't be surprised about disproportionally huge number of motorized infantry units: this is because in this game, I culturally-absorbed nearly 40 cities into my empire, and "dismissed" nearly as many on top of that. "Dismissed" - when culturally flipped, there's an option to dismiss a revolting AI city instead of making it your own, and there are good reasons to dismiss many of such cities, especially when playing strategically with further cultural advancements in mind. City positioning - is key, and AI cities are often in way too messy places. But i digress; back to army. With some 1...4 "free" mechanized infantry units given upon each of ~80 cultural city flips i had, I ended up with ~200 extra Motorized Infantry, you see. And I did not dismiss any of those, too: no reason to, because in this inflation-free game, my empire was having ample ~1k income ever since 3rd millenium - even while running 100% commerce into Culture at all times. As for overall purpose - most of my army was built preparing for unexpected, but still always anticipated, agression from AIs. I created two specialized cities which were dedicated to building all these units fast and well: one (Portland) with national wonders which provide +4 exp and free Medic upgrade, thus dishing out level 4 (10 exp points total upon creation) land units and fleet; and the other (Seattle) had +200% military unit building speed from Heroic Epic and Ironworks, resulting in so high production (before GW) that it was able to spit out one plane every turn - alternating between fighters and bombers. As it had exactly enough production for that. Mighty efficient combo! And yes, my army composition is bomber-heavy - on purpose: I prefer the birds to artillery, especially in vanilla Civ4. Even while bombers can not weaken enemy units more than to 50% health - nothing is as handy as couple dozens stealth bombers covering all aproaches possible, with overpowering numbers of Bombers plus solid Jet Fighter cover, of course. Indeed, modern warfare - is all about air superiority.

Largest AI army (Rome's), and largest remaining bits of Roman empire:
12777 - Rome.jpg

Note: at some point, I was quite amuzed to see this; at some point well after I was done shaping up my own military for my own reasons, that is. Because, as you can see - there are some striking similarities and striking differences:
- both he and I ended up with very similar number of modern armor, gunships, jet fighters and navy ships (alas I think my "heavier" proportion of battleships - is better);
- even our number of transports is exactly the same - 11! %) And yet,
- he's slacking on stealth bombers, big-time, instead investing massively into artillery. Which is pretty inferiour, I think;
- I maintained token numbers of infantry units, mainly upgraded from earlier-eras, while good part of his army - is non-motorized infantry. Slow is bad, even when defending, but he's still using 'em.
And yep, that's population 7 roman city near the bottom - largest AI city remaining on the map. But it does not work any tiles - all of its population is instead supported by Great Merchant specialists, each providing +1 food to the city. Even in that, Rome well surpassed all other AI nations. Good old Rome... :)

And for the last screenshot of this report - here' Sacramento, the largest of my cities for thousands years, and a true crown jewel of my 125-cities' large, all deserts-and-seas, empire:
12777 - Sacramento.jpg

It just happened to have this heart-shaped working area. I was kinda stunned when I realized what it looks like, yep. %) Next, in the bottom-right corner, you may notice the number of "great people points" required to generate the next great person - 128,000 points. With a regular desert and/or ocean city, even under Pacifism, generating some 6...24 points per turn - it takes some 5...20 millenia, now, for any usual city to fill the bar. Wild. ;) Also, this Sacramento completely dominates Karakorum city - partially visible 3 clicks to the south, - which is one of 11 surviving roman cities. And about Sacramento's size - sure, I could put many more Great Merchants into it, thus increasing its size much further, but 21 population and 42 food - is just so right. Why 42? Because 42! The answer to everything, you know. :lol: And if you don't know - well, google "42" maybe. 42! :D
Spoiler The remainder of the original post :

This time, i am playing it in Civilization 4 (1.7.4.0, Civ version 174, Final Release). Currently, it's the year 2613 AD. And things are all good, so far.

The game is base Civilization 4 (no BtS units) with no mods and only one alteration of XML files: I removed inflation (set to 0). Obviously, inflation was never meant to remain a manageable feature over many thousands years post-year-2000AD gameplay. No choice but to remove it, for this kind of a game. Fortunately, it was easy to do.

Everything else is vanilla. Including global warming, which have started by 2155 AD. My nation did everything right about it: no coal plants built whatsoever, no nuclear plants whatsoever, and I've led the UN in banning nuclear proliferation long before any nation could even start Manhattan project. And it stays banned ever since. As I now have the majority of UN votes, it will remain banned forever, too. However, other nations on the map - used coal and nuclear power plants (amazingly, the latter does not require Manhattan project; "wow", yeah). And of course, while i saved lots of forests - other nations were not so eco-friendly. Global warming - have started.

I plan to endure it. Even if all land become desert, and most of my people will die to it - the survivors will live on. My empire became so large, and its economy efficiency so good, that i'm sure i'll be able to enforce long-term peace even in complete-desert world. Good old Arakis would be hella jealous, yep! :)


But, i have three questions, which is one of three things i made this topic for:

1. is it possible for global warming to stop happening, provided that no nuclear weapons were ever detonated, and I'm running recycling centers in every city I own? The pace of global warming, so far, is rather slow. Could it slow down further and eventually stop, if sufficiently much of world's forests would grow back?

// I've read that 40% of the land masses covered by forests - prevents global warming completely, in Civilization 4. But would it work retroactively, if I'll be lucky to have that much forest to regrow - before most of the world turns into deserts? So far, after 458 years of global warming, way less than 10% of all land was turned into deserts, and something like 10%...15% of all land is still covered by forests (mostly in tundra belts and inside my empire's originally-settled territories). New forests grow up few times less often than new deserts appear, though. It's probably too late. :( But i'll try anyway.

2. can I speed up reforestation (which happens by new forests popping up on their own, on non-improved tiles capable of supporting forests), if I remove / plunder non-road improvements from tiles which are not being worked by any city?

3.
is there a way in vanilla Civilization 4 to deconstruct buildings?

// I captured many cities from other nations via culture spread, and I will capture even more. Even by 2613, way more than half of each of the both continents - are now under my banner. But, many of those AI-built cities have coal plants, and some - nuclear plants. I found no way to remove them. Don't want no mods just for this, though.


Another thing I made this thread for - I hope to keep adding to it, documenting my Eternal Peace story as I go.

And the third thing - is to answer any questions about this game style and approach, which I do - "Eternal Peace". If there will be any. Anything like which settings I used, how I prevented wars and deterioration of diplomatic relations, details of cultural expansion I practice, how I defended myself when good old Montezuma declared war on me in 2004AD, etc. I'll be happy to share all details I can.


P.S. In my many previous Eternal Peace attempts in other Civilization games - from Civ2, through to Civilization: Beyond Earth, then in FreeCiv and some other fan-made games, as well, - I always eventually failed to achieve it: sometimes due to some in-game feature like out-of-control inflation, sometimes it was unavoidable game crash beyond certain specific date (shame - Civ:BE could otherwise be it!), sometimes it was limited game length (one of titles, i remember, has hard-coded limit of 3500AD), etc. But this time, I hope this is the one. I hope to one day reach 10,000AD - and I hope to go beyond even that. Wish me luck! :)

P.P.S. Being back to playing Civ4 lately, after playing all the other games of Civilization series, I now developed the feeling that Civ4 is the most true one of them all - true to the spirit of the whole idea of Civilization game; the idea as it was created all the way back in Civilization 1. Which I also played, way back in late 1990s. Played normally - not the "Eternal Peace way". Great memories! But, I wonder: is there anyone else here who have this same feeling? Late stuff like Civ5 and especially Civ6, in particular, just "don't click" with me...
P.S. Si Vis Pacem - Para Bellum! :smoke:
 
Last edited:
Fun project! Except the global warming part. I hate that mechanic; I don't know the exact details but I never seem to be able to do anything about it other than end the game before it becomes a problem. Back in Civ 3, if I kept my pollution well in hand and didn't do anything ridiculous, I could stave it off indefinitely. In Civ 4 it doesn't seem to matter even if I'm running clean energy everywhere, the global warming just keeps coming. And there's no strategy to it either, it just comes and says "Boom, say goodbye to this tile!" and you just have to deal with it. Incredibly obnoxious.

I am no expert as to the nitty gritty details, but from my experience, I'd answer your questions as follows:
  1. I don't think so
  2. I am >90% certain that forests do not grow on roaded tiles - they need to be completely bare
  3. I don't think so
 
Thanks!

Indeed i am no expert also, but i know some few peculiar details about it. And i have some good news for you! :)

1st, it _always_ matters if we run clean energy or not, in Civ4, because the chance of a tile becoming a desert is calculated every turn roughly this way: 1st, check if more than 40% of all land tiles are forested, and if so, then global warming just can't happen; 2nd, if less than 40% of the world is forests, then check how much pollution is produced in the world, and if it's below certain treshold (if memory serves, a total of something like ~100 "green faces" from all cities combined) - then global warming just can't happen; and only 3rd, if global pollution is above that treshold - then generate random number and compare it to the amount of world pollution above said pollution treshold. Which means, the higher global pollution is - the more chance there is that global warming hits will happen more and more often. Which means, if your nation produces 0 pollution, then as soon as global warming starts, the whole world is hit by it way less often; and so, your own lands are hit by global warming way less often, too.

2nd, i believe that there should be at least two strategies to defeat already-started global warming in Civ4. In short:

- 1st strategy, pretty late-game: prevent any nuke usage by banning it via UN, and keep it banned; run your empire clean; conquer all but very few cities in the world, and install recycling centers there ASAP. The goal is to reduce total world pollution to the level below above-mentioned treshold, which should reduce per-turn chance of global warming events back to 0;

- 2nd strategy, early-game, only viable on initially much-forested maps (though, those are many): do early expansion and conquer most of the land before AIs should chop any many forests; keep the forests. As long as 40% of more of the world are covered by forests and/or jungle, any amount of pollution and nuke usage won't result in any global warming.

Also, thanks for trying to answer my questions. And again - good news: i know that forests can regrow on tiles with roads / railroads, yet no production-boosting other improvements. I've seen it happen, more than once, in my current game! :) I've seen, repeatedly, how "forest has grown near city" notification, generated among others, creates a temporary icon when observing the area with such a tile, and that icon pointed exactly to now-forested tile with a railroad on it. :)
 
Roaded tiles have half the chance of forest/jungle regrow vs empty tiles.
 
iirc Recycling Centers unintuitively don't help with Global Warming (that's something I vaguely remember reading a long time ago, so it would need further investigating before you take my word for it).

You can speed up forest growth by buillding Forest Preserves on them. They can grow back on roaded tiles, but roads halve the chance of that happening (or is it 1/4 chance?).

You don't need to remove Coal Plants: they will stop polluting (and be rendered inactive) if you build a "replacement" power plant (best are hydro plants, but Nuclear can do as well, although there's the meltdown risk to consider...).
Unfortunately, Factories themselves are there to stay...
You could try using spies to sabotage Coal mines and Oil wells? (Sea oil would be an issue though... not sure Privateers could live long enough to pillage them in a modern world).
 
Roaded tiles have half the chance of forest/jungle regrow vs empty tiles.
Oh, i never knew this one. Thanks! :thumbsup: Off i go to switch auto tradework building on all workers. But it seems, there is no way to remove already-built railroads? My gunships pillage other improvements inside my territory just fine, even city ruins, but "Pillage" button just disappears when it's only a railroad left.

iirc Recycling Centers unintuitively don't help with Global Warming (that's something I vaguely remember reading a long time ago, so it would need further investigating before you take my word for it).

You can speed up forest growth by buillding Forest Preserves on them. They can grow back on roaded tiles, but roads halve the chance of that happening (or is it 1/4 chance?).

You don't need to remove Coal Plants: they will stop polluting (and be rendered inactive) if you build a "replacement" power plant (best are hydro plants, but Nuclear can do as well, although there's the meltdown risk to consider...).
Unfortunately, Factories themselves are there to stay...
You could try using spies to sabotage Coal mines and Oil wells? (Sea oil would be an issue though... not sure Privateers could live long enough to pillage them in a modern world).
I vaguely remember something like that about recycling centers too, but i build 'em anyway. Can't hurt...

Forest Preserves are not available: it's one of BtS features, and like i said in 1st post, i play vanilla Civ4 (no BtS). I like it simple. Plus, it's a long, long game, which means all AIs are fully developed. So, in BtS, it'd be cruise missiles, attack subs, AI _spies_... :cringe: Still, thank you for trying to help! in BtS, this would definitely be a must-do.

Much interesting about Coal Plants! My culture captured about 2 dozen or so AI cities so far, and many of them have Coal Plants, no access to hydro power, most of them are on another continent (so no free power from the wonder), and in most cases, AIs did not build a Nuclear Plant, too. Should i do it? I've read varying reports about nuclear plants - some players get many meltdowns per game, others - none and never. Today, i reached 2700AD in my game, and i did not have a single meltdown yet. Quite a few of cities i captured from AI - have both Coal and Nuclear plants in them. So, i probably should add Nuclear Plants to cities with only Coal Plants, given all the above, right? What you think?

And no, can't use spies to sabotage 'em. That's another BtS feature - in vanilla Civ4, the only thing my spy can do in an AI city - is to "steal plans", which makes all units and cities of that AI visible. Same for privateers, too: also BtS unit, so i can't use 'em.

That said, i guess there is - technically - a way to remove Coal Plants (as well as nuclear plants) from the world, in vanilla Civ4: have a war with an AI, let it capture a city with a Coal Plant (or anything else undesirable), then recapture the city and burn it. Does the job, right? :mischief: But sadly, for my Eternal Peace game, even that is not an option. Because, naturally, i play it peacefully - which means i never declare any war on any AI, fighting only if they declared war on me; and i never raze any AI city (except one single barbarian city very early in the game, but barbarians can't be reasoned with and never do peace treaties, so it had to be done), too.

P.S. Guess it will end up being Arakis-like world in my game, after few more millenia at this pace - sand everywhere. Well, it's kinda appropriate, too: global warming is where we go in real world, and indeed, there's lots of talk but little action to shut all those Coal Plants down. Nope, they keep building new ones even as we speak, eh. :sad:
 
I'm learning a lot from this thread!
Happy to help!

And to this end, I figured out a guess to the mystery of Nuclear Plant meltdowns, while playing my game. Which mystery was mentioned above: why some players get a lot of meltdowns every game, while other players never get them even through multiple games, not even once?

I suspect that meltdowns of Nuclear Plants can only happen are times more likely in cities which have a Nuclear Plant but do NOT have a Coal Plant nor a Hydro Plant. Here's why:
Spoiler deets :

- reason #1: I've started to get meltdowns in my game - 1st happened some time in 2700s, 2nd happened just now in 2991AD. Thing is, in both cases those cities did not have any other power plant in them - only Nuclear Plants were installed;

- reason #2: if this is indeed how meltdowns work, then it would explain perfectly why some players get many meltdowns, and some get none: simply, some players have a habit to build Coal and/or Hydro plants in all their cities before they discover the tech to build Nuclear Plants, while other players beeline to nuclear tech (possibly, for offensive reasons) and skip building Coal Plants. With Hydro Plants being often not available in any given city, these players end up with many cities running Nuclear Plants only, and then they start to get meltdowns;

- reasons #3: it'd be logical. Real-world nuclear power plants - actually require some external power from a country's power grid to function properly, and that external power is usually generated exactly by coal and/or hydro power plants capable of "on-demand" power generation.

After browsing all my cities, I found that among all my cities which were culturally captured from AIs, I have the total of 6 cities with Nuclear Plants, but only 2 of those cities also have a Coal Plant in them. Zero nuclear plants in all cities I built myself. For now, I reloaded an earlier save in both cases of those meltdowns, which keeps meltdown chances maximized during future turns. So far, the chance for both meltdowns to happen in 4 cities without a Coal Plant (out of 6 cities total) - was this: (4/6)^2 = 44.4%. Not any conclusive, yet. But already "slightly confirming" my suspicion.

I'll keep observing how meltdowns will be happening as I go, and pick up more statistics about it. And also, there is one other thing about meltdowns, which I now know for sure: free power from the Three Gorges Dam wonder (free power to every city on the continent) - does not prevent meltdown of a Nuclear Plant, as one of meltdowns happened in my game - was in a city which had that free power.

Edits:

3rd meltdown happened 3241 AD, again in a city without any other than Nuclear power plant. For 3 in a row, odds for that to just be a coincedence: (4/6)^3 = 29.6%;
4th meltdown by 3384 AD, same; odds for 4 in a row: (4/6)^4 = 19.8%;
5th meltdown by 3449 AD, same; odds for 5: (4/6)^5 = 13.2%;
6th meltdown by 3537 AD, still the same - always in one of cities without Coal nor Hydro plants. Odds for 6 in a row: (4/6)^6 = 8.8%;
7th meltdown by 4071 AD broke the pattern: it happened in a city which had both Nuclear and Coal power plants.

And most interestingly, preventing Nuclear Plant meltdowns by having a Coal Plant in the same city - seems to work (by reducing the odds of a meltdown occuring) for the player, but not for AIs. Very 1st nuclear meltdown in an AI city i've noticed, some time early in 3500sh AD - was in a city which had a Coal Plant few turns prior to the meltdown, and it even kept the Coal Plant after the meltdown, too. So it wasn't that that AI sold the Coal Plant being short on money and _then_ had a meltdown - the Coal Plant survived the blast. And of course, nukes were always banned and Manhattan Project was never built, too. So it wasn't a missile hit - it was a meltdown 100%. Which, basically, means that if any AI is allowed to reach the tech required to build Nuclear Plants, and then to build some, and allowed to keep cities with those long enough - then in the long run, global warming is unstoppable due to meltdowns in AI cities. Except if over 40% of the world is forests / jungles, that is. Fat chance for that late-game, though...
Edits: updating per later discoveries in my game.
 
Last edited:
P.S. Guess it will end up being Arakis-like world in my game, after few more millenia at this pace - sand everywhere. Well, it's kinda appropriate, too: global warming is where we go in real world, and indeed, there's lots of talk but little action to shut all those Coal Plants down. Nope, they keep building new ones even as we speak, eh. :sad:
Not to get too far OT, but real-world global warming doesn't turn everything into deserts; it dumps rain in some places making other places deserts. And some places alternate between getting too much and too little rain.
 
Not to get too far OT, but real-world global warming doesn't turn everything into deserts; it dumps rain in some places making other places deserts. And some places alternate between getting too much and too little rain.
For now, sure. But if it gets far enough, everything properly turns into deserts - exactly like it happened on Venus: all water evaporates, whole planet becomes arid and toxic desert. Even oceans. Fortunately, Earth is too far for that to happen - Stefan-Boltzman law is our savior. I think. I _think_... Back to the topic, though.

I made two properly amazing discoveries about nuclear meltdowns, in the game!

Discovery 1: Nuclear Plant meltdowns' effect of massive increase of global warming - is temporary. It lasts ~200 years, is particularly severe during 1st 100 years after a meltdown - but then, it disappears.

Spoiler Proof (with graphs): :

Below are two graphs proving it, from my game: 1st is 2000 turns of crop yield, and 2nd is 2000 turns of power. Both from 2200 AD to 4200 AD:

Crop yield is one direct indication of how quickly global warming kills food-producing tiles into desert. On the graph, you can see that is was not dropping steadily. Indeed, I had centuries with more than 1 global warming hit per turn averaged, at times 3 global warming hits in a single turn - but I also had other centuries which had ~5...10 times less global warming hits! I.e., merely 10...20 global warming hits in a century. Thing is, I culturally captured and built very few cities after ~2700 AD; those are small bumps up on the power graph. This means that I had quite insignificant increases of food production between 2700 AD and 4200 AD, and main influence on the shape of my crop yield line was made by global warming. And yet, it is that period from 2700 AD to 4200 AD when i had those "waves" of global warming - and they are visible as periods of near-stable crop yields vs periods of relatively rapid decreases of it:
2199-4199_crop-yield.jpg


2199-4199_power.jpg


Note that AI crop yield graphs are far less informative here, because each AI empire lost much or most of their crop yield during those 2000 turns not to global warming hits, but to my cultural borders advancing. They all lost half or more of their lands to it, and thus that much crop production also. This is why AIs' crop yield graphs are barely influenced by those periods of intense, meltdwon-induced global warming: for their small - relatively to mine - empires, the main decrease was me, not global warming. On most of the lands i grabbed from them, however, i do not produce any crops: most of grabbed lands are simply beyond working areas of my own cities. Working radius is 2, but cultural radius of any legendary city is 6; and by 4200 AD, all but couple of my cities are legendary-culture indeed.

And up to 4200 AD, I was running my empire without a single automated improvement-building by any worker - manual improvements only. So lots of lands just remained non-improved - for forests to grow there. Lots of forest grew indeed, but lots were destroyed by global warming hits. So, now by 4200 AD, I decided to send a dozen workers to automated improvement building, as forests clearly won't manage to help anymore, with nearly half of all lands now being deserts, where forests can't ever grow... :sad:

The peace was absolute during those times, too: no wars whatsoever between 2777 AD and 4200 AD, not even between AIs. Russians were wiped out by India in 2777 AD, and remaining AIs keep seeing no gain in waging any wars. Means, there were no power changes nor crop yield changes due to any city conquests - not by me, nor by any AI.

Thus, I can fathom no other reason for those waves of global warming than this one: AI cities popped a meltdown once in a few centuries, sometimes two of them in a single century - and then whole map suffered accelerated global warming during next couple centuries. Then the effects of those meltdowns ended, and the usual, quite slow "industrial pollution-based" pace of global warming - resumed; and with AIs not re-building Nuclear Plants after each meltdown - each meltdown kills the Nuclear Plant which produced it, 100% of the time, - i still hope that soon enough, those meltdown-based intense periods of global warming will stop happening.

Sidenote: quite amusingly, 6 remainig AIs just sit on their butts and get slowly consumed by my culture during the last ~1500 years. Alas, by 4200 AD, most AI cities are stationing so many troops that it is mathematically impossible to have those revolt, even after these cities are completely surrounded by my lands and all other factors which increase revolt chance - are applied. Ironically, each culturally-flipped AI city - pushes all AI troops which were stationed in it into remaining AI nearest city or few. Normally, such units are then sent to some war by the AI - but when there are no wars, they just sit in such cities forever. Which means, in this Eternal Peace game, last few cities of any significant AI empire - inevitably end up with dozens of modern ground units, which produces well over 400 "anti-revolt" points in every such AI city. Which is how many anti-revolt power, by my rough estimate, is completely un-flippable culturally. Remember, it's vanilla Civ4 - no fancy spy stuff. Meaning, only war effort could capture such cities in my game - but of course, i don't declare wars, as i go peacefully in this game. This is how these AIs are still there even by 4200 AD, yep. :crazyeye: I don't mind, though. The only enabled win condition in my game - is diplomacy, and by 4200 AD, my empire has something like 70% of all votes in UN. So, I just keep "abstaining" when UN secretary election is offered, and without my votes, AIs can't do anything in UN. I could "win" this game way over 2000 turns ago, of course - but that's not my goal. So, them little AIs can stay there for all eternity. As long as they don't produce endless stream of meltdowns, of course. Which, i found, they don't:
Discovery 2: AI cities which had a meltdown - do not rebuild Nuclear Plants which were destroyed by those meltdowns.

Because one of completely squeezed by my culture AI cities, the one which had a meltdown in early 3500s - i kept an eye on. For almost 7 centuries after that meltdown, the city remained without a Nuclear Plant. Together with the above discovery #1, and having relatively quite few AI cities with Nuclear Plants to begin with, and some of those already being done with having a meltdown - there is still hope for my game to not end up completely Arakis-like.
 
Last edited:
Global warming in v1.74 should be really simplistic according to the source code: CvGame::doGlobalWarming
Only based on NukesExploded, a count that includes meltdowns and never decreases. And no indication in CvCity::doMeltdown that other power sources stop Nuclear Plants from exploding – stubbornly goes through all buildings in the city. Forests should not matter at all. There are NukesExploded independent GW rolls, each with a constant probability of 20% to convert one random non-city land tile to desert. If the tile is already desert, then there is no GW event. In this way, GW will slow down on its own. Which may explain plateaux in the desertification rate a while after a spike caused by a meltdown. But I don't see any timer that would cause a meltdown to be forgotten.

Apparently, it's only since BtS update 3.17 that forests affect GW:
Global warming in forests and jungles just removes those features, without turning plot to desert
Global warming takes into account the number of forests and jungles in the world
Global Warming is affected by unhealthy buildings in addition to nuclear explosions.
The change history of the source code confirms this; until 3.17, there was no substantial change to CvGame::doGlobalWarming. I've been playing with a GW mod for so long that I feel a little insecure about asserting anything about the original GW behavior. But it seems quite unequivocal to me code-wise. Pre-2008 CFC threads (v3.17 came out in May 2008) also align with this. This one is from Dec 2007 – unsuspecting of any upcoming changes. And a summer 2008 thread.

Preventing AI access to Uranium would seem like the only way to prevent GW entirely in the (very) long run.
 
Global warming in v1.74 should be really simplistic according to the source code: CvGame::doGlobalWarming
Only based on NukesExploded, a count that includes meltdowns and never decreases. And no indication in CvCity::doMeltdown that other power sources stop Nuclear Plants from exploding – stubbornly goes through all buildings in the city. Forests should not matter at all. There are NukesExploded independent GW rolls, each with a constant probability of 20% to convert one random non-city land tile to desert. If the tile is already desert, then there is no GW event. In this way, GW will slow down on its own. Which may explain plateaux in the desertification rate a while after a spike caused by a meltdown. But I don't see any timer that would cause a meltdown to be forgotten.

Apparently, it's only since BtS update 3.17 that forests affect GW:The change history of the source code confirms this; until 3.17, there was no substantial change to CvGame::doGlobalWarming. I've been playing with a GW mod for so long that I feel a little insecure about asserting anything about the original GW behavior. But it seems quite unequivocal to me code-wise. Pre-2008 CFC threads (v3.17 came out in May 2008) also align with this. This one is from Dec 2007 – unsuspecting of any upcoming changes. And a summer 2008 thread.

Preventing AI access to Uranium would seem like the only way to prevent GW entirely in the (very) long run.
"Should be" it is, but it isn't really simplistic. My proof is pretty conclusive, and personally, i'm fully convinced. And, gameplay takes preference over code-digging. I suspect there some other parts of the code in the game which do some few things about it - outside mentioned function. Perhaps in places which are not exposed to code digging, even?

There were at least three meltdowns in AI cities between 2700 AD and 4200 AD - i've seen and cleared most of fallout from them, as it spilled into my cultural borders. And as i said and demonstrated with crop yield graph, there were repeated periods of fast global warming with ~1 GW hit per turn average - and slow global warming, with 0.1...0.2 GW hits per turn average. And as you can see on the graph, those were long periods - hundreds of years each. Repeated periods, so can't be "GW hits less often as more and more land becomes desert" thing. Plus, the intensity difference is just way too huge - 5...10-fold. And statistics tell us that for such a long series of rolls - many hundreds turns with multiple rolls being done per turn - all that being "just a coincedence" is miserably, astronomically unlikely.

Most interesting about the forests' effect on GW being a BtS feature, though! I'll dig into this further. Thanks a lot for this info! :goodjob:

As for preventing AIs' access to Uranium - sure, i do what i can, which is directing much of my cultural expansion towards AI Uranium sources. But like i said, not all AI cities can be culturally flipped, because with enough troops in an AI city it just becomes immune to cultural revolt, and them AIs keep getting troops piled together in remaining cities the more of their cities i flip; and of course, declaring war is not an option, too. For now, i have only one AI left with one Uranium source - way too deep into its territory. Sadly, other AIs had enough time to built some nuclear plants before i've cut off their access to the stuff, too... %)
 
there were repeated periods of fast global warming with ~1 GW hit per turn average - and slow global warming, with 0.1...0.2 GW hits per turn average.
Not consistent with the source code, I agree. If it's simply 20% chance per meltdown (or rather already a good deal less due to existing deserts), then, once there have been a few meltdowns, another one won't increase the expected number of GW events dramatically (nor would a forgotten meltdown make a huge difference). Not sure how accurate your sense of the events per turn is. The fairly rapid decrease of your crop yield toward the end of the 4th millennium is an interesting phenomenon, and I don't disagree that, short of another explanation, it would have to be attributed to GW. Still a somewhat murky piece of evidence. Since GW doesn't destroy bonus resources, I reckon that food production could become incresingly dependent on food resource yields (and maybe water tiles too) as GW gets really bad. This could perhaps (in part) explain the flat tail – but not the steep decrease beforehand.

The DLL function for exploding a nuke and for changing the nuke-exploded count are exposed to Python and exported for access by the EXE. However, a text search of mine found no call locations in the Python scripts, and Dependency Walker, a tool for analyzing DLL calls in binary executables, showed no calls in the Civ 4 EXE either. (Not unusual for DLL functions to be externally accessible like this without actually being used.) There's a question whether the DLL source provided by Firaxis along with its game updates has really always been the code that they've compiled the DLL and EXE from – i.e. not somehow out of date. I generally think inspecting the code can be very nearly conclusive, but it needs to be done carefully and arguably by multiple people to have full assurance. And I've only looked at this ... until the matter seemed clear enough, maybe just for 15 minutes in total.

Regarding AI garrison movements: Closed borders could keep large stacks trapped once culturally enclosed. (Maybe that's already what you have.)
 
...
This could perhaps (in part) explain the flat tail – but not the steep decrease beforehand.
...
I generally think inspecting the code can be very nearly conclusive, but it needs to be done carefully and arguably by multiple people to have full assurance. And I've only looked at this ... until the matter seemed clear enough, maybe just for 15 minutes in total.

Regarding AI garrison movements: Closed borders could keep large stacks trapped once culturally enclosed. (Maybe that's already what you have.)
2nd steep decrease, you mean? 1st quite steep decrease happened starting ~2900AD, and then GW slowed enough for the graph to nearly level for quite some centuries. If that 1st clear slow-down of GW could be explained by any reasons you considered, then indeed we'd never see the 2nd steep decrease, yes. And vice versa: if that 2nd steep decrease would be how it goes, then we'd never see 1st almost-level period of the graph. I specifically voiced all the relevant conditions in the thread post above: tried real hard to find another explanation for this before posting, poking it this and that way in my mind; but found no other than to conclude that meltdown effect on GW - is finite length.
...
Completely agree about the code. Except, instead of "very nearly", i'd rather say "in most cases". Anyhow, seeing those real fine details from you, i had one more wild guess here: perhaps, this is not even an intended function or outdated code from Firaxis slipping in, but some kind of a bug / glitch? Something like this: them AIs had several revolutions during that 2700 AD ... 4200 AD period in my game. Perhaps that somehow zeroes "meltdown happened" flag or something? Or perhaps, number of cities / map size somehow glitches it out at some point? I'm running Intel 8400 CPU here, which for Civ4 is a total powerhouse, but with some 79 cities in my empire now, there's more than 2 seconds delay every time i open Domestic Advisor interface. Thank God for arrow keys in city view, yep - those, switch cities instantly. I.e., some kind of a bug / glitch which wasn't detected by testers and players simply because such long and such peaceful games - are not the mainstream way to play the game. Possible?
...
Yep, i closed them borders to all AIs ages ago. But i tried to open them as an experiment, to see what happens, few centuries ago - nothing changed in couple dozens turns; them AIs kept all the troops unevenly distributed among their remaining cities. Via diplomacy, each AI had all "declare war on" options for other AIs disabled, with lines like "we have nothing to gain from it" popping up when hovering my mouse over those. They just don't wanna move anywhere no matter open borders or not... :D
 
True, bugs are also always a possibility and could cause the nukes-exploded count to get changed by (seemingly) unrelated code. Also makes sense that playing for so long could run into obscure bugs.

A quick test using the built-in AIPlay developer feature hasn't been terribly illuminating. Tracking the nukes-exploded count via Python (taking advantage of it being exposed by the DLL), it did seem that desertification was surprisingly fast in a (50-turn) period with a lot of explosions (10 - apparently some nuclear war) and surprisingly slow in a (100-turn) period with no increase of the explosion count. Though both numbers also didn't seem that far off the expectation based on the code. I'll spare us the details. And the crop yield graph of the dominant AI civ was shrinking pretty much linearly so long as no conquests were taking place, bottoming out surprisingly abruptly toward the end. Anyway, given the amount of time and attention you've put into your game (even with autoplay, getting to year 2700 already took some patience – and the world was already 85% desert by then), you're in a better position than anyone to speculate on the dynamics.
 
True, bugs are also always a possibility and could cause the nukes-exploded count to get changed by (seemingly) unrelated code. Also makes sense that playing for so long could run into obscure bugs.

A quick test using the built-in AIPlay developer feature hasn't been terribly illuminating. Tracking the nukes-exploded count via Python (taking advantage of it being exposed by the DLL), it did seem that desertification was surprisingly fast in a (50-turn) period with a lot of explosions (10 - apparently some nuclear war) and surprisingly slow in a (100-turn) period with no increase of the explosion count. Though both numbers also didn't seem that far off the expectation based on the code. I'll spare us the details. And the crop yield graph of the dominant AI civ was shrinking pretty much linearly so long as no conquests were taking place, bottoming out surprisingly abruptly toward the end. Anyway, given the amount of time and attention you've put into your game (even with autoplay, getting to year 2700 already took some patience – and the world was already 85% desert by then), you're in a better position than anyone to speculate on the dynamics.
So it does seem there's something weird going on. Most interesting. Thanks for testing it, much appreciated - I lack such skills myself. :thumbsup:

As for speculating, there's one more thing - related to what you said: during my game, I've observed both AI lands as well as lands which are within my cultural borders but outside any city's working area getting GW hits far less often then any lands within any of my cities' working radius. Something like 2...3 times less often. At some point, when most of my cities already had multiple tiles desertified within their working radius - most AI cities had none or 1 desert tile. Later on, though, this difference roughly evened out, which I guess was my cities starting to have many GW rolled into already-desert tiles, while mostly-untouched AI cities had most GW hits actually creating new deserts for them.

Going through 4362 AD now - slowed down for all the work of building over a dozen new cities; started that right after 4200 AD. To ramp up cultural pressure on remaining AI cities best I can. Which means, further graphs and estimations have to wait until i'd do few millenia more - 1st, this new new wave of my cultural pressure should largely take its effect, and 2nd, gotta play for another long period of largely "most things remain as is" to observe GW progression more. After few millenia of AI piling up all the culture, it now takes several centures of maxed out cultural pressure to culturally flip any amount of their lands... %)

P.S. One of AI cities now has over 700 anti-revolt points worth of troops: about two dozens ground vehicles, roughly three dozen gunships, lots of infantries. One of two largest AIs, but isolated city far from AI's main lands. This bunch will sit there forever, and 700 is definitely impossible to revolt through, no matter how much cultural pressure on the city, i believe. And this city has a nuclear power plant, too... Sigh. :rolleyes:
 
Global warming in v1.74 should be really simplistic according to the source code: CvGame::doGlobalWarming
Only based on NukesExploded, a count that includes meltdowns and never decreases. And no indication in CvCity::doMeltdown that other power sources stop Nuclear Plants from exploding – stubbornly goes through all buildings in the city. ...
I now have one more piece of evidence that there is more complexity to GW than that in one other, different regard. If you're interested - see under the spoiler.
Spoiler GW tile selection is not completely random :

Three sreenshots below:
- India's capital and nearby area by 4700 AD, which by this time has roughly half of its working area still not desertified, while other half and more tiles to the east - are desertified, all by GW;
- same area in 2150 AD (last save i have before 1st GW hit happened in the game);
- whole map by 4700 AD.

What you can see on those - is that somehow, somewhere, there is some code which makes GW to prefer targeting a tile which is already near some other desert tile. Because if that would not be the case, then chances for Delhi to have such a distinct "half non-desertified, other half desertified" lands near it by GW (with zero desert tiles near it to begin with) - are practically zero. It'd require odds akin to dropping 30+ coins and getting a bunch of them to the west all showing faces while the bunch to the east almost all showing tails. Corresponding probability theory formula is tough to calculate, but this chance is something like 0.00sh percent or less, i'm sure.

I see the same tendency all over the map, too, but not as stunningly clear. While existing deserts somehow increase GW hits' chance to happen into non-desert tiles nearby, still, from time to time, quite infrequently, GW also will hit a tile which does not border any existing desert, too. Just significantly less often. Which is how the desert near Delhi started - I remember seeing a couple desert tiles added by GW to the spot about 3 tiles to the east from Delhi, about couple thousands years ago, into otherwise completely fertile land mass. I remember that with certainty because I was kinda jealous a bit, at the time, too: "That lucky bastard Gandhi still has all of his capital city's working tiles intact, while my cities are already suffering from GW quite much. Not fair!". %)

Anyhow, 3rd screenshot shows how by 4700 AD, i got dozens of complete desert areas which are 2...5 tiles wide, with still some 1...4 tiles wide non-desertified-yet areas between 'em, and relatively few places where desert and non-desert tiles are somewhat evenly mixed. The latter kind of regions tend to turn into complete deserts pretty quickly.

P.S. I even have a feeling, by now, that the code which does this - takes into account how many already-desert tiles are near the tile which is the next GW target. I.e., the more deserts are near any given non-desert tile - the more likely it is that the GW will consume that non-desert tile soon.
Delhi_4700.jpg


Delhi_2150.jpg


Worldi_4700.jpg

 

Attachments

  • Delhi_4700.jpg
    Delhi_4700.jpg
    898.6 KB · Views: 10
It could also be a simple artifact of the order in which the tiles are processed, or those 30 coins may all have dropped showing heads. :lol:
There's a tendency to read too much into game observations (I remember a few weeks ago a guy claiming his observations proved you got a combat bonus when attacking units facing away).

For that behaviour to exist, it would need to have been coded in, it simply cannot be the result of a "bug" (ie, unintended behaviour).
From @f1rpo's reading of the code, it's not there: the chance of conversion is a flat 20% (ie, non dependant of adjacent desert tiles), the tile selection is random (the only criterion being a non-city tile: no trace of a check for adjacent desert tiles).
 
It could also be a simple artifact of the order in which the tiles are processed, or those 30 coins may all have dropped showing heads. :lol:
There's a tendency to read too much into game observations (I remember a few weeks ago a guy claiming his observations proved you got a combat bonus when attacking units facing away).

For that behaviour to exist, it would need to have been coded in, it simply cannot be the result of a "bug" (ie, unintended behaviour).
From @f1rpo's reading of the code, it's not there: the chance of conversion is a flat 20% (ie, non dependant of adjacent desert tiles), the tile selection is random (the only criterion being a non-city tile: no trace of a check for adjacent desert tiles).
What you said - is exactly because i provided all the details about it, including the world map. It ain't a single case, those 30 coins.

As for why this behaviour exists - i have not a slightest idea, myself. I took a look into the code f1rpo provided, and i agree with you: there is no clear sign there that adjucency to existing desert is anyhow a factor. Yet, i am not competent enough to figure out how all the used by that code functions actually work - including "pick a random tile" one. I know that any "randomize" function in coding actually involves dozens of actual CPU operations, though. What exactly happens there - is beyond my understnading. While on the other hand, on that worldmap, i see - and now you also can see - what the game is doing. It's as visible a thing as the code f1rpo provided, to me.
 
I have a pretty clear idea of what a plausible random sequence of coin flips should look like, but I'm not used to looking at randomized 2D patterns. [Edit: OK, a silly thing to say for a Civ player; each game's map is filled with random 2D patterns. But most of them aren't uniformly distributed.] My sense is that the unevenness may get a lot more pronounced; something something the curse of dimensionality. I've run a little Python script (written mostly by a chat AI and relying on the Python random module for the PRNG) that starts with a grid of 8x8 "X"s and then successively chooses one at random and flips it to a blank character – until exactly half of the Xs are gone. Here are six of those half-cleared grids:
Spoiler :
Code:
X           X X
        X   X  
X X   X   X X X
    X     X    
X   X X X X    
X   X X X X    
  X     X X X  
X X   X X     X



  X X   X X X X
      X     X X
X X   X   X   X
X         X X  
    X   X X   X
  X X         X
X X   X X      
X X X     X    



X X   X X X    
  X   X     X X
  X X X     X  
X   X   X X    
X X X X X      
X X     X X    
  X           X
    X X   X   X



X X X X X X   X
            X  
X   X X   X X  
X X X X     X  
  X X       X  
X   X X   X    
  X X       X  
X X       X X  



X       X X   X
X     X       X
  X       X   X
X X     X   X X
  X     X X X  
    X       X  
X X X   X   X X
  X X X X X    



X   X X X      
X   X          
X X   X   X X  
X     X X   X X
X   X X        
    X     X   X
X X   X   X X  
      X X X X X
The fifth, for example, has a big (3x5) cleared block in the middle. Several of them have long portions of the outer edge left untouched. Your screenshot doesn't blatantly fall out of this pattern to my eye.
Edit3: Let's also go from Chess to a Go-sized grid ...
Spoiler :
Code:
X   X     X         X X   X   X X   X
  X     X     X X   X X   X   X      
  X X     X X X     X X         X X  
X   X X   X X   X X X X X X     X X  
  X         X X X             X X   X
  X   X   X   X   X X X X X         X
    X X   X   X X X X   X            
X     X X       X X X   X   X X   X  
X X X X X X       X     X X       X X
X     X X X   X X     X     X   X X  
    X X   X X   X   X     X X X X X  
X X X X   X         X     X X X X X  
    X   X   X X X X X X X X X        
X X   X         X         X X X X X  
X X           X   X X   X X     X   X
                X X   X X X     X X X
X   X       X X X X               X X
    X   X     X   X       X X X X X X
X     X         X X X   X X   X X

There is certainly one kink in the syncRandPlot function: It does not first collect all valid tiles and then choose one at random, but it chooses a random coordinate ignoring all selection criteria, checks the criteria and tries again until it happens to pick a valid tile – or until it has failed 100 times. That's the iTimeout default value defined in CvMap.h. Once the number of valid tiles becomes very small, especially on very large maps, it may well happen that the timeout is reached before finding any nondesert, noncity land tile. But this should only affect the (very) late stages of GW. It might explain the surprisingly abrupt flattening of the descending crop yield curve that I've observed in my automated test. This timeout mechanism could also be an intentional design decision (rather than just an attempt to save processing time) to generally allow per-tile events to get canceled that can only affect a very narrow set of tiles. Ensures a degree of unpredictability. After rewriting syncRandPlot in my own mod (so that a tile is guaranteed to be found), I ran into the problem that Barbarians kept respawning too rapidly where they had just been killed when very few eligible tiles were left.

Edit2: The GameCore code may also be in your own Civ 4 install folder, in a subfolder CvGameCoreDLL. But I'm not sure if all editions of the game come with the source code. So (maybe) I don't need to be relied on for providing the source code in its correct version. The game updaters, e.g. for v1.74, also used to include the (presumably) latest version of the source code and (presumably) replaced any older existing version of the code with the latest when applying the update.
 
Last edited:
Top Bottom