Advanced Civ

Hi f1rpo,
Thank you for the wonderful work! I played this mod some time back around v0.96 and it gave me great enjoyment.

I was just looking over the changes that you have made since then, and I have one quick question about the SPI: in the manual there is a list of specific scripts for which it is used, but it also says that in principle it should be used for any script that doesn't provide its own starting site implementation. An upper limit on the size of the map (12000 tiles) is also mentioned, beyond which SPI will not "engage" after the initial default implementation. The Global Highlands script is in this context given as an example of a map which on world size HUGE is too big (perhaps that is why it is not in the list of official scripts for which SPI is intended, since it meets the other condition - not to override any of the starting plot functions). My question: would I be correct in expecting that SPI would be used for a script like that on smaller world sizes? That is, is my impression that the 12000-tile limit is the actual criterion for whether to do SPI - rather than the name of the script - correct?
 
@Fimbulvetr: Thanks for stopping by.
Apart from decisions about war and peace, I haven't attempted anything ambitious with the AI. The Caveman2Cosmos team still has some bigger plans, but they're kind of digging themselves out of a hole. K-Mod also got included in Realism Invictus and History Rewritten, which are still being developed. Civ 4 Reimagined is also based on K-Mod, got started around the same time as AdvCiv and is still alive. Agree that it would've been a pity for karadoc's work to just slip into oblivion. Oh, and the Rhye's-and-Fall world might still adopt (part of) AdvCiv.

@Jorunkun: No problem, I had assumed that you might be busy until sometime in August. I agree that the coastlines are too detailed, I don't think any of the parameters in the script have much impact on that. For example, there is
Code:
#Chance for plates to grow. Higher chance tends to make more regular
#shapes. Lower chance makes more irregular shapes and takes longer.
# advc: Was 0.3, both. Also randomized now.
self.plateGrowthChanceX = 0.5
self.plateGrowthChanceY = 0.5
... but the effect isn't really as described, I guess because large continents usually consist of multiple plates. Or maybe another part of the algorithm adds (a lot of) noise. I think I'd have to add some debug output to get a better intuitive understanding of how it works.

I've played a bit more with your settings too, Totestra, combined with the v1.07 DLL:
t300.jpg
That really is its own thing. In 75 turns of exploration, my 2 Warriors only managed to explore half of this enormous, near-100% forested disk. Didn't meet another civ until turn 150 or so. Mainly just produced Workers (which mostly chopped forests and improved resources - the rivers connect most cities for free), Settlers, Monuments and Axemen against the Barbarians. I'm not ahead by much in score, but I've 12 cities vs. France's 7. Still, not too bad on the part of the AI, for Monarch. Russia got overwhelmed by the Barbarians. I was a little late in producing Galleys, so I took a bit of a beating from the sea, briefly losing two cities to naval landings. Barbarians didn't feel crazily strong overall despite all the space, but it's probably a bit too much for the AI. I'll try to let the AI treat maps with a lot of space the same way as Raging Barbarians; there's already some code for dealing with that option.
 
@nkp6: v1.06 has moved the tile limit from the DLL into GlobalDefines_advc.xml (SPI_PLOT_LIMIT) and also the time limit (SPI_TIME_LIMIT_PERCENT) after which the iterations stop even if the solution doesn't look very good. Yes, Global Highlands should work if the limits are raised or if Large size is used. The dimensions set by the script, as you seem to be aware, are:
Spoiler :
Code:
def getGridSize(argsList):
	"Enlarge the grids! Need extra land to cover that 'wasted' by all the peaks!"
	grid_sizes = {
		WorldSizeTypes.WORLDSIZE_DUEL:		(13,8),
		WorldSizeTypes.WORLDSIZE_TINY:		(16,10),
		WorldSizeTypes.WORLDSIZE_SMALL:		(21,13),
		WorldSizeTypes.WORLDSIZE_STANDARD:	(26,16),
		WorldSizeTypes.WORLDSIZE_LARGE:		(32,20),
		WorldSizeTypes.WORLDSIZE_HUGE:		(38,24)
	}
These are 4x4 cells of tiles, so Huge maps have 14592 tiles.
[...] The Global Highlands script is in this context given as an example of a map which on world size HUGE is too big (perhaps that is why it is not in the list of official scripts for which SPI is intended, since it meets the other condition - not to override any of the starting plot functions). [...]
I probably assumed that Global Highlands overrides assignStartingPlots just like Highlands does. Not the case, so I'm going to add Global Highlands to the list (with a caveat about Huge world size). I'll also add one line of log output somewhere just so that a player can tell whether SPI was used; I've wondered that too at times, and then had to first enable the SPI log in the DLL.
That is, is my impression that the 12000-tile limit is the actual criterion for whether to do SPI - rather than the name of the script - correct?
Right, I haven't hardcoded the names.
 
Thanks for trying out Totestra - I've just lost a game (finished 2nd) on your settings, on a low-ocean huge PerfectMongoose, with 16 Civs.

In the actual playing and what results from the mechanics of the game, the experience is very similar. But you have less space, and every city location is broadly equal, and the number of viable strategic playstyles is lessened and the traditional default strategy becomes more dominant. It feels a lot more like traditional civ, with everybody starting close to each other, a short rex, and then wars with your neighbors.

PerfM is certainly a more balanced, a fairer map. But I miss the extra width, and gradual morphing of terrain types; everything is a lot more blurry and jumbled up.

But as for the GotM, if you feel that PerfM is the map to use, I would. Maybe we could go with standard sea-level and only a dozen or so civs, to make things a little less crowded?
 
[...] I've just lost a game (finished 2nd) on your settings, on a low-ocean huge PerfectMongoose, with 16 Civs. [...]
Thanks for doing this dry run, a whole (Marathon) game even. I gather that the modifier changes didn't make a world of a difference. (I'd consider an indifferent reaction a success here since it's not something that you felt needed fixing, as far as I'm aware.) Could you provide me with a savegame or some screenshots? I'm mainly wondering if starts were perhaps unusually close together – which wouldn't really be so unusual because sometimes a small continent or two are left empty initially. So that I have a better idea of what a "short" expansion phase is in your book.
[...] But you have less space, and every city location is broadly equal, and the number of viable strategic playstyles is lessened and the traditional default strategy becomes more dominant. It feels a lot more like traditional civ, with everybody starting close to each other, a short rex, and then wars with your neighbors.
City locations broadly equal in terms of their overall yields? Power level? I mean, the cities in my Totestra game were highly uniform ... (Very different cities existed, no doubt, on other continents.) And the non-viable strategies would be ... peaceful ones? Marathon, even with the 250% modifiers, sure stacks the deck against those. Or just slow expansion? That should generally be feasible with the settings you've used, but will often necessitate war before the midgame. By the way – if I've provided enough info about my Totestra game to answer this at all –, do you think you would have approached that map very differently? I had no resources for speeding up wonders, no islands in sight, seemed to me like the only thing to do was to REX at least until my borders would reach another civ (which has just happened after 300 turns). I guess I could've founded a religion (dicy with so many civs in the game; early wonders too). But it seems pretty clear that just grabbing the land was the most powerful play, with maintenance relatively low on Huge maps and workers cheap.
But as for the GotM, if you feel that PerfM is the map to use, I would. Maybe we could go with standard sea-level and only a dozen or so civs, to make things a little less crowded?
If we want more space, then, sure fewer civs will work. Medium sea level tends to result in snakier continents, so it might be better to stick with Low and go down to 14 or 15. Could also increase the grid size a bit, which would just be a matter of adding
Spoiler :
Code:
def getNumPlotsPercent(argsList):
	[iWorldSize] = argsList
	if iWorldSize < 0:
		return 100
	sizeModifiers = {
		WorldSizeTypes.WORLDSIZE_DUEL:		100,
		WorldSizeTypes.WORLDSIZE_TINY:		100,
		WorldSizeTypes.WORLDSIZE_SMALL:		100,
		WorldSizeTypes.WORLDSIZE_STANDARD:	100,
		WorldSizeTypes.WORLDSIZE_LARGE:		100,
		WorldSizeTypes.WORLDSIZE_HUGE:		100
	}
	return sizeModifiers[iWorldSize]
to the PM script (after the import statements) on your end, replacing the 100 percent for Huge size with, say, 115. (If the grid is made even bigger than that, then the SPI_... variables mentioned in my previous post will have to be increased as well.)
 
There's an option "Opacity on Water" on the Map tab of the BUG menu. Should be medium by default. Certain colors might be difficult to see at all on water with that setting.
 
As for the GotM, I'd like to summon @SantaFlagship, @keldath and anyone else interested and propose the following:

- Huge / Marathon (yes, this leads to 50+ hour games, but for me that's kinda the point - I'm most interested in how the first half of the game plays out, so no worries if some don't finish ... the game is often somewhat decided by this point anyway)
- No Goodie Huts, No Events, No Tech Trading
- cultural victory off (it's a dominated strategy under these settings)
- Monarch difficulty (FYI, Monarch is actually pretty tough with No Tech Trading - I only win about 1/3 of my games on these settings)
Sorry for being away for ages! Better late than never, I suppose. The proposed settings, with a PerfectMongoose map, look great! Reading the discussion and notes on Totestra vs PerfectMongoose, I take it everyone is on board with the latter? If you want a less crowded map, Jorunkun, doing something like f1rpo proposed,
If we want more space, then, sure fewer civs will work. Medium sea level tends to result in snakier continents, so it might be better to stick with Low and go down to 14 or 15.
is absolutely fine by me. If we go with PerfectMongoose, it's only fair if you get to decide on this one.

Will the GoTM game be played with v1.07rc? I am pretty excited to try out the new changes, in particular the modified Marathon game speed.
 
Hello everyone,

on my last day of work before a two week holiday. I'm ready to kick off a GotM with the following settings:

AdvCiv 1.07 rc
12 Random Civs / Leaders, random personalities
Huge Perfect Mongoose map with default settings (temperate, normal sea level, etc)
Monarch difficulty, marathon speed
No tech trading, vassal states, tribal villages, random events; no cultural victories

The recommended density is 16 players, so expect the map to be a bit spacier and the game to be more strategic: In your REX you can still do all the things you'd do on faster and denser settings (such as an early rush on a neighbor) but it requires more planning and trade-offs and peaceful building is a more viable alternative.

If I get three go-aheads, I'll generate the map, and post a 4000 AD save and a screenshot here. Customarily, GotM players post about the opening and their intended strategy before starting to play, and many give updates on their progress at fixed marks, frex 2nd city founded, 1000 AD, etc. before they post the final save. This is of course optional, but is what makes the format so appealing - alternative histories, comparing strategies and outcomes.

I'll promote it over on reddit, ask people to join if you don't mind. We'll mention that this is a mod with a small but dedicated community and invite people to try it with us.

Sound like a plan?
 
Fine with me. At least to outsiders, the reduced player count shouldn't feel strange as the default in BtS even lower. Attaching some sample maps with your proposed settings for illustration (the human player is Arabia in light green). That's with one more tweak to the map script: continents are discouraged near the edges of the map – to make continents that cross the edges of the minimap less likely. I've released that along with a few small unreleated changes as v1.07 – so we can forget about the release candidate. Notable: The Demographics tab had shown garbled text on turn 0 (bugfix) and, relevant to @nkp6's question about the iterative starting position algorithm: PythonDbg.log now says "Running Starting Position Iteration" when the algorithm is used. (Not really a Python thing, but that logfile is easy for players to enable.)
If I get three go-aheads, I'll generate the map, and post a 4000 AD save and a screenshot here.
This should imo go into a separate thread, arguably in the AdvCiv subforum. (Not sure if by "here" you mean this very thread.)
 

Attachments

  • Civ4ScreenShot0000.JPG
    Civ4ScreenShot0000.JPG
    238.9 KB · Views: 49
  • Civ4ScreenShot0001.JPG
    Civ4ScreenShot0001.JPG
    208.9 KB · Views: 43
  • Civ4ScreenShot0002.JPG
    Civ4ScreenShot0002.JPG
    242.4 KB · Views: 42
  • Civ4ScreenShot0003.JPG
    Civ4ScreenShot0003.JPG
    226.5 KB · Views: 45
Reading the play along was intriguing. It seems there may be some changes incoming. May I suggest a small nitpicky detail? Workboat starts with fish and 4 hammer cities seem very strong on marathon because improving sea resources didn‘t slow as much as everything else.
It seems like a rare edge case, but workboat first starts could use a nerf on marathon imho.

Is city maintenance behaving as expected in marathon?

Also: What map script do you generally suggest to produce interesting games?
Which of the options chosen for the play along (no tribal villages, no tech trading, no random events) favor the AI?
Which would you recommend (including map settings) for a newer player who looks for more challenge without more AI cheats?

Thank you for all the work and dedication put into the mod!
 
It seems there may be some changes incoming.
Yes, a bunch of tweaks and minor bugfixes. I haven't started going through my list, but I think I will soon. (I have fwiw just implemented an unrelated change that eliminates puzzling trade connections between rivers that are two tiles apart.)
Workboat starts with fish and 4 hammer cities seem very strong on marathon because improving sea resources didn‘t slow as much as everything else.
It seems like a rare edge case, but workboat first starts could use a nerf on marathon imho.
Work Boat and a coastal start (which typically has seafood) and Fishing as a starting tech and, ideally, a Forest Plains Hill in the city radius. That's indeed pretty narrow. Producing the Work Boat extra fast by working a production-only tile means that the city won't grow – which is normally an advantage of Work Boat over Worker. Well, with population growth being slow on Marathon (and one Warrior -arguably- being sufficient for exploration), growing to size 2 before a Worker seems undesirable anyway. Perhaps most importantly, Worker is also cheap on Marathon, so I don't see how Work Boat is being advantaged over Worker. Even if it is, to some extent, it wouldn't worry me because Work Boat starts are imo rather undesirable on Normal speed. Edit: Your point is that worker speed is affected by Marathon while Fishing Boats still only take a single turn; I see. Still, I don't think it's a problem on the bottom line; at least not one that would justify a cost increase for Work Boat.
Is city maintenance behaving as expected in marathon?
Distance maintenance seemed too low in the play-along game, but I guess that's an issue with Huge map size. Game speed currently affects maintenance only through an inflation adjustment.

Regarding game settings, it sounds like you intend to play on Marathon, but I'm not sure. Marathon favors warfare and an experienced human player runs circles around the AI when it comes to that. Here's a recent discussion about that – not involving mods, but I don't think the changes in AdvCiv make a fundamental difference.

What map script do you generally suggest to produce interesting games?
Fractal is the default in my mind. I find the uncertainty about the number and size of the continents interesting. PerfectMongoose is arguably even better at that and is, in many ways, more realistic, but the maps can turn out too watery (too many small continents, chokepoints created by bays), negatively affecting game balance and AI. There's been progress on that front; I think the script plays pretty well by now. Not sure how much the Earth-like climate (less locally varied than Fractal) affects fairness.
Which of the options chosen for the play along (no tribal villages, no tech trading, no random events) favor the AI?
The reluctance of the AI to trade techs, especially the "not just yet" response, is in my estimation a major handicap imposed on the AI – by necessity, I think, because, if the AI were allowed to trade rationally, tech would spread around very rapidly. No Tech Trading levels the playing field, i.e. makes the game more difficult for the human player. (Unless the human player is, for some reason, unable to trade techs aggressively: starts isolated or falls far behind before tech trading becomes available.) Without tech trading, however, war trades and peace trades (reparations or brokered peace) are rarely affordable. AdvCiv lets the AI trade cities at peacetime, but the conditions for that are pretty narrow, so it's not a great help. The "fair and forthright trade," "shared discoveries" and "traded with our enemy" diplo modifiers also don't play much of a role without tech trading, and the AI will rarely be able to ask for a gift or demand tribute, so those diplo modifiers, too, are diminished. As a result, Friendly attitude becomes very difficult to reach without a shared religion and Annoyed attitude also somewhat rare without warfare. I don't think diplomacy becomes entirely bland without tech trading, but it's a major element of gameplay removed and not a favor to the AI in terms of making the AI fun to interact with. There is certainly also much to criticize about the gameplay that results from tech trades being allowed, but, to me, it's the lesser evil. Keeping the player count moderate reins in the pervasiveness of tech trading.

The human player should be better at locating Tribal Villages than the AI, and, below Monarch difficulty, the AI starts exploring a few turns later than the human player because it won't leave its first city unguarded. (But the AI does start with 10 free production at the capital, so another Warrior doesn't take long.) On Emperor and Immortal, the AI gets 2 Archers and 1 Scout. In BtS, it's even 2 Scouts and, on Immortal, 3 Archers; that'll clearly make Tribal Villages benefit the AI more than the human. With just 1 Archer and 1 Scout for exploration on Emperor in AdvCiv, I don't think the AI gets much of an advantage from Tribal Villages being enabled. For me, it's more a matter of early exploration being driven too much by Tribal Villages when they're enabled or being unexciting when they're disabled; and Tribal Villages sometimes leading players down unusual tech paths vs. introducing too much (random) unfairness – despite the substantial reduction of gold and research rewards in AdvCiv.

The AI logic for dealing with Random Events is very primitive, but, then, most Random Events don't require any particular response or preparation and just aren't very impactful. (Maybe that's their saving grace. I find most events ill-conceived and poorly implemented, so I've been treating this feature as irredeemable and have only fixed a few major issues.)
Which would you recommend (including map settings) for a newer player who looks for more challenge without more AI cheats?
I would recommend the mod's default settings, i.e. Normal speed, Fractal, Standard size, Medium sea level, 8 players, or possibly Small size, Low sea level, 7 players –, all options unchecked. Well, see above about Random Events. The mod adds some options for an extra or special challenge (Rise & Fall, Start Points as Handicap, No Slavery), but I don't suppose those are recommendable for a "newer" player. If Marathon is a must, then perhaps Large size, Medium sea level, 11 players – although this will magnify the role of tech trading.

Regarding AI cheats, on Monarch I still find the AI freebies and discounts pretty subtle. On Emperor, the mod turns part of the AI discounts into slight cost increases for the human player. At least this makes things more transparent (and it reins in the tech pace and makes the map less crowded with military units). Immortal is where the game feels heavily and overtly lopsided to me.

To find the right settings, it could also help to look at some maps in Debug mode (Ctrl+Z, requires the chipotle cheat code in My Games\Beyond the Sword\CivilizationIV.ini) and perhaps to run AI Auto Play (with the map revealed) for one or two hundred turns. Also helpful for getting an impression of the AI behavior.

Oh, and there's already a thread in the AdvCiv subforum with some opinions about game options.

edit: clarity
 
Last edited:
Great mod, I love that the AI is much more cautious and calculating in its war decisions. Your mod is significantly more difficult than vanilla K-mod and I was wondering if it had anything to do with the idifficulty setting in the Handicapinfo.xml file that increases by 10 for each level. Can you elaborate on this setting as to what exactly it does. Also a few suggestions that I have done to make the game better imo. Mod the Great temples to give -1 science in addition to +1 gold, which will let you field larger armies, but not run away with the game founding multiple religions. This reflects historical church dogma towards science and is a better balance for game balance. Also I would love super forts with automatic irrigation like a city as I play with the only settle by fresh water turned on to prevent massive city spamming and to keep the barbs coming out of the unsettled wilderness areas. Nothing compares to K-mod and your mod makes it better. Picked up AOW Planetfall not to long ago but while still an entertaining game, severely lacking in the AI department, so I'm back to the old steady. Thanks for the hard work and keep it going. Maybe we can get 1/2 a century out of this game. :)
 
Great mod, I love that the AI is much more cautious and calculating in its war decisions. Your mod is significantly more difficult than vanilla K-mod and I was wondering if it had anything to do with the idifficulty setting in the Handicapinfo.xml file that increases by 10 for each level. Can you elaborate on this setting as to what exactly it does.
Thanks. iDifficulty is pretty much just a technicality, an annotation, say, that summarizes how challenging all the other modifiers are taken together. The mod uses this for the game end score and for averaging difficulty levels in team and multiplayer games. BtS had relied on the number and order of difficulty levels in XML for this – which had made it awkward to add additional difficulty levels. AdvCiv also makes some few adjustments to AI behavior based on iDifficulty:
• less likely on higher difficulty to offer gifts to a human player that is falling behind
• tribute demands for low-value items disabled except on low difficulty
• AI somewhat more willing to pay for peace on lower difficulty
• sub-Monarch AI slightly more confident in strength in numbers for tactical decisions (though still less so than the original AI)
Mod the Great temples to give -1 science in addition to +1 gold, which will let you field larger armies, but not run away with the game founding multiple religions. This reflects historical church dogma towards science and is a better balance for game balance.
So a Shrine would merely convert research to gold one for one? Couldn't I just lower my research slider to get the same effect? Conventional wisdom for BtS is that Shrines are often worth conquering but rarely worth building (see e.g. this thread).

I believe the mod changes the calculation a bit, mainly, by making the AI less interested in founding early religions. I would agree that Shrines don't scale well with the map size and that conquering Shrines feels a little cheap – letting the original owner do all the work –, especially when the Shrine is pretty far away from the conqueror's homeland; doesn't feel historically resonant then either, i.e. when the religion is really alien to the conqueror. I had been considering to halve the Shrine income when the Shrine religion is not the state religion (or maybe only when running a different state religion). Founding multiple religions in a row should, ideally, also be more difficult, but, primarily, because the AI struggles with juggling multiple religions, and because games tend to be more interesting when more than a couple of religions compete for adherents.

As for adjusting the Shrine income to the map size, the production cost of Missionaries would then arguably also need an adjustment because, if the income on Huge maps is e.g. only 0.5 GPT per city, then spread via 40-production Missionaries becomes pretty unattractive.

Well, I don't think the balance issues are severe enough to justify such significant changes at this point of the mod's life. (I had, at one point, meant to implement a package of balance tweaks to religion.)
Also I would love super forts with automatic irrigation like a city as I play with the only settle by fresh water turned on to prevent massive city spamming and to keep the barbs coming out of the unsettled wilderness areas.
I'm not a fan of Super Forts; I'd sooner (but no plans for this currently) experiment with changes to the maintenance formula to make remote outposts (i.e. proper cities) less costly so long as their population stays small – but not too cheap because that would enable rapid early expansion. Maybe you could relax your freshwater rule a little by applying it only to some terrain types (assuming that this is implemented simply through Civ4TerrainInfos.xml). Or maybe you could tweak the map scripts so that more rivers or lakes get placed ...
 
I don't like superforts making irrigation either.
However, I would like to see irrigation pass through a city on a hill or desert at something like Steam Power (you can make lifting pumps).
 
[...] I would like to see irrigation pass through a city on a hill or desert at something like Steam Power (you can make lifting pumps).
For Desert, that's already how it works (without any tech requirement), also in BtS. The game doesn't check all the Farm conditions, only checks for flat terrain. A check for a positive food yield would be difficult to perform because the city raises the food yield to 2 and, even if we don't want that to count, there may have been a Flood Plains that got removed by the city. So requiring a tech for Desert and Snow irrigation would require some cleanup, and adding an effect to Steam Power just for Hill cities isn't worth the added complexity to me. (My sense of realism also isn't offended by Hills being insurmountable obstacles to irrigation.) Nice to see that you're still around btw.
 
If someone wanted to add a building, a Desalination plant would be another option. Placed at Biology maybe? Requires coastal, and Power. Gives a city fresh water. Probably not really a case for this mod, but someone else might like the idea.

---
And now we wander off topic...
Nice to see that you're still around btw.
Thanks. Dang, I've been playing Civ IV as long as I had played the original (1991), and hanging out here the whole time. I don't post much anymore, but I stop by at least once a month.
Master of Orion (now updated as the freeware Remnants of the Precursors), and MoO2 (with this years fan patch) I still play as well. That's it for 4x games. I own civ 5 & 6, I don't have 100 hours in either. Same with MoO3&4. And more other 4x games than I can count that come and go.

For non-4x I still fire up the 1987 Dungeon Master on the Atari ST emulator every year or so, or it's sequel. (and now it has a great Android port).
Showed some kids (in their 30s..) at work Trek-80 a few months ago. Despite not having played it in 40+ years, it came back pretty quick.

Get off my lawn :)
 
Top Bottom