Advanced Civ

[keldath]

Pepo,

from xp, its touch to merge existing code with AdvCiv. f1rpo, changed enormous amount of files and code.
many places total rewrites.
i have been keeping up with advc for the past 2 years on a daily basis.
merge it with revdcm base code / and2 / c2c is a hell of a work i think.
the other way around should be, merge existing code to advc. even then, many hours of work.
my Doto is advc with additions to some extent, though still vanilla flavor.

 

[f1rpo]

Thanks, Pepo.

Hyphotetically speaking, how difficult would it be to integrate your work with other mods regarding Ai? For example with LOR, besides the revolution module, would it require a conisderable effort to adapt ?

A lot of the AI changes are all over the AI classes. This is, in particular, true about most of the AI changes inherited from K-Mod. Merging those into another DLL mod is not feasible I think. The "LoR SDK ModMod" consequently only adopts some select bits of K-Mod code. The code for AI decisions on war and peace is mostly contained in separate classes, but still isn't really portable because it relies on the K-Mod/AdvCiv pathfinder. Moreover, I've refactored the codebase a lot, and that makes all the code somewhat difficult to port to other mods.

Merging another DLL into AdvCiv seems like the saner approach. (Agree with keldath – didn't see his post as I was writing mine.) This has been attempted with Rhye's and Fall of the Greek World:

[...] Anyways, that'll be after I'm finished with the AdvCiv merge, which will take a while - probably not nearly as long as 2.0, but I expect a month or two depending on my motivation and free time. There's still plenty of DLL code I haven't merged yet, and I've only started with the XML part this weekend.

(That was in February; no updates since then.)

As for LoR specifically, according to my crib sheet, that's mostly:
• Better BUG AI (BULL + BBAI)
• DCM (stack vs. stack combat, ranged/ archer bombardment plus some combat rule changes relevant for the late game)
• Revolution
• Orion's Inquisition, Super Spies, Influence Driven War, Wolfshanze's mod, Varietas Delectat, legendary military units
• added content ("unofficial expansion")

Perhaps one could whittle that down to a subset of essential components. It could also help to merge with keldath's DotO (instead of AdvCiv), which already includes a ranged combat system. As for the UI, it would seem wisest to ditch BULL and stick to the AdvCiv UI for the time being, i.e. BUG plus customized parts of BULL plus novel UI tweaks. Merging Varietas Delectat isn't a trivial task, but should be doable (see this post).

I don't suppose that the actual AI behavior would have to be adapted much; the merge would mostly be a (big) technical task.

 

[Pepo]

Oh well good to know. I have been doing some modifications of my own on LoR and one thing that irks me is the Ai, specially on regards to war declaration and pathfinding. As my knowlegde of DLL is close to 0, i would leave it as it is for the time being. Thanks a lot for both of your anwsers

 

[keldath]

lots of new updates
my vacation from modding depends on your updates...so, its over now

 

[raystuttgart]

Congratulations to your new release.

 

[f1rpo]

Thanks keldath, ray. These won't be the last updates either. I'm still not happy with the frequency of AI-initiated wars. Still a bit too few early on. The crowdedness of the map has a big impact on that. In my tests, it's mostly Pangaea that often fails to produce any pre-Medieval wars, and just decreasing the dimensions (which are greater in the mod than in BtS) a tiny bit already seems to do the trick. And late-game wars are probably still too frequent, especially on large maps.

Also, and perhaps more importantly, I've just gotten a bug report in a private message. Posting this here in case that someone else runs into the same issue:

[...] I'm facing some problems with new releases: when starting a single player game, in leader selection, Civ4 crashes (suddenly) if I select some leader (Dutch, Egypt, Khmer ..., but most are working fine).
With AdvCiv v0.99b all was working fine, but with 1.00 to 1.03 the problem is present... [...] (I'm writing a private message, but if you prefer I can write this issue in a thread...) [...]

Looks like I hadn't really tested the Play Now/ Play Scenario screens; I guess other players don't use them much either. That would also be my suggested workaround – to use the Custom Game/ Custom Scenario screens, at least when trying to select a leader with the Creative trait. Turns out that the usual check for whether a game has already been started doesn't work on the Play Now screen, and a change in v1.00 to the game text for the Creative trait (which the Play Now screen requests without even displaying it) then causes the crash. Can only be fixed in the DLL. I'll include the fix in the next update, maybe in early December.

 

[gi@c]

That would also be my suggested workaround – to use the Custom Game/ Custom Scenario screens, at least when trying to select a leader with the Creative trait. Turns out that the usual check for whether a game has already been started doesn't work on the Play Now screen, and a change in v1.00 to the game text for the Creative trait (which the Play Now screen requests without even displaying it) then causes the crash. Can only be fixed in the DLL. I'll include the fix in the next update, maybe in early December.

Thank you for the fast reply and workaround!

 

[Jorunkun]

Been playing a couple of adv civ games on my usual settings - huge totestra, marathon, 18 civs, monarch, no tech trading, no vassals - and wanted to again thank f1rpo for his excellent work on this mod.

AdvCiv is the closest thing to the civ of my dreams: Still basically civ, but with better AI that doesn't rely just on numerical bonuses and a few sensible tweaks that improve balance and eliminate dominant strategies. Kudos!

Haven't come across any major stability issues, but here's a few things I've noticed:

1. On huge maps, the AI is still founding a lot of sub-optimal cities. I think it made better choices in "better AI" and "K-Mod". Not sure what has changed, but there are still too many cities one tile removed from the coast and cities are often spaced sub-optimally

2.The AI is very reluctant to go to war before riflemen. Maybe this is because of the changes to diplomacy, where attitudes oscillate more around zero (except in the case of shared religion and favorite civics)? I've played maybe a dozen games on the latest version and have not seen a single axemen-era rush.

3. The United Nations elections have had some issues where a chairman or policy got the required majority but the result wasn't enacted

Also, I found out the hard way that completed spaceship components can be sabotaged - destroyed - by spies, which I wasn't aware was possible, thinking that they were classified as something like wonders and thus immune to espionage once built. Is this a bug or a feature?

 

[DarkLunaPhantom]

Also, I found out the hard way that completed spaceship components can be sabotaged - destroyed - by spies, which I wasn't aware was possible, thinking that they were classified as something like wonders and thus immune to espionage once built. Is this a bug or a feature?

It's a feature that also exists in BTS. Finished spaceship parts can be destroyed before launch. Additionally, if you lose your capital after launch, your spaceship is destroyed.

 

[<Nexus>]

It's a feature that also exists in BTS. Finished spaceship parts can be destroyed before launch. Additionally, if you lose your capital after launch, your spaceship is destroyed.

Wow! I never knew that either.
Is that also true about other projects too? And what if someone moves his capital before it would be destroyed? I mean you have an almost finished spaceship, build the Palace in an other city and a little later your formal capital is destroyed. Does that also destroye your spaceship?

 

[DarkLunaPhantom]

Is that also true about other projects too? And what if someone moves his capital before it would be destroyed? I mean you have an almost finished spaceship, build the Palace in an other city and a little later your formal capital is destroyed. Does that also destroye your spaceship?

This is true only for spaceship. Moving the capital yourself doesn't affect the spaceship, the spaceship is destroyed only if your current capital is conquered or razed by another player.

 

[f1rpo]

1. On huge maps, the AI is still founding a lot of sub-optimal cities. I think it made better choices in "better AI" and "K-Mod". Not sure what has changed, but there are still too many cities one tile removed from the coast and cities are often spaced sub-optimally

It's conceivable that the AI has trouble with exploration near potential city sites on Huge maps. The dense vegetation on Totestra might exacerbate that problem. The AI may then not be aware of seafood when founding a city. The long stretches of barren terrain on Huge maps and, again, especially on Totestra can also lead to fairly long distances in between adjacent cities, and the AI evaluation has problems weighing undesirably long gaps against a locally optimal position. I doubt that K-Mod places cities more effectively on any map; I see lots of poorly placed cities in K-Mod games. It's also possible that the issue is more one of opinion than of simplistic or erroneous AI programming. Some players seem to regard cities one tile away from the coast as rarely ever justifiable. Let me dig up some posts on that subject ... 1 (final quote box) | 2 (first quote box) | 3 (fourth spoiler box)
In any case, it would be helpful to have a screenshot (or savegame) from the point of view of an AI civ that is about to found a questionable city. For example, the attached screenshot shows a non-coastal AI city with a workable Crab resource. That city is, in my estimation, perfectly fine. Well, might've been good to wait 3 more turns for Bronze Working: there's a Copper resource just outside the city radius.

2.The AI is very reluctant to go to war before riflemen. Maybe this is because of the changes to diplomacy, where attitudes oscillate more around zero (except in the case of shared religion and favorite civics)? I've played maybe a dozen games on the latest version and have not seen a single axemen-era rush.

I don't think attitude is usually decisive for early wars. As I wrote in my last post, early warfare isn't quite as common as I'd like it to be in v1.03, and I've made a few more tweaks for v1.04. I might still add a mechanism that makes the AI a little more inclined to start a war when there hasn't been one in a long time and vice versa. However, an Axeman rush still isn't going to happen when civs aren't close enough to each other. I don't know what land-sea ratio Totestra uses and whether you use the New World option, but, just looking at the map dimensions, there are about 40% more tiles than the AdvCiv default for Huge world size. Starting around AD 1000 (i.e. well before Riflemen), wars have been quite common in my tests with v1.03 (and really any version of this mod). Don't know what's going on there; are those maps that enormous?

3. The United Nations elections have had some issues where a chairman or policy got the required majority but the result wasn't enacted

Could it have been vetoed? Holy war (but that's not a UN resolution) compels only full members in the mod. If there's actually a bug, then I'll have to wait for more specific information.

Thanks for this update about your experience with the mod and your kind words. Re-reading your post from August, has the Barbarian activity been less unbalanced since v1.00?

Also, I'd rather have the old barb behavior back: on large maps there is huge variance now in who gets the grease and who does not based on the surrounding terrain.

Regarding the Spaceship – for a bit of trivia – I've adopted a fragment of AI code from the LoR SDK mod a couple of years ago. Comment in the code: "if we're going for a space victory let's quietly relocate our capital away from the coast." I see this happen in all-AI games now and then. A little goofy and not always smart, but I think a thing like this can make the AI look a little streetwise.

Edit: fixed a link

 

[TheOtherOne]

2.The AI is very reluctant to go to war before riflemen. Maybe this is because of the changes to diplomacy, where attitudes oscillate more around zero (except in the case of shared religion and favorite civics)? I've played maybe a dozen games on the latest version and have not seen a single axemen-era rush.

I tend to play on small maps with low water and one extra civ. I find the AI is often quite hostile and very willing to take advantage of low power, undefended cities, current conflicts, etc. It is usually quite brutal and I've learned that if I start in the middle of a continent in a resource rich area I'm going to be facing non-stop conflict. I also don't play with the aggressive AI, so my experiences are with f1rpo's reduction in aggression. Interesting why we are getting different experiences.

Edit: I'm only now updating from version 0.97 so maybe changes have been made since then . . .

Edited edit: I've played a game through to the later middle ages (gunpowder and knights) on the latest version. So far there should have been several wars, especially as I play on maps somewhat denser than normal. 6 civs on a small continent in this case.

I was somewhat weaker than my neighbours and lagging technologically. We were all squished together with no room to expand yet I was the only one declaring war. They certainly defended their cities with the same tenacity with counter attacks and attemps to destroy critical resources, but they didn't decare war either against me or each other.

AI's attitude to me certainly was more positive (or, more accurately, less negative), but in previous versions having a somewhat positive AI didn't protect me from their desire to XXX.

I played a couple more games only investing in just enough military to protect from barbarians (raging barbarian setting) and the AI civs did declare war on me though not as early as I'd expect in the 0.97.
I know you said you wanted to reduce dog-pilling on weaker civs and games in 0.97 were brutal at times but I wouldn't be too worried about that. I was a little frustrated by how easily once-friendly civs would turn on me, but when I look at realworld geopolitics maybe that is somewhat appropriate : /

 

[TheOtherOne]

So I've downloaded the updated version of advciv and I'm reading through the changes. Some of my feedback:

V1:
I think it is fine to have barbarian cities with hill defensive bonuses. They are pretty tough to conquer so are likely to survive until catapults. I think this isn't a negative, I find it makes the game more interesting by having these hold-outs.

I like that you've made the diplomatic con of being a different religion is tied to the AI revealing cities. It means the effects aren't felt until civs start interacting with each other resulting in culture clashes. Nice change.


That's it. Sad to hear you aren't planning on doing much more. Though, now that you are looking at historical starts does this mean you will move onto something different (within civ4) rather than an end to modding?

 

[f1rpo]

So I've downloaded the updated version of advciv and I'm reading through the changes. Some of my feedback: [...]I think it is fine to have barbarian cities with hill defensive bonuses. They are pretty tough to conquer so are likely to survive until catapults. I think this isn't a negative, I find it makes the game more interesting by having these hold-outs.

Thanks, sorry to take so long to respond. Barbarian cities can still happen to appear on Hills, and the extra production that Plains Hills provide to the city tile is still taken into account. Probably not a significant change. The negative side of tough Barbarian cities is that they can randomly make a player's life a lot harder and that the AI isn't good at dealing with any Barbarian cities – will send too few units or too many or too far away or ignore nearby cities for too long. The AI had also been trying to "choke" cities economically that turned out to be too well defended. I hope I've at least fixed that issue (in some recent update).

I like that you've made the diplomatic con of being a different religion is tied to the AI revealing cities. It means the effects aren't felt until civs start interacting with each other resulting in culture clashes. Nice change.

Right, that was my thinking too. However, I wanted to avoid incentives for hiding cities from the AI because that might seem a little silly: "We mustn't let them see our disturbing cultic practices" – something out of a Lovecraft novel. So I'm actually letting the AI cheat with visibility and count all cities of the "heathen" civ for whom the diplo midifier is computed, and only unrevealed cities of third parties sharing the heathen religion are disregarded. I hope this still has the desired effect.

Sad to hear you aren't planning on doing much more.

The upside is that I won't be introducing many more bugs. Lots of fixes for self-introduced bugs in those old change notes, as you must've noticed. I don't think v1.05 will fix any major bug. (Although, a couple of sync issues in network games might qualify ...)

Though, now that you are looking at historical starts does this mean you will move onto something different (within civ4) rather than an end to modding?

On my own, probably not. No real change in plans since a year ago:

Once v1.0 is out, there'll probably be a new bug or two to fix, but, beyond that, I can't see myself as the driving force for further development. A couple of users have voiced interest in collaborating on more far-reaching gameplay changes, [...] which sounds like fun, but such a project will move along very slowly when every participant contributes only sporadically.

 

[DeepWell]

I think it may be worth a mention here. So I have completed a graphics mod for AdvCiv. It's an unitart mod that extends the ethnic styles all the way to the modern units. I have uploaded it to the subforum Ethnic Units Extended.

- - - - -

The upside is that I won't be introducing many more bugs.

That's a big upside imo

 

[keldath]

hey hey,

noticed 2 errors on the bug options.
take a look.

 

[f1rpo]

@keldath: Thanks. I can't readily reproduce either of them. Is this with the current v1.06 on GitHub? Or that merged into your own mod? Have Python scripts been reloaded due to a change at runtime? In any case, I think I'll have to lean on you for the debugging; I'll write a private message.

 

[f1rpo]

I'm attaching a beta version of AdvCiv 1.06. Don't want to upload it to the database straight away because some the UI changes need to be tested on operating systems other than my own. Tentative release notes: ... [edit2: deleted, see updated notes for release candidate here]

Edit: Attached screenshots to show the effect of the HUD scaling option. (I took those before figuring out how to shrink the resource bubbles.)

 

[keldath]

Amazing work sir,as usual.

I cant wait to merge the latest batch and run it .

Well done