AdvCiv-SAS (Simple Advanced Strategy)

[civ4-advciv-oracle-bug]

Update regarding next development version(s):

Unlike what i said just before, it seems i'll be developping AdvCiv-SAS a bit to quite a bit more at least if not more (i don't know for sure yet i mean), so it may all be a bust and i don't develop it much more, or i may continue more.

In all cases here is the latest AI enhancement i added in AdvCiv-SAS 5161 (development version) (commit https://github.com/wonderingabout/AdvCiv-SAS/commit/632163dc4b360293e2cf7eaf389bacc843bc44e1), with gemini 3 pro's help thanks anyways etc.:

- (Requires AdvCiv-SAS 5161+) **Aggressive tech trade discounts for widely known techs:** AIs now sell technologies known by ≥1 of their rivals at **heavy discounts** (e.g. 25% of the original cost/value), recognizing that it is better to secure **some gold** immediately via a cheap sale than to hold out and risk getting **0 gold** if a rival sells it first; implemented in `CvPlayerAI::AI_dealVal`. To compensate for this higher trading efficiency, **AI research costs** (`iAIResearchPercent`) have been adjusted (reduced/increased at higher/lower difficulties).

For details with screenshots of before/after, see KI#83:

 

[civ4-advciv-oracle-bug]

I also made a nice UI change so that the foreign advisors' screens now show more or less up to 12 players now without needing to scroll! Plus it is now more dynamic, so even if dimensions are changed, the more elements in the screen will follow instead of staying awkwardly stuck at old position with very suboptimal display but anyways etc.

It can be much enhanced still and is still the initial version (fit more techs/bonuses, expand the military advisor as well, etc.) (commit: https://github.com/wonderingabout/AdvCiv-SAS/commit/75085838bc151fa222e6ef127dfa7e143d2eb57f)

For now only in the development version.

If you want to test it, see: https://github.com/wonderingabout/A...ck_Install_Setup_Guide.md#development-version

 

[civ4-advciv-oracle-bug]

Small update on the progress: i have more or less finalized the foreign advisor's tech sub screen and the military advisor's screen UI relook and enhancement (including a refactor to make elements move mroe dynamically relative to screen size rather than absolutely anchored in a position which was less optimal), thanks to very nice help of gemini 3 pro thanks, and my own adjustements and empirical review or/and such.

Now displays much more information while still preserving as of now key relevant information for display (scoreboard with power ratios of rivals if hovered on, commerce sliders and their commerce values, map, etc.) without having to scroll or much less so than before.

Added related panels as i could etc. i mean but anyways etc.

Looks very nice so far but anyways etc.

 

[civ4-advciv-oracle-bug]

I also just added the button (i.e. icon) of the current thing we are building in the domestic advisor (F1 key) with the help of gemini 3 pro, which could very impressively succeed it at first try!

And it looks very nice thanks a lot hehe gemini 3 pro or/and such but anyways etc.

update: further beautify domestic advisor and some foreign advisor tabs: add iExtraWidth, iExtraHeight, iExtraY

 

[civ4-advciv-oracle-bug]

I also updated or replaced these ai-updated images with the help of nano banana pro, the result is amazing!!!

too bad civ4 can't show it though but still insanely good in a good way i mean if i may say but anyways etc.

See for details or related info anyways etc.: https://github.com/wonderingabout/AdvCiv-SAS/blob/tech-rework/README.md#ai-generated-images

 

[civ4-advciv-oracle-bug]

another very nice code i managed to add with gemini 3 pro and claude sonnet 4.5 and me too but anyways etc.: show icons for "willing to become vassal" and "vassal" rivals

it was not easy (at least to me) but together we made it xd i mean if i may say but anyways etc.

this fixes this issue i had named example 145 in base advciv 1.12 (using same save file from that time as well for these screenshots since they are handy and available for this need xd i mean if may say thanks to me too xd but anyways etc.), of this being very tedious or/and annoying to check or easily missed if not paying attention i mean but anyways etc.
Update: now also added a tooltip (on hover), in [CvDLLWidgetData.cpp](/CvGameCoreDLL/CvDLLWidgetData.cpp) with gemini 3 pro's help as well thanks anyways etc.
And now also using our own advciv-sas save file (see drive link for details anyways etc.)

commits:

- https://github.com/wonderingabout/AdvCiv-SAS/commit/e7d5d4d75f1db55bc785f8972556ff1a3ed8eec2
- https://github.com/wonderingabout/AdvCiv-SAS/commit/39fcce8ec0d9ebf9bf6d0a79ce604252bf4d434b
- https://github.com/wonderingabout/AdvCiv-SAS/commit/daa528a5571ef110a069fbb6da9b3e0d51be326f

See for details: https://github.com/wonderingabout/A...d-vassal-icons-in-foreign-advisors-glance-tab

 

[civ4-advciv-oracle-bug]

Also added a new specialists breakdown in the city screen, that helps avoid tedious hovering to see this info, with the very nice help of gemini 3 pro and chatgpt 5.2 and possibly 5.1 as well thanks but anyways etc.

Did also other beautification like adding a + great person N (icon) for each building that has it in its row in city screen, or/and tweaks/enhancements as well if any but anyways etc. The rework is still unfinished and can be polished further if i implement any more things but anyways etc.

commits:

- https://github.com/wonderingabout/AdvCiv-SAS/commit/9cd573cbecf65e55a4e86dc6a1db386dc754231b
- https://github.com/wonderingabout/AdvCiv-SAS/commit/8a47389a1793cf0bf9169165bf6562487e5c07b4

 

[civ4-advciv-oracle-bug]

i also just added the option to have extra rows in the city screen's production chooser (tunable in sas defines: https://github.com/wonderingabout/A...sets/XML/GlobalDefines_advciv_sas.xml#L42-L51)

adjust as you see fit, hopefully helps reduce tedious scrolling and increases visibility, at the cost of not seeing the selected units bar though but anyways etc.

Code added with the help of chatgpt 5.2 thanks and my help too i mean xd if i may say but anyways etc.

edit: note: outside display of the production chooser bar (by clicking without entering the city screen) is unaffected by the number of rows (behaves the same as before or if disabled) so even with extra rows it can still be used as the old way but anyways etc.

 

[civ4-advciv-oracle-bug]

i did also more beautification and expanded the foreign advisor's relation tab's web's radius and made it dynamic to the foreign advisor screen's resolution, with gemini 3 pro's help + also beautified the foreign advisor's cities tab

 

[winkler283]

I was looking at some of your changes to see if it would be possible to backport them into K-Mod.

Some of the bugfixes based on LLM suggestions seem questionable.

Example: https://github.com/wonderingabout/AdvCiv-SAS/commit/7f434b7fe0ae56f1e35f522a1d6f5107a9be8710

Why add all that code instead of fixing the root cause here?

return (*eBestSpreadUnit != NULL);

The fix is NO_UNIT instead of NULL.

return (*eBestSpreadUnit != NO_UNIT);

 

[civ4-advciv-oracle-bug]

edit3: also i looked into it and it was already in base advciv and so most likely both correct and not added my me (which makes sense since i didnt put a comment there) see below message

hello, i don't know too much about C++ so there may be some mistakes, however as a general rule i tried to be conservative and thorough as i could, and it is also why i tried to document any change exhaustively as much as i could or thought was relevant.

This change is one of the few where i only fixed some asserts that were firing a debug, without having otherwise encountered an issue myself.
Even though i don't know that much C++, isn't *eBestSpreadUnit a pointer (edit: actually it may be a reference rather, dusty C++ knowledge that again i don't know too much about xd i mean. But even then i find it hard to believe AI would make a silly mistake like this, so most likely i think it should be NULL and not NO_UNIT, but again tell me as i might be mistaken. Edit 2: no it seems this is indeed a pointer: https://www.geeksforgeeks.org/cpp/pointers-vs-references-cpp/), and NO_UNIT an enum?

If so it might make sense to leave it as such no? (again i don't know enough to tell about these more than that though and i may or not be mistaken here)

Hopefully if i made a mistake somehow people can see point it to me based on comments i left about said changes and possibly fix it if i find how.
In this specific case the assert stopped firing and it didn't seem like there were issues after this but i am open to feedback i mean though anyways etc.

Outside of these few asserts i fixed in debug without having encountered an issue in release otherwise myself, i believe there are plenty of fixes you may want to add to k-mod.

Some real bugs like infinite workboat loops (although i don't know if it was introduced in K-Mod, base advciv, or possibly even before that maybe) or such, or some smarter logic.

You may want to read the main guide or known issues listing to quickly pinpoint whichever fixes you'd possibly want to add in your mod (or K-mod):

- main changes guide: https://github.com/wonderingabout/A..._AdvCiv-SAS/Docs/README_Main_Changes_Guide.md
- known issues: https://github.com/wonderingabout/A...cs/README_Known_Issues_In_Base_AdvCiv_Civ4.md

if you are knowledgeable about C++, i have in this branch (https://github.com/wonderingabout/A...-related-crash-tentative-changes-stillcrashes) a real reproducible crash seemingly related to CvSelectionGroup::plot that i'd love to have a fix on, but don't know how, it happens rarely but consistently.

Thanks in all cases,

 

[civ4-advciv-oracle-bug]

also i looked into it and this was already in base advciv and so most likely both correct and not added my me (which makes sense since i didnt put a comment there) see below message

here is a screenshot of base advciv code showing it already existing (and as said before, even though i don't know a lot of C++, this is a pointer so a null check makes total sense here i think xd (but tell me if wrong i mean)) thanks,

and base advciv's github repo link as well: https://github.com/f1rpo/AdvCiv/blob/master/CvGameCoreDLL/CvCityAI.cpp#L9459

as you can see it is already part of base advciv, and makes sense from little i can tell as said before, so not added by me, and most likely not a mistake (AI would never make such a basic mistake as well, the kind of mistakes it does is more like complicated ones, not basic theory from the longest time i used it, it always made plausible mistakes, never basic theory ones from what i can tell i mean but anyways etc.) at all, but tell me if wrong i mean but anyways etc.

 

[winkler283]

This change is one of the few where i only fixed some asserts that were firing a debug, without having otherwise encountered an issue myself.
Even though i don't know that much C++, isn't *eBestSpreadUnit a pointer, and NO_UNIT an enum?

eBestSpreadUnit is a pointer to a UnitTypes value. In many programming Languages you write *eBestSpreadUnit to dereference a pointer to access the value it is pointing to.

If the function fails to find a suitable unit *eBestSpreadUnit is NO_UNIT which has the value -1. The original code checks if *eBestSpreadUnit is 0 but since -1 isn't 0 the function returns true instead of false. Replacing NULL with NO_UNIT makes it return false if the function doesn't find a suitable unit.

 

[civ4-advciv-oracle-bug]

Thanks for the explanation, then i will abstain from forming a conclusion as i don't know enough to tell.

If you think it should be so, i'd suggest testing it and seeing what happens.
On my end, i can only rely on AI or what other known knowledgeable source says so i'd rather leave it as such, especially since again i did not add this code myself.

The assert happened fairly frequently as well and was annoying (unless i was super unlucky multiple times), so i find it hard to believe both the AI was wrong as well as base advciv and K-Mod maintainers (whoever wrote this code) for so long, if so there may be a reason why it was as such.

I am not sure fixing the assert was necessary, but did as chatgpt suggested after reviewing it, and it seemed harmless afterwards.

If you can confirm this i may add it then, but without a reliable source and not knowing i'd rather leave it as is (again the issue was silent in release, we are only talking about an assert in debug i mean, so possibly there might be nothig to fix at all to begin with (i don't know to tell)). Thanks for your feedback though and open to further details if any but it seems to be as such for this specific issue thanks.

Although it is a minor issue, maybe @Nightinggale or some other knowledgeable could give some feedback if they see this as well, if not i'd leave it as such without further info, thanks for your feedback though and if you have other issues tell me then thanks

 

[civ4-advciv-oracle-bug]

with that said i mean, i am doing this tech advisor rework following recent ui rework commits and last about having more dynamic coordinates (that this code was actually the base for before i implemented it just now but anyways etc.).

the idea is to, for the most part for the panel:

- unlock more bottom space (ideally remove the half rows that are hard to read and not necessary anymore if we have more rows)
- show the side rivals research information (no need to toggle in/out)
- use more of the left side and top space as well as long as we can see the commerce sliders and such

then for the grid elements themeselves (i.e. for me i mean the techs here but anyways etc.), i have noticed we can gain some space by moving them to the left, which allows to entirely better the last tech column; not sure i'll go this way as we need enough space for the first column to remain readable, but it is a quick enhancement that looks quite good as well as long as not overdone i'd say as the space seems a bit too much i think/feel or whatever word it is but anyways etc.

is unfinished, but it would look something like this

btw i have found another base advciv bug it seems, as the chivalry tech timer does not refresh (need to click on another tech to research then click back on it to show its timer but anyways etc.)
edit: not sure if i had encountered it in base advciv, but i did had found weird tech advisor and other UI issues (example 175 for example in the following link of base advciv issues i had compiled (no pun but anyways etc.) at that time anyways etc.): https://forums.civfanatics.com/thre...-misc-suggestions.695343/page-7#post-16782814

edit: with the half rows removed, it would look like this (uncommitted still in the development version (for now only in this temporary branch https://github.com/wonderingabout/AdvCiv-SAS/tree/tech-advisor-ui-rework-2 anyways etc.), which i think is quite nice actually! But anyways etc.

 

[civ4-advciv-oracle-bug]

also, i added a sas define so players can set the tech advisor's width as they prefer (4 different values possible but anyways etc.) depending on how much they want of the scoreboard to be visible.

from sas defines (https://github.com/wonderingabout/A...ework/Assets/XML/GlobalDefines_advciv_sas.xml), as of now:

Spoiler sas defines values for tech advisor's width (0 (as of now default), 1, 2, or 3) :

<!-- custom: choose how wide the tech tree should be horizontally:
-> 0 (default): the tech tree goes from leftmost part of the screen (tightly compressed left margin as there is less room), to about halfway in the scoreboard on the right: we see leader name, the tech they are researching (if we know it), their tech %
-> 1: the tech tree goes from the leftmost part of the screen (usual left margin though since there is room) to about half of the scoreboard on the right: we see about half of the player names on the scoreboard unless theirs is extremely short (e.g. "Joao II" not visible, most are visible partially), plus the tech they are researching (if we know it), their tech %.
-> 2: the tech tree goes from the leftmost part of the screen (usual left margin though since there is room) to about 2 thirds of the scoreboard on the right: we don't see player names on the scoreboard anymore, but we see the tech players are researching (if we know it) and their tech %.
-> 3: the tech tree goes from the leftmost part of the screen (usual left margin though since there is room) to the rightmost part of the screen on the right: it provides maximum immersion visually as potentially distracting info is removed, but we don't know anymore their tech (and the world map is now hidden as well).
-->
<Define>
<DefineName>SAS_CV_TECH_CHOOSER_HORIZONTAL_DEPTH</DefineName>
<iDefineIntVal>0</iDefineIntVal>
</Define>

commit: https://github.com/wonderingabout/AdvCiv-SAS/commit/1aeb7b55b9081b84da3248b4c481991eb1a522c0 (with other fixes or tweaks or other changes in the commit but anyways etc.)

output example below anyways etc.

next i would adjust the techs' rows i guess if i do it, but it may be quite the extensive rework, so maybe i'll continue doing UI reworks rather and possibly minimal tech adjustments, and then after next release since we have more extra room vertically and no longer need to have a tight clutter may do a more extensive rework then (if i do it hopefully but not promised nor guaranteed xd i mean but anyways etc.), with some ideas from gemini pro 3; the current version is so far still just a draft i mean if i'd ever do it more but anyways etc.

 

[winkler283]

On my end, i can only rely on AI or what other known knowledgeable source says so i'd rather leave it as such, especially since again i did not add this code myself.

Never C++ Compilers give you a warning about this kind of issue.

Compiler Explorer - C++ (x86-64 clang (trunk))

// Type your code here, or load an example. enum UnitTypes{ NO_UNIT = -1 }; bool test(UnitTypes* unit) { return *unit != NULL; }

compiler-explorer.com

The CivIV modding community uses a ancient version of the MSVC compiler which has less advanced diagnostics.

Being able to ignore asserts is also not helpful. If the code doesn't handle something and a assert gets triggered the program should crash forcing the developer to fix the code.

 

[civ4-advciv-oracle-bug]

Like i said, i am in no position to tell since i don't know enough about these, as for this technical question you should really ask more knowledgeable people than me and go with whatever you find best fit.

What i can tell you since i added plenty C++ code (with AI help), is that regularly there are asserts that are just there just because, in the existing civ4 code; i saw plenty K-Mod humbly written by K-Mod maker themselves as very bad or not ideal, and while i don't remember in detail, i guess BBAI code too. Base advciv also has asserts that are there just to check stuff, really as long as it works and is sane i have no issue with an assert firing or not, it's not an issue ingame (the base advciv assert was something related to conquered players, it was litterally there just to check if the code fires nothing more).

Also i'm not too interested in technical debate, as i don't know enough, and i'm more focused on making functional and clean code as much as possible. Since i didn't add this code myself, i have no reason to change it, i only tried to fix the assert firing (while btw adding a nice != NO_UNIT guard that was missing as chatgpt nicely noted).

You can find examples of very effective code i wrote in the settler and to some extent the worker rework as well (edit: even though it's a bit messier, the core logic is super smart and much much stronger than base advciv one), they are much much stronger and smarter in advciv-sas. I don't think there is that much value in arguing too deep about technicalities, what matters most is code works, is clean enough, and improves players' experience (i want a stronger and smarter AI opponent the most, or QoL tweaks).

On that end i'm doing the UI reworks, have a look at them if you'd want them in your mod, i believe they are quite nice as well.

I'll try to look quick to show you but if i cant find it quick it's not too valuable of me to look too deep into it, but really having seen plenty civ4 code myself so far, the code is not always nice and shiny, but it's effective and clean enough, i like it tbh but anyways etc. , hopefully helps.

trying to look quick into it: can't find it quickly enough and not so worth the time to look for me, but i believe you can find it fairly quickly if you run latest base advciv 1.12 with a debug dll, around turn 150s (and plenty other asserts firing before you even start a game). I would suggest not spending too much time on these however do as you see fit, on my end i hope i provided enough info to help you with whatever you were wanting to do, as for me i'm implementing UI things, tech rework, or doing whatever else atm, thanks.

 

[civ4-advciv-oracle-bug]

and with that said, i have made my draft changes (quite conservative for now if not always or not but anyways etc.) for next i.e. this release i think (next = the one i'll make after 5160 i mean but anyways etc).; which i am satisfied with it seems/i'd say but anyways etc.

with the very nice help of chatgpt 5.2 and gemini 3 pro who helped me a lot thanks

so far only ancient and classical have been reworked and it's a draft as i said so ont implemented but logic mostly figured out and enhanced i mean but anyways etc. Notably music theory is removed as ancient was bloated and medieval a bit too thin, hoepfully balances pace as well for longer medieval as i'd want and fater late classical as bit slow as well + other balance changes in tech tree as shown i mean but anyways etc

added old one of previous commit i mean for comparison as well for vs fighto xd i mean if i may say but anyways etc.

 

[Osakasayama]

I tried a game on one of the recent commits. Here's my thoughts.

IMO, the AI has become smarter but most importantly, less frustrating compared to older ADV iterations. I'm relieved that they aren't doing stupid archer rushes anymore, and that they peace out properly after wiping their stack.
Not sure if I like the other changes though, especially the tech tree, buildings, and unit changes.

I don't expect you to implement these right now, but it will be great if you prioritize the bolded ones as they're making the game somewhat unplayable.

Units​

• Too many units have bonues on banal terrain & features, like plains grasslands etc. This is just confusing and cluttering.
  • The bonuses don't make sense for many units too, e.g. Samurai having a tundra bonus
• Why does the Quechua and Jaguar now require metal?
• Since crossbows require metal, civs without it are stuck with 4 str "longbows" until muskets.
  • If you need a justification, I'm pretty sure trebs and catapults use as much metal IRL but they're resourceless
• What is the point of the 2str "Archer" unit? Why would one build it over the warrior or the scout?
  • I know they have different terrain modifiers and stuff, but since the AI isn't retardedly aggressive early game anymore, one should just wait till the 4str "Longbows" come online.
• "Ancient Maceman" is lame. "Warrior" (or "Militia" in RFC DoC) was fine, why did you bother changing it?
• "Bazooka" should be "Anti-tank." Bazooka is the name of the Americanweapon used in WWII, not the military units using them.
  • They also need to be buffed so they'd have more than 50% against tanks (BtS issue).
• The Dutch Switf worker is barely distinguishable from the Indian worker. Can either of them be changed (restore the East Indiaman or that Longbow UU in RFC-DoC)?
• Camel units shouldn't be as powerful as horses outside of deserts, for the sake of historicity.

Suggestions​

• Remove the 2str Archer, rename the 4str Longbow to "Archer"
• Add some sort of a "Bombard" unit in between trebs and cannons, since Gunpowder tech doesn't unlock anything rn
• Buff Pikes, so they'd be a unit in between Maces and the 12 str Muskets
• Nerf Dragoon base strength (to prevent AI overbuild. Pretty sure line infantry caught up with cavalry by Napoleonic - Victorian eras)
• Nerf War Elephants (BTS issue)
• Nerf SAM infantry base strength to something closer to Anti-Tanks (BtS issue, to prevent AI from overbuilding it)

AI​

• AI don't colonize overseas continents as fast as they used to?
• AI isn't building wonders at all

Misc​

• The unlimited Engineer perk from Wage Labour civic is OP, it allows so many GEs = 1 turn wonders
• Ports are OP, should be delayed till later eras (e.g. astronomy or steam engine)
• Feitoria is missing the free merchant bonus from its generic building
• Unit icon hover menu doesn't display all the unit bonuses, e.g. matchups against unit types / classes
• I guess its WIP but the tech tree has many choke points where only 1-2 techs are researchable
• Astronomy comes way too early. Galleons shouldn't be a thing until gunpowder, at least.
• I guess this is also WIP, but the build icon list in the city menu are bugged