The proper release will have at least one more bugfix: The relations boost from the favorite civic starts at +0 instead of +1. This isn't intentional; I had only meant to reduce the maximal relations boost for some leaders (which is working too). This has been broken since v1.01 and was also responsible for the FavoriteCivicsDetector Python errors reported by @Jorunkun – the BUG utility for figuring out favorite AI civics when playing with Random Personalities relies on favorite civics always boosting relations by at least +1. On that note, it's pretty silly that players (or BUG) have to infer and guess what the favorite civics are when it's supposedly something that the AI leaders are trying to promote. Should probably be based on the visible leader head, not the hidden personality; favorite religion also. But I don't think I'll change that for v1.06, if ever.
As an aside: I use a mod-mod of AdvCiv which includes ~20 changed/new files and have not found a really good way to merge these mods yet, other than checking on a file-by-file basis using winmerge. Any advice to streamline things would be highly welcome.
If you're comfortable with Git (I guess you wouldn't ask if you were) or wish to familiarize yourself with that more, then you could use whichever Git client you like to clone my GitHub repository (just the most recently changed branch would be enough, currently "1.06"), apply your changes and commit them in a local branch, and let Git merge my latest changes whenever you want to incorporate them. Only conflicting changes need to be handled manually, i.e. when I've changed just the same lines of code that differ in your branch. You could also, from the outset or at a later point, fork my repository on GitHub (in the browser, will require a user account) and push your mod-mod for public access or as a private backup.
I could post some screenshots, not exactly as a guide, but at least as an example. Caveat: I'm not really savvy either. And I don't know if I'll update the mod often enough in the future to justify this initial investment of time on your part.
For WinMerge, @keldath might have some advice; he's been keeping up to date with AdvCiv for quite some time, using mainly WinMerge. Though I don't suppose there's much that can be optimized about that workflow(?). GitHub can list files that have changed between two commits, e.g. https://github.com/f1rpo/AdvCiv/compare/2b044408dcb8b6fd6974f5ebb7501251c510917c...1.06
comparing the latest release candidate (commit ID before the three dots) with the latest commit on branch 1.06. But that's not really something that WinMerge doesn't do as well.
well, yup, im sort of an expert in keeping up with advc.
winmerge is my superpower.
i do everything with it.
my method is,
every few days, download and copy the updated advc repo and the relevant branch.
then, i compare it, to the previous iteration i saved up.
this comparison shows me all the differences between my latest updates to my code,
i save the advc folder as dates, so ill know when was the last merge i did.
while winmerge is opened with the difference between latest commits to last one i saved,
in my other half of the screen, i got my files compared with the latest commits version .
then i know what to merge and which line according to the avdc new to old difference.
this is how i do it in a very fast way, for 2 years or so
basically i got many many "build" of advc , as i dont delete, i archive.
sometimes i can update the latest folder every few days, sometimes, a month, depends how connected i am and in the mood to do some civ4 code.
in addition, i keep my own git with the updates and branches.
Thanks f1rpo and keldath. I occasionaly use git for work, but I don't want to interfere accounts and did not figure in a reasonable time how to avoid this [btw, git is strange, great in some ways, but incredibly obscure in others]. I'll try as keldath suggested.
It's always been that way in AdvCiv. I had looked at the code before and didn't see a problem, but, having taken another look now, I think the problem is that the scoreboard columns with just a gamefont icon aren't affected by the mod's increased font size and therefore have a different line height and need to be positioned at a slight offset. Anyway, an offset for those columns does the trick.
Can't reproduce that; screenshot attached. I've tried a smaller screen size too, same result. Also haven't been able to reproduce the issue that you mentioned in a PM with the GP bar not appearing unless at least two specialists are assigned that sounds like it might be related. Edit: part of reply moved to private conversation
Sorry for the long radio silence, had to travel for work.
>> Delighted to see the latest RC and fixes - will install it right away and give it a spin this weekend. Thanks, f1pro - and kudos for sussing out the favcivdetector bug!
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)
I'm open to trying Pefect Mongoose, but would ask your advice on the number of civs. I play Totestra with 18 civs because it works out about the same (a little above) land-to-civs ratio that you get on the standard map size with 8 civs. With f1pro's the latest changes to map scripts, 16 may be a better fit. That said, I think Totestra maps are more Earth-like wrt rivers, landmass shapes and island chains and kinda like having the extra challenge that comes with the more extreme clustering of terrain and resources and would urge you to give it a try. Preferences, suggestions?
Re: Vassals, I prefer No Vassals as the game gets a lot more difficult with them given these settings, but if this is a deal-breaker, we can switch them back on.
Don't have strong preference on AI personality - I set it to random because I feel it constrains the game knowing that, frex, if you meet Tokugawa, you know that he will, by default, not open his borders. Random probably weakens AI performance, but not knowing their playstyle also weakens yours, which to me is good trade-off. YMMV.
Maybe we can wait for the next full release and give this a whirl in August? Would be happy to invite others over on the main forum and on reddit if you are interested - part of the reason I proposed this was to popularize Adv Civ.
Thoughts?
PS: It warms my heart to see the discussion here picking up again - pleasure to meet all you loonies.
hey,
well i had never played Gotm, how does it work?
August is fine, assuming i understand how to play
p.s i prefer standard sizes, my pc cant handle huge, plus it takes a much longer game and commitment....
hey,
well i had never played Gotm, how does it work?
August is fine, assuming i understand how to play
p.s i prefer standard sizes, my pc cant handle huge, plus it takes a much longer game and commitment....
Basically, we all start from the same 4000 AD save and compare notes as we play.
There are quite elaborate rules for the official GotM about how and when to post updates, and a tradition of discussing strategies before the game starts (given just the screenshot of the starting location) but what I'm looking to get out of it is to see how different people approach the game in different ways, and how the world develops on different tangents as a consequence. It'd just be neat to see the same game / world played out side by side. I'm sure we'll all learn something about our respective playstyles, and how Adv Civ works.
As for huge, are you sure your PC can't handle it? What's the issue?
Maybe we can wait for the next full release and give this a whirl in August? Would be happy to invite others over on the main forum and on reddit if you are interested - part of the reason I proposed this was to popularize Adv Civ.
Full release done. August still sounds good, I could get some more tweaks to Marathon and Barbarians done by then. I'll write another post about that and Totestra/ Mongoose later (today or tomorrow).
For those who have followed the release candidates, here's the (short) Git commit history since v1.06rc2. Edit: Maybe I should better point this change out explicitly:
1.06 release notes said:
• The mod's name is written into replay files, meaning that the BtS Hall Of Fame (HoF) screen will no longer list replays created by AdvCiv. There is an option on the System tab of the BUG menu for restoring the behavior from previous versions of AdvCiv, i.e. to omit the mod name so that the BtS HoF will list AdvCiv replays. I've come to think that players may find it intrusive that AdvCiv sneaks its replays into the BtS HoF by default. [advc.106i]
I'm open to trying Pefect Mongoose, but would ask your advice on the number of civs. I play Totestra with 18 civs because it works out about the same (a little above) land-to-civs ratio that you get on the standard map size with 8 civs. With f1pro's the latest changes to map scripts, 16 may be a better fit.
I was going to write that, since I've already made all the decisions about the non-customizable parts of the mod, I really don't need any say over the game settings. However ...
Would be happy to invite others over on the main forum and on reddit if you are interested - part of the reason I proposed this was to popularize Adv Civ.
... if you phrase it like this, a third-party map script that (unlike most map scripts) takes charge of resource placement and starting locations strikes me as a flawed way of introducing the mod. The algorithm for placing starting locations is a major feature of AdvCiv. I also don't see how Totestra with 18 civs (without starts being restricted to the Old World) lands just "a little above" a standard map in terms of crowdedness. As far as I can tell, it's nearly 50% more resources per civ than the AdvCiv standard settings and closer to 100% more land per civ. So that's also too far out in my book, seeing that the available space for expansion has major implications for the game's dramatic arc – and for Barbarian placement, another focal point of AdvCiv.
Here are some statistics about Totestra, Fractal and PerfectMongoose maps. This should be helpful for discussing player counts and differences between the scripts. I've regenerated each of map once to get a sense of the variance too.
Spoiler:
Totestra, 18 civs, Huge, Medium sea, Start Anywhere, Allow Pangaea
[...] I think Totestra maps are more Earth-like wrt rivers, landmass shapes and island chains and kinda like having the extra challenge that comes with the more extreme clustering of terrain and resources and would urge you to give it a try.
I see your point about the clustering. The landmass generator is essentially the same in PerfectMongoose, which is to say that neither script has changed the performTectonics function from PerfectWorld 2 much. While some parameters do differ, I believe the maps differ mainly due to the smaller grid size in (the AdvCiv version of) PerfectMongoose. (That is, so long as the optional PerfectWorld 3 land generator based on Perlin noise isn't used.) Well, I'm not sure if a large enough PerfectMongoose map would also have those Aleutian island chains; maybe I should look into that. Generally, all of these scripts are just customized and enhanced versions of PerfectWorld 2. One issue with Totestra that is easily visible on this screenshot is that often only one starting location is non-coastal. This is because the starting location algorithm tries (hard) to maximize the (path) distances between the starting locations. Another thing that irks me in the screenshot is that the tropics are represented as almost 100% Grassland river; that's realistic only in some superficial sense. And then there's the absence of a tree line near the poles.
I hope this post is not coming across as, say, verbosity for the sake of assertiveness. I intend to (partially) participate in this community game either way. At the end of the day, if it's you sharing your favorite way of playing Civ, then the settings have to be true to that, and it's very cool that AdvCiv has a part in this. I guess I'd just like to suggest that AdvCiv + Totestra, with your settings, is so far removed from the standard experience of AdvCiv that it's really its own thing.
To this end, it would be nice if we could arrive at a compromise. (Though, again, if that's really bridge too far, then let's not compromise.) Without Totestra or another external script and without changing the mod, Low-sea PerfectMongoose is probably the closest we can get to what you have in mind. Here's an example of that:
To get a first-hand impression, I've played about 250 turns on such a map, though I used the Perlin noise land generator for this test and Emperor difficulty – other than that, with your proposed game options. Here's a screenshot on turn 248:
A short play report – perhaps helpful for gauging whether you can get on board with this, and also relevant for some changes to Marathon speed that I'm considering.
Spoiler:
Disabling "Wait at End of Turn" helped with the flow during the first 100 turns. Early scouting is a pretty boring affair when there is neither time pressure nor goody huts. Just need to make sure not to run the free initial Scout (as Persia) into an animal. Well, can't really ensure that. At one point, it was luckily spared by a Bear.
Just optimizing city placement and tech paths without having to pay much attention to the AI (abundant space, no tech trading) has its charm. In this game, Frederick (whose true personality I don't know), actually started just 10 moves away from me. Close enough for me to place my 3rd city, Susa, in a potentially contested spot. Shortly after that, I rushed him with Immortals. AdvCiv has nerfed those a little bit, but taking Berlin this way has still felt like shooting fish in a barrel. Not sure how much the disparity between unit production/ movement costs and research costs has contributed to this. Maybe it's mainly just a case of Frederick not having a strategic resource nearby. I razed another German city later to make room for Tarsus, and I feel I could still quickly produce some more Immortals whenever I want to take more cities. But it's probably more worthwhile to expand westward and leave Frederick in place as a harmless, jungle-dwelling buffer. The peculiar mountain corridor to the west, along with early access to Immortals, has made the Barbarians very easy to handle thus far.
Compared with Totestra, the uniformity of the terrain is not as extreme here. There is a pretty big temperate, mountainous region in the center of the screenshot, but I've also already expanded into the jungle and may expand into the semi-arid region in the southwest. I would be open to making the terrain more uniform, if I can figure out how, maybe as a PerfectMongoose option, or just for this one occasion. I don't think this conflicts with the goals of the mod. Could also tweak it for more reliably bulky landmasses (it varies quite a bit currently). Resources being placed pretty far apart (more so than on BtS Fractal) could be amended by disabling the "SUBLINEAR_BONUS_QUANTITIES" setting in XML. Should then perhaps go to 18 civs. The river distribution doesn't seem fundamentally different from Totestra here. A lot of them in the jungle, some in the temperate part, none in the southwest (tropical desert/ rain shadow thing).
Regarding Marathon speed, my tentative intuition is that the 300% modifier should be reduced to 250% while leaving the 200% modifiers alone. The modifier for the Barbarian placement rate (not speed-adjusted in BtS) should arguably be 200%, i.e. half as fast as on Normal speed. Currently it's 300%; I don't know why I went with that, probably out of carelessness. The placement rate should match the unit production speed of the civs. Maybe a case could be made for a somewhat slower rate, seeing that the civs will also want to construct some buildings in between dealing with the Barbarians. But I think it would be better to err on the side of dangerous Barbarians because, in BtS, Barbarians are brutal on Marathon. Would like players to feel that, while much more manageable and reasonable than in BtS, Marathon Barbarians are still tough in AdvCiv. Another consideration is that Marathon is often combined with Huge world size. The default civ count (16 on Huge) gives each civ a bit more space than on smaller map sizes, so Marathon games perhaps tend to have more room for Barbarians. Maybe that justifies a 225% adjustment. But that's still a 33% increase in Barbarian activity over the status quo, and I don't want to ruin those Totestra games:
On the barb side, I think the settings are tough but playable. The AI certainly suffers from them as much as the player does. In every game, I've had two or more "failed states" due to early barb invasions. These are Civs that have had one of their cities taken (often from galleys) and stay on the back foot for many centuries thereafter. I think this is fun, but barbs sure make for a more swingy game. If you lose one city to them, it can be a major setback and they are really hard to dislodge once they build walls or connect a strategic resource. But again, I like it like this.
But I also don't want to adjust Barbarian activity to the crowdedness of the map because I would like players to have some control over Barbarian activity through the map settings.
If you were to make barbs a little easier, you might want to consider turning off barb galleys spawning a new crew in the FOW and only allowing them to be crewed in barb cities. Also, you might want to have a lower chance for spawning axes and swords until later in the game.
I've already implemented, for v1.07, an increased probability of spawning Barbarian Archers at the expense of metal-based Barbarians. And I intend to place some checks and balances on units placed aboard Galleys. Hopefully, these changes will keep the Barbarian threat at a tolerable level even on (very) sparsely populated maps.
On a different note, I think there's a bit of an issue with the AI not taking full advantage of units being relatively cheap on Marathon. Much of the AI code for deciding how many units to produce doesn't look at the price tag; it just says e.g. "this city is important, we want at least 3 defenders here." But if they're cheap, perhaps 4 should be aimed at. Same for the size of city attack stacks. So I think I'll want to adjust those numbers to (or toward) the ratio of unit production modifier divided by tech cost modifier. Unit production modifiers from the difficulty level are already taken into account in those code locations. A slight adjustment to the personality-based "BuildUnitProb" might also be in order for Marathon; this has a big impact on how many units the AI produces overall. That said, if the 300% modifiers are reduced to 250%, then I might want to adjust AI unit production just by 10% to 15%; not such a big deal.
I'm also thinking about adjusting the culture level thresholds. The delay between finishing a Monument and getting a border pop feels very long (already feels long on Normal speed); Perhaps culture levels, except Legendary, should use the 200% modifier. Admittedly, growing extra population to work the expanded city radius is just as slow (300% growth modifier), but, often, the player wants to assign existing population to tiles in the outer ring.
Opinions on these ideas? Or better wait with that until after our game?
Well, I'll play fast. Those 250 turns weren't too bad. Things get exponentially worse, of course, but another 350 turns would already be AD 1550 if my Excel sheet is correct. By the way, would have to do the game year progression for Marathon over if I reduce the 300% modifiers.
Don't have strong preference on AI personality - I set it to random because I feel it constrains the game knowing that, frex, if you meet Tokugawa, you know that he will, by default, not open his borders. Random probably weakens AI performance, but not knowing their playstyle also weakens yours, which to me is good trade-off. YMMV.
To be clear, the personalities just get shuffled around, so the AI leaders are not getting, say, incoherent personalities. Actual randomized personalities would be a nice option. I don't much care for deducing who each AI leader "truly" is, but this doesn't totally ruin the option for me, so it's fine. The AI isn't going to guess at who is who, but won't cheat either, so it'll e.g. assume that everyone might start a war at Pleased attitude. That puts the AI at a very slight disadvantage, but, I agree, human players not being able to take full advantage of knowing the AI personalities may more than make up for that.
Thanks, I've found the problem. Fortunately doesn't let players upgrade their units without paying, so I think I'll delay the fix until v1.07, i.e. until next month hopefully.
I've implemented pretty much all the stuff I had outlined in that long post about Marathon Perfect World games. Could release that as AdvCiv 1.07, release candidate attached. But I'll wait with that until @Jorunkun has weighed in.
[Edit: Draft of release notes deleted now that v1.07 has been properly released.]
An addendum regarding Totestra and PerfectMongoose: I've run a diff of Perfect World 2.06 and Totestra, and, so long as most of the custom map options aren't used, the differences are very few and minor. So the discussion about PerfectMongoose vs. Totestra is really one about the merits of my changes to the script and those that LunarMongoose backported from Perfect World 3 (Civ 5 version of Perfect World). Looking at it this way, sure, I'd much rather tweak those changes than fully revert to Perfect World 2 for any purpose.
@f1rpo Sorry for the long pause, work interfered again - and thanks for your well-reasoned post on maps, and the latest changes to marathon. I have a day off today and will try your new RC and the map settings you recommend for a huge Low-sea PerfectMongoose map.
My main gripe with Mongoose is that the smaller huge map size - or rather, resolution - makes the landmasses look swollen or blurry, with less clearly delineated coastlines and islets. Totestra maps are somehow more crisp. But, having seen your numerical comparison, I agree that resource placement and clustering as well as starting locations are problematic with Totestra - having played some 50 or so games under my settings I've observed that the start is the single biggest driver of who eventually comes out on top, and disparities are often huge.
So even now someone tries to improve AI? That's awesome.
Even though I lost the HDD with my dev setup more than a decade ago, and barely played Civ4 since, I had still always hoped someone would at least merge in K-mod's improvements.
While I do have a new favorite game now, I'm sure I'll play Civ4 again at some point, and when do, I'll definitely use AdvCiv
This site uses cookies to help personalise content, tailor your experience and to keep you logged in if you register.
By continuing to use this site, you are consenting to our use of cookies.
