The Touhou Mod Development Thread

[nvdr]

Actually regarding balance of UAs, I'd like to bring up Yukari. All of her uniques are basically designed to allow her units to use the gap concept, but she doesn't really get any bonuses that make her viable when you compare her to civs like Nitori, Utsuho, Eirin, Makai, etc.

Sure she's got some distinct advantages in terms of being able to do conquest, since she has indirect fire by default and can paradrop really early (including her boats somehow) but otherwise she's really underwhelming for any other victory, and she still pales in comparison to Utsuho, Nitori and Suika for war simply because their uniques yield larger advantages.


Also, while it's only a small thing, I think Alice actually could work as a "poccy" civ. Although it'd probably require some reworking of her uniques since they're designed more for a normal civ at the moment.


don't nerf anyone that's OP just buff anyone that's underpowered imo

 

[bouncymischa]

don't nerf anyone that's OP just buff anyone that's underpowered imo

I could write a lengthy essay on my opinions regarding rebalancing underpowered and overpowered things... X3

In the end, though, I prefer to just balance towards the middle, particularly if one considers the official civs to be the desired balance.

After a handful of games against and as the Human Village and Makai I'm getting convinced they're just a touch overpowered. Makai's extra production is a big deal, and stronger cities mean it's that much harder to stop them before they go out of control, even if I go to stop them ASAP. I'm not sure what Keine's problem is but she becomes a diplomatic/tourism/domination powerhouse in ways I've never seen any other AI achieve.

Moriya Shrine and Oni Clans are fun to play with/against, the Kishins are a fun way to get golden ages from war.

With regards to Makai, I've been debating if I should tweak them a bit. The production bonus was added at the last minute, so I never really tested it as fully as the other aspects. On the other hand, I think the production bonus is good for her, given the discussions we had regarding how Shinki works better with a desert start and Petra, rather than tundra.

So! I could remove the city strength bonus -- it was originally just tossed in to help Shinki get into the later game when her tourism bonuses started kicking in, so it's mostly something of an afterthought. Another idea I've been considering is getting rid of the production bonus from the Crystal Forest, so it just produces culture and gold. It would help tone down her production output a bit, while still making Tundra as appealing (or possibly moreso) than a desert start for her.

As for Keine, it's kind of odd because she doesn't really have any super-bonuses... just a collection of weaker ones. I suppose between the food and gold bonuses, it just gives her a really open-ended playstyle where she can go whatever direction she wants...

I guess it just goes to show you shouldn't underestimate humans?

 

[bouncymischa]

So... currently, I'm making progress with Marisa!

I'm feeling some concern that she's not going to synergize well, though, as her current uniques feel a bit scattered. Reading up on Marisa, two aspects to her stood out in my mind -- her ability to learn from her opponents and replicate her attacks, and the nature of the Forest of Magic to deter both humans and youkai from venturing into it, particularly due to the numerous mushrooms there that form the basis of Marisa's magic. There is also Marisa's tendency to steal things from people to consider, as well...

I'm currently using "Stealing the Precious Thing" as her UA, giving her spies a bonus to steal technologies, and she researches techs already researched by her friends faster. It does seem to replicate the "duplicates her enemies attacks" aspect of Marisa nicely.

The Atelier UB currently works kind of the like the Seidi UB from the Sami civilization -- it reveals the presence of Mushroom Patches in surrounding forest, jungle, or marsh tiles. I haven't figured out exactly what the Mushroom Patches will do, but they'll provide both enhanced yields as well as a defensive bonus, in the form of a debuff applied to any enemy units that pass near them. I've got art for a variety of mushroom patches, so considering having three random types that generate different yields -- most likely food, production, and science.

Marisa's UU is a Witch unit... but I'm not entirely sure what to have them do. One idea was to have them venture into foreign territory like the Trickster, and then get a gold/science boost when they returned to friendly territory, to reflect Marisa's tendency to steal things. The main downside is I'm not sure the AI could really use this effectively, and it would likely mean making yet another Scout UU. One alternative that came to mind was to instead focus on Marisa's spellcasting, and make the Witch a Cannon UU (since I'm planning for the Mushroom Patches to upgrade with Chemistry, making it an appealing tech). As a Cannon replacement, they'd likely be able to fire without setting up, making them more mobile, and it would play into the "defensive" nature of the Forest of Magic. I've also considered having them act like Missile Carriers, able to carry a limited number of "Master Spark" spells. Another possibility is that Witches stationed in cities could reveal additional Mushroom Patches...

So which would be the best direction to go with Marisa? A defensive civ that can catch up on technologies fast, using Witches to bombard enemies into submission while protected with their abundant Mushroom Patches? Or should Witches be more mobile, scout units that can steal money and science from other civs -- but may not work as well for the AI? I think the main question at this point is exactly how I want to set up with Witch UUs. The Cannon version works better with the idea of a defensive civ, and I haven't really made one of those yet. On the other hand, the "stealing" Scout version might be more like what people imagine Marisa is like... although less synergistic and being somewhat similar to the Trickster...

 

[nvdr]

The Atelier UB currently works kind of the like the Seidi UB from the Sami civilization -- it reveals the presence of Mushroom Patches in surrounding forest, jungle, or marsh tiles. I haven't figured out exactly what the Mushroom Patches will do, but they'll provide both enhanced yields as well as a defensive bonus, in the form of a debuff applied to any enemy units that pass near them. I've got art for a variety of mushroom patches, so considering having three random types that generate different yields -- most likely food, production, and science.

The randomized mushrooms seem like the best idea to me, because I like the idea of not just having one yield from them, after all, Marisa wouldn't use just one kind of shroom.

Another thought would be to have the shrooms be like the Blackfoot civ's Buffalo and have bonuses come from improving the shrooms.

Either way, Marisa's shrooms end up mostly being stolen from other custom civs and I think that's amazing.

Marisa's UU is a Witch unit... but I'm not entirely sure what to have them do. One idea was to have them venture into foreign territory like the Trickster, and then get a gold/science boost when they returned to friendly territory, to reflect Marisa's tendency to steal things. The main downside is I'm not sure the AI could really use this effectively, and it would likely mean making yet another Scout UU. One alternative that came to mind was to instead focus on Marisa's spellcasting, and make the Witch a Cannon UU (since I'm planning for the Mushroom Patches to upgrade with Chemistry, making it an appealing tech). As a Cannon replacement, they'd likely be able to fire without setting up, making them more mobile, and it would play into the "defensive" nature of the Forest of Magic. I've also considered having them act like Missile Carriers, able to carry a limited number of "Master Spark" spells. Another possibility is that Witches stationed in cities could reveal additional Mushroom Patches...

I do quite like the idea of the "missile carrier" witch, although I think it perhaps could get more use with Mima, assuming the "buildable" spells had more variety with her.

The cannon idea seems good as well, after all, the Master Spark (well, Final Spark, really) is literally called a Magicannon. It could definitely work very well as a defensive thing given what you're apparently planning for the forest concept.


Another thought regarding cheaper tech: I have no idea if it's possible, but it'd be cool if she got the bonus from neighbors or opponents in a war as well. (Maybe even a small chance to instantly acquire a tech known by your foe that you don't know?)

I could write a lengthy essay on my opinions regarding rebalancing underpowered and overpowered things... X3

In the end, though, I prefer to just balance towards the middle, particularly if one considers the official civs to be the desired balance.

I'm mostly joking really, although it did go with me commenting on my perception that Yukari is less powerful than a lot of the other touhous. (Which seems absurd when you consider that she's traditionally in the absolute top tier of overpowered in canon.)

 

[bouncymischa]

The randomized mushrooms seem like the best idea to me, because I like the idea of not just having one yield from them, after all, Marisa wouldn't use just one kind of shroom.

Another thought would be to have the shrooms be like the Blackfoot civ's Buffalo and have bonuses come from improving the shrooms.

Either way, Marisa's shrooms end up mostly being stolen from other custom civs and I think that's amazing.

To be honest, pretty much all of my code ends up being stolen from somewhere -- I'm rather like Marisa, in that regard. X3

Got the randomized mushrooms working. Now you can potentially get food mushrooms, production mushrooms, or science mushrooms.

I do quite like the idea of the "missile carrier" witch, although I think it perhaps could get more use with Mima, assuming the "buildable" spells had more variety with her.

The cannon idea seems good as well, after all, the Master Spark (well, Final Spark, really) is literally called a Magicannon. It could definitely work very well as a defensive thing given what you're apparently planning for the forest concept.


Another thought regarding cheaper tech: I have no idea if it's possible, but it'd be cool if she got the bonus from neighbors or opponents in a war as well. (Maybe even a small chance to instantly acquire a tech known by your foe that you don't know?)

Thinking about it, I remembered I'd considered doing the "missile carrier" idea with Mima and her Sorceresses... but I'm not sure it'd work with Marisa, because I'm not entirely sure how to spawn a unit being carried. With Mima, I considered having the units spawn at the capital, and then you could rebase them onto the Sorceresses... but it seemed weird. So I'll likely hold off on that for now. The Witch will likely be strong enough as-is, as it's rather similar to a very early rocket artillery... X3

I've been considering having the tech-bonus come from enemies rather than friends, particularly people you've fought in wars... but it'd be harder to code, so it's something I may have to build up to. As it stands, I have a preliminary version of Marisa up and running, although it feels rather basic. I'll have to sit down and play through it to see how it goes...

I'm mostly joking really, although it did go with me commenting on my perception that Yukari is less powerful than a lot of the other touhous. (Which seems absurd when you consider that she's traditionally in the absolute top tier of overpowered in canon.)

Heheh... I figured as much. It just reminded me of lengthy arguments I've had about game balance before.

As for Yukari, you probably do have a point... Unfortunately, aranicar hasn't posted here lately, so it'd be difficult to discuss changes that could be made. Perhaps I'll have to see if I can get ahold of him on Steam...

 

[nvdr]

I posit a joke Overpowered "Canon" Yukari who has a ton of ridiculous bonuses even if the real one gets a buff.

(Also it's just personal opinion but I think that whole three pack needs better city names, even if they can't ever be as good as Utsuho's list.)


The thing about "rebasing" spells is fine, I think you can just write it off as a limitation of the engine of a game that was never really designed for the concept being implemented.

 

[bouncymischa]

I've got Yumemi almost finished up... at this point, I just need to do some tweaking with her trait (I'm not entirely certain the war effect is working properly, so might change how I do it a bit), and decide if I'm happy with her custom promotion icons. Otherwise, I think everything is pretty much ready to go.

I've uploaded her current version here if anyone wants to test it out.

I think she's looking pretty cool, though!

EDIT: Just remembered, some other things I still need to fix... need to solve the problem of the v2.0 and v3.0 Research Assistants appearing in the tech tree, for one. (I'm certain there was another, but it's slipped my mind again...)

EDIT2: Do still need to do my alpha atlas, I think. @__@

 

[nvdr]

I'd like to report that not only did the Lab models not render once built, but other improvements didn't either, and then my game crashed after meeting Cirno. That said it's probably my stupidly large mod list causing a conflict in both cases, but I figured I'd mention it all the same. I'll give it another go with a smaller mod list soon.

EDIT: Locked up at turn 130 this time. No idea where to find a log to try to provide more of an idea why it happened.

EDIT 2: So why aren't Ruukoto and Mimichan on the spy list? (Or alternately adding them as hidden cameo units. Mimichan would make a great Nuclear Missile replacer.)

EDIT 3: Manged to finish the game thanks to autosaves, and I'm thinking I need to play on something higher than prince in order to really judge things anymore. I snowball way too hard. (Maybe I should try making a second city sometime as well.)

 

[bouncymischa]

Hmm... so did the lab graphics ever work for you? I know I had some problems earlier when DarkScythe was testing them, as I hadn't uploaded all of the textures properly, but they should've been working now.

As for Ruukoto and Mimi-chan... I completely forgot about them! I'll have to add them in.

How did the game snowball? I'll confess, I tend to play on Prince myself, but since my playstyle probably isn't anywhere near optimal (I tend to expand fairly fast), I rarely snowball.

EDIT: I think I found the source of the error for the labs -- I accidentally defined the standard lab graphic as BaseLab2.fxsxml instead of BaseLab.fxsxml the way it's supposed to be. I don't think that'd cause problems for other improvements (and the other lab models appear to render correctly), so not sure if that was the source of your crashes...

Do you have logging activated? The Database log may give useful information about errors, although the Lua log can reveal errors in Lua scripts.

EDIT2: Fixed lab error. Added Ruukoto and Mimi-chan to the spy list. Fixed an error where Yumemi's diplomacy image wasn't appearing. Updated the "science from wars" ability to be more explicit about how much science you were getting, and give a more direct bonus than the roundabout method I was using before.

Gotta go run some errands, but I'll zip up the revised version and post it this afternoon.

 

[nvdr]

Hmm... so did the lab graphics ever work for you? I know I had some problems earlier when DarkScythe was testing them, as I hadn't uploaded all of the textures properly, but they should've been working now.

EDIT: I think I found the source of the error for the labs -- I accidentally defined the standard lab graphic as BaseLab2.fxsxml instead of BaseLab.fxsxml the way it's supposed to be. I don't think that'd cause problems for other improvements (and the other lab models appear to render correctly), so not sure if that was the source of your crashes...

The upgraded versions did render properly, but I think you've pretty obviously sorted that out. As for the other improvements not rendering, it was one instance (the game crashed before I got to test that further) of one that I could only build thanks to a mod. (The tipi, if that's at all useful.)

How did the game snowball? I'll confess, I tend to play on Prince myself, but since my playstyle probably isn't anywhere near optimal (I tend to expand fairly fast), I rarely snowball.

It's just a thing I seem to do a lot, I've been able to do it as Alice and I don't really think there's any part of Yumemi that made it any easier to do. Although I was in the Modern era when Nitori and Kanako, who are notably science focused, were still in the Renaissance. To me, all this means is that it's time to move up to King, not that Yumemi is especially unbalanced.

Regarding my playstyle, I just play kinda passively and defensively with a single city and just focus on growing as much as I can. A single size 50+ city can do ridiculous things. (I think Probability Space Hypervessel was in the high 70s when I won by science.)

Do you have logging activated? The Database log may give useful information about errors, although the Lua log can reveal errors in Lua scripts.

I really should look into how to do that if I'm gonna continue helping test things, especially when it's now causing crashes. (Something Shinki, Mamizou and Yuuka have not done, I'll note.)

EDIT2: Fixed lab error. Added Ruukoto and Mimi-chan to the spy list. Fixed an error where Yumemi's diplomacy image wasn't appearing. Updated the "science from wars" ability to be more explicit about how much science you were getting, and give a more direct bonus than the roundabout method I was using before.

Gotta go run some errands, but I'll zip up the revised version and post it this afternoon.

Mimi-chan being on the spy list is amazing. I can only imagine how effective an ICBM is at espionage.

 

[bouncymischa]

It's just a thing I seem to do a lot, I've been able to do it as Alice and I don't really think there's any part of Yumemi that made it any easier to do. Although I was in the Modern era when Nitori and Kanako, who are notably science focused, were still in the Renaissance. To me, all this means is that it's time to move up to King, not that Yumemi is especially unbalanced.

Regarding my playstyle, I just play kinda passively and defensively with a single city and just focus on growing as much as I can. A single size 50+ city can do ridiculous things. (I think Probability Space Hypervessel was in the high 70s when I won by science.)

That could be a big part of it -- I've always tended to favor more sprawling empires. My brother and I would go extremely wide in Civ 3 games, and it's one of the reasons I never could really get into Civ 4 -- in my experience, it tended to be pretty brutal on expansion. Civ 5 isn't quite so bad, although I'm finding I still have to temper my expansionist urges a bit if I want to be effective at climbing the tech or social trees.


Anyways, finally uploaded an updated version of Yumemi here. Should have all of the changes I mentioned above.

Vicevirtuoso has pointed me in the direction of cleaning up the Research Assistant line in the tech tree, and I'm tweaking the "science from wars" effect a bit so it scales with game size and drops decimal values. Once I get all that done and tested, it should be ready to go, although I still haven't tested playing against Yumemi to know how effective she might be.

EDIT: Thanks to Vicevirtuoso's help, the v2.0 and v3.0 Research Assistants no longer show up in the tech tree. I noticed I still need to fix their "Historical Info" text, though...

 

[nvdr]

I have discovered a small issue: Laboratories I build after researching Education don't seem to change into the upgraded versions, even after saving and reloading. Pre-Education labs change just fine.

 

[bouncymischa]

I have discovered a small issue: Laboratories I build after researching Education don't seem to change into the upgraded versions, even after saving and reloading. Pre-Education labs change just fine.

It will depend partly on where you build them -- Labs built next to Hubs won't upgrade. If you build a Lab off by itself, though, it should still upgrade.

 

[DeletedSettler]

Once I get all that done and tested, it should be ready to go, although I still haven't tested playing against Yumemi to know how effective she might be..

I ran a Prince, Oval map game as Vicevirtuoso's Liquid Ocelot and some various Gensokyo residents (he'd fit right in!). Yumemi seemed like she was holding her own, wasn't dead last at least, although she was getting picked on a bit and lost one of her cities. Might just be since she was neighbors with Suika, though. I'm not entirely sure if she used the UI or not. She wasn't a tech leader, surprisingly, even though there were lots of wars to boost her science. Goddamn Shinki was the tech leader, of course.

Also, she denounced and spoke to me about my warmongering, which was a surprise since I figured she wouldn't have any warmonger hate at all.

 

[bouncymischa]

I reeeeeeally need to start playing some games against her, I suppose. And against the new Shinki, for that matter... >__<

I have a suspicion that Yumemi actually requires a bit of finesse to use properly, particularly to get use out of her various science boosts. I had the Research Assistants set to UNITAI_EXPLORE, which might've confused the AI about how to use them, so I'll try adjusting them to UNITAI_ATTACK, instead -- hopefully this'll cause Yumemi to use them more in combat and to fight barbarians. I'll also adjust her unique promotions to make them a bit more effective.

I'm debating whether or not I should increase the science yield on the Laboratories...

Still also considering toning down Shinki, as I think her production bonuses make her quite a wondermonger... Honestly, with no science or growth bonuses at all, I'm not sure how she becomes such a runaway... X3

 

[nvdr]

I think her production bonuses make her quite a wondermonger... Honestly, with no science or growth bonuses at all, I'm not sure how she becomes such a runaway... X3

Answered your own question, really. I'm pretty sure that done right, you can get the Great Library, Temple of Artemis and Hanging Gardens, which are all early game and provide exactly those science and growth bonuses.

Owning a lot of wonders can beat out virtually any bonus anyone else has and is a prime element of my normal one city runaway strategy. On a difficulty like Prince, the sheer weight of the bonuses you get from wonderspamming can make you steamroll any AI easily. And the AI, for all its faults, knows how to wonderspam and Shinki can do that job very well.

It's basically the wonder version of the GTI thing I did with Eirin. Just spam stuff that gives bonuses until you win. Although I'll be fair, Shinki doesn't have a way to make every land tile near her city give at least 2 food and 6 other yield along with wonder bonuses.


Also, I had another Yumemi game crash today. Once again I have no idea of the cause, but I do have a save a few turns before it, so I'll look into how to log it to see if it's from her or if it's just other mods not playing nice. (The crash happens a few turns into a war between Yuuka and Mamizou, so it could be related to some other mod.)

I'll also see about running a game with Yumemi as an AI and see how she does.

 

[bouncymischa]

Answered your own question, really. I'm pretty sure that done right, you can get the Great Library, Temple of Artemis and Hanging Gardens, which are all early game and provide exactly those science and growth bonuses.

Owning a lot of wonders can beat out virtually any bonus anyone else has and is a prime element of my normal one city runaway strategy. On a difficulty like Prince, the sheer weight of the bonuses you get from wonderspamming can make you steamroll any AI easily. And the AI, for all its faults, knows how to wonderspam and Shinki can do that job very well.

It's basically the wonder version of the GTI thing I did with Eirin. Just spam stuff that gives bonuses until you win. Although I'll be fair, Shinki doesn't have a way to make every land tile near her city give at least 2 food and 6 other yield along with wonder bonuses.

So I suppose I shouldn't feel too bad?

You may have a point there, though. One of the things I noticed in my games against Shinki was that she tended to struggle to generate Tourism a bit, so I believe I increased her Wonder flavor in the hopes she'd build more Wonders, and hence hopefully generate more Tourism. (I can't remember if Wonders add inherent Tourism or not... although I'm starting to suspect not.) It also suits her nature as a "builder" -- she did make an entire dimension, after all...

Still, as you say, wondermongering can be an easy road to success. Playing a hotseat game with my brother over the weekend, Brazil is currently the runaway, grabbing up every wonder they can. Of course, they currently have nearly everyone declaring war against them, and only ever built something like 3 or 4 cities, so they're position at the top of the heap may not last long. But I can see how successfully wondermongering can catapult a civ into the lead. It's also amusing because I checked Brazil's flavors, and they're only set to 5 for FLAVOR_WONDER. I suspect Civ randomizes flavors a bit, so Brazil likely just ended up higher than normal, leading to their rampant wondermongering... but it is interesting. Since I suspect I set Shinki to a higher wonder flavor, it's not too surprising she's a wondermonger.

So I suppose the question is, what changes to make? I could tone down Shinki's wonder flavor a bit so she isn't quite so aggressive at grabbing wonders... but it does kind of suit her character. I still suspect the best options may be to streamline the UA (removing the somewhat superfluous city defense bonus) and the Crystal Forest yields -- I do find a certain elegance in simplicity, and having the Crystal Forest only generate culture and gold, rather than generating three yields, does have a certain appeal.

Also, I had another Yumemi game crash today. Once again I have no idea of the cause, but I do have a save a few turns before it, so I'll look into how to log it to see if it's from her or if it's just other mods not playing nice. (The crash happens a few turns into a war between Yuuka and Mamizou, so it could be related to some other mod.)

I'll also see about running a game with Yumemi as an AI and see how she does.

Any more details about exactly what might be causing the crashes would be extremely helpful -- I've been playing Yumemi myself lately, and haven't had any issues with crashes, so I'm still wondering if there's conflicts between other mods causing them.

I'd also really like more information on how well Yumemi does as an AI. I'm currently considering a possible tweak in which the AI might receive a science bonus from labs early in the game (i.e. before the boost from Education). Two Science seems to be the point that the AI likes building labs, so I'm worried the AI won't build labs before Education, setting back Yumemi's progress. I also suspect the AI can't make full use of Research Assistants in farming barbarians, so an early science boost to compensate for that might be what they need...

 

[DeletedSettler]

I think there's a 2-both-ways randomness in AI's flavor, so 5 can be 3-7, along with just a general hint of randomness in their decision-making. At least that's my understanding for why Gandhi's a 12 for nukes: to ensure he'd always be at 10 in practice.

Popping the hood open to check around, Shinki's wonder competitiveness and wonder flavor is 7. She's a little more into wonders than Pedro is, who's 6 and 5. She's also 8 on culture emphasis compared to Pedro's 7.

I'll do another game or two and better pay attention to how Yumemi works out.

 

[nvdr]

So I turned on logging, and after the crash in the previously mentioned Yumemi game I looked at the log files and lua.log contains about 2 MB of variations on this:

[1757469.656] YumemiImprovementReplace: found improvement -1
[1757469.656] YumemiImprovementReplace: Citadel is 19

I presume the fact that it's an absurd amount of this, over and over, might have something to do with the crash?

If pastebin didn't have a limit of 500kb per paste I'd share the whole thing, but it feels somewhat redundant when it's basically just hundreds of lines of that.

EDIT: Turns out I had a Dropbox account I wasn't using, here's the full duplicate of the lua.log file.

 

[DarkScythe]

So I turned on logging, and after the crash in the previously mentioned Yumemi game I looked at the log files and lua.log contains about 2 MB of variations on this:



I presume the fact that it's an absurd amount of this, over and over, might have something to do with the crash?

If pastebin didn't have a limit of 500kb per paste I'd share the whole thing, but it feels somewhat redundant when it's basically just hundreds of lines of that.

This would not cause a crash.

My mod outputs thousands of lines of debug text (will be disabled once it actually goes live, but well, I'm still in dev..) I've had logs grow to around 30MB or so after a bit of play time, and my game didn't crash. Turn times slowed down a bit, though.

You'll need to upload the whole log for us to check what's going on.