New Hotfix Version (12-15)

[Workerspam]

Huh. Used IGE to check out what happened in one of my runaway games with ilteroi's experimental 12/15 AI build and on Turn 144 not a single AI civ has built a monument in any of their cities.

 

[Ziad]

I'm currently playing Deity/Marathon as Inca and despite having a somewhat of an amazing start (lots of mountains, hills, and mining resources), I'm barely ahead of the AI as a whole in Renaissance when usually a start like this would be an auto win. Indonesia, Poland, and China giving me a run for my money.

I like this build so far and I hope the AI keeps up. I'll keep playing this to see how it goes.

 

[Ziad]

Huh. Used IGE to check out what happened in one of my runaway games with ilteroi's experimental 12/15 AI build and on Turn 144 not a single AI civ has built a monument in any of their cities.

I also noticed that some civs took a ridiculously long time to get social policies rolling. Some civs also founded a pantheon like 200 turns in on Marathon (which is just weird)

Dunno if relevant to this thread, but every single Tradition civ is behind by 4 policies of all the Progress/Authority civs. If they aren't building monuments, then that might be one reason why

 

[Txurce]

Using ilteroi's hotfix and playing Immortal, going for an SV, I have lost my last two games at the very end to runaway civs that won by... culture. I hadn't seen a CV by the AI in months.

In my most recent game, I actually had as many policies and Great Works until late, which is interesting. In that game, I stayed even or ahead in tech all the way, though. In the prior game, in which the runaway was off to the races early on, I could never catch up -- just stay even in spurts. My experience tells me that the late-game mechanics enhance the leaders more because everything enhances you if you're huge. It's not so much about getting good tech first, since that's also how you catch up. The issue is the advantages of size. If you have the equivalent of those advantages early due to good breaks, an AI at higher levels is probably going to pull away. (And it's much more likely to happen in my games, because I play with continents separated by deep water.) I'm also starting to come around to the belief that this is okay. It's what Civ has usually been like, historically. The anomaly is being able to win with a tall civ, or even a good-sized wide empire, if an AI dominates half the globe. Because if an AI civ is that dominant, it should win... and if it loses, it's because the human gamed it.

 

[Infixo]

as a general rule, building the cheapest buildings first is one that would most often mimick my usual game behavior (not that im an authority on building stuff, but i feel like i am!)

I think my point can be best summarized as. "The Monument should outweigh later buildings if the AI is playing well"

Still. It sounds like it should be weighted the other way.

I've tweaked some parameters of that process and initial results are very promising. Puppets build Monuments, Walls, Barracks, Libraries, etc. all missing early infrastructure. Well-developed cities do more or less the same as before, but older buildings appear more often, also cheap support buildings like Garden, Forge, Ligthouse are chosen more often.
If somebody wants to do a little testing, I attach a modified DLL.
You also need to add the following lines to any SQL file loaded along VP

-- ai building scoring
UPDATE Defines
SET Value = 0.03 -- org. 0.015
WHERE Name = 'AI_PRODUCTION_WEIGHT_MOD_PER_TURN_LEFT' AND EXISTS (SELECT * FROM COMMUNITY WHERE Type='COMMUNITY_CORE_BALANCE' AND Value= 1 );
UPDATE Defines
SET Value = 0.3 -- org. 0.15
WHERE Name = 'AI_PRODUCTION_WEIGHT_BASE_MOD' AND EXISTS (SELECT * FROM COMMUNITY WHERE Type='COMMUNITY_CORE_BALANCE' AND Value= 1 );
-- xp scaling on/off
INSERT INTO CustomModOptions (Name, Value, Class)
VALUES ('BALANCE_CORE_SCALING_XP', 1, 6);

 

[Gazebo]

I've tweaked some parameters of that process and initial results are very promising. Puppets build Monuments, Walls, Barracks, Libraries, etc. all missing early infrastructure. Well-developed cities do more or less the same as before, but older buildings appear more often, also cheap support buildings like Garden, Forge, Ligthouse are chosen more often.
If somebody wants to do a little testing, I attach a modified DLL.
You also need to add the following lines to any SQL file loaded along VP

-- ai building scoring
UPDATE Defines
SET Value = 0.03 -- org. 0.015
WHERE Name = 'AI_PRODUCTION_WEIGHT_MOD_PER_TURN_LEFT' AND EXISTS (SELECT * FROM COMMUNITY WHERE Type='COMMUNITY_CORE_BALANCE' AND Value= 1 );
UPDATE Defines
SET Value = 0.3 -- org. 0.15
WHERE Name = 'AI_PRODUCTION_WEIGHT_BASE_MOD' AND EXISTS (SELECT * FROM COMMUNITY WHERE Type='COMMUNITY_CORE_BALANCE' AND Value= 1 );
-- xp scaling on/off
INSERT INTO CustomModOptions (Name, Value, Class)
VALUES ('BALANCE_CORE_SCALING_XP', 1, 6);

I've been working on bugs and optimizing this yesterday and today, tread lightly in the DLL so you don't step on toes.

Edit: also be careful with those defines move them too much and the AI will ignore wonders.
G

 

[ElliotS]

and the AI will ignore wonders.

*eyes light up and picks egypt*

 

[Infixo]

I've been working on bugs and optimizing this yesterday and today, tread lightly in the DLL so you don't step on toes.
Edit: also be careful with those defines move them too much and the AI will ignore wonders.
G

I’ve ran 130 turns from the start, all Wonders were built timely. But i suppose that once they start to really be costly, taking 15+ turns to build, AI might choose something else to build from time to time.

 

[Gazebo]

I’ve ran 130 turns from the start, all Wonders were built timely. But i suppose that once they start to really be costly, taking 15+ turns to build, AI might choose something else to build from time to time.

Yep - backing them off to .2 is probably more fair. Any more than that and 10+ turn items start to get sloughed off.


G

 

[Bromar1]

Yep - backing them off to .2 is probably more fair. Any more than that and 10+ turn items start to get sloughed off.


G

Any ETA on when the next patch will be? I wanna start a new game but it's hard with a new version so tantalizingly close!

 

[Infixo]

Yep - backing them off to .2 is probably more fair. Any more than that and 10+ turn items start to get sloughed off.
G

Yeah, that is the idea. Those 10+ buildings were scored more than those with 1-3 turns. 0.2 is too little, 0.5 is too much, i did some calculations. Probably 0.25 should be ok. It needs that old-era boost anyway. Actually, that era boost is more important, since there is a bug, that results in „no-differentiation” effect.
It is also possible that those defines could stay untouched. Because buildings cost more with eras passing, then those two mechanisms provide similar effects.

 

[Infixo]

Do we have any idea on handicap values a, b, c?
My observations are that a should be less than 1, like 0.7-0.8 so AI will not get crazy in late eras, and b more than 1.0, like 1.5, so effects will start to show mid game. C is only imporant up to medival, so current 4 is ok, could also be 5.

 

[Gothic_Empire]

Odd to me but I can see the argument for it. But would you really build walls before monument in your capital? That is just a bad play.

The AI is actually struggling with culture in my games, I'm often hitting policies 5 and 6 before they do. I bet if they build monument second instead of a warrior, they might even be able to knock me down a difficulty.

It could be confirmation bias on my part. And actually, now that I'm thinking about it, I keep getting influential status with civs when I'm not doing much to build my tourism.

With respect, I find that a really weird build order. What units are you building that you'd want them before Monuments? A second-build Worker has basically nothing useful to do, a second-build Warrior might be tangentially useful but is not more so than a Monument. I genuinely don't think that's a play-style difference, I think that's poor play, at least as far as I can see it.

Pathfinder | Pathfinder or situationally, Pathfinder | Archer/Spear. Early exploration is worth a lot, and it's often difficult to quantify (and pretty gambly), and 1-2 mobile medic units are VERY valuable in early warfare. Naturally, though, I'll switch over to a monument if I can't explore for some reason.

I win regularly on emperor. Since the AI has apparently not been building its culture buildings, we'll see if that holds true.

 

[Infixo]

Since the AI has apparently not been building its culture buildings, we'll see if that holds true.

It is true. If Monument is not built quickly enough, it gets stuck in the queue and AI „forgets” about it for some time.
In that quick test game I mentioned above, Monument was usually built as one of first 4-5 things, in many cities as 1st or 2nd. China surprised me the most. Lady Wu built 4 monuments in 4 cities, as first choice. At turn 130 she had 8 policies (!) and 19 techs. On King diff, btw. The next civ had 5 policies only.

 

[Asterix Rage]

@ Infixo

-- xp scaling on/off INSERT INTO CustomModOptions (Name, Value, Class) VALUES ('BALANCE_CORE_SCALING_XP', 1, 6);

I'm not sure to understand corrrectly.
'BALANCE_CORE_SCALING_XP', 0, 6 means "xp scales with game speed" will be disable?
Relevant with VP(1-3) ?

 

[Infixo]

@ Infixo
I'm not sure to understand corrrectly.
'BALANCE_CORE_SCALING_XP', 0, 6 means "xp scales with game speed" will be disable?
Relevant with VP(1-3) ?

Yes. 0 - disabled. 1 - enabled. In 1-3 default should be 1.

 

[Gazebo]

Yes. 0 - disabled. 1 - enabled. In 1-3 default should be 1.

Default is set in CBO.

G

 

[CrabHelmet]

Pathfinder | Pathfinder or situationally, Pathfinder | Archer/Spear. Early exploration is worth a lot, and it's often difficult to quantify (and pretty gambly), and 1-2 mobile medic units are VERY valuable in early warfare. Naturally, though, I'll switch over to a monument if I can't explore for some reason.

I win regularly on emperor. Since the AI has apparently not been building its culture buildings, we'll see if that holds true.

Ah, I forgot how useful Pathfinders were on lower difficulties. The AI has so many extra units on Deity that if you get 2 ruins, that was a reasonable performance, so I don't invest in a second Pathfinder.

 

[MorphBer]

Next game i will play on standart speed. Or try to play for Shaka. Cause on epic speed with new exp system - it’s pain to play. Dont even think on playing maraphon....
Melee units are fine on epic, i still can easy get march promotion(swordsmans are still the best) or supply promo for naval. And then get one or two promotions during game - thats enough for me.
But ALL range units now are a piece of ***. Earning promotion for dromons/archery lines looks like grinding mobs in some MMORPG. I remember how in one mmorpg i killid 895 wolfes to get a lvl, 2 or 3 hours of pressing 5 buttons )
In my current game i feel the same with my cruisers... Maybe there is a some easy way to scale the number of hitpoints?

Can't share your observation, you need to get that oil monopoly

 

[Ziad]

I love the XP change on Marathon. My units actually take a while to level up so I have to really nurture them if I want them to become gods.

I play with the upgraded Generals mod so combat legit feels like I'm playing Rome Total War sometimes lol