New Beta Version - February 18th (2-18)

[Gazebo]

I will say this - beyond the handicap bugfix, the AI was making some really bad decisions early game, esp. with workers. Fixing that was like shifting them from 1st to 2nd gear early game.


G

 

[Zuizgond]

[...]

Thanks a lot Recursive! Very clear

 

[nDook]

Been climbing in difficulty these past few weeks without much of a hitch between so I'm excited to see how the ai improves from these changes.

 

[Gidoza]

Bonus for Immortal difficulty, Ancient and Classical era:

Era = 1
Base = 7
A = 340
B = 240
C = 160

Yields = Base * ((C * Era * Era) + (B * Era) + A) / 100

Yields = 7 * ((340 * 1 * 1) + (240 * 1) + 160) / 100

Yields = 7 * (340 + 240 + 160) / 100

Yields = 7 * 740 / 100

Yields = 5180 / 100

Yields = 51.8 (rounded down to 51)

Yield Bonus: +51 Gold, Science, Culture, Golden Age Points

Bonus triggers:
- Creating a city
- Capturing a city
- Winning a war (warscore above 25)
- Constructing a World Wonder
- Birth of a Great Person
- Completing a trade route to a major civ
- Entering a new era

If bonus was not triggered by a Great Person's birth or a trade route completion, then all cities also gain +51 Food and Production.

If bonus was triggered by entering a new era, double the yields to 102.

As Era goes up (2 in Medieval, 3 in Renaissance, etc.), based on the AI's current era, the bonuses will increase exponentially.

I would drop the bonus from generating Great People and finishing Trade Routes, then I might be OK with this setup. For two reasons: 1. GP generation and trade routes are way too variable, so you'll get huge differences between different AI Civs. 2. These effects occur far too frequently compared to all the others. I consider this both unreasonable and unnecessary as a bonus against human players.

 

[crdvis16]

I would drop the bonus from generating Great People and finishing Trade Routes, then I might be OK with this setup. For two reasons: 1. GP generation and trade routes are way too variable, so you'll get huge differences between different AI Civs. 2. These effects occur far too frequently compared to all the others. I consider this both unreasonable and unnecessary as a bonus against human players.

I think you have to be careful that your list of events that lead to bonuses doesn't upset the balance of other things. For instance, removing great person bonuses but leaving settling/capturing cities would harm an AI that is going small/tall (like how the AI often plays Korea) relative to AI going wide.

 

[Stalker0]

I think you have to be careful that your list of events that lead to bonuses doesn't upset the balance of other things. For instance, removing great person bonuses but leaving settling/capturing cities would harm an AI that is going small/tall (like how the AI often plays Korea) relative to AI going wide.

I agree, I think the TR component is more of a target. This also means that AIs left alone get significantly more bonuses as their TRs get to finish, as opposed to AIs that get declared on and their TRs cut.

 

[Gidoza]

I think you have to be careful that your list of events that lead to bonuses doesn't upset the balance of other things. For instance, removing great person bonuses but leaving settling/capturing cities would harm an AI that is going small/tall (like how the AI often plays Korea) relative to AI going wide.

But that's pretty weird and variable too. Why not just drop ALL of these items and instead just offer something that's clean across all playstyles, like a smaller bonus for every technology research?

 

[Gazebo]

But that's pretty weird and variable too. Why not just drop ALL of these items and instead just offer something that's clean across all playstyles, like a smaller bonus for every technology research?

The current iteration generates more variability.

G

 

[CrazyG]

Is it intentional that you can buy horsemen without barracks, but you need a barracks to buy Mandekalu cavalry?

I would drop the bonus from generating Great People and finishing Trade Routes, then I might be OK with this setup. For two reasons: 1. GP generation and trade routes are way too variable, so you'll get huge differences between different AI Civs. 2. These effects occur far too frequently compared to all the others. I consider this both unreasonable and unnecessary as a bonus against human players.

I see the appeal of trade routes, it is something that scales slowly over the game. But it must make Petra or Colossus really good for the AI.

I'm thinking that several AI that are constantly a problem to deal with on Deity have trade route things going on, so they might complete them more often than others, for example I think that trade routes to foreign civs get pillaged a lot more often than internal or to city state, and Germany is frequently the leader of the pack on high difficulties.

 

[Gidoza]

The current iteration generates more variability.

G

The fact that each Civ is unique and takes entirely different Policy paths isn't already unique enough? The AIs need a boost that buffers them, not this odd thing that makes them even more unpredictably different from one another.

 

[Heinz_Guderian]

Quite amazing how often these patches come out. Will try it out!

Thanks Gazebo and the dev team

 

[Gazebo]

Is it intentional that you can buy horsemen without barracks, but you need a barracks to buy Mandekalu cavalry?


I see the appeal of trade routes, it is something that scales slowly over the game. But it must make Petra or Colossus really good for the AI.

I'm thinking that several AI that are constantly a problem to deal with on Deity have trade route things going on, so they might complete them more often than others, for example I think that trade routes to foreign civs get pillaged a lot more often than internal or to city state, and Germany is frequently the leader of the pack on high difficulties.

Mande used to be knight, is just a leftover error. I'll fix.

 

[Crag Hack]

The civilopedia still claims that yields are compared to empire yields rather than to global yields.

The civilopedia needs a major overhaul and many corrections. And also need to include the equations and the mathematics behind it. One equation is worth 1000 words.

 

[Moi Magnus]

The fact that each Civ is unique and takes entirely different Policy paths isn't already unique enough? The AIs need a boost that buffers them, not this odd thing that makes them even more unpredictably different from one another.

I disagree. When you apply buff, you have to be careful to not unbalance the different strategies. For example, if you give big bonuses when a city is conquered, and no other bonus, the only AIs that will benefit from those bonuses are the aggressive ones, which will just destroy the variability of the game, making the the top AI of each game necessarily warmongers.

The different policy path are balanced around an unbuffed player, so as much as possible, every buff should affect all the possible policy path equaly.
So we're searching from bonuses that affect everyone equaly (trade routes), together with as many bonuses that affect wide AIs (founding cities and conquering them) than bonuses that affect tall AIs (GP).

The only way out of those "odd things" would be to go away from current design of "rewarding AIs that do good more than AIs that do poorly", and just give raw yields per turn to every AI, increasing with time (not number of techs, because that would favour some AIs, increase with raw number of turns instead). But that doesn't look really appealing to me.

The civilopedia still claims that yields are compared to empire yields rather than to global yields.

The civilopedia needs a major overhaul and many corrections. And also need to include the equations and the mathematics behind it. One equation is worth 1000 words.

Inaccuracies in the civilopedia count as bugs, and can be reported as such on Github (https://github.com/LoneGazebo/Community-Patch-DLL/issues).

I'm not sure how well equations can be put in the civilopedia. And as for an overhaul, you're not the first to try to launch such a project, but nobody is motivated enough to spend his personnal time going through it and checking every info to correct them. That's the point of community projects, nobody is assigned mandatory tasks. But feel free to contribute to help us on that.

 

[LifeOfBrian]

Could someone please help me and tell me which line(s) of code should I change in my local files to stop the AI from getting the bonuses from settling their capital? Thank you!

 

[crdvis16]

But that's pretty weird and variable too. Why not just drop ALL of these items and instead just offer something that's clean across all playstyles, like a smaller bonus for every technology research?

I'd probably favor something that's clean across playstyles and, if possible, doesn't favor frontrunner AI. A bonus every X turns perhaps?

I imagine making that change could lead to serious reverberations, though. The effect on Venice if trade routes no longer give their bonus, for instance. I suppose those effects might be known by comparing AI performance at settler difficulty vs deity maybe?

 

[Stalker0]

But it must make Petra or Colossus really good for the AI.

Honestly the bonuses from an extra TR I don't think matter nearly as much as the timing of TR pillages. The right war timing could pillage 5-6 TRS right before they complete....which is an incredible amount of bonus for the AI. I also agree with you it makes ITRs a lot stronger for the AI in general due to its safety.

 

[Crag Hack]

Honestly the bonuses from an extra TR I don't think matter nearly as much as the timing of TR pillages. The right war timing could pillage 5-6 TRS right before they complete....which is an incredible amount of bonus for the AI. I also agree with you it makes ITRs a lot stronger for the AI in general due to its safety.

It has been discussed before that it is nice idea to be able to stop TRs and re-rout them, but turns out is programming nightmare to make the AI capable of doing the same. I always believed (without proof) that the AI somehow cheats and re-routs their TRs anyway. I have seen TRs just pass in a certain spot and I setup an ambush there for them. But they never appear again.

 

[Stalker0]

Balance aside, here is why I dislike the AI bonuses on Turn 1 (aka when they city settle). I feel like its the opposite problem of the Archer Rush.

To me, human vs AI is a competition of a human's superior tactics/optimizations vs the AI's bonuses. When the AI's bonus are ramping up in the mid game, it is up to me to have developed a superior strategy, warfare tactics, and build order optimization in order to counter. A "fair" fight.

The archer rush was so abusive because it allowed the human to flex its tactical superiority before the AI bonuses kicked in...an unfair fight.

Likewise, the AI getting bonuses on Turn 1 allows them to start a snowball effect before I have had a chance to utilize my optimizations. It gives them incredible advantages in forward settling, religion, and early wonder production...and that early in the game there isn't that much I can do to counter. Again, an "unfair fight".

Its similar to how I feel about the AI getting free promotions (generally fair, a compensation vs my superior warfare tactics) vs AI pathfinders getting promotions on turn 1 (unfair, ability to scout the map and take ruins before I could possibly counter).

 

[SupremoBean]

Hey, love the mod and appreciate all the work that's gone into it.

Has anyone else been getting constant crashes with the new patch? Around turn 60 or so in every game I've played on this version (5 or 6 times so far), the game will close for no apparent reason. It's not hanging and stopping responding so much as alt-f4ing itself suddenly. I've tried different settings and such and disabled all mods besides the VP ones but it keeps happening. Nothing like this has happened before for me and I've played a couple of games on each of the most recent betas, including the previous one.

Not sure if this is worth submitting a bug report over or not but I'd be happy to do one if so.