NPC Factions RANT

[MaitreBouh]

I like the npc factions a lot. The war vs peace is balanced because the war gives you rewards and because you have about the same work to deal with the npc faction : at peace, you have quests and demands ; at war you have attacking armies. Either way you need an army at home to deal with them.

I like it because they are the chaos of the world you must deal with. That they require you attention while other matters already keep you busy is their whole point and their most interesting feature.

They actually are the maintenance cost of your empire.

 

[The Founder]

At the moment, I think the AI just gets too many free resources too quickly. Too much cosmite means that they spam out way too many colonizers too quickly. Too much influence means the covert ops and NPC faction purchase spam. Too much energy in general means that the player can't really get ahead at all in military, so they're bottom of the rankings and that means that the diplomatic AI is not interested in dealing with you diplomatically because you're a 'easy' target.

Game AI still has the same problem it had 1970 in Pong:
It can not plan ahead.*
I do not mean it is bad at planning ahead. It can not plan ahead at all. It has the planning ability of a falling slice if pizza.

Those resource are the "AI inability to plan" offset.
Having it ramp up does not sound like a good idea. This is a 4x game. Early game bonsues/results snowballing is the name of the game. If anything, the bonus should decay.


*Memory and the resulting planing (in)ability is one big issue of AI design. And one giant roadblock to AGI. We barely had some experimental breaktrhough with GPT2. If it is even feasible to port it (resource cost on a consumer computer), it will take decades for that to come to gaming.

 

[Hunnuli]

True I guess if it were limited maybe it should be per player so AI's with cheat resources can't just buy them all. Though I dunno do they receive cheated influence? But yes I can also see it being annoying to micro units across the map. I was kind of thinking maybe the map generator could be tweaked to make sure all players have at least some dwellings nearby their start position like they currently are guaranteed to get a settlement. But just a delay on the unit buying would be better than what we have now.

Re: Influence. AI, on hardest difficulty, gets 20 flat income of influence and 2x influence multiplier. https://forum.paradoxplaza.com/foru...ets-at-each-difficulty.1232671/#post-25741569 assuming that table is right anyways. But yeah that is a lot of influence the AI gets. Not really any competing going to happen there.

At the moment, I think the AI just gets too many free resources too quickly. Too much cosmite means that they spam out way too many colonizers too quickly. Too much influence means the covert ops and NPC faction purchase spam. Too much energy in general means that the player can't really get ahead at all in military, so they're bottom of the rankings and that means that the diplomatic AI is not interested in dealing with you diplomatically because you're a 'easy' target.

The energy I haven't actually seen too much issue of when I play. It is always the Cosmite and Influence that seems to have the most influence on things. Such as the Dwelling unit spam and Energy Siphon spam. The rest just feels like par for course given that AI's need bonuses to keep up. The problem with the two resources mentioned is how much of the balance for those resources rely on them being limited. The AI bonus just throws the balance of those resources out. Which is probably intentional since it seems like the AI can't really manage these type of limited resources. Does lead to the aforementioned NPC Unit spam and Energy Siphon spam though. The per turn limit on buying NPC units would limit the one as well as keep the player from pulling shenanigans.

 

[Vind7]

I think player must got some bonuses from destroyed heavily modded enemy units. Fighting 3-module equipped stacks can be tedious with little reward. Maybe some cosmite or additional research. Right now destroying 3-4 2.2k stacks yilds same reward as killing 10 low-tier units as rewards is not based on units firepower. Killing 20 units will get you one reward 20/10 and thats it.

 

[Elgareth]

Having it ramp up does not sound like a good idea. This is a 4x game. Early game bonsues/results snowballing is the name of the game. If anything, the bonus should decay.

Huh? Having it ramp up means no/very low Bonus at the start of the game, increasing as turns go on. Decaying means starting at the max. Bonus, then lessening it as turns go on. So a decaying Bonus would lead to larger early game bonuses than a ramping up Bonus.

And I think ramping up is correct here. At the early game the Player is very limited in what he can do. You only have one army, and only one colony. There are no big decisions to make, just creep and get settlers to go near cosmite. It's in the later stages of the game where a coherent Research plan, well placed colonies and synergizing army compositions give Players the edge over the AI (along with move Patterns that make sense... ), so there the AI would Need help.

 

[The Founder]

Huh? Having it ramp up means no/very low Bonus at the start of the game, increasing as turns go on. Decaying means starting at the max. Bonus, then lessening it as turns go on. So a decaying Bonus would lead to larger early game bonuses than a ramping up Bonus.

And I think ramping up is correct here.

Early game bonus is what leads to mid-lategame power. And the later the game, the less "Artifical stupidity offset" is nessesary. The income already aquired will be enough.

At least intuitively, I would asume decaying would be the better approach.

 

[Elgareth]

Early game bonus is what leads to mid-lategame power. And the later the game, the less "Artifical stupidity offset" is nessesary. The income already aquired will be enough.

At least intuitively, I would asume decaying would be the better approach.

I'd say the exact opposite.
Let's assume the ressource Bonus goes from 0 to 100.
Now a ramping Bonus would maybe give the AI 0 for the first 10 turns, then 10 for the next 10 turns, then 20 and so on.
So it'd be as limited as the Player in the early game. They can't buy faction Units early, spam colonizers or anything.
The longer the game goes, the Player will naturally be more effective than the AI (the reason the AI Needs those extra ressources in the first place).
As the Player will thus snowball harder than the AI once the more effective early and midgame come to fruition, the AI gets a bigger and bigger Bonus to aid its snowballing.

A decaying Bonus would give the AI 100 for the first 10 turns, then 90 etc.
So it could spam colonizers, faction Units, armies etc., leading to the current Problem that it's almost impossible to get any Goals as pioneer, as the AI doesn't have to wait for cosmite or influence ever, and can regularly insta-buy production.
This would technically lead to a harder snowball in the midgame, one the Player can never outspeed because he simply is limited by his ressources, and the only reason the AI doesn't stomp the Player into the ground by sheer number and power of modded Units is its stupidity to not attack most of the time.

So a decaying Bonus would give the Player a massive disadvantage in the early-game, Balance it out in the midgame, and a massive Advantage in the lategame (unless the AI somehow managed to cripple the Player).
While a ramping up Bonus would ideally Offset the AI stupidity more and more, the more complicated the turns become and thus the AI will fall farther and farther behind without help, as Efficiency becomes more and more important.

 

[TheDarkMaster]

Ideally, a ramping bonus also gives a proper difficulty curve. The early game shouldn't be difficult, it's the late game that should be. However that isn't what happens, since the player is in a somewhat despriate situation from day 1, but the moment the get even slightly ahead you can pretty much coast to victory. If those bonuses disappeared over time, all you'd need to do is survive the first 50 turns, then you just win.

With a ramp, the AI gets progressively more and more powerful as the game goes on, both helping it keep up with the player and letting it put up a challenge in the end game. Stellaris has the option to either give the AI flat bonuses immediately or scale them up over the course of the game. It would be nice at least to have the option for this, even if the game's general difficulty isn't built and balanced around the AI ramping.

 

[fdjw88]

I think player must got some bonuses from destroyed heavily modded enemy units. Fighting 3-module equipped stacks can be tedious with little reward. Maybe some cosmite or additional research. Right now destroying 3-4 2.2k stacks yilds same reward as killing 10 low-tier units as rewards is not based on units firepower. Killing 20 units will get you one reward 20/10 and thats it.

this is a great point! one of the main reason i refuse to deal with the NPC quest is the fact that their rewards sucks from the start to finish. and this problem is particularly awful in the late game. the NPC wants you to kill these 2000+ fighting power monster armies, but the rewards are barely worth the effort. the NPC quest needs to give me a reason to divert my attentions from the wars i have with the AI and try to finish their quest. i also find out that the AI empires are not very interested in NPC quest either. i see these monster stacks with gold cups above them all over the map, sitting there for 10 turns plus, and the AIs never kill them. the rewards need to be scaled up, and i agree with this reply that reward player some cosmites could be a good start. but for me personally, my cosmite income in the mid to late game always outpace my energy income.

 

[fdjw88]

Ideally, a ramping bonus also gives a proper difficulty curve. The early game shouldn't be difficult, it's the late game that should be. However that isn't what happens, since the player is in a somewhat despriate situation from day 1, but the moment the get even slightly ahead you can pretty much coast to victory. If those bonuses disappeared over time, all you'd need to do is survive the first 50 turns, then you just win.

With a ramp, the AI gets progressively more and more powerful as the game goes on, both helping it keep up with the player and letting it put up a challenge in the end game. Stellaris has the option to either give the AI flat bonuses immediately or scale them up over the course of the game. It would be nice at least to have the option for this, even if the game's general difficulty isn't built and balanced around the AI ramping.

agreed, i think the game right now employs a decaying bonus system. because from playing hard difficulty, the early game is always the hardest part. i think once i get pass turn 40 or something, when i have a decent size army (the AI still has significantly more stacks than i do mind you), i feel like the game just get much easier. despite my small army size, my units are very well modded, and all units in my army have great synergy with one another.

even on hard difficulty, some unit composition that the AI uses just don't make sense. for example, whenever there is an Amazon AI, you always see a stack army of 6 with 4 Biomencers, 1 hero and another random unit. why would you make an army like that? sleep and scans are good abilities, but sleep has very limited range, and many tier I units can out range the Biomencers. i think this is something that the devs can work on.

 

[Elgareth]

agreed, i think the game right now employs a decaying bonus system. because from playing hard difficulty, the early game is always the hardest part. i think once i get pass turn 40 or something, when i have a decent size army (the AI still has significantly more stacks than i do mind you), i feel like the game just get much easier. despite my small army size, my units are very well modded, and all units in my army have great synergy with one another.

even on hard difficulty, some unit composition that the AI uses just don't make sense. for example, whenever there is an Amazon AI, you always see a stack army of 6 with 4 Biomencers, 1 hero and another random unit. why would you make an army like that? sleep and scans are good abilities, but sleep has very limited range, and many tier I units can out range the Biomencers. i think this is something that the devs can work on.

That, and AIs basically never use anything T3/4, apart from faction Units. I never saw an enemy AIs Tyrannodon...or artillery. Which makes City sieges a bit weird since the defending AI still Comes to you, losing all their siege Bonus like Towers and defensible positions...

 

[fdjw88]

That, and AIs basically never use anything T3/4, apart from faction Units. I never saw an enemy AIs Tyrannodon...or artillery. Which makes City sieges a bit weird since the defending AI still Comes to you, losing all their siege Bonus like Towers and defensible positions...

true! the AIs do use tier 3 units, but using tier 3 artillery is indeed none heard of. during one of my game players, i did see a vanguard AI using Laser tanks, but i have never seen an AI empire using Walkers. i have only seen NPC armies using Walkers. i do agree that if the AI has a few siege units, the battles will definitely become more challenging. based on my experience, i also see AI failing at making a good army composition, so we might see a full stack of 6 artillery. given that AI has higher income than player, but they are not good with tactics and strategies.

 

[Sifer2]

Just to chime in on the scaling thing. That's one of my favorite settings in Stellaris. It's surprising so few strategy games offer it. It's either give the AI massive cheat bonus from the start or not at all. Even though like mentioned the players superior intelligence doesn't do much in the very early game when he has few armies. It's how lot's of good decisions early on can lead to a snowballed strong economy in mid to late game that usually wins it for the player. So giving the AI an increasing scaling bonus as the game goes on seems like the ideal way to make the game stay interesting. Especially since you could justify giving the AI even more insane resource boosts as the game goes on without making the early game impossible.

 

[The Founder]

this is a great point! one of the main reason i refuse to deal with the NPC quest is the fact that their rewards sucks from the start to finish. and this problem is particularly awful in the late game. the NPC wants you to kill these 2000+ fighting power monster armies, but the rewards are barely worth the effort.

All things already scale to game Progression (Number of turns used):
The NPC quests
The Armies for NPC quests
The Landmark and Pickup guards

If ther is too little reward, this is due to the quest itself giving to little. Some extra reward based on enemy mods seems wrong.

 

[daniel060]

All things already scale to game Progression (Number of turns used):
The NPC quests
The Armies for NPC quests
The Landmark and Pickup guards

If ther is too little reward, this is due to the quest itself giving to little. Some extra reward based on enemy mods seems wrong.

The number of turns that's gone by has very little bearing on how hard the quest is to achieve--you could have experienced a hard early game or an easy one, and how is the computer to know? The numerical strength of the army is far more relevant. I want rewards to feel proportional to the effort expended in achieving them, and defeating a powerful army should yield an awesome reward. Defeating an army on turn 50 as opposed to turn 30 doesn't take an additional effort except insofar as the turn 50 army might be stronger. So even then, army strength is the critical measure, not what turn it is.

 

[daniel060]

Just came back to the game after a long hiatus to see if things were any better, and man, this is STILL BROKEN!!! The NPC factions' demands and ridiculous war mechanics are so not fun that it makes this game almost not worth playing.

 

[DragonZee]

Just came back to the game after a long hiatus to see if things were any better, and man, this is STILL BROKEN!!! The NPC factions' demands and ridiculous war mechanics are so not fun that it makes this game almost not worth playing.

Mmm, did you know that you can control NPC faction behaviour after Tyrannosaurus update?

"NPC Dwelling Mode - This can be used to turn down (or up!) how often dwellings make demands and offer quests. It also controls how often they launch invasions when you are at war with them. At its lowest setting, it will also remove the Rivalry system, so doing a quest for one faction will no longer reduce your reputation with other factions."

I tested this and their demands were almost non-existant...

 

[Fluksen]

I tend to agree that the faction quests can be a bit annoying, but you can always click "decline" and/or turn down their behavious in the settings as @DragonZee pointed out. Still this system could certainly be improved, but have any of you gone for a diplomatic victory yet? you need several dwellings integrated for that. If you tune down the faction quests it'll become nearly impossible to reach that state quickly enough to be viable (and you still need cities near that actually can integrate them).

 

[DragonZee]

but have any of you gone for a diplomatic victory yet? you need several dwellings integrated for that. If you tune down the faction quests it'll become nearly impossible to reach that state quickly enough to be viable (and you still need cities near that actually can integrate them).

Judging by the achievement I have, I did it Can't exactly remember how I did it though, but I think I had at least one NPC faction integrated

 

[TheDarkMaster]

I tend to agree that the faction quests can be a bit annoying, but you can always click "decline" and/or turn down their behavious in the settings as @DragonZee pointed out. Still this system could certainly be improved, but have any of you gone for a diplomatic victory yet? you need several dwellings integrated for that. If you tune down the faction quests it'll become nearly impossible to reach that state quickly enough to be viable (and you still need cities near that actually can integrate them).

That isn't actually the hard part about that victory type. Once you have some experience with the game and consistently get up to 4 cities by turn 20, you should have no trouble keeping up with the quests and making good friends with the NPCs. Especially now that happiness events can give you extra influence.

The hard part is getting up to a virtuous reputation, due to the player AI's tendency to force you into wars or just attack you because they either think you're an easy target or too powerful. It's almost impossible to become virtuous without several pacts with several different AIs, but none of them ever play nice with each other. So you form one defensive pact or alliance and that AI calls you into wars against everyone else.