Bad AI Diplomacy Is Holding This Game Back

Some Easy Improvements to Diplomacy

The good news is that this problem is very easily fixable.

In general I like this game. However, I would never have purchased it (not for 50 bucks) if I knew how bad the AI was going to be. If they made a couple easy fixes/refinements, I'd consider buying DLC. But right now diplomacy with AI factions, and between them, is so extraordinarily dumb. I'm not even bothered by the tactics or strategy AI. The diplomacy AI sucks. I've seen plenty of bad reviews on steam referring to this problem. And it can be vastly improved very easily inside the system already in place.

Diplomacy in GalCiv3 is very one dimensional, and not very dynamic. If I have the smallest military, then friendly personalities who share borders with me will declare war on me, and aggressive personalities will declare war on me even if they are on the other side of a huge map. And the war will never end, even if they lose all the ships they send for hundreds of turns. What they require in a trade deal to make peace seems to be dependent on the difference between their military power and mine, which means if I'm the smallest and they're the largest, even offering every planet, station, and ship I have is not enough sometimes. They won't ask for peace until I create a larger military than them. That seems to be their only consideration. This can be pretty irritating when other factions continue to declare war on me because I still have the smallest military, and why would they declare war on anyone other than the weakest faction?

I know there are other factors, but they don't seem to weigh very heavily, and there are plenty of factors that don't seem to be considered at all. I've played several games where I have the weakest military, and even several hundred turns in (I always play the slowest pace settings), with a dozen factions, the only wars in the galaxy are against me. Basically the resulting feel of the game is that if you have the smallest military, then everyone hates you and declares war on you (defacto uniting against you), and if you have the largest military, then everyone loves you (or at least hates you less and tolerates everything you do). Very dumb game design when the opposite should be true. The stronger your military becomes, the more factions should be uniting to fight you.

Strong factions picking on weak factions is realistic. This alone doesn't bother me. It's the cost of pursuing a non-military strategy. Although it feels silly when it seems factions rarely declare war on any faction other than the weakest one. The real problem is that GalCiv3 is missing a lot of other realistic diplomatic behavior that would balance out the game play.

This can be fixed with very little code if a couple more factors are considered.

Every faction should be vying to become the most powerful faction. That means the most powerful factions should view each other as rivals, and should likewise be more friendly and supportive of that faction's enemies. If I'm a weak faction in a war with the largest faction, why is no one providing me with aid? There should be other factions very interested in weakening the top dog. This used to happen in GalCiv2. Other factions would gift ships if you were fighting one of their rivals. 

Furthermore, if #2 faction wants to declare war on #1 faction, but can't because their military isn't strong enough, then they should consider the military strength of any other faction fighting against the #1 faction. When the #1 faction is in a war with any other faction, that should be seen as an opportunity for the #2 and #3 factions to deal a blow to the top dog.

When I'm a weak player, at war with the strongest, and all the enemy's ships are in my territory (instead of defending their own), I always wonder why no other faction is taking advantage of that opportunity. The answer is simple. AI factions aren't really trying to win. They grow like plants, following dumb scripts, and occasionally bump into each other or choke each other out. They are not even remotely opportunistic, ambitious, or forward thinking. They ignore the reality that if another faction reaches victory conditions before they do, then they lose. If they were trying to win, they'd take appropriate action to prevent other factions from reaching victory conditions first. You could spend a lot of time creating an intricate AI that weighs and analyzes each factions proximity to victory conditions against the costs of different interventions. But for now, it would be good enough just to consider each factions 'Faction Power' ranking.

All of the structure is already in place to make AI diplomacy much more intelligent very easily with the + - system. So here's a quick list of what could be added.

1) The most powerful factions should see each other as rivals. (add some -'s for each other)

2) The most powerful faction should have some extra animosity toward them from everyone.

3) Every faction should have more positive views of any faction that is fighting their 'rivals'. (common enemy is currently considered, but only applies if two factions are actually at war with the same faction. rivals aren't necessarily at war)

*The 3 changes above will do a lot to turn the game into a competitive struggle for the top without actually changing anything outside the + - system. The changes below are also easy to make, but involve calculating some other data.

4) When deciding whether to declare war on a particular faction, the AI should consider the military power rating of every faction already fighting that faction. (this will make the AI more opportunistic)

5) Weaker factions should attempt to form loose unofficial alliances (short of official military alliances) to challenge stronger players. For some reason, other factions only want my help fighting weak factions like minor races. I've never had another faction ask for help to bring down a larger faction.

6) Maybe less important, but maybe the AI factions should consider their losses compared to enemy losses in a given war when considering peace. I really feel like I'm playing against a potato when I destroy incoming fleets for 100 turns, never lose a planet, and the AI faction refuses to reconsider the benefits of the war they started because my military rating is still considerably lower than theirs.

7) For that matter, maybe a faction's kill/loss ratio, and the weight of those kills and losses, should figure into their military rating. This would result in AI factions re-evaluating your military strength if you were holding your own against a stronger military, which would make the AI seem more intelligent and also leave the player feeling that their achievement is respected, the same way people would look again at a skinny kid who just beat up a big dude. This could deter other factions from declaring war, and persuade factions to end wars that aren't going well. If the other changes were implemented, this would also cause AI factions to view you as a threat a little earlier on, creating a greater challenge for the player if they are disliked and other factions see the need to unite earlier. 

8) If you want a diplomacy system even more dynamic and lively that doesn't revolve as much around the most powerful factions, easily implement an actual 'rival' system. For each faction, the faction they view most negatively is their rival, and gets an extra '-' (if you don't want factions to switch rivals quite as often). Every faction who is at war with their rival, or who also views that faction as a rival, gets an extra '+'. Rivals will tend to be factions who share borders, which is fine, but something as simple as a powerful faction being reduced could result in shaking up and refactoring the geopolitical playing field.

 

Right now, in my experience, games tend to be either impossible, or too easy to enjoy, depending on the difficulty setting I choose, or what stage of the game I am in. Some games are very challenging until I persevere and become somewhat powerful (that part is cool), after which the game stops being challenging at all (the long boring part). I end up rising to the top, no one wants to mess with me or even could mess with me if they wanted to, and all AI factions sit back while I eliminate them one by one. Making some or all of these changes would result in a game that is challenging during more stages of the game and under more difficulty settings, while also being impossible under fewer circumstances and under fewer difficulty settings. The difficulty of each game will scale better with the player power rating, and it will feel more like playing against a vibrant set of living competitive opponents instead of a pile of vegetables.

11,088 views 8 replies
Reply #1 Top

"This can be fixed with very little code if a couple more factors are considered."

^^^^

I find this really humorous since you have no idea how hard or easy it may be to code what you want. 


The diplomacy is an ongoing work in progress and will always have room to improve. Try some of the mods created by players which make diplomacy and ai engagement more interesting while we wait for the expansion which will change the entire game from the ground up. 


Cheers!

Reply #2 Top

"This can be fixed with very little code if a couple more factors are considered."

^^^^

I find this really humorous since you have no idea how hard or easy it may be to code what you want. 

 

I'm not sure why you would assume that I have no idea how hard it would be to code this, or why you think that it wouldn't be easy to do so.

It's true that I haven't looked at the source code for GalCiv3. But I've been programming for a while. I may not know exactly how things are implemented, but a lot of what has been implemented is clearly visible to the player, and those are the things I recommended adjusting and adding to.

If the AI can check to see if you have ships in their territory illegally, and adjust their attitude toward you accordingly, then certainly they can check to see if you are the faction with the highest power ranking, and adjust their attitude toward you accordingly. If this takes more than a couple lines of code to add into the current architecture then that says some pretty bad things about the current architecture. But I'm willing to bet that the architecture is just fine, that the game developers are competent people, and that this feature could be added in a couple minutes if the devs are slow and careful about it.

Likewise, with the exception of #5, each change I recommended would require only the comparison of a few points of data. If they implemented #7, they could almost get away with skipping #6 because the slow increase in military rating would eventually make the enemy want to end the war anyway (but both 6 and 7 would still be better). #4 would only require subtracting the military ratings of common enemies from the faction in question (although if it were me I'd only subtract half the military ratings of common enemies). #5 is the only one that might be a little more work than the rest, depending on how they actually set up the code that has AI factions seeking help in war. But honestly, you could skip #5 if you implemented numbers 1-4 and simply weighted more heavily the animosity toward the top faction.

Anyways, I'm sure the devs are busy working on important stuff. I'm not trying to criticize them or the game (okay, yeah I'm criticizing the game), I'm just trying to persuade the devs to put a little love into the aspect of the game that I feel is lacking the most. And I'm offering possible solutions to persuade them that the large benefits are more than worth the small investment of time.

I'll check out the mods.

Is there an expansion coming that will change the game from the ground up? I couldn't tell if that was sarcastic.

Cheers

Reply #3 Top

There is an expansion that will fundamentally change the entire game and how everything is played right now. This expansion is going to be called Crusade and will removed the ability to control planets at the local level and will change how the empire works. It is a rewrite of how the game is now. Crusade is slated to be released Spring 2017. 

Reply #4 Top

Diplomacy feels like some creature that is still mostly skeleton waiting to be fleshed out. There is structure there on which to hang muscle and finesse, but they haven't really been implemented yet. Interesting things are promised and no one knows what it will finally look like, but the general mood in the monstrous delivery room is hopeful. I try not to pace while waiting, myself.

Reply #5 Top

You are much better off sticking with: Diplomacy is one dimensional. It is simple and to the point.

Because it is, and yes that is a flaw in the game. There seems to be one AI behind it all, because of how it makes decisions.

The drengin should pick on weaker races, exactly as they do now. A more orderly peaceful race might instead be happier when you don't have an overwhelming military, and fear you getting one. With good diplomacy it might ask you to reduce your military, with amazing diplomacy it might ask to form an independent peacekeeping force. That's aiming for the stars though, were just shooting for fun. :D

While the word variety has been sought by several for some time. There are also people that say things like, no AI should declare war on you if it can't win. So you have a tug of war between those who want purely logical decisions, and those who want differing personalities. At the moment the logical decision crowd have the ball in their court, unfortunately too far in their court at the expense of fun, variety or personality.

For me I'd love a chaotic race who I can't predict at all, an enigma, whose personality is only revealed to you as you play, and changes each game. I'd also like a race that will fight you even if they see it as a hopeless cause, heck i'd design an underdog race based off the bajorans, the maquis, or the rebel alliance that get bonuses for being the underdog, and have them look to excel in that role, taking offense if they feel challenged. You could have a slow to grow but really powerful race, that does diplomacy very slowly in hibernation cycles, building tall, meaning that every offer has to be considered carefully else the malus or penalty will exist for a long time. Or a race that is so alien it has no diplomacy at all, it has to do its diplomacy through manipulating others. Or a race that can't start wars, they can only join allies, and are always looking to make peace. A race that twists planets, the longer they are on them, the less other races can use them, meaning everyone really, really hates them - All these types of restrictions at least would add more personality to the races, rather than just 

Military > Military. Goal of Diplomacy. Seems lackluster as to what could be there.

On the simple side:
Tech > Tech, could be more important to a science loving race, could show a kindred spirit who would rather ally with you, you get the idea.

Reply #6 Top

I think it should be a balance between the values.

 

Military + Tech + Resources

 

Where the weight of each could possibly scale.

For example Dregin would weight Military more and Tech/Resources less.

 

Then it's a simple comparision.

 

For your chaotic race you could just randomly determine the weights at check time. 

 

You could also add in modifiers.

So formula could look something like the following

 

If ((my.military +/- my.mil-modifier) +

     (my.tech +/- my.tech-modifier) + 

     (my.resources +/- my.res-modifier) > 

     (target.military +/- target.mil-modifier) +

     (target.tech +/- target.tech-modifier) + 

     (target.resources +/- target.res-modifier)) then attack

 

Resources could just be Mined resources or Mined resources + credits. as Mined resources I can create a lot of custom ships.

 

Another thought I would consider is total.military production. For example if a target has 3x my production at the time of the check I better have a quick solution to destroy them or in y rounds they could consume me.

 

Other factors to take into account is if target is at war with other civs, if it is then possibly divide each value (minimally resources/production) by the number of civs attacking. One time I was at war with 9 other civs, and I was clearly at more of a disadvantage than if I was only being attacked by a single civ.

 

Another factor that should be taken into account is the distance. I know in a single player game, it's basically all the AI against me, but to have an AI that is more than 1-2 factions away (ready more than 30 hexes) should not attempt to attack. First they can't easily reach me and by the time they build up a bridge of bases I simply just have to destroy the last bridge and again they need to send a "slow" constructor to build another bridge or go to war with neighbor faction just to reach me. This I think should be one of the first things addressed, as it seems to happen way too often.

 

The computation could easily get complex, especially if they improve diplomacy where my allies actually support me in times of war. Then enemies would need to take any/all allies into account as well. But what could be done is you create a function/method which computes an individual "war" value, then just pass in targets and allies return a list of values, then possibly just sum them all upto determine if I (or AI) have a chance to win.

 

Reply #7 Top

Well if someone asked me what i would prefer is a challenging game or faction differentiation. I would prefer a challenging game. They could program factions to play character. 

Altarians are good. Meaning they could aid underdog factions to help improve life. Terrans are usually resource driven or are working a combination of whoever has the most power if they are hostile towards. By answering with a coldwar embargo and a arms race. Drengin and snathi are conquerors. The yor wants to exterminate everyone who is notmechanicle. The torians are paranoid with ever aren't human. The arceans are trying to reestablish the arcean empire. The thalans are trying to turn everyone against humans. The iridium are nuetral. 

 

Reply #8 Top

Enemy declares war TOO far.

This is one of the issues which should be easy to check.

My long range scout ran into on of their scouts and shortly after he declared war on me.

Which for me is fine, because it will be a good 50+ rounds(at 4 movement per round ) before I even see one of his ships. And that is even if he can reach me. Any ship he sends to me is either going to be complete Life support or a LOT of bases.

And I am sure there is someone closer to him which could deserve his focus.

The AI should simply check the distance to the target and do some type of calculation based on their current life support range.