Poll:Is the AI capable of defending itself properly/Should Defensive Pacts be more common?

[Txurce]

For players that remain competitive, this would only make things more difficult(that's the idea, anyways). I play with 22 civs and most of them become irrelevant and up for grabs as the game goes on, due to failed wars and runaways grabbing up everything important. An increase in DPs would hopefully improve the balance of powers, thus making it more difficult for a runaway of any sort to exploit everything.

I agree with most of what you said, but it does seem like runaways would be a bigger problem on 22-civ games than those with a standard number. I don't think Gazebo balances for huge maps, 22 civs, etc... but I could be wrong.

If someone plays poorly enough that prince is the hardest AI they can compete with, are we sure that their units are more valuable?

Additionally on lower difficulties you can easily keep up with the AI making units, because the way difficulties work means the AI should keep pace.

Their units are about as valuable for them on Prince as yours are for you on Deity.

You and I can keep up easily on lower difficulties, but I've seen plenty of posts from people who can't create a strong all-around empire at those levels. Or who prefer playing peacefully, but then have to deal with constant wars from aggressive AI civs. I'll be these latter folks would love to have more DPs, too.

The point of making your units seem bigger isn't some sort of 'handicap' as you imply. It's suggested as a way to make the AI play better. It's not a question of if your units are better than the AI's on immortal and deity. It's a fact that you'll fight better than them. So having the AI think you're weak and launch a suicidal war, or not approach with a defensive pact, because your numbers are too low in the eyes of the AI is a real problem.

It's only on Deity (and to a tiny extent Immortal) that the player needs to be exploitative to win. You can't go for a peaceful science victory on Deity. You can't really do a peaceful ANYTHING on deity. The AI gets such massive bonuses that you need to screw with them to win.

This leads to the problem. The play CAN screw with the AI on deity, because their units fight so much better than the AI's due to lots of experience and practice. That's not reflected in their military scores at all.

There's no way a brand new player could compete on Deity. The only people who play it have practiced for a decent amount of time on this an similar games. That should be reflected in military scores, because it's relevant for the AI's decision making.

You're saying that although the only way the human can win on Deity is militarily, and even then only by employing his superior skill, the lack of DPs is "a real problem" for the AI. That is hard for me to imagine. Are you saying you're beating the AI on Deity too easily?

Regardless, here is how I see the effect of more DPs on Deity:

1) you have the DP. The AI is less likely to attack you. But you're playing for domination, so this gives you the advantage of increasing your defense vs civ A while being on offense vs civ B.

2) the AI has the DP. This does nothing to slow them down from attacking you due to a misread of your strength, which is what you've largely been arguing. (They would be safer from you if you attack, though. So if you're wiping the floor with the AI on Deity, then I could understand why you would want to give them -- not you -- more DPs.)

 

[CrazyG]

That's a problem for the AI on any level. For example:

I can hold off the AI militarily with a bare-bones military on Emperor, and go on to win some sort of victory. The AI attacks me fruitlessly because it thinks I'm weak. I'm sure the same miscalculation can be made all the way down the line to Prince. So should we encourage more human/AI DPs for all levels? I think that would just make the game easier for the human.

On Diety its really different. They will attack with just warriors when you have a spearmen UU and multiple already built. His units don't even attack anything because its suicide to hit my city and suicide to hit my units. This leads to a ridiculous and immersion breaking dance off between my pictish warriors and his warriors. And even if I slaughter those warriors his score is still higher. Its not a bare bone military having crap score, its almost any military. I've had enough army to take a capital and still not been able to pledge to protect city states.

I view it as their needing to a balance between strategies, and right now on Diety certain strategies aren't options, not necessarily for lack of merit but because the calculations block you from trying. Tribute from city states is a great example, its unrealistic how difficult it is, for early game its almost exclusive to civs with spearmen UUs (Zulu with 6 spears could not earlier today). I think EliotS is suggesting doing something specific to Diety (and immortal to an extent) to adjust calculations, and I agree.

It's only on Deity (and to a tiny extent Immortal) that the player needs to be exploitative to win. You can't go for a peaceful science victory on Deity. You can't really do a peaceful ANYTHING on deity. The AI gets such massive bonuses that you need to screw with them to win.

I have won a 0 war game on Deity before as Korea. In general I agree with you, but this is an exaggeration.

 

[ElliotS]

I agree with most of what you said, but it does seem like runaways would be a bigger problem on 22-civ games than those with a standard number. I don't think Gazebo balances for huge maps, 22 civs, etc... but I could be wrong.



Their units are about as valuable for them on Prince as yours are for you on Deity.

You and I can keep up easily on lower difficulties, but I've seen plenty of posts from people who can't create a strong all-around empire at those levels. Or who prefer playing peacefully, but then have to deal with constant wars from aggressive AI civs. I'll be these latter folks would love to have more DPs, too.

The AI's skill with units doesn't improve as difficulty goes up AFAIK. The player's does.

Unless you're telling my that someone who can play on prince as their hardest level is as efficient in their unit usage as someone who plays on deity then saying "Their units are about as valuable for them on Prince as yours are for you on Deity." is patently untrue.

If my units are worth 2 or 3 of an AI's units, the prince player's units are worth 1 to 1.5 I'd guess. That's not "about as valuable" in any sense of the phrase.

You're saying that although the only way the human can win on Deity is militarily, and even then only by employing his superior skill, the lack of DPs is "a real problem" for the AI. That is hard for me to imagine. Are you saying you're beating the AI on Deity too easily?

Regardless, here is how I see the effect of more DPs on Deity:

1) you have the DP. The AI is less likely to attack you. But you're playing for domination, so this gives you the advantage of increasing your defense vs civ A while being on offense vs civ B.

2) the AI has the DP. This does nothing to slow them down from attacking you due to a misread of your strength, which is what you've largely been arguing. (They would be safer from you if you attack, though. So if you're wiping the floor with the AI on Deity, then I could understand why you would want to give them -- not you -- more DPs.)

No, that's not what I'm saying. I'm saying that while the mechanic works fairly well on lower difficulties, the AI will launch suicidal wars, underestimate my threat, or refuse to ally when it will help them at higher difficulties. That's a serious problem.

While I think that defensive pacts should be more common, and would like to see more alliances to overcome such pacts or runaway civs, that's another issue. The balance implications of making the player's units count for more on higher difficulties are as follows:

-AI is less likely to suicidally attack you. This should be seen as a positive imo, but occasionally a bunch of AIs will decide it's a good idea to attack me because I'm "so weak" and I'll actually lose.
-You can actually pledge protection when you're a military force to be reckoned with. I'll bet anything that a 1.5x multiplier wouldn't allow the player to pledge to protect without significant effort regardless, and if it does then 1.5x is too much.
-You can demand tribute more reliably on deity. This is actually really important imo.
-The AI should be more likely to offer friendship and defensive pacts, and less likely to declare suicidal wars.
-The AI should properly evaluate you as a military threat, and make plans to counter it.

Aside from the last one, many of these make the game easier. However they feel more like bug-fixes for a mechanic that doesn't work well in deity than anything.

I think we should treat defensive pact logic and coalition/alliance logic as a separate issue, that uses the player's military strength but isn't directly related. The AI's perception of the player's military strength is more far-reaching than that. So please comment on them separately.

On to the other half; Pacts and coalitions:

I suppose part of my problem with the AI is that while it likes to eat the weak, it never seems willing to team up to break the strong. That's been an important part of history, and is important in multiplayer gameplay.

So the idea of defensive pacts and coalitions/alliances is to bring the AI more in line with how human players tend to act. Two people don't plan on fighting each other, so they band together for mutual benefit. Meanwhile two warmonger civs might need to team up to overcome a powerful defensive alliance.

That sounds both more exciting and more realistic (historically and human-player wise) than the current behavior, so it seems in line with the goals of the project.

 

[Gazebo]

Lots of theory here, very little of it grounded in the AI's actual code. The AI absolutely understands runaway civs and the utility of DPs to counter warmongers, but let's be clear: the AI isn't out to give you a 'good time.' Being competitive is about more than just being a thorn in the human's side or suddenly being obtuse when someone is about to win the game. These kinds of design decisions are what currently cripple the Civ 6 agenda AI.

 

[ElliotS]

Lots of theory here, very little of it grounded in the AI's actual code. The AI absolutely understands runaway civs and the utility of DPs to counter warmongers, but let's be clear: the AI isn't out to give you a 'good time.' Being competitive is about more than just being a thorn in the human's side or suddenly being obtuse when someone is about to win the game. These kinds of design decisions are what currently cripple the Civ 6 agenda AI.

The only thing I'd ask to be different for the player vs. AI is the calculation for the player's military strength on deity. That causes some serious balance issues.

The thing with pacts and increased willingness to gang-up on powerful civs applies to both humans and AI. It's not that I want the AI to make my goals easier/harder, but rather that I thought the goal of the mod was to make the AI play more human. The almost complete lack of defensive pacts and ganging up on runaways is an extremely AI-like behavior.

Regardless of that for points both I and CrazyG have mentioned: (largely regarding tribute and pledges) the player's AI score needs an buff/adjustment on deity.

 

[Jdoug312]

I play with 22 civs and most of them become irrelevant and up for grabs as the game goes on, due to failed wars and runaways grabbing up everything important.

Have you been able to win domination victories with 22 civs in the game? I couldn't conquer fast enough to make it work; penalties from resistance, unhappiness (war weariness and under developed cities), and DPs/chain DoWs bogged me down each time I've tried.

 

[ashendashin]

Lots of theory here, very little of it grounded in the AI's actual code. The AI absolutely understands runaway civs and the utility of DPs to counter warmongers, but let's be clear: the AI isn't out to give you a 'good time.' Being competitive is about more than just being a thorn in the human's side or suddenly being obtuse when someone is about to win the game. These kinds of design decisions are what currently cripple the Civ 6 agenda AI.

This isn't about counters or just being a nuisance. It's to increase balance. Right now, there are too many underdogs that consistently get eaten up and rarely launch a war that doesn't cripple them. Better relations and cooperation between the AI is akin to what a player would do, and for good reason. Most civs just can't take full advantage of their unique abilities because they end up becoming irrelevant for the same reasons over and over. Before balancing the civs themselves, there should be more work on their decisions, no? Or maybe diplomacy is horribly more complex than we think and you've already tested similar things.

Have you been able to win domination victories with 22 civs in the game? I couldn't conquer fast enough to make it work; penalties from resistance, unhappiness (war weariness and under developed cities), and DPs/chain DoWs bogged me down each time I've tried.

I've came close a fair few times. Things bog down a lot once I'm managing 30+ cities, so I start looking at other victories. I think it's one of the easier ones, but far too tedious with 22 civs, especially when the other victories are so much easier with half the world conquered. The idea is to already have a strong stance before ideologies become a problem. Having tanks to quickly storm everything helps too.

 

[Gazebo]

Or maybe diplomacy is horribly more complex than we think and you've already tested similar things.
.

It is. There's a lot more at play here than just 'make the AI more competitive.' Changing values in one area can and will make it more likely to break diplo elsewhere.

 

[ashendashin]

It is. There's a lot more at play here than just 'make the AI more competitive.' Changing values in one area can and will make it more likely to break diplo elsewhere.

Oh well. This really will remain as a difficult to answer balance after all. I still love VP as is anyways.

 

[CrazyG]

The only thing I'd ask to be different for the player vs. AI is the calculation for the player's military strength on deity. That causes some serious balance issues.

The thing with pacts and increased willingness to gang-up on powerful civs applies to both humans and AI. It's not that I want the AI to make my goals easier/harder, but rather that I thought the goal of the mod was to make the AI play more human. The almost complete lack of defensive pacts and ganging up on runaways is an extremely AI-like behavior.

Regardless of that for points both I and CrazyG have mentioned: (largely regarding tribute and pledges) the player's AI score needs an buff/adjustment on deity.

Just something I'll point out, I've played a lot of FFA strategy games in my life and its pretty rare for humans to actually successfully gang up against a leader and swing it. Human behavior is normally to try to gang up, either fail to unite (and games of entirely good players have failed to coalition), try the war and surrender when it fails, or end up letting the #2 guy win

I can see unintended consequences for this as well. Recently I was going for science, but I'm influential in 5 civs, and I'm only exotic in Portugal (side note nice to Portugal finally doing well), am I suddenly going to have to war everybody? Same thing with science, what sets off the coalition? I can skip the apollo project, I can wait to attach spaceship parts, I can sit on my scientists and bulb them all the same turn. For diplomatic they do a pretty job of proposing resolutions to hurt me (sanction, decolonize, repeal world religion).

One possible spot for improvement here would be more logic for conquering of city states that are my allies or have my embassy. Does the AI have logic to steal an embassy via citadel?

 

[Txurce]

On Diety its really different. They will attack with just warriors when you have a spearmen UU and multiple already built. His units don't even attack anything because its suicide to hit my city and suicide to hit my units. This leads to a ridiculous and immersion breaking dance off between my pictish warriors and his warriors. And even if I slaughter those warriors his score is still higher. Its not a bare bone military having crap score, its almost any military. I've had enough army to take a capital and still not been able to pledge to protect city states.

I view it as their needing to a balance between strategies, and right now on Diety certain strategies aren't options, not necessarily for lack of merit but because the calculations block you from trying. Tribute from city states is a great example, its unrealistic how difficult it is, for early game its almost exclusive to civs with spearmen UUs (Zulu with 6 spears could not earlier today). I think EliotS is suggesting doing something specific to Diety (and immortal to an extent) to adjust calculations, and I agree.

Everything you mentioned falls into the bucket of why I don't play Deity. In short, the AI bonuses lead to a warped game (compared to just about all other levels) that preclude the strategies you've mentioned (including inability to get tribute), and ones you didn't mention (like having a decent shot at most wonders).

Players not as good as you feel the same way about the levels beyond their reach: the AI has too many advantages, and conversely, underestimates the human tactical mind. You can find a steady stream of posts on the CP site voicing their version of your complaint. Using the tribute issue as an example, you are flat-out asking for a handicap, so you can better enjoy the game. I would argue that this is just one of the handicaps of playing Deity, just like the huge Wonder deficit -- or playing against monster armies, with you the one civ that for some immersion-breaking reason can't get your people to reach equivalent production levels.

 

[ElliotS]

Everything you mentioned falls into the bucket of why I don't play Deity. In short, the AI bonuses lead to a warped game (compared to just about all other levels) that preclude the strategies you've mentioned (including inability to get tribute), and ones you didn't mention (like having a decent shot at most wonders).

Players not as good as you feel the same way about the levels beyond their reach: the AI has too many advantages, and conversely, underestimates the human tactical mind. You can find a steady stream of posts on the CP site voicing their version of your complaint. Using the tribute issue as an example, you are flat-out asking for a handicap, so you can better enjoy the game. I would argue that this is just one of the handicaps of playing Deity, just like the huge Wonder deficit -- or playing against monster armies, with you the one civ that for some immersion-breaking reason can't get your people to reach equivalent production levels.

I suppose there could be a whole thread dedicated to deity balance, and probably should be at some point.

The problem for me is that Immortal has flat out become too easy, and Deity has said warped game aspect feel. I'm fine with the AI being bigger and badder than Immortal, it's expected that they'll have more units, culture, tech faster etc.

What I can't stand is that certain mechanics that work in every other level don't work.

The largest aspect of this is in regards to demanding tribute. It really, really hurts authority players.

So I'd like to see some change in regards to that at the very least.

As for it being 'flat out asking for a handicap' that's only because it's impossible to demand early tribute without a UU spearman. It's one thing to find something difficult because it's beyond your level, it's another to have entire mechanics become neigh unusable.

Anyone complaining about such on a lower difficulty is factually wrong, because I could go do that.

So if I'm wrong and someone can show me them consistently demanding tribute on deity like I can on immortal or below without a UU, I'll admit I'm wrong and retract my suggestions. I really don't think it's possible though.

And if it isn't possible then we can agree that there needs to be some change, because it's a core game mechanic integral to Authority. No?

 

[CrazyG]

Everything you mentioned falls into the bucket of why I don't play Deity. In short, the AI bonuses lead to a warped game (compared to just about all other levels) that preclude the strategies you've mentioned (including inability to get tribute), and ones you didn't mention (like having a decent shot at most wonders).

Players not as good as you feel the same way about the levels beyond their reach: the AI has too many advantages, and conversely, underestimates the human tactical mind. You can find a steady stream of posts on the CP site voicing their version of your complaint. Using the tribute issue as an example, you are flat-out asking for a handicap, so you can better enjoy the game. I would argue that this is just one of the handicaps of playing Deity, just like the huge Wonder deficit -- or playing against monster armies, with you the one civ that for some immersion-breaking reason can't get your people to reach equivalent production levels.

My first question is why do feel compelled to disagree with a change specific to Deity if you don't play on it? And if you don't play on it aren't you just speculating? BTW you have a decent shot at any wonder on Deity, provided a decent starting position, you just have to give up more to get them. Recently I built all of Stonehenge, Mausoleum, Great Library, Hanging Gardens, and Oracle on Deity, that isn't a wonder deficit.

I might have to give up a faith heavy pantheon to get Mausoleum of Hali, but I can still get it. In a situation where I build 1 monument and after that just spear men I couldn't get tribute (and I don't think dropping the monument would have let me either). There is a difference between having a larger trade off, and not having the option at all

 

[ishmael_]

In what kind of situations do you see defensive pacts? I don't remember seeing any AI suggesting a defensive pact to me nor seeing one in place between the AI factions (in vanilla or vox populi). Are they more common in deity difficulty or games with a larger number of civs? Or maybe I just miss them? I play immortal with the standard number of civs (8, right?)

I am in favor for more collaboration between civs, especially the ones that have become irrelevant and/or are threatened by the runaway civ. Sometimes I get attacked by more than one civ if I'm close to a cultural or scientific victory (yay!), but cooperation in defense seems more rare. But defensive pacts would be just an extra, a nice-to-have feature. More importantly I'd like if the AI could be improved to not attack me when they have no chance to achieve anything meaningful by attacking. Don't get me wrong, I do love it when I don't have to be the attacker. I can just take their cities in response to their actions: "I was just defending myself".

I get a feeling that often the weak civ attacks me after being bribed, if that observation helps. Or maybe they sometimes make a deal and the other party attacks me after 10 turns? (I also think this situation did improve in some patch not that long ago, so I'm not complaining.)

 

[Txurce]

I suppose there could be a whole thread dedicated to deity balance, and probably should be at some point.

The problem for me is that Immortal has flat out become too easy, and Deity has said warped game aspect feel.

What I can't stand is that certain mechanics that work in every other level don't work.

The largest aspect of this is in regards to demanding tribute. It really, really hurts authority players.

So I'd like to see some change in regards to that at the very least.

And if it isn't possible then we can agree that there needs to be some change, because it's a core game mechanic integral to Authority. No?

I could say that legitimate access to Wonders is a core mechanic intergral to Tradition, and you can't play Tradition on Deity either, right? But I get it. I think there should be threads devoted to all relevant difficulty levels, and that they would be a much better place to discuss what in your case at least are inter-related issues. Deity certainly looks screwed up to me.

My first question is why do feel compelled to disagree with a change specific to Deity if you don't play on it? And if you don't play on it aren't you just speculating? BTW you have a decent shot at any wonder on Deity, provided a decent starting position, you just have to give up more to get them. Recently I built all of Stonehenge, Mausoleum, Great Library, Hanging Gardens, and Oracle on Deity, that isn't a wonder deficit.

I might have to give up a faith heavy pantheon to get Mausoleum of Hali, but I can still get it. In a situation where I build 1 monument and after that just spear men I couldn't get tribute (and I don't think dropping the monument would have let me either). There is a difference between having a larger trade off, and not having the option at all

First question: because I don't like insuffciently integrated, inelegant changes.

Second question: yes, I am speculating. So are you. That's what we do here.

When you win a few non-dom games on Deity starting with Tradition and Progress, then I'll take not being able to raise tribute a little more seriously. I think Deity has much bigger fish to fry that.

 

[ElliotS]

In what kind of situations do you see defensive pacts? I don't remember seeing any AI suggesting a defensive pact to me nor seeing one in place between the AI factions (in vanilla or vox populi). Are they more common in deity difficulty or games with a larger number of civs? Or maybe I just miss them? I play immortal with the standard number of civs (8, right?)

I am in favor for more collaboration between civs, especially the ones that have become irrelevant and/or are threatened by the runaway civ. Sometimes I get attacked by more than one civ if I'm close to a cultural or scientific victory (yay!), but cooperation in defense seems more rare. But defensive pacts would be just an extra, a nice-to-have feature. More importantly I'd like if the AI could be improved to not attack me when they have no chance to achieve anything meaningful by attacking. Don't get me wrong, I do love it when I don't have to be the attacker. I can just take their cities in response to their actions: "I was just defending myself".

I get a feeling that often the weak civ attacks me after being bribed, if that observation helps. Or maybe they sometimes make a deal and the other party attacks me after 10 turns? (I also think this situation did improve in some patch not that long ago, so I'm not complaining.)

It's not just you. I know G keeps saying it's really complex, but I've as I said I've seen maybe 1 or 2 defensive pacts offered TO me in ~20 games at 7 and 8 in recent memory that went late game. Both times I was WAY ahead and the one steamrolling everyone, so I obviously didn't accept.

I've never once seen the AI form a defensive pact, even when it's the only chance they have of stopping my completely transparent domination victory. I don't care how complex the formula is if the result is the same as a one line formula to effectively disable all defensive pacts.

Maybe my experiences are an aberration, but I've seen a lot of people with similar feelings and I've never once heard anyone complain about too plentiful defensive pacts.

So has anyone else had a different experience, or is the physical result that defensive pacts are effectively non-existent for everyone else too. (Regardless of how complex the formula is.)

 

[Txurce]

I've never once seen the AI form a defensive pact, even when it's the only chance they have of stopping my completely transparent domination victory. I don't care how complex the formula is if the result is the same as a one line formula to effectively disable all defensive pacts.

I see the AI forming DPs in almost every game I play.

 

[ashendashin]

I also see DPs in my games. Just not very often or particularly useful.

 

[ElliotS]

I created a thread for Deity discussion, please move the discussion on that over there.

 

[CrazyG]

I could say that legitimate access to Wonders is a core mechanic intergral to Tradition, and you can't play Tradition on Deity either, right? But I get it. I think there should be threads devoted to all relevant difficulty levels, and that they would be a much better place to discuss what in your case at least are inter-related issues. Deity certainly looks screwed up to me.



First question: because I don't like insuffciently integrated, inelegant changes.

Second question: yes, I am speculating. So are you. That's what we do here.

When you win a few non-dom games on Deity starting with Tradition and Progress, then I'll take not being able to raise tribute a little more seriously. I think Deity has much bigger fish to fry that.

I've won more often with Tradition that Authority on Deity. If you look in the tradition China science victory thread, I summarized a Deity game I won after choosing tradition and building several wonders. There is a difference between playing several games on a difficulty and posting thoughts from that experience and speculation which leads to your very warped opinion of the eighth difficulty. We can continue in the Deity thread if you like