I am really bummed about the lack of relationship modifiers information

Ahriman said:

This argument drives me nuts.
In the real world, leaders most certainly have a detailed breakdown of their relationships with other powers. They know which issues are important, which of their actions are pissing other countries off, which of their actions the other guys like, what actions might improve relations, and most importantly, whether the other country is likely to go to war with you.
They have diplomats, foreign affairs departments, and intelligence services.
They read each others' newspapers and talk to the elites.
Its not just two guys talking, who don't know anything about each other's countries. That's not Great Power Diplomacy.

There have never been truly unexpected wars between sedentary nation-states. (There are some unexpected wars when a bunch of nomads or migrating tribes show up on your doorstep).

Click to expand...

You still get indications. The AI will likely say "don't settle near my borders" or "don't attack this city-state". You'll be able to do the same. That's far more like real life if we want to continue the tired realism argument.

My personal opinion is that, assuming that the AI has a random behaviour, you will learn how to read their opinion of you the more you play the game.

You'll know that if you have a cooperation pact with a leader and that leader help you during war, then he's friendly; you can also see which deals he's prepared to do with you to understand how your relationship is.

I personally welcome the change as I think it will make it a more challenging game.

Well we saw yesterday that Greg was able to tell exactly how much gold he needed to spend to get Monaco to "cautious" or "friendly" (I forget the exact terms), and what benefits it would give and for how long. So it seems that where it's important (relationships with city states), the information is available.

But I think the idea in Civ 5 is that diplo with the actual other civs is always at some level of war, whether it be cold war or actually declared war. Because the AIs are all trying to win this time round. And the Civ 4 thing of knowing the exact numerical "friendship" level so you can look it up on a chart of the things they'll do at each level is not on any more. You should expect to be attacked opportunistically and somewhat on the whim of the other leaders (i.e. with a random component), which is neither unrealistic, nor unplayable IMO.

In Civ 4 it's tremendously expensive to maintain standing armies for no reason, so the benefit of gaming the diplomacy system into a situation where you're entirely safe from DOW is huge. But it looks like Civ 5 is designed so that maintaining an army is completely natural and doesn't carry such an opportunity cost.

We should move away from Civ 4 thinking on this (like so much else!) Diplo will be important but no so much that you can base an entire game plan on it. All well and good, I says. The first time I leave a border city undefended and a previously-neutral AI says "thank you very much" and invades it, I will applaud them for good play!

How about this - in the real world, your cannot make strategic decision without data. So... it is (drum roll please)... MORE REALISTIC... to have the modifiers

So, so, very tired of reading the word "realistic" in threads about game design. "realism" is a childish game design principle (there I said it).

In the live video they spoke of how the AI takes into account the "deal history", which had its own screen, when making diplomatic decisions.

I'm thinking it's possible (i.e. I'm just guessing) that there aren't even such a thing as diplo modifiers anymore. That is, the separate AI leaders might have differing parameters associated with them (based on xml for example) but when they want to determine whether or not to accept a deal they run some algorithm on the "deal history". This means that if you ask them how they feel about another civ, they might generate that relationship rating on the fly. But then, if that could be done, why not just automate it and make that info shown to the player?
I'm going to send myself in circles here.

My point is, it's quite possible for the AIs to be designed in a way fairly different to civ4's and that modifiers as they are known in civ4 may simply not make sense in civ5. It's a stretch, I admit, but it's plausible if there's any substance to the hype about the newly designed AI.

I am looking forward to the diplomacy in Civ 5. I understand some of the concerns presented, but I think it will be interesting to play with AI who's relationships can't be explicitly defined.

It would be nice if it's like playing with another human player. They will work with you, when it benefits them, but they may backstab if/when the opportunity presents itself, and that opportunity will give them a way to win the game. I don't think it a good idea to say "oh, my relationship with Monty is +12, he will never attack me (LOL @ dumb AI), and therefore I can leave that border undefended for my war with India". It’s too transparent when you can just look a table and see the precise numerical value of relationships between all leaders with you and themselves.

In the gameplay demo yesterday, Greg indicated that new civs were not likely to join into a Pact of Cooperation when they first meet you. You can also get an idea on just how a leader feels about you from the tone of the deals they offer you. If they are demanding a one-sided deal, they are probably not your BFF. I would also guess that leaving your boarder with Monty undefended is unwise no matter how much he smiles.

It would be nice to know just how set up and interpret deals with the AI. Well he always prefers more of a good thing, or are diminishing returns coded into the deals? I think this type of information should be provided in the game manual and in the civpedia.

VitalyB said:

I loved in Civ4 the fact that I could easily see the relationship between me and other civilization and what caused it to be that way. That answered a lot of answers I had on "Has that thing I did angered/pleased them?" Moreover, I could see the relations between various civilization and decide upon them who I should ally/denounce.

Now it seems all that data is gone. Aside of active pacts and such, it is impossible to know who thinks what of you and what your relations are.

Is there any data or info that helps me out there with that?

Click to expand...

I think it's entirely realistic to understand what things they're mad about.

Maybe not actually quantify how these factors affect your relationship, but letting you know that the problems are there is realistic and, IMHO, really important to diplomacy

PieceOfMind said:

In the live video they spoke of how the AI takes into account the "deal history", which had its own screen, when making diplomatic decisions.

I'm thinking it's possible (i.e. I'm just guessing) that there aren't even such a thing as diplo modifiers anymore. That is, the separate AI leaders might have differing parameters associated with them (based on xml for example) but when they want to determine whether or not to accept a deal they run some algorithm on the "deal history". This means that if you ask them how they feel about another civ, they might generate that relationship rating on the fly. But then, if that could be done, why not just automate it and make that info shown to the player?
I'm going to send myself in circles here.

My point is, it's quite possible for the AIs to be designed in a way fairly different to civ4's and that modifiers as they are known in civ4 may simply not make sense in civ5. It's a stretch, I admit, but it's plausible if there's any substance to the hype about the newly designed AI.

Click to expand...

That's kind of what I've been arguing.

Suppose their decision whether or not to accept a deal is based on a trust value and a strategic need value. Trust would be "Deal history - untrustworthy things you've done". Strategic value could be a whole number of things. Suppose you want to trade iron, the strategic value could be how much he needs iron, but it could also be how much does he want to trade with you (maybe he wants to declare war on you and doesn't want to trade his horses).

Certainly, they shouldn't give you the strategic value numbers. That would completely handicap the AI. But suppose they gave you the trust numbers. If the AI still doesn't trade you, you kinda know that he's planning something against you. Also, if the AI has a high trust number, maybe you'll try and invade him, since the numbers have told you he doesn't see it coming.

Clearly numbers physically have to exist in some form (the AI is not making its moves completely randomly each turn), but I do think, in overall context of how it plays, it doesn't make sense to give you these numbers. This isn't Civ4 where the idea was to build relationships with some AIs and make enemies with others. Relationship building is with City-States. Interaction is the key with other AIs. You can ally with them to help both of you or to hurt a third party, but it's not a list of actions done specifically to improve relations, therefore, it makes no sense to list a series of actions to say why the AI is working with you.

migkillertwo said:

I think it's entirely realistic to understand what things they're mad about.

Maybe not actually quantify how these factors affect your relationship, but letting you know that the problems are there is realistic and, IMHO, really important to diplomacy

Click to expand...

I would guess that if the AI is angry with you, you will know why. But what if they do something unexpectedly hostile when they have no reason to be "mad"? It could be because they pragmatically view you as a threat, or easy pickings at a point when they want to expand, or various other reasons. As Louis XXIV more or less wrote, there is a logic - fuzzy or clear - underlying it... but it might not be as obvious as it was before. Which I think is a good thing!

I can understand why people are frustrated by the lack of modifiers, but unlike that V8 (or whatever it is) commercial, people don't walk around with a fluctuating number above their heads. The removal of the modifiers, I think, is intended to add to the game's realism.

You loose the advantage over the AI of being able to look into it's head which was kind of unfair! You'll just have to watch their behaviour, get to know their "flavours" and guess!

I massively prefer the black box AI. Makes it more like playing real opponents which is sort-of the point.

I'm more concerned about the related problem of having very little way to interact with an AI in many situations.

Especially anyone you meet later in the game and on another continent, it seems like the system is going to lead to little interaction and even less positive interaction. You won't have any history with a civ on another continent. When they first meet at that point in the game it's probably even more likely to have reluctance/neither side cares about tech pacts or secrecy agreements against other civs across the sea and so on.

Resource trading is certainly much more limited than in the past, and tech trading is gone, and other major diplo interactions an AI would care about are gone, so I think we'll have the problem of utter insensitivity towards distant players. At best you could go out of your way to argue over an overseas city-state or we get the two AI another continent each asking for secrecy pacts against the other one but that hardly counts for much.

I was under the impression the AI would be a little more forthcoming with suggestions about what it likes/doesn't like. The "not settle cities" is all right but not very specific - I understand why it might not have been in the preview but I would have though some more specific things - like that same order about a specific area, or discussions of resources/military, might occasionally pop up. But then again I haven't set my expectations too high about everything here, if it's simple and not broken that's at least enough.

but not very specific

Click to expand...

This is precisely the problem.

I get a message from Alexander saying "don't settle near me".

Ok. What does that mean? Which tiles can't I settle in?
For how long?
If he builds a new city near me, and I settle near that, 100 turns after the original message, does that still piss him off?

Obviously, you'd have more interaction with Civs you've been established with longer, that's just common sense (reflective of both good gameplay and realism). But new civs will probably ask you for help against a rival on the same continent. Then you can help him and make a new friend (along with new enemy from the other guy).

Remember the Escapist Magazine review. He visited the new continent and everybody asked for pacts of secrecy with you. You can play them off each other or choose sides if you're secure at home, but want to gain influence somewhere else.

The only situation where an A.I. shouldn't attack you or attempt to beat you is if you'll obviouslly kick their ass.

An A.I. should never sit back and let you reach a point of beating them because you were nice to them. Ever. They're supposed to be our competition. Maybe some A.I. can have a "Well... he was so nice to me this game, I think I'll let him win" personality trait... but it shouldn't be a coded part of the game.

The reason why the modifiers aren't shown to us is because in civ5 they're irrelevant, period. The diplomatic process is intuitive. You give someone some iron, you give them money, you join them in war, you do anything for them then naturally, they might like you.

However it comes with the exact same risk of playing with any opponent. Trying to befriend someone is the equivalent of gifting some of your own resources in the hopes you won't be an active target. It is, and always should be, a gamble.

Further, coding the A.I. so they actually perform as opponents or players is the only way to ensure they aren't taken advantage of, because in the previous civ4 design, I, who always achieves a conquest victory, will make nice with X amount of civs and without logic or reason stab them in the back when my superiority is guaranteed. Wait... there is a logical reason; I want to win the game.

Under a system where modifiers govern the A.I.'s behavior, ala civ4; where certain A.I.'s won't ever declare war on you if they reach a certain status... is completely abusable. It prevents the A.I. from making the same decisions we can make: Being friends with someone for such a long time and then betraying them with the only rationale being that they've become a threat to our win, or the next on the chopping block to achieve our win.

As such, there's two things you can do; Have the modifiers control the A.I. and show them to us.... with which the former is quite simply stupid. Or; Have the modifiers not control the A.I. and show them to us... with which the latter is quite simply irrelevant ~ Because we'd never hear the end of "OMG I had +15 with Napoleon but the SoB attacked me for no reason".

Either the modifiers are relevant (poor design, imo) and thus we need to see them... Or they're not relevant at all (great design, as it keeps the A.I. competitive and less of a waiting puppet for the player to control) and thus seeing them serves zero purpose to the player.

I'm going to go out on a limb and say the real reason we don't have the A.I. modifiers on display is because the A.I. aren't governed by them. Thus, the information is moot.

Ahriman said:

This is precisely the problem.

I get a message from Alexander saying "don't settle near me".

Ok. What does that mean? Which tiles can't I settle in?
For how long?
If he builds a new city near me, and I settle near that, 100 turns after the original message, does that still piss him off?

Click to expand...

Again, I think it's fairly intuitive. Don't settle near him. Ever. The more you do, the closer you get, the more likely you will become more of a threat in his eyes. Why do you need a turn timer? Why do you need a spreadsheet? Imagine you're playing against... Me, for example. If You build near me, you risk becoming my opponent, like any other player. If I say "Hey, stay away from me" Then obviously I don't want you near me. If an A.I. says this, then you settle in a different direction, or you settle in their direction, becoming more and more prepared for a confrontation the closer you get to them. It's pretty simple. "Stay away from me" means stay away from them. The closer you get, be it 20 tiles away or 10... the more likely they're going to consider you a threat and/or want to attack you.

This time the AI players will not be so easily manipulated. Before you could easily appease them before you took them out. Now they will be less easy to bluff, nor will you be able to keep them off you, with the exact minimum resources.

I thought everyone was against the dumbing down of the game? Surprise! It just got more complex.

Yea, this is the exact opposite of dumbing down. Diplomacy in Civ 4 is basically a math equation that can be solved. The modifiers are transparent, leaders always will react the same way (within the defined random rolls). You can have 100% certainty that Ghandi won't attack when you get his modifiers to "pleased", and that will be true every game. Now, AIs will be completely different in positive ways:
1. You won't be able to "know" how the AIs will react for certain. Between relationships not being displayed and the AIs "flavors" shifting between games, you'll have to feel stuff out, not solve it. You don't "know" whether napoleon is going to be an absolutely unrepentant warmonger, or merely a more aggressive cultural player. You'll have to judge these things, which is a skill that can really only be developed through play.
2: The AI won't be at such a huge diplomatic disadvantage when compared to the player: The AI in Civ 4 has no "relationship modifiers" for the human player, and really has no idea what the human could do. Giving players complete knowledge of the AIs behavior is not a good thing, it just leads to exploits (Hence how many immortal+ games are more about juggling modifiers in diplomacy than "winning").
3: Most of the AIs in Civ 4 are essentially incapable of winning. Only some AIs even threaten victory, and most of those are incompetent at accomplishing them (Not playing defensively on their capitals/cultural cities). Having the AI be less predictable and always focused on victory makes them more interesting and challenging opponents. You don't want a "KILL THE HUMAN" style of play, but Civ 4 goes too far in the "fun AI" direction of "coddle the human" gameplay. There's a problem when the AI, at it's full potential, is incapable of doing basic actions that could seal it victory. Like a space race civ fortifying its capital properly. Or a cultural civ preparing for a defensive war over its culture cities. Or a domination civ being 'diploed' out of an easy conquest by someone giving them a worthless city.

King Jason said:

The only situation where an A.I. shouldn't attack you or attempt to beat you is if you'll obviouslly kick their ass.

An A.I. should never sit back and let you reach a point of beating them because you were nice to them.

Click to expand...

This is wrong from here on out, in that while it is a valid opinion, it is specifically not how any games in the civ series have been designed. I do hope players with this mindset can continue to enjoy (and mod!) the game though.

The AI won't be at such a huge diplomatic disadvantage when compared to the player: The AI in Civ 4 has no "relationship modifiers" for the human player, and really has no idea what the human could do. Giving players complete knowledge of the AIs behavior is not a good thing, it just leads to exploits (Hence how many immortal+ games are more about juggling modifiers in diplomacy than "winning").

Click to expand...

I'd also argue this is isn't very accurate, and an AI that behaves more randomly and without set modifiers can often be even more abusable. I have no clue if you were experienced with civ3 for instance - I know a lot of folks really aren't though, but an AI that didn't have obvious modifiers or anything was just as abusable because it was very very easy to exploit them against each other.

The only way to prevent this in a game without modifiers is to make the AI specifically behave differently towards the human (bad) or give massive penalties/bonuses that put the human at a disadvantage anyway (so results in different but not better gameplay than civ4 again, for instance).

Plus I don't think modifiers and personalities are really gone. The AI seem to have far too distinguishable personalities from what we've seen already to argue that's fully the case.

There is very very little evidence as well that the "cheesy" human wins won't still be common in civ5. Seems like a human could still easily beat many an AI that could have/should have won otherwise by culture or diplo still anwyay.

The numbers were a bit of a crutch for dealing with bad AI. It was totally gamey. Hopefully this just means the AI is better (and/or the game's design is better). Just use you intuition.

The only situation where an A.I. shouldn't attack you or attempt to beat you is if you'll obviouslly kick their ass.

Click to expand...

This is a reasonable opinion, I just disagree with it completely.
Thats not how real historic Great Powers act.
I want to play a game which makes me feel like I'm playing a version of history, not to be constantly reminded that its a game where the AI acts like a deathmatch bot.

[And: the whole idea of something being a "gamble".... if you want something to be random, fine, then show me the probabilities. I can't "gamble" in a rational fashion if I don't know anything about the odds.]

Don't settle near him

Click to expand...

Define near.
2 tiles? 3 tiles?

Ever.

Click to expand...

I have to remember some guys demand for hundreds of turns, even if I stop playing and reload the game a week later?

How is that intuitive?