Excellent diplomacy analysis just popped-up on the 2k forums of Civ5

[ohioastronomy]

What actually happens in this case is you get branded a warmonger and nobody will trade with you, or only at exorbitant prices -- 3 of your luxuries for 1 of their own, that sort of thing. Your happiness goes to crap, your cities stop growing, and pretty soon everyone's outstripped you in science, gold, and military forces.

This is kind of bizarre, really. Half the threads griping about Civ 5 diplomacy claim it's unfair that other civs treat you as a warmonger for destroying entire nations and the negative effects make it too hard to win as a conqueror, and now we get claims that you can destroy entire nations and the other AIs don't even notice.

Except, oddly enough, that doesn't match my experience at all. The benefits from trade just aren't good enough - if they actually mattered it would be a different issue.
Nor does it match my specific beef with the broken diplomatic system in this game. I can either struggle with a poor and opaque system or just throw up my hands and do what the game wants you to do (be a warmonger.)

I'm perfectly OK with the game treating you as a threat if you conquer people - but not because they decided to drop a city down next to you at random, or "you're trying to win the same way I am", or because I'm about to launch a spaceship.

Other people expressing their unhappiness with the severity of the warmonger trigger (which is exessive) are not contradictory; those are just distinct problems.

 

[King Patrick]

Except, oddly enough, that doesn't match my experience at all. The benefits from trade just aren't good enough - if they actually mattered it would be a different issue.
Nor does it match my specific beef with the broken diplomatic system in this game. I can either struggle with a poor and opaque system or just throw up my hands and do what the game wants you to do (be a warmonger.)

I'm perfectly OK with the game treating you as a threat if you conquer people - but not because they decided to drop a city down next to you at random, or "you're trying to win the same way I am", or because I'm about to launch a spaceship.

Other people expressing their unhappiness with the severity of the warmonger trigger (which is exessive) are not contradictory; those are just distinct problems.

So the AI is broken because the penalties for being a warmonger are too light while the AI is broken because the penalties for being a warmonger are also too harsh. Okay, then.
Moderator Action: You could have formulated this in a much better and not trollish way.

 

[ohioastronomy]

So the AI is broken because the penalties for being a warmonger are too light while the AI is broken because the penalties for being a warmonger are also too harsh. Okay, then.

I try to actually understand what other people write. It's a useful skill; I'd suggest trying it.
Moderator Action: You could have formulated this in a much better and not trollish way.

 

[Diamondeye]

I can walk on eggshells (as you describe)...or I can do something simpler. Take out everyone near me at the start (before they get difficult to deal with), and it doesn't matter what more distant people think about me because the AI can't handle naval invasions. It's a better way to win, always, than the other route. I don't need to worry about random acts of aggression and obscure diplomacy, after all.

This seems a redundant reply to an answer about how to win peacefully... "Yes but I could also kill them!". Yes. But that was not the point. It is perfectly possible to win this game peacefully, and waging war will stigmatize you diplomatically. Which means the best option is to keep warring....

 

[mlund]

The real issues stem from making the AI play more like a human player. Humans know that this is a competitive game. If anyone else wins then you lose. Then means there are no permanent alliances, no altruism, and nothing matters other than whether a player is increasing or decreasing your chances to win.

The flip side is that A.I. is still trying to simulate an in-game persona as well. No human player cares if you are a "warmonger" or beat down random city-states no one has ever heard of half a world away. They care about how likely it is that you disrupt some of their plans by attacking their interests (cities, allies, etc.) and they care how likely it'll be that you out-pace their own victory condition.

Because of the conflict between simulationist and gamist concerns in the A.I. it feels like the A.I. players are irrational beasts. If the A.I. only cared about how much power you projected in terms of victory conditions and threatening their assets, I think the complaints would be much less of "the A.I. is crazy" and far more of "all the A.I. gang up on me all the time because I play better than them." Frankly, I'd rather have it that way, though. That way the A.I. turning on other A.I. would be governed by sensible threat levels, instead of things like perma-war triggers than competitive human players wouldn't bother with.

"Oh, Alexander is going to build the U.N. in 10 turns and win by diplomacy? I don't care - you stole my cows!" <- That's the kind of thing that makes people tear their hair out when they start to accept the idea that the A.I. is just being competitive in its diplomatic stances.

Also, the fact that A.I. is globally informed of your nefarious deeds even when they have no contact with your victims is completely unacceptable in terms of game mechanics or simulation. Players are not Jedi. They do not feel a "great disturbance in the force" when you kick over an otherwise undiscovered sand-castle in a remote corner of the world. Even if word gets back through common chatter it should be handled as base rumor with far less weight than if a player had their own diplomatic knowledge of the incident in question. I don't want to be informed out in Mongolia somewhere that "Unknown Civilization" just horse-whipped Rasuga (who I never met) and suddenly know England to be a warmongering menace when I encounter them 100 turns later. Same goes for the A.I. knowing what I did.

So, in short summary:

- Make the A.I. gamist or simulationist - but not both, make it an Advanced Setting toggle or something
- End perma-war status in competitive scenarios
- Limit out-of-character knowledge transfers that impact diplomacy

That's what I want from the Diplomacy A.I.

- Marty Lund

 

[Civsassin]

If I'm going for a space race win, for instance, I don't want the computer to just roll over. But I want them to compete "appropriately* - that is, try to build their own spaceships, or culture, or diplomacy, or whatever. Some nations might choose combat - but the whole point of a peaceful victory condition is that you aren't forced to always conquer the planet.

So what do you do when you find you are seriously behind and have not good way to catch back up? Do you go to war or reload and try again? Do you keep the level so low that the AI is inept enough, so you can always win? The AI exploits weaknesses as human opponents do. Compete "appropriately" sounds like I don't like it if I am weak and the AI exploits my weaknesses. Here's the answer: play on settler you can even try warlord if settler gets boring. You should rarely if ever go to war, and you can peacefully build your spaceship and peacefully enjoy the victory screen.

Your argument doesn't make any sense.

Moderator Action: You could have formulated this in a better, less trollish way.

 

[krc]

The real issues stem from making the AI play more like a human player. Humans know that this is a competitive game. If anyone else wins then you lose. Then means there are no permanent alliances, no altruism, and nothing matters other than whether a player is increasing or decreasing your chances to win.

- Marty Lund

Well...

A long time ago (in a galaxy that's quite close, actually), I occasionally played a (board) game called Diplomacy that involved actual human beings interacting in the same physical location. I can recall one game in particular that had evolved into a three-on-three competition between two alliances. Because of the state of the world at the time, my nation was on the front lines and my two allies were offering support. But that meant that I was starting to grow larger than they were. The key point in the game came when (suspecting that I was about to be back-stabbed), I offered a province to each one of my allies in order to keep the alliance alive and our sizes roughly comparable. The game ended when our three-way alliance destroyed our three opponents and then declined to go to war with one another.

So, it is possible even for competitive games among humans to develop permanent alliances....

 

[mlund]

So, it is possible even for competitive games among humans to develop permanent alliances....

I've played Diplomacy many times. There's no such thing as a permanent alliance in Diplomacy unless someone is being rather unsportsmanlike about the whole thing and rolling over for a buddy. Even in your anecdote you'll notice that what kept your alliance together was the concession of territory to maintain a balance of spoils among your alliance. (For those who haven't played, the only mechanic power in Diplomacy the only power is simple numeric military superiority, and supply center territories have a 1-1 relationship with military units - all of which have the same combat strength rating.)

It is fine for A.I. to team up when it is mutually beneficial to acquire spoils or tear down a mutual threat (usually the human player at mid-and-lower difficulties). The only time the back-stab and gang-pile against someone who is an imminent threat for victory should not come from a gamist A.I. is when it sees itself as already eliminated from contention and makes some sort of sympathy or spite-driven attempt at playing Kingmaker. That should be a simple weighing of who they hate the most or like the most (that guy took half my cities, the guy was always a fair trading partner, etc.) No A.I. that's still in contention should ever pull a punch on somebody (AI or Human) that's threatening to close out the game - especially to pursue some eternal vendetta over a 3rd party's cow pasture being stolen 150 turns ago ...

I guess what I'm saying is the sorting order of relevance is screwed up. Gamist A.I. needs to prioritize all other players in terms of threats and opportunities in terms of victory. Once victory is off the table, survival becomes top priority. If survival is not in question then like/spite should be a factor. The problem with mixing Simulationist and Gamist A.I. is like/spite can jump out in front of victory or self-preservation as concerns, leading to "irrational" A.I. behavior. The worst is when the A.I. swings wildly between how it treats the Human player vs. the other A.I. - IE, chain DoW on a human trying to build the U.N. but completely unwilling to go to war with another A.I. building the U.N.

- Marty Lund

 

[Pac-Dragon]

Voluntarily quitting a game (that wasn't the result of a stalemate) sounds like a pretty rare exception rather than a standard game of Diplomacy. If you had players who didn't care about winning, any war game could end the moment all of the remaining the players felt they liked each other too much to keep fighting.

I think mlund's point about competetive human players is still valid despite the bizzare way you've had a few board games end.

 

[JoeBas]

The defenders of this game seem to miss a key point: there are a lot of us who really want a way to win games like Civ without warfare.

It's there, and called Civ V.

I win about 85-90% of the games I play, and I've won a grand total of ONE domination victory - my first game, when I was mainly play-testing the 1upt thing for myself.

Usually win diplo or science, or sometimes culture. Never win domination, and have had games with no declared wars the entire game.

Play on a large or huge map, with extra civs... taking over the world is VERY difficult, from a happiness management standpoint.

 

[ohioastronomy]

So what do you do when you find you are seriously behind and have not good way to catch back up? Do you go to war or reload and try again? Do you keep the level so low that the AI is inept enough, so you can always win? The AI exploits weaknesses as human opponents do. Compete "appropriately" sounds like I don't like it if I am weak and the AI exploits my weaknesses. Here's the answer: play on settler you can even try warlord if settler gets boring. You should rarely if ever go to war, and you can peacefully build your spaceship and peacefully enjoy the victory screen.

Your argument doesn't make any sense.

What you're saying is that there should be absolutely no peaceful victory option; whenever a player is close to winning the AI should all declare war on them.

If that is the game - why have any non-military victory condition at all? A good game would give me viable and distinct victory paths - and I could win or lose at any of them. So when it looks as if I'm jumping into a tech lead the AIs retool and fight me that way, for example.

As far as the difficulty level swipe by you goes (that's pretty insulting, by the way) - Civ 5 isn't even challenging to me on Immortal. And I've always detested the "stack the dice" approach used on Deity. I don't want to play a game of Chess where I have no queen or rooks; I want to play a game where I have a challenge when both of us have all of our pieces.

By the way - I see a very consistent refusal in the defenders of this game on this thread: none of you appear to be making any sort of good faith attempt to understand what I'm saying.

I think that it's broken to have victory conditions which are supposed to be peaceful ones where the AI declares war on you when you're "too close."

I think that it's broken to have the AI attack you when they think "you're trying to win the game the same way they are."

I think that it's broken to have it be so difficult to repair bad relations - which can be caused by random coin flips.

I think that it's broken to have the game reward war so strongly (no war weariness for thousand year conflicts) and peace so weakly (essentially no trade income).

I think that it's broken to be unable to have a true diplomatic win and a true strong (NATO-like) alliance.

I think that a game with obscure and capricious diplomacy strongly favors a play style that "cuts the Gordian knot" - since it's difficult to have peaceful neighbors, just don't have neighbors. Then choose your own adventure.

 

[Babri]

I think that it's broken to have victory conditions which are supposed to be peaceful ones where the AI declares war on you when you're "too close."

I think that it's broken to have the AI attack you when they think "you're trying to win the game the same way they are."

I think that it's broken to have it be so difficult to repair bad relations - which can be caused by random coin flips.

I think that it's broken to have the game reward war so strongly (no war weariness for thousand year conflicts) and peace so weakly (essentially no trade income).

I think that it's broken to be unable to have a true diplomatic win

I agree with all statements except the 1st one. ciV needs a lot of work in diplomacy.

 

[King Patrick]

I think that it's broken to have victory conditions which are supposed to be peaceful ones where the AI declares war on you when you're "too close."

I think that it's broken to have the AI attack you when they think "you're trying to win the game the same way they are."

I think that it's broken to have it be so difficult to repair bad relations - which can be caused by random coin flips.

I think that it's broken to have the game reward war so strongly (no war weariness for thousand year conflicts) and peace so weakly (essentially no trade income).

I think that it's broken to be unable to have a true diplomatic win and a true strong (NATO-like) alliance.

I think that a game with obscure and capricious diplomacy strongly favors a play style that "cuts the Gordian knot" - since it's difficult to have peaceful neighbors, just don't have neighbors. Then choose your own adventure.

Is it possible to set the number of AIs to zero? Seems like that might solve all your problems. Then you can build your empire to your heart's content and pursue any victory conditions you choose, without ever having to worry about the other civs trying to stop you, "forcing" you to conquer them, etc.

Moderator Action: Please don't troll other users.

 

[mlund]

What you're saying is that there should be absolutely no peaceful victory option; whenever a player is close to winning the AI should all declare war on them.

Likewise whenever an A.I. is close to winning all the other AI should declare war on them. If someone else wins, you lose. The AI should not play to lose. Without a pacifist mod military action will always be on the table for other players (AI or Human) as at least a means of last resort to avoid losing.

If that is the game - why have any non-military victory condition at all?

One of the universal functions of government is to secure the liberties of its citizens by protecting its borders from foreign aggressors. You don't have to win Domination Victory, but you do have to defend your citizens by disabling aggressors.

A good game would give me viable and distinct victory paths - and I could win or lose at any of them.

A good game puts the survival responsibilities of your civilization in your lap. Most civilizations, historically, got wiped out by invasion. If you don't want to have to defend yourself from rivals (and make no mistake, those other Great Civilizations are explicitly stated to be your rivals) then you need a Sim City mod of some sort.

So when it looks as if I'm jumping into a tech lead the AIs retool and fight me that way, for example.

Considering the amount of infrastructure involved, the time consumption, and the effect that terrain and civilization abilities have on science output you're just asking for the AI to lose gracefully instead of seriously impede your chances at winning.

By the way - I see a very consistent refusal in the defenders of this game on this thread: none of you appear to be making any sort of good faith attempt to understand what I'm saying.

I understand what you've been writing, anyway. You're completely keyed onto an invalid Straw Man Argument that because the AI will attempt to beat your face in if you make a break for the finish line victory conditions other than Domination are "broken." That claim is false on its face and you should abandon it if you want your arguments to be taken seriously.

I think that it's broken to have victory conditions which are supposed to be peaceful ones where the AI declares war on you when you're "too close."

Not a bug - definitely a feature that's Working As Intended.

I think that it's broken to have the AI attack you when they think "you're trying to win the game the same way they are."

This is probably engineered badly. If you try to win the same way they do but are doing it worse they shouldn't give a hoot. If it looks obvious that you'll out-race them, though, they need to start adjusting their strategy to out-pace you. Part of that is going to be trying to trip you up by hurting you diplomatically and militarily. War-by-proxy should be especially popular for opponents who are far away from you and / or need to focus on their economy for Science or Culture victories. Likewise they should do so vs. rival AI Civs as well.

I think that it's broken to have it be so difficult to repair bad relations - which can be caused by random coin flips.

Totally agreed. If you are more of an asset than an hurdle to their advancement they should stay on good terms until it becomes necessary to dispose of you.

I think that it's broken to have the game reward war so strongly (no war weariness for thousand year conflicts) and peace so weakly (essentially no trade income).

War works. The cost of war comes from economic damage. It's true in real life too. War-weariness generally only results from military campaigns when the costs hits the proles in terms of taxes and conscription.

I think that it's broken to be unable to have a true diplomatic win and a true strong (NATO-like) alliance.

It is a game. Nobody should be expected to volunteer for themselves to lose a game. Even NATO-like alliances only exist due to mutual threats. When you become the threat (of losing) you're out of the club.

I think that a game with obscure and capricious diplomacy strongly favors a play style that "cuts the Gordian knot" - since it's difficult to have peaceful neighbors, just don't have neighbors. Then choose your own adventure.

One thing I find missing is that the game gives you very little opportunity to hide information and deceive other civilizations into believing you aren't a threat until it is too late. Players enjoy too much perfect knowledge. There's also no non-military way to put road-blocks in the way of another Civilization's internal affairs. It would definitely be more interesting if you had to actually build a diplomatic / intelligence network through other Civs to obtain this knowledge - perhaps as a sub-function of Culture. Then there could be options to attack other civilizations non-militarily through subversion and espionage. Maybe it could be as inelegant as Civ 2/3 Diplomats - a unit that travels even without Open Borders and attacking it without being at war causes a huge reputation hit. The effectiveness of your subversion attacks using Diplomats (Spies) would be based on your Culture score.

- Marty Lund

 

[r_rolo1]

Voluntarily quitting a game (that wasn't the result of a stalemate) sounds like a pretty rare exception rather than a standard game of Diplomacy. If you had players who didn't care about winning, any war game could end the moment all of the remaining the players felt they liked each other too much to keep fighting.

I think mlund's point about competetive human players is still valid despite the bizzare way you've had a few board games end.

The issue with the line of thought that mlund is following is that , if streched to it's logical conclusion, leads to a game where diplomacy is not needed: if everyone is a coldish manouver master, you can get pretty much everything you want from them by simply not attacking them at the time you think it is better not to. For that you might even be in a Always War game

x-post

@ mlund

Likewise whenever an A.I. is close to winning all the other AI should declare war on them. If someone else wins, you lose. The AI should not play to lose. Without a pacifist mod military action will always be on the table for other players (AI or Human) as at least a means of last resort to avoid losing.

That is fine and dandy ... in a 1:1 game. Think on this: if ( as you point well in the subsequent entries of the same post ) war works but it has a cost, a really smart player ( human or not ) would think in a situation like the described above: "Should I attack that guy, or should I wait to see/ manouver someone to attack him while I stay in home eating popcorns and maybe go there in the last straws to be on the winners table ( like USA did in WWI ... or Spain in WW II )?" Same benefits, less cost

 

[mlund]

The issue with the line of thought that mlund is following is that , if streched to it's logical conclusion, leads to a game where diplomacy is not needed: if everyone is a coldish manouver master, you can get pretty much everything you want from them by simply not attacking them at the time you think it is better not to. For that you might even be in a Always War game

I don't really understand the way you constructed your sentence, but I'll try to follow along. I don't see any issue with all the players (AI or Human) operating in their own self-interests and full knowledge that they are just players in a game rather than people trying to role-play Caesar or Ghandi or whatever. Everyone is trying to beat you (and everyone else) 100% of the time. You just need to give them good reason (or deceive them into thinking there is good reason) to work with you for the time being rather than against you.

That's exactly how you play a game. The tricky part is handling imperfect knowledge, choices beyond your control, and random chance.

That is fine and dandy ... in a 1:1 game. Think on this: if ( as you point well in the subsequent entries of the same post ) war works but it has a cost, a really smart player ( human or not ) would think in a situation like the described above: "Should I attack that guy, or should I wait to see/ manouver someone to attack him while I stay in home eating popcorns and maybe go there in the last straws to be on the winners table ( like USA did in WWI ... or Spain in WW II )?" Same benefits, less cost

Yes, that's how it should be. The A.I. already has the idea of bribing City States to use as allies in wartime, it should probably be able to handle the idea of sub-contracting wars to another Civilization when it would be more efficient. Bonus points if the A.I. sends Alexander over to mess up your Apollo Program, garnering him the major warmonger title, and then denounces and DoW's Alex to sack Sparta and Athens after Greece makes peace with you. The day the A.I. manages to beat a human player via United Nations, Culture Victory, or Space Race due to a proxy-war + back-stab city-grab everybody should get together and buy Firaxis a beer.

- Marty Lund

 

[RedneckRapture]

All my science, diplomacy, and cultural victories have ALWAYS had war involved in them. The key is knowing how you're going to win and plan accordingly.

If you're going to get a domination victory, don't worry about any alliances lasting. Make short term alliances to deal with one threat at a time. By the time everyone is mad at you, you'll own at least 2 continents and have everyone outstripped by a couple of hundred points.

If you're going for a cultural or science victory, have an army large enough to defend, but not invade. Have a group of friends, not all the AI will remain friendly to you, but there will be ones who will. You may avoid war simply by having allies in place who would make life difficult for your enemy.

For diplomacy, defend your city states. You should have a force that is mobile enough to respond to threats to your city states. Having this force will allow you to pounce any time one AI wipes another AI out. The liberated AI will vote for you.

Above all else, remember this when looking at diplomacy in this game. There has never, EVER been a civilization in the history of mankind that has not gone to war with another civilization and in a lot of cases wiped one out. No matter what victory you are after, you will NEVER be free from war.

Also, griping that other countries won't vote for you in the UN unless you've liberated them is just plain foolish. Look at the world today. Experts agree the United States is on top. But NOBODY is going to petition their government or pass legislation applying for statehood in the US. There is simply too much pride in the individual nations for them to do that. The same thing is going on in Civ 5.

 

[ohioastronomy]

The secret to the Civ series has always been that people enjoy it in different ways. Some people like a simulation, for instance, and others like a sandbox. The "always war" argument, however, also fails as a game mechanic. I think that a lot of folks posting here really don't have a lot of exposure to the insights from game theory, and what we've learned in game design. If you want a wargame, great. But if you want replay value you'd like distinct and equally powerful options, and they don't need to involve war.

To take a concrete example, look at the spaceship win. If the AI will attack you if "winning" then the rules of the game demand conflict. The structure of a 4X single player game favor the attacker (the defender is doing other things, while the attacker maximizes development, gets the largest possible military, and strikes at the perfect time for them.) So the structure of the game, in effect, provides a strong incentive to always attack your neighbors - after all, they're going to attack you no matter what if you're "winning."

By contrast, imagine a game (like the GalCiv series) where peaceful trade is profitable and helps research, or where extended wars can wreck your economy and warmongers get taken down a peg (but not eliminated) by collective action (e.g. Europa Universalis 3.) In this game you want to avoid conflict with your neighbors and tech away; warmongers are in a race to see if they can conquer everyone else before the other players can out-tech and out-spend them enough to stop them.

Now the rules in the latter sort of game do prevent last-minute backstabs - but you're trading the loss of such choices against a much richer set of approaches. By contrast, the stab-the-leader game model has a far more restricted set of possibilities.

 

[r_rolo1]

I don't really understand the way you constructed your sentence, but I'll try to follow along. I don't see any issue with all the players (AI or Human) operating in their own self-interests and full knowledge that they are just players in a game rather than people trying to role-play Caesar or Ghandi or whatever. Everyone is trying to beat you (and everyone else) 100% of the time. You just need to give them good reason (or deceive them into thinking there is good reason) to work with you for the time being rather than against you.

That's exactly how you play a game. The tricky part is handling imperfect knowledge, choices beyond your control, and random chance.

I'm simply pointing that, if are playing a competitive game, you have no real need for diplomacy ( this from a game designer perspective, that is ). You can "cooperate" with a enemy by simply not attacking him at a time and still act in your own self interest ( there are tons of strategy games in Always war enviroment ... even some previous versions of Civ for mobile phones are like that ). So if you actually want a competitive game, you don't need diplomacy at all...

That was the point i wanted to convey above. Using the fact that X game is competitive or not to support any diplo system is a fallacy, because competitiveness and diplomacy are unrelated . OFC that they can work against each other, but that is another issue, that can't be separated of all the other aspects of the game ...

Yes, that's how it should be. The A.I. already has the idea of bribing City States to use as allies in wartime, it should probably be able to handle the idea of sub-contracting wars to another Civilization when it would be more efficient. Bonus points if the A.I. sends Alexander over to mess up your Apollo Program, garnering him the major warmonger title, and then denounces and DoW's Alex to sack Sparta and Athens after Greece makes peace with you. The day the A.I. manages to beat a human player via United Nations, Culture Victory, or Space Race due to a proxy-war + back-stab city-grab everybody should get together and buy Firaxis a beer.

- Marty Lund

Note, I'm not asking anything impossible. I'm just asking that the Ai knows how to count rifles and to analize the propensity of wars between other civs ( that , knowing XML values and/or the knowledge they already have of the wars between the parts, is a mere spreadsheet analysis ). it is not perfect, but it would neither add a lot of strain in hardware usage or in code lines while making the AI far more smarter.

Anyway, my point is that being competitive means that you have to be minimally competive also in smartness and going "Bomb Iran" as a reflexive gag everytime someone does something that menaces your win is not a particualrly good strategy if there are more than 2 players. In fact, it is only a good strat if you have a situation where one of the players has far more predictive power of the oponents moves than the rest of the players ( aka smart player vs dumb ones )

 

[mlund]

The "always war" argument, however, also fails as a game mechanic. I think that a lot of folks posting here really don't have a lot of exposure to the insights from game theory, and what we've learned in game design. If you want a wargame, great. But if you want replay value you'd like distinct and equally powerful options, and they don't need to involve war.

I disagree with the last part. While you should not need to play the conqueror to win the basic conceits of the game demands that you should always have to at least furnish a deterrent force and repel last-ditch invasions made to attempt to thwart a win if you are militarily vulnerable.

By contrast, imagine a game (like the GalCiv series) where peaceful trade is profitable and helps research, or where extended wars can wreck your economy and warmongers get taken down a peg (but not eliminated) by collective action (e.g. Europa Universalis 3.) In this game you want to avoid conflict with your neighbors and tech away; warmongers are in a race to see if they can conquer everyone else before the other players can out-tech and out-spend them enough to stop them.

That sounds great. Civilization 5 would definitely benefit from inter-civilization trade routes being a core economic factor like in Civilization 4. That would provide an ongoing economic drain when constantly warring with your neighbors, and it would provide a simple economic attack to throw against the leader - IE, everyone Embargoes you.

Now the rules in the latter sort of game do prevent last-minute backstabs - but you're trading the loss of such choices against a much richer set of approaches. By contrast, the stab-the-leader game model has a far more restricted set of possibilities.

I don't see a need to trade off. Stab-the-leader is how the game should be played if you have to boil the AI down to a simple concept that's as close to a competitive Human player as you can make it. Using trade and culture as additional knives to shove into the Scourge's back is definitely a bonus.

Also, it allows for more covert development when hanging around below the radar. Instead of rushing Space or Conquest maybe you slow-play your strategy, hording cash, keeping in second place across the board so when you instigate mass DoW on the person "closest" to victory you can just sprint to the finish independently. That's how you normally win in Twilight Imperium, after all - stay strong across the board, make someone else a distraction, and then bubble-victory.

I'm simply pointing that, if are playing a competitive game, you have no real need for diplomacy ( this from a game designer perspective, that is ). You can "cooperate" with a enemy by simply not attacking him at a time and still act in your own self interest ( there are tons of strategy games in Always war enviroment ... even some previous versions of Civ for mobile phones are like that ). So if you actually want a competitive game, you don't need diplomacy at all...

Perhaps we're confusing the definition of the word "Diplomacy" here. Nation-states engage in diplomacy all the time. A diplomatic corps has a singular mission - advance the interests of their nation. Even in eras where nations have been direct rivals vying for supremacy diplomacy has still boiled down to hammering out terms under which both nations believe their best interest are being served.

Using the fact that X game is competitive or not to support any diplo system is a fallacy, because competitiveness and diplomacy are unrelated.

The competitive nature of the situation and the nature of the game have everything to do with the appropriate nature of diplomacy. In a zero-sum game with 1 winner and X losers diplomacy is about temporary alliances of opportunity and everyone should be aware of it. In something other than a zero-sum game with a clear winner and a heap of losers diplomacy completely changes. You can have situations with "haves" and "have mores" cooperating and all coming away with more than they started, as opposed to one "have" and a bunch of "have nots" as you have in Civilization.

The real restriction on the nature of competitive diplomacy in Civilization is that second place is just the first loser. In such a situation you always throw the proverbial Blue Shell unless the guy in first place is already on his way out and you're waiting to bag the next leader.

Anyway, my point is that being competitive means that you have to be minimally competive also in smartness and going "Bomb Iran" as a reflexive gag everytime someone does something that menaces your win is not a particualrly good strategy if there are more than 2 players. In fact, it is only a good strat if you have a situation where one of the players has far more predictive power of the oponents moves than the rest of the players ( aka smart player vs dumb ones )

I agree. The AI should be looking beyond simply "stab the leader" to a plan of "stab everyone in front of me in the correct order so I can win" with a fall-back plan of "stab/boost the guy who hurt/helped me the most now since I can't possibly win."

- Marty Lund