Civilization 5 Announced

[Kylearan]

Hi,

But players will crack the formulas involving AI diplomacy very quickly, and then playing the game optimally forces the use of third party mods and/or references.

Being forced to use reverse-engineered cheat sheets and take notes of your diplo actions just to infer the diplo status of your opponents would suck, yes. However, if Krill is right and they do randomize parts of it, it *might* work - although I fail to see how you could do this without making the AI reactions totally arbitrary and thus the whole diplo mechanics useless. After all, previous incarnations of Civ put emphasis on fixed personalities of the leaders, e.g. Monty tending to be a warmonger in every game, and only adding a small random fluctuation to the internal "won't declare war" and other, similar thresholds would feel lame. Let's hope they got a little bit more creative...

Why would you want to remove feedback from diplomacy - isn't it a good thing to be able to see what effect your actions have, and be able to make informed decisions as a result?

Not if you want the AIs to be as much human-like as possible. Humans cannot be modeled with a simple cause-and-effect mechanic, you only get a rough idea and have sometimes wrong expectations of how a human opponent will treat you in a MP game, regardless of how many free techs you gift them and how many long-term resource deals you have with them.

Making an informed decision about how a human reacts can only be a rough, error-prone heuristic, because humans are more than the simple finite state machine the Civ IV AIs are (or at least we haven't understood all human states yet...). If they somehow manage to simulate this without becoming too random, it works for me.

-Kylearan

[Speaker]

nother big change is the combat system. Units are no longer destroyed if they lose a battle, which means that civs can spend much more resources on maintaining their armies as opposed to cranking out new units. The combat that we saw took place completely outside of a city, and positioning and terrain are much more important. "In the past, combat revolved around stacks [of units], which our fans affectionately call 'stacks of doom,'" Shirk said. "We wanted to pull combat out of the cities, and make every unit important." No two units can occupy the same tile, even friendly ones, so positioning on the battlefield becomes very important. Ranged units, like archers, are used to soften up the front lines from up to two hexes away, but they are vulnerable to attacks from melee units. The result is an emphasis on battlefield tactics instead of most Civ games which favored the civ that was able to crank out the most units

I have never understood why the emphasis should be on tactics and not on production. Civ is not a combat simulator, nor is it a wargame. The combat systems, interface and strategic options are far too simple for that. Civ has always been about empire management, and being able to balance manufacturing and commerce with unit building is at the core. Why exactly is the "Stack of Doom" a bad thing? Because the AI can't handle it? This is the same AI who spends valuable worker turns in the early game roading random tiles instead of prioritizing food and growth, and won't send out a settler because it's not safe, but never prebuilds roads to speed along a settler in a safe manner.

I do not have high hopes for Civ5.

[Krill]

But players will crack the formulas involving AI diplomacy very quickly, and then playing the game optimally forces the use of third party mods and/or references.

Not necessarily, it depends on the implementation. Depending on the randomization that Firaxis has stated will affect every AI leadership category, it might be that all the expert cracking comes down to: X leader needs between y and z land ceded to him for him to not declare war etc. And if Firaxis were smart, they'd make it so that the AI will never decided to not declare war on a player (ie the "Won't declare at pleased" BS that we see now).

As far as I'm concerned, taking information away from the player and hiding it away is almost never a sign of good game design. Why would you want to remove feedback from diplomacy - isn't it a good thing to be able to see what effect your actions have, and be able to make informed decisions as a result? I don't get it.

I don't see how the current implementation is much fun either.

I gift away a tech, get +2 relations with Zara, only 5 more (sic) to go and then I'm never going to get declared on. Giving the player too much information isn't better than giving the player too little.

Hey, I'm all for making the AIs more aggressive in Civ5 (it's a little too easy to pacify them in Civ4), but that's something altogether different from removing diplomatic feedback and making it impossible to know how and why they feel the way they do about your civ.

The important bit is how the feedback is delivered IMO. I don't like the plus/minus system being exposed, because all it does turn diplomacy into a bean counting exercise, just like civ 3 MM was. Hiding away the plus/minus system, and changing it so that it is different in every game, for every leader, is a plus for me. Not giving any feedback on how diplo interactions with the AI is not good, but so far I've not heard that Firaxis are planning on hiding all feedback, just the plus/minus system.

+1
I'm sure there is a good side to this, but to me this is adding in a random element in a bad way. Ok so "just keep a credible defense all of the time" is the counter, but now "fastest finish" type players will be forced to gamble on a token defense with their Friendly neighbor. I don't think "hope against a random AI diceroll not impacted by any player action" is a good game feature.

When you play against me, will I rush you, or will I try to co-operate with you so we can both build? I think the comparison between SP and MP is critical for how this might work. In SP, you meet a civ, know who it is, and everything that you have to do appease it so you can crush it later. In MP, you can't do that, because humans are devious bastards. I think giving the AI that ability is one step towards making an AI that isn't a pushover, and increases the replayability of the game.

I don't see how the idea can be argued against from this perspective tbh.

[Sullla]

But you aren't playing humans, Kylearan. The AI is never going to act humanlike, not anytime soon at least. Given that fundamental baseline for diplomatic interaction, it's better to make it clear to the player how and why the AI thinks, so that you can make strategic choices. Don't get me wrong, it would be wonderful if you could create an AI personality which was humanlike without turning their interactions into pure dice rolls and arbitrary actions, I'm just dubious that it can be pulled off.

Actually, if I'm running a Single-Player game, I think I'd rather play against a logical AI opponent, who doesn't make batshit crazy moves, than a humanline opponent who stabs you in the back in the endgame while attempting to win. Yeah, that sounds well and good in theory, but it doesn't make for a very fun game in practice.

[Kylearan]

Hi,

But you aren't playing humans, Kylearan. The AI is never going to act humanlike, not anytime soon at least. Given that fundamental baseline for diplomatic interaction, it's better to make it clear to the player how and why the AI thinks, so that you can make strategic choices.

You have a point there, and despite my previous post I'm not sure I like the new diplo system. I'll reserve judgement for when the game comes out and give Firaxis the benefit of the doubt until then.

But like Krill I don't like how easy it is to manipulate the AIs once you know how they worked, and how completely predictable they are. And while I agree with you that the AI won't be able to act humanlike (whatever that means) anytime soon, I'm happy for a crude simulation until then and will suspend disbelief if it makes the game less predictable without feeling too random. Temporarily exposing my southern flank because my units are needed elsewhere for an attack on my northern neighbour should be a gambit, not an optimal choice just because I know my southern neighbour is friendly and thus will never declare war against me.

-Kylearan

[T-hawk]

Temporarily exposing my southern flank because my units are needed elsewhere for an attack on my northern neighbour should be a gambit, not an optimal choice just because I know my southern neighbour is friendly and thus will never declare war against me.

Well, except that that's realistic. The US can leave its northern border undefended since it knows that Canada is friendly and will never declare war.

Anyway, let's all turn the hype machine down a few notches, and see what is actually in the game before we go condemning or worshiping it.

[Krill]

Well, except that that's realistic. The US can leave its northern border undefended since it knows that Canada is friendly and will never declare war.

Anyway, let's all turn the hype machine down a few notches, and see what is actually in the game before we go condemning or worshiping it.

Completely off topic, but isn't that your primary role T-hawk? You seem to say that in every other post...

[Speaker]

Completely off topic, but isn't that your primary role T-hawk? You seem to say that in every other post...

Isn't what his primary role? Being a nice guy and extremely knowledgeable about game mechanics? Yeah, that's T-Hawk's primary role.

[Krill]

Isn't what his primary role? Being a nice guy and extremely knowledgeable about game mechanics? Yeah, that's T-Hawk's primary role.

Telling everyone to turn the hype and hyperbole machine down a few notches.

[MWIN]

How about adding more modifiers to +/- rule set… Like 1) you are friendly with us, but your military is too weak 2) You are friendly, but our global strategy says we need to fight you in order to win . 3) You are friendly, but you are fighting and winning two other wars and acquiring lot of new cities. We cannot afford this.