Report Questionable Behavior

[Uncle_Joe]

The unconditional statement about committing to all out attacks is too strong. I think this after reading a large number of posts from players who wanted to see the AIs launch devastating attacks. Many players requested AIs that go to war more often. Sometimes the result is going to be a benefit to the warring party, sometimes not. Maybe more often not. However, this was what was called for by many posters.

Speaking for myself, I find totally peaceful games really boring. If the only reason AIs go to war is because I've manipulated them, that's boring too. Sometimes you need that insane leader that goes to war, even though the poor result can be predicted. This happens in real life

I dont think people really were looking for an AI to turn 'berserker' on them.

But yes, I understand what you are saying. IMO though, its not good for the game as a whole. If the AI wants to attack a neighbor and secure some resources or some expansion room or even just to hurt a competitor, I'm all fine with that. But it was making no efforts to consolidate its gains and was running itself into the ground to sustain the attacks.

And as I said, the only way to resist such an attack is to be spamming military left and right yourself. Sure, that makes the game more difficult, but then again I could always play with 'All AIs vs the player' too and it would be even more difficult. But it dramatically changes the game from one of 'Civilization' to a simple wargame.

So to sum it up, yes, the AIs should attack (and they always have IMO), but no, they shouldnt commit 100% to the attack...even if they win, they lose in the long run. The only way the 'win' is if their target is the human player and they make the game untennable for him. Beyond that, it simply eliminating competition and weakening itself in the long run.

This happens in real life

And in real life, nations dont prepare properly for war and trust in diplomacy or luck or divine providence not to be killed off either. Yet when the AIs do the same, its criticized for being 'stupid' and 'not knowing how to defend'. Either the AIs have to be playing 'to win' or they have to be playing 'to feel like other nations'. If all of the AIs are trying to win, then you shouldnt see either of these cases because neither is beneficial in the long term (although both have happened time and again in history).

Therein is the same problem I noted above. The game cant be tweaked to play smart and play 'real' in the same build. Because in the real world, nations tend to a lot of things that aren't really smart. My fear with the current AI is that it is no longer acting as opposing nation states but instead, as simply another player looking to keep the player from winning.

Do you happen to have any saves from this game? Particularly during Alex's wars with the other AI's (I'd like to see his economic stats).

I've attached all of the autosaves I had. I dont know how far back they go or if they will show you want to see but you can take a look.

What does "modified difficulty" mean?

I've tweaked the heck out of it. Essentially the level is still called 'Noble', but the parameters are closer to Monarch/Emperor in terms of Inflation/Maintenance/Upkeep etc. The AI doesnt get many bonuses either (although they are close to the 'Better Handicaps' file).

What I was shooting for was a level where the economic side matches the higher difficulties but the AI bonuses were not drastic.

By the way I'm thinking of tying unit training levels to aggressive AI... so aggressive AI would be militaristic AI.

Can you elaborate on this? Do you mean on the 'Aggressive AI' setting that the AI would 'spam' units but not so much on the standard setting?

 

[Blake]

Thanks for the saves.

I've made a new thread for AI Aggression levels, please move the discussion on this topic over to there.

 

[Doomed_UK]

Just finished (gave up) my first game on the new build (27/1).
3 continents, 6 AI
Americans and me on one continent. After a short while they went for me and would not let up. I could stop thier penertration into my lands, but had any counter attacking stacks wiped out, so fairly stalemate. (neither of us had siege weapons).
Unfortunately my tech research was slowed to a real crawl, and I checked the world builder - the 3 AI's on the big continent were so far ahead in tech it was not worth continuing (I don't normally ever give up, but the constant war grind and no hope of winning got to me)
On the third continent the Mongols had wiped out the Malinese.

Checking the power graph it was interesting to see that the Mongols had the most power up until the defeat of Mansa, and then kept thier power level the same - obviously adopting a different stratagy after winning the war.

edit: sorry - posted this in wrong area

 

[glokkonn]

@Uncle Joe

I gather you have never modified a large computer program.

Blake and Iustus are working at breakneck speed making many many changes to a game that sorely needs the changes.

The first set of changes - 2.08 - thats the cherry picking part of the changes. Change all the easy stuff.

The second set of changes is quite a bit tougher. Now the AI makes many more units. OK that wasn't too hard. Now the hard part is: use these units in an effective way. This is a substantially tougher project.

Yes they need to have the AIs make more siege units for collateral damage, make more offensive units etc etc etc. All the factual stuff you said... yes it all needs to be done, accounted for, and taken care of. However this is not a trivial project they are undertaking.

And yes what this means now is that sometimes the side effects of given changes will be worse for the game than the bonus the change was supposed to give to the AI. Thats part of programming too. We are not doing beta testing - more like alpha testing. They WILL make the AIs use all of this military in an effective way.

And then I guarantee that you will be forced back a difficulty level or two And with lower AI bonuses, there will be smaller stacks of AI units to contend with!

 

[Todd Hawks]

No one - least of all Uncle Joe - claimed that Blake and Iustus had an easy "job" or that they weren't doing their best to make it work.

That doesn't change the fact (and neither will your imaginary "guarantees") that Better AI has gone worse lately. And if we don't tell Blake and Iustus, how are they supposed to know? If they noticed themselves or knew exactly what the problem was, they surely would have changed it for the better instead of making it worse, right? There's only so much testing one or two persons can do, after all. Adopting a "it's all right, all will be fluffy and sweet sooner or later" attitude never did any good and it won't in this case either.
Especially considering Uncle Joe only brought up valid and reasonable concerns. In a polite manner, I might add. So far, I fully support his point of view.

 

[Quagga]

Has anybody else noticed Barbarians being a bit more aimless than usual? In the game I'm playing (07-01-25 build), the Barbs hung around outside my property well after the usual date when they rush in. In fact, they haven't stepped into my boundaries yet and its the Medieval Era.

Just a fluke, or something that has changed...?

 

[Ishon]

Or you just built the Great Wall and missed this crucial fact.

 

[Quagga]

Or you just built the Great Wall and missed this crucial fact.

Hee, hee. Yes, I built it in my sleep. JK!

 

[Ralgar]

I could stop thier penertration into my lands, but had any counter attacking stacks wiped out, so fairly stalemate.

I lost my first 25/01 game too! (And i reloaded a couple times... )

Monarch, Terra, Standard. Non aggressive AI.

I (Carthago) expanded quckly, very good start position for my early cities with plenty of production and resources. My south neighbour was Mehmet and i hit him hard in an early war and conquered his capital (with great lighthouse) and two other cities. Then I gained tech lead and adopted Hinduism to tigh myself to the most powerful civ, the French (Ludwig). Then Montezuma attacked me. Brutal. After I lost two key border cities i reloaded.

Second attempt:
I expanded quickly, and this time I made friends with Mehmet and especially Psycho-Montezuma and settled all my cities by myself. I built reasonable Military, in all cities a few Units. I didn't choose a religion (which sucks, because you can't take the 25% production bonus or great persons take longer) and have big positive modifiers with almost all Civs. Tech race is okay, the french only have five cities but are leading.

Then Stalin attacks me, the turn before he was pleased with me. Thanks to my early focus on production i manage to throw back his attack (btw: Numidian Cavalry sucks against normal horse riders) and with knights, trebuchets, pikeman and maceman i invade the russion land. Ask for peace. Stalin wants one of my cities. Slaughter medium size stacks of him. No peace. Conquer a city from him. No peace, but Mehmet declares war on me. I produce every turn knights, but Mehmet has musketman. After some time i give up, i am not losing, but war against two enemys is crippling me serious.

It was great fun! Good job on the 25/1 build

But i don't know what i could have done better. The other AI sucked as friends, my best friend Ludwig never helped me. Only the Mongols could be bribed to war.

Then, if you are in a long stalemate with an enemy your research is hurted. Which means you never gain enough Techs to bribe a builder Civ to war. Or get your enemys to a truce. Which sucks. Or i suck. Back to prince i think...

Long story short:
It should be easier after a long war to sign for a ceasefire.

 

[Arlborn]

Long story short:
It should be easier after a long war to sign for a ceasefire.

Blake changed it, and for good. The reason is simple(not sure it was your case exactely), if the AI really thinks it still has more to gain in the war, it wont want peace then.

 

[jkp1187]

Blake changed it, and for good. The reason is simple(not sure it was your case exactely), if the AI really thinks it still has more to gain in the war, it wont want peace then.

I have to agree with this sentiment. The AI shouldn't show any mercy; you wouldn't.

 

[Uncle_Joe]

The AI shouldn't show any mercy; you wouldn't.

IMO, the AIs should act according to the way they were originally intended, not as bloodthirsty killers. 'Better' AI should be just that...better efficiency of the original intent, not changing the way the AIs operate to such point that the game of Civilization is lost in the shuffle.

Civilization was not intended to be solely a wargame or a game of military conquest. Changes like these and the unit spamming and the all-out commitments to war threaten to reduce the game to just that...a game of military conquest.

I dont want to see this turning into something that is geared simply towards a military style of play. That is not making the AI 'better', it changing the focus of the game. IMO, these types of changes are not good for the game at all.

That said, I dont believe the AI should give up in a war simply because a certain amount of time has expired etc. If it has no WW and is still making gains at reasonable cost (and not following behind in the tech and development race), then by all means it should continue the war. But if its cities are under blockade and their land is being torn up by pillaging etc then continuing to grind out even a slightly successful war is not good strategy. A human player may do it out of spite or whatever, but that doesnt make it too terribly bright.

 

[Arlborn]

That said, I dont believe the AI should give up in a war simply because a certain amount of time has expired etc. If it has no WW and is still making gains at reasonable cost (and not following behind in the tech and development race), then by all means it should continue the war. But if its cities are under blockade and their land is being torn up by pillaging etc then continuing to grind out even a slightly successful war is not good strategy. A human player may do it out of spite or whatever, but that doesnt make it too terribly bright.

Im guessing here but I think it's probably al ittle hard ot the AI to determine exactely when its better stop.. But in vanilla it would stop after 30 tuns(was it?) doesent matter what. That is a little dumb if you are just before the capitol of your enemy for example..
The thing is, the going for war should depend more of their personality, but after they are in a war, it should be for win or at least not lose, right?

And did you see my reply in the other thread aobut agressiviness? The AI in my continent is not agressive AT ALL, maybe because they are both 'peacefull' and as we didnt meet any other AI yet, they dont get more power cuz they think they are fine in comparation with the others that they already met..Or something like that.


Sorry if didnt make much sense, my english skills are not 100% and my keyboard skill is near 0.

 

[Ishon]

I find the question "what should the AI do" quite annoying. The answer is simple:

The AI should pursue the goal which is its victory in the game (no matter what type of victory it would be). I don't think it should have only one type of victory in "mind", it should simply do its best to win.

So: should the AI show mercy or not? Lol, obviously not. It should pursue the GOAL = victory. It should end the war it is winning only if it pays. Eg., it expects some difficulties in close future (say, the enemy is about to develop a new kind of weapon that might enable it to take back a city or two). Another reason for such a cease-fire would be the need to build up more units just to continue the war after several turns. A reason could be also the size of the empire and the need to develop the new acquisitions so that they don't pull the national economy too much down.

I absolutely disagree with the argument that the game "should let the player d o X and Y, so the AI shouldn't be too aggressive towards the player". The AI should pursue victory, period.

The reason why overinvestment in military is dumb is simply that it induces general arms race. But if the arms race actually harms the other civilizations more than it hurts us, it pays to stimulate the race. If my economy can sustain 100 units and nobody else can sustain this much troops, I might go for the arms race. I might not do so if other countries have weaker economies, but they are involved in military pacts that provide enough protection to them even if each of said states can only upkeep 50 units without much harm to its economy.

I mean, the AI should analyze what is best to it and do what is best to it. If some player has problems about it, they should reduce the difficulty level.

 

[Elandal]

The "AI should pursue victory" point of view is only one. It leads to SP becoming MP game with AI opponents.
The other point of view is that the AI is meant to be an obstacle to the human player, not an opponent.

I think the Aggressive AI behaviour is a good change. Maybe in time that little checkbox will be what separates MP-type AI from SP-type AI - one that is going for a win and also prevention of victory by someone else, the other a more docile version.

 

[Arlborn]

The "AI should pursue victory" point of view is only one. It leads to SP becoming MP game with AI opponents.
The other point of view is that the AI is meant to be an obstacle to the human player, not an opponent.

I think the Aggressive AI behaviour is a good change. Maybe in time that little checkbox will be what separates MP-type AI from SP-type AI - one that is going for a win and also prevention of victory by someone else, the other a more docile version.

Im my opnion(MINE), that is just stupid.

Anyway if the AI is winnig too much over somebody, that person clearly needs to go a level down.

The objective of this as far as I know is to make AI better. One of things is to make it also able to win, hence the Cultural victory.

Now if somebody want the AI only as an 'obstacle', well, maybe its not the right project for them?

And because of that Im 99% sure Firaxis wont introduce much of it in the game anymore. So who dont like a 'better'(still needs improviment, but I mean an AI actually trying to win and not only being an annoyance and doing nothing good) AI, well...

Thread:What is Better AI .

A little part of it:

"Better AI improves the AI's decision making, enabling the AI to put up a stronger game with less reliance on handicaps and ultimately to play a balanced game more like a human."

 

[Uncle_Joe]

The AI should pursue victory, period.

OK, so by that rationale, why bother to have 'relations' with the AI? Why bother trying to develop good relations or share a religion or anything of the sort because if the AI is playing only to 'win', then none of that would matter a whit. If it 'thinks' it can get something out of a war with you, then it should attack regardless of previous dealings? After all, human players would never care about 'favorite civics' or 'years of peace' or anything else. So you are saying all of that should be thrown out for the AIs too?? (because thats exactly where this is leading).

Sure, that might make the AIs 'better' in some ways, but its throwing out crucial concepts of the game of CIVILIZATION. All of those modifiers and attitudes and whatnot are there for a REASON....to give the illusion that you are dealing with rival nation states and NOT just another random player out there trying to 'win the game'.

And this is EXACTLY why I proposed branching the product between those who want the AI to try to win at any cost and those who still want to play Civilization, but without the bugs and poorly optimized AI economic routines.

Dont get me wrong, I CAN play a militaristic style of game and I CAN win when the AIs are building piles of units and trying only to win the game. But its no longer the SAME GAME as Civ4. My impression was that the Better AI was not going to change fundamental elements of the game, but making the AIs play solely to 'win' regardless of other circumstance is COMPLETELY throwing the core concepts of Civilization out in favor a game of conquest.

 

[Arlborn]

OK, so by that rationale, why bother to have 'relations' with the AI? Why bother trying to develop good relations or share a religion or anything of the sort because if the AI is playing only to 'win', then none of that would matter a whit. If it 'thinks' it can get something out of a war with you, then it should attack regardless of previous dealings? After all, human players would never care about 'favorite civics' or 'years of peace' or anything else. So you are saying all of that should be thrown out for the AIs too?? (because thats exactly where this is leading).

Sure, that might make the AIs 'better' in some ways, but its throwing out crucial concepts of the game of CIVILIZATION. All of those modifiers and attitudes and whatnot are there for a REASON....to give the illusion that you are dealing with rival nation states and NOT just another random player out there trying to 'win the game'.

And this is EXACTLY why I proposed branching the product between those who want the AI to try to win at any cost and those who still want to play Civilization, but without the bugs and poorly optimized AI economic routines.

Dont get me wrong, I CAN play a militaristic style of game and I CAN win when the AIs are building piles of units and trying only to win the game. But its no longer the SAME GAME as Civ4. My impression was that the Better AI was not going to change fundamental elements of the game, but making the AIs play solely to 'win' regardless of other circumstance is COMPLETELY throwing the core concepts of Civilization out in favor a game of conquest.

I probably didnt make myself clear enough as I was afraid.
Here is one of my phrases:
"The thing is, the going for war should depend more of their personality, but after they are in a war, it should be for win or at least not lose, right?"

I meant that, the AI shouldent go in a war against their personality or against a friend(unless the AI has a backstabber personality). But if the AIis in a war at the moment, doesent matter the reason of the war, then YES it should get as much as it can, after all its in a war anyway no? If its in a war, the AI decided to attack according with its personallity and relationships, or got attacked. Either way its against an enemy and should try to win.


Did I make myself clear now?

 

[Uncle_Joe]

I probably didnt make myself clear enough as I was afraid.

and

I meant that, the AI shouldent go in a war against their personality or against a friend(unless the AI has a backstabber personality).

You are quite clear, dont worry. But what you are saying is not what some seem to be wanting and that is that the AI should play like a human player and disregard all the same things a human player might with the sole goal of victory.

I freely accept that there are players out there who desire this and think it would make for a better game. I just disagree. I believe it would make a different game, but not necessarily a better game. What drew me to the mod was the fact that gameplay was not going to be altered. But making radical changes to AI behavior (including rampant unit spamming) IS altering the gameplay quite a bit.

Look at it this way. Going from Warlords 1.0 to Warlords 2.08, there is marked improvement in the AI. The 2.08 AI runs its econ better and manages teching and production etc better than 1.0. But the gameplay is completely UNCHANGED.

Going from 2.08 to the 1/25 build, the game does not play even remotely closely. Military production HAS TO BE your main focus. Otherwise you will be crushed. There is no other way around it. You either 'keep up with the Joneses' or you lose. And just like the AI, if you are going to build a huge military, then there is a need to get some payoff from it. And that means that I'm far more likely to warmonger in 1/25 than in 2.08 (but I'm sure not everyone feels that way).

What I would like to see is the logical progression from 1.0 to 2.08 to the next build where the AI is better optimized, but the entire way the game must be played is not so radically different (ie military stressed above everything else).

And having the AI's behavior changed to 'win at any cost' is going even that much further off the baselines of the game established in Civ Vanilla, Warlords 1.0, and 2.08. In all of those upgrades and changes, the core gameplay did not radically change as it has between 2.08 and the last couple of builds. I, personally, do not want to see it to continue moving even farther off. At the same time, I recognize that apparently others DO want to see that. Hence again, my simple proposition for a more optimized 2.08.

 

[Arlborn]

I think the game shouldent go againt their personality, BUT they should be able to win in some situations even when they personality is quite hum..Ah well you got my meaning I hope.


Man, in vanilla I ALWAYS had the worse power graph and NEVER got dogpilled for example. Nor any near of get wiped out. That is kinda dull even if no agg. AI is on.

I always hated the fact that AI only tried Space Ship victories, and not even all of them.


And so on..For me that is the meaning of this project.

Maybe right now its not 100%, but it will get there eventually in my opnion