BtS WAY to easy???

[frob2900]

The Aggressive AI in BTS is NOTHING like the Aggressive AI in Warlords. It is no longer simply more biased towards the player. It should be named Oppurtunistic AI.

http://apolyton.net/forums/showthread.php?s=&threadid=168663

There's the link where Blake describes exactly what Aggressive AI is in BTS.

Thanks Helpful info!

Aggressive AI no longer causes the AI to have a relationship penalty with humans. Basically it can be said that the AI expects things to get aggressive. The pacifist AI's aren't actually that much more likely to declare war, they just keep larger armies on hand as to not be easy victims. The naturally militaristic AI's go crazy. In any case if you neglect your army, any AI will notice and with it's larger power will be more likely to declare war and come for you. It is more likely you'll get declared upon, especially if you don't change your playstyle...


The default AI is a bit of a sandbox, you can employ the strategy you want and the AI may interfere with your plans... but on Aggressive AI, the AI can DICTATE your strategy! If Alexander is going to invade you, then you damn well prepare an army or you're going to taken out of the game! Even with the best prepeardedness if you fail to avoid a dogpile you're probably a goner.

On the normal setting, you just play, it's casual.
On aggressive AI, you adapt, it's hardcore.

Sounds pretty good I take back anything bad I said about Aggressive AI.

 

[DaviddesJ]

It seems pretty clear that the AI handicaps at the highest levels were reduced too much, in an excess of optimism about what the AI can do. Either they should go back up some, or we need to add a level above Deity again.

From what I've seen of Aggressive AI (not a whole lot), I'd personally rather have larger handicaps, instead. And then there's still the opportunity for people to do both.

 

[frob2900]

It seems pretty clear that the AI handicaps at the highest levels were reduced too much, in an excess of optimism about what the AI can do. Either they should go back up some, or we need to add a level above Deity again.

That's what I am guessing too. While I agree fully in principle with the posters who say that the AI behaviour should be sorted out rather than "increasing the bonuses", the point here is that bonuses have been massively decreased for BtS, so there is no harm in restoring some of them if necessary.

[EDIT] But then again:
http://forums.civfanatics.com/showpost.php?p=5809824&postcount=181
Solver's unofficial patch has addressed some of the issues raised

 

[Snaaty]

Some points I like to mention:

1. The AI is teching to slow on higer levels, no matter how you look at it in BtS. Simply take my save and peek around at the other conti... ....if I remember right some not even have CoL at around 1000 AD. You don´t need to play as England, beeline to rifles or a good strat. to take them out... ...simply kill them with knight and maces if you want (at 1000 AD)

2. Although the aggressive AI might solve some problems (forcing you to have more troops for example), it doesn´t fix the geneal problem:
I don´t want to meet an AI on emperor level or above around 1000 AD running around with swords...

3. Another point against the aggressive AI is, that I (and maybe some other players) like to play a builderish style. I don´t want to produce troops like hell from the beginning of the game or be engaged in endless wars...

4. I don´t like the AI going MORE troop crazy as it already does. Like uberfish mentioned some posts ago, it really IS like "banging your head against the wall" if you have to kill tons of outdated units before beeing able to advance into AI territory (not difficult, but time consuming and basically only annoying)

5. And, last but not least, I don´t want to play on deity all the time. With the starting bonuses on deity the early game just isn´t much fun and you ALWAYS depend on luck surviving it without beeing attacked (and killed) early. I want an AI that can keep up more or less decent on emperor and immortal, so some things have to be done here

...

Therefore I still favour the solution of giving the AI some boni back

 

[uberfish]

I have trouble with the mace/knight/treb approach actually. I find by the time you accumulate a late-medieval army the AI does usually get to engineering quite fast, Castles slow things down and while you are bombarding the AI emergency whips longbows out. Then you get to capture a size 3 city or something stupid.

I think the AI should train slightly less troops during medieval and spend the resources on tech instead, because I think it is currently unbalanced in favour of skipping medieval entirely.

 

[SevenSpirits]

So i'm a pretty new player; before BTS came out I'd only finished one game on each level <= Noble, in Warlords 2.08. I was looking forward to my next game, because I thought I might not have a ridiculous tech lead for the entire game.

My first game with BTS, I play on Prince and it's a walk in the park. OK, fine.

I try Monarch next. It's the same! I still have a big tech lead, and I don't even want to bother finishing the game. I guess I'll try Emperor next, but this is silly.

The problem is that the AIs play too similarly.

They do research at different speeds to one another, but they all go for the same techs! In a game with trading, this is really, really stupid. If you research a tech that no one else has, you can trade it to everyone else for one tech each. You gain N techs this way on the N other players. If you research something that is already known, it's a huge waste of resources. If more than one other player has it, any of them should be willing to sell it to you for cheap.

Every human player knows this. And the human player is willing to make much "worse" trades than the AIs, because he/she knows that trading is so strong in a multiplayer game. Note that if the AIs played smarter and were willing to trade at a "loss", they would lose even more horribly, but only because they all research the same stuff. If you can effectively trade with many players (not just the human player), trading is very powerful for you.

It's not just research paths where the AIs could benefit from more variety. Regarding Aggressive vs Normal AI settings, the human player has an advantage either way, because he/she knows what ALL the AIs are going to do. Actually, in any case where all the AIs do something, it's bad for them because CIV is to a large extent a game of counters. If all the AIs go scissors, the human player can pick the metaphorical rock and just crush all of them at once.

The best aspect of the AIs as they are is that they have different personalities. The game would be much worse if that wasn't the case. However, this does not go nearly far enough.

The AIs need to have more varying research paths.

They need to be more apt to beeline. Beelining tends to give you techs no one else has, at the expense of techs that everyone else has and should be fighting to trade to you at discount prices.

They need to prioritize techs that no one has, and everyone will want to trade for.

They need to be a lot more willing to trade techs to make a profit.

And finally, there needs to be a setting where the AIs are willing to be more opportunistic (like Aggressive AI) without the mindless spamming of outdated units.

 

[futurehermit]

With the AI spending so much on espionage, they should be knowing what other civs are researching so, yes, it begs the question: Why are they researching what they could be trading for

 

[InvisibleStalke]

Great point - knowing what they are researching is a really huge change. So huge its almost a difficulty level gained in itself. I plan my tech path to maximise my trading opportunities knowing what the AIs will have to trade when I finish my own research.

I'm not sure there is an easy fix to this though. Having the AI's deliberately take different research paths without an overall strategic understanding (which an AI will never have) might just open up even more exploits for the player. Right now I can plan my own research to maximize trading, but if all the AIs are researching Banking that limits what I can trade for.

Personally I'd like to see the AI's build less units in peacetime. But have a big push on producing units just before a war or if attacked. Whether they think they have untrustworthy neighbours might lead them to build more units. Empasizing buildings more would raise their tech rate and their espionage rate without trying.

Do the AI ever do anything constructive with their spies? Unless we are talking late term space race, there are only two missions that improve your own civilization significantly - steal tech and incite revolt in a city you are about to capture. The rest just have harrassment value - and you can't win the game by devoting your commerce output into just harrassing one opponent (excepting the late stages of space victories).

 

[Quornix]

Some points I like to mention:

1. The AI is teching to slow on higer levels, no matter how you look at it in BtS. Simply take my save and peek around at the other conti... ....if I remember right some not even have CoL at around 1000 AD. You don´t need to play as England, beeline to rifles or a good strat. to take them out... ...simply kill them with knight and maces if you want (at 1000 AD)

2. Although the aggressive AI might solve some problems (forcing you to have more troops for example), it doesn´t fix the geneal problem:
I don´t want to meet an AI on emperor level or above around 1000 AD running around with swords...

If you're just getting maces at 1000 AD, that's not really a problem, though. And if you're playing with Aggressive AI, you're going to be less likely to have them, as you've spent more on units instead of settlers/buildings, and more on upkeep rather than science.

3. Another point against the aggressive AI is, that I (and maybe some other players) like to play a builderish style. I don´t want to produce troops like hell from the beginning of the game or be engaged in endless wars...

Then play on a lower level where you don't actually need to spend every hammer on troops to keep up with the AIs militarily. Making the AI keep up with you rather than crush you when you're constantly running 80% - 90% on the slider, with a warrior in each city seems ridiculous.

4. I don´t like the AI going MORE troop crazy as it already does. Like uberfish mentioned some posts ago, it really IS like "banging your head against the wall" if you have to kill tons of outdated units before being able to advance into AI territory (not difficult, but time consuming and basically only annoying)

You're looking at the number of troops the AI has on Deity, and then adding even more for Aggressive AI. Believe me, if you were playing with Aggressive AI, you wouldn't be playing on Deity (or maybe you would, because you're just that good, but we have people playing on Noble normally jumping up to Emperor).

5. And, last but not least, I don´t want to play on deity all the time. With the starting bonuses on deity the early game just isn´t much fun and you ALWAYS depend on luck surviving it without beeing attacked (and killed) early. I want an AI that can keep up more or less decent on emperor and immortal, so some things have to be done here

...

Therefore I still favour the solution of giving the AI some boni back

Have you tried Aggressive AI? I don't actually play on those levels (was just getting to Monarch pre-BTS), but I've blown away Emperor without Aggressive AI, and do fairly well on Prince with it. But as per Blake, if you're not playing with Aggressive AI, you're handicapping the AI.

There are definitely changes that need to be made (and from the quick look I took at Solver's list, a lot of them are there), but the Aggressive AI should be the base from which they are made, especially for Monarch+. Find your level there, then worry about boosting the AI.

 

[Quornix]

With the AI spending so much on espionage, they should be knowing what other civs are researching so, yes, it begs the question: Why are they researching what they could be trading for

Yeah. I'd love to see a big improvement in tech choosing by the AI. Active tech decisions by the AI tend to be exclusively worker techs. Especially given their tendency to invest in Espionage, using the information they gain seems like it would boost things significantly.

 

[DaviddesJ]

There are definitely changes that need to be made (and from the quick look I took at Solver's list, a lot of them are there), but the Aggressive AI should be the base from which they are made, especially for Monarch+. Find your level there, then worry about boosting the AI.

But he doesn't want to. I don't question that Aggressive AI makes the game harder, but so do AI research and upgrade bonuses. One is not "better" than the other; it's a matter of taste.

I could make the Deity AIs rush the player right at the beginning of the game, when their military advantage is greatest, and then they would always win, except for isolated starts. But would that make the game better?

 

[FrancisMarion]

Even though in theory the AI should be teching slower on Aggressive AI, I've generally found them to be teching faster because 1. I have to build more troops to keep up and 2. Often times the AI will crush a weak neighbor early and become a powerhouse through the land they get.

I think some of the people complaining about an Easy BTS will be pleasantly surprised by turning on Aggressive AI. I mean I tried playing without aggressive AI for 1 game (because I still can't beat Emperor + Aggressive AI), but I quit in disgust after easily axe rushing my neighbor and gaining a huge tech lead through the land I got.

Like I said before, naming it Aggressive AI has lead to undue prejudice in my mind. Oppurtunistic AI or Comps Who Aren't Your . .. .. .. .. . Anymore AI would be more accurate.

 

[sveint]

But he doesn't want to. I don't question that Aggressive AI makes the game harder, but so do AI research and upgrade bonuses. One is not "better" than the other; it's a matter of taste.

Deity should be a special case - I do think it wouldn't hurt if the AI bonuses were increased on that level. Or perhaps there needs to be an "Insane" or "Impossible" level abovie diety.

 

[Quornix]

But he doesn't want to. I don't question that Aggressive AI makes the game harder, but so do AI research and upgrade bonuses. One is not "better" than the other; it's a matter of taste.

I could make the Deity AIs rush the player right at the beginning of the game, when their military advantage is greatest, and then they would always win, except for isolated starts. But would that make the game better?

And I could make the Deity AIs unable to build military units. Would that make the game better? Aggressive AI is neither of those options. Aggressive AI is an extremely well-crafted option by the person who created the definitive AI mod for Warlords, and was put fully in charge of the AI development for BTS. If he says that the default is a "sandbox" game, and that if he'd had the time to spare, he'd have petitioned to make "Peaceful AI" the option, rather than "Aggressive AI", I'd take his word on it.

And if I'm an Immortal to Deity level player, I'm not going to complain that it's too easy when I'm using the sandbox version. I'd drop down to Emperor or something, and try to work my way back up. Aggressive AI is well balanced, more so than the default.

[EDIT] This does not mean that the default version can not be balanced as well, but XML changes will affect both. But humans will always be better at specializing, beelining, and exploiting weaknesses. The Aggressive AI imposes a penalty (being invaded) for overspecialization, rather than an all-around boost, requiring even more specialization.

 

[gaiko]

The problem is that the AIs play too similarly.

They do research at different speeds to one another, but they all go for the same techs! In a game with trading, this is really, really stupid. If you research a tech that no one else has, you can trade it to everyone else for one tech each. You gain N techs this way on the N other players. If you research something that is already known, it's a huge waste of resources. If more than one other player has it, any of them should be willing to sell it to you for cheap.

The best aspect of the AIs as they are is that they have different personalities. The game would be much worse if that wasn't the case. However, this does not go nearly far enough.

The AIs need to have more varying research paths.

SevenSpirits makes an excellent point. Using espionage I've noticed this as well - the AI's are far more likely to beeline Economics and skip the Liberalism race, meaning that race is almost always a slam-dunk for me. Maybe they "know" they are too ******** in BtS to win it anymore? At least that is how it feels.

I wouldn't say the AIs' don't beeline - besides the Econ example mentioned, they quite ofter do the Constit/Democ -> Commu beeline (for the espionage bldgs no doubt). The problem as mentioned is that the majority of AI's will all do the _same_ beeline, game after game.

To me this is even more of a problem than the question of overall AI teching slowness for exactly the reason SevenSpirits mentions - it precludes inter-AI tech-swapping opportunities. Adjusting this would open up a way to address overall AI tech slowness without introducing AI "cheat" bonuses, including of course indiscriminate tech-whoring without regard to diplomatic relations.

Can't agree though that all the AI's are _always_ behind me in every game: most times 1 or 2 are always on par with me (Monarch) - perhaps I've relaxed my tech pace in adaptation to the overall AI slowness. But in my present game, where I'm trying to focus on teching using SE/bulbing techniques without extraneous elements: no wonders outside Oracle/GL, no espionage slider use, no GAs, no shrines, minimal wars, random leaders, large/Epic - one AI (Hannibal) has _consistently_ had the tech lead over me, at least by 250AD. And I think there is another unknown AI lurking out there that is also doing well with tech.

Now this is one of the _few_ times an AI has gotten ahead of me, but it validates my experience in any case once again: some of the AI's will manage to stay roughly on par with you in almost all games. Now what I'd like to see are those couple of lead AI's be consistently _ahead_ of me at Monarch and the laggards not so far behind. Like I'm used to.

 

[DaviddesJ]

Aggressive AI is an extremely well-crafted option

I have no doubt it's "well crafted". I just don't think making the AIs more aggressive is a good way to make the game harder, for me. It's entirely a matter of taste.

 

[gaiko]

Oh and as for Aggressive AI, I'm with aelf on this. Far from a "solution", it is certain to exacerbate the root cause of any lack of challenge - the perceived slow tech pace - by making the overall AI tech even slower. That, and not the reported general unwillingness of the AI to use those big accumulations of units that is no doubt addressed by Aggressive AI, is the root issue.

If the AI can both adequately tech _and_ keep piles of units just sitting around, more power to it (though we'd probably suspect cheat bonuses).

But competitive unit spam ain't fun.

 

[jihe]

I don't understand why people are against Aggressive AI. I mean aggressive is just a word, in reality I think the option makes the AI play more like a human would play, which is always a great step forward. I mean when would a human player ever sit on his/her ass with a huge stack of swords and axes while another civ is racing to liberalism for rifles and guarding his cities with a couple of archers? Come on people the real issue here is not that the AI techs too slowly, it's because the human player is allowed to neglect his/her military in pursue of tech advantage. I honestly think that aggressive AI is a good thing because that's how humans play on high levels. Of course you can also be a builder if you are surrounded by Gandhi and MM, but no way should someone like Alex let you get away with that.

 

[DaviddesJ]

I honestly think that aggressive AI is a good thing because that's how humans play on high levels.

It's not how I play. I don't build up a big pile of troops and then decide whether to attack with them based on my neighbor's power formula.

But I also don't want the AIs to play like humans. Again, whether that's a good thing or not depends entirely on personal preference. I like the fact that they have diplomatic attitudes, and ways of relating to the human and each other, that can be manipulated, and are part of the game.

 

[Quornix]

Oh and as for Aggressive AI, I'm with aelf on this. Far from a "solution", it is certain to exacerbate the root cause of any lack of challenge - the perceived slow tech pace - by making the overall AI tech even slower. That, and not the reported general unwillingness of the AI to use those big accumulations of units that is no doubt addressed by Aggressive AI, is the root issue.

If the AI can both adequately tech _and_ keep piles of units just sitting around, more power to it (though we'd probably suspect cheat bonuses).

But competitive unit spam ain't fun.

The tech rate only matters as a relative measure, though. Aggressive AI slows you down, rather than speeding up the AI. If you're looking at an absolute date comparison, Warlords AIs are obviously going to tech faster, as they get better bonuses. But let's face it, as absolute date comparisons go, a 1650 space launch is ridiculous.

The unwillingness of the AI to use the big stacks is also a reason it techs so slowly -- unit upkeep. If those units are dying in wars, they aren't being paid for anymore. Aggressive AI doesn't slow down tech rate very much -- unit spam is not the primary hallmark of Aggressive AI. They build more units, but run through a lot more as well. Aggressive AI also makes it more likely for an AI to do the same thing a human does -- actually conquer an enemy, take their land, and use that to speed up their tech/conquest rate.