Aggressive AI = The Real Civ?

[Spearthrower]

Yes but you will tech better on normal settings as well, as with Agg AI you are forced to built more units.

Relatively, I will tech better on AGG:AI... I mean relative to the AI as I wont keep loads of obsolete units around - fewer, more modern units provide a higher power graph... so for less, I will actually look tougher. Meanwhile the AI is driving its economy into the ground with terrible infrastructure (have you ever actually looked in the WB at how poorly it develops now?) and masses of useless troops - little better than free XP for me when it attacks.

He also states that even peacefull AI's will keep larger military so it will be harder to attack early.

Which is my point: I dont attack early - never have and never will unless there's just too good a situation to refuse it. The AI overall in BTS was designed to counter the rusher, it's just more so on AGG:AI - the determined rusher will still achieve it no matter what. Further, the AI now responds far worse to the non-rusher and AGG:AI compounds that.

We all know that that is the most common way humans win. Axe or sword rush take some cities early expand easier. If that is stoped or you pay a heavy price for it it will obviously be more difficult for humans.

See? As I said, people work on that assumption and now they have a more difficult time with BTS - personally, and I have seen others agree, BtS actually made the whole game easier for me as it doesnt counter my style of play well (REX and build). I out-tech and get quality cities while the AI rex's crappy cities and builds no infrastructure, maintains too many troops and allows its tech to bottom out.

My first thought is that these settings make Imperialistic a better trait, as early land grab is more important.

Yeah, it's not a terrible trait - but chopping out settlers largely negates the value of it. I prefer it for the GG's!

 

[Indiansmoke]

Which is my point: I dont attack early - never have and never will unless there's just too good a situation to refuse it. The AI overall in BTS was designed to counter the rusher, it's just more so on AGG:AI - the determined rusher will still achieve it no matter what. Further, the AI now responds far worse to the non-rusher and AGG:AI compounds that.

This is interesting, expanding without war, could be achieved under certain circomstances I guess, but in maps like pangea and difficulties above Prince, it is very difficult to get more than 5 cities before running out of space.
I would think more so in Agg AI when the cities will have to have some defenders.

My experience is that it is difficult to build a tech lead with just 5 cities, maybe Agg AI is different.

The determined rusher can and will achieve it but the cost in teching will be propably much higher than in normal settings.

 

[Spearthrower]

Yes, I totally 100% accept that settings play a huge part in any discussion like this....

Really, that is the beauty of civ - there are so many settings to play with that suit different people.

Personally, I think that AGG:AI is designed for standard settings.... anything larger or longer and it gets horribly bogged down in a vicious circle of expenses, while they player just skims the cream constantly.

I normally play Monarch (or recently Emperor with AGG:AI)/Marathon (or sometimes Epic)/Huge - Map type depends a lot on what I want to do that game, or I leave it on Random.

Playing with AGG:AI on with these settings makes me feel like I am playing Prince or Noble.... except that the AI has a more credible military in the early stages of the game. Personally, I rex like crazy, but I get my economy under control as quickly as possible.

When an AI inevitably declares war (I tend to only keep around mid-way on the power graph because I don't need to stop individual AI's attacking, just not all AI's gang-banging me!) I drag the war on until I take the cities of theirs that I want.... being a builder is a focus, it doesn't mean that I don't take cities by the sword too! I even sometimes declare war - but it's normally later in the game and it's specifically for strategic purposes.

I would say I "rush" a neighbouring civ perhaps 1 out of 20 games (generally due to either having picked an early UU civ or through opportunism - not from a deep-seated strategy) this is largely because in such a huge map, a neighbour's capital is so far away, there are plenty of more affordable city sites nearby.

I can normally achieve 8-10 cities before my economy is squealing and I have to stop and dig myself out. Meanwhile the AI is doing the same but it takes 10 times as long to pull its socks up.

These aren't entirely flaws with AGG:AI.... they are predominantly things that have been introduced with BtS - they are just compounded by AGG:AI on non-standard sizes and speeds.

Before anyone says it, I am not against Blake or his AI - I used to use it exclusively on Warlords, but I think there are a lot of failings in the present incarnation... I do however, fully expect them to be dealt with in due course.

 

[Indiansmoke]

For a challenge then try Normal speed, Pangea, Monarch, Agg AI.

Epic and Marathon are way easier for humans

 

[Spearthrower]

But I dont like normal speed - I find it uninteresting.... anyway, there are a number of things that are easier / harder either way, it's not clean cut.... but I do agree that overall faster speeds are more difficult due to pressures of micromanagement and because battles play more of a role at slower speeds and the human is better at battles.

It's largely irrelevent though, because you just need to compare AGG:AI on/off with the setting you normally use to judge if it's more challenging or not.

 

[Indiansmoke]

Last night I did try my first game with Agg AI. It was fractal normal speed Monarch difficulty.

I (Frederick) was on a continent with 5 others (H. Capac, Toguwama, J. Ceasar, Hatty & Hanibal) 2 others were on a separate continent.

I have to say that the start was not any different for any normal game, but when the map started to fill and borders were getting closer, big, threatening stacks of units made their appearence. I was forced to abandon any thoughts about wonders or peacefull tecking and built a strong military....and I mean strong in all borders not just for one opponnet......quite different from normal games.

I will have to finish the game to get a complete picture, but it sure looks tougher so far...as in normal settings by this time I would be cruising away from the AI.

 

[sagelnarg]

(REX and build).

Sorry, noob question.
What is that "REX" you keep talking about? I wish to learn more about it, it sounds awesome.

 

[EdCase]

REX =Rapid EXpansion.

In previous incarnations of Civ..taken to the extreme,it was referred to as ICS.

 

[vilemerchant]

Sorry, noob question.
What is that "REX" you keep talking about? I wish to learn more about it, it sounds awesome.

It basically refers to the 'expansion' phase of the game when the civs are settling the map.

 

[SevenSpirits]

Since I was curious, I looked through the code for any references to agressive AI. The differences are mostly: it goes for culture strategies less; it goes for attack strategies more; and it thinks it needs more defenders. Here's the complete list:

-Less likely to build culture-focused buildings.
-2/3 as likely to build wonders.
-less likely to go for culture victory.
-Willing to spend more gpt on upkeep.
-keeps 1.5x as many "floating defenders"
-more likely to use "war" strategies: "crush", "dagger", "OWABWNW"
-much more willing to sabotage improvements and destroy buildings in the case where it has no war planned and is above cautious.
-takes attitude less into account when making choices for random events.
-overestimates its own power by a third.
-more likely to declare war.
-more likely to pursue a promising domination possibility.
-willing to poach.
-slight differences in war tactics.

To be honest, I think it really just needs to value good military techs a lot more, and quantity of units (especially specialized defender units) less. Personally I would never build more than one city defender type unit per city, unless I was not planning on attacking anyone ever. Every other unit you build is both an attacker and a defender.

 

[Lance of Llanwy]

I think part of the problem is the AI just seems to not be managing its economy well. Period, Agg AI or Normal AI. It seems to spend a lot on the espionage slider, but fails to invest in the infrastructure needed to make that effective.

 

[sagelnarg]

REX =Rapid EXpansion.

In previous incarnations of Civ..taken to the extreme,it was referred to as ICS.

Ah thank you. Now I know!

 

[peter450]

My big issue with aggressive AI is that it simply builds massive stacks, on longer games epic /marathon the stacks can get so big units get listed all the way to the top of the screen an beyond to who knows were.

Rather than warfare being about carefully chosen promotions an well positioned stacks tactics (these are still important, but less so in my opinion, due to the massive troops number involved) it seems more about huge numbers of troops, i do like some aspects of agg AI, i like the fact the AI is more likely to go for domination type wins, however with the new patch looking like it will fix the main area of weakness for normal AI (the fact it does'nt build a strong enough military initially to stop itself being overun) it looks like agg AI will be somthing i leave off in future, it's a shame they cant incoporate the best parts of both AI's into the standard AI, which plays a well rounded game, agg civ's in standard AI games should behave more like they do in Agg AI mode (although perhaps not going quite as overboard on stack sizes & military spend that they do in agg AI which can be at a level that seems excessive to say the least), this would mean a variety of civ's each pursuing differant goals to victory, an more differance's in the persona's of aggressive an peaceful type civ's