The game is just TOO easy!!!!

[ironanvil1]

TW should go the same wasy as Galactic Civilisations 2. That game has so many difficulty settings it's hard to remember them all, I think it has ten or more. They range from pathetically easy all the way up to hellishly difficult.

GalCiv 2 is a good way ti go, in that the higher difficulties actually made the AI smarter rather than just adding additional resources to a moron.

It also has the various choices of what you want included in each game.

I don't want an AI that's necessarily more aggressive, just one that will surprise you on occasion with sneakiness rather than just spamming elite stacks. Stuff like cutting you off from your allies, building up it's own coalition and isolating you diplomatically before invading, that kind of thing.

[zarkis]

There are several factors which are contributing to the overall weak campaign AI performance.

(1) Inability to launch naval invasions or transport troops between hotspots.
(2) Wasting their units in endless small scale raids.
(3) AI countries are locked up in endless wars between each other without any progress, which drains their economy. That may be the reason why many factions have a worse financial rating later in the game then in the beginning (exeptions are Spain, England, Holland and France)
(4) Insufficient recruiting behavior, which seems random and not controlled by any plan.
(5) Sometimes strange attack moves, where the AI attacks a stronger force with a much weaker force.
(6) Inability to retreat on the campaign map when facing impossible odds (a feature even RTW had).

So far, I have no idication that the campaign AI was improved or worked on. Maybe they run out of time, the whole stuff is unfinished, and we will get a better AI opponent with patches. Maybe CA failed to acknowledge that this area needs some serious work. However, from the interviews they did, I had the impression, that they were working not only on the battle AI (which is not without flaws but quite improved over RTW/MTW2 in my opinion) but were designing a new campaign AI too. For example they talked about the campaign AI telling the battle AI what to do in certain situations and implementing a more goal based system overall. So far, I can't see any of that. It's a pity because ETW has many neat new design features, great graphics and the battles are quite fun.

[Rufus]

I think the person who brought up the Prestige campaign raises a good point. If you're going for the Prestige victory in addition to your normal victory conditions, you have to research a lot of the technologies and philosophies that might give rise to a revolution within your own home territory. This doesn't necessarily negate the problems others have identified, but it could make the grand campaign harder.

[Alexander the Pretty Good]

Heh, at least ETW got JAG to post.

Now, what to do if ETW is too easy but Galciv is boring because it doesn't have tactical battles...

[Spino]

GalCiv 2 is a good way ti go, in that the higher difficulties actually made the AI smarter rather than just adding additional resources to a moron.

It also has the various choices of what you want included in each game.

I don't want an AI that's necessarily more aggressive, just one that will surprise you on occasion with sneakiness rather than just spamming elite stacks. Stuff like cutting you off from your allies, building up it's own coalition and isolating you diplomatically before invading, that kind of thing.

Well most people who cry for a better AI opponent simply want one that 1) behaves sensibly and 2) behaves more like a human than a jacked up ENIAC on overdrive. Understandably 1) is far easier to achieve than 2) but the fact that some developers can do both leaves us wondering why the rest don't make a bigger effort to do so as well.

The lackluster AI in E:TW and other strategy games only illustrates the massive differences in design approach between most developers and Stardock. When Brad Wardell & company set down to design the Galactic Civilizations series they deliberately set out to create an extremely effective AI opponent for the single player game. A strong AI was considered one of their top priorities and they applied an extraordinary amount of intellect and resources in order to achieve their goal. As many people have mentioned, this effort SHOWS. The fact that Brad Wardell cut his teeth as an AI programmer didn't hurt either...

The folks at Stardock knew that if they provided a challenging AI that did not need to resort to cheating in order to make a good show of it would go over well with gamers and they were right!

[Dogfish]

When Brad Wardell & company set down to design the Galactic Civilizations series they deliberately set out to create an extremely effective AI opponent for the single player game. A strong AI was considered one of their top priorities and they applied an extraordinary amount of intellect and resources in order to achieve their goal. As many people have mentioned, this effort SHOWS. The fact that Brad Wardell cut his teeth as an AI programmer didn't hurt either...

The folks at Stardock knew that if they provided a challenging AI that did not need to resort to cheating in order to make a good show of it would go over well with gamers and they were right!

I read that and can't help but think back to reading an interview where somebody at CA was bragging about how they spent a couple of years getting the water to look right.



Yes, CA, the screenshots are very pretty. Now please make a decent AI and the game will be in the upper echelon as it was expected and hoped to be.