First of all, I have to say that I am very impressed what I've seen here - especially if you're considering how 'small' this team actually is.

Have the deveopers already an idea about the AI? Will it be fully dynamic (would be cool) or is it more scripted...perhaps even the AI could be somehow changed by the user (mission editor or via scripting)

Thanks.

I presume you're talking about the campaign engine AI? I think it'll be fairly simple deterministic AI that will choose it's advance vector based upon some simple factors. It will consider it's intelligence for the next node and decided whether it has the forces to attack or whether it will wait for reinforcements.

So essentially it will be a dynamic campaign with an 'AI' that makes decisions based upon factors.

Cheers

I'm 5 foot 10 inches. I'm not small.

I'm borrowing some AI from a wargame I wrote in the 80s, it's as good today as it was back then. Don't want to talk about how it works in detail, some things are better left to wonder at. If Google told you how pages were ranked it would ruin the service. Suffice to say it should (in theory) not always make the obvious move. If you smack it at the same point it should take the hint and given enough resources apply pressure elsewhere.

I'll be testing it as a stand-alone program before it goes into the game. Easier to speed test the mechanics and not cause problems to/or get sidetracked by other issues in the game.

Ok, I herby confirm that your team is not small


Actually I was not referring to the campaign AI (although it is nice that you give some input here).

I was thinking of the tactic AI: For example: Will it be capable to do shoot&hide manoevers (eg. insurgents at the rooftop with a SA-7)...will tanks try to conduct defensive manoevers like creating smog, split their formation and that stuff?


[/quote]

Some tanks will pop smoke, US tanks will attempt to pop smoke but instead shower themselves with beer

Some units will have an evasion function, so long as it's not silly and works in co-op. Can't say more than that till we have it working. If we have processing power to have behaviors for autonomous characters interacting with compounds and other structures then maybe. But we're an Apache Combat Sim first and foremost.

My personal take based on a case study of a AAA game:

What's likely to happen, if we spend an additional month or two developing AI behaviours for different kinds of formations, a player is probably going to smack them all down in seconds with a full Hellfire loadout before the AI has had a chance to twitch thus rendering the exercise and additional development time pointless.

For example, Halo(2 I think it was) had great AI, but hardly anyone ever got to see it and consider the AI was terrible. The reason was simply that players are overpowered and often killed the AI before they are able to exhibit any of that cool flanking and communication behaviour.

So doing things like popping smoke, formation changes and calling for help, even ducking and rolling are things that are easy to do. Running around looking for hiding spots requires dedicated pathfinding for different unit types.

Some units will get special treatment but it all eats time. Just have to see what works. It's a good question though and I'll certainly worry about it.

Thats sounds good Flexman.
Another simple idea would be to let activate some kind of retreat waypoint when the unit strength is low.
I think even some evasive movements like you describe would make the game much more interesting compared to EECH (just sitting around a road doing nothing) especially when you fly low.

Idea noted. That could work too. Cheers.