**Very Low on Priority List May be Revisited In The Future**
well the name give it all away i think some AI players would be good so that if you cant get enough players together for a realtime or just a normal game you could use a computer player to fill the space
would this be hard to programme?
just a thought anyway
i think it definately would be hard to program. the ai would have to be programmed so it can
conquer continents
trade cards
reinforce sensible places to reinforce
and basically think like a human, and then we would get the compliants "the dice favours the ai" blah blah blah
Freaking impossible! I bet it'd be anyway. Sort of a masters level thing, or a PhD dissertation.
Well... basically you could play it with statistics. You have easy statistics for the chance of winning a roll, and you could program in statistics for... say, someone has 30 dudes parked on your border, and he keeps plunking down more every turn, statistically he's threatening you. But then you'd get into questions of quality, judgement calls... you'd quickly need to program a personality in. If I had to make the call between Europe and Africa, and it were an equal call, I'd go for Europe every time, but that's only because it's position is superior to the deadend of Africa. If no one was in S. America, I might go for Africa... You'd need to build a statistical model of the entire board. I guess it would be like chess computers, who work from an analysis of thousands of historical games, just like humans do.
Hmm...
Of course, remember, that it took bazillions of dollars and years of research to build a computer that could consistently defeat a player in chess. Barring some secret resources of lackattack, what you're saying here is, let's create an idiot, someone that will play predictibly and stupidly, and then assign points for defeating them.
I've given this enough patient thought to dismiss it as ludicrous without sounding abrupt, haven't I?
A certain website has modifiable ai.. you choose their agressiveness, and they work fine... I'm not posting a link dammit.. I dont think ti was too hard to program though,.
One of the best part about a world domination game like this...is that you get to play against real people. With real, unique, and sometimes dumb strategies. It's wonderful and horrible all wrapped up in one, and I wouldn't have it any other way.
AndyDufresne wrote:One of the best part about a world domination game like this...is that you get to play against real people. With real, unique, and sometimes dumb strategies. It's wonderful and horrible all wrapped up in one, and I wouldn't have it any other way.
AndyDufresne wrote:One of the best part about a world domination game like this...is that you get to play against real people. With real, unique, and sometimes dumb strategies. It's wonderful and horrible all wrapped up in one, and I wouldn't have it any other way.
--Andy
I agree but if they did incorperate this, no points for AI games. I can see some unscrupulous people abusing it for stats. (AI=oxymoron)
Silence is golden & Duct tape is silver.
Subsiste Sermonem Statim
I will probably never have time to program AI players, but some day I could fathom people having bot accounts.
I suppose you could already write the logic in JavaScript and make a Grease Monkey script to take your moves. Or write a stand-alone program to automatically login and play. It is not against the rules or anything.
I would want an AI that learns and can change strategy on AoE it took me ages to write a script that considered everything it had everything like if 20 archers in castle blah blah blah
If regicide targer = king etc.
It takes ages, This was just for 1 of them
It isn't worth the time It took me 120 hours overall according to the program and I was planning on making 100 of them it would take 5 days no sleep to complete it and Lack would also have to make something to read the program
Just isn't worth it
‹max is gr8› so you're a tee-total healthy-eating sex-addict?
‹New_rules› Everyone has some bad habits
(4th Jan 2010)
I don't think AI would be hard to program at all. Although I think it would need a slightly different script depending on the map, and it would take a team to keep it up to date with the speed of the foundry lately.
But mostly it would run off of statistics logic with the dice. With maybe a low random chance of attacking when the odds aren't in the AI's favor.
You could even go so far as to have different types of AI that it could randomly decide to be. Turtles, Breakers, and what not. (Turtles would be taking one continent and then forting until they can steamroll everyone) (Breakers would be always stopping other people from getting continents when it can even if it is bad for them)
Team AI would be really hard though. I don't know how that would work at all... And it is a very high level thing. I might be willing to take it on in three or four years when I am better at programming, lol.
Well, then, I guess some of us should put together some AI code, host it on their machines at home, and have that AI log in through its own account -- these accounts will be declared by their owners in a special AI Forum. BTW, Happy Easter!
Coleman wrote:I don't think AI would be hard to program at all. Although I think it would need a slightly different script depending on the map, and it would take a team to keep it up to date with the speed of the foundry lately.
Ideally you wouldn't make it map-specific, you would give it various heuristics it would use to determine the best move based on the borders it has and so forth. If you did it right it'd play well on all the maps.
You know guys. Have you ever played Lux ? They have been using AI players for years, maybe someone could talk to them about how to program them for conquer club.
