EC3 New Idea #1 - OPEN SOURCE AI

[raingoon]

"Open source AI." -- DaveV

[Travathian]

Amen!

I second this idea.

[korn469]

in order to start some discussion...

davev

if asked why out of the five things to put on the new ideas thread why would this idea belong? what are the greatest strength in adding this idea? and what if any weaknesses or exploits does this idea have?

also when you say open source AI do you mean that firaxis should develop the AI using the open source development model like linux, or do you feel that they should just include some kind of deep AI scripting program, when the users get to define how they want the AI to act?

[apocalyptica]

easy. post about it on /. and get those guys involved in the development. you'll have an AI that you can't beat in three months.

[TheTron]

reasons for open source AI
1 10,000 heads are better than 1.
2 ability to change and improve the AI after release as new tactics evolve.
3 ability to use different versions of AI for each game and hopefully each different AI for each nation in each game.
4 programmer contests, each AI plays a different nation and all nations are computer controlled and game plays itself at hi speed, may the best AI win.
5 fewer complaints about the AI, if you think you can do it better, prove it.
6 Smarter or quicker AI for faster and slower machines.
7 In online games, a custom AI could play for you and let you take a break, and the AI would actually play like you.

by the way, this would be a great idea.

[DaveV]

Final summary:

Despite the low number of posts, I still think this is a great idea, and would do a lot to improve the longevity and enjoyability of Civ III. No matter how well the "out of the box" AI is written, it will be limited in its adaptability. Many people on this site say that they don't play single-player Civ II any more because it's too easy. After years of discussion, the tactics for beating the AI are well known. Imagine instead that the user community (proven in the past to have great resources of talent, intiative, and altruism) could update and improve the AI. Such an improved AI could not only enhance the single-player experience, but remove much of the micromanagement burden by *competently* automating some of the decision making.

Firaxis can best decide on the method of implementation - anything from a completely open source code model, to a limited-access dll file, to a text file containing situation/action rules. The greater the amount of AI behavior open to modification, the greater the possiblity of improvement.

Ideally, the game could include multiple AIs, operating from different rulebases; then the player would not be able to predict the AIs' behavior so well. AI-only games could allow programmers to compare their AI algorithms.

Biggest drawbacks of this idea: Firaxis spends a lot of programmer time on a feature which many of the end users will never use. Sloppy coding by end users could cause strange game behavior which would be reported back to Firaxis as a bug.

[The Joker]

If this idea could be implemented well it could be one of the things that would truly make Civ3 the best game ever.

Absolutely great idea.

[The Mad Viking]

This is the single most important improvement possible. Civ2 AI stands for Artificial Idiocy. I don't think more than a day could have been spent programming the AI in Civ2. When you think that Deep Blue can beat the world's greatest chess player, and that a cheap chess program can beat a really good rec player, the AI really stinks.

I guess they could make a brilliant, out of the box AI without it being open source... but I sort of think that the act of disclosing the algorithms will make the programmers to try harder to make a brilliant AI engine.

That this lets you automate your own micromanagement style could be a big bonus.

I would add that special anti-cheat precautions might be advisable, to avoid both tampering and spying on the other civs AI.
<font size=1 face=Arial color=444444>[This message has been edited by The Mad Viking (edited March 08, 2000).]</font>