Tic Tac Toe - AI (efficient algorithm?)

[Roland_Yonaba]

But it's a good feel, that I could come back with more questions.
(I'm sure.)

No problem.

Yeah, exactly how I meant it. This would be possible with a few of variables and values. Sounds easy.

Well, I'm not that sure. But at first, try to focus on writing the game itself. Once done, you can come back to the network support.

[Petunien]

According to my knowledge, network works basically like:

1) Player X choose a field. The new values (selected field and maybe other values like gametime) will be sent.
2) Player Y receive the new values, the program works accordingly and waits for Player Y's selection.

And so on, or?

Basically...

[Xgoff]

Check for Minimax algorithm, it can be used for AI in a Tic Tac Toe game.
But it'll need some skills to be implemented in Lua, though.

minimax can't be that hard to implement in lua, considering there's an implementation of it on that wiki page

[Robin]

So to add to what Roland said:

2) Robin, you wrote it in Python, is it difficult to "translate" it bzw. write it in Lua/LÖVE?

It is not really difficult, but it will take certainly more effort than just writing it again for your specific purpose.

Minimax isn't complicated: you either choose the highest value or the lowest value, and that for every option you have.

3) Are all 255,168 (Wikipedia) game-situations covered with Minimax?

Yes. Minimax finds all possible games and evaluates them. That's why you shouldn't use Minimax for chess, which has something like 10^70 possible games...

4) How does it work with network?
Are there "normal" values like numbers, true/false and so on, that can be sent and further processed to/by another PC as well?

Not sure what you mean here, Roland probably wrote a better answer than I would write here.

[Jasoco]

Man, I wish I still had the code for the TTT game I built in Visual Basic way back in the day. I was extremely proud of the AI system I built in. I had three skills. Easy, which just picked random blocks. Hard, which actually used the most logical move to block your move. And Medium which was a hybrid and randomly chose between random and logical. I can't remember how I made it so smart. I was new to programming so for me back then figuring out how to do it by myself was pretty big. So sad that back then we didn't have the luxury of cheap USB backup hard drives or Internet cloud-based file hosting sites. I lost so much cool stuff I wish I had at least copied to a floppy disk or something.

[Adamantos]

Here is the optimal solution for this game
http://xkcd.com/832/

[Roland_Yonaba]

Here is the optimal solution for this game
http://xkcd.com/832/

Lol.
I had a clue when I first saw "xkcd" in the url...

EDIT:
Another algorithm that can be used for solving a Tic-Tac-Toe is Backtracking.
Useful resources can be found on SiteExperts, csStanford and Duke.

[Petunien]

I'm trying and reading a lot of stuff. Only so you don't think I gave up.
But many of the examples are written in Java, so I have to understand the basic of that language before I can understand anything.
Bad.

@Roland
Thanks for the other links, I saw them just now.
It looks easier than Minimax, at least in my opinion.