Thanks fogmiester I will give this a shot in the labs tomorrow, much appreciated.

Your a legend

Cheers

Kal

Don't be surprised if another "omg Java is such a pile of wank" thread appears relatively soon.

I thought me and Java were finished, but I've been given a chess game to make in the dreaded language.

I have to make it so that the one app can have two players playin each other over the network or one player against the app itself.

Fun times.

First step: learn how to play chess!

Setting up the rules and establishing a 2 player game shouldn't be too hard.

But do you have to write your own AI algorithms for the 1 player game?!

Wouldn't that be a monumental task?

Thats the kind of thing that people in universities spend decades studying and improving.

For our equivalent we only had to provide trivial AI, that could make a possible move, or we could use known algorithms.

It would make sense to download a chess AI API. That would make it a lot easier anyway.

Or just randomly move any available pieces onto any available random squares.

That's a good question actually.

I've read the brief, and there is little mention of it except for "A computer player is a requirement of this assignment".

I will post in our department's forum to ask my lecturer.

We won't be expected to produce a miracle machine capable of defeating the grand chess master, that's for sure!

Sorry for semi-hijacking this thread kalisclark, I'm finished now! And good luck with your assignment!

Let me know what the result is regarding the computer player. I might join you in this project and see if I can create something also

For something of that complexity.... not really

You make a fuzzy solution such as a neural net for machine learning, then set it off a load of times with AI vs mutated AI, whichever wins is deemed better

it's a way to do it, but you might not get the truely best AI if the mutations aren't extreme enough to start with. It's possibly more time consuming depending on how quickly the game can run... but its a more realistic solution

Nick: http://www.chessbin.com/

You can rewrite the stuff there into Java.

Don't ask why but I was thinking about something like this a few days ago.

All you needs is the rules. Then you set off the computer playing itself and each time store the moves and the outcome of the game. Then play it hundreds of thousands of times and each time weight the moves in favour of the ones with the best outcomes (somehow).

I don't know how you would do it in practice though.

Cheers for that link fin.

It turns out that we do have to code the computer player, but only very simply. It has to be defensive, and avoid being taken. If more than one piece is in danger, it has to save the highest ranked piece.

It doesn't have to attempt to win in anyway.

So pretty straight forward.