PDA

View Full Version : othello AI



laasunde
10-24-2003, 07:11 AM
Afternoon folks.

I was recently given a project at Uni to make a othello game. Must admit that I find the game rather boring and I am far from a master of it. Obviously it would help a lot when making the AI knowing the in's and out's of the game.

What I do know is that I need to use alpha-beta pruning to find the best next move for the computer.

My biggest problem is assigning each board node a value that reflects how good or bad it's state is. Examples :
- A very simple approce would be just to count how many position you have and subtract that with the amount of position you opponent has.
- Give each position a value, for instance a corner is more worth than a position in the middle of the board. But the thing is, the value of a position might change during the cours of a game. Sounds difficult to set a static value for a position that at all times should accuratly describes the stat of a board.
- Always try to minimize the amount of possible moves your opponent has and try to maximize the amount of moves you (the computer has).

I was also thinking that I might need to divide the game into two or three stages : start, middle and finish and have a different strategy for each stage. Or maybe try to combine two of the strategy's above.

I'd appreciate any input on this matter.

PJYelton
10-24-2003, 08:50 AM
I personally haven't done an othello game but I have done others with alpha beta pruning and have seen programs that played othello. All of them had a constant value for the squares that never changed, very high for the corners, kind of high for the sides (except for the side squares right next to the corners, these were worth negative points if I remember), etc. I would recommend writing it with constant values that don't change, and if you feel the AI isn't playing very well then try tweaking it to changing values. But I think you'll be surprised how well it plays without it.

As far as the exact values you should make each square, google around. There are a lot of sites that talk about AI for chess, go, othello, etc.

laasunde
11-02-2003, 02:21 PM
Originally posted by PJYelton
As far as the exact values you should make each square, google around. There are a lot of sites that talk about AI for chess, go, othello, etc.

Been searching like crazy for a site that gives exact values for each position but havnt found one. Would appreciate any help.

Cheers :)

confuted
11-02-2003, 02:37 PM
http://www.ai-junkie.com/ai-junkie.html There are forums there...