Chess program designed to learn from player interaction. Made with flash and PHP server technology, the finished version will (in theory) be able to beat all players that aren't as good as players that have played the system often.
George Jempty says: hey Frank: go back to watching American Idol 3:10pm 27 Jan 07
Frank says: "A chess genius is a human being who focuses vast, little-understood mental gifts and labors on an ultimately trivial human enterprise." - George Steiner 12:10am 19 Jan 07
Christo says: The game now uses a database to log all moves for all game types. The popularity for a move is compounded whenever the same move is performed for a certain board layout. If a move doesn't exist, the chess client picks a move for the AI player. This allows for rudimentary training of the AI by human players. If playing the computer, press the 'take back move' button if the AI move is a bad one. Playing alternatively as white and black against the AI will also increase the training. 4:00pm 29 Nov 06
Christo says: Loads of bugs at this stage. Currently the available moves for the AI is calculated and one picked out a hat, so it's very dumb. The aim is to create a pseudo type B chess computer that uses previous game play data to create responses. This will mean it will need a server database (currently it's a standalone swf) 4:19pm 12 Nov 06
How do I castle dude?
12:38pm 10 Mar 07
hey Frank: go back to watching American Idol
3:10pm 27 Jan 07
"A chess genius is a human being who focuses vast, little-understood mental gifts and labors on an ultimately trivial human enterprise." - George Steiner
12:10am 19 Jan 07
The game now uses a database to log all moves for all game types. The popularity for a move is compounded whenever the same move is performed for a certain board layout. If a move doesn't exist, the chess client picks a move for the AI player. This allows for rudimentary training of the AI by human players. If playing the computer, press the 'take back move' button if the AI move is a bad one. Playing alternatively as white and black against the AI will also increase the training.
4:00pm 29 Nov 06
Loads of bugs at this stage. Currently the available moves for the AI is calculated and one picked out a hat, so it's very dumb. The aim is to create a pseudo type B chess computer that uses previous game play data to create responses. This will mean it will need a server database (currently it's a standalone swf)
4:19pm 12 Nov 06