Re: gnomine solver

[David Bellot <bellot stat berkeley edu>]

> In principle we could replace the current hint function with one of
> these algorithms, but keep in mind that the current hint function is not
> designed to give the best move, it is designed to reveal a square that
> gives a lot of information to the player. Having said that, it doesn't
> give the most possible information to the player.
>  
Why don't you turn this function into a 2-player gnomine. The two 
players (one is a human, the other is a computer) starts with different 
games. The window is splitted into two parts, one part with the board 
for each player. The computer is slowed down in order not to win too 
quickly and the winner is the first to solve the game. As the human 
player can see the progress of the computer, his/her stress increases 
during the game.

Another kind of game would be to play on the same board against the 
computer.
At each turn, a player chooses a square. Depending on the number of 
squares which are erased, the player can play one or more turn after. If 
at the next try, the player discover less squares than the previous try, 
then his/her number of tries is decreased so that to avoid a cumulating 
effect. Then you must be lucky to continue otherwise it's computer's 
turn and so on...

David