Re: Mancala solved!

[Chris Nash <chris_nash@prodigy.net>]

Thanks Bernhard for this...

> His set of rules is slightly different from the set of rules used for
> "wari" on Pbem.
> Firstly: Here a capture, which empties all oponent houses is allowed and
> ends the game (as the oponent doesnt have any moves left).
> Secondly: Not in all "deadlocked" cases the remaining beans are split
> evenly. There are things like a Two-One-split. There are no half beans.

Yes, I did notice a few things about their method that were confusing - mainly it appears that the
"value" of the game for their analysis depends only on the remaining position, not the current
score (it is a very different situation if the score is 24-16 than 20-20 with the same stones
remaining on this server).
 
> Its quite a nice effort to have solved the game, but its also sad (in my
> opinion). It will be less exciting to play if a perfect player exists.

I agree that it is very sad. It is also a "brute force" approach to list all the 800,000,000,000
possible positions, we have learnt very little except that it is possible to explore such a game
tree with enough powerful computers. It is good to know that there is not a certain win for player
1 or player 2, but it is not good to know a perfect player exists. I hope it does not destroy the
game.

- Chris