My chess application on X10

[zen-r]

I like your idea of logging in, playing against someone else, & saving. This would also be a much easier facility for you to program in, since the computer wouldn't have to work out any moves itself.

One stage on (& when your program is much cleverer) would be to then ask the human to decide if he had just played against another human, or the computer!

When developing the software to calculate moves, don't forget that not all moves need to be calculated. Certainly in the early stages of the game, a lot of set moves could be programmed in from classic game strategies. So the computer could, for example, go for a classic Karpov or Kasparov opening, playing as many moves as it can get away with until its competitor forces a deviation. And the computer could always be on the look-out for known, pre-programmed positions for which it already knows the winning moves

(Edit: I've just added an extra thought to my last post here)

[fguy64]

I like your idea of logging in, playing against someone else, & saving. This would also be a much easier facility for you to program in, since the computer wouldn't have to work out any moves itself.

One stage on (& when your program is much cleverer) would be to then ask the human to decide if he had just played against another human, or the computer!

When developing the software to calculate moves, don't forget that not all moves need to be calculated. Certainly in the early stages of the game, a lot of set moves could be programmed in from classic game strategies. So the computer could, for example, go for a classic Karpov or Kasparov opening, playing as many moves as it can get away with until its competitor forces a deviation. And the computer could always be on the look-out for known, pre-programmed positions for which it already knows the winning moves

(Edit: I've just added an extra thought to my last post here)

true enough about pre-programming standard opening moves to save on calculation time. However, I am of the opinion that if a person does not really understand the position that results from a particular opening sequence, then it is not really productive to spend time learning the sequence. It's a common mistake among humans to learn certain opening sequences without having any real understanding of the resulting positions. I think the same applies to chess programs, at least to this chess program. When I do manage to give my program the ability to choose a good opening position over a bad position, then I'll start dealing with pre-defined openings.

Problem is, there are just too many different ways to play the opening moves well.

I'll always remember the following passage from a chess book I read many years ago, loosely quoted it said "...there are more ways to play the first 12 moves of a game of chess than there are raindrops that would fall on greater Chicago in over a decade." Of course not all of those ways are good, but you get the idea.

All this being said, no doubt it will be a useful learning experience to set up a small opening database.

[ichwar]

hmmmm, as soon as I've got some time, I shall try and beat it on lvl 3. I think I could do that since I can beat the shredder engine on lvl 1.

[fguy64]

hmmmm, as soon as I've got some time, I shall try and beat it on lvl 3. I think I could do that since I can beat the shredder engine on lvl 1.

It should be a piece of cake. It has zero concept of strategy. I'd like to make it better, but that requires some advanced recursive algorithms that at this point just make my head spin. There won't be an significant improvements until late in the year.