metax wrote:Don wrote:Hey Peter,
When are you going to start your own chess program? I think you should go ahead and try. My first program was horrible and so will yours, but that is ok - nothing prevents you from improving it as you go or scrapping it later for a rewrite.
Start with the most basic thing and go from there. Don't copy some public domain code but figure out the data structure for yourself - copying code is like using a calculator instead of first learning how to add and subtract for yourself.
Do you do any programming? If not, now is a good time to learn. Mabye by the time your are 15 or so you will have a very strong program if you start learning C now!
Don
I am about to lose my title as youngest chess programmer here in the forum...
When I was 15 or 16 there were no home computers. I did however "write" a paper chess program, a list of instructions that if followed correctly led to a move (with the aid of a coin flip.) The instructions basically tried to post pieces to the center and avoid blunders. I think the first "instruction" was to deliver (any) checkmate (mate in one only) if one is possible - so it did require a little bit of human skill to interpret this without error - a lot of very weak players cannot reliably find checkmating moves.
The instructions did not require sophisticated look ahead, but I think everything tactical was based on knowing what was up-attacked, and if something was defended - I did not try to take it very far because it would be increasing difficult to "execute" correctly. So it would not move into a situation where a lower valued piece could capture it or leave anything hanging outright - although it could easily be defeated with simple exchanges. Most of move logic involved avoiding blunders primarily and secondarily trying to center pieces. If I'm am remember correctly the "positional logic" was to try to center pieces in a specific order, the d and e pawns first, then knights, then bishops, then rooks and queens. So if you could play e2 to e4 or d2 to d4 that was best, then e2e3 or d2d3 etc. After that, knight moves that centered the knight more (or the most.) I had tie break rules to cover moves that were equal, but in general you played the first move that applied.
I wanted it to not be too tedious to execute and once you got the hang of it you could calculate a move pretty quickly. And it was surprisingly good when compared to the average high school kid. I could be remembering a lot of this wrong, but those were the general principles. The skill in this has more to do with how to make this simple and non-tedious. It could have been made significantly stronger with more rules, but it was not feasibly to have to spend 2 or 3 minutes figuring out each move.
For this to be fun you must be able to execute the move in a few seconds with practice.
I have often considered making up such a program, memorizing the instructions and then going to local chess tournaments with this in your head to get a rating for it! And you could improve it little by little based on your experience with it, the primary factor being that it cannot be too complicated to be unreliable.
Of course I could not do this now, as I am too strong and it would ruin my reputation as a player - not that I'm that good, but I would still be a few hundred ELO stronger than this and some of the moves produced would look so alien that even weak players would flinch.
Don
