Author: Inmann Werner
Date: 04:41:39 10/27/99
Go up one level in this thread
On October 27, 1999 at 03:58:07, James Robertson wrote: >Nicolas, > >When I first started writing my chess program, I had never read any books, had >no internet connection, no nothing. I had no programming training, no people to >guide me; basically, all I had was the QBasic interpreter that came with MSDOS >and my own brain. So, my first program searched a tree of exactly W^D. > >After I wrote my program (it could search two plies in 3 minutes on my 486, and >'handled' captures by just extending a ply if the last move was a capture), I >found a book by (I think) Monty Newborn. It contained the first computer chess >stuff I ever read, and I remember sitting on my couch working out trees with a >pencil and paper for about three hours trying to figure out why the 'alpha-beta' >algorithm worked. Eventually, it did work, and my program searched 3 plies in >three minutes. :) > >A few months later I bought myself a C++ compiler, and learned C++ from scratch. >Another while later my family got an internet connection and a whole new chess >world opened up, one that I have been thrilled with and wouldn't give up for >anything. > >The whole point of this is, I took a chess program step by step; I would >encourage you to do the same. Write a search algorithm (maybe not even >recursive; my first wasn't) that just searches W^D trees. If you can do it >without looking at anybody else's code, all the better. Chess programming is >like math; you have to 'do all the homework problems', and you're guaranteed a >good grade. Add the alpha-beta algorithm next. I added it to my QBasic program >before I had seen the source to any other program, so it was a completely >original implementation; while clumsy, it ensured I understood what was >happening completely. It took me days of work and about 50 emails to Tom >Kerrigan to figure out how a quiescence search works. But since I did take the >time to learn how it works, I can debug my code when something goes wrong. :) > >Good luck, and keep at it! Aren't you 15? I was 15 when I first started chess >programming. (2 years ago!! Man, time flies.) > >James :-))))))) I was a "bit" older when I started. My first prog did not work, cause i did not get the quiet search idea :-( But I got a book explaining some things, and whow, it calculated moves!!! You can take a little help, no problem, but you must "be in control of your code", or you will have no chance to improve. You have to understand each line of your prog (maybe except of the TB Code of Eugene :-) ), or you will "fail low". You must understand, what a pv is, where it comes from (you know it in the end of search!) and then the program lines come automatic (and with the needed bugs) I sometimes looked at Crafty Code (as Bob says) and sometimes understood what happend. But if I did not understand it, I did not use it that time. Werner P.S: James, your prog got rather strong. But watch out, I will be back :-) (Arnold Schwarzenegger?) Best wishes for Leiden, hope you get a good score!
This page took 0 seconds to execute
Last modified: Thu, 15 Apr 21 08:11:13 -0700
Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.