Author: Tord Romstad
Date: 09:18:41 02/15/04
Go up one level in this thread
On February 15, 2004 at 02:42:42, Dann Corbit wrote: >On February 14, 2004 at 04:48:43, Tord Romstad wrote: >[snip] >>>The most important thing will be to write good algorithms. But you have to have >>>enough knowledge added so that the program won't play like an idiot. If you >>>tell your program to do nothing but count wood, 19 plies won't be enough to beat >>>anyone. >> >>You don't need to write any algorithms at all to create a reasonably strong >>chess >>program. The published algorithms are sufficient to get quite far. It is >>correct >>that you need some chess knowledge, but I think knowledge on the level of a >>1500 rated player is enough. >> >>All you need is to read a few basic chess books and to implement the known >>algorithms. Nothing more than ordinary chess and programming skills are >>required. But it takes lots of work, of course. > >Yes, I meant to "implement" rather than invent the algorithms. But one person's >alpha-beta can be four times faster than another person's. That's enough to >make a huge difference. It is, but achieving this factor of four requires nothing except lots of patience and tuning. Being a good programmer might help you get there quicker, but it is by no means necessary. The mistake almost all beginning chess programmers do, of course, is that they spend countless hours trying to make their N/s count a few percent higher, rather than reducing the size of their search trees. Tord
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.