Author: Stuart Cracraft
Date: 15:56:54 09/19/04
Go up one level in this thread
On September 19, 2004 at 18:33:42, martin fierz wrote: >On September 19, 2004 at 14:44:13, Uri Blass wrote: > >>On September 19, 2004 at 14:03:52, martin fierz wrote: > >[snip] > >>>areas of improvement. but i think that from now on, progress is not so simple >>>any more. >> >>I think that progress is still simple and it is possible to improve your program >>by 50 elo only by changing less than 20 lines in the code(the problem is only to >>think about the right ideas). > >i think that is just the point: up to a certain point, you can make very fast >progress by implementing all the standard ideas. you start out with an >alpha-beta searcher with a simple evaluation. then you add a hashtable, you do >all kinds of tricks to improve move ordering, you add a qsearch, you add a SEE, >you improve your eval for king safety and with everything you do you gain lots >of rating points. but from then on, it is not so easy any more... > >cheers > martin I agree and disagree. I think there are a bag full of tricks that every programmer has and isn't talking about, even here. Also, that bag of tricks doesn't necessarily apply to any other chess program except his own. So discussing it is really a no brainer and wise! Stuart
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.