Author: Tord Romstad
Date: 11:31:13 10/08/03
Go up one level in this thread
On October 08, 2003 at 08:42:53, Uri Blass wrote: >On October 08, 2003 at 07:58:46, Tord Romstad wrote: > >> >>"Stupid" is perhaps not the best word, but I am not a programmer, and I lack >>the required skills to make bitboards work well (or so I think -- I have >>not given it a try so far). > >I do not think that it is very relevant if you have a job as a programmer. Of course not, and that was not really what I meant. I was thinking about skills, knowledge and mindset rather than profession. >I do not think that most other people do not try to teach their program the same >information that you try to teach your program so comparison with other >who use bitboards is meaningless. Perhaps. Who knows. >Other people usually do not try to write big evaluation like you. > >I also see that you are very fast in implementing ideas and only few hours after >somebody post idea about the botvinik extension you already implement it with >good results. Implementing new ideas quickly is easier when you keep things simple, you know. :-) >>I think chess programming is simple enough that it is possible to get quite >>far by using only very simple and straightforward techniques, and that is >>the route I am following. > >I doubt if that is the route you follow It is, at least with regard to low-level stuff like data structures, move generation and things like that. I don't mind doing complicated things as long as I do not have to make my hands dirty with low-level details. >and I do not think that MTD is a very simple and straightforward technique. It is no harder to implement than most other alpha-beta variants, I think. The only problem is that you need a working transposition table before you write the search, unlike PVS, where you can write the search first and add the transposition table later. Assuming a well-working transposition table, MTD is remarkably simple and elegant. > I do not claim that this is superior to any other >>approaches, but for laymen like me it is the easiest way to achieve a >>working and reasonably strong engine without spending too much time and >>effort. > >The easiest way to achieve a working and reasonably strong engine without >spending too much time and effort does not seem to write a big evaluation >function. It depends of your definition of "reasonably strong", I guess. To me, a master-level program is good enough, and my results at ICC indicate that I'm already there. >It also does not seem to be your target otherwise you could easily do your >engine better by doing it more materialistic. If maximum strength were my goal, you would probably be right, yes. 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.