Author: Bob Durrett
Date: 09:23:21 12/06/02
Go up one level in this thread
On December 06, 2002 at 11:51:32, Uri Blass wrote: >On December 06, 2002 at 10:54:32, Bob Durrett wrote: > >>On December 06, 2002 at 10:27:44, Uri Blass wrote: >> >>>On December 06, 2002 at 09:58:06, Bob Durrett wrote: >>> >>>>On December 06, 2002 at 09:34:16, Uri Blass wrote: >>>> >>>><snp> >>>> >>>>>Improvement in search rules can help also in non tactical positions. >>>>>Searching deeper helps also when there is no combination. >>>>> >>>>>Uri >>>> >>>>Please realize that I am currently a user and not a programmer. To be more >>>>accurate, I am an obsolete archaic Fortran programmer, sort of. Hence the words >>>>"Ignorant Question." >>>> >>>>There have been many "tricks," such as alpha/beta, discussed here. The >>>>*primary* purpose of these tricks seems to be to avoid evaluation of unimportant >>>>portions of the tree of moves [with root at the position of the last move >>>>played]. >>>> >>>>But, to the best of my recollection, this is the first time I have seen the >>>>words "search rules" used here at CCC. What are they, and what are they good >>>>for? >>>> >>>>Bob D. >>> >>>I use these words not because of reading them somewhere but because it is >>>logical to use these words. >>> >>>search rules tell the program where to extend and where to prune and they can >>>help not only when winning material is possible. >>> >>>Uri >> >>OK. Thanks. >> >>I guess, in the normal course of events, that "search rules" come before "search >>algorithms." First you must set up a set of "rules," and then try to come up >>with an algorithm which implements and inforces these rules in some reasonable >>manner. Only after that is done does the programmer turn his/her attention to >>coding. >> >>Please let me know if I'm still confused. : ) >> >>Bob D. > >Maybe there is a misunderstanding. Well, it is probably me. >I do not think about search rules as something different than search >algorithms. I have always thought of an "algorithm" as being like a "procedure." First you accomplish this, then that, etc., with conditional decisions sometimes. My impression was that your "search rules" were something that came before the process of deciding which rule to implement first, which second, etc. I saw an algorithm as being more like a finished product, ready for coding. > >The process of finding better algorithms is also not like you describe. >First you need to discover better search algorithms and you need good ideas how >to think about that problem. Well, I guess the *first* thing that must happen is to get an "idea." Sort of like "a gleam in the father's eye" which occurs before conception. : ) > >Uri
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.