Author: Miguel A. Ballicora
Date: 07:32:32 01/08/01
Go up one level in this thread
On January 08, 2001 at 08:01:06, Severi Salminen wrote: >>>My program (it does not have a name yet) beat TSCP too! at the end of the >>>message there is the game which looks interesting. >>> >> >>Forgot to say, game was played on a AMD-K6-II 400 mhz both programs >>run in Winboard, game in 60 minutes. > >Congratulations, doesn't it feel good! It was me who "first" beat TSCP. One Thanks! >question (I'm not familiar with terminology): what is "Hash table for refutation >and transposition"? Is this the normal hash table in all programs (not in mine >yet). I understand the "transposition", but what is "refutation"? My sources of information to call it in that way is chapter I read in the "Enciclopaedia of Artificial Intelligence", may be it is old terminology. That is the only thing written in hardcopy I read. The rest of my sources were/are r.g.c.c and CCC. "transposition" means that if I find a position that has been analyzed before, I return the score found previously without searching. If I find the position but the depth stored is not enough to return without searching, the table it is used to find "refutations" (or killer), using the suggested move first (in my case, before even generating moves). I think everybody does both. According to this article, old programs did only "refutation" since you need less memory. You don't have to store a score, depth and the hashkey could be smaller. >And have you tried your program's evaluation without mobility factor? It should >make it faster and maybe without affecting the precision of your eval(). Try it >and if possible make 200 or more games (verison with mobility vs. version >without) and report here at CCC. It would be very interesting to see results! My case could be even worse than Jose. If his program become stupid, my become headless. Mobility, for now, it is almots all the evaluation. I don't have piece-square tables. The good thing about bitboards is that mobility is very easy to program (the hard part, the bitboards, has been done already). I believe that it is not very slow. I think that my bottleneck is in the movegeneration/makemove/unmakemove. At one point I will have to do a serious profiling and rewriting. Maybe I will have to redesign the way I update the bitboards. Miguel > >Severi
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.