Author: Robert Hyatt
Date: 06:08:34 08/06/99
Go up one level in this thread
On August 06, 1999 at 03:42:51, Dennis Breuker wrote: >On August 05, 1999 at 13:03:07, Eugene Nalimov wrote: > >>On August 05, 1999 at 05:23:05, Dennis Breuker wrote: >> >>>I haven't read CCC for a long time, and saw some reference to >>>Nalimov's EGTB. My question is: what is so special about Nalimov's EGTB? >>>Is it different from the "normal" ones like Edwards' and Thompson's? >>> >>>Dennis Breuker (Dennis.Breuker@nospam.capgemini.nl) >> >>(1) I generated *full* set of 3+2 TBs; generator (hopefully) can handle all 4+1, >>too, if somebody wants them. > >What do you mean exactly by 3+2 and 4+1? 3 pieces and 2 pawns, 4 pieces and 1 >pawn? > these files have 5 pieces max, including two kings. 3 + 2 means one side has 3 pieces (including the king) while the other side has only two. 4 + 1 means one side has a king and three pieces, the other side has just a king. We didn't do these files, because they are simple wins anyway,simple enough to be able to search the results... >>(2) En passant is handled correctly. >>(3) They are 2 times less than Edwards' when uncompressed, 8 times less when >>compressed; and compressed format (unlike Thopson's) allows fast probing. > >Wow, this is great! > >>(4) TB format and probing code allows TB probing "in the tree", not only at the >>root. >>(5) Source of probing code is available, so you can easily incorporate it into >>your chess program. Some programmers had done that in ~2 hours. >>(6) You can easily download generator itself, as well as all 3+3 TBs, from >>ftp://ftp.cis.uab.edu/pub/hyatt. > >I will take a look at this in due time. > >>(7) A lot of programs already support them, and more will be released this fall. > >More? What can we expect? > >>Eugene > >Thanks for the rapid response, >Dennis (Dennis.Breuker@nospam.capgemini.nl)
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.